Custom Query (619 matches)

Filters
 
Or
 
  
 
Columns

Show under each result:


Ticket Created Resolution Component Reporter Owner Modified
#2807 2 years ago fixed admin Sebastian Huber Amar Takhar 18 months ago
Summary

rtems-docs repository is not known to trac

#3098 21 months ago fixed admin Chris Johns Amar Takhar 6 months ago
Summary

Add new RTEMS repos to github.

Description

Please add:

  1. rtems-docs.git
  2. rtems-release.git

to our github repos.

#3135 20 months ago fixed admin Sebastian Huber amar@… 18 months ago
Summary

Devel mailing list doesn't work and Git push impossible due to disk full

Description

I got this:

git push
Counting objects: 18, done.
Delta compression using up to 12 threads.
Compressing objects: 100% (17/17), done.
Writing objects: 100% (18/18), 1.68 KiB | 0 bytes/s, done.
Total 18 (delta 16), reused 0 (delta 0)
remote: error: file write error (No space left on device)
remote: fatal: unable to write sha1 file
error: remote unpack failed: unpack-objects abnormal exit
To ssh://dispatch.rtems.org/data/git/rtems.git
 ! [remote rejected]       upstream -> master (unpacker error)
error: failed to push some refs to 'ssh://sebh@dispatch.rtems.org/data/git/rtems.git'

We have on dispatch.rtems.org:

Filesystem        Size    Used   Avail Capacity  Mounted on
/dev/gpt/root0     88G     82G   -600M   101%    / 
#3203 18 months ago fixed admin Amar Takhar Amar Takhar 6 months ago
Summary

Upgrade trac to fix numerous problems.

Description

There are a ton of issues going on with trac that need to be resolved. The two major ones are:

  • The ticket commenter emails people who aren't a trac user. This may require a custom modification.
  • The always_email setting is taking things too literally and always sending emails even if it shouldn't.
  • The Git plug-in consistently spins, floods the jail with processes then the site dies.
  • Frequent, strange and random crashes.
    • This one is not a huge deal since it's just that request process users won't even notice when this happens.

Upgrading trac is a weeklong project usually I will start preparing for it and update here.

If anyone has any feature requests now is the time to do it!

#3227 18 months ago worksforme admin Joel Sherrill Chris Johns 17 months ago
Summary

sb-check fails on Msys2 64-bit

Description

There must be a recent change to msys2 which is breaking things. I installed the 64-bit version from https://msys2.github.io/ per the instructions at https://docs.rtems.org/branches/master/user/hosts/index.html#microsoft-windows

$ ./source-builder/sb-check
error: no hosts defaults found; please add

After adding some prints, I learned this:

$ ./source-builder/sb-check
posix
made it
MSYS_NT-10.0
error: no hosts defaults found; please add

I filled in options.py and windows.py to recognize this as MSYS2. I was then able to run sb-check. But it wasn't happy. Apparently the pacman command in the User's Guide is missing some packages based on newer versions:

$ ./source-builder/sb-check
posix
MSYS_NT-10.0
RTEMS Source Builder - Check, 5 (8b30eb3f440a modified)
error: exe: not found: (__ar) ar
error: exe: not found: (__as) as
error: exe: not found: (__cc) x86_64-w64-mingw32-gcc
error: exe: not found: (__cxx) x86_64-w64-mingw32-g++
error: exe: not found: (__ld) ld
error: exe: not found: (__nm) nm
error: exe: not found: (__objcopy) objcopy
error: exe: not found: (__objdump) objdump
error: exe: not found: (__ranlib) ranlib
Environment is not correctly set up

I installed binutils explcitly with pacman and then sb-check is complaining about gcc. I did a find to locate the gcc's installed:

$ find / -name "*gcc.*"
/home/jrs007/.ssh/id_rsa_gcc.pub
/mingw64/bin/gcc.exe
/mingw64/bin/x86_64-w64-mingw32-gcc.exe
/mingw64/lib/gcc/x86_64-w64-mingw32/6.2.0/include/stdint-gcc.h
/mingw64/lib/gcc/x86_64-w64-mingw32/6.2.0/libgcc.a
/mingw64/share/info/gcc.info.gz
/mingw64/share/man/man1/gcc.1.gz
/usr/share/vim/vim80/compiler/gcc.vim
find: failed to read file names from file system at or below ‘/’: No such file or directory

jrs007@JRS-OAR-Laptop MINGW64 /c/opt/rtems/rsb/source-builder
$ /mingw64/bin/x86_64-w64-mingw32-gcc.exe --version
x86_64-w64-mingw32-gcc.exe (Rev2, Built by MSYS2 project) 6.2.0
Copyright (C) 2016 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
#3238 17 months ago fixed admin Chris Johns Amar Takhar 6 months ago
Summary

Git push to Trac with more than one commit does not update tickets.

Description

The git push to trac hook does not queue or handle a number of commits in a push. As a result updates to tickets can be missed.

#3386 13 months ago invalid admin Chris Johns Amar Takhar 6 months ago
Summary

Trac's git changeset browsing is suspect.

Description

It is critical this interface works because we have moved to Trac for release notes and the release notes contain links to the changesets because we reference the tickets in the commits.

Some requests work:

  1. 900c40730dbee34cd7a6f1c03c80896951bf1b9c/rtems
  2. d8de6b9dbe4ab1ef375ecce55e8bfb1028c5dd13/rtems
  3. 9704efb4ec088a472842cbc9bc46392685ebc806/rtems

and others do not:

  1. 2afb22b7e1ebcbe40373ff7e0efae7d207c655a9/rtems

Notes:

  • items 2. and 3. are either side of the changeset a. in the commit history of RTEMS.
  • Clicking on 3. and then the Next Changeset link also fails.
#3415 12 months ago fixed admin Joel Sherrill Chris Johns 6 months ago
Summary

Add examples and tests as components

Description

It seems as if we should have tests and examples as components.

#3421 12 months ago wontfix admin Joel Sherrill chrisj@… 6 months ago
Summary

New Trac components for Coverage and Trace

Description

Both coverage and tracing are large enough areas that lumping them into tools or other random categories makes work on them harder to trac.

Please add coverage and tracing. Coverage could be a subcategory of tools.T Tracing could be a standalone component. It has target and tool components.

#3423 12 months ago duplicate admin Joel Sherrill 12 months ago
Summary

examples-v2: m68k/powerpc BSPs undefined reference to _Thread_Life_action_handler

Description

fat_ramdisk is failing to link on multiple m68k and powerpc BSPs. Errors below the list of BSPs

m68k-av5282 m68k-gen68340 m68k-gen68360 m68k-gen68360_040 m68k-mcf5206elite m68k-mcf52235 m68k-mcf5225x m68k-mcf5235 m68k-mcf5329 m68k-mrm332 m68k-pgh360 m68k-uC5282 powerpc-mpc8260ads powerpc-qemuppc powerpc-qoriq_e6500_64 powerpc-ss555

[20/20] Processing rtrace: build/m68k-rtems5-av5282/filesystem/fat_ramdisk/init.c.4.o build/m68k-rtems5-av5282/filesystem/fat_ramdisk/fs-root-tar.c.4.o -> build/m68k-rtems5-av5282/filesystem/fat_ramdisk/fat_ramdisk.texe
/home/joel/rtems-work/tools/5/bin/m68k-rtems5-gcc: /tmp/ccIRjaaa.o: In function `__wrap__Thread_Life_action_handler':
/home/joel/rtems-work/tools/5/bin/m68k-rtems5-gcc: /tmp/cckrhaaa.c:1248: undefined reference to `_Thread_Life_action_handler'
/home/joel/rtems-work/tools/5/bin/m68k-rtems5-gcc: collect2: error: ld returned 1 exit status
error: linking: Linker error
#3555 6 months ago fixed admin Amar Takhar Amar Takhar 6 months ago
Summary

IRC bots need to be registered to join #rtems

Description

Due to the spam on Freenode only registered users can join #rtems. The bots both need accounts now.

#3557 6 months ago fixed admin Amar Takhar Amar Takhar 6 months ago
Summary

Test ticket

Description

Using this as a test ticket to test out my fix.

#3558 6 months ago fixed admin Amar Takhar Amar Takhar 6 months ago
Summary

Update TracSpamFilter?

Description

Updated to the latest Trac Spam Filter and upgraded captcha to v2 to avoid any errors. This was reported a while back and should fix any issues.

There are more please open a new ticket.

#3559 6 months ago fixed admin Amar Takhar Amar Takhar 6 months ago
Summary

Fix NavAdd? plugin.

Description

I had no idea but this had gotten removed in the last upgrade I've re-added it.

This makes a few changes to the navigation:

  • "New Ticket" now goes to /wiki/NewTicket
  • There is a new button "New Ticket (direct)" in the upper right for those who want to directly go to creating a ticket.
  • "My Tickets" used to go to a query but now goes to the new wiki:MyTickets page.

These changes existed years ago when NavAdd? was working I opened this ticket in case anyone has complaints about it coming back if not I will close it in a few days.

#3560 6 months ago fixed admin Amar Takhar Amar Takhar 6 months ago
Summary

Fix FlexibleAssignTo?

Description

When we first went to trac we had restrictions on the 'owner' to developers only. When track was upgraded this broke completely but all the code to handle this was already in place.

I took the time to fix it today so we have dropdowns again.

I've created this ticket to see if anyone has an issue with this should we keep it? Drop it? I know it's been years but it was our original choice.

See any ticket the 'reassign to' and on a new ticket the 'assign to' is now a dropdown. These are based on trac permissions so we can always add more if we need it but it really should be restricted to having a project member be the owner so we can ensure tickets are closed and sorted properly.

#3561 6 months ago fixed admin Amar Takhar Amar Takhar 5 months ago
Summary

Migrate to CommitTicketUpdater?

Description

The old script was ancient and outdated. I've now killed it off and moved to the internal system described here:

https://trac.edgewall.org/wiki/CommitTicketUpdater

This should handle all scenarios if it does not please let me know. I will leave this ticket open for a week or so.

#3579 6 months ago fixed admin Chris Johns Chris Johns <chrisj@…> 6 months ago
Summary

testsuite's rtems-test-check.py python version support

Description

This command used in the testsuite needs to find a suitable python or the build system needs to find it and invoke it with that python.

#3620 5 months ago fixed admin Sebastian Huber Amar Takhar 4 months ago
Summary

CommitTicketUpdater? does not process commits in order

Description

The new CommitTicketUpdater? does not process commits in order. For example see:

https://devel.rtems.org/ticket/3598#comment:40 https://devel.rtems.org/ticket/3598#comment:41

Compare with Git commit order:

https://git.rtems.org/rtems/log/?id=eaa5ea84eaf1b3dab72d7a7a6578f0dc59e55396&qt=range&q=1947449a5d6f01a44ccc61eda3e78ef7e06da952..5fc727fe77a632f9df38161a8474007dab020608

#3649 5 months ago fixed admin Joel Sherrill Amar Takhar 2 months ago
Summary

Error with IRC anouncing in examples-v2 commits.

Description
remote: 1: mail vc@rtems.org 
remote: 2: update github 
remote: 4: IRC 
remote: usage: <file with rev-list> <repo name without .git> 
remote: 5: Buildbot
To ssh://joel@dispatch.rtems.org/data/git/examples-v2.git
   ced6542..276a025  am -> master 
#2385 4 years ago fixed arch/arm Chris Johns Sebastian Huber 18 months ago
Summary

Warning from commit "bsps/arm: Do not use ARM_ARCH_7A"

Description

This change https://git.rtems.org/rtems/commit/?h=4.11&id=d0733bb8 generate a warning in user code. The warning is:

.../arm-errata.h:45:1: warning: 'in line' is not at beginning of declaration [-Wold-style-declaration]
 static bool inline arm_errata_is_applicable_processor_errata_764369(void)
 ^
#2449 3 years ago fixed arch/arm Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove arm/gba BSP

Description

Remove the arm/gba BSP per the instructions at ​​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: This BSP is for very old Nintendo hardware and required the use of either a simulator or hard to obtain programmable game cartridge. Nintendo was aggressive in shutting down resellers of those cartridges. There is no real console input and it is hard to automate testing. This was a useful BSP when there were few ARM BSPs but with the Pi, Beagle, etc. these days are long past.

#2450 3 years ago fixed arch/arm Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove arm/nds

Description

Remove the arm/nds BSP per the instructions at ​​​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: This BSP is for very old Nintendo hardware and required the use of either a simulator or hard to obtain programmable game cartridge. Nintendo was aggressive in shutting down resellers of those cartridges. There is no real console input and it is hard to automate testing. This was a useful BSP when there were few ARM BSPs but with the Pi, Beagle, etc. these days are long past.

#2451 3 years ago fixed arch/arm Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove arm/gp32 BSP

Description

Remove the arm/gp32 BSP per the instructions at ​​​​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: This BSP is for an open source alternative to the Gameboy Advance introduced in 2001. Wikipedia notes that 30K units were sold but it has been unavailable since 2007. This was a useful BSP when there were few ARM BSPs and the openness was interesting but with the Pi, Beagle, etc. these days are long past.

#2529 3 years ago fixed arch/arm Sebastian Huber Sebastian Huber 18 months ago
Summary

BSP for the Atmel SAM V71/V70/E70/S70 chip platform

Description

http://www.atmel.com/products/microcontrollers/arm/sam-v-mcus.aspx

#2560 3 years ago fixed arch/arm Sebastian Huber 18 months ago
Summary

smdk2410 is broken due to gp32 removal

Description

The smdk2410 BSPs use files of the removed gp32 BSP.

[f2a228b2cb5ce376c56ae8d767084b92f2822af0/rtems]

#2576 3 years ago fixed arch/arm Joel Sherrill Joel Sherrill <joel@…> 7 months ago
Summary

arm/lpc176x: linker script update (add KEEP() sections)

Description

This BSP's linker script does not include KEEP() directives and thus cannot have per-function and per-data element section support enabled.

The preferred solution is to convert the BSP to use a shared base linker script. The acceptable solution is to add the proper KEEP directives to the existing linker script(s).

Shared linker scripts for the arm, m68k, and sparc have the proper KEEP sections and can serve as examples.

#2685 3 years ago fixed arch/arm David Binderman Sebastian Huber 18 months ago
Summary

c/src/lib/libbsp/arm/atsam/network/if_atsam.c:409: possible bad if statement

Description

rtems/c/src/lib/libbsp/arm/atsam/network/if_atsam.c:409]: (style) Redundant condition: If 'phy <= 0', the comparison 'phy <= 31' is always true.

Source code is

if ((phy <= 0) && (phy <= 31)) {

/*

  • invalid phy number */

Maybe better code

if ((phy <= 0)
(phy >= 31)) {

/*

  • invalid phy number */
#2775 3 years ago fixed arch/arm Chris Johns Chris Johns <chrisj@…> 18 months ago
Summary

ARM CP15 arm_cp15_set_translation_table_entries fails if TTB in read-only memory

Description

If the TTB is held in the text section and the section is set to read-only, and cached when booting no section change happen at run time because the table cannot be written too to change. The table cannot be changed unless the MMU is disabled.

I suggest the MMU be disabled, the table updated and then the MMU enabled.

Note, the issue only shows up on real hardware, qemu does not complain.

#2816 2 years ago fixed arch/arm Joel Sherrill Sebastian Huber 18 months ago
Summary

Many ARM BSPs have Static Assert

Description

Hi

With the recent commits, many BSPs on the master do not build. They error out with this static assert:

In file included from ../../../../cpukit/../../../gumstix/lib/include/rtems/score/types.h:23:0,

from ../../../../cpukit/../../../gumstix/lib/include/rtems/score/cpu.h:32, from ../../../../cpukit/../../../gumstix/lib/include/rtems/system.h:23, from ../../../../../../../../rtems/c/src/../../cpukit/score/cpu/arm/cpu.c:29:

../../../../cpukit/../../../gumstix/lib/include/rtems/score/basedefs.h:241:5: error: static assertion failed: "ARM_CONTEXT_CONTROL_ISR_DISPATCH_DISABLE"

_Static_assert(cond, # msg)

../../../../../../../../rtems/c/src/../../cpukit/score/cpu/arm/cpu.c:54:3: note: in expansion of macro 'RTEMS_STATIC_ASSERT'

RTEMS_STATIC_ASSERT(

The list of BSPs is:

arm1136jfs arm1136js arm7tdmi arm920 csb336 csb337 csb637 edb7312 gumstix kit637_v6 lpc2362 lpc23xx_tli800 lpc24xx_ea lpc24xx_ncs_ram lpc24xx_ncs_rom_ext lpc24xx_ncs_rom_int lpc24xx_plx800_ram lpc24xx_plx800_rom_int lpc32xx_mzx lpc32xx_mzx_stage_1 lpc32xx_mzx_stage_2 lpc32xx_phycore raspberrypi rtl22xx rtl22xx_t smdk2410

#2824 2 years ago fixed arch/arm Joel Sherrill Sebastian Huber 18 months ago
Summary

arm/lpc23xx_tli800 no longer links tar01

Description

Normally I would add the test to those skipped but I thought this deserved a second look. Should this BSP be able to run this test? I thought it was a fairly beefy board.

arm-rtems4.12-gcc -B../../../../../lpc23xx_tli800/lib/ -specs bsp_specs -qrtems -mcpu=arm7tdmi-s -mthumb -Os -g -ffunction-sections -fdata-sections -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -Wl,--gc-sections -mcpu=arm7tdmi-s -mthumb -DHAVE_XZ=1 -o tar01.exe init.o test_cat.o initial_filesystem_tar.o initial_filesystem_tar_gz.o initial_filesystem_tar_xz.o -lrtemscpu -lz /data/home/joel/rtems-4.11-work/tools/4.12/bin/../lib/gcc/arm-rtems4.12/6.2.1/../../../../arm-rtems4.12/bin/ld: tar01.exe section .data' will not fit in region ROM_INT' /data/home/joel/rtems-4.11-work/tools/4.12/bin/../lib/gcc/arm-rtems4.12/6.2.1/../../../../arm-rtems4.12/bin/ld: region `ROM_INT' overflowed by 36 bytes collect2: error: ld returned 1 exit status gmake[7]: * [tar01.exe] Error 1

#2826 2 years ago fixed arch/arm Chris Johns 18 months ago
Summary

arm_cp15_get_translation_table_base_control_register warning.

Description

arm_cp15_get_translation_table_base_control_register in c/src/lib/libcpu/arm/shared/include/arm-cp15.h returns a pointer however ttb_cr is not a pointer"

../../cpukit/../../../xilinx_zynq_zedboard/lib/include/libcpu/arm-cp15.h: In function 'arm_cp15_get_translation_table_base_control_register':
../../cpukit/../../../xilinx_zynq_zedboard/lib/include/libcpu/arm-cp15.h:401:10: warning: return makes pointer from integer without a cast [-Wint-conversion]
   return ttb_cr;
          ^~~~~~
#2868 2 years ago fixed arch/arm David Binderman Gedare Bloom <gedare@…> 18 months ago
Summary

src/c/src/lib/libbsp/arm/smdk2410/smc/smc.c: 3 * pointless local variables ?

Description

[src/c/src/lib/libbsp/arm/smdk2410/smc/smc.c:235]: (style) Variable 'cnt1' is modified but its new value is never used. [src/c/src/lib/libbsp/arm/smdk2410/smc/smc.c:243]: (style) Variable 'cnt2' is modified but its new value is never used. [src/c/src/lib/libbsp/arm/smdk2410/smc/smc.c:246]: (style) Variable 'cnt3' is modified but its new value is never used.

$ egrep "cnt1|cnt2|cnt3" src/c/src/lib/libbsp/arm/smdk2410/smc/smc.c

uint32_t pblock, i, j, lblock, zone, count, cnt1, cnt2, cnt3; cnt1 = 0; cnt2 = 0; cnt3 = 0;

cnt1++;

cnt2++; cnt3++;

$

Maybe someone left some debug code in ?

#2873 2 years ago fixed arch/arm David Binderman Gedare Bloom <gedare@…> 18 months ago
Summary

src/c/src/lib/libbsp/arm/raspberrypi/i2c/i2c.c:320: defective error checking ?

Description

src/c/src/lib/libbsp/arm/raspberrypi/i2c/i2c.c:320]: (style) Checking if unsigned variable 'rv' is less than zero.

Source code is

rv = rpi_i2c_setup_transfer(bus);

if ( rv < 0 ) {

but

uint32_t rv = 0;

and

static int rpi_i2c_setup_transfer(rpi_i2c_bus *bus)

Suggest put return value into an int local variable, then sanity check it, then assign it to rv.

#2883 2 years ago fixed arch/arm David Binderman Pavel Pisa 18 months ago
Summary

src/c/src/lib/libbsp/arm/tms570/console/tms570-sci.c:248: strange expression ?

Description

src/c/src/lib/libbsp/arm/tms570/console/tms570-sci.c:248]: (style) Same expression on both sides of '|'.

Source code is

uint32_t flr_tx_ready = TMS570_SCI_FLR_TX_EMPTY | TMS570_SCI_FLR_TX_EMPTY;

#2954 2 years ago fixed arch/arm Sebastian Huber Sebastian Huber 18 months ago
Summary

ARM: Optimize context switch

Description

Set CPU_ENABLE_ROBUST_THREAD_DISPATCH to TRUE. In this case the interrupts are always enabled during a context switch even after interrupt processing (see #2751). Remove the CPSR from the context control since it contains only volatile bits.

#3007 2 years ago fixed arch/arm munster joel.sherrill@… 18 months ago
Summary

ARM caching issues

Description

There are two problems with the caching on ARM:

  • In cases where the buffer is not aligned to line boundary at the beginning or the end, the invalidate operation would lose modifications done on the adjacent data. This applies to both L1 and L2 caches.
  • The L2C-310 cache management operations use excessive locking. According to manual, the used operations (Clean Line by PA, Clean and Invalidate Line by PA, Cache Sync) are atomic and do not require locking.

I have attached the proposed patch.

#3011 2 years ago worksforme arch/arm Arturo Pérez Gedare 18 months ago
Summary

Error compiling xilinx_zynq_zedboard.

Description

I encountered an error compiling the xilinx_zynq_zedboard BSP. I am using a built of the RSB that I compiled in December. With that built of the RSB I could built this BSP several times until I did a git pull of the RTEMS repo two weeks ago. Today I updated my repos of the RTEMS and RSB sources, I rebuilt the RSB and I tried to built again the xilinx_zynq_zedboard BSP, encountering the same error:

gmake[6]: * No rule to make target posix/include/sys/mman.h', needed by ../cpukit/../../../xilinx_zynq_zedboard/lib/include/sys/mman.h'. Stop.

#3100 21 months ago fixed arch/arm Chris Johns Chris Johns 18 months ago
Summary

Add Xilinx AXI I2C driver

Description

Add a Xilinx AXI I2C driver.

#3152 19 months ago wontfix arch/arm Chris Johns Chris Johns 6 months ago
Summary

Beaglebone Black crashes on u-boot master build.

Description

The crash with a Linux type image and no FDT is:

] ## Booting kernel from Legacy Image at 82000000 ...
]    Image Name:   RTEMS
]    Image Type:   ARM Linux Kernel Image (gzip compressed)
]    Data Size:    60886 Bytes = 59.5 KiB
]    Load Address: 80000000
]    Entry Point:  80000000
]    Verifying Checksum ... OK
]    Uncompressing Kernel Image ... OK
] 
] Starting kernel ...
] 
] data abort
] 
]     MAYBE you should read doc/README.arm-unaligned-accesses
] 
] pc : [<8000010c>]        lr : [<800000ac>]
] sp : 80101000  ip : 0000000c   fp : 9f35ac28
] r10: 9f3ad0f4  r9 : 00000000   r8 : 9f238f40
] r7 : 00000000  r6 : 80000100   r5 : 00000e05  r4 : 60000193
] r3 : 9f238fe0  r2 : 80000100   r1 : 00000e05  r0 : 60000193
] Flags: nzcv  IRQs off  FIQs on  Mode SVC_32
] Resetting CPU ...

and the code is:

BSP_START_TEXT_SECTION void bsp_start_hook_0(void)
{
}
80000104:       e12fff1e        bx      lr

80000108 <bsp_start_hook_1>:
BSP_START_TEXT_SECTION static inline arm_a8core_start_set_vector_base(void)
{
  /*
   * Do not use bsp_vector_table_begin == 0, since this will get optimized away.
  */
  if (bsp_vector_table_end != bsp_vector_table_size) {
80000108:       e3002040        movw    r2, #64 ; 0x40
8000010c:       e3003040        movw    r3, #64 ; 0x40
80000110:       e3482000        movt    r2, #32768      ; 0x8000
80000114:       e3403000        movt    r3, #0
80000118:       e1520003        cmp     r2, r3
#3173 19 months ago fixed arch/arm Chris Johns Chris Johns 15 months ago
Summary

XIlinx AXI I2C driver IP race condition causes clock glitch.

Description

The Xilinx AXI I2C IP has a race condition when the PIRQ read FIFO level is reached and the clock is throttling.

#3249 17 months ago fixed arch/arm Joel Sherrill Sebastian Huber 17 months ago
Summary

imx7 does not link getentropy01 test on master

Description

arm-rtems5-gcc -B../../../../../imx7/lib/ -specs bsp_specs -qrtems -march=armv7-a -mthumb -mfpu=neon -mfloat-abi=hard -mtune=cortex-a7 -O2 -g -ffunction-sections -fdata-sections -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -Wl,--gc-sections -march=armv7-a -mthumb -mfpu=neon -mfloat-abi=hard -mtune=cortex-a7 -Wl,--wrap=printf -Wl,--wrap=puts -Wl,--wrap=putchar -o getentropy01.exe init.o init.o: In function `test_getentropy': /home/joel/rtems-work/rtems-testing/rtems/build-arm-imx7-rtems/arm-rtems5/c/imx7/testsuites/libtests/getentropy01/../../../../../../../rtems/c/src/../../testsuites/libtests/getentropy01/init.c:57: undefined reference to `getentropy' /home/joel/rtems-work/rtems-testing/rtems/build-arm-imx7-rtems/arm-rtems5/c/imx7/testsuites/libtests/getentropy01/../../../../../../../rtems/c/src/../../testsuites/libtests/getentropy01/init.c:59: undefined reference to `getentropy'

#3290 15 months ago fixed arch/arm Sebastian Huber Sebastian Huber 15 months ago
Summary

Add device tree support to Altera/Intel? Cyclone V BSP

#3305 14 months ago fixed arch/arm Joel Sherrill Joel Sherrill 14 months ago
Summary

Add paravirtualization support to ARM

Description

The Arm port does not currently have paravirtualization support.

#3352 13 months ago fixed arch/arm Joel Sherrill Sebastian Huber 13 months ago
Summary

Warning in all lpc176x variants

Description

bsps/arm/lpc176x/include/bsp.h defines OPERATION_COUNT in an attempt to override the autoconf generated constant. This conflicts and results in this warning:

/home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/arm/lpc176x/include/bsp.h:42:0: warning: "OPERATION_COUNT" redefined

I understand why this is lowered by the BSP but the mechanism used is not good. And if the include file order is different between tests, you could get the BSP value or the autoconf generated value based on the order.

This warning needs to be fixed and a safer mechanism for a BSP to override OPERATION_COUNT defined.

My first suggestion is to use BSP_OPERATION_COUNT and add logic to one of the common test .h files to undef OPERATION_COUNT and redefine it to BSP_OPERATION_COUNT if it is defined.

A safer option might be to change the name of the autoconf generated variable to OPERATION_COUNT_DEFAULT and rely on logic in a common test support .h to define OPERATION_COUNT to OPERATION_COUNT_DEFAULT or BSP_OPERATION_DEFAULT.

#3502 8 months ago fixed arch/arm Joel Sherrill Sebastian Huber 8 months ago
Summary

PL111_LCD_CONTROL_LCD_BPP_16 Redefined

Description

The constant PL111_LCD_CONTROL_LCD_BPP_16 is defined twice in the file bsps/arm/include/bsp/arm-pl111-regs.h:

#define PL111_LCD_CONTROL_LCD_BPP_16 0x04U #define PL111_LCD_CONTROL_LCD_BPP_24 0x05U #define PL111_LCD_CONTROL_LCD_BPP_16 0x06U #define PL111_LCD_CONTROL_LCD_BPP_12 0x07U

Given the context, I am guessing the first one should be BPP_32 but since Sebastian added the file, I am assuming he has docs and can answer this question for sure.

#3525 7 months ago fixed arch/arm Sebastian Huber Sebastian Huber 7 months ago
Summary

Add MMC/SDCard support for i.MX 7Dual BSP

Description

Port device drivers from FreeBSD for i.MX 7Dual uSDHC module.

#3667 4 months ago fixed arch/arm Sebastian Huber Sebastian Huber 3 months ago
Summary

Support data cache disable on ARMv7-AR

#3673 3 months ago fixed arch/arm Joel Sherrill Sebastian Huber 3 months ago
Summary

xilinx_zynq_a9_qemu - fails to link psxconfig01

Description

This is with POSIX disabled.

/home/joel/rtems-work/rtems-testing/rtems/build-arm-xilinx_zynq_a9_qemu-rtems/arm-rtems5/c/xilinx_zynq_a9_qemu/testsuites/psxtests/../../../../../../rtems/c/src/../../testsuites/psxtests/psxconfig01/init.c:499: undefined reference to `timer_create' collect2: error: ld returned 1 exit status

#3674 3 months ago fixed arch/arm Joel Sherrill Sebastian Huber 3 months ago
Summary

Raspberry Pi Fails to Build

Description

../../../../../../../../rtems/c/src/lib/libbsp/arm/raspberrypi/../../../../../../bsps/arm/shared/cache/cache-v7ar-disable-data.S: Assembler messages: ../../../../../../../../rtems/c/src/lib/libbsp/arm/raspberrypi/../../../../../../bsps/arm/shared/cache/cache-v7ar-disable-data.S:47: Error: selected processor does not support `dmb' in ARM mode ../../../../../../../../rtems/c/src/lib/libbsp/arm/raspberrypi/../../../../../../bsps/arm/shared/cache/cache-v7ar-disable-data.S:53: Error: selected processor does not support `isb' in ARM mode ../../../../../../../../rtems/c/src/lib/libbsp/arm/raspberrypi/../../../../../../bsps/arm/shared/cache/cache-v7ar-disable-data.S:77: Error: selected processor does not support `isb' in ARM mode ../../../../../../../../rtems/c/src/lib/libbsp/arm/raspberrypi/../../../../../../bsps/arm/shared/cache/cache-v7ar-disable-data.S:85: Error: invalid constant (3ff) after fixup ../../../../../../../../rtems/c/src/lib/libbsp/arm/raspberrypi/../../../../../../bsps/arm/shared/cache/cache-v7ar-disable-data.S:92: Error: invalid constant (7fff) after fixup gmake[6]: * [cache-v7ar-disable-data.o] Error 1

#3699 2 months ago fixed arch/arm Kinsey Moore Sebastian Huber 2 months ago
Summary

Wrong system register specified for ARM virtual timer value retrieval

Description

In arm_cp15_get_counter_pl1_virtual_timer_value() in cpukit/score/cpu/arm/include/libcpu/arm-cp15.h, the system register specified by "p15, 0, %[val], c14, c2, 0" is actually the system register for the physical timer value. This should be "p15, 0, %[val], c14, c3, 0" for the virtual timer value as used in the setter.

#3724 6 weeks ago fixed arch/arm Sebastian Huber Sebastian Huber 5 weeks ago
Summary

bsp/lpc24xx: Convert SSP driver to Linux API

#3725 6 weeks ago fixed arch/arm Sebastian Huber Sebastian Huber 5 weeks ago
Summary

bsp/lpc24xx: Convert I2C driver to Linux API

Description
#3346 13 months ago fixed arch/bfin Joel Sherrill Joel Sherrill 13 months ago
Summary

bf533.h

Description

TIMER_STATUS, TIMER< DISABLE, and TIMER_ENABLE are defined in bf52x.h and in bf533.h. Disable second definition in full bf533 register set list and add a sanity check to ensure it stays the same.

In file included from /home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/bfin/TLL6527M/include/bsp.h:28:0,

from ../../../../../rtems/c/src/libchip/display/disp_hcms29xx.c:26:

/home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/bfin/include/bf52x.h:43:0: warning: "TIMER_STATUS" redefined

#define TIMER_STATUS 0xffc00648

#3201 18 months ago fixed arch/epiphany Joel Sherrill Joel Sherrill <joel@…> 18 months ago
Summary

epiphany tools checksum error

Description

I assume this is a side-effect of the recent checksum changes. If that's the case, it just needs to be updated. Otherwise, it is a more serious error.

download: (full) https://github.com/adapteva/epiphany-gcc/archive/f7051762470c42ce7f01baa7edeb113d51c7dd72.zip -> sources/f7051762470c42ce7f01baa7edeb113d51c7dd72.zip download: https://github.com/adapteva/epiphany-gcc/archive/f7051762470c42ce7f01baa7edeb113d51c7dd72.zip -> sources/f7051762470c42ce7f01baa7edeb113d51c7dd72.zip

redirect: https://codeload.github.com/adapteva/epiphany-gcc/zip/f7051762470c42ce7f01baa7edeb113d51c7dd72 redirect: https://codeload.github.com/adapteva/epiphany-gcc/zip/f7051762470c42ce7f01baa7edeb113d51c7dd72

checksums: f7051762470c42ce7f01baa7edeb113d51c7dd72.zip: 4d911e7bff4f1827dd7712669d20e4a1bf02806df0fae113ff0e7d13466bef2e => 2b2034fd12f2fd5108205ade66400c175ede8cef8141a38ae03fc78bf2d65325 warning: checksum error: f7051762470c42ce7f01baa7edeb113d51c7dd72.zip error: checksum failure file: sources/f7051762470c42ce7f01baa7edeb113d51c7dd72.zip

#2468 3 years ago fixed arch/i386 Joel Sherrill Sebastian Huber 18 months ago
Summary

Add Thread Local Storage (TLS) support on x86

Description

The x86 is one of the architectures which does not support Thread Local Storage. Each architecture requires target architecture specific support to provide this standard language feature.

Details on the implementation by the compiler may be found at http://wiki.osdev.org/Thread_Local_Storage.

Based on this information, I think a segment register needs to be added to the thread context and some hooks to the TLS implemented.

Architecture information on TLS implementation should be added to the CPU Supplement document as this is part of the ABI and context switch.

As part of effort, the documentation for the general procedure of adding target specific TLS support should be added to the porting guide or reviewed.

#2634 3 years ago fixed arch/i386 Joel Sherrill Pavel Pisa <ppisa@…> 18 months ago
Summary

New warning in pc386 VESA driver

Description

Pavel.. can you look into this?

./../../../../../../../rtems/c/src/lib/libbsp/i386/pc386/console/fb_vesa_rm.c: In function 'find_mode_using_EDID': ../../../../../../../../rtems/c/src/lib/libbsp/i386/pc386/console/fb_vesa_rm.c:502:13: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing]

if (*(uint16_t*)&edid.STI[index] == EDID_STI_DescriptorUnused) ~

#2663 3 years ago wontfix arch/i386 Joel Sherrill Sebastian Huber 18 months ago
Summary

pc386 BSP has complex dependencies

Description

In 4.11, the minimum executable did not include open() and close() because the methods rtems_libio_post_driver() and rtems_libio_exit() were not included in the executable. On the master, these two methods are showing up in minimum and pulling in these methods.

The dependency chain used to be if the console driver was installed, we needed to open and close stdin, stdout, and stderr. Now even without the console configured these are included.

FWIW the minimum size looks pretty good on the master for sis. Fixing this would likely drop it at least another 5%.

#3172 19 months ago fixed arch/i386 Chris Johns Chris Johns 18 months ago
Summary

i386 PC BSP does not reset when bsp_reset is called.

Description

Removal of the Edison support removed the standard PC reset using the keyboard controller rather then the specific Edison support.

#3349 13 months ago fixed arch/i386 Joel Sherrill Joel Sherrill 13 months ago
Summary

pc386 edid.h invalid macro names

Description

Minus sign not underbar in macro name.

-#define DVS_HDMI-a 0x2 -#define DVS_HDMI-b 0x3 +#define DVS_HDMI_a 0x2 +#define DVS_HDMI_b 0x3

#3410 12 months ago fixed arch/i386 Sebastian Huber Sebastian Huber 12 months ago
Summary

Remove bin2boot program used by i386 BSPs

Description

For which boot loader is this? Can it be removed?

The sources have no copyright information.

#3672 3 months ago fixed arch/i386 Joel Sherrill 3 months ago
Summary

No i386 BSP can link all tests after cache manager changes

Description

/data/home/joel/rtems-work/tools/5/bin/../lib/gcc/i386-rtems5/7.4.0/../../../../i386-rtems5/bin/ld: ./../../lib/libbsp/i386/pc386/librtemsbsp.a(cache.o): in function `rtems_cache_invalidate_entire_instruction': /home/joel/rtems-work/rtems-testing/rtems/build-i386-pc486-rtems/i386-rtems5/c/pc486/lib/libbsp/i386/pc386/../../../../../../../../rtems/c/src/lib/libbsp/i386/pc386/../../../../../../bsps/i386/shared/cache/../../../shared/cache/cacheimpl.h:350: undefined reference to `_CPU_cache_invalidate_entire_instruction'

#3599 6 months ago fixed arch/m32c Sebastian Huber Sebastian Huber 5 months ago
Summary

Remove m32c architecture port

Description

The m32c architecture port is incomplete, e.g. important features such as interrupt support are missing. It never run on real hardware. The tools are out dated and unmaintained. There are no known users:

https://lists.rtems.org/pipermail/users/2018-January/031991.html

#2444 3 years ago fixed arch/m68k Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove m68k/mvme136 BSP

Description

Remove the m68k/mvme136 BSP per the instructions at https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: Although it is sad to see the BSP for the board that RTEMS was developed for be removed, this board was current in 1988-9. It has only 1MB RAM, 2 UARTS, and no NIC. It is unlikely to be available and without a NIC, isn't that useful.

#2445 3 years ago fixed arch/m68k Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove m68k/sim68000 BSP

Description

Remove the m68k/sim68000 BSP per the instructions at ​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: This is a BSP for a simulator named BSVC (http://www4.ncsu.edu/~bwmott/bsvc/) that was never under a truly free license and has not been updated in a decade. Although a decent tool, it was extremely slow.

#2513 3 years ago fixed arch/m68k Joel Sherrill 18 months ago
Summary

Remove m68k/idp BSP

Description

Remove the m68k/idp BSP per the instructions at ​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

#2546 3 years ago fixed arch/m68k Joel Sherrill Joel Sherrill 18 months ago
Summary

Obsolete idp BSP

Description

Obsolete and remove the m68k/idp BSP before the 4.12 release.

#3344 13 months ago fixed arch/m68k Joel Sherrill Joel Sherrill 13 months ago
Summary

mcf5272/mcf5272.h Timer3 Duplicate Definition

Description

This .h file uses the same macro names for two blocks of INT macros. My assumption given that the second looks to be a different INT, is that it should not be INT3 again but INT3.

--- a/bsps/m68k/include/mcf5272/mcf5272.h +++ b/bsps/m68k/include/mcf5272/mcf5272.h @@ -88,9 +88,9 @@

#define MCF5272_ICR1_INT3_PI (bit(23)) #define MCF5272_ICR1_INT3_IPL(x) ((x) << 20) #define MCF5272_ICR1_INT3_MASK ((7) << 20)

-#define MCF5272_ICR1_INT3_PI (bit(19)) -#define MCF5272_ICR1_INT3_IPL(x) ((x) << 16) -#define MCF5272_ICR1_INT3_MASK ((7) << 16) +#define MCF5272_ICR1_INT4_PI (bit(19)) +#define MCF5272_ICR1_INT4_IPL(x) ((x) << 16) +#define MCF5272_ICR1_INT4_MASK ((7) << 16)

#2448 3 years ago fixed arch/mips Joel Sherrill 18 months ago
Summary

Remove mips/mongoose BSP

Description

Remove the m32r/m32rsim BSP per the instructions at ​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: This is a radiation hardened MIPS R3000 CPU that has only been used by a few missions. After discussions with various NASA and commercial engineers, we have learned that it is no longer considered an option for new missions and has not an option for a considerable length of time. The missions still underway (including New Horizons) are locked down on very old versions of their development infrastructure including hosts.

#3202 18 months ago fixed arch/or1k Joel Sherrill Joel Sherrill <joel@…> 18 months ago
Summary

or1k tools build error

Description

I assume this is a side-effect of recent checksum changes. Otherwise, there is a serious problem.

ownload: (full) https://git.rtems.org/rtems-tools/plain/tools/4.12/gdb/gdb-7.11-sis-leon2-leon3.diff -> patches/gdb-7.11-sis-leon2-leon3.diff download: https://git.rtems.org/rtems-tools/plain/tools/4.12/gdb/gdb-7.11-sis-leon2-leon3.diff -> patches/gdb-7.11-sis-leon2-leon3.diff checksums: gdb-7.11-sis-leon2-leon3.diff: 0b8b2a23c7d1592315fe0130188f457c80f8b1e26645535bed091a5e0671682dc44a1987d00e6939a1b1c562c7579404db43183e666c29c2b479446aa61ca4f6 => 4c44afec9c00a45b9322d787da3796f3294f207ddae9fe9faab3327b6991ac75 warning: checksum error: gdb-7.11-sis-leon2-leon3.diff error: checksum failure file: patches/gdb-7.11-sis-leon2-leon3.diff

#3602 5 months ago fixed arch/or1k Sebastian Huber Sebastian Huber 5 months ago
Summary

Update or1k tool chain to use the upstream GCC

#2306 4 years ago fixed arch/powerpc Joel Sherrill 18 months ago
Summary

powerpc/mvme5500/vectors/exceptionhandler.c uses task variables

Description

I am addressing almost all uses of deprecated methods. They are mostly calls to rtems_clock_get() which can be easily corrected or test code which will be removed when the deprecated feature is removed.

This BSP however has what appears to be a unique feature -- the ability for a thread to add a unique exception fault handler. My inclination is to rip this out but I am not doing it now. I am just turning off deprecated warnings for the file.

#2457 3 years ago fixed arch/powerpc Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove powerpc/ep1a BSP

Description

Remove the powerpc/ep1a BSP per the instructions at ​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: AFAIK this BSP was only used on a program supported by OAR. That program is no longer in active development and is completely frozen. If that situation changes, the BSP can be resurrected. It was introduced 10 years ago and has not has a modification other than general maintenance in the last four years.

#2458 3 years ago fixed arch/powerpc Joel Sherrill Ralph Holmes <ralph@…> 18 months ago
Summary

Remove powerpc/score603e BSP

Description

Remove the powerpc/score603e BSP per the instructions at ​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

Rationale: AFAIK this BSP was only used on a program supported by OAR. That program is no longer in active development and is completely frozen. If that situation changes, the BSP can be resurrected. It was introduced in 1999 and has not has a modification other than general maintenance in the last six years.

#2503 3 years ago fixed arch/powerpc Aun-Ali Zaidi joel.sherrill@… 18 months ago
Summary

mvme5500 BSP: Exception Handler uses deprecated Notepads.

Description

The MVME5500 BSP uses Notepads in its exception handler and #2493 removes them. This is obviously not portable and requires a rewrite.

#2553 3 years ago fixed arch/powerpc Nick Withers Sebastian Huber 18 months ago
Summary

[mvme3100] boot_card() broken by 37030e38

Description

Backtrace with 37030e38* (with a fatal exception handler installed):

#0  fatal_extension (the_source=INTERNAL_ERROR_CORE, is_internal=true, the_error=23) at init.c:425
#1  0x000a6d18 in _User_extensions_Iterate (arg=arg@entry=0x1ab6348, visitor=0xa6c70 <_User_extensions_Fatal_visitor>) at ../../../../../../../rtems/c/src/../../cpukit/score/src/userextiterate.c:155
#2  0x000a1c90 in _User_extensions_Fatal (error=23, is_internal=true, source=INTERNAL_ERROR_CORE) at ../../cpukit/../../../mvme3100/lib/include/rtems/score/userextimpl.h:254
#3  _Terminate (the_source=the_source@entry=INTERNAL_ERROR_CORE, is_internal=is_internal@entry=true, the_error=the_error@entry=23) at ../../../../../../../rtems/c/src/../../cpukit/score/src/interr.c:44
#4  0x0007ad40 in RTEMS_Malloc_Initialize (areas=areas@entry=0x1ab6398, area_count=area_count@entry=1, extend=extend@entry=0x0 <preload>)
    at ../../../../../../../rtems/c/src/../../cpukit/libcsupport/src/malloc_initialize.c:53
#5  0x0005cf84 in bsp_work_area_initialize_default (area_size=<optimized out>, area_begin=<optimized out>) at ../../../../../.././mvme3100/lib/include/bsp/bootcard.h:183
#6  bsp_work_area_initialize () at ../../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mvme3100/../../powerpc/shared/startup/bspgetworkarea.c:23
#7  0x0005cec4 in boot_card (cmdline=<optimized out>) at ../../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mvme3100/../../shared/bootcard.c:80
#8  0x00003294 in __rtems_entry_point () at ../../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mvme3100/start/start.S:89

If I reverse the changes to c/src/lib/libbsp/shared/bootcard.c it works again.

* With RSB e3b9fb68 and the following RTEMS patches:

#2800 2 years ago fixed arch/powerpc Joel Sherrill Sebastian Huber 18 months ago
Summary

qoriq variants failing to build

Description

The following qoriq variants fail to build:

powerpc-qoriq_core_0 powerpc-qoriq_core_1 powerpc-qoriq_p1020rdb

All fail with this error:

../../../../../.././qoriq_core_0/lib/include/bsp/qoriq.h:407:3: error: conflicting types for 'qoriq_gpio'
 } qoriq_gpio;
   ^~~~~~~~~~
../../../../../.././qoriq_core_0/lib/include/bsp/qoriq.h:184:3: note: previous declaration of 'qoriq_gpio' was here
 } qoriq_gpio;
   ^~~~~~~~~~

Found during full build sweep based on this commit:

commit bb9f09f34c9bdcf4d2631a1fd317bcefd8426efb
Author: Sebastian Huber <sebastian.huber@embedded-brains.de>
Date:   Mon Oct 31 13:07:34 2016 +0100

    posix: Fix timer interval
    
    Do not overwrite timer interval with initial interval in
    _POSIX_Timer_Insert().
    
    Close #2798.

#2805 2 years ago fixed arch/powerpc Sebastian Huber Sebastian Huber 18 months ago
Summary

Use SPRG0 on PowerPC for current per-CPU control (SMP only)

Description

Add _CPU_Get_current_per_CPU_control() on SMP configurations as an optimization for PowerPC. Use SPRG0 for the current per-CPU control. This reduces the code size a bit and is slightly faster in some benchmarks.

#2819 2 years ago fixed arch/powerpc Joel Sherrill Sebastian Huber 18 months ago
Summary

powerpc-ss555 does not compile on master

Description

Recent changes broke this configuration:

gmake[6]: Entering directory `/data/home/joel/rtems-4.11-work/rtems-testing/rtems/build-powerpc-ss555-rtems/powerpc-rtems4.12/c/ss555/testsuites/samples/hello' powerpc-rtems4.12-gcc -B../../../../../ss555/lib/ -specs bsp_specs -qrtems -DHAVE_CONFIG_H -I. -I../../../../../../../rtems/c/src/../../testsuites/samples/hello -I.. -mcpu=505 -Dmpc555 -O2 -g -fno-keep-inline-functions -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -MT init.o -MD -MP -MF .deps/init.Tpo -c -o init.o ../../../../../../../rtems/c/src/../../testsuites/samples/hello/init.c In file included from ../../../../../ss555/lib/include/rtems/score/percpu.h:22:0,

from ../../../../../ss555/lib/include/rtems/confdefs.h:32, from ../../../../../../../rtems/c/src/../../testsuites/samples/hello/init.c:51:

../../../../../ss555/lib/include/rtems/score/cpuimpl.h:196:3: error: conflicting types for 'CPU_Interrupt_frame'

} CPU_Interrupt_frame;

In file included from ../../../../../ss555/lib/include/bsp/irq.h:28:0,

from ../../../../../ss555/lib/include/bsp.h:31, from ../../../../../../../rtems/c/src/../../testsuites/samples/hello/init.c:17:

../../../../../ss555/lib/include/libcpu/irq.h:193:3: note: previous declaration of 'CPU_Interrupt_frame' was here

} CPU_Interrupt_frame;

gmake[6]: * [init.o] Error 1 gmake[6]: Target `all' not remade because of errors.

#2874 2 years ago fixed arch/powerpc David Binderman Sebastian Huber <sebastian.huber@…> 18 months ago
Summary

src/c/src/lib/libbsp/powerpc/beatnik/marvell/gt_timer.c: 4 * pointless check ?

Description

[src/c/src/lib/libbsp/powerpc/beatnik/marvell/gt_timer.c:102]: (style) Checking if unsigned variable 'timer' is less than zero. [src/c/src/lib/libbsp/powerpc/beatnik/marvell/gt_timer.c:109]: (style) Checking if unsigned variable 'timer' is less than zero. [src/c/src/lib/libbsp/powerpc/beatnik/marvell/gt_timer.c:117]: (style) Checking if unsigned variable 'timer' is less than zero. [src/c/src/lib/libbsp/powerpc/beatnik/marvell/gt_timer.c:128]: (style) Checking if unsigned variable 'timer' is less than zero.

Parameter "timer" is only ever type uint32_t, so any check < 0 seem pointless.

#3015 2 years ago fixed arch/powerpc phongvanpham joel.sherrill@… 18 months ago
Summary

Add support for IBM PPC 750 chip

Description

Currently MPC750 chip is supported. However, PPC750 (from IBM) is very close to MPC750 except minor differences. Enclosed is the delta to support PPC750.

#3082 21 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 3 months ago
Summary

Add 64-bit support for PowerPC

Description

The QorIQ chips have more than 4GiB of memory available.

#3085 21 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 13 months ago
Summary

Add hypervisor support for QorIQ BSPs

Description

NXP provides a hypervisor (Topaz) for the QorIQ platform.

https://www.xes-inc.com/wp-content/uploads/2016/03/NXP-Embedded-Hypervisor-for-QorIQ-Overview.pdf

#3245 17 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 17 months ago
Summary

Replace BSP_panic() with rtems_panic()

Description

Due to a new rtems_panic() implementation, it is possible to replace the PowerPC-specific BSP_panic() with rtems_panic(). Remove BSP_panic() implementations.

#3246 17 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 17 months ago
Summary

Remove _BSP_Fatal_error()

Description

BSPs can use the bsp_fatal_extension() to provide BSP-specific fatal error handling. There is no need for a _BSP_Fatal_error().

#3255 17 months ago duplicate arch/powerpc Joel Sherrill Sebastian Huber 17 months ago
Summary

Warnings on 64-bit targets

Description

This occurred on the 64 bit PowerPC and SPARC64 targets.

log/powerpc-qoriq_e6500_64.log:../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:57:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] log/powerpc-qoriq_e6500_64.log:../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:64:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] log/powerpc-qoriq_e6500_64.log:../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:488:11: warning: format '%u' expects argument of type 'unsigned int', but argument 5 has type 'long int' [-Wformat=] log/powerpc-qoriq_e6500_64.log:../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:536:13: warning: format '%u' expects argument of type 'unsigned int', but argument 2 has type 'long int' [-Wformat=]

#3268 16 months ago fixed arch/powerpc Chris Johns 15 months ago
Summary

PowerPC BSP include naming mess.

Description

The PowerPC BSP family headers need some refactoring for the RTEMS 5 release. The BSP family relies on the preinstall process to get suitable headers installed to work and removing preinstall exposes this. The specific issue appears with irq.h when building the no-preinstall branch. There is a PowerPC BSP family header and a number of BSPs also have an irq.h which overrides families header. The code has #include <bsp/irq.h> and the header used depends on the include order on the GCC command line. This is fragile for any user. These headers needs to be moved to BSP specific paths, for example #include <mvme3100/irq.h>.

#3270 16 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 15 months ago
Summary

Remove unused support for MPC505

Description

There is some support for MPC505 in libcpu, however, I cannot find a BSP for this code. Remove this apparently dead code.

#3306 14 months ago fixed arch/powerpc Joel Sherrill Joel Sherrill <joel@…> 14 months ago
Summary

Add paravirtualization support to PowerPC

Description

The PowerPC port does not currently have paravirtualization support.

#3307 14 months ago fixed arch/powerpc Joel Sherrill Joel Sherrill 7 months ago
Summary

PowerPC linkcmds.base missing wildcards on some sections

Description

Some sections were missing sections. Wildcards needed to be added.

#3340 14 months ago fixed arch/powerpc Joel Sherrill Sebastian Huber 13 months ago
Summary

gen83xx warning for macros redefined

Description

log/powerpc-hsc_cm01.log:/home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/powerpc/gen83xx/include/bsp/hwreg_vals.h:244:0: warning: "FPGA_START" redefined log/powerpc-hsc_cm01.log:/home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/powerpc/gen83xx/include/bsp/hwreg_vals.h:246:0: warning: "FPGA_SIZE" redefined

Looking at the code, it is pretty clear that the macros are redefined. Unfortunately one of the three has a different value the second time:

======================================== /* fpga BCSR register */ #define FPGA_START 0xF8000000 #define FPGA_SIZE 0x8000 #define FPGA_END (FPGA_START+FPGA_SIZE-1)

/*

  • working values for various registers, used in start/start.S */

/* fpga config 16 MB size */ #define FPGA_CONFIG_START 0xF8000000 #define FPGA_CONFIG_SIZE 0x01000000 /* fpga register 8 MB size */ #define FPGA_REGISTER_START 0xF9000000 #define FPGA_REGISTER_SIZE 0x00800000 /* fpga fifo 8 MB size */ #define FPGA_FIFO_START 0xF9800000 #define FPGA_FIFO_SIZE 0x00800000

#define FPGA_START (FPGA_CONFIG_START) fpga window size 32 MByte #define FPGA_SIZE (0x02000000) #define FPGA_END (FPGA_START+FPGA_SIZE-1)

========================================

#3345 13 months ago fixed arch/powerpc Joel Sherrill Joel Sherrill 13 months ago
Summary

mvme3100 spaces needed around quote in macro definitions in bsp.h

Description

Various BSP_I2c_XXX_DEV_NAME macros have a stray " at the end of the first parameter.

#3348 13 months ago fixed arch/powerpc Joel Sherrill Joel Sherrill 13 months ago
Summary

beatnick:spaces needed around quote in macro definitions in bsp.h

Description

Macros need spaces around ","

#3354 13 months ago fixed arch/powerpc Joel Sherrill Joel Sherrill 13 months ago
Summary

PowerPC BSPs duplicate PAGE_MASK, etc redefinition

Description

The following BSPs:

powerpc-beatnik powerpc-mcp750 powerpc-mtx603e powerpc-mvme2100 powerpc-mvme2307 powerpc-mvme5500 powerpc-qemuprep-altivec powerpc-qemuprep

use bsps/powerpc/include/libcpu/page.h which defines _ALIGN, PAGE_MASK, and PAGE_SIZE. These are defined by <machine/param.h>. I think the solution is to delete the versions in libcpu/page.h. Comments appreciated.

=================== In file included from ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mvme5500/../../powerpc/shared/startup/pgtbl_setup.c:3:0: /home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/powerpc/include/libcpu/page.h:22:0: warning: "PAGE_MASK" redefined

#define PAGE_MASK (~(PAGE_SIZE-1))

In file included from /data/home/joel/rtems-work/tools/5/powerpc-rtems5/include/sys/_cpuset.h:36:0,

from /data/home/joel/rtems-work/tools/5/powerpc-rtems5/include/sys/cpuset.h:45, from /data/home/joel/rtems-work/tools/5/powerpc-rtems5/include/sys/_pthreadtypes.h:24, from /data/home/joel/rtems-work/tools/5/powerpc-rtems5/include/sys/types.h:239, from /data/home/joel/rtems-work/tools/5/powerpc-rtems5/include/sys/time.h:43, from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/score/timestamp.h:43, from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/score/thread.h:36, from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/score/heap.h:22, from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/rtems/types.h:26, from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems.h:31, from ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mvme5500/../../powerpc/shared/startup/pgtbl_setup.c:1:

/data/home/joel/rtems-work/tools/5/powerpc-rtems5/include/machine/param.h:70:0: note: this is the location of the previous definition

#define PAGE_MASK (PAGE_SIZE - 1)

#3411 12 months ago fixed arch/powerpc Joel Sherrill Sebastian Huber 12 months ago
Summary

qemuppc does not install linkcmds.base

Description

examples-v2 fail to compile qemuppc because linkcmds.base is not installed.

They build OK for sparc/erc32. This must be a minor glitch from the build system changes.

#3501 8 months ago fixed arch/powerpc Joel Sherrill Christian Mauderer 8 months ago
Summary

MSR_RI defined multiple places

Description

Two files define the MSR_RI macro. Since one is a register name on the PowerPC, this shows up on 48 PowerPC BSPs. This is one example.

log/powerpc-qoriq_e6500_64.log:../../../../../../rtems/c/src/../../cpukit/dev/serial/sc16is752-regs.h:117:0: warning: "MSR_RI" redefined log/powerpc-qoriq_e6500_64.log: #define MSR_RI (1u << 6) log/powerpc-qoriq_e6500_64.log: #define MSR_RI (1<<1) /* Recoverable Exception */ log/powerpc-qoriq_e6500_64.log:../../../../../../rtems/c/src/../../cpukit/dev/serial/sc16is752-regs.h:117:0: warning: "MSR_RI" redefined log/powerpc-qoriq_e6500_64.log: #define MSR_RI (1u << 6) log/powerpc-qoriq_e6500_64.log: #define MSR_RI (1<<1) /* Recoverable Exception */

#3504 8 months ago fixed arch/powerpc Joel Sherrill Sebastian Huber 8 months ago
Summary

Warning and formatting in bsps/powerpc/mpc55xxevb/dev/dspi.c

Description

This is a printf format warning. Also the file is formatted with tabs and not two spaces.

In file included from ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mpc55xxevb/../../../../../../bsps/powerpc/mpc55xxevb/dev/dspi.c:32:0: ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mpc55xxevb/../../../../../../bsps/powerpc/mpc55xxevb/dev/dspi.c: In function 'mpc55xx_dspi_edma_done': /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/status-checks.h:88:23: warning: format '%x' expects argument of type 'unsigned int', but argument 3 has type 'uint32_t {aka long unsigned int}' [-Wformat=]

RTEMS_SYSLOG_PRINT( "%s: " fmt, func, ##VA_ARGS)

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/status-checks.h:76:15: note: in definition of macro 'RTEMS_SYSLOG_PRINT'

printk( fmt, ##VA_ARGS)

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/status-checks.h:109:3: note: in expansion of macro 'RTEMS_SYSLOG'

RTEMS_SYSLOG( "Error: " fmt, ##VA_ARGS) ~

../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mpc55xxevb/../../../../../../bsps/powerpc/mpc55xxevb/dev/dspi.c:122:3: note: in expansion of macro 'RTEMS_SYSLOG_ERROR'

RTEMS_SYSLOG_ERROR( "eDMA error: 0x%08x\n", error_status); ~

../../../../../../../../rtems/c/src/lib/libbsp/powerpc/mpc55xxevb/../../../../../../bsps/powerpc/mpc55xxevb/dev/dspi.c:122:41: note: format string is defined here

RTEMS_SYSLOG_ERROR( "eDMA error: 0x%08x\n", error_status);

~ %08lx

#3505 8 months ago wontfix arch/powerpc Joel Sherrill Sebastian Huber 6 months ago
Summary

powerpc/virtex redefined warning

Description

This looks like ppc403 and ppc405 are both defined but I am not seeing source of the ppc405 definition. The warning is in this section of code:

#if defined (ppc403) #define exisr 0x040 /* DCR: external interrupt status register */ #define exier 0x042 /* DCR: external interrupt enable register */ #endif /* ppc403 */ #if defined(ppc405) #define exisr 0x0C0 /* DCR: external interrupt status register */ #define exier 0x0C2 /* DCR: external interrupt enable register */ #endif /* ppc405 */

In file included from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/score/percpu.h:25:0,

from ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/virtex/../../../../../../bsps/powerpc/shared/exceptions/ppc_exc_async_normal.S:16:

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/score/cpu/powerpc/include/rtems/asm.h:228:0: warning: "exisr" redefined

#define exisr 0x0C0 /* DCR: external interrupt status register */

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/score/cpu/powerpc/include/rtems/asm.h:224:0: note: this is the location of the previous definition

#define exisr 0x040 /* DCR: external interrupt status register */

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/score/cpu/powerpc/include/rtems/asm.h:229:0: warning: "exier" redefined

#define exier 0x0C2 /* DCR: external interrupt enable register */

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/score/cpu/powerpc/include/rtems/asm.h:225:0: note: this is the location of the previous definition

#define exier 0x042 /* DCR: external interrupt enable register */

#3511 8 months ago duplicate arch/powerpc Joel Sherrill Sebastian Huber 8 months ago
Summary

int/pointer size warnings in powerpc-qoriq_e6500_64

Description

These all look suspiciously like real issues:

$ grep warning log/powerpc-qoriq_e6500_64.log ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/qoriq/../../../../../../bsps/powerpc/qoriq/start/bspstart.c:173:5: warning: passing argument 1 of 'qoriq_initialize_exceptions' makes pointer from integer without a cast [-Wint-conversion] ../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:57:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] ../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:64:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] ../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:488:11: warning: format '%u' expects argument of type 'unsigned int', but argument 5 has type 'long int' [-Wformat=] ../../../../../../rtems/c/src/../../cpukit/libmisc/rtems-fdt/rtems-fdt-shell.c:536:13: warning: format '%u' expects argument of type 'unsigned int', but argument 2 has type 'long int' [-Wformat=] ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/qoriq/../../../../../../bsps/powerpc/shared/exceptions/ppc_exc_alignment.c:28:25: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] ../../../../../../../../rtems/c/src/lib/libbsp/powerpc/qoriq/../../../../../../bsps/powerpc/shared/exceptions/ppc_exc_initialize.c:38:10: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]

#3513 7 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 7 months ago
Summary

Convert tqm8xx console driver to new Termios API

#3549 6 months ago fixed arch/powerpc Sebastian Huber Sebastian Huber 6 months ago
Summary

Obsolete powerpc/virtex BSP

Description

This BSP is quite old (was added 1995), unmaintained and likely without users:

https://lists.rtems.org/pipermail/users/2018-September/032557.html

#3109 20 months ago fixed arch/riscv Sebastian Huber Hesham Almatary 8 months ago
Summary

Add RISC-V support

Description

Add RISC-V 32-bit tool chain to RSB consisting of Binutils, GCC, Newlib and GDB. Add CPU port and a basic simulator BSP.

#3433 11 months ago fixed arch/riscv Sebastian Huber Sebastian Huber 3 months ago
Summary

Add SMP support for RISC-V

Description

The project includes the following tasks:

  • add CPU counter support
  • add context validation code
  • add BSP support for Qemu
  • add support for device tree provided by Qemu
  • fix all unexpected test suite failures running on Qemu
  • add build system support to enable an SMP build
  • add SMP implementation
#3452 10 months ago fixed arch/riscv Sebastian Huber Sebastian Huber 8 months ago
Summary

Update RISC-V tool chain to support standard 64-bit chips

Description

First step is to include this bug fix in Binutils:

https://sourceware.org/bugzilla/show_bug.cgi?id=23244

Second step is a multilib update. Third step is a merge of the riscv32 and riscv64 tool chains into a single riscv tool chain.

#3453 10 months ago fixed arch/riscv Sebastian Huber Sebastian Huber 10 months ago
Summary

Add RISC-V GDB

#3478 9 months ago fixed arch/riscv Joel Sherrill Sebastian Huber 8 months ago
Summary

RISCV BSP Tester Cleanup Needed

Description

rtems-tools currently has the following bsp testing configurations:

$ find . -name "*riscv*ini"
./tester/rtems/rtems-bsps-riscv64.ini
./tester/rtems/testing/bsps/riscv64_generic.ini
./tester/rtems/testing/bsps/riscv_generic.ini
./tester/rtems/rtems-bsps-riscv32.ini

rtems-bsps.ini does not include the riscv.

tester/rtems/rtems-bsps-tiers.ini does not list the riscv

#2325 4 years ago fixed arch/sparc Sebastian Huber Daniel Hellstrom 18 months ago
Summary

Broken console driver infrastructure for SPARC

Description

The stuff in "c/src/lib/libbsp/sparc/shared/uart/cons.c" should get removed and the new Termios device API should be used instead (see also rtems_termios_device_install()).

#2355 4 years ago fixed arch/sparc Sebastian Huber Daniel Hellstrom 17 months ago
Summary

SPARC: Several shared drivers are not SMP ready

Description

Several drivers in c/src/lib/libbsp/sparc/shared/ use interrupt disable/enable for low-level mutual exclusion. This is not enough on SMP configurations.

#2363 4 years ago duplicate arch/sparc Sebastian Huber 15 months ago
Summary

SPARC: Silent FP context corruption possible

Description

On uni-processor configurations the post-switch actions (e.g. signal handlers) and context switch extensions may silently corrupt the floating point context. Set test sptests/spcontext01.

This problem exists for many years and might be working as intended. It is possible to fix this issue using the SPARC_USE_SAFE_FP_SUPPORT option. This is already used for the SMP configurations. The disadavantage is that this disables the deferred floating point support.

#2684 3 years ago fixed arch/sparc David Binderman Sebastian Huber 18 months ago
Summary

rtems/c/src/lib/libbsp/sparc/leon3/clock/ckinit.c:122: duplicate if

Description

rtems/c/src/lib/libbsp/sparc/leon3/clock/ckinit.c:122]: (style) Expression is always false because 'else if' condition matches previous condition at line 116.

Source code is

} else if (state == 1) {

unsigned int ks = 1U << 5;

state = 0;

irqmp_ts->control = ks | s1_s2 | (unsigned int) clkirq;

} else if (state == 1) {

#2726 3 years ago fixed arch/sparc Sebastian Huber Daniel Hellstrom 18 months ago
Summary

grascs.c: Questionable use of binary semaphore

Description

Use a simple binary semaphore or binary semaphore with inherit priority instead.

c/src/lib/libbsp/sparc/shared/ascs/grascs.c-    
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-  /* Create semaphores for blocking ASCS_TC/TM functions */
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-  if(rtems_semaphore_create(rtems_build_name('A','S','C','0'),1,
c/src/lib/libbsp/sparc/shared/ascs/grascs.c:                        (RTEMS_FIFO|RTEMS_BINARY_SEMAPHORE|
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-                         RTEMS_NO_INHERIT_PRIORITY|RTEMS_LOCAL|
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-                         RTEMS_NO_PRIORITY_CEILING), 0,
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-                        &cfg->tcsem1) != RTEMS_SUCCESSFUL) {
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-    goto init_error2;
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-  }
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-  if(rtems_semaphore_create(rtems_build_name('A','S','C','2'),0,
c/src/lib/libbsp/sparc/shared/ascs/grascs.c:                        (RTEMS_FIFO|RTEMS_BINARY_SEMAPHORE|
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-                         RTEMS_NO_INHERIT_PRIORITY|RTEMS_LOCAL|
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-                         RTEMS_NO_PRIORITY_CEILING), 0,
c/src/lib/libbsp/sparc/shared/ascs/grascs.c-                        &cfg->tcsem2) != RTEMS_SUCCESSFUL) {
#2810 2 years ago fixed arch/sparc Joel Sherrill Joel Sherrill 18 months ago
Summary

Remove sparc/sis BSP variant

Description

As discussed in the following thread, the sparc/sis BSP variant is no longer necessary and can be removed.

https://lists.rtems.org/pipermail/devel/2016-November/016383.html

This ticket is to track that removal.

#2820 2 years ago fixed arch/sparc Joel Sherrill Sebastian Huber 18 months ago
Summary

All SPARC64 BSPs do not Build on master

Description

Recent changes broke all builds:

sparc64-rtems4.12-gcc --pipe -DHAVE_CONFIG_H -I../../.. -I../../../../cpukit/../../../usiii/lib/include -mcpu=ultrasparc3 -DUS3 -DSUN4U -g -O2 -ffunction-sections -fdata-sections -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -MT libscorecpu_a-sparc64-exception-frame-print.o -MD -MP -MF .deps/libscorecpu_a-sparc64-exception-frame-print.Tpo -c -o libscorecpu_a-sparc64-exception-frame-print.o test -f 'sparc64-exception-frame-print.c' || echo '../../../../../../../../rtems/c/src/../../cpukit/score/cpu/sparc64/'sparc64-exception-frame-print.c sparc64-rtems4.12-gcc --pipe -DHAVE_CONFIG_H -I../../.. -I../../../../cpukit/../../../usiii/lib/include -mcpu=ultrasparc3 -DUS3 -DSUN4U -g -O2 -ffunction-sections -fdata-sections -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -MT ../no_cpu/libscorecpu_a-cpucounterread.o -MD -MP -MF ../no_cpu/.deps/libscorecpu_a-cpucounterread.Tpo -c -o ../no_cpu/libscorecpu_a-cpucounterread.o test -f '../no_cpu/cpucounterread.c' || echo '../../../../../../../../rtems/c/src/../../cpukit/score/cpu/sparc64/'../no_cpu/cpucounterread.c In file included from ../../../../../../../../rtems/c/src/../../cpukit/score/cpu/sparc64/../no_cpu/cpucounterread.c:15:0: ../../../../cpukit/../../../usiii/lib/include/rtems/score/cpu.h: In function '_CPU_ISR_Is_enabled': ../../../../cpukit/../../../usiii/lib/include/rtems/score/cpu.h:759:12: error: 'psr' undeclared (first use in this function)

return ( psr & SPARC_PSTATE_IE_MASK ) != 0;

../../../../cpukit/../../../usiii/lib/include/rtems/score/cpu.h:759:12: note: each undeclared identifier is reported only once for each function it appears in gmake[7]: * no_cpu/libscorecpu_a-cpucounterread.o Error 1 mv -f .deps/libscorecpu_a-context.Tpo .deps/libscorecpu_a-context.Po In file included from ../../../../cpukit/../../../usiii/lib/include/rtems/system.h:23:0,

#2878 2 years ago fixed arch/sparc David Binderman Daniel Hellstrom 18 months ago
Summary

src/c/src/lib/libbsp/sparc/shared/can/occan.c:1573: broken error checking ?

Description

src/c/src/lib/libbsp/sparc/shared/can/occan.c:1573]: (style) Checking if unsigned variable 'speed=pelican_speed_auto(can)' is less than zero.

Source code is

if ( (speed=pelican_speed_auto(can)) < 0 ){

/* failed */ return RTEMS_IO_ERROR;

}

but

unsigned int speed;

and

static int pelican_speed_auto(occan_priv *priv);

I am not sure which C compiler gets using in rtems, but I do know that gcc compiler flag -Wtype-limits will flag this kind of problem.

#2905 2 years ago fixed arch/sparc Tanu Hari Dixit Daniel Hellstrom 18 months ago
Summary

Merge LEON

Description

Merge LEON

Students: Past, Present, and Potential Students

Status: Some work may be done.

Introduction: Merge Leon RTEMS and mainstream RTEMS merge the Leon DriverManager, samples, and LEON-RTEMS into mainstream RTEMS.

Goal: Concise statement of the overall goal of the project. Refine this initial statement to include: project deliverables (code, docs, testing), required/suggested methodology, standards of quality, possible goal extensions beyond the main objective.

Requirements: List the requirements and level of expertise you estimate are required by the developer tackling this project will have to have: Required level of programming language(s), specific areas of RTEMS or tools, level of familiarity with RTEMS, cross-development, GNU/Linux, etx., development/documentation/testing tools, mathematical/algorithmic background, other desirable skills.

Resources: Current RTEMS developers, papers, etc that may help you in this project.

Acknowledgements

  • who helped and did work

Miscellaneous Sections

As the project progresses, you will need to add build instructions, etc and this page will evolve from a project description into a HOWTO.

References

  • TBD

Other sections: If you have more to say about the project that doesn't fit in the proposed sections of this template, feel free to add other sections at will.

#2924 2 years ago fixed arch/sparc Joel Sherrill Daniel Hellstrom 18 months ago
Summary

Warnings in SPARC BSPs

Description

As of today, the following warnings exist for SPARC BSPs.

log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/console/debugputs.c:43:5: warning: this 'while' clause does not guard... [-Wmisleading-indentation] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/amba/ahbstat.c:156:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/amba/ahbstat.c:156:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/drvmgr/ambapp_bus.c:647:4: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/drvmgr/ambapp_bus.c:647:4: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/drvmgr/leon2_amba_bus.c:168:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/drvmgr/leon2_amba_bus.c:168:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/irq/genirq.c:244:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/irq/genirq.c:244:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/spw/grspw_router.c:213:4: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/spw/grspw_router.c:213:4: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon2.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon2/../../sparc/shared/uart/apbuart.c:574:21: warning: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'apbuart_priv * {aka struct <anonymous> *}' [-Wformat=] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/amba/ahbstat.c:156:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/amba/ahbstat.c:156:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/amba/ahbstat.c:156:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/amba/ahbstat.c:156:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/drvmgr/ambapp_bus.c:647:4: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/drvmgr/ambapp_bus.c:647:4: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/drvmgr/ambapp_bus.c:647:4: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/drvmgr/ambapp_bus.c:647:4: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/irq/genirq.c:244:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/irq/genirq.c:244:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/irq/genirq.c:244:3: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/irq/genirq.c:244:3: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/spw/grspw_router.c:213:4: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/spw/grspw_router.c:213:4: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/spw/grspw_router.c:213:4: warning: implicit declaration of function 'printk' [-Wimplicit-function-declaration] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/spw/grspw_router.c:213:4: warning: nested extern declaration of 'printk' [-Wnested-externs] log/sparc-leon3.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/uart/apbuart.c:574:21: warning: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'apbuart_priv * {aka struct <anonymous> *}' [-Wformat=] log/sparc-ngmp.log:../../../../../../../../rtems/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/uart/apbuart.c:574:21: warning: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'apbuart_priv * {aka struct <anonymous> *}' [-Wformat=]

#3006 2 years ago fixed arch/sparc Chris Johns Daniel Hellstrom 18 months ago
Summary

SPARC LEON3 BSP SMP build is broken.

Description

The rtems-bsp-builder failure output is:

   2 smp-debug sparc/leon3 build:
      configure: /opt/work/chris/rtems/kernel/rtems.git/configure --target\
      =sparc-rtems4.12 --enable-rtemsbsp=leon3 --prefix=/opt/rtems/4.12\
      --enable-debug --enable-smp --enable-tests
      error: c/src/lib/libbsp/sparc/shared/spw/grspw_pkt.c:61:2 error:
      #error SMP mode not compatible with these interrupt lock primitives

The BSP builder command line is:

RTEMS Tools Project - RTEMS Kernel BSP Builder, 4.12.not_released
command: /opt/work/rtems/4.12/bin/rtems-bsp-builder --rtems-\
         tools=/build/rtems/tools/4.12\
         --rtems=/opt/work/chris/rtems/kernel/rtems.git --build=smp-debug\
         --log=x
#3072 22 months ago fixed arch/sparc Sebastian Huber Daniel Hellstrom 18 months ago
Summary

Declaration of global functions in driver source files

Description

There are declarations of global functions in various driver source files, e.g.

c/src/lib/libbsp/sparc/shared/drvmgr/ambapp_bus_grlib.c c/src/lib/libbsp/sparc/shared/drvmgr/ambapp_bus.c

The declaration should move to a header file or a static function should be used.

#3077 21 months ago fixed arch/sparc Sebastian Huber Sebastian Huber 18 months ago
Summary

SPARC: Add lazy floating point context switching

Description

The uniprocessor floating point context switching is unsafe, e.g. it is possible to silently corrupt the floating point context. The SMP floating point switching is safe, however, it doesn't use a deferred switch. Implement lazy floating point switching in uniprocessor configurations. This fixes test case spcontext01.

#3080 21 months ago fixed arch/sparc Sebastian Huber Sebastian Huber 18 months ago
Summary

Infinite loop in SPARC rtems_invalidate_multiple_instruction_lines()

Description

A

#define CPU_INSTRUCTION_CACHE_ALIGNMENT 0

is not a good idea in case the default range functions are used.

#3260 17 months ago fixed arch/sparc Sebastian Huber Sebastian Huber 16 months ago
Summary

libpci depends on BSP-specific header files

Description

The libpci is currently only used on SPARC. It is in cpukit, so BSP-specific header files are not allowed. Unfortunately this is not the case for libpci. However, it seems the the routines depending on BSP_PCI_BIG_ENDIAN are not used:

grep -r 'pci_[ldst]\{2\}_' . ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint16_t pci_ld_le16(volatile uint16_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_le16(volatile uint16_t *addr, uint16_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint32_t pci_ld_le32(volatile uint32_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_le32(volatile uint32_t *addr, uint32_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint16_t pci_ld_be16(volatile uint16_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_be16(volatile uint16_t *addr, uint16_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint32_t pci_ld_be32(volatile uint32_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_be32(volatile uint32_t *addr, uint32_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint16_t pci_ld_le16(volatile uint16_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_le16(volatile uint16_t *addr, uint16_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint32_t pci_ld_le32(volatile uint32_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_le32(volatile uint32_t *addr, uint32_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint16_t pci_ld_be16(volatile uint16_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_be16(volatile uint16_t *addr, uint16_t val) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE uint32_t pci_ld_be32(volatile uint32_t *addr) ./cpukit/libpci/pci/access.h:RTEMS_INLINE_ROUTINE void pci_st_be32(volatile uint32_t *addr, uint32_t val)

Is this dead code and can I remove BSP_PCI_BIG_ENDIAN?

Another issue is the use of a BSP-specific interrupt API in cpukit/libpci/pci/irq.h.

The missing functions should be added to <rtems/irq-extension.h>.

#3341 14 months ago fixed arch/sparc64 Joel Sherrill Gedare 13 months ago
Summary

sparc64: Macro Redefined

Description

log/sparc64-usiii.log:/home/joel/rtems-work/rtems-testing/rtems/rtems/bsps/sparc64/include/arch/stack.h:56:0: warning: "STACK_BIAS" redefined

This is defined in two header files with the same value. Not sure what the proper fix is.

#2442 3 years ago fixed bsps Joel Sherrill joel.sherrill@… 18 months ago
Summary

Remove avrtest BSP

Description

Remove the avr/avrtest BSP per the instructions at https://devel.rtems.org/wiki/Developer/Removing_a_BSP

#2447 3 years ago fixed bsps Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove m32r/m32rsim

Description

Remove the m32r/m32rsim BSP per the instructions at https://devel.rtems.org/wiki/Developer/Removing_a_BSP

#2453 3 years ago fixed bsps Joel Sherrill Aun-Ali Zaidi <admin@…> 18 months ago
Summary

Remove h8300/h8sim BSP

Description

Remove the h8300/h8sim BSP per the instructions at ​https://devel.rtems.org/wiki/Developer/Removing_a_BSP

#2542 3 years ago fixed bsps Joel Sherrill joel.sherrill@… 18 months ago
Summary

Review cxx_iostream size change per function-section changes

Description

It looks like some BSPs with their own linkcmds may have shrunk too much. Norm appears to be 50% while some went to 75-80%. A second look after catching the pattern indicates that the KEEP() section requirements in the linker scripts were not correct and it was missed.

Rather than reverting a bunch of patches, this ticket is to review all function-section patches from this one back in time for cxx_iostream shrinking too much.

commit 6d21c13e5094d490280a941cf0e8333f91f85715 Author: Ralph Holmes <ralph@…> Date: Sat Jan 23 21:15:40 2016 +0000

powerpc/gen5200: Add per-section compilation and linking support.

For the brs5l BSP variant:

#2737 3 years ago fixed bsps Sebastian Huber Sebastian Huber 18 months ago
Summary

Add CLOCK_DRIVER_USE_ONLY_BOOT_PROCESSOR

Description

Add CLOCK_DRIVER_USE_ONLY_BOOT_PROCESSOR clock driver option. If defined, then do the clock tick processing on the boot processor on behalf of all other processors. Currently, this is intended as a workaround for a Qemu shortcoming on ARM.

#2741 3 years ago fixed bsps Joel Sherrill Chris Johns 18 months ago
Summary

New warning from printf plugin changes

Description

../../../../../../../rtems/c/src/lib/libcpu/powerpc/mpc6xx/mmu/pte121.c: In function 'whatPrintf': ../../../../../../../rtems/c/src/lib/libcpu/powerpc/mpc6xx/mmu/pte121.c:189:46: warning: pointer type mismatch in conditional expression

return _Thread_Executing ? (PrintF) printf : printk;

#3014 2 years ago fixed bsps phongvanpham joel.sherrill@… 18 months ago
Summary

interrupt vector indexing is assuming BSP_INTERRUPT_VECTOR_MIN = 0 for this code.

Description

Someone implement bsp_interrupt_handler_index() forgot to update this delta in rtems\c\src\lib\libbsp\shared\src\irq-generic.c:bsp_interrupt_allocate_handler_index(). See attachment.

#3016 2 years ago fixed bsps phongvanpham joel.sherrill@… 18 months ago
Summary

missing a couple register names + a #ifndef ASM around serial.h inclusion

Description

In rtems\c\src\libchip\serial\ns16550_p.h, need to add a couple register and #ifndef around serial.h

#3071 22 months ago fixed bsps Sebastian Huber Sebastian Huber 18 months ago
Summary

Create an interrupt server for every processor in the system

Description

Create an interrupt server for every processor in the system with a one-to-one thread processor affinity. This allows load balancing for interrupt processing. Add support routines to customize the setup after initialization.

#3090 21 months ago fixed bsps Sebastian Huber Sebastian Huber 15 months ago
Summary

Add BSP for i.MX 7

#3122 20 months ago fixed bsps Sebastian Huber Sebastian Huber 11 days ago
Summary

Simplify and unify BSP_output_char

Description

The BSP_output_char should output a char and not mingle with high level processing, e.g. '\n' to '\r\n' translation. Move this translation to rtems_putc(). Remove it from all the BSP_output_char implementations.

#3139 19 months ago fixed bsps Sebastian Huber Sebastian Huber 18 months ago
Summary

Remove old ISR parameter from Clock_driver_support_install_isr() and make it optional

Description

The old ISR is not used by the clock driver shell.

#3219 18 months ago fixed bsps Chris Johns Chris Johns 7 months ago
Summary

Zynq BSP missing linker option --gc-sections

Description

This Zynq BSP is missing this option.

#3318 14 months ago fixed bsps mfletcher Sebastian Huber 14 months ago
Summary

Improve INTERNAL_ERROR_THREAD_EXITTED to show the id and thread name

Description

It might be more helpful i the case of a thread exit to output some information about that thread to make tracking it down simpler.

This example works ok.

static void thread_exitted_print_info(rtems_tcb *tcb) {

printf("Thread exited: %s (id %d)\n", tcb->Object.name, tcb->Object.id)

}

/* In your configuration: */ #define CONFIGURE_INITIAL_EXTENSIONS \

{ .thread_exitted = thread_exitted_print_info }

#3728 4 weeks ago fixed bsps Chris Johns Chris Johns 3 weeks ago
Summary

Set small data seciton to max size for mvme5500 and motorola_powerpc BSPs

Description

These are large memory targets that can support libdl. Make the small data memory the maximum size.

#2407 4 years ago fixed build Sebastian Huber Joel Sherrill 18 months ago
Summary

Enable function and data sections

Description

In order to reduce the size of executables it is beneficial to put all global functions and data into separate sections. This enables the linker to perform a garbage collection which removes all items not directly referenced. The following steps are necessary:

  1. Modify the build system to use the following compiler and linker flags:

CFLAGS += -ffunction-sections -fdata-sections LDFLAGS += -Wl,--gc-sections

  1. Review all linker command files and ensure that linker sets and global constructor sections are not affected by the garbage collection (e.g. use the KEEP() directive of GNU ld).
#2641 3 years ago fixed build aurelio Chris Johns 13 months ago
Summary

configure: enable-rtemsbsp doesn't warn if bsp does not exist

Description

When running configure with an incorrect bsp name the script does not cause an error. You can ever run make without getting any warning message.

The script should check the name of the bsp and continue only if it is a valid bsp. On the other hand if the bsp name given by the user is invalid the script should prompt a message.

#2701 3 years ago fixed build printk Amar Takhar 18 months ago
Summary

Rename asm file with .S(upper case) ext. name

Description

The are some asm file with .s ext. name, .s and .S is different for gnu as, the pre processed produce .s file from .S. In a word, .S can use #define .s can not. KBuild clean .s files when make clean. I have submit a patch to devel, but blocked. Too big patch.

#2951 2 years ago fixed build alexgerbor chrisj@… 18 months ago
Summary

Error path in rtems-gcc-6.3.0-newlib-2.5.0.20170228-1.cfg

Description

newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff in path https://git.rtems.org/rtems-tools/plain/tools/4.12/newlib/arm/ not found

In rtems-gcc-6.3.0-newlib-2.5.0.20170228-1.cfg there error: %patch add newlib %{rtems_newlib_patches}/arm/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff

Log from execute /source-builder/sb-set-builder:

download: (full) https://git.rtems.org/rtems-tools/plain/tools/4.12/newlib/arm/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff -> patches/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff download: https://git.rtems.org/rtems-tools/plain/tools/4.12/newlib/arm/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff -> patches/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff download: no ssl context download: https://git.rtems.org/rtems-tools/plain/tools/4.12/newlib/arm/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff: error: HTTP Error 404: Not found error: downloading https://git.rtems.org/rtems-tools/plain/tools/4.12/newlib/arm/newlib-ARM-Optimize-IEEE-754-sqrt-implementation.diff: all paths have failed, giving up

#3079 21 months ago fixed build Sebastian Huber Sebastian Huber <sebastian.huber@…> 18 months ago
Summary

Ada tests do not build

Description

A "make" in the top level build directory does not build the Ada tests (used to work some weeks ago). A "make" in ./sparc-rtems4.12/c/erc32/ada-tests for example works.

The configuration step seems to work:

configure: configuring in ada-tests
configure: running /bin/sh '../../../../../rtems/c/src/ada-tests/configure'  '--prefix=/home/joel/rtems-4.11-work/bsp-install' '--host=sparc-rtems4.12' '--build=x86_64-pc-linux-gnu' '--target=sparc-rtems4.12' '--enable-smp' '--disable-profiling' '--disable-multiprocessing' '--enable-rtems-debug' '--enable-cxx' '--disable-rdbg' '--enable-maintainer-mode' '--enable-tests' '--enable-networking' '--enable-posix' '--disable-itron' '--disable-deprecated' '--enable-ada' '--enable-expada' 'SIMSPARC_FAST_IDLE=1' '--with-target-subdir=sparc-rtems4.12' '--exec-prefix=/home/joel/rtems-4.11-work/bsp-install/sparc-rtems4.12' '--includedir=/home/joel/rtems-4.11-work/bsp-install/sparc-rtems4.12/include' 'build_alias=x86_64-pc-linux-gnu' 'host_alias=sparc-rtems4.12' 'target_alias=sparc-rtems4.12' '--with-project-root=../../' '--with-project-top=../../' 'RTEMS_BSP=erc32' 'RTEMS_CPU_MODEL=erc32' 'RTEMS_BSP_FAMILY=erc32' 'CFLAGS=-mcpu=cypress -O2 -g -ffunction-sections -fdata-sections -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs' '--enable-rtems-root=../'    '--enable-project-root=../../../erc32'    '--with-project-top=../../../'    '--enable-rtemsbsp=erc32'    --cache-file=/dev/null --srcdir=../../../../../rtems/c/src/ada-tests
configure: loading site script /usr/share/site/x86_64-unknown-linux-gnu
checking for gmake... gmake

However:

gmake[4]: Entering directory '/scratch/git-rtems-testing/rtems/build-sparc-erc32-rtems/sparc-rtems4.12/c/erc32/ada-tests'
gmake[4]: Nothing to be done for 'all-am'.
#3083 21 months ago fixed build Joel Sherrill Chris Johns 13 months ago
Summary

parallel make not working

Description

As reported on the mailing list, parallel make on the master is broken.

#3084 21 months ago fixed build Joel Sherrill Chris Johns 18 months ago
Summary

Makefile recipe override warning has returned

Description

The makefile overrides recipe warnings had disappeared with Chris' rework to improve parallelism. Unfortunately, one Makefile has had the warning return. To reproduce, complete a build with all tests enabled, then just type make >/dev/null at the top of the build tree

[joel@rtbf64c rtems-work]$ ./build_bsp sparc erc32 Using rtems for RTEMS source

real 5m4.247s user 5m58.188s sys 1m34.959s 0 [joel@rtbf64c rtems-work]$ cd b-erc32/ [joel@rtbf64c b-erc32]$ make >/dev/null Makefile:653: warning: overriding recipe for target `spprofiling01' Makefile:653: warning: ignoring old recipe for target `spprofiling01'

#3180 19 months ago fixed build Joel Sherrill Sebastian Huber 6 months ago
Summary

ar warning: u' modifier ignored since D' is the default (see `U')

Description

Remove this warning

sparc-rtems4.12-ar: `u' modifier ignored since `D' is the default (see `U')
#3181 19 months ago fixed build Joel Sherrill Joel Sherrill <joel@…> 18 months ago
Summary

Various cc1plus warnings for "valid for C/ObjC but not for C++"

Description

Fix these

cc1plus: warning: command line option '-Wmissing-prototypes' is valid for C/ObjC but not for C++
cc1plus: warning: command line option '-Wimplicit-function-declaration' is valid for C/ObjC but not for C++
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
cc1plus: warning: command line option '-Wnested-externs' is valid for C/ObjC but not for C++

#3187 19 months ago fixed build Joel Sherrill Sebastian Huber 18 months ago
Summary

smptests/Makefile.am Issues

Description

There appear to be two issues in smptests/Makefile.am:

  1. smppsxaffinity0[12] are listed twice. This causes a warning. I am unsure if they should always be built or only when POSIX is enabled.
  1. smppsxmutex01 is listed in the "HAS_POSIX" section. Aren't POSIX mutexes always on now so this test should not be inside the conditional?
#3247 17 months ago fixed build Sebastian Huber Sebastian Huber 17 months ago
Summary

Remove BSP-specific defaults for RTEMS_BSP_CLEANUP_OPTIONS()

Description

Remove BSP-specific defaults for RTEMS_BSP_CLEANUP_OPTIONS() to simplify the BSP configuration and documentation. Change default to:

BSP_PRESS_KEY_FOR_RESET=0 BSP_RESET_BOARD_AT_EXIT=1 BSP_PRINT_EXCEPTION_CONTEXT=1

#3248 17 months ago fixed build Sebastian Huber Sebastian Huber 17 months ago
Summary

Add BSP_VERBOSE_FATAL_EXTENSION to RTEMS_BSP_CLEANUP_OPTIONS

Description

Add BSP_VERBOSE_FATAL_EXTENSION to RTEMS_BSP_CLEANUP_OPTIONS to optionally print the RTEMS version, the fatal source and the fatal code in the shared bsp_fatal_extension().

#3254 17 months ago fixed build Sebastian Huber 4 months ago
Summary

Reorganize header files to avoid "make preinstall"

#3285 15 months ago fixed build Sebastian Huber Sebastian Huber 9 months ago
Summary

Reorganize BSP source directory

Description

Now, that all BSP header files are in

  • bsps/include
  • bsps/@RTEMS_CPU@/include
  • bsps/@RTEMS_CPU@/@RTEMS_BSP_FAMILY@/include

we should also move the BSP sources to this new directory tree. How do we want to organize the BSP sources in bsps/@RTEMS_CPU@/@RTEMS_BSP_FAMILY@?

  • include (this is already there, see #3254)
  • config
    • somebsp.cfg
  • start (everything required to run a minimal application without devices)
    • start.S
    • bspstart.c
    • bspsmp.c
    • linkcmds
  • cache (everything for the cache controller support)
  • irq (everything for the interrupt controller support)
  • console (everything for the console driver)
  • clock (everything for the clock driver
  • i2c (everything for the I2C driver)
  • spi (everything for the SPI driver)
  • net (legacy network stack drivers)
  • mpci (RTEMS_MULTIPROCESSING support)
  • rtc (everything for the RTC driver)
  • ata (everything for the ATA driver)
  • contrib (import of external sources)
    • The layout of external sources should be used as is if possible.
#3328 14 months ago fixed build Amaan Cheval 14 months ago
Summary

bootstrap uses non-POSIX compliant echo -e

Description

On certain shells, the "-e" option is not supported, and causes echo to output the flag along with the quoted text.

-> % sh
$ echo -e "foo bar"
-e foo bar
$

This varies by shell, and is not even consistent between sh or bash.

It was introduced while removing the make preinstall stage here, and may still work on most shells, though it didn't for me on sh on Ubuntu 16.04 LTS (4.4.0-78-generic x86_64 GNU/Linux) - as far as I can tell, this bug hasn't made it to any releases yet, so just fixing it on master should be enough.

A patch is attached.

Reference to the POSIX standard which confirms that -n is the only argument supported.

http://pubs.opengroup.org/onlinepubs/9699919799/utilities/echo.html

Link to POSIX for printf(1):

http://pubs.opengroup.org/onlinepubs/9699919799/utilities/printf.html

#3375 13 months ago fixed build Sebastian Huber Sebastian Huber 6 months ago
Summary

Remove command line pre-processor defines

Description

Command line defines defined by the build system make it difficult get a consistent view of the sources from other entities, e.g. static code analysis, code editors and reviews.

Command line defines are currently used here:

c/src/lib/libbsp/mips/hurricane/Makefile.am:libbsp_a_CPPFLAGS = $(AM_CPPFLAGS) -DRM52XX
c/src/lib/libbsp/mips/rbtx4938/Makefile.am:libbsp_a_CPPFLAGS = $(AM_CPPFLAGS) -DTX49
c/src/lib/libbsp/mips/jmr3904/Makefile.am:libbsp_a_CPPFLAGS = $(AM_CPPFLAGS) -DTX39
cpukit/pppd/Makefile.am:libpppd_a_CPPFLAGS = $(AM_CPPFLAGS) -D__BSD_VISIBLE -I$(srcdir)/../libmd
cpukit/libfs/Makefile.am:libjffs2_a_CPPFLAGS += -D__ECOS
cpukit/libfs/Makefile.am:libjffs2_a_CPPFLAGS += '-DKBUILD_MODNAME="JFFS2"'
cpukit/mghttpd/Makefile.am:# libmghttpd_a_CPPFLAGS += -DHAVE_MD5
cpukit/mghttpd/Makefile.am:libmghttpd_a_CPPFLAGS += -DNO_SSL -DNO_POPEN -DNO_CGI -DUSE_WEBSOCKET
cpukit/librpc/Makefile.am:librpc_CPPFLAGS = -D_RPC_read=read -D_RPC_write=write -D_RPC_close=close \
cpukit/libnetworking/Makefile.am:libnetworking_CPPFLAGS = -DINET -DNFS \
cpukit/libnetworking/Makefile.am:libc_CPPFLAGS = -DNOPOLL -DNOSELECT -D__BSD_VISIBLE -D_THREAD_SAFE
cpukit/libnetworking/Makefile.am:lib_CPPFLAGS = -DNOPOLL -DNOSELECT
cpukit/libnetworking/Makefile.am:lib_a_CPPFLAGS = $(AM_CPPFLAGS) $(lib_CPPFLAGS) -D__BSD_VISIBLE
cpukit/libdl/Makefile.am:libdl_a_CPPFLAGS = $(AM_CPPFLAGS) -DRTEMS_RTL_RAP_LOADER=1 -DRTEMS_RTL_ELF_LOADER=1
#3382 13 months ago fixed build Chris Johns Chris Johns 6 months ago
Summary

Testsuite Makefile merge to one per group of tests

Description

Merge the nested Makefile.am files into a single file per group of tests.

A single Makefile.am for all tests is not practical at this point in time because a test is an estimated 7 lines and with over 750 tests this means the file would be too big and a conflict hot spot.

#3383 13 months ago fixed build Chris Johns Chris Johns 13 months ago
Summary

Require --enable-rtemsbsp with --enable-smp or --enable-multiprocessor

Description

There is a limited number of BSPs that support SMP or MP so using the BSP wildcard will result in a failed build. Require the user provide a BSP.

#3385 13 months ago fixed build Chris Johns Chris Johns 13 months ago
Summary

Generate an error if RTEMS's gcc is not found when the user runs configure

Description

Generate an error when the user runs configure if one cannot be found in the path.

#3506 8 months ago fixed build Joel Sherrill Chris Johns 6 months ago
Summary

waf for building RTEMS applications needs updating

Description

Because there are no native tools in the RTEMS tree anymore, the RTEMS install point will not have a bin/ directory. If the --rtems-tools and --rtems directories are different, the sanity check by waf configure for ${rtems}/bin fails. See examples-v2.

+ ./waf configure -v --rtems=/home/joel/rtems-work/bsp-install --rtems-tools=/home/joel/rtems-work/tools/5 --rtems-bsps=powerpc/qemuppc
Setting top to                           : /home/joel/rtems-work/examples-v2 
Setting out to                           : /home/joel/rtems-work/examples-v2/build 
RTEMS path is not valid. No bin directory found.
(complete log in /home/joel/rtems-work/examples-v2/build/config.log)

#3553 6 months ago fixed build Joel Sherrill 6 months ago
Summary

rtems-libbsd Missing waf in Top Directory

Description

At least examples-v2 and rtems-libbsd use waf to build. examples-v2 has a copy of waf known to work for the users' convenience. rtems-libsd is missing one. Add one to rtems-libbsd.

Also (if there are other repos using waf), make sure they have a copy of waf also.

#2893 2 years ago fixed config Sebastian Huber Sebastian Huber 18 months ago
Summary

Remove CONFIGURE_SMP_APPLICATION

Description

SMP support must be enabled with the CONFIGURE_SMP_APPLICATION configuration option. Remove this option and enable the SMP support if CONFIGURE_SMP_MAXIMUM_PROCESSORS > 1.

#2894 2 years ago fixed config Sebastian Huber Sebastian Huber 18 months ago
Summary

Rename CONFIGURE_SMP_MAXIMUM_PROCESSORS to CONFIGURE_MAXIMUM_PROCESSORS

Description

Rename CONFIGURE_SMP_MAXIMUM_PROCESSORS to CONFIGURE_MAXIMUM_PROCESSORS since the SMP part is superfluous.

#2895 2 years ago fixed config Sebastian Huber Sebastian Huber 18 months ago
Summary

Prefix the confdefs.h internal defines with an underscore

Description

Prefix the confdefs.h internal defines with an underscore to simplify the review of this extremely complex header file.

#3063 22 months ago fixed config Sebastian Huber Sebastian Huber 18 months ago
Summary

Make the EDF scheduler the default SMP scheduler

Description

The EDF SMP scheduler supports simple thread processor affinities (see #3059) with a small run-time overhead. The current default SMP scheduler lacks support for thread processor affinities at all. The EDF SMP scheduler offers a good feature set for most applications. So, use it by default. Run-time libraries like libgomp, MTAPI, work stealing schedulers, language interpreters (e.g. Erlang virtual machine), etc. use a one-to-one thread processor affinity for example.

#3325 14 months ago fixed config Sebastian Huber Sebastian Huber 14 months ago
Summary

Simplify clustered scheduler configuration

Description

Improve the scheduler configuration documentation according to user review.

Do not use names derived from scheduler implementation details. Instead use names derived from the scheduler configuration or documentation. Provide defines for backward compatibility.

#3389 13 months ago fixed config Joel Sherrill 6 months ago
Summary

Warning flags have disappeared with recent autoconf changes

Description

As of March 30, the compiler invocations had warnings flags. As of today (4/11), there are no warnings flag on most of the compiler invocations.

Something has been lost in the updates.

#3434 11 months ago fixed config Sebastian Huber Sebastian Huber 9 months ago
Summary

Add CONFIGURE_MINIMUM_POSIX_THREAD_STACK_SIZE configuration option

#3435 11 months ago fixed config Sebastian Huber Sebastian Huber 8 months ago
Summary

Add test case for CONFIGURE_BSP_PREREQUISITE_DRIVERS configuration option

Description

This configuration option is untested

#3480 9 months ago fixed config Sebastian Huber Sebastian Huber 6 months ago
Summary

CONFIGURE_MINIMUM_TASK_STACK_SIZE may affect CONFIGURE_INTERRUPT_STACK_SIZE

Description

In case an application defines CONFIGURE_MINIMUM_TASK_STACK_SIZE, then this may change the CONFIGURE_INTERRUPT_STACK_SIZE as well:

#ifndef CONFIGURE_INTERRUPT_STACK_SIZE
  #ifdef BSP_INTERRUPT_STACK_SIZE
    #define CONFIGURE_INTERRUPT_STACK_SIZE BSP_INTERRUPT_STACK_SIZE
  #else
    #define CONFIGURE_INTERRUPT_STACK_SIZE CONFIGURE_MINIMUM_TASK_STACK_SIZE
  #endif
#endif

I think this is not what a user expects.

#3488 9 months ago fixed config Sebastian Huber Sebastian Huber 7 months ago
Summary

Remove CONFIGURE_HAS_OWN_MOUNT_TABLE

Description

RTEMS has the configuration option CONFIGURE_HAS_OWN_MOUNT_TABLE since 1999. This configuration option is broken since RTEMS 4.11. Remove this broken configuration option.

#3489 9 months ago fixed config Sebastian Huber Sebastian Huber 7 months ago
Summary

Obsolete CONFIGURE_HAS_OWN_CONFIGURATION_TABLE

Description

Obsolete the CONFIGURE_HAS_OWN_CONFIGURATION_TABLE configuration option. The RTEMS configuration should be done via explicit configuration options to allow more freedom for implementation changes.

#3490 9 months ago fixed config Sebastian Huber Sebastian Huber 2 weeks ago
Summary

Remove CONFIGURE_HAS_OWN_CONFIGURATION_TABLE

Description

This configuration option was obsoleted in RTEMS 5.1.

#3520 7 months ago fixed config Sebastian Huber Sebastian Huber 7 months ago
Summary

Remove CONFIGURE_HAS_OWN_FILESYSTEM_TABLE

Description

This configuration is untested and undocumented. Remove it to avoid a potential exposure of internal data structures to the application domain.

#3528 7 months ago fixed config Sebastian Huber Sebastian Huber 7 months ago
Summary

Remove undocumented and untested CONFIGURE_MAXIMUM_PTYS

Description

Remove the undocumented and untested CONFIGURE_MAXIMUM_PTYS configuration option. Add a rtems_telnetd_config_table::client_maximum member to the Telnet configuration.

#3239 17 months ago fixed dev Sebastian Huber Sebastian Huber 17 months ago
Summary

Add getentropy() implementation provided by each BSP

Description

The getentropy() system call was introduced by OpenBSD and is now also available on glibc 2.25 and later. It is used for example by arc4random_buf(). Which in turn is used by various cryptographic functions.

#3436 11 months ago fixed dev Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove clock driver Clock_driver_support_shutdown_hardware() hook

Description

Most applications use a clock driver and most BSPs use the clock driver framework provided by clockimpl.h. This framework offers a Clock_driver_support_shutdown_hardware() hook which is used like this.

#ifdef Clock_driver_support_shutdown_hardware
/**
 *  @brief Clock_exit
 *
 *  This routine allows the clock driver to exit by masking the interrupt and
 *  disabling the clock's counter.
 */
void Clock_exit( void )
{
  Clock_driver_support_shutdown_hardware();

  /* do not restore old vector */
}
#endif
...
#ifdef Clock_driver_support_shutdown_hardware
  atexit( Clock_exit );
#endif

The aim is to stop clock tick interrupts at some late point in the exit() procedure.

The use of atexit() pulls in malloc() which pulls in errno. It is incompatible with the intention of the CONFIGURE_DISABLE_NEWLIB_REENTRANCY configuration option.

The exit() function must be called from thread context, so accompanied clock tick interrupts should cause no harm. On the contrary, someone may assume a normal operating system operation, e.g. working timeouts.

Remove the superfluous Clock_driver_support_shutdown_hardware() hook.

#2840 2 years ago fixed dev/serial Sebastian Huber Sebastian Huber 14 months ago
Summary

Use self-contained mutexes for Termios framework

Description

Use C11 mutexes instead of Classic semaphores as a performance optimization and to simplify the application configuration.

A performance of Classic semaphores vs. C11 mutexes was measured on the arm/atsam BSP. A NXP SC16IS752 was connected via SPI. The RTEMS application used one task to read from the device and write it immediately back (look back via task). A development system constantly transmitted data at 115200 bits per second.

CPU usage by function with Classic semaphores:

name______________________________________|ratio___|1%_____2%________5%_____10%_____20%_______50%___100|
CPU_Thread_Idle_body                      | 22.454%|=================================                  |
atsam_spi_setup_transfer                  |  6.767%|====================                               |
Objects_Get                               |  5.859%|===================                                |
atsam_spi_interrupt                       |  4.483%|================                                   |
Event_Seize                               |  3.867%|==============                                     |
rtems_termios_enqueue_raw_characters      |  3.804%|==============                                     |
Timecounter_Binuptime                     |  3.715%|==============                                     |
Scheduler_priority_Block                  |  3.104%|============                                       |
rtems_semaphore_release                   |  3.018%|============                                       |
Scheduler_priority_Unblock                |  2.901%|===========                                        |
rtems_termios_read_tty                    |  2.777%|===========                                        |
ARMV7M_NVIC_Interrupt_dispatch            |  2.750%|===========                                        |
rtems_semaphore_obtain                    |  2.627%|==========                                         |
Thread_Do_dispatch                        |  2.351%|=========                                          |
ARMV7M_Interrupt_service_leave            |  2.086%|========                                           |
iproc                                     |  1.919%|=======                                            |
CPU_Context_switch                        |  

CPU usage by function with C11 mutexes:

name______________________________________|ratio___|1%_____2%________5%_____10%_____20%_______50%___100|
CPU_Thread_Idle_body                      | 33.395%|======================================             |
atsam_spi_setup_transfer                  |  6.061%|===================                                |
atsam_spi_interrupt                       |  4.690%|================                                   |
Mutex_recursive_Release                   |  3.011%|============                                       |
Event_Seize                               |  2.955%|===========                                        |
ARMV7M_NVIC_Interrupt_dispatch            |  2.885%|===========                                        |
rtems_termios_enqueue_raw_characters      |  2.771%|===========                                        |
rtems_termios_read_tty                    |  2.722%|===========                                        |
Timecounter_Binuptime                     |  2.653%|==========                                         |
Thread_Do_dispatch                        |  2.240%|========                                           |
Scheduler_priority_Block                  |  2.112%|========                                           |
ARMV7M_Interrupt_service_leave            |  2.100%|========                                           |
Scheduler_priority_Unblock                |  1.919%|=======                                            |
Mutex_recursive_Acquire                   |  1.876%|======                                             |
iproc                                     |  1.773%|======                                             |
CPU_Context_switch                        |  

The change resulted in 10% more total idle time on the system.

#3198 18 months ago fixed dev/serial Sebastian Huber Sebastian Huber 18 months ago
Summary

Add lazy update of line control and baud divisor to NS16550 serial driver

Description

Updates of the line control and baud divisor while transfers are in progress may lead to unpredictable behaviour on some chips. Perform the updates only if necessary.

#3218 18 months ago fixed dev/serial Sebastian Huber Sebastian Huber 18 months ago
Summary

Termios canonical mode (ICANON) does not return input line by line

Description

In canonical mode, input is made available line by line. We must stop the canonical buffer filling upon reception of an end-of-line character.

#3320 14 months ago fixed dev/serial Sebastian Huber Sebastian Huber <sebastian.huber@…> 10 months ago
Summary

Add a simple task console driver

Description

The default console driver for tests is the simple console driver. It uses a polled output via rtems_putc() done directly in the context of the executing thread. This is a problem for timing sensitive tests. Add a simple task console driver.

.. index:: CONFIGURE_APPLICATION_NEEDS_SIMPLE_TASK_CONSOLE_DRIVER

.. _CONFIGURE_APPLICATION_NEEDS_SIMPLE_TASK_CONSOLE_DRIVER:

CONFIGURE_APPLICATION_NEEDS_SIMPLE_TASK_CONSOLE_DRIVER
------------------------------------------------------

CONSTANT:
    ``CONFIGURE_APPLICATION_NEEDS_SIMPLE_TASK_CONSOLE_DRIVER``

DATA TYPE:
    Boolean feature macro.

RANGE:
    Defined or undefined.

DEFAULT VALUE:
    This is not defined by default.

DESCRIPTION:
    ``CONFIGURE_APPLICATION_NEEDS_SIMPLE_TASK_CONSOLE_DRIVER`` is defined if
    the application wishes to include the Simple Task Console Device Driver.

NOTES:
    This device driver is responsible for providing the :file:`/dev/console`
    device file.  This device is used to initialize the standard input, output,
    and error file descriptors.

    This device driver reads via ``getchark()``.

    This device driver writes into a write buffer.  The count of characters
    written into the write buffer is returned.  It might be less than the
    requested count, in case the write buffer is full.  The write is
    non-blocking and may be called from interrupt context.  A dedicated task
    reads from the write buffer and outputs the characters via
    ``rtems_putc()``.  This task runs with the least important priority.  The
    write buffer size is 2047 characters and it is not configurable.

    Use ``fsync(STDOUT_FILENO)`` or ``fdatasync(STDOUT_FILENO)`` to drain the
    write buffer.

    The Termios framework is not used.  There is no support to change device
    settings, e.g.  baud, stop bits, parity, etc.

    The

    * ``CONFIGURE_APPLICATION_NEEDS_CONSOLE_DRIVER``,

    * ``CONFIGURE_APPLICATION_NEEDS_SIMPLE_CONSOLE_DRIVER``, and

    * ``CONFIGURE_APPLICATION_NEEDS_SIMPLE_TASK_CONSOLE_DRIVER``

    configuration options are mutually exclusive.
#3526 7 months ago fixed dev/serial Sebastian Huber Sebastian Huber 7 months ago
Summary

Convert PTY driver to new Termios API

Description
#2198 4 years ago fixed doc Gedare Chris Johns 10 months ago
Summary

Automate doxygen build

Description

The doxygen builds are no longer being generated nightly.

#2367 4 years ago fixed doc mw Sebastian Huber <sebastian.huber@…> 18 months ago
Summary

Documentation of User Extensions needs more information

Description

The documentation for User Extension sets in the C User's Guide could use some clarification. It discusses the multiple sets of extensions, but it is unclear as to whether or not the extensions are added or replaced when rtems_extension_create() is called. There is a section - 22.2.4 (Order of Invocation) that does discuss the operation of the sets, but it only makes sense once the reader understands that the sets are, in fact, additive.

#2557 3 years ago fixed doc Amar Takhar Amar Takhar 18 months ago
Summary

Add word splitting to print output

#2680 3 years ago fixed doc Joel Sherrill Chris Johns 18 months ago
Summary

Add pthread_setconcurrency() and pthread_getconcurrency()

Description

* Code merged. Ticket changed to documentation to remind us to add documentation when master documentation reopens in new format.

We only require the simple implementation documented here:

http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_getconcurrency.html

This is required for FACE Conformance.

#2693 3 years ago fixed doc Joel Sherrill Chris Johns 18 months ago
Summary

Update doc to reflect obsoleting rtems_clock_get()

Description

Ticket to reflect documentation change needed on the master but not on 4.11. When new documentation format is available for master, this needs to be accounted for.

#2770 3 years ago fixed doc Christian Mauderer Sebastian Huber 18 months ago
Summary

Missing documentation for RTEMS_LINKER_ROSET_CONTENT and RTEMS_LINKER_RWSET_CONTENT

Description

Currently the two macros

RTEMS_LINKER_ROSET_CONTENT RTEMS_LINKER_RWSET_CONTENT

are not documented. This should be added as soon as the doc repo is ready for it.

The macros have been introduced in this commit:

https://git.rtems.org/rtems/commit/?id=5fe6d07ad5690e3d9c6445ca3a465a700a5a5015

#2825 2 years ago fixed doc Sebastian Huber 18 months ago
Summary

Improve the fatal error handling chapter of the user manual

Description

At least replace the "Document me" markers with something useful.

#2845 2 years ago fixed doc Sebastian Huber 18 months ago
Summary

Add I2C framework documentation

Description

The new I2C framework lacks documentation.

#2849 2 years ago fixed doc Sebastian Huber 18 months ago
Summary

ATA/IDE support in RTEMS is out-dated

Description

The ATA/IDE support in RTEMS is out-dated. New platforms should consider to use the SATA support provided by FreeBSD via libbsd.

Update the documentation accordingly.

#2850 2 years ago fixed doc Sebastian Huber 18 months ago
Summary

Driver manual covers non-existent Analog Driver

Description

Remove this chapter from the documentation.

#2851 2 years ago fixed doc Sebastian Huber 18 months ago
Summary

Driver manual covers non-existent Discrete Driver

Description

Remove this chapter from the documentation.

#2853 2 years ago fixed doc Sebastian Huber 18 months ago
Summary

Driver manual covers non-existent Non-Volatile Memory Driver

Description

Remove this chapter from the documentation.

#2862 2 years ago fixed doc Chris Johns Chris Johns 18 months ago
Summary

docs.rtems.org Add support to ReST format releases.

Description

Add support to the releases section of the web site to handle ReST packages. The catalogues have a legacy field for texinfo docs.

The 4.11.0 and 4.11.1 releases need to have a catalogue added because this did not exist when those releases were created.

#2863 2 years ago duplicate doc Joel Sherrill Joel Sherrill 18 months ago
Summary

Update POSIX 1003.1 Compliance Guide for ReST

Description

The POSIX 1003.1 Compliance Guide should be auto-generated from a spreadsheet into the ReST format. My vague recollection is that we used shell scripts to do this for texinfo output.

I will have to decipher what we used to do and define a new procedure.

This impacts 4.11 and newer. One issue is having correct information for what methods are present on a branch. The FACE Conformance Test Suite can be used for ~800 of the methods.

#2864 2 years ago fixed doc Chris Johns 18 months ago
Summary

docs.rtems.org Automatic update of branches content when a rtems-doc.git change is made.

Description

Add support to automatically update the branches when a git commit happens.

#2865 2 years ago fixed doc Chris Johns Chris Johns 18 months ago
Summary

Coverpage installed when building the docs repeats catalogue.xml entries

Description

The catalogue repeats entries.

#2906 2 years ago fixed doc Chris Johns 18 months ago
Summary

rtems-doc waf configure does not detect sphinxcontrib.bibtex status

Description

The rtems-docs requires the Sphinx contribution extension for bibtex to build.

A configure check should be added to see if the extension is installed and an error raised or the documentation conditionally built.

It is not clear to me if bibtex needs Tex Alive. Requiring Tex on all hosts to build the documentation is a regression so the extension will need to be removed or a fall back position taken where degraded quality documentation in HTML is created.

#3008 2 years ago fixed doc Hassan Karim chrisj@… 18 months ago
Summary

missing pax causes install failures

Description

I have tried to install sparc bsp=erc32 on 4 different builds. 3 failed, and 1 flawlessly installed. The others all seem to fail somewhere during make install of test suites. Each reports one missing config problem or another.

I believe the problem resulted in a missing package, pax & libbsd-dev on Ubuntu 12.04.5 LTS (GNU/Linux 3.2.0-126-virtual x86_64)

I hadn't seen this exact problem because I normally update & upgrade as soon as I get a new image. Pressed for time, I skipped it. So, I am not sure if we need to update the documentation to directly include pax, since it is directly called in configure and breaks if not present.

https://docs.rtems.org/rsb/#_host_setups Under this section, 11.1.5. Ubuntu

Add pax to this line $ sudo apt-get build-dep binutils gcc g++ gdb unzip git python2.7-dev pax

#3130 20 months ago fixed doc Chris Johns chrisj@… 18 months ago
Summary

RTEMS Doxygen.in latex output does not build

Description

Doxygen latex output on sync.rtems.org does not build.

Does latex output build on any host? If so which hosts and what tool combination.

If it does not build we should consider defaulting the setting for latex output to "no".

#3141 19 months ago fixed doc Chris Johns Chris Johns 18 months ago
Summary

Change the BSP Howto's name to something smaller.

Description

The BSP Howto's current name is:

RTEMS BSP and Device Driver Development Guide

This is long and causes problems in the PDF output. Change the name to:

RTEMS BSP and Driver Guide

#3177 19 months ago fixed doc Joel Sherrill Chris Johns 18 months ago
Summary

Replace/update POSIX Compliance Guide

Description

The POSIX Compliance Guide was never converted from texinfo. Beyond that, it is out of date and follows the outline of the printed version of the POSIX standard which no one sees anymore. This ticket proposes:

  1. new Sphinx document
  2. contents generated from POSIX API tracking spreadsheet (CSV)
  3. outline per .h file, not functional area
  4. use bullets, not tables so easier to format
#3207 18 months ago fixed doc Joel Sherrill 17 months ago
Summary

Supported Architectures Page is out of date

Description

https://devel.rtems.org/wiki/TBR/UserManual/SupportedCPUs is out of date. I have the information to update it if that's what we want to do.

I don't know the best way to provide this broad view from 4.6 up on what architectures are supported. The wiki seems OK.

#3215 18 months ago fixed doc Joel Sherrill Joel Sherrill 18 months ago
Summary

Configuring a System Still Includes Notepads and Has Wrong Heading

Description

This section has the wrong heading and needs to be deleted anyway.

24.8.2. Specify Maximum Classic API Timers
CONSTANT:
CONFIGURE_ENABLE_CLASSIC_API_NOTEPADS
#3229 18 months ago fixed doc Chris Johns Chris Johns 6 months ago
Summary

Add index to all documents.

Description

Indexes currently do not work. Fix this adding them to all documents.

Add index entries where possible.

#3232 18 months ago fixed doc Chris Johns Chris Johns 18 months ago
Summary

Use of .. include:: in the User Manual should be changed.

Description

This tricket for Sphinx highlights an issue when using .. include::, we should be using .. toctree:::

https://github.com/sphinx-doc/sphinx/issues/3432

#3234 18 months ago invalid doc Joel Sherrill Joel Sherrill 18 months ago
Summary

Quick Start Instructions Inconsistent

Description

In section 5 of the User's Manual, the clone of rtems-source-builder has you clone it into rsb but the sb-bootstrap command is based on cloning it into the rsb subdirectory.

#3261 17 months ago fixed doc Frédéric Jouault Frédéric Jouault <f.jouault@…> 4 months ago
Summary

A couple of documentation typos

Description

This patch includes a couple of typo corrections in the C User Manual.

#3283 15 months ago fixed doc Joel Sherrill Chris Johns 15 months ago
Summary

Bad URL in OpenOCD/Xilinx_Zynq Wiki Page

Description

https://devel.rtems.org/wiki/Debugging/OpenOCD/Xilinx_Zynq has a link to the Zedboard Processor Debug Adapter. I think the URL has changed to this but would like someone more knowledgeable to confirm that before it is changed.

http://zedboard.org/accessories/zedboard-processor-debug-adapter

#3309 14 months ago fixed doc Chris Johns 6 months ago
Summary

rtems_task_create's initial_mode SMP update

Description

The initial_mode cannot have the non-preempt flag or an interrupt level set or an RTEMS_UNSATISFIED error is returned. This is not documented in the directive.

#3454 10 months ago fixed doc Vidushi Vashishth Vidushi Vashishth 2 months ago
Summary

Tracing Framework Documentation in User Manual

Description

1) Write up a new chapter in the user manual regarding the existing tracing framework in RTEMS. Include a description of the components of the tracing framework and the various techniques used to generate traces currently. Add explanatory demonstrations and samples.

2) Expand the chapter to include CTF generation (currently under development) as it evolves.

#3498 9 months ago fixed doc Jens Schweikhardt Sebastian Huber 8 months ago
Summary

Command and Variable Index is empty

Description

The Command and Variable Index, https://docs.rtems.org/branches/master/cpu-supplement/command.html does not contain any commands or variables.

If this chapter does not apply to RTEMS it should not be generated. If it does apply, it should contain what its title promises.

#3499 9 months ago fixed doc Jens Schweikhardt Sebastian Huber 8 months ago
Summary

The "Index" chapter is empty

Description

The "Index" chapter, https://docs.rtems.org/branches/master/cpu-supplement/genindex.html is empty.

This chapter should contain a usable and helpful index or not be generated at all.

#3503 8 months ago fixed doc Chris Johns Chris Johns 8 months ago
Summary

PDF Documentation is missing an index

Description

The PDF generated documents have an empty index.

#3517 7 months ago fixed doc Joel Sherrill Joel Sherrill <joel@…> 7 months ago
Summary

RSB Ubuntu Host Requirements Missing Some

Description

Add libncurses5-dev and zlib1g-dev to Ubuntu apt-get instructions

Also bison and flex seemed to be missing per one of the persons trying it.

#3518 7 months ago fixed doc Joel Sherrill Chris Johns 7 months ago
Summary

RSB MacOS Nits

Description

The MacOS section of the RSB manual has some minor things that need to be fixed:

  • says Serria when it should be Sierra (I think).
  • has +sb-check+ which indicates a formatting error

Please review as a Mac user and make sure that's it. :)

#3538 7 months ago fixed doc Joel Sherrill Joel Sherrill 7 months ago
Summary

Classic API Barrier Wait Section Title Has Wrong Name

Description

The section title says obtain not wait.

Likely also applies to 4.11.

#3669 4 months ago fixed doc Chris Johns Amar Takhar 3 months ago
Summary

rtems-docs.git does not build with Sphinx 1.8.2 and 1.8.3

Description

The docs do not build with Sphinx 1.8. Recent posts indicate some changes to support unicode indexes via xindy have caused this and the solution being discussed is to use the generate Makefile ...

https://github.com/rtfd/readthedocs.org/issues/4454

The need to use the Makefile is debatable however what it contains is important as it defines what needs to happen.

This recent issue can be seen in the Tex generated file for the User Manual (user.tex). It contains:

\item[{Waf\index{Waf@\spxentry{Waf}|spxpagem}\phantomsection\label{\detokenize{glossary/index:term-waf}}}] \leavevmode
Waf build system.  For more information see \sphinxurl{http://www.waf.io/}

Our current build uses pdflatex directly and there is an error as spxpagem is not defined.

If you inspect a version 1.8 generated Makefile the command latexmk is used. This wraps the PDF generation so the correct number of passes are performed. Using this tool should be considered.

The Makefile contains:

export XINDYOPTS = -L english -C utf8  -M sphinx.xdy

The sphinx.xdy contains the needed spxpagem. I can only conclude sphinx needs to be built with xindy because the reference is always generated.

The problem for building FreeBSD is xindy is not an available command.

#3424 12 months ago fixed examples Joel Sherrill Joel Sherrill 6 months ago
Summary

examples-v2: no MIPS BSPs pass configuration step

Description
Checking for program 'ar'                  : /home/joel/rtems-work/tools/5/bin/mips-rtems5-ar 
Checking for program 'g++, c++'            : /home/joel/rtems-work/tools/5/bin/mips-rtems5-g++ 
Checking for program 'ar'                  : /home/joel/rtems-work/tools/5/bin/mips-rtems5-ar 
Checking for program 'gas, gcc'            : /home/joel/rtems-work/tools/5/bin/mips-rtems5-gcc 
Checking for program 'ar'                  : /home/joel/rtems-work/tools/5/bin/mips-rtems5-ar 
Compiler version (mips-rtems5-gcc)         : 7.3.0 20180125 (RTEMS 5, RSB 6d9c77c77d271d1fc2dfe8493d6713930b52a6dd, Newlib 3.0.0) 
Checking for RTEMS CPU options header      : started 
-> processing test results                 : 1 test failed 
One of the tests has failed, read config.log for more information
(complete log in /data/home/joel/rtems-work/examples-v2/build/config.log)
+ check_fatal 1 'failed waf configure - examples-v2 on rbtx4925'
#3670 4 months ago fixed examples Chris Johns joel@… 4 months ago
Summary

examples-v2 uses deprecated or obsolete RTEMS interfaces

Description

The examples need to be change to use what ever is the newer method of doing something.

../../filesystem/fat_ramdisk/init.c:46:3: warning: 'rtems_blkdev_generic_open' is deprecated [-Wdeprecated-declarations] 
   open_entry:           rtems_blkdev_generic_open,
   ^~~~~~~~~~
In file included from /opt/work/chris/rtems/kernel/5/arm-rtems5/xilinx_zynq_zedboard/lib/include/rtems/ramdisk.h:17:0,
                 from ../../filesystem/fat_ramdisk/init.c:16:
/opt/work/chris/rtems/kernel/5/arm-rtems5/xilinx_zynq_zedboard/lib/include/rtems/blkdev.h:408:1: note: declared here
 rtems_blkdev_generic_open( 
 ^~~~~~~~~~~~~~~~~~~~~~~~~ 
../../filesystem/fat_ramdisk/init.c:47:3: warning: 'rtems_blkdev_generic_close' is deprecated [-Wdeprecated-declarations]
   close_entry:          rtems_blkdev_generic_close,                                        
   ^~~~~~~~~~~                          
In file included from ../../ticker/low_ticker/init.c:88:0:
/opt/work/chris/rtems/kernel/5/arm-rtems5/xilinx_zynq_zedboard/lib/include/rtems/confdefs.h:3276:4: warning: #warning "The CONFIGURE_TERMIOS_DISABLED configuration option is obsolete since RTEMS 5.1" [-Wcpp]
   #warning "The CONFIGURE_TERMIOS_DISABLED configuration option is obsolete since RTEMS 5.1"
    ^~~~~~~
#1247 12 years ago fixed fs strauman Chris Johns 18 months ago
Summary

RTEMS does not implement locks needed by multithreaded newlib

Description

multi-threaded newlib protects a number of internal data structures (as of newlib-1.15 these comprise:

  • global list of FILE objects
  • stdio FILE object initializer
  • individual FILEs [since FILEs with the exception of stdin/stdout/stderr are per-process entities]
  • global hash table used by telldir/seekdir
  • individual DIR structures (opendir/readdir)
  • atexit list
  • list of environment variables
  • global timezone variable

) using mutexes. It expects the OS to implement these locks but defaults to not using locking if the OS does not provide an implementation. Currently, RTEMS does *not* provide its own implementation of 'sys/lock.h' and therefore vital data structures in newlib are currently *unprotected* (with the exception of environment variables -- 'envlock.c' had been added to RTEMS a while ago but since then, newlib has introduced more locks and a general OS interface which - once implemented - will obsolete 'envlock.c').

Note that while semantics of having no protection for individual FILE objects may be tolerable, having no protection for global newlib data structures such as lists of FILEs is not acceptable.

I am currently working on an implementation which should be available shortly.

#1662 9 years ago fixed fs Bharath Suri Chris Johns 15 months ago
Summary

termios.c: semaphore not deleted, consequently resulting in failure of rtems_termios_open

Description

The semaphore osem is still in use in rtems_termios_close while an attempt to delete it is made and hence is not deleted. Consequently, it results in a RTEMS_TOO_MANY on rtems_semaphore_create, which further results in failure of rtems_termios_open.

#1971 7 years ago fixed fs nopscmn Chris Johns 15 months ago
Summary

Memory leak in tmpfile()

Description

Hello,

I use tmpfile() in my RTEMS application, and found that it has memory leak. I wrote small test application (see attachment), that output memory usage:

Memory usage before: Number of used blocks: 12 Largest used block: 1288 Total bytes used: 3628

Memory used after: Number of used blocks: 1013 Largest used block: 1288 Total bytes used: 112064

By 1000 iteration, each call tmpfile() cause memory leak about 108 bytes.

#2844 2 years ago fixed fs Sebastian Huber Sebastian Huber 18 months ago
Summary

JFFS2: Add IO controls to get filesystem instance information and force a garbage collection

Description

Some applications need to control the garbage collection of the JFFS2 filesystem. For example during bootloader to application transitions with execute in place flashes (XIP).

#3089 21 months ago fixed fs Fan Deng Fan Deng 13 months ago
Summary

Inconsistent blocking addressing in RFS

Description

Background There are two ways to address a block in RFS:

  1. Via a single 32bit block number (bno)
  2. Via a group number(gno) and a bit offset (bit)

They should be fully convertible (1-1 mapping). In other words, the equation to convert 1 to 2 should be unique within the RFS implementation.

The bug The RFS implementation contains two different conversions between 1 and 2.

Details

  1. In rtems_rfs_group_bitmap_alloc (rtems-rfs-group.c, line 172)
    bno = gno * group_blocks + bit
    
  1. In rtems_rfs_group_bitmap_alloc (rtems-rfs-group.c, line 228)
    bno = gno * group_blocks + bit + 1 (via rtems_rfs_group_block() function)
    
  1. In rtems_rfs_group_bitmap_free (rtems-rfs-group.c, line 283)
    bno = gno * group_blocks + bit + 1 (RTEMS_RFS_SUPERBLOCK_SIZE)
    
  1. In rtems_rfs_group_bitmap_test (rtems-rfs-group.c, line 332)
    bno = gno * group_blocks + bit
    

To summarize, the implementation contains two ways of converting a bno to a (gno, bit) pair:

Either:

bno = gno * group_blocks + bit

Or:

bno = gno * group_blocks + bit + 1

The Fix The RFS implementation should consistently convert a bno to a (gno, bit) pair with:

bno = gno * group_blocks + bit + RTEMS_RFS_SUPERBLOCK_SIZE

This is because the superblock is not accounted for in the block bitmaps. So places to change:

  1. rtems-rfs-group.c: all references to the conversion must be updated to use RTEMS_RFS_SUPERBLOCK_SIZE explicitly.
  1. rtems_rfs_group_block converts the pair to bno via:
    #define rtems_rfs_group_block(_g, _b) (((_g)->base) + (_b))
    

(_g)->base is calculated via rtems-rfs-format.c from:

#define rtems_rfs_fs_block(_fs, _grp, _blk) \
  ((((_fs)->group_blocks) * (_grp)) + (_blk) + 1)

The "+ 1" part should really be "+ RTEMS_RFS_SUPERBLOCK_SIZE" to be logically correct. As RTEMS_RFS_SUPERBLOCK_SIZE itself has a comment saying:

/**
 * Number of blocks in the superblock. Yes I know it is a superblock and not
 * superblocks but if for any reason this needs to change it is handled.
 */
#define RTEMS_RFS_SUPERBLOCK_SIZE (1)
#3132 20 months ago fixed fs Sebastian Huber Sebastian Huber 17 months ago
Summary

Add reference counting to file descriptors

Description

The use of a file descriptor after or during a close() operation may result in a use after free. Finding such errors in applications is difficult. Especially in SMP systems using the highly dynamic libbsd network stack.

The file descriptor objects reside in a table with a application configuration defined size. So, the storage of a file descriptor object is always present, only the referenced file system node may change over time. The file system nodes may use an internal reference counting, which is independent of the file descriptors.

To implement reference counting for the file descriptors add a bit field for the reference count to the rtems_libio_t::flags and use atomic operations to maintain the flags.

Each operation using a file descriptor should perform a sequence like this:

int op( int fd, ... )
{
  rtems_libio_t *iop;
  unsigned int   flags;

  if ( (uint32_t) fd >= rtems_libio_number_iops ) {
    rtems_set_errno_and_return_minus_one( EBADF );
  }

  iop = rtems_libio_iop( fd );
  flags = rtems_libio_iop_hold( iop );

  if ( ( flags & LIBIO_FLAGS_OPEN ) == 0 ) {
    rtems_libio_iop_drop( _iop );
    rtems_set_errno_and_return_minus_one( EBADF );
  }

  do_op( iop, ... );
  rtems_libio_iop_drop( iop );
  return 0;
}

A close() should return -1 with EBUSY in case the file descriptor is referenced. In this case, no close operation will be performed.

#3133 20 months ago fixed fs Sebastian Huber Sebastian Huber 18 months ago
Summary

Remove rtems_libio_t::driver

Description

Remove unused rtems_libio_t::driver member.

#3134 20 months ago fixed fs Sebastian Huber Sebastian Huber 18 months ago
Summary

Remove LIBIO_FLAGS_CREATE

Description

Remove unused LIBIO_FLAGS_CREATE flag.

#3136 20 months ago fixed fs Sebastian Huber Sebastian Huber 18 months ago
Summary

Use FIFO for file descriptor free list

Description

Currently, the free list of file descriptors is organized as a LIFO. In erroneous systems which use a file descriptor after a call to close(), this increases the likelihood that this error is undetected due to the prompt re-use of the file descriptor. The use of a FIFO has the benefit that free file descriptors remain on the free list as long as possible. This increases the time frame in which an invalid use of a closed file descriptor returns an error status.

#3694 2 months ago fixed fs Joel Sherrill Gedare 6 weeks ago
Summary

shm_open has logically unreachable code (Coverity ID: 1399706, 1399714)

Description

Coverity ID: 1399706 and 1399714 File: shmopen.c Method: shm_open for first

   	dead_error_condition: The condition oflag & 0 cannot be true.
289  if ( oflag & O_RDONLY ) {
   	
CID 1399706 (#1 of 1): Logically dead code (DEADCODE)
dead_error_line: Execution cannot reach this statement: flags |= 2U;.
290    flags |= LIBIO_FLAGS_READ;
291  } else {

URL: https://scan5.coverity.com/reports.htm#v29811/p10069/fileInstanceId=153084281&defectInstanceId=42558012&mergedDefectId=1399706&fileStart=1&fileEnd=250

Same issue at other place in same file:

197  int flags;
   	dead_error_condition: The condition oflag & 0 cannot be true.
198  if ( oflag & O_RDONLY ) {
   	
CID 1399714 (#1 of 1): Logically dead code (DEADCODE)
dead_error_line: Execution cannot reach this statement: flags = 4;.
199    flags = RTEMS_FS_PERMS_READ;
200  } else {
#2727 3 years ago fixed fs/fat Sebastian Huber Sebastian Huber 18 months ago
Summary

FAT file systems use wrong semaphore for mutual exclusion

Description
cpukit/libfs/src/dosfs/msdos_initsupp.c-
cpukit/libfs/src/dosfs/msdos_initsupp.c-    sc = rtems_semaphore_create(3,
cpukit/libfs/src/dosfs/msdos_initsupp.c-                                1,
cpukit/libfs/src/dosfs/msdos_initsupp.c:                                RTEMS_BINARY_SEMAPHORE | RTEMS_FIFO,
cpukit/libfs/src/dosfs/msdos_initsupp.c-                                0,
cpukit/libfs/src/dosfs/msdos_initsupp.c-                                &fs_info->vol_sema);
cpukit/libfs/src/dosfs/msdos_initsupp.c

Should use a binary semaphore with inherit priority.

#3003 2 years ago fixed fs/fat munster Sebastian Huber 18 months ago
Summary

FAT does not support clusters bigger than 32K

Description

When used with 64KiB clusters, the FAT driver will loop forever in cpukit/libfs/src/dosfs/fat.c, line 580. This happens because struct fat_vol_s declares bytes per cluster variable as uint16_t bpc, whereas it can be as big as 256KiB.

Here is a link for Linux FAT driver which doesn't make any assumption about cluster size: http://lxr.free-electrons.com/source/fs/fat/inode.c?v=2.6.24#L1262

#3465 10 months ago fixed fs/jaffs2 Sebastian Huber Sebastian Huber 7 months ago
Summary

Integrate all changes from Linux v3.11 to v4.17 made in the JFFS2 sources

Description

The original import version of the JFFS2 sources was Linux v3.11 (September 2013). Update the JFFS2 sources to Linux v4.17.

The Git command to generate the patches is:

git format-patch v3.11..v4.17 -- include/uapi/linux/jffs2.h fs/jffs2/LICENCE fs/jffs2/acl.h fs/jffs2/build.c fs/jffs2/compr.c fs/jffs2/compr.h fs/jffs2/compr_rtime.c fs/jffs2/compr_rubin.c fs/jffs2/compr_zlib.c fs/jffs2/debug.c fs/jffs2/debug.h fs/jffs2/erase.c fs/jffs2/gc.c fs/jffs2/jffs2_fs_i.h fs/jffs2/jffs2_fs_sb.h fs/jffs2/nodelist.c fs/jffs2/nodelist.h fs/jffs2/nodemgmt.c fs/jffs2/read.c fs/jffs2/readinode.c fs/jffs2/scan.c fs/jffs2/summary.h fs/jffs2/write.c fs/jffs2/xattr.h

We need a source file transformation in the patches:

sed -i 's%/fs/jffs2%/cpukit/libfs/src/jffs2/src%g' 00*

To support the first commit:

From e8bbeeb755a077cfc0f814b07739f9225642d65c Mon Sep 17 00:00:00 2001
From: Cody P Schafer <cody@linux.vnet.ibm.com>
Date: Thu, 23 Jan 2014 15:56:11 -0800
Subject: [PATCH 01/24] fs/jffs2: use rbtree postorder iteration helper instead
 of opencoding

Use rbtree_postorder_for_each_entry_safe() to destroy the rbtree instead
of opencoding an alternate postorder iteration that modifies the tree

Signed-off-by: Cody P Schafer <cody@linux.vnet.ibm.com>
Cc: Michel Lespinasse <walken@google.com>
Cc: Jan Kara <jack@suse.cz>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

we have to a postorder iterator to the red-black tree support code.

The remaining 23 patches are easy to apply.

#3484 9 months ago fixed fs/rfs Sebastian Huber Sebastian Huber 9 months ago
Summary

RFS: Remove stray call of rtems_disk_release() in rtems_rfs_buffer_sync()

Description

The function rtems_rfs_buffer_sync() erroneously calls rtems_disk_release(). This screws up the reference counting of the disk.

#3217 18 months ago fixed lib Chris Johns Chris Johns 13 months ago
Summary

Add RTEMS version, build and tools details to tests

Description

Published test results need the RTEMS version, how it is built and the tools used to build the kernel and tests.

#3240 17 months ago fixed lib Andrei Chichak Chris Johns 17 months ago
Summary

cpukit/libmisc/stackchk/check.c stack addresses formatted incorrectly.

Description

The function Stack_check_Dump_threads_usage displays the stack high, low, and current pointers incorrectly.

Instead of displaying these pointers in conventional hex format, the values have a proper prefix of 0x, but the pointer value is displayed in decimal.

The incorrect inttypes.h formatting define was used.

#3266 16 months ago duplicate lib Chris Johns 16 months ago
Summary

cpukit/libpci references BSP headers.

Description

On the no-preinstall branch of https://git.rtems.org/chrisj/rtems.git/ the build fails with:

sparc-rtems5-gcc --pipe -DHAVE_CONFIG_H   -I.. -I/opt/work/chris/rtems/kernel/bsps/beagleboneblack/sparc-rtems5/c/erc32/include -I/opt/work/chris/rtems/kernel/rtems.git/cpukit/include -I/opt/work/chris/rtems/kernel/rtems.git/cpukit/score/cpu/sparc/include   -mcpu=cypress -O2 -g -ffunction-sections -fdata-sections -MT pci_access.o -MD -MP -MF $depbase.Tpo -c -o pci_access.o /opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/libpci/pci_access.c &&\
mv -f $depbase.Tpo $depbase.Po
In file included from /opt/work/chris/rtems/kernel/rtems.git/cpukit/include/pci.h:23:0,
                 from /opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/libpci/pci_access.c:10:
/opt/work/chris/rtems/kernel/rtems.git/cpukit/include/pci/access.h:16:10: fatal error: libcpu/byteorder.h: No such file or directory
 #include <libcpu/byteorder.h>
          ^~~~~~~~~~~~~~~~~~~~

This header is found under:

$ find . -name byteorder.h
./bsps/powerpc/include/libcpu/byteorder.h
./bsps/sparc/include/libcpu/byteorder.h
./bsps/i386/include/libcpu/byteorder.h
#3267 16 months ago fixed lib Chris Johns 13 months ago
Summary

rtems/status-checks.h calls printk without including the needed header.

Description
/opt/work/chris/rtems/kernel/rtems.git/cpukit/include/rtems/status-checks.h:74:7: warning: implicit declaration of function 'printk'; did you mean 'printf'? [-Wimplicit-function-declaration]
       printk( fmt, ##__VA_ARGS__)
       ^
/opt/work/chris/rtems/kernel/rtems.git/cpukit/include/rtems/status-checks.h:86:3: note: in expansion of macro 'RTEMS_SYSLOG_PRINT'
   RTEMS_SYSLOG_PRINT( "%s: " fmt, __func__, ##__VA_ARGS__)
   ^~~~~~~~~~~~~~~~~~
/opt/work/chris/rtems/kernel/rtems.git/cpukit/include/rtems/status-checks.h:107:3: note: in expansion of macro 'RTEMS_SYSLOG'
   RTEMS_SYSLOG( "Error: " fmt, ##__VA_ARGS__)
   ^~~~~~~~~~~~
/opt/work/chris/rtems/kernel/rtems.git/cpukit/include/rtems/status-checks.h:113:3: note: in expansion of macro 'RTEMS_SYSLOG_ERROR'
   RTEMS_SYSLOG_ERROR( "SC = %i: %s\n", (int) sc, msg);
   ^~~~~~~~~~~~~~~~~~
/opt/work/chris/rtems/kernel/rtems.git/cpukit/include/rtems/status-checks.h:152:5: note: in expansion of macro 'RTEMS_SYSLOG_ERROR_WITH_SC'
     RTEMS_SYSLOG_ERROR_WITH_SC( sc, msg); \
     ^~~~~~~~~~~~~~~~~~~~~~~~~~
/opt/work/chris/rtems/kernel/rtems.git/c/src/lib/libbsp/lm32/milkymist/../../lm32/shared/milkymist_gpio/gpio.c:57:5: note: in expansion of macro 'RTEMS_CHECK_SC'
     RTEMS_CHECK_SC(sc, "create GPIO device");
     ^~~~~~~~~~~~~~
#3323 14 months ago fixed lib Chris Johns Chris Johns 13 months ago
Summary

mhttpd's http etag can result in invalid caching in a browser.

Description

The mhttp's http etag uses the mtime and file length and this can cause subtle issues if a target has no RTC or it is incorrect and files are being copied without preserving the mtime or changes happen that do not change the length.

The cp and untar code do not update a file's time.

Add support for an etag callback so a user can manage the tag, ie MD5 or something similar.

#3471 9 months ago fixed lib Sebastian Huber Sebastian Huber 9 months ago
Summary

Update libfdt as of date 2018-07-09

Description

The initial import of libfdt was in 2015. Update it to the version as of date 2018-07-09.

#3529 7 months ago fixed lib Sebastian Huber Sebastian Huber 6 months ago
Summary

Fix issues raised by Coverity Scan for Telnet server

#3535 7 months ago fixed lib Sebastian Huber Sebastian Huber 7 months ago
Summary

Remove stdin, stdout, stderr convenience routines for CEXP

Description

These functions should be moved to a general CEXP support.

#3542 7 months ago fixed lib Sebastian Huber Sebastian Huber 7 months ago
Summary

Remove keep_stdio feature from Telnet service

Description

The Telnet service started via rtems_telnetd_start() has a keep_stdio feature. This just a task and executes the command function in a loop. For this kind of service we do not library support. This can be done by an application task on its own. Remove this functionality and provide only the real Telnet services.

#3543 7 months ago fixed lib Sebastian Huber Sebastian Huber 6 months ago
Summary

Change Telnet server to allocate most resources during initialization

Description

The Telnet server currently creates the resources needed for a client connection on demand. Allocate most resources during initialization to avoid sporadic resource shortage issues.

#3545 6 months ago fixed lib Sebastian Huber Sebastian Huber 6 months ago
Summary

Support O_DIRECTORY open() flag

Description

Use this flag in opendir().

#3546 6 months ago fixed lib Sebastian Huber Sebastian Huber 6 months ago
Summary

Support O_NOFOLLOW open() flag

Description
#3547 6 months ago fixed lib Sebastian Huber Sebastian Huber 6 months ago
Summary

Support O_CLOEXEC open() flag

Description

This is a POSIX flag. Make sure its use causes no open failure.

#3551 6 months ago fixed lib Sebastian Huber Sebastian Huber 6 months ago
Summary

Move default configuration to separate library

Description

An RTEMS application default configuration is contained in cpukit/libmisc/dummy/default-configuration.c. This default configuration is contained in librtemscpu.a. This has at least two problems:

  1. Application configuration errors may pull in the default configuration which in turn leads to multiply define symbols error. This is quite confusing. You have to consult the linker map file to figure out what cased the pull in of the default configurations. You need to know what a linker map file is and how you generate it with your build system. This is not very user friendly.
  1. It prevents the use of default configuration items for each subsystem in librtemscpu.a. This can be used to reduce the size of the configuration itself.

Proposed change: Move the default configuration to a separate library, e.g. librtemsdefaultconfig.a.

#3666 4 months ago fixed lib Sebastian Huber Sebastian Huber 3 months ago
Summary

Add support for C++17 std::aligned_alloc

Description

In C++17 there is a std::aligned_alloc():

https://en.cppreference.com/w/cpp/memory/c/aligned_alloc

Unfortunately, it doesn't work with RTEMS currently:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85904

Provide aligned_alloc() and memalign() (as a strong alias to aligned_alloo()) by RTEMS.

#3684 3 months ago fixed lib Chris Johns Chris Johns 2 months ago
Summary

rtems_print_buffer is broken

Description

Consider this call:

#include <rtems/dumpbuf.h>
rtems_print_buffer ((const unsigned char *) "\x12\x23\x56\x78", 4);

On psim you get:

1f 2f 5f 7f                                     |.#Vx            |
#3358 13 months ago fixed lib/block Sebastian Huber Sebastian Huber 5 months ago
Summary

Deprecate rtems_disk_create_phys(), etc.

Description

There are currently two implementations of a block device (disk). Deprecate the legacy rtems_disk_create_phys(), etc. implementation. Remove all RTEMS internal uses except in the block01 test. Add RTEMS_DEPRECATED attribute to API.

#3510 8 months ago fixed lib/block Joel Sherrill Sebastian Huber 8 months ago
Summary

ATA driver uses deprecated rtems_blkdev services

Description

This shows up building fileio on the following BSPs.

i386/pc386 i386/pc486 i386/pc586 i386/pc586-sse i386/pc686 i386/pcp4 powerpc/brs5l powerpc/brs6l powerpc/dp2 powerpc/icecube powerpc/pm520_cr825 powerpc/pm520_ze30

#3733 3 weeks ago fixed lib/debugger Chris Johns Chris Johns 2 weeks ago
Summary

Add general reg support to libdebugger

Description

Testing master on a Zynq reports:

(gdb) target remote 10.10.5.45:1122                        
Remote debugging using 10.10.5.45:1122             
Truncated register 19 in remote 'g' packet

It looks to me like gdb is now smart enough to know this ARM arch has a NEON and floating point registers:

(gdb) maint print registers
Name         Nr  Rel Offset    Size  Type
 r0            0    0      0       4 uint32_t
 r1            1    1      4       4 uint32_t
 r2            2    2      8       4 uint32_t
 r3            3    3     12       4 uint32_t
 r4            4    4     16       4 uint32_t
 r5            5    5     20       4 uint32_t
 r6            6    6     24       4 uint32_t
 r7            7    7     28       4 uint32_t
 r8            8    8     32       4 uint32_t
 r9            9    9     36       4 uint32_t
 r10          10   10     40       4 uint32_t
 r11          11   11     44       4 uint32_t
 r12          12   12     48       4 uint32_t
 sp           13   13     52       4 *1
 lr           14   14     56       4 uint32_t
 pc           15   15     60       4 *1
 f0           16   16     64      12 _arm_ext
 f1           17   17     76      12 _arm_ext
 f2           18   18     88      12 _arm_ext
 f3           19   19    100      12 _arm_ext
 f4           20   20    112      12 _arm_ext
 f5           21   21    124      12 _arm_ext
 f6           22   22    136      12 _arm_ext
 f7           23   23    148      12 _arm_ext
 fps          24   24    160       4 uint32_t
 cpsr         25   25    164       4 uint32_t

The target support in libdebugger is a simple array of 32bit ints. This needs to change to handle registers at various offsets. The lack of fp regs was a simplification at the time I first implement this server. Loos like I need to sort this out.

#2754 3 years ago fixed lib/dl Patrick Gauvin Chris Johns 18 months ago
Summary

no .strtab section

Description

dlopen on the object generated by libfoo.cpp in the attached test case fails and results in the error no .strtab section. readelf shows that the section is present, though:

readelf -S libfoo.o | grep strtab
(standard input):97:  [92] .shstrtab         STRTAB          00000000 001fb0 00040c 00      0   0  1
(standard input):99:  [94] .strtab           STRTAB          00000000 0018b0 00019e 00      0   0  1

Steps to Reproduce (you may have to edit BSP_DIR in the Makefile):

make clean all
qemu-system-arm -m 256M -M xilinx-zynq-a9 -serial null -serial mon:stdio \
    -nographic -no-reboot -kernel libdl-strtab-test.exe

Expected Output:

TEST BEGIN
dlopen: no .strtab section
assertion "handle != NULL" failed: file "libdl-strtab-test.c", line 46, function: POSIX_Init

Development Environment:

  • RTEMS Version: 4.11 (Branch "4.11", commit 3f72dda6ee518d3ea04341ad4df079ecb1895ef7) with the dlerror patches from #2747, and the attached ARM PREL31 support patch (I will be making a separate ticket for this with test code soon).
  • System Type: ARM Cortex-A9, xilinx_zynq_a9_qemu BSP
  • GCC Version:

arm-rtems4.11-gcc (GCC) 4.9.3 20150626 (RTEMS 4.11, RSB 1675a733536d1aec2020011e5e522497a442561a (HEAD, origin/4.11, 4.11), Newlib 2.2.0.20150423)

  • RTEMS Configure Options:

--target=arm-rtems4.11 --enable-rtemsbsp="xilinx_zynq_a9_qemu xilinx_zynq_zedboard xilinx_zynq_csp_cots xilinx_zynq_csp_hybrid" --enable-tests=samples --enable-posix --prefix=$HOME/development/rtems/4.11 --disable-networking

#2802 2 years ago fixed lib/dl Sebastian Huber Chris Johns 18 months ago
Summary

Test "libdl (RTL) 5" fails on SPARC targets

Description

On GR740 I get:

rtl: RELOC_32 0x60ae8 @ 0x86edc in /dl-o5.o
rtl: relocation: .rela.eh_frame, syms:.symtab
rtl: rela: sym:__gxx_personality_v0(20)=00001dec type:3  off:00000013 addend:0

CPU 0:  IU in error mode (tt = 0x07, mem address not aligned)
        0x0001fa9c: c4040000  ld  [%l0], %g2  <rtems_rtl_elf_relocate_rela+204>
CPU 1:  Power down mode
CPU 2:  Power down mode
CPU 3:  Power down mode

On GR712RC I get:

rtl: WDISP_30 0x7ffe2ccd @ 0x40087108 in /dl-o5.o
rtl: relocation: .rela.gcc_except_table.exception_dl, syms:.symtab
rtl: rela: sym:_ZTISt9exception(32)=40060ae8 type:3  off:00000034 addend:0
rtl: RELOC_32 0x40060ae8 @ 0x400871b4 in /dl-o5.o
rtl: relocation: .rela.eh_frame, syms:.symtab
rtl: rela: sym:__gxx_personality_v0(20)=40001dec type:3  off:00000013 addend:0

Target resets now.

#2922 2 years ago fixed lib/dl Chris Johns Chris Johns 5 months ago
Summary

libdl unresolved externals that use more than one block or multiple entries corrupts.

Description

If there are lots of unresolved externals the unresolved block structure get confused and the can lock up looking for a new slot to add an unresolved external.

#2959 2 years ago fixed lib/dl Chris Johns Chris Johns 18 months ago
Summary

arm/libdl: C++ exception index tables may not be ordered correctly

Description

The ARM EXIDX sections have the SHF_LINK_ORDER flag set and this is not honored by libdl which means the section order in the ELF file needs to be the correct order of the functions in the address map.

Add support to libdl to follow the link-to order.

#3298 15 months ago fixed lib/dl Chris Johns Chris Johns 2 months ago
Summary

dlerror non-conformance

Description

This is a port of the 4.11 patches from #2747 to master. Please refer to that ticket for details.

#3604 5 months ago fixed lib/dl Joseph Hickey Chris Johns 5 months ago
Summary

RTL Unresolved Symbols from common section on i386/pc686 (cloned)

Description

Cloned from #3527:


By default GCC puts uninitialized global variables into a common section in the ELF file. When attempting to load the resulting ELF file at runtime using dlopen(), these global symbols are not resolved as expected.

The RTL reports unresolved symbols, and runtime code that take the address of the global get NULL instead.

This is reproducible using the libtests/dl01 example by adding a global variable to the module code. I will attach a patch that replicates the issue.

Test platform is QEMU using pc686 BSP, RTEMS source version 4.11.3 (latest on 4.11 git branch as of this writing)

#3612 5 months ago fixed lib/dl Chris Johns Chris Johns 5 months ago
Summary

RTL unresolved compaction does not update string indexes after removing a string

Description

The RTL unresolved compaction does not update the string indexes when compacting.

#3625 5 months ago fixed lib/dl Kevin Gordon Chris Johns 5 months ago
Summary

RTL Allows Unloading a Module other Modules Depend Upon (cloned)

Description

Cloned from #3195:


Consider the following example using ELF .o files from compiled source files module-0.c and module-1.c from ticket #3194:

module-0.o is loaded using dlopen() with no code or data dependencies. module-1.o is loaded using dlopen() with both code and data dependencies on module-0 which are resolved by RTL.

The RTL function dlcose() returns no error when module-0 is unloaded, when it should return an error and not unload module-0. This becomes quite dangerous because a subsequent call to module1Function1() in the currently-loaded module-1.o, which accesses shared_resource_0[ ] and calls module0Function0(), will result in an unexpected trap on qemu or the call succeeding with the correct return value on hardware when it should not.

The erroneous successful unload() of module-0 aside, it appears as though the resources are not actually deleted and I believe this ticket is related to tickets #3192 and #3194.

Architecture is sparc-leon3 using both the RTEMS 4.11.1 public release and rtems master @f043b9bd3bf25626fb1a311dd7fa041eacc68adc with rtems-source-builder @55f2d69e9b67cde23d61375fa34ef5b0f04a985d.

Execution environments are qemu-system-sparc and LEON3 UT700 hardware.

#3692 2 months ago fixed lib/dl Chris Johns Chris Johns 2 months ago
Summary

libdl does not honour write unlock/lock for sections

Description

The allocator does no honour write unlock and lock for read-only sections as it should. This can used to write protect executable memory.

#2633 3 years ago fixed network/legacy joguin Sebastian Huber 18 months ago
Summary

waf build failed for rtems-libbsd

Description

The rtems-libbsd failed when building with waf. Here is the output:

[488/845] Compiling freebsd/sys/kern/subr_taskqueue.c In file included from /home/josh/development/rtems/bsps/4.12/i386-rtems4.12/pc386/lib/include/rtems/score/threadimpl.h:36:0,

from ../../freebsd/sys/kern/subr_sleepqueue.c:91:

/home/josh/development/rtems/bsps/4.12/i386-rtems4.12/pc386/lib/include/rtems/score/watchdogimpl.h: In function '_Watchdog_Per_CPU_insert_relative': /home/josh/development/rtems/bsps/4.12/i386-rtems4.12/pc386/lib/include/rtems/score/watchdogimpl.h:356:18: error: 'struct <anonymous>' has no member named '_bsd_ticks'; did you mean 'ticks'?

cpu->Watchdog.ticks + ticks

In file included from ../../freebsd/sys/kern/subr_sleepqueue.c:62:0: ../../freebsd/sys/kern/subr_sleepqueue.c: In function 'sleepq_set_timeout': ../../freebsd/sys/kern/subr_sleepqueue.c:424:29: error: 'Thread_Timer_information {aka struct <anonymous>}' has no member named 'state'

BSD_ASSERT(executing->Timer.state == WATCHDOG_INACTIVE);

../../freebsd/sys/kern/subr_sleepqueue.c:424:2: note: in expansion of macro 'BSD_ASSERT'

BSD_ASSERT(executing->Timer.state == WATCHDOG_INACTIVE); ~

../../freebsd/sys/kern/subr_sleepqueue.c:425:2: error: too many arguments to function '_Watchdog_Initialize'

_Watchdog_Initialize(&executing->Timer, sleepq_timeout, ~

In file included from /home/josh/development/rtems/bsps/4.12/i386-rtems4.12/pc386/lib/include/rtems/score/threadimpl.h:36:0,

from ../../freebsd/sys/kern/subr_sleepqueue.c:91:

/home/josh/development/rtems/bsps/4.12/i386-rtems4.12/pc386/lib/include/rtems/score/watchdogimpl.h:178:27: note: declared here

RTEMS_INLINE_ROUTINE void _Watchdog_Initialize(

~

Waf: Leaving directory `/home/josh/development/rtems/rtems-libbsd/build/i386-rtems4.12-pc386' Build failed

#2694 3 years ago worksforme network/legacy Joel Sherrill Sebastian Huber 18 months ago
Summary

linking issue for htonl, etc when using -std=c99

Description

When -std=c99 is on the compile line, there is a linking error for undefined references to htonl, htons, ntohl, and ntohs. This test case is just for htonl but others should be similar.

This likely impacts the 4.11 branch of rtems-libbsd as well but I was testing on master.

Test case ====================== #include <arpa/inet.h>

int main(

int argc, char argv

) {

uint32_t v = (uint32_t) argc; uint32_t rc = htonl(v); return v;

} ======================

This script was what I used to find what caused the linking error to go away.

====================== RTEMS_MAKEFILE_PATH=/home/joel/rtems-4.11-work/tools/4.12/i386-rtems4.12/pc586/

i386-rtems4.12-gcc -std=c99 \

-B${RTEMS_MAKEFILE_PATH}/lib -specs bsp_specs -qrtems \ -D_XOPEN_SOURCE=600 -DUSE_SVID main.c -lbsd -lm -lbsd

i386-rtems4.12-gcc -std=c99 \

-B${RTEMS_MAKEFILE_PATH}/lib -specs bsp_specs -qrtems \ -DUSE_SVID main.c -lbsd -lm -lbsd

i386-rtems4.12-gcc -std=c99 \

-B${RTEMS_MAKEFILE_PATH}/lib -specs bsp_specs -qrtems \ main.c -lbsd -lm -lbsd

i386-rtems4.12-gcc \

-B${RTEMS_MAKEFILE_PATH}/lib -specs bsp_specs -qrtems \ main.c -lbsd -lm -lbsd

======================

#2748 3 years ago fixed network/legacy Sebastian Huber Sebastian Huber <sebastian.huber@…> 18 months ago
Summary

Move RTEMS-specific socket wake-up to RTEMS-specific <rtems/rtems_bsdnet.h>

Description

The <sys/socket.h> contains definitions for the RTEMS-specific socket wake-up support. Move this stuff to <rtems/rtems_bsdnet.h> since this feature is not present in standard network stacks. Portable applications should not use it.

#3390 13 months ago fixed network/legacy Sebastian Huber Sebastian Huber 13 months ago
Summary

NFS: Remove support for cexp

Description

There is some support for cexp and tests in the NFS client directory:

cpukit/libfs/src/nfsclient/src/cexphelp.c cpukit/libfs/src/nfsclient/src/dirutils.c cpukit/libfs/src/nfsclient/src/nfs.modini.c cpukit/libfs/src/nfsclient/src/nfsTest.c cpukit/libfs/src/nfsclient/src/rpcio.modini.c

There are also some *.rel files installed. This stuff is probably unused. If it is still in use it should move elsewhere, e.g. some general cexp support outside of the main RTEMS sources. Dead/untested code should not be present in the RTEMS code base.

See also:

https://lists.rtems.org/pipermail/users/2018-April/032182.html

#3419 12 months ago fixed network/legacy Sebastian Huber Sebastian Huber 12 months ago
Summary

Always build network services (tftpfs, ftpfs, ftpd, telnetd, libdebugger)

Description

Always build network services (tftpfs, ftpfs, ftpd, telnetd, libdebugger) which only depend on the POSIX socket API (provided by Newlib header files) as support libraries. Remove them from libbsd.

The network services must reside in dedicated libraries to avoid a cyclic dependency between libbsd.a and librtemscpu.a.

#3277 15 months ago fixed network/libbsd Sebastian Huber Sebastian Huber 15 months ago
Summary

QorIQ: Add MAC-less DPAA driver to libbsd

Description

The SDK Linux DPAA driver supports a so called MAC-less interface driver. This driver allows Ethernet communication between guest systems of a hypervisor.

#3472 9 months ago fixed network/libbsd Sebastian Huber Sebastian Huber 3 months ago
Summary

Update of libbsd to a version close to the FreeBSD 12 release

Description

The FreeBSD project is about to prepare the FreeBSD 12 release soon:

https://www.freebsd.org/releases/12.0R/schedule.html

Use this time frame to update the libbsd stepwise to a FreeBSD trunk version close to the FreeBSD 12 release.

#3522 7 months ago fixed network/libbsd Sebastian Huber Sebastian Huber 7 months ago
Summary

Update mDNSResponder to Apple version v878.30.4

Description

Download

mDNSResponder-561.1.1.tar.gz mDNSResponder-567.tar.gz mDNSResponder-576.30.4.tar.gz mDNSResponder-625.41.2.tar.gz mDNSResponder-765.1.2.tar.gz mDNSResponder-765.20.4.tar.gz mDNSResponder-765.30.11.tar.gz mDNSResponder-765.50.9.tar.gz mDNSResponder-878.1.1.tar.gz mDNSResponder-878.20.3.tar.gz mDNSResponder-878.30.4.tar.gz

from

https://opensource.apple.com/tarballs/mDNSResponder/

Merge each update into libbsd/mDNSResponder.

#3523 7 months ago fixed network/libbsd Sebastian Huber Sebastian Huber 7 months ago
Summary

Add FEC network interface driver for TQM8XX

Description

Import legacy network driver and port it to libbsd.

#3554 6 months ago fixed network/libbsd Joel Sherrill Sebastian Huber 3 months ago
Summary

rtems-libbsd README.waf Needs an Update Sweep

Description

It is out of date at least by mentioning 4.12 instead of 5. If there are other nits or issues, they need to be addressed while updating the release info.

#2135 6 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

times() and _times() are subject to integer overflows

Description

The clock_t type is defined as unsigned long for RTEMS in Newlib. With a 1ms clock tick an overflow happens after 7 days on 32-bit long targets.

#2514 3 years ago fixed posix Sebastian Huber Sebastian Huber 6 months ago
Summary

Make POSIX API mandatory (except signals and the sporadic server)

Description

The POSIX API is currently a build-time configuration option. In general it is beneficial to avoid build-time configuration options since this reduces the testing scope.

Applications not using the POSIX API should observe only a minimal overhead due to this change.

This enhancement depends on #2408.

#2606 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

alarm() uses seconds watchdog and thus is affected by clock changes

Description

alarm() uses _Watchdog_Insert_seconds() and thus is affected by clock changes, e.g. via _TOD_Set(). This is wrong. The POSIX documentation is not that clear since it talks only about "realtime seconds". However, the FreeBSD implementation uses the uptime. This is also in line with the RTEMS ualarm() and nanosleep().

#2608 3 years ago fixed posix Joel Sherrill Gedare 17 months ago
Summary

POSIX Condition Variables Clock Attribute Support

Description

I am beginning to add support for the clock attribute to POSIX condition variables.

http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_condattr_getclock.html

Since the clock can't be a CPU time clock, that leaves CLOCK_MONOTONIC and CLOCK_REALTIME. The thread queue is based on CLOCK_MONOTONIC and does not have an option to use CLOCK_REALTIME. Threads and timers waiting on CLOCK_REALTIME should be impacted by time of day changes.

https://docs.google.com/document/d/1GsGer0t84p-nUfZFim4Ty0LTDYNhgKBvlwip_gLQjTY/edit?usp=sharing is a Google doc with my notes so far in it on POSIX clocks. I will move it to the Wiki as it turns into something more concrete than notes and reflects plans/code.

So the first issue is how best to alter the thread queue to support using either clock source? And what does that do to the current ticks based API since you proposed different time representations for the ticks (relative/monotonic) and seconds (absolute/realtime) structures?

#2625 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Use one lookup tree per-thread for the POSIX keys

Description

Currently a global lookup tree is used for all the POSIX key/value pairs. On SMP configurations this is a bottleneck. Use one lookup tree per thread instead.

#2674 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

CORE spinlock implementation is next to be useless

Description

Lets assume we have two tasks A and B. Task A acquires a CORE spinlock. Now B somehow executes and tries to acquire the same CORE spinlock, then no progress can be made.

Alternative implementation:

Disable thread dispatching and interrupts while owning the spinlock. Forbid blocking calls while owning the spinlock.

Drawback: The test cases of the Linux Test Project would fail:

https://github.com/linux-test-project/ltp/blob/master/testcases/open_posix_testsuite/conformance/interfaces/pthread_spin_lock/1-2.c

Optimization: User provided storage space for pthread_spin_t. In line with POSIX:

"Only the object referenced by lock may be used for performing synchronization."

http://pubs.opengroup.org/onlinepubs/009695399/functions/pthread_spin_destroy.html

#2689 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

POSIX key destructors must be called during thread restart

Description

POSIX key destructors must be called during thread restart. Just like the POSIX cleanup handlers. This ensures that the TLS object destructors are called during thread restart for example. It is important for the global construction, which uses a thread restart to run the Init task in a clean environment.

#2702 3 years ago fixed posix Sebastian Huber Sebastian Huber 16 months ago
Summary

Remove descriptor objects for POSIX message queues

Description

The mq_open() function returns a descriptor to a POSIX message queue object identified by a name. This is similar to sem_open(). In contrast to the POSIX semaphore the POSIX message queues use a separate object for the descriptor. This extra object is superfluous, since the object identifier can be used directly for this purpose, just like for the semaphores.

#2714 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

A pthread_detach() does not lead to a resource reclamation

Description

According to POSIX a pthread_detach() should lead to a resource reclamation if the thread is already cancelled.

#2732 3 years ago fixed posix Gedare Gedare 18 months ago
Summary

Add clock_nanosleep()

Description

The clock_nanosleep function is provided to enable specifying the clock source (CLOCK_REALTIME or CLOCK_MONOTONIC) and to control whether or not to use an absolute or relative reference point via TIMER_ABSTIME flag.

See also: http://pubs.opengroup.org/onlinepubs/009695399/functions/clock_nanosleep.html

#2734 3 years ago fixed posix Sebastian Huber Sebastian Huber <sebastian.huber@…> 18 months ago
Summary

pthread_setschedprio() is missing

Description

See also

http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_setschedprio.html

and

http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_08_04_01

In particular the distinction to pthread_setschedparam() (SCHED_FIFO, item 7.).

Prototype is defined in Newlib provide <pthread.h>.

#2735 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

pthread_setschedparam() sets the priority not according to POSIX

Description

See also

http://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_08_04_01

In particular the distinction to pthread_setschedprio() (SCHED_FIFO, item 7.).

#2736 3 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

pthread_getschedparam() returns wrong priority values

Description

See also

http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_getschedparam.html

"The priority value returned from pthread_getschedparam() shall be the value specified by the most recent pthread_setschedparam(), pthread_setschedprio(), or pthread_create() call affecting the target thread. It shall not reflect any temporary adjustments to its priority as a result of any priority inheritance or ceiling functions."

#2745 3 years ago fixed posix Gedare Gedare 18 months ago
Summary

Use clock from pthread_condattr in pthread_cond_timedwait

Description

For pthread_cond_timedwait, the condition variable shall have a clock attribute which specifies the clock that shall be used to measure the time specified by the abstime argument. RTEMS currently does not honor the clock attribute.

See http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_cond_timedwait.html

#2798 2 years ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Fix POSIX timer interval

Description

See also:

https://lists.rtems.org/pipermail/users/2016-October/030714.html

Just back to RTEMS after a long time. I've build the last version (Git) 
with last compiler (4.12) and looks like a simple POSIX timer  doen't 
work anymore... It should display "Signal 14" every second but period is 
veryyyy short...and display very fast :(

Any idea? any change in the way to use POSIX with RTEMS?

See my program attached (it works fine with Linux).

I have tested with Raspberry Pi and QEMU/i386.

thx by advance
#2859 2 years ago fixed posix Gedare Gedare 18 months ago
Summary

Implement POSIX Shared Memory Objects

Description

POSIX Shared Memory is a widely used API for inter-process communication. The functions in the API include:

#3112 20 months ago fixed posix Sebastian Huber Sebastian Huber 17 months ago
Summary

POSIX: Make pthread_mutex_t self-contained

Description

Change the POSIX mutex into a self-contained object using <sys/lock.h>, e.g.

typedef struct {
  struct _Mutex_recursive_Control  _mutex;
  unsigned int                     _flags;
  struct _Scheduler_Control       *_scheduler;
  __uint64_t                       _priority_ceiling;
} pthread_mutex_t;
#3113 20 months ago fixed posix Sebastian Huber Sebastian Huber 17 months ago
Summary

POSIX: Make pthread_cond_t self-contained

Description

Change the POSIX condition variable into a self-contained object using <sys/lock.h>, e.g.

typedef struct {
  struct _Condition_Control  _condition;
  pthread_mutex_t           *_mutex;
  clockid_t                  _clock;
} pthread_cond_t;
#3114 20 months ago fixed posix Sebastian Huber Sebastian Huber 17 months ago
Summary

POSIX: Make pthread_barrier_t self-contained

Description

Change the POSIX barrier into a self-contained object using <sys/lock.h>, e.g.

typedef struct {
  struct _Thread_queue_Queue  _queue;
  unsigned int                _flags;
  unsigned int                _count;
} pthread_barrier_t;
#3115 20 months ago fixed posix Sebastian Huber Sebastian Huber 17 months ago
Summary

POSIX: Make pthread_rwlock_t self-contained

Description

Change the POSIX read-write lock into a self-contained object using <sys/lock.h>, e.g.

typedef struct {
  struct _Thread_queue_Queue _queue;
  unsigned int               _flags;
  unsigned int               _readers;
} pthread_rwlock_t;
#3116 20 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

POSIX: Make sem_t self-contained

Description

Change the POSIX semaphore into a self-contained object using <sys/lock.h>, e.g.

typedef struct {
  struct _Semaphore_Control _sem;
} sem_t;
#3125 20 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Accept PTHREAD_PROCESS_SHARED for POSIX mutexes

Description

Since we have only one process, sharing between processes is trivial.

#3126 20 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Accept PTHREAD_PROCESS_SHARED for POSIX barriers

Description

Since we have only one process, sharing between processes is trivial.

#3137 20 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Accept PTHREAD_PROCESS_SHARED for POSIX condition variables

Description

Since we have only one process, sharing between processes is trivial.

#3142 19 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

POSIX: Reduce size of pthread_once_t and make it zero-initialized

Description

A zero-initialized pthread_once_t reduces the ROM usage of RTEMS applications, since the global pthread_once_t objects may reside in the BSS section.

#3148 19 months ago fixed posix Chris Johns joel.sherrill@… 18 months ago
Summary

PSXRDWRV Test failure on Beaglebone Black

Description

Running rtems-test with a recent u-boot and a current master this failure is reported:

] RTEMS Beagleboard: am335x-based
]
]
] *** BEGIN OF TEST PSXRDWRV ***
] writev bad file descriptor -- EBADF
] writev error 1: 22=Invalid argument
] Error during error test!!!!
#3153 19 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Accept PTHREAD_PROCESS_SHARED for POSIX rwlocks

Description

Since we have only one process, sharing between processes is trivial.

#3168 19 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Simplify POSIX_API_Control

Description

There is no need to have a copy of the thread attributes used for the pthread_create() in POSIX_API_Control::Attributes. This is at least in line with Linux.

#3174 19 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Remove rtems_pthread_attribute_compare()

Description

The rtems_pthread_attribute_compare() function is undocumented and used only in one test. Move it to the test.

#3179 19 months ago fixed posix Joel Sherrill Sebastian Huber 18 months ago
Summary

New warnings from Time Changes

Description

New warnings after picking up your recent commits. How are you checking for warnings?

../../../../../../rtems/c/src/../../cpukit/posix/src/pthreadattrdefault.c:58:5: warning: initialization discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
     &_POSIX_Threads_Default_attributes.affinitysetpreallocated,
     ^
../../../../../../rtems/c/src/../../cpukit/posix/src/adjtime.c: In function 'adjtime':
../../../../../../rtems/c/src/../../cpukit/posix/src/adjtime.c:85:16: warning: passing argument 1 of '_TOD_Adjust' from incompatible pointer type [-Wincompatible-pointer-types]
   _TOD_Adjust( &delta_as_timestamp );
                ^
In file included from ../../../../../../rtems/c/src/../../cpukit/posix/src/adjtime.c:28:0:
../../cpukit/../../../erc32/lib/include/rtems/score/todimpl.h:287:6: note: expected 'const struct timespec *' but argument is of type 'Timestamp_Control * {aka long long int *}'
 void _TOD_Adjust(
      ^~~~~~~~~~~
sparc-rtems4.12-ar: `u' modifier ignored since `D' is the default (see `U')
sparc-rtems4.12-ar: `u' modifier ignored since `D' is the default (see `U')
sparc-rtems4.12-ar: `u' modifier ignored since `D' is the default (see `U')
sparc-rtems4.12-ar: `u' modifier ignored since `D' is the default (see `U')
../../../../../../rtems/c/src/../../cpukit/libcsupport/src/__times.c: In function '_times':
../../../../../../rtems/c/src/../../cpukit/libcsupport/src/__times.c:60:31: warning: passing argument 1 of '_TOD_Get_zero_based_uptime' from incompatible pointer type [-Wincompatible-pointer-types]
   _TOD_Get_zero_based_uptime( &binuptime );
                               ^
In file included from ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/__times.c:35:0:
../../cpukit/../../../erc32/lib/include/rtems/score/todimpl.h:215:20: note: expected 'Timestamp_Control * {aka long long int *}' but argument is of type 'struct bintime *'
 static inline void _TOD_Get_zero_based_uptime(
                    ^~~~~~~~~~~~~~~~~~~~~~~~~~
../../../../../../rtems/c/src/../../cpukit/libcsupport/src/__times.c:71:55: warning: passing argument 2 of '_Thread_Get_CPU_time_used' from incompatible pointer type [-Wincompatible-pointer-types]
   _Thread_Get_CPU_time_used( _Thread_Get_executing(), &bin_cpu_time_used );
                                                       ^
In file included from ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/__times.c:37:0:
../../cpukit/../../../erc32/lib/include/rtems/score/threadimpl.h:906:6: note: expected 'Timestamp_Control * {aka long long int *}' but argument is of type 'struct bintime *'
 void _Thread_Get_CPU_time_used(
      ^~~~~~~~~~~~~~~~~~~~~~~~~
#3182 19 months ago fixed posix Sebastian Huber Sebastian Huber 13 months ago
Summary

CLOCK_REALTIME timeout implementation is not POSIX compliant

Description

Changes of the CLOCK_REALTIME must trigger absolute timeouts. This is not the case. There is no test for this in RTEMS.

#3205 18 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Relative timespec timeouts are subject to integer overflows

Description

As a best-effort approach, a very large relative timeout should result in the maximum monotonic watchdog value and not in an undefined integer overflow.

#3211 18 months ago fixed posix Sebastian Huber Sebastian Huber 18 months ago
Summary

Fix pthread_create() with user provided stack

Description

In case the user provides a stack with address and size, then do not alter the stack size.

#3265 16 months ago fixed posix Sebastian Huber Sebastian Huber 15 months ago
Summary

Use second one based uptime for CLOCK_MONOTONIC for FreeBSD compatibility

Description

This simplifies the CLOCK_MONOTONIC based time services. It is potentially important for libbsd.

#3334 14 months ago fixed posix Stavros Passas Sebastian Huber 2 months ago
Summary

deadlock in _once()

Description

RTEMS threads getting locked up when using certain c++ functionality. Issue happens for example when std::future is combined with std::async.

Investigating deeper, seems like this happens if std::async executes before std::future gets scheduled to run. Both of these create a pthread_once instance.

_once() uses a common semaphore for all calls, thus the first function (async.get usually) gets the lock, calls its “init” function (which blocks until the second function has completed. After this, std::future also uses pthread_once to execute, but because the lock is already taken, it also blocks, casing a deadlock.

Attached you can find a test application that reproduces the deadlock.

#3342 14 months ago fixed posix Joel Sherrill Joel Sherrill 9 months ago
Summary

pthread_setschedparam() has incorrect prototype

Description

We are missing the const on the third parameter. This requires a change to newlib and RTEMS. The correct prototype is:

int pthread_setschedparam(

pthread_t thread, int policy, const struct sched_param *param

)

#3343 14 months ago fixed posix Joel Sherrill Joel Sherrill 9 months ago
Summary

pthread_mutex_getprioceiling() has incorrect prototype

Description

We are missing the const and restrict on the first parameter. This requires a change to newlib and RTEMS. The correct prototype is:

int pthread_mutex_getprioceiling(

const pthread_mutex_t *restrict mutex, int *prioceiling

)

#3491 9 months ago fixed posix Sebastian Huber Sebastian Huber 8 months ago
Summary

Align mprotect() prototype with POSIX

Description

The correct prototype is:

int	mprotect(void *, size_t, int);
#3626 5 months ago fixed posix Joel Sherrill Sebastian Huber 5 months ago
Summary

sigtimedwait() needed when POSIX is disabled

Description

When POSIX is disabled, psxmsgq01 does not link. Should we enable sigtimedwait() when POSIX is disabled or disable this test?

/data/home/joel/rtems-work/tools/5/bin/../lib/gcc/powerpc-rtems5/7.3.0/../../../../powerpc-rtems5/bin/ld: psxmsgq01/psxmsgq01-init.o: in function `wait_for_signal':
/home/joel/rtems-work/rtems-testing/rtems/build-powerpc-ss555-rtems/powerpc-rtems5/c/ss555/testsuites/psxtests/../../../../../../rtems/c/src/../../testsuites/psxtests/psxmsgq01/init.c:932: undefined reference to `sigtimedwait'
#2459 3 years ago fixed rtems Sebastian Huber Sebastian Huber 18 months ago
Summary

Add rtems_chain_get_first_unprotected() to chain API

#2749 3 years ago fixed rtems Sebastian Huber Sebastian Huber 18 months ago
Summary

rtems_task_set_scheduler() has insufficient parameters

Description

Task priorities are only valid within a scheduler instance. The rtems_task_set_scheduler() directive moves a task from one scheduler instance to another using the current priority of the thread. However, the current task priority of the source scheduler instance is undefined in the target scheduler instance. Add a third parameter to specify the priority.

/**
 * @brief Sets the scheduler instance of a task.
 *
 * Initially, the scheduler instance of a task is set to the scheduler instance
 * of the task that created it.  This directive allows to move a task from its
 * current scheduler instance to another specified by the scheduler identifier.
 *
 * @param[in] task_id Identifier of the task.  Use @ref RTEMS_SELF to select
 *   the executing task.
 * @param[in] scheduler_id Identifier of the scheduler instance.
 * @param[in] priority The task priority with respect to the new scheduler
 *   instance.  The real and initial priority of the task is set to this value.
 *   The initial priority is used by rtems_task_restart() for example.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_ILLEGAL_ON_REMOTE_OBJECT Directive is illegal on remote tasks.
 * @retval RTEMS_INVALID_ID Invalid task or scheduler identifier.
 * @retval RTEMS_INVALID_PRIORITY Invalid priority.
 * @retval RTEMS_RESOURCE_IN_USE The task owns resources which deny a scheduler
 *   change.
 *
 * @see rtems_scheduler_ident().
 */
rtems_status_code rtems_task_set_scheduler(
  rtems_id            task_id,
  rtems_id            scheduler_id,
  rtems_task_priority priority
);
#2784 3 years ago fixed rtems Sebastian Huber Sebastian Huber 18 months ago
Summary

Add function to get the current priority of a task by scheduler instance

Description
/**
 * @brief Gets the current priority of the specified task with respect to the
 * specified scheduler instance.
 *
 * The current priority reflects temporary priority adjustments due to locking
 * protocols, the rate-monotonic objects on some schedulers and other
 * mechnisms.
 *
 * @param[in] task_id Identifier of the task.  Use @ref RTEMS_SELF to select
 *   the executing task.
 * @param[in] scheduler_id Identifier of the scheduler instance.
 * @param[out] priority Returns the current priority of the specified task with
 *   respect to the specified scheduler instance.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_ILLEGAL_ON_REMOTE_OBJECT Directive is illegal on remote tasks.
 * @retval RTEMS_INVALID_ADDRESS The priority parameter is @c NULL.
 * @retval RTEMS_INVALID_ID Invalid task or scheduler identifier.
 * @retval RTEMS_NOT_DEFINED The task has no priority within the specified
 *   scheduler instance.
 *
 * @see rtems_scheduler_ident().
 */
rtems_status_code rtems_task_get_priority(
  rtems_id             task_id,
  rtems_id             scheduler_id,
  rtems_task_priority *priority
);
#2797 2 years ago fixed rtems Sebastian Huber Sebastian Huber 18 months ago
Summary

Add ability to add/remove processors to/from a scheduler instance

Description

The scheduler configuration is done at link-time. In order to support run-time re-configuration add functions to dd/remove processors to/from a scheduler instance.

/**
 * @brief Adds a processor the set of processors owned by the scheduler.
 *
 * Must be called from task context.  This operation obtains and releases the
 * objects allocator lock.
 *
 * @param[in] scheduler_id Identifier of the scheduler.
 * @param[in] cpu_index Index of the processor to add.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_INVALID_ID Invalid scheduler identifier.
 * @retval RTEMS_NOT_CONFIGURED The processor is not configured to be used by
 *   the application.
 * @retval RTEMS_INCORRECT_STATE The processor is configured to be used by
 *   the application, however, it is not available.
 */
rtems_status_code rtems_scheduler_add_processor(
  rtems_id scheduler_id,
  uint32_t cpu_index
);

/**
 * @brief Removes a processor from set of processors owned by the scheduler.
 *
 * Must be called from task context.  This operation obtains and releases the
 * objects allocator lock.  Removing a processor from a scheduler is a complex
 * operation that involves all tasks in the system.
 *
 * @param[in] scheduler_id Identifier of the scheduler.
 * @param[in] cpu_index Index of the processor to add.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_INVALID_ID Invalid scheduler identifier.
 * @retval RTEMS_INVALID_NUMBER The processor is not owned by the scheduler.
 * @retval RTEMS_RESOURCE_IN_USE The set of processors owned by the scheduler
 *   would be empty after the processor removal and there exists a non-idle
 *   task that uses this scheduler as its home scheduler.
 */
rtems_status_code rtems_scheduler_remove_processor(
  rtems_id scheduler_id,
  uint32_t cpu_index
);
#3069 22 months ago fixed rtems Sebastian Huber Sebastian Huber 18 months ago
Summary

Add rtems_scheduler_ident_by_processor()

Description
/**
 * @brief Identifies a scheduler by a processor index.
 *
 * @param[in] cpu_index The processor index.
 * @param[out] id The scheduler identifier associated with the processor index.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_INVALID_ADDRESS The @a id parameter is @c NULL.
 * @retval RTEMS_INVALID_NAME Invalid processor index.
 * @retval RTEMS_INCORRECT_STATE The processor index is valid, however, this
 *   processor is not owned by a scheduler.
 */
#3070 22 months ago fixed rtems Sebastian Huber Sebastian Huber 18 months ago
Summary

Add rtems_scheduler_ident_by_processor_set()

Description
/**
 * @brief Identifies a scheduler by a processor set.
 *
 * The scheduler is selected according to the highest numbered online processor
 * in the specified processor set.
 *
 * @param[in] cpusetsize Size of the specified processor set buffer in
 *   bytes.  This value must be positive.
 * @param[out] cpuset The processor set to identify the scheduler.
 * @param[out] id The scheduler identifier associated with the processor set.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_INVALID_ADDRESS The @a id parameter is @c NULL.
 * @retval RTEMS_INVALID_SIZE Invalid processor set size.
 * @retval RTEMS_INVALID_NAME The processor set contains no online processor.
 * @retval RTEMS_INCORRECT_STATE The processor set is valid, however, the
 *   highest numbered online processor in the specified processor set is not
 *   owned by a scheduler.
 */
rtems_status_code rtems_scheduler_ident_by_processor_set(
  size_t     cpusetsize,
  cpu_set_t *cpuset,
  rtems_id  *id
);
#3235 17 months ago fixed rtems Sebastian Huber Sebastian Huber 17 months ago
Summary

Fix rtems_semaphore_flush() for priority inheritance semaphores

Description

The _Semaphore_Get_operations() must return the proper operations for priority inheritance semaphores.

Add a test case for rtems_semaphore_flush() with priority inheritance.

#3244 17 months ago fixed rtems Sebastian Huber Sebastian Huber 11 months ago
Summary

Change rtems_panic() implementation and document this function

Description

The current rtems_panic() implementation is quite heavy weight. It depends on _exit() which calls the global destructors. It uses fprintf(stderr, ...) for output which depends on an initialized console device and the complex fprintf().

Introduce a new fatal source RTEMS_FATAL_SOURCE_PANIC for rtems_panic() and output via printk().

Document this function in Fatal Manager chapter.

Replace all BSP_panic() with rtems_panic().

#3350 13 months ago fixed rtems Joel Sherrill Sebastian Huber 13 months ago
Summary

sptimecounter02 warning due to defining _KERNEL and disabling part of <sys/time.h>

Description

The bottom of <sys/time.h> is protected by ifndef _KERNEL where gettimeofday() is prototyped. sptimecounter02 is the only test which trips this.

In file included from /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/confdefs.h:323:0,

from ../../../../../../../rtems/c/src/../../testsuites/sptests/sptimecounter02/init.c:268:

/home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/imfs.h: In function 'IMFS_update_atime': /home/joel/rtems-work/rtems-testing/rtems/rtems/cpukit/include/rtems/imfs.h:345:3: warning: implicit declaration of function 'gettimeofday' [-Wimplicit-function-declaration]

gettimeofday( &now, 0 ); ~

#3482 9 months ago fixed rtems Sebastian Huber Sebastian Huber 9 months ago
Summary

Relax the buffer alignment required by rtems_partition_create()

Description

Buffer alignment required by rtems_partition_create() is too strict since it is checked via _Addresses_Is_aligned() which is

RTEMS_INLINE_ROUTINE bool _Addresses_Is_aligned (
  const void *address
)
{
#if (CPU_ALIGNMENT == 0)
    return true;
#else
    return (((uintptr_t)address % CPU_ALIGNMENT) == 0);
#endif
}

The CPU_ALIGNMENT must take long double and vector data type alignment requirements into account. For the partition maintenance only pointer alignment is required. The user should ensure that its buffer is suitable for the items it wants to manage. The user should not be burdened to provide buffers with the maximum architecture alignment, e.g. why need a 16 byte aligned buffer if you want to manage items with 4 byte integers only?

#3486 9 months ago fixed rtems Sebastian Huber Sebastian Huber 9 months ago
Summary

Use uintptr_t and size_t instead of uint32_t in rtems_partition_create()

Description

Use uintptr_t to specify the length of the partition buffer area instead of uint32_t. This is in line with rtems_region_create(). On 64-bit targets, the length may exceed 4GiB. Use size_t for the buffer size, since on some targets the single object size is less than the overall address range, e.g. m32c sizeof(uintptr_t) > sizeof(size_t).

#3533 7 months ago fixed rtems Sebastian Huber Sebastian Huber 7 months ago
Summary

Add rtems_task_exit()

Description

The rtems_task_delete(RTEMS_SELF) function does not return. In order to aid compilers and static analysis tools provide an rtems_task_exit() function which can be specified as a no return function.

void rtems_task_exit(void) RTEMS_NO_RETURN;

This is similar to the POSIX equivalent.

void pthread_exit (void *__value_ptr) __dead2;
#3583 6 months ago fixed rtems Sebastian Huber Sebastian Huber 5 months ago
Summary

Add rtems_malloc() and rtems_calloc()

Description

The standard C/POSIX functions malloc() and calloc() set errno in case of an error. A dependency to errno pulls in getreent() which pulls in a lot of data structures and functions. This is an issue in low level code especially in the area of a basic board support package initialization and device drivers.

Provide rtems_malloc() and rtems_calloc() functions declared in <rtems/malloc.h> which do the same as the corresponding C/POSIX functions except setting errno.

The posix_memalign() and aligned_alloc() functions do not have this issue with the errno.

#3587 6 months ago fixed rtems Sebastian Huber Sebastian Huber 6 months ago
Summary

Deprecate rtems_context

Description

The rtems_context typedef as no corresponding API. A user can do nothing with it. It is only used in cpukit/libmisc/monitor/mon-monitor.c and cpukit/libmisc/monitor/mon-editor.c in RTEMS. Deprecate it in this release.

#3589 6 months ago fixed rtems Sebastian Huber Sebastian Huber 6 months ago
Summary

Deprecate rtems_context_fp

Description

The rtems_context_fp typedef as no corresponding API. A user can do nothing with it. It is only used in cpukit/libmisc/monitor/mon-editor.c in RTEMS. Deprecate it in this release.

#3591 6 months ago fixed rtems Sebastian Huber Sebastian Huber 6 months ago
Summary

Deprecate region_information_block

Description

The region_information_block typedef as no corresponding API. It has no proper namespace prefix. A user can do nothing with it. It is only used in cpukit/libmisc/cpuuse/cpuusagetop.c and cpukit/libmisc/shell/main_mallocinfo.c in RTEMS. Deprecate it in this release.

#3593 6 months ago fixed rtems Sebastian Huber Sebastian Huber 6 months ago
Summary

Deprecate rtems_thread_cpu_usage_t

Description

The rtems_thread_cpu_usage_t typedef as no corresponding API. It violates the POSIX namespace. A user can do nothing with it. It is only used in cpukit/include/rtems/rtems/ratemon.h in RTEMS. Deprecate it in this release.

#3595 6 months ago fixed rtems Sebastian Huber Sebastian Huber 6 months ago
Summary

Deprecate rtems_rate_monotonic_period_time_t

Description

The rtems_rate_monotonic_period_time_t typedef as no corresponding API. It violates the POSIX namespace. A user can do nothing with it. It is only used in cpukit/include/rtems/rtems/ratemon.h in RTEMS. Deprecate it in this release.

#3598 6 months ago fixed rtems Sebastian Huber Sebastian Huber 5 months ago
Summary

Move internal types of API objects to separate header file

Description

The <rtems.h> header file still exposes a lot of implementation details via the definition of internal data structures, e.g. the *_Control structures of the API objects. They are only necessary for the application configuration. Move them to separate header files. Currently we have:

  • <rtems/rtems/XYZ.h>
  • <rtems/rtems/XYZimpl.h>

Use

  • <rtems/rtems/XYZdata.h>

for this new header file.

Potential new header files are:

  • rtems/extensiondata.h
  • rtems/rtems/asrdata.h
  • rtems/rtems/barrierdata.h
  • rtems/rtems/dpmemdata.h
  • rtems/rtems/eventdata.h
  • rtems/rtems/messagedata.h
  • rtems/rtems/partdata.h
  • rtems/rtems/ratemondata.h
  • rtems/rtems/regiondata.h
  • rtems/rtems/semdata.h
  • rtems/rtems/tasksdata.h
  • rtems/rtems/timerdata.h
#3636 5 months ago fixed rtems Sebastian Huber Sebastian Huber 4 months ago
Summary

Add rtems_scheduler_get_maximum_priority()

Description

The maximum task priority depends on the scheduler instance. It is a configuration parameter. Add a function to get it at runtime.

/**
 * @brief Gets the maximum task priority of the specified scheduler instance.
 *
 * @param[in] scheduler_id Identifier of the scheduler instance.
 * @param[out] priority Pointer to a task priority value.
 *
 * @retval RTEMS_SUCCESSFUL Successful operation.
 * @retval RTEMS_INVALID_ADDRESS The @a priority parameter is @c NULL.
 * @retval RTEMS_INVALID_ID Invalid scheduler instance identifier.
 */
rtems_status_code rtems_scheduler_get_maximum_priority(
  rtems_id             scheduler_id,
  rtems_task_priority *priority
);
#3637 5 months ago fixed rtems Sebastian Huber Sebastian Huber 5 months ago
Summary

Fix rtems_task_restart() argument type

Description

The argument type must be rtems_task_argument in rtems_task_restart() similar to rtems_task_start(). This is a severe issue on 64-bit targets since it prevents to pass pointer values to the task.

#3731 3 weeks ago fixed rtems Sebastian Huber Sebastian Huber 2 weeks ago
Summary

Add rtems_scheduler_get_processor()

Description

Add rtems_scheduler_get_processor() as a replacement for rtems_get_current_processor(). The rtems_get_current_processor() is a bit orphaned. Adopt it by the Scheduler Manager. This is in line with the glibc sched_getcpu() function.

#3732 3 weeks ago fixed rtems Sebastian Huber Sebastian Huber 13 days ago
Summary

Add rtems_scheduler_get_processor_maximum()

Description

Add rtems_scheduler_get_processor_maximum() as a replacement for rtems_get_processor_count(). The rtems_get_processor_count() is a bit orphaned. Adopt it by the Scheduler Manager. The count is also misleading, since the processor set may have gaps and the actual count of online processors may be less than the value returned by rtems_get_processor_count().

#1747 8 years ago fixed score Chris Johns Sebastian Huber 18 months ago
Summary

Heap extend allows discontinuous memory regions.

Description

The check in (cpukit/ChangeLog) states:

2010-06-07 Sebastian Huber <sebastian.huber@…>

  • score/src/heapextend.c: Implemented support for scattered heap areas.

The heap cannot support scattered blocks because the _Heap_Is_block_in_heap assumes the region is continuous between the first and last blocks of the heap. Making the gaps in the regions passed to the heap extend call used is questionable and makes the _Heap_Is_block_in_heap test not really perform the task it's name states. This is an issue because it is this check that determines if a heap free of NULL should proceed. This issue is covered in another PR.

I also wonder about a heap free call to an address that maps to one of the "in-use" gap regions. The previous heap code knew if an address was in the heap and therefore it was kind of safe to probe for a valid block. This assumption is now not valid.

The former heap extend code:

http://www.rtems.org/viewvc/rtems/cpukit/score/src/heapextend.c?revision=1.7&view=markup

clearly states the type of memory that can be added to an existing heap. The current code has no restrictions. The user manual is not great in this area. It would also be useful if comments are added to the heap extend code.

The heap extend code is used by the rtems_region_extend call and this call clearly states in the manual that the memory region must be continuous. If this has changed we should discuss the API change and make better note of it. I also suspect the testsuite will need additions to test any API changes.

#2132 6 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

<rtems/score/basedefs.h> superfluously includes <limits.h>

Description

In older RTEMS versions <rtems.h> provided <limits.h> indirectly. The include of <limits.h> was added to not break application source files that relied on this accidentally.

We may remove this include in the future.

#2133 6 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

<rtems/score/basedefs.h> superfluously includes <string.h>

Description

In older RTEMS versions <rtems.h> provided <string.h> indirectly. The include of <string.h> was added to not break application source files that relied on this accidentally.

We may remove this include in the future.

#2173 5 years ago fixed score Sebastian Huber Joel Sherrill 18 months ago
Summary

Potential integer overflow problem in EDF scheduler

Description

On 2014-03-21 14:46, Gedare Bloom wrote:> On Fri, Mar 21, 2014 at 9:43 AM, Sebastian Huber

<sebastian.huber@…> wrote:

[...]

I have another question regarding the EDF scheduler. Does this work in case _Watchdog_Ticks_since_boot overflows?

No. For this, I think we need to use "deadline folding" which is just modulo arithmetic.

void _Scheduler_EDF_Release_job(

Thread_Control *the_thread, uint32_t deadline

) {

Priority_Control new_priority;

if (deadline) {

/* Initializing or shifting deadline. */ new_priority = (_Watchdog_Ticks_since_boot + deadline)

& ~SCHEDULER_EDF_PRIO_MSB;

} else {

/* Switch back to background priority. */ new_priority = the_thread->Start.initial_priority;

}

the_thread->real_priority = new_priority; _Thread_Change_priority(the_thread, new_priority, true);

}

_Watchdog_Ticks_since_boot us uint32_t and overflows after 49 days with a one millisecond clock tick.

#2176 5 years ago wontfix score Jeffrey Hill Joel Sherrill 18 months ago
Summary

fishy behavior in termios tx task mode

Description

I have a look around in the drivers in the various BSPs and I notice that none of the termios drivers appear to transmit characters at task level even if they are running in termios task mode. Maybe all (most) of them send characters in the ISR. If there was a large frame of characters to send then this could lock out task activity for too long.

FWIW, I had a closer look at this today, and maybe something is fishy in the termios code when the TX part of termios runs in task driven mode. It seems that in task mode if the UART can accept characters immediately in the write routine then we wouldn’t need to turn on any interrupts at all. The write routine would need to somehow tell termios how many characters it sent; presumably this would occur by calling rtems_termios_dequeue_characters in the driver's write function. I see in the code that this tries to work, rtems_termios_dequeue_characters posts the semaphore of termios tx and increases the characters sent count of termios. However after the write routine returns it goes badly.

If the transmitter runs in termios TASK mode and the driver's write routine does not immediately enable an interrupt, then it returns to the code below in rtems_termios_puts and it sets the transmitter to rob_busy. After that the termios tx daemon proceeds to step through all of the characters remaining (I think that I see this in the debugger) and discards them because the transmitter stays in rob_busy state.

It's also probably odd that termios calls the write function with interrupts disabled when it is in task driven mode; we could loop outputting a large frame of characters in the write routine at task level with interrupts globally disabled.

if (tty->rawOutBufState == rob_idle) {

/* check, whether XOFF has been received */ if (!(tty->flow_ctrl & FL_ORCVXOF)) {

(*tty->device.write)(tty->minor,

(char *)&tty->rawOutBuf.theBuf[tty->rawOutBuf.Tail],1);

} else {

/* remember that output has been stopped due to flow ctrl*/ tty->flow_ctrl |= FL_OSTOP;

} tty->rawOutBufState = rob_busy;

}

[debug]#0 rtems_termios_puts (_buf=0x9ca60, len=1, tty=0x11509c) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libcsupport/src/termios.c:677 [debug]#1 0x00015c68 in oproc (c=99 'c', tty=0x11509c) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libcsupport/src/termios.c:748 [debug]#2 0x00015d9c in rtems_termios_write (arg=0x9cad8) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libcsupport/src/termios.c:770 [debug]#3 0x00002ef0 in console_write (major=0, minor=0, arg=0x9cad8) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/lib/libbsp/nios2/altera-sys-config/./console/console.c:171 [debug]#4 0x0006c328 in rtems_io_write (major=0, minor=0, argument=0x9cad8) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/sapi/src/iowrite.c:47 [debug]#5 0x00068044 in device_write (iop=0x114988, buffer=0x3366d4, count=44) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libfs/src/imfs/deviceio.c:160 [debug]#6 0x00017ef4 in write (fd=1, buffer=0x3366d4, count=44) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libcsupport/src/write.c:51 [debug]#7 0x0007f078 in _write_r (ptr=0x9d638, fd=1, buf=0x3366d4, nbytes=44) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libcsupport/src/write_r.c:38 [debug]#8 0x0006d488 in _fflush_r (ptr=0x9d638, fp=0x9d98c) at ../../../../../../nios2-rtems/altera/altera11.0/gnu-tools/rtems-patched/gcc-4.1/newlib/libc/stdio/fflush.c:214 [debug]#9 0x000754e8 in sfvwrite_r (ptr=0x9d638, fp=0x9d98c, uio=0x9cbe0) at ../../../../../../nios2-rtems/altera/altera11.0/gnu-tools/rtems-patched/gcc-4.1/newlib/libc/stdio/fvwrite.c:257 [debug]#10 0x0007830c in sprint_r (ptr=0x9d638, fp=0x20, uio=0x9cbe0) at ../../../../../../nios2-rtems/altera/altera11.0/gnu-tools/rtems-patched/gcc-4.1/newlib/libc/stdio/vfprintf.c:322 [debug]#11 0x00072fc8 in _vfprintf_r (data=0x9d638, fp=0x9d98c, fmt0=<value optimized out>, ap=0x0) at ../../../../../../nios2-rtems/altera/altera11.0/gnu-tools/rtems-patched/gcc-4.1/newlib/libc/stdio/vfprintf.c:1501 [debug]#12 0x0006dce8 in printf (fmt=0x1 "") at ../../../../../../nios2-rtems/altera/altera11.0/gnu-tools/rtems-patched/gcc-4.1/newlib/libc/stdio/printf.c:52 [debug]#13 0x00057de0 in bootpc_init (update_files=false, forever=true) at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libnetworking/nfs/bootp_subr.c:997 [debug]#14 0x0002960c in rtems_bsdnet_do_bootp () at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libnetworking/rtems/rtems_bootp.c:23 [debug]#15 0x0002b218 in rtems_bsdnet_initialize_network () at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/libnetworking/rtems/rtems_glue.c:980 [debug]#16 0x00001528 in Init (ignored=565432) at init.c:47 [debug]#17 0x0006ccb0 in _Thread_Handler () at /home/hill/nios2-rtems/rtems/rtems-git/rtems-4.10/c/src/../../cpukit/score/src/threadhandler.c:145 [debug]#18 0x0006cc28 in _Thread_Is_heir (the_thread=0x6cc28) at ../../cpukit/../../../altera-sys-config/lib/include/rtems/score/thread.inl:82 [debug]Backtrace stopped: frame did not save the PC

#2350 4 years ago fixed score Sebastian Huber 18 months ago
Summary

One watchdog ticks header per scheduler instance

Description

Currently there is one watchdog header for all ticks based watchdogs. This is not scalable. For example on the Freescale T4240 platform with 24-processors we observe in the smptests/smpwakeafter01 test a maximum thread dispatch disabled time of 3.8ms on processor 0 and 1.7ms on the other processors.

<PerCPUProfilingReport processorIndex="0">
  <MaxThreadDispatchDisabledTime unit="ns">3807457</MaxThreadDispatchDisabledTime>
  <MeanThreadDispatchDisabledTime unit="ns">124091</MeanThreadDispatchDisabledTime>
  <TotalThreadDispatchDisabledTime unit="ns">1706880473</TotalThreadDispatchDisabledTime>
  <ThreadDispatchDisabledCount>13755</ThreadDispatchDisabledCount>
  <MaxInterruptDelay unit="ns">0</MaxInterruptDelay>
  <MaxInterruptTime unit="ns">24661</MaxInterruptTime>
  <MeanInterruptTime unit="ns">10148</MeanInterruptTime>
  <TotalInterruptTime unit="ns">127682501</TotalInterruptTime>
  <InterruptCount>12582</InterruptCount>
</PerCPUProfilingReport>
<PerCPUProfilingReport processorIndex="1">
  <MaxThreadDispatchDisabledTime unit="ns">1715826</MaxThreadDispatchDisabledTime>
  <MeanThreadDispatchDisabledTime unit="ns">102805</MeanThreadDispatchDisabledTime>
  <TotalThreadDispatchDisabledTime unit="ns">1884937615</TotalThreadDispatchDisabledTime>
  <ThreadDispatchDisabledCount>18335</ThreadDispatchDisabledCount>
  <MaxInterruptDelay unit="ns">0</MaxInterruptDelay>
  <MaxInterruptTime unit="ns">47</MaxInterruptTime>
  <MeanInterruptTime unit="ns">12</MeanInterruptTime>
  <TotalInterruptTime unit="ns">8299</TotalInterruptTime>
  <InterruptCount>664</InterruptCount>
</PerCPUProfilingReport>
<SMPLockProfilingReport name="Watchdog">
  <MaxAcquireTime unit="ns">47020</MaxAcquireTime>
  <MaxSectionTime unit="ns">2709</MaxSectionTime>
  <MeanAcquireTime unit="ns">31</MeanAcquireTime>
  <MeanSectionTime unit="ns">52</MeanSectionTime>
  <TotalAcquireTime unit="ns">990203330</TotalAcquireTime>
  <TotalSectionTime unit="ns">1674926849</TotalSectionTime>
  <UsageCount>31604848</UsageCount>
  <ContentionCount initialQueueLength="0">10574</ContentionCount>
  <ContentionCount initialQueueLength="1">8168</ContentionCount>
  <ContentionCount initialQueueLength="2">8578</ContentionCount>
  <ContentionCount initialQueueLength="3">31577528</ContentionCount>
</SMPLockProfilingReport>

The watchdog lock is highly contended and since the watchdog insert procedure acquires and releases the lock during the iteration of the watchdog chain several times, this yields the high thread dispatch disabled times.

To get rid of this bottleneck we should move the watchdog context into the scheduler context to use one watchdog context per scheduler instance. Take care that active watchdogs move in case of a scheduler change of a thread.

#2366 4 years ago wontfix score Joel Sherrill 18 months ago
Summary

Create a Public API for the Atomic Operations

Description

Ticket #2364 regarded use of a pthread mutex in three graphics driver as basically an atomic flag to ensure only one open() was active at a time. This created an unnecessary dependency on the POSIX API being enabled. I changed the code to use score Atomic flags.

This highlighted the need for a public Atomic API.

The existing tests could be converted to the public API, a macro wrapper written for Classic API Atomics, and documentation added. This may be enough to be a small GSOC project.

#2412 4 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Improved priority inheritance implementation

Description

Problem

The RTEMS mutexes implement only a very simple approximation of the priority inheritance protocol. The real priority of a thread is only restored once it releases its last mutex. Lets consider this scenario. We have a file system instance protected by one mutex (e.g. JFFS2) and a dynamic memory allocator protected by another mutex. A low priority thread performs writes some log data into a file, thus it acquires the file system instance mutex. The file system allocates dynamic memory. Now a high priority thread interrupts and tries to allocate dynamic memory. The allocator mutex is already owned, so the priority of the low priority thread is raised to the priority of the high priority thread. The memory allocation completes and the allocator mutex is released, since the low priority thread still owns the file system instance mutex it continues to execute with the high priority (the high priority thread is not scheduled). It may now perform complex and long file system operations (e.g. garbage collection, polled flash erase and write functions) with a high priority.

Functional requirements

  • The mutex shall use the priority inheritance protocol to prevent priority inversion. On SMP configurations OMIP shall be used.
  • The mutex shall allow vertical nesting (a thread owns multiple mutexes).
  • The mutex shall allow horizontal nesting (a thread waits for ownership of a mutex those owner waits for ownership of a mutex, and so on).
  • Threads from one scheduler instance shall wait in priority order. The highest priority thread shall be dequeued first.
  • The highest priority waiting thread of each scheduler instance shall wait in FIFO order.
  • The mutex shall provide an acquire operation with timeout.
  • In case a mutex is released, then the previous owner shall no longer use the priorities inherited by this mutex.
  • In case a mutex acquire operation timeout occurs, then the current owner of the mutex shall no longer use the priorities inherited by the acquiring thread.
  • The order of the mutex release operations may differ from the order of the mutex acquire operations.
  • Priority changes not originating due to the priority inheritance protocol shall take place immediately.
  • Deadlock shall be detected. In case a deadlock would occur an error status shall be returned or a fatal error shall be generated.
  • Deadlocks at application level shall not lead to a deadlock at operating system level.

Performance requirements

  • The mutex acquire operation shall use only object-specific locks in case the mutex is not owned currently.
  • The mutex release operation shall use only object-specific locks in case no threads wait for ownership of this mutex.

Invariants

  • A mutex shall be owned by at most one thread.
  • A thread shall wait for ownership of at most one mutex.

Possible implementation

Use a recursive data structure to determine the highest priority available to a thread for each scheduler instance, e.g.

typedef struct Thread_Priority_node {
  Priority_Control current_priority;
  Priority_Control real_priority;
  struct Thread_Priority_node *owner;
  RBTree_Node Node;
  RBTree_Control Inherited_priorities;
} Thread_Priority_node;

typedef struct {
  ...
  Thread_Priority_node *priority_nodes; /* One per scheduler instances */
  ...
} Thread_Control;

Initially a thread has a priority node reflecting its real priority. The Thread_Priority_node::owner is NULL. The Thread_Priority_node::current_priority is set to the real priority. The Thread_Priority_node::Inherited_priorities is empty.

In case the thread must wait for ownership of a mutex, then it enqueues its priority node in Thread_Priority_node::Inherited_priorities of the mutex owner.

In case the thread is dequeued from the wait queue of a mutex, then it dequeues its priority node in Thread_Priority_node::Inherited_priorities of the previous mutex owner (ownership transfer) or the current mutex owner (acquire timeout).

In case the minimum of Thread_Priority_node::real_priority and Thread_Priority_node::Inherited_priorities changes, then Thread_Priority_node::current_priority is updated. In case the Thread_Priority_node::owner its not NULL, the priority change propagates to the owner, and so on. In case Thread_Priority_node::current_priority changes, the corresponding scheduler is notified.

The biggest issue is the locking on SMP configurations in case of recursive minimum updates.

Somehow we must connect this to the scheduler helping protocol for OMIP. We may have to replace the return value based scheduler operations with a pre-context-switch action. Due to some recent implementation changes the run-time of the _Thread_Dispatch() function is no longer average-case performance critical.

#2554 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

New watchdog handler implementation

Description

Background

The watchdog handler uses delta chains. The insert operation has a O(n) worst-case time complexity with n being the count of watchdogs in the delta chain. In each step of the insert operation, the SMP lock of the corresponding watchdog header is acquired and released. The profiling data obtain by test program smptests/smpwakeafter01 showed that the current implementation leads to unacceptable latencies, thus it should be replaced by something else.

The use cases for the watchdog handler fall roughly into two categories.

  • Timeouts - used to detect if some operations needs more time than expected. Since the unexpected happens hopefully rarely, timeout timers are usually removed before they expire. The critical operations are insert and removal. They are important for the performance of a network stack.
  • Timers - used to carry out some work in the future. They usually expire and need a high resolution. An example user is a time driven scheduler, e.g. rate-monotonic or EDF.

One approach is to use a red-black tree with the expiration time as the key. This leads to O(log(n)) worst-case insert and removal operations. For each operation it is sufficient to acquire and release the lock only once. The drawback is that a 64-bit integer type must be used for the intervals to avoid a potential overflow of the key values. With a system tick interval of 1ns the system could run more than 500 years before an overflow happens. The EDF scheduler would also profit from a 64-bit interval representation, see #2173.

An alternative is the use of a timer wheel based algorithm which is used in Linux and FreeBSD for example. A timer wheel based algorithm offers O(1) worst-case time complexity for insert and removal operations. The drawback is that the run-time of the watchdog tick procedure is somewhat unpredictable due to the use of a hash table or cascading.

Which approach should we choose? Since the watchdog serves the timeout and timer services in RTEMS we have to make some trade-offs. We recommend to use the red-black tree approach which offers a more predictable run-time behaviour and sacrifice the constant insert and removal operations offered by the timer wheel algorithms, see also https://www.kernel.org/doc/ols/2006/ols2006v1-pages-333-346.pdf. We can reuse the red-black tree support already used for the thread priority queues.

The new watchdog handler implementation is a prerequisite to eliminate the Giant lock in the Classic Timer manager.

Implementation

Change the _Watchdog_Ticks_since_boot to a 64-bit integer type. Keep the Watchdog_Interval at 32-bit for backward compatibility. Replace the delta chains with a red-black tree. Use the ticks for timers with a relative expiration time. Use struct timespec or struct bintime for timers with an absolute expiration time. This has the benefit that we do not have to adjust the data structures in case the absolute time changes, e.g. due to NTP. It simplifies the POSIX timer services, since no conversion to ticks is necessary.

#2555 3 years ago fixed score Sebastian Huber Sebastian Huber 16 months ago
Summary

Eliminate the Giant lock

Description

Background

The standard approach to turn a uni-processor operating system into an SMP-capable operating system is to encapsulate virtually the complete operating system state and protect it by one global recursive lock - the Giant lock. Thus, at most one processor can modify the operating system state at a time. Under Giant lock protection interrupt disable/enable critical sections still work. This approach is easy to realize and leads to something that runs on SMP with a minimal effort. Unfortunately, such an operating system does not scale with the processor count and offers very poor performance. It is quite useless for real applications.

The first steps to get rid of the Giant lock are tackled with the introduction of fine grained locking for the scheduler, watchdog handler, timestamps, thread queues, events, semaphores and message queues. The Giant lock is still used in a couple of places, e.g. all other objects using thread queues, thread life cycle changes (termination, restart) and scheduler changes. It is a straight forward task to eliminate it entirely, but it is also somewhat labour intensive since a lot of code must be changed.

Implementation

Eliminate all remaining code areas that use

  • _ISR_Disable(),
  • _ISR_Enable(), and
  • _ISR_Flash().

Direct users of these macros are

  • _Chain_Append(),
  • _Chain_Extract(),
  • _Chain_Get(),
  • _Chain_Insert(), and
  • _Chain_Prepend().

Each spot must be dealt with individually. Once this is done, delete these macros since they are now superfluous. Rename _ISR_Disable_without_giant() into _ISR_Locale_disable(). Rename _ISR_Enable_without_giant() into _ISR_Locale_enable().

Eliminate all remaining code areas that use

  • _Thread_Disable_dispatch() and
  • _Thread_Enable_dispatch().

A prominent user of this functions is _Objects_Get(). The following components are affected by these functions

  • Classic barrier handler,
  • Classic dual-ported memory handler,
  • Classic message queue handler,
  • Classic partition handler,
  • Classic rate-monotonic handler,
  • Classic region handler,
  • Classic semaphore handler,
  • Classic timer handler,
  • extension handler,
  • IO manager,
  • multi-processing management,
  • objects management,
  • POSIX barrier handler,
  • POSIX condition handler,
  • POSIX key handler,
  • POSIX message queue handler,
  • POSIX mutex handler,
  • POSIX rwlock handler,
  • POSIX semaphore handler,
  • POSIX spinlock handler,
  • POSIX timer handler,
  • signals,
  • thread cancellation,
  • thread life-cycle changes, and
  • thread scheduler changes.

Once this is done, delete _Thread_Disable_dispatch() and _Thread_Enable_dispatch(). As a side-effect the Giant lock will be removed.

#2556 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Implement the O(m) Independence-Preserving Protocol (OMIP)

Description

Background

The O(m) Independence-Preserving Protocol (OMIP) is a generalization of the priority inheritance protocol to clustered scheduling which avoids the non-preemptive sections present with priority boosting. The m denotes the number of processors in the system. Its implementation requires an extension of the scheduler helping protocol already used for the MrsP semaphores. However, the current implementation of the scheduler helping protocol has two major issues, see Catellani, Sebastiano, Luca Bonato, Sebastian Huber, and Enrico Mezzetti: Challenges in the Imple- mentation of MrsP. In Reliable Software Technologies - Ada-Europe 2015, pages 179–195, 2015. Firstly, the run-time of some scheduler operations depend on the size of the resource dependency tree. Secondly, the scheduler operations of threads which don't use shared resources must deal with the scheduler helping protocol in case an owner of a shared resource is somehow involved.

To illustrate the second issue, let us look at the following example. We have a system with eight processors and two L2 caches. We assign processor 0 to a partition P for latency sensitive real-time tasks (e.g. sensor and actuator handling), processors 1, 2 and 3 are assigned to a cluster CA and the remaining processors are assigned to a cluster CB for soft real-time worker tasks. The worker tasks use a shared resource, e.g. a file system for data storage. Let us suppose a task R of partition P sends a message to the workers. This may make a waiting worker ready, which in turn pre-empts the owner of a shared resource. In this case the scheduler helping protocol takes action and is carried out by the task R. This contradicts the intended isolation of scheduler instances.

The reason for this unfortunate coupling is a design issue of the scheduler helping protocol implementation. Some scheduler operations may return a thread in need of help. For example, if a thread is unblocked which pre-empts an owner of a shared resource, then the pre-empted thread is returned. Once a thread in need of help is returned, the ask for help operation of the scheduler is executed. An alternative to this return value based approach is the introduction of a pre-emption intervention during thread dispatching. Threads taking part in the scheduler helping protocol indicate this with a positive resource count value. In case a thread dispatch occurs and pre-empts an owner of a shared resource, the scheduler ask for help operation is invoked. So, the work is carried out on behalf of the thread which takes part in the scheduler helping protocol.

To overcome the first issue, an improved resource dependency tracking is required. One approach is to use a recursive red-black tree based data structure, see #2412.

Implementation

There are several steps necessary to implement OMIP.

  • Introduce per-scheduler locks.
  • Enable context switches with interrupts enabled.
  • Add a pre-emption intervention to the thread dispatch.
  • Add a table for priority nodes to the thread control block. For each scheduler instance there is one priority node.
  • Update the table in case the thread blocks on a resource, a timeout while waiting for a resource occurs, or ownership of a resource is transferred to the thread.
  • Use this table in the pre-emption intervention.
  • Update the MrsP implementation to the new infrastructure.

Currently, only one scheduler lock for all scheduler instances is used. This simplified the MrsP implementation and due to the presence of a Giant lock, this was not an issue. With the elimination of the Giant lock, however, we need one scheduler lock per scheduler instance to really profit from a decoupled system due to clustered scheduling.

The current implementation of thread dispatching has some implications with respect to the interrupt latency. It is crucial to preserve the system invariant that a thread can execute on at most one processor in the system at a time. This is accomplished with a boolean indicator in the thread context. The processor architecture specific context switch code will mark that a thread context is no longer executing and waits that the heir context stopped execution before it restores the heir context and resumes execution of the heir thread (the boolean indicator is basically a TTAS lock). So, there is one point in time in which a processor is without a thread. This is essential to avoid cyclic dependencies in case multiple threads migrate at once. Otherwise some supervising entity is necessary to prevent deadlocks. Such a global supervisor would lead to scalability problems so this approach is not used. Currently the context switch is performed with interrupts disabled. Thus in case the heir thread is currently executing on another processor, the time of disabled interrupts is prolonged since one processor has to wait for another processor to make progress.

If we add pre-emption intervention to the thread dispatch sequence, then there is an even greater need to avoid this issue with the interrupt latency. Interrupts normally store the context of the interrupted thread on its stack. In case a thread is marked as not executing, we must not use its thread stack to store such an interrupt context. We cannot use the heir stack before it stopped execution on another processor. If we enable interrupts during this transition, then we have to provide an alternative thread independent stack for interrupts in this time frame.

The pre-emption intervention should be added to _Thread_Do_dispatch() before the heir is read and perform the following pseudo-code actions.

pre_emption_intervention(executing):
	if executing.resource_count > 0:
		executing.lock()
		if executing.is_ready():
			for scheduler in executing.schedulers:
				scheduler.lock()
			if !executing.is_scheduled():
				for scheduler in executing.schedulers:
					scheduler.ask_for_help(executing)
			for scheduler in executing.schedulers:
				scheduler.unlock()
		else if executing.active_help_level > 0:
			idle.use(executing.scheduler_node)
		executing.unlock()

The scheduler help operation affects multiple scheduler instances. In terms of locking we have only two options,

  • use a global scheduler lock, or
  • obtain multiple per-scheduler locks at once.

A global scheduler lock is not an option. To avoid deadlocks obtain the per-scheduler locks in a fixed order. However, in this case the per-scheduler locks will observe different worst-case and average-case acquire times (depending on the order).

Use a recursive data structure to determine the highest priority available to a thread for each scheduler instance, e.g.

typedef struct Thread_Priority_node {
	Priority_Control current_priority;
	Priority_Control real_priority;
	struct Thread_Priority_node *owner;
	RBTree_Node Node;
	RBTree_Control Inherited_priorities;
} Thread_Priority_node;

typedef struct {
	...
	Thread_Priority_node *priority_nodes; /* One per scheduler instances */
	...
} Thread_Control;

Initially a thread has a priority node reflecting its real priority. The Thread_Priority_node::owner is NULL. The Thread_Priority_node::current_priority is set to the real priority. The Thread_Priority_node::Inherited_priorities is empty.

In case the thread must wait for ownership of a mutex, then it enqueues its priority node in Thread_Priority_node::Inherited_priorities of the mutex owner.

In case the thread is dequeued from the wait queue of a mutex, then it dequeues its priority node in Thread_Priority_node::Inherited_priorities of the previous mutex owner (ownership transfer) or the current mutex owner (acquire timeout).

In case the minimum of the Thread_Priority_node::real_priority and the Thread_Priority_node::Inherited_priorities changes, then Thread_Priority_node::current_priority is updated. In case the Thread_Priority_node::owner its not NULL, the priority change propagates to the owner, and so on. In case Thread_Priority_node::current_priority changes, the corresponding scheduler is notified.

Use the thread lock to protect the priority nodes.

#2559 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Delete the EXTERN pattern

Description

Change the following pattern: some.h:

#ifndef SOME_XYZ_EXTERN
#define SOME_XYZ_EXTERN extern
#endif
SOME_XYZ_EXTERN type xyz;

some_xyz.c:

#define SOME_XYZ_EXTERN
#include <some.h>

into: some.h:

extern type xyz;

some_xyz.c:

#include <some.h>
type xyz;

See discussion:

https://lists.rtems.org/pipermail/devel/2016-January/013506.html

Update Developer/Coding/Conventions accordingly.

#2617 3 years ago fixed score Joel Sherrill Sebastian Huber 18 months ago
Summary

rtems_heap_allocate_aligned_with_boundary() body and prototype inconsistent

Description

The first parameter is size_t in the .h and uintptr_t in the body. This resulted in a compiler error on the m32c. But it is an inconsistency which should be fixed even if no architecture complained.

The malloc.h header file has this:

void *rtems_heap_allocate_aligned_with_boundary(

size_t size, uintptr_t alignment, uintptr_t boundary

);

malloc_deferred.c has this:

void *rtems_heap_allocate_aligned_with_boundary(

uintptr_t size, uintptr_t alignment, uintptr_t boundary

)

#2626 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Unify thread cancel/join and delete

Description

The rtems_task_delete() is basically a pthread_cancel() plus pthread_join(). Unify the implementation and introduce a _Thread_Cancel() and _Thread_Join() to be used by both APIs. Get rid of the Giant lock for thread delete.

#2627 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Fix CPU time used for threads on SMP

Description

The CPU time used of a thread is currently maintained per-processor mostly during _Thread_Dispatch(). However, on SMP configurations the actual processor of a thread is difficult to figure out since thread dispatching is a highly asynchronous process (e.g. via inter-processor interrupts). Only the intended processor of a thread is known to the scheduler easily. Do the CPU usage accounting during thread heir updates in the context of the scheduler operations. Provide a function to get the CPU usage of a thread using proper locks to get a consistent value.

#2628 3 years ago wontfix score Sebastian Huber Sebastian Huber 17 months ago
Summary

Avoid home-grown condition variable implementation in the Classic Regions

Description

The Classic Region manager enables users to wait until memory is available to satisfy an allocation request. This is done through special purpose code that basically implements a condition variable.

#2631 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Use an ISR lock to protect the state of Classic Rate Monotonic objects

Description

The state of Classic Rate Monotonic is currently protected by the Giant lock and ISR disable sections. Use a per-object ISR lock to protect state changes instead.

#2664 3 years ago duplicate score Joel Sherrill Sebastian Huber 18 months ago
Summary

spclock_err02

Description

New test failure on sis but likely all targets.

* BEGIN OF TEST SPCLOCK_ERR 2 * TA1 - rtems_io_close - RTEMS_INVALID_NUMBER TA1 - rtems_io_control - RTEMS_INVALID_NUMBER TA1 - rtems_io_initialize - RTEMS_INVALID_NUMBER TA1 - rtems_io_open - RTEMS_INVALID_NUMBER TA1 - rtems_io_read - RTEMS_INVALID_NUMBER TA1 - rtems_io_write - RTEMS_INVALID_NUMBER TA1 - rtems_clock_set - 23:59:59 12/31/2000 - RTEMS_SUCCESSFUL TA1 - rtems_clock_get_tod - 00:00:00 01/01/2001 - RTEMS_SUCCESSFUL TA1 - rtems_clock_set - 23:59:59 12/31/1999 - RTEMS_SUCCESSFUL TA1 - rtems_clock_get_tod - 00:00:00 01/01/2000 - RTEMS_SUCCESSFUL assertion "ticks < 0x400000000" failed: file "../../cpukit/../../../sis/lib/include/rtems/score/watchdogimpl.h", line 316, function: _Watchdog_Ticks_from_timespec

Breakpoint 1, _Terminate (the_source=the_source@entry=RTEMS_FATAL_SOURCE_ASSERT,

is_internal=is_internal@entry=false, the_error=the_error@entry=33694096) at ../../../../../../rtems/c/src/../../cpukit/score/src/interr.c:36

36 { (gdb) bt #0 _Terminate (the_source=the_source@entry=RTEMS_FATAL_SOURCE_ASSERT,

is_internal=is_internal@entry=false, the_error=the_error@entry=33694096) at ../../../../../../rtems/c/src/../../cpukit/score/src/interr.c:36

#1 0x0200aed4 in rtems_fatal (source=source@entry=RTEMS_FATAL_SOURCE_ASSERT,

error=error@entry=33694096) at ../../../../../../rtems/c/src/../../cpukit/sapi/src/fatal2.c:34

#2 0x02004a9c in assert_func (

file=file@entry=0x201a650 "../../cpukit/../../../sis/lib/include/rtems/score/watchdogimpl.h", line=line@entry=316, func=func@entry=0x201a6d0 <func.3277> "_Watchdog_Ticks_from_timespec", failedexpr=failedexpr@entry=0x201a638 "ticks < 0x400000000") at ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/assert.c:52

#3 0x0200bbf8 in _Watchdog_Ticks_from_timespec (ts=0x2022210)

at ../../cpukit/../../../sis/lib/include/rtems/score/watchdogimpl.h:316

#4 _TOD_Set_with_timestamp (tod_as_timestamp=tod_as_timestamp@entry=0x2022280)

at ../../../../../../rtems/c/src/../../cpukit/score/src/coretodset.c:40

#5 0x02009880 in rtems_clock_set (tod=tod@entry=0x2022304)

at ../../../../../../rtems/c/src/../../cpukit/rtems/src/clockset.c:42

#6 0x02001818 in Init (argument=<optimized out>)

at ../../../../../../../rtems/c/src/../../testsuites/sptests/spclock_err02/init.c:93

#7 0x0200fcbc in _Thread_Entry_adaptor_numeric (executing=0x201fb90)

at ../../../../../../rtems/c/src/../../cpukit/score/src/threadentryadaptornumeric---Type <return> to continue, or q <return> to quit---

.c:25 #8 0x02012e0c in _Thread_Handler ()

at ../../../../../../rtems/c/src/../../cpukit/score/src/threadhandler.c:93

#9 0x02012d60 in _Thread_Handler ()

at ../../../../../../rtems/c/src/../../cpukit/score/src/threadhandler.c:29

(gdb) l init.c:93 88 status = rtems_clock_get_tod( &time ); 89 directive_failed( status, "rtems_clock_get_tod" ); 90 print_time( "TA1 - rtems_clock_get_tod - ", &time, " - RTEMS_SUCCESSFUL\n" ); 91 92 build_time( &time, 12, 31, 2100, 23, 59, 59, 0 ); 93 status = rtems_clock_set( &time ); 94 directive_failed( status, "rtems_clock_set" ); 95 print_time( "TA1 - rtems_clock_set - ", &time, " - RTEMS_SUCCESSFUL\n" ); 96 status = rtems_task_wake_after( rtems_clock_get_ticks_per_second() ); 97 status = rtems_clock_get_tod( &time ); (gdb)

#2692 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

User extensions execution order must be clarified

Description

The implemented and documented execution order of some user extensions disagree. Intended behaviour must be tested. Documentation must be updated accordingly.

#2725 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Classic binary semaphores without a locking protocol can be released by everyone

Description

The Classic binary semaphores without a locking protocol can be released by everyone, e.g. in contrast to the POSIX mutexes (all variants) or the Classic binary semphores with priority inheritance or ceiling, there is no owner check in the release path.

This behaviour is a bit unexpected and not documented.

The following test case fails in case an owner check is added:

*** BEGIN OF TEST SP 42 ***
Exercising blocking discipline w/extract in FIFO order 
Exercising blocking discipline w/unblock in FIFO order
TA00 - unblocked - OK

rtems_semaphore_delete FAILED -- expected (RTEMS_SUCCESSFUL) got (RTEMS_RESOURCE_IN_USE)

This is actually a bug in the test, since an available mutex is released again.

#2728 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Pipes use wrong semaphore for mutual exclusion

Description
cpukit/libfs/src/pipe/fifo.c-      sc = rtems_semaphore_create(
cpukit/libfs/src/pipe/fifo.c-        rtems_build_name('P', 'I', 'P', 'E'),
cpukit/libfs/src/pipe/fifo.c-        1,
cpukit/libfs/src/pipe/fifo.c:        RTEMS_BINARY_SEMAPHORE | RTEMS_INHERIT_PRIORITY | RTEMS_PRIORITY,
cpukit/libfs/src/pipe/fifo.c-        RTEMS_NO_PRIORITY,
cpukit/libfs/src/pipe/fifo.c-        &pipe_semaphore
cpukit/libfs/src/pipe/fifo.c-      );

Should use a binary semaphore with inherit priority instead.

#2729 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

TFTP client uses wrong semaphore for mutual exclusion

Description
cpukit/libnetworking/lib/tftpDriver.c-    rtems_build_name('T', 'F', 'T', 'P'),
cpukit/libnetworking/lib/tftpDriver.c-    1,
cpukit/libnetworking/lib/tftpDriver.c-    RTEMS_FIFO |
cpukit/libnetworking/lib/tftpDriver.c:    RTEMS_BINARY_SEMAPHORE |
cpukit/libnetworking/lib/tftpDriver.c-    RTEMS_NO_INHERIT_PRIORITY |
cpukit/libnetworking/lib/tftpDriver.c-    RTEMS_NO_PRIORITY_CEILING |
cpukit/libnetworking/lib/tftpDriver.c-    RTEMS_LOCAL,

Should use a binary semaphore with inherit priority.

#2740 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Suboptimal type for Timestamp_Control

Description

Currently we have

typedef struct bintime Timestamp_Control;

this type offers more precision than needed. Maybe use sbintime_t (also known as int64_t) instead to simplify computations.

#2751 3 years ago fixed score Sebastian Huber Sebastian Huber 10 months ago
Summary

Thread dispatch via interrupt is broken at least on ARM and PowerPC

Description

The ARM and PowerPC interrupt epilogues call _Thread_Dispatch() with interrupts disabled (counter example: SPARC).

On SMP configurations, since inter-processor interrupts set the thread dispatch necessary indicator this prevents a thread dispatch notification in post-switch handlers (which all run with interrupts disabled).

On all configurations, this is a serious issue for the interrupt latency.

#2752 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Relax execution enviroment for thread begin extensions

Description

Currently, the thread begin extensions are invoked with thread dispatching disabled. There is an explanation for this in the code

  /*
   * Take care that 'begin' extensions get to complete before
   * 'switch' extensions can run.  This means must keep dispatch
   * disabled until all 'begin' extensions complete.
   */
  _User_extensions_Thread_begin( executing );

However, the switch extension is always invoked before the thread begin extension for all threads except the initialization thread. A thread dispatch disabled contexts drastically limits the work which can be carried out in the thread begin extensions. It is for example not possible to call malloc(), create POSIX keys or access C++ thread local storage.

The thread begin extension should execute in a normal thread context. Thread begin extensions that are disturbed by a thread dispatch should deal with this locally.

With the availability of C++ thread local storage in RTEMS being able to pre-initialize such objects in the thread begin extension would be quite handy.

#2765 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Application level deadlocks may lead to SMP lock level deadlocks

Description

Due to a missing deadlock detection application level deadlocks may lead to SMP lock level deadlocks.

#2771 3 years ago wontfix score Chris Johns 18 months ago
Summary

Empty C++ file with just <rtems.h> does not compile with HEAD.

Description

I have an application that does not build.

The following C++ file:

$ cat t1.cpp
#include <rtems.h>

does not compile with git head 5fe6d07ad5690e3d9c6445ca3a465a700a5a5015 on Zynq ARM. Build with:

$ /opt/work/rtems/4.12/bin/arm-rtems4.12-g++ \
  -B/opt/work/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib \
  -B/opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib \
  -specs bsp_specs -qrtems \
  -march=armv7-a -mthumb -mfpu=neon -mfloat-abi=hard -mtune=cortex-a9 \
  -g -O2 -DNDEBUG -std=c++11 \
  -Werror -Wall -Wextra \
  -o t1.o \
  -c t1.cpp

Some (too much to post) of the output is:

In file included from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/thread.h:36:0,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/heap.h:22,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/rtems/types.h:26,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems.h:31,
                 from t1.cpp:1:
/opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/timestamp.h: In function 'void _Timestamp_Set(Timestamp_Control*, time_t, long int)':
/opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/timestamp.h:78:33: error: 'timespec2bintime' was not declared in this scope
   timespec2bintime( &_ts, _time );
                                 ^
/opt/work/si/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/timestamp.h: In function 'void _Timestamp_Set_to_zero(Timestamp_Control*)':
/opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/timestamp.h:94:8: error: invalid use of incomplete type 'Timestamp_Control {aka struct bintime}'
   _time->sec = 0;
        ^~
In file included from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/time.h:299:0,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/timestamp.h:43,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/thread.h:36,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/score/heap.h:22,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems/rtems/types.h:26,
                 from /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/include/rtems.h:31,
                 from t1.cpp:1:
/opt/work/rtems/4.12/arm-rtems4.12/include/machine/_time.h:40:15: note: forward declaration of 'Timestamp_Control {aka struct bintime}'
 extern struct bintime _Timecounter_Boottimebin;
               ^~~~~~~

If '-std=c++11' is removed or replaced with '-std=gnu++11' the error becomes:

arm-rtems4.12-g++: fatal error: /opt/work/bsps/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib/bsp_specs: attempt to rename spec 'endfile' to already defined spec 'old_endfile'
#2776 3 years ago fixed score Alexander Krutwig Sebastian Huber 18 months ago
Summary

SPI Framework

Description

Development of a SPI framework which shall be used for further SPI bus and device drivers. The framework shall be developed using the i2c framework as a template. It shall export the Linux Userspace SPI API.

#2788 3 years ago wontfix score Chris Johns Chris Johns 18 months ago
Summary

RTEMS I2C API only defines Standard-mode (Sm) speed as a default.

Description

The RTEMS I2C API as defined in cpukit/dev/include/dev/i2c/i2c.h only defines the bus speed as Standard-mode (Sm) as defined by the I2C standard. This is set as I2C_BUS_CLOCK_DEFAULT. The default speed is defined by the hardware, ie the devices connected, and not this API.

The API should define the speeds as defined in the I2C standard and there should be no default. Drivers like the Cadence driver for the Zynq should be modified to require the bus speed be provided and all future drivers need to provide the speed.

#2790 3 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Linker sets broken with GCC 7

Description

See also:

https://gcc.gnu.org/ml/gcc/2016-09/msg00114.html

The

#define MAKEGCCNOTKNOWTHEADDRESS(ptr) asm("":"+r"(ptr))

is probably the best option. It works probably also with link-time optimization.

#2795 3 years ago fixed score Kuan Gedare Bloom <gedare@…> 18 months ago
Summary

Overrun Handling for general real-time models

Description

In the current implementation, if a task period is time out, the next call of rtems_rate_monotonic_period() will only release one following job and manipulate the task period with the calling moment + the next length of period. With the assumption that implicit/constraint deadline and hard real-time model, the above mechanism is okay.

However, it may not be applicable for general task models, e.g., soft real-time task, arbitrary deadline, mixed-criticality system [1-4]. It is usually assumed that multiple task jobs of a task are executed in a first-come-first-serve manner. Thus, it is sufficient to release the second task job at the moment the first task job finishes according to a strictly periodic release pattern. The current design in fact shifts the release pattern of periodic/sporadic tasks. Since there maybe more than one postponed jobs due to the preemption, these postponed jobs that should be released are never released to the system.

Although there is no standard requirement in reality for deadline misses, with this enhancement, the postponed jobs will be released with the correct number without periodic release shifting. This way of handling is already widely considered in academia from 90s [2] until now [3] or even on multicores as well [4].

I refine the following four files and handle this requirement individually. The overhead seems to me negligible. cpukit/rtems/include/rtems/rtems/ratemon.h cpukit/rtems/include/rtems/rtems/ratemonimpl.h cpukit/rtems/src/ratemontimeout.c cpukit/rtems/src/ratemonperiod.c I have tested the enhancement on Qemu and Raspberry Pi Model B+ with corresponding BSPs.

I believe this patch as a basis is required for further use for more general real-time task models. This enhancement only affect those timeout cases without changing any behaviour in normal cases. This enhancement is accepted in workshop mixed-criticality (WMC 2016) along with RTSS'16 this year [5].

To demonstrate the differences, a heuristic example is prepared in testsuites/sptests/sprmsched01 to show the benefit of the enhancement: Given two tasks with implicit deadline that task deadline is equal to its period. Task 1 period is 10000 ticks, whereas task 2 is 2000 ticks. Task 1 has the execution time 6000 ticks, and task 2 has 1000 ticks. Assume Task 1 has a higher priority than task 2. Task 1 only executes 2 times. In the expected result, we can observe that the postponed jobs are continuously released till there is no postponed job left, and the task period will still keep as it is. (Job 3-7 in task 2 are postponed jobs)

[1] Buttazzo et al., Soft Real-Time Systems: Predictability vs. Efficiency, Springer 2005, ​http://www.springer.com/gp/book/9780387237015 [2] Lehoczky et al., Fixed priority scheduling of periodic task sets with arbitrary deadlines, RTSS 1990, ​http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=128748 [3] Georg von der Brüggen et al., Systems with Dynamic Real-Time Guarantees in Uncertain and Faulty Execution Environments, RTSS'16, accepted. [4] Huang et al., Response time bounds for sporadic arbitrary-deadline tasks under global fixed-priority scheduling on multiprocessors, RTNS 2015, ​http://dl.acm.org/citation.cfm?doid=2597457.2597459 [5] Chen et al., Overrun Handling for Mixed-Criticality Support in RTEMS, WMC 2016, accepted.

#2803 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Get rid of CPU_BIG_ENDIAN and CPU_LITTLE_ENDIAN

Description

The remaining uses of the CPU port defines CPU_BIG_ENDIAN and CPU_LITTLE_ENDIAN should be replaced by the BSD (also available in glibc) BYTE_ORDER.

#2808 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Conditionally provide rtems_interrupt_frame

Description

Provide rtems_interrupt_frame only if CPU_ISR_PASSES_FRAME_POINTER is defined to TRUE.

#2809 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Reduce interrupt latency on SMP configurations during thread dispatch

Description

Currently we have this situation:

https://docs.rtems.org/doc-current/share/rtems/html/c_user/Symmetric-Multiprocessing-Services-Thread-Dispatch-Details.html#Symmetric-Multiprocessing-Services-Thread-Dispatch-Details

"On SMP systems, scheduling decisions on one processor must be propagated to other processors through inter-processor interrupts. So, a thread dispatch which must be carried out on another processor happens not instantaneous. Thus several thread dispatch requests might be in the air and it is possible that some of them may be out of date before the corresponding processor has time to deal with them. The thread dispatch mechanism uses three per-processor variables,

  • the executing thread,
  • the heir thread, and
  • an boolean flag indicating if a thread dispatch is necessary or not.

Updates of the heir thread and the thread dispatch necessary indicator are synchronized via explicit memory barriers without the use of locks. A thread can be an heir thread on at most one processor in the system. The thread context is protected by a TTAS lock embedded in the context to ensure that it is used on at most one processor at a time. The thread post-switch actions use a per-processor lock. This implementation turned out to be quite efficient and no lock contention was observed in the test suite.

The current implementation of thread dispatching has some implications with respect to the interrupt latency. It is crucial to preserve the system invariant that a thread can execute on at most one processor in the system at a time. This is accomplished with a boolean indicator in the thread context. The processor architecture specific context switch code will mark that a thread context is no longer executing and waits that the heir context stopped execution before it restores the heir context and resumes execution of the heir thread (the boolean indicator is basically a TTAS lock). So, there is one point in time in which a processor is without a thread. This is essential to avoid cyclic dependencies in case multiple threads migrate at once. Otherwise some supervising entity is necessary to prevent deadlocks. Such a global supervisor would lead to scalability problems so this approach is not used. Currently the context switch is performed with interrupts disabled. Thus in case the heir thread is currently executing on another processor, the time of disabled interrupts is prolonged since one processor has to wait for another processor to make progress.

It is difficult to avoid this issue with the interrupt latency since interrupts normally store the context of the interrupted thread on its stack. In case a thread is marked as not executing, we must not use its thread stack to store such an interrupt context. We cannot use the heir stack before it stopped execution on another processor. If we enable interrupts during this transition, then we have to provide an alternative thread independent stack for interrupts in this time frame. This issue needs further investigation.

The problematic situation occurs in case we have a thread which executes with thread dispatching disabled and should execute on another processor (e.g. it is an heir thread on another processor). In this case the interrupts on this other processor are disabled until the thread enables thread dispatching and starts the thread dispatch sequence. The scheduler (an exception is the scheduler with thread processor affinity support) tries to avoid such a situation and checks if a new scheduled thread already executes on a processor. In case the assigned processor differs from the processor on which the thread already executes and this processor is a member of the processor set managed by this scheduler instance, it will reassign the processors to keep the already executing thread in place. Therefore normal scheduler requests will not lead to such a situation. Explicit thread migration requests, however, can lead to this situation. Explicit thread migrations may occur due to the scheduler helping protocol or explicit scheduler instance changes. The situation can also be provoked by interrupts which suspend and resume threads multiple times and produce stale asynchronous thread dispatch requests in the system."

Add an interrupt frame to the per-CPU control which can be used during context switches on SMP configurations.

#2811 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

More robust thread dispatching on SMP and ARM Cortex-M

Description

On SMP configurations, it is a fatal error to call blocking operating system with interrupts disabled, since this prevents delivery of inter-processor interrupts. This could lead to executing threads which are not allowed to execute resulting in undefined behaviour.

The ARM Cortex-M port has a similar problem, since the interrupt state is not a part of the thread context.

Add a new CPU port function:

/**
 * @brief Returns true if interrupts are enabled in the specified ISR level,
 * otherwise returns false.
 *
 * @param[in] level The ISR level.
 *
 * @retval true Interrupts are enabled in the ISR level.
 * @retval false Otherwise.
 */
RTEMS_INLINE_ROUTINE bool _CPU_ISR_Is_enabled( uint32_t level )
{
  return false;
}

Use this function to ensure that _Thread_Do_dispatch() is called with an interrupt level with enabled interrupts, otherwise call _Terminate().

#2836 2 years ago fixed score Joel Sherrill Joel Sherrill 18 months ago
Summary

Add posix_devctl()

Description

The posix_devctl() method is defined in POSIX 1003.26 and required by the FACE POSIX profiles. The only use case that needs to be supported is FIONBIO on sockets per the FACE Technical Standard.

ioctl() is not a standardized method per POSIX and is not included in the FACE Profiles.

Making operations non-blocking can also be done with fcntl() but due to RTOS qualification concerns, fcntl() is not included in the more stringent FACE profiles. Specifically, it is not in the Safety Base profile which matches the RTEMS POSIX capabilities.

This requires adding the <devctl.h> file to newlib. That has been done. I am testing my implementation but a tool update will be needed before this can be pushed to the community. This is OK because we have other reasons to move to a new gcc and newlib version soon.

#2838 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Termios task driven mode should use mutex for device operations

Description

Termios has a task driven mode (TERMIOS_TASK_DRIVEN). This mode aims to avoid long sections with disabled interrupts. This is only partly implemented since the device level state is still protected by disabled interrupts. Use a mutex to protect the device level state in task driven mode to fix this issue.

#2839 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Add new interrupt server driven Termios mode

Description

Add a new new interrupt server driven Termios mode (TERMIOS_IRQ_DRIVEN). This mode is identical to the interrupt driven mode except that a mutex is used for device level locking. The intended use case for this mode are device drivers that use the interrupt server, e.g. SPI or I2C connected devices.

#2841 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Add NXP SC16IS752 serial device driver

Description

Add Termios device driver for NXP SC16IS752 (RS232/RS485 over SPI or I2C).

#2858 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Add user defined thread names

Description

Add user defined thread names to ease debugging, enhance the system diagnostics and improve compatibility to other systems, e.g. Linux and FreeBSD.

Implement pthread_setname_np() and pthread_getname_np(). Add CONFIGURE_MAXIMUM_THREAD_NAME_SIZE to the application configuration options. Add a application configuration dependent storage area for thread names to the thread control block.

#2877 2 years ago fixed score Stavros Passas joel.sherrill@… 18 months ago
Summary

DHCP client fails on complex networks

Description

What happens is that on networks with more than one DHCP servers, or on networks that use multiple vlans it can happen that After a DHCP discover our client broadcasts, it receives multiple offers, which is perfectly fine based on the DHCP RFC.

However our implementation of a DHCP client, expects a linear execution flow:

  1. Broadcast a DHCP discover;
  2. Wait for a DHCP offer;
  3. Transmit a DHCP request;
  4. Wait for a DHCP ack;

However the network stack is not cleaned between the reception of a DHCP offer and a DHCP ack, so if multiple offers are received, just the first one will be processed during the "Receive DHCP offer" phase, and the next one will be received when we expect a "DHCP ack", which makes our implementation assume the DHCP handshake is invalid and fail. Thus we restart the network and retry the whole process from the beginning which will cause the same issue again.

This issue that is present from when I remember in RTEMS, definitely from 4.10 up to now.

#2879 2 years ago fixed score David Binderman Chris Johns 18 months ago
Summary

src/cpukit/libdebugger/rtems-debugger-server.c: four problems

Description

1.

src/cpukit/libdebugger/rtems-debugger-server.c:1306]: (style) Redundant condition: extended. '!extended || (extended && check_pid(pid))' is equivalent to '!extended || check_pid(pid)'

Suggest simplify.

2.

src/cpukit/libdebugger/rtems-debugger-server.c:1858]: (warning) Possible null pointer dereference: rtems_debugger

Source code is

  if (r < 0) {
    rtems_printf(printer, "error: rtems-db: remote begin: %s: %s\n",
                 rtems_debugger->remote->name, strerror(errno));
    free(rtems_debugger);
    rtems_debugger = NULL;
  }

  /*
   * Reset at the end of the session.
   */
  rtems_debugger->flags = 0;

Suggest adding return -1 inside the if.

3.

src/cpukit/libdebugger/rtems-debugger-server.c:906]: (style) Redundant condition: extended. '!extended || (extended && check_pid(pid))' is equivalent to '!extended || check_pid(pid)'

Duplicate.

4.

src/cpukit/libdebugger/rtems-debugger-server.c:956]: (warning) Char literal compared with pointer 'p'. Did you intend to dereference it?

Source code is

  while (p != NULL && p != '\0') {

Maybe better code

  while (p != NULL && *p != '\0') {
#2885 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Fix rtems_rate_monotonic_postponed_job_count() prototype

Description

rtems_rate_monotonic_postponed_job_count() should return an RTEMS status code. It should be renamed to rtems_rate_monotonic_get_postponed_job_count() or rtems_rate_monotonic_get_postponed_jobs() (similar to rtems_rate_monotonic_get_statistics()).

#2889 2 years ago fixed score Stavros Passas Stavros Passas <stavros.passas@…> 18 months ago
Summary

RTEMS_STACK_CHECKER_EXTENSION has incomplete definition

Description

The extension for the stack checker defines 8 entries, while the structure for RTEMS extensions gets 9 arguments. This causes warnings to appear on applications compiled with -Werror, or similar flags.

The handler that is missing is for the terminate callback, so a 0 entry would be enough to fix this error.

#2890 2 years ago fixed score Stavros Passas Sebastian Huber <sebastian.huber@…> 18 months ago
Summary

_RBTree_Initialize_node generates warnings

Description

Currently, when _RBTree_Initialize_node is used, it generates warnings of unused variables. I traced the issue down to the variable being used only if RTEMS_DEBUG is set.

Thus, the argument should be marked as RTEMS_UNUSED, if RTENS_DEBUG is not set.

#2897 2 years ago fixed score Kevin Kirspel Needs Funding 18 months ago
Summary

Update termios.h to match the latest FREEBSD definitions

Description

The FREEBSD struct termios separates out the input and output baud rates into individual speed_t variables. It also supports more flag options. The big benefit is that the rtems-libbsd tty code can be ported cleanly without the need to fixup the input and output baud rates. This should be a transparent change unless someone was manipulating the baud rates directly through the c_cflags (not using the cfgetispeed, cfgetospeed, cfsetispeed, cfsetospeed functions).

#2909 2 years ago fixed score Sebastian Huber Chris Johns 18 months ago
Summary

xz: Support for 64-bit CRC is build although XZ_USE_CRC64 is not defined

Description

This leads to:

../../../../../../rtems/c/src/../../cpukit/libmisc/xz/xz_crc64.c:21:16: warning: no previous prototype for 'xz_crc64_init' [-Wmissing-prototypes] ../../../../../../rtems/c/src/../../cpukit/libmisc/xz/xz_crc64.c:40:20: warning: no previous prototype for 'xz_crc64' [-Wmissing-prototypes]

We should enable the 64-bit CRC or remove this file from the build.

#2912 2 years ago fixed score Sebastian Huber Chris Johns 18 months ago
Summary

libdebugger: control reaches end of non-void function

Description

../../../../../../rtems/c/src/../../cpukit/libdebugger/rtems-debugger-server.c:393:1: warning: control reaches end of non-void function [-Wreturn-type] ../../../../../../rtems/c/src/../../cpukit/libdebugger/rtems-debugger-server.c:405:1: warning: control reaches end of non-void function [-Wreturn-type]

#2916 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

termios: Change receive callback invocation to enable select() and poll() support

Description

Call the receive callback in case a read will succeed without to block. This enables the use of the receive callback for a poll() and select() support. Increase raw input buffer size to allow buffering of one line.

#2917 2 years ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

termios: Make write POSIX compatible

Description

Currently only blocking read/write operations are implemented. A blocking write must transfer at least one character. It should not wait for the device for the second character and so on.

#2923 2 years ago wontfix score Joel Sherrill Sebastian Huber 18 months ago
Summary

Questionable Code in resource_snapshot.c

Description

Coverity URL:

https://scan5.coverity.com/reports.htm#v29808/p10069/fileInstanceId=108958257&defectInstanceId=30877219&mergedDefectId=1399703

Coverity Scan doesn't like taking the address of a single uint32_t and then treating it as an array pointer. I think the code is likely doing what is intended but there is an implicit linkable between the order of the fields in the structure here and the class numbers used in the object IDs. Is there a way to improve this code and reduce the linkage?

141

  1. address_of: Taking address with &snapshot->active_posix_keys yields a singleton pointer.
  2. assign: Assigning: active = &snapshot->active_posix_keys.

142 active = &snapshot->active_posix_keys; 143

  1. Condition i < 19U /* sizeof (objects_info_table) / sizeof (objects_info_table[0]) */, taking true branch.

144 for (i = 0; i < RTEMS_ARRAY_SIZE(objects_info_table); ++i) { 145 const Objects_Information *information; 146 147 information = _Objects_Get_information( 148 objects_info_table[i].api, 149 objects_info_table[i].cls 150 ); 151

  1. Condition information != NULL, taking true branch.

152 if (information != NULL) { CID 1399703 (#1 of 1): Out-of-bounds access (ARRAY_VS_SINGLETON)

  1. ptr_arith: Using active as an array. This might corrupt or misinterpret adjacent memory locations.

153 active[i] = _Objects_Active_count(information);

#2925 2 years ago fixed score Joel Sherrill Chris Johns 6 months ago
Summary

Warnings in rtl-obj-cache.c on some targets

Description

These warnings are on m68k but not sparc, mips powerpc, or arm. Looks like inttypes.h should be used.

../../../../../../rtems/c/src/../../cpukit/libdl/rtl-obj-cache.c:175:47: warning: format '%d' expects argument of type 'int', but argument 3 has type 'long unsigned int' [-Wformat=] ../../../../../../rtems/c/src/../../cpukit/libdl/rtl-obj-cache.c:175:85: warning: format '%d' expects argument of type 'int', but argument 7 has type 'long unsigned int' [-Wformat=] ../../../../../../rtems/c/src/../../cpukit/libdl/rtl-obj-cache.c:81:67: warning: format '%d' expects argument of type 'int', but argument 7 has type 'long unsigned int' [-Wformat=] ../../../../../../rtems/c/src/../../cpukit/libdl/rtl-obj-cache.c:81:81: warning: format '%d' expects argument of type 'int', but argument 9 has type 'long unsigned int' [-Wformat=] ~

#2930 2 years ago fixed score Joel Sherrill Daniel Hellstrom 18 months ago
Summary

Coverity Reports Out of Bounds Read in drvmgr_print.c

Description

https://scan5.coverity.com/reports.htm#v29808/p10069/fileInstanceId=109359850&defectInstanceId=30967449&mergedDefectId=1399730

354 printf(" DRIVER ID: 0x%llx\n", drv->drv_id);

  1. Condition drv->name, taking true branch.

355 printf(" NAME: %s\n", drv->name ? drv->name : "NO_NAME"); 356 printf(" BUS TYPE: %d\n", drv->bus_type); 357 printf(" OPERATIONS:\n");

  1. alias: Assigning: ppfunc = &drv->ops->init[0]. ppfunc now points to element 0 of drv->ops->init (which consists of 4 4-byte elements).
  2. Condition i < 6U /* sizeof (struct drvmgr_drv_ops) / sizeof (void (*)(void)) */, taking true branch.
  3. Condition i < 6U /* sizeof (struct drvmgr_drv_ops) / sizeof (void (*)(void)) */, taking true branch.
  4. cond_at_most: Checking i < 6U implies that i may be up to 5 on the true branch.

358 for (i = 0, ppfunc = (fun_ptr *)&drv->ops->init[0]; 359 i < DRVMGR_OPS_NUM(struct drvmgr_drv_ops); i++)

  1. Jumping back to the beginning of the loop.

CID 1399730 (#1 of 1): Out-of-bounds read (OVERRUN)

  1. overrun-local: Overrunning array of 4 4-byte elements at element index 5 (byte offset 20) by dereferencing pointer ppfunc + i.

360 printf(" %s %p\n", drv_ops_names[i], ppfunc[i]); 361 printf(" NO. DEVICES: %d\n", drv->dev_cnt); 362

#2935 2 years ago wontfix score Martin Aberg joel.sherrill@… 18 months ago
Summary

Termios task driven mode not compatible with SMP

Description

When the Termios task driven functioning mode is used, rtems_termios_open_tty() calls rtems_task_create() with RTEMS_NO_PREEMPT in the initial task mode parameter. RTEMS_NO_PREEMPT is not supported on SMP.

rtems_task_create() returns RTEMS_UNSATISFIED in this SMP scenario and Termios ends up in rtems_fatal_error_occurred().

Termios starts the RX and TX tasks successfully on SMP if RTEMS_NO_PREEMPT is removed from the initial task modes of these tasks. However, I suspect there may be assumptions on the NO_PREEMPT mode for the RX and TX tasks in other parts of Termios.

#2957 2 years ago fixed score Sebastian Huber Gedare 18 months ago
Summary

Shared memory support internal locking is broken

Description

The top level lock is an ISR lock (interrupt disable/enable or SMP lock) and the low level lock is potentially a mutex. The problem is exposed by test psxshm02:

#0  _Terminate (the_source=INTERNAL_ERROR_CORE, the_error=31) at ../../../../../../rtems/c/src/../../cpukit/score/src/interr.c:35
#1  0x00111654 in _Internal_error (core_error=INTERNAL_ERROR_BAD_THREAD_DISPATCH_ENVIRONMENT) at ../../../../../../rtems/c/src/../../cpukit/score/src/interr.c:52
#2  0x00117010 in _Thread_Do_dispatch (cpu_self=0x2035c0 <_Per_CPU_Information>, level=1611071955) at ../../../../../../rtems/c/src/../../cpukit/score/src/threaddispatch.c:190
#3  0x0011a568 in _Thread_Dispatch_enable (cpu_self=0x2035c0 <_Per_CPU_Information>) at ../../cpukit/../../../realview_pbx_a9_qemu/lib/include/rtems/score/threaddispatch.h:227
#4  0x0011b6c4 in _Thread_Change_life (clear=THREAD_LIFE_PROTECTED, set=THREAD_LIFE_PROTECTED, ignore=(unknown: 0)) at ../../../../../../rtems/c/src/../../cpukit/score/src/threadrestart.c:684
#5  0x0011b6ea in _Thread_Set_life_protection (state=THREAD_LIFE_PROTECTED) at ../../../../../../rtems/c/src/../../cpukit/score/src/threadrestart.c:691
#6  0x0010f3dc in _API_Mutex_Lock (the_mutex=0x2037d8) at ../../../../../../rtems/c/src/../../cpukit/score/src/apimutexlock.c:31
#7  0x001050a0 in _RTEMS_Lock_allocator () at ../../cpukit/../../../realview_pbx_a9_qemu/lib/include/rtems/score/apimutex.h:120
#8  0x00105442 in rtems_heap_allocate_aligned_with_boundary (size=10004, alignment=0, boundary=0) at ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/malloc_deferred.c:89
#9  0x001055a6 in malloc (size=10004) at ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/malloc.c:39
#10 0x0011e820 in realloc (ptr=0x0, size=10004) at ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/realloc.c:62
#11 0x0010b1a2 in _POSIX_Shm_Object_resize_from_heap (shm_obj=0x204870, size=10004) at ../../../../../../rtems/c/src/../../cpukit/posix/src/shmheap.c:59
#12 0x0010b6ac in shm_ftruncate (iop=0x202cf8 <rtems_libio_iops+168>, length=10004) at ../../../../../../rtems/c/src/../../cpukit/posix/src/shmopen.c:83
#13 0x00104cfc in ftruncate (fd=3, length=10004) at ../../../../../../rtems/c/src/../../cpukit/libcsupport/src/ftruncate.c:37
#14 0x001008e0 in POSIX_Init (argument=0x0) at ../../../../../../../rtems/c/src/../../testsuites/psxtests/psxshm02/init.c:54
#15 0x001201ee in _Thread_Entry_adaptor_pointer (executing=0x2041a8) at ../../../../../../rtems/c/src/../../cpukit/score/src/threadentryadaptorpointer.c:25
#16 0x00120302 in _Thread_Handler () at ../../../../../../rtems/c/src/../../cpukit/score/src/threadhandler.c:88
#2993 2 years ago fixed score Chris Johns joel.sherrill@… 18 months ago
Summary

SMP assert in _Thread_Executing in libdebugger

Description

The target code in libdebugger has support to recover from exceptions related to invalid memory accesses. GDB may request the server access memory on the target that results in an exception. The exception occurs on the server's remote connection thread and the server needs to recover and return and error to GDB.

Running the debugger01 test with an SMP build of RTEMS and libbsd for xilinx_zedboard and issuing bt in GDB results in:

*** LIBBSD DEBUGGER 1 TEST ***                                                                                                                                                                                                                                                                            [144/1950]
shell:cannot set terminal attributes(/dev/console)

RTEMS Shell on /devn/ecxounss0o:l e<.R TUEsMeS  'Nheexlups'  dteov ilcies>t
 ccogmemma0n:d s<.C
adence CGEM Gigabit Ethernet Interface> on nexus0
miibus0: <MII bus> on cgem0
[/] # e1000phy0: <Marvell 88E1512 Gigabit PHY> PHY 0 on miibus0
e1000phy0:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT-FDX, 1000baseT-FDX-master, auto
cgem0: Ethernet address: fa:69:35:9e:04:2f
zy7_slcr0: <Zynq-7000 slcr block> on nexus0
[zone: udpcb] kern.ipc.maxsockets limit reached
notice: cgem0: link state changed to DOWN
add host 10.10.5.1: gateway cgem0
add net default: gateway 10.10.5.1
rtems-db: remote running
rtems-db: tcp remote: listing on port: 1122
notice: cgem0: link state changed to UP

rtems-db: tcp remote: connect host: 10.10.5.2
rtems-db: arm debug: (v3.0) ARMv7 [v7, all CP14 registers] breakpoints:5 watchpoints:3
assertion "cpu_self->thread_dispatch_disable_level != 0 || _ISR_Get_level() != 0" failed: file "../../cpukit/../../../xilinx_zynq_zedboard/lib/include/rtems/score/percpu.h", line 630, function: _Per_CPU_Get

If I enable TARGET_DEBUG in libdebugger and apply the attached patch I can create the assert with DIE_ON_ASSERT set to 1. The output is:

 rtems-db: tcp remote: connect host: 10.10.5.2
rtems-db: arm debug: (v3.0) ARMv7 [v7, all CP14 registers] breakpoints:5 watchpoints:3
[} frame = 005664EC sig=1 vector=4 ifsr=00000000 pra=0024173A
[}  R0 = 00000158  R1 = 00000004  R2 = 00000001  R3 = 0041AB64
[}  R4 = 00000158  R5 = 00000004  R6 = 00000000  R7 = 005606A4
[}  R8 = 00000016  R9 = 00000001 R10 = 00000006 R11 = 0041AB64
[} R12 = 00560658  SP = 00566540  LR = 00000FFD  PC = 00241736
[}  CPSR = 08010173 ----Q--A-FT GE:0 IT:01 M:13 SVC
[} target exception: 0 0 0
assertion "cpu_self->thread_dispatch_disable_level != 0 || _ISR_Get_level() != 0" failed: file "../../cpukit/../../../xilinx_zynq_zedboard/lib/include/rtems/score/percpu.h", line 630, function: _Per_CPU_Get

and set to {{{0}}:

rtems-db: tcp remote: connect host: 10.10.5.2
rtems-db: arm debug: (v3.0) ARMv7 [v7, all CP14 registers] breakpoints:5 watchpoints:3
[} frame = 005664EC sig=1 vector=4 ifsr=00000000 pra=0024173A
[}  R0 = 00000158  R1 = 00000004  R2 = 00000001  R3 = 0041AB64
[}  R4 = 00000158  R5 = 00000004  R6 = 00000000  R7 = 005606A4
[}  R8 = 00000016  R9 = 00000001 R10 = 00000006 R11 = 0041AB64
[} R12 = 00560658  SP = 00566540  LR = 00000FFD  PC = 00241736
[}  CPSR = 08010173 ----Q--A-FT GE:0 IT:01 M:13 SVC
[} target exception: 0 0 0
[} tid:0A01000A: thread:0041F5B0 frame:005664EC
[} server access fault
[} frame = 005664EC sig=1 vector=4 ifsr=00000000 pra=0024173A
[}  R0 = 00000158  R1 = 00000004  R2 = 00000001  R3 = 0041AB64
[}  R4 = 00000158  R5 = 00000004  R6 = 00000000  R7 = 005606A4
[}  R8 = 00000016  R9 = 00000001 R10 = 00000006 R11 = 0041AB64
[} R12 = 00560658  SP = 00566540  LR = 00000FFD  PC = 00241736
[}  CPSR = 08010173 ----Q--A-FT GE:0 IT:01 M:13 SVC
[} target exception: 0 0 0
[} tid:0A01000A: thread:0041F5B0 frame:005664EC
[} server access fault

The following lines first two values are cpu_self->thread_dispatch_disable_level and _ISR_Get_level() which are both 0 so I cannot see a reason the assert is happening:

[} target exception: 0 0 0
#2997 2 years ago fixed score Chris Johns Chris Johns 18 months ago
Summary

Monitor config command does not handle unlimited objects.

Description

Running the console's config command with unlimited objects gives:

[/] # config
INITIAL (startup) Configuration Info
------------------------------------------------------------------------------
WORKSPACE      start: 0x800f0173;  size: 0x374c8
TIME           usec/tick: 10000;  tick/timeslice: 50;  tick/sec: 100
MAXIMUMS       tasks: -2147483614;  timers: -2147483616;  sems: -2147483609;  que's: -2147483616;  ext's: 1
               partitions: -2147483616;  regions: -2147483616;  ports: -2147483616;  periods: -2147483616
#3000 2 years ago fixed score Chris Johns Joel Sherrill 3 months ago
Summary

Setting interrupt level in the mode arg on SMP returns RTEMS_UNSATISFIED

Description

If for any reason a user sets the interrupt level in the mode on an SMP build the error RTEMS_UNSATISFIED is returned. The documentation indicates this is a lack of stack and this confusing.

The reason this happens is the SMP check for an interrupt level being set is in the score's _Thread_Initialize. I propose that and is_preemptible check be converted to an assert and checks be added to the Classic API to catch these errors and report suitable error codes.

There is no meaningful error code available without abusing an existing one so I propose adding RTEMS_INVALID_MODE.

#3001 2 years ago fixed score Chris Johns Sebastian Huber 18 months ago
Summary

SMP build of RTEMS Testsuite does not set CONFIGURE_MAXIMUM_PROCESSORS

Description

The default setting for CONFIGURE_MAXIMUM_PROCESSORS is 1 and this means rtems_configuration_is_smp_enabled() returns false. Only the smptests set the maximum processor count to CPU_COUNT and therefore run in SMP mode.

If SMP is not running in an SMP build when running the tests are the tests really reporting a true indication of the of the system?

I would expect we have the API tests, libtests and fstests running with SMP enabled in an SMP build.

#3017 2 years ago worksforme score phongvanpham joel.sherrill@… 18 months ago
Summary

improvement in pci.h

Description

In pci.h, there are references to BSP_pci_configuration data structure which is in pci.c. However, in this file, there are also references to detect_host_bridge () in detect_raven_bridge.c. For folks that are just interested in pci_read_config_dword() + its brothers, all they need is to include pci.h and content for where BSP_pci_configuration is defined. The rest of the stuff in pci.c should be separate. Or in another word, data structures and #defines involving with BSP_pci_configuration needs to be in separate files rather all stuffed in pci.c

I currently do not need this functionality for my BSP (nor do I able to test it), so I cannot modify code and submit. It is best someone who can test the code to make the code change. Or else, just shelf it under the table and/or close this ticket.

#3056 22 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Add EDF SMP scheduler

Description

The current SMP schedulers are all fixed-priority schedulers. Add a job-level fixed priority scheduler (EDF).

#3059 22 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Add a simple processor affinity support to the EDF SMP scheduler

Description

Add support to the EDF SMP scheduler to honour one-to-one and one-to-all thread processor affinities. Use one ready queue for threads with a one-to-all affinity. Use one ready queue for each of the one-to-one threads for each processor. Since a red-black tree is used for the ready queues, the space overhead of one pointer per ready queue is small.

#3101 21 months ago fixed score Chris Johns Chris Johns 18 months ago
Summary

Add I2C Drivers for LM25066A, TMP112, ADS1113 and ADS1115

Description

Add drivers for:

  1. LM25066A
  2. TMP112
  3. ADS1113
  4. ADS1115
#3117 20 months ago fixed score Sebastian Huber Sebastian Huber 13 months ago
Summary

score: Optimize _Thread_queue_Enqueue() timeout handling

Description

Use the Thread_queue_Context::enqueue_callout to do the timeout handling. This avoids the switch statement in _Thread_queue_Timeout(). It removes the thread queue dependency to _Thread_Timeout().

#3124 20 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Ignore pshared attribute for POSIX semaphores

Description

Since we have only one process, sharing between processes is trivial.

#3140 19 months ago fixed score Chris Johns joel.sherrill@… 18 months ago
Summary

CPU Kit broken with --enable-rtems-debug

Description

Building with:

..../rtems.git/configure --target=arm-rtems4.12 --prefix=/opt/work/chris/rtems/kernel/4.12 --disable-networking --enable-rtemsbsp=beagleboneblack --enable-maintainer-mode --enable-rtems-debug

results in an error:

gmake[5]: Entering directory '/opt/work/chris/rtems/kernel/bsps/beagleboneblack/arm-rtems4.12/c/beagleboneblack/cpukit/score'
arm-rtems4.12-gcc --pipe -DHAVE_CONFIG_H   -I.. -I../../cpukit/../../../beagleboneblack/lib/include   -mcpu=cortex-a8 -O2 -g -ffunction-sections -fdata-sections -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -MT src/libscore_a-semaphore.o -MD -MP -MF src/.deps/libscore_a-semaphore.Tpo -c -o src/libscore_a-semaphore.o `test -f 'src/semaphore.c' || echo '/opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/score/'`src/semaphore.c
In file included from /opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/score/src/semaphore.c:21:0:
/opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/score/src/semaphore.c: In function '_Semaphore_Post':
/opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/score/src/semaphore.c:134:27: error: 'UINT_MAX' undeclared (first use in this function); did you mean 'UINT8_MAX'?
     _Assert( sem->count < UINT_MAX );
                           ^
../../cpukit/../../../beagleboneblack/lib/include/rtems/score/assert.h:67:12: note: in definition of macro '_Assert'
        ( ( _e ) ? \
            ^~
/opt/work/chris/rtems/kernel/rtems.git/c/src/../../cpukit/score/src/semaphore.c:134:27: note: each undeclared identifier is reported only once for each function it appears in
     _Assert( sem->count < UINT_MAX );
                           ^
../../cpukit/../../../beagleboneblack/lib/include/rtems/score/assert.h:67:12: note: in definition of macro '_Assert'
        ( ( _e ) ? \
            ^~
gmake[5]: *** [Makefile:4571: src/libscore_a-semaphore.o] Error 1

We need the rtems-bsp-builder to be run on a regular basis to catch these errors.

Tools are:

$ /opt/work/rtems/4.12/bin/arm-rtems4.12-gcc --version
arm-rtems4.12-gcc (GCC) 7.2.0 20170814 (RTEMS 4.12, RSB e6d0a8bae6d16eba605370ca11a5928b797820bb-modified, Newlib 2.5.0.20170818)
#3163 19 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Add I2C device driver for temperature sensor LM75A

#3167 19 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Internal status codes must not depend on RTEMS_POSIX_API

Description

The internal status codes encode a Classic rtems_status_code and error codes used by the POSIX and C11/C++11 APIs. In case the POSIX API is disabled, the C11/C++11 support must still work.

#3175 19 months ago fixed score Sebastian Huber Sebastian Huber 6 months ago
Summary

Merge FreeBSD timecounter changes from 2015-01-20 to now

#3185 19 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Change uptime seconds to int32_t

Description

The use of a 64-bit integer for the uptime seconds is overkill. Use int32_t instead for _Timecounter_Time_uptime. Change derived APIs, e.g. rtems_clock_get_uptime_seconds().

#3216 18 months ago fixed score Sebastian Huber Sebastian Huber 18 months ago
Summary

Replace vprintk() implementation

Description

The current vprintk() implementation has a questionable licence header, lacks support for the 'z' and 'j' format specifiers, is not robust against invalid format specifiers, uses a global variable for output. Replace it with a stripped down version of the FreeBSD kernel kvprintf() function.

#3236 17 months ago fixed score Sebastian Huber Sebastian Huber 17 months ago
Summary

Fix thread queue owner priority update in _Thread_queue_Flush_critical()

Description

The thread queue extract operations performed by the _Thread_queue_Flush_critical() may result in a priority change of the thread queue owner. Carry out the scheduler priority update operation. This is especially important in SMP configurations.

#3237 17 months ago fixed score Sebastian Huber Sebastian Huber 17 months ago
Summary

Fix priority ceiling updates

Description

We must not clear the priority updates in _Thread_queue_Extract_locked() since this function is used by the priority ceiling surrender operations after the ceiling priority handover from the previous owner to the new owner. This is especially important in SMP configurations.

Move the _Thread_queue_Context_clear_priority_updates() invocation to the callers.

#3243 17 months ago fixed score Sebastian Huber Sebastian Huber 7 months ago
Summary

Simplify global construction

Description

For the SMP support the global construction was changed to use an approach with a thread restart after global construction. With this implementation thread-local objects and POSIX keys initialized during global construction are not present in the initialization thread (main thread). This is not in line with what users familiar with GNU/Linux or FreeBSD would expect. See for example:

https://lists.rtems.org/pipermail/users/2017-July/031525.html

#3264 16 months ago fixed score Sebastian Huber Sebastian Huber 15 months ago
Summary

Add monotonic watchdog based on uptime

Description

The CLOCK_MONOTONIC time services use currently the clock tick based watchdog. This is a problem in case the uptime (measured via the timercounter) and the ticks since boot drift away (measured via the clock ticks). Introduce a new watchdog which uses the uptime. The memory overhead is quite small (two pointers per processor).

#3327 14 months ago fixed score Joel Sherrill 13 months ago
Summary

Eliminate score/cpu/*/.../types.h

Description

Each port contains a types.h file. It universally defines one type (CPU_Uint32ptr) that is required. Some of the types.h files define a CPU specific simple vectored ISR handler prototype.

  • Move the CPU_Uint32ptr typedef to cpu.h
  • If unused, delete the ISR handler prototype. If used, move to cpu.h
#3432 11 months ago wontfix score Joel Sherrill 7 months ago
Summary

Remove Simple SMP Priority Scheduler

Description

This scheduler was the first SMP scheduler added. It was created to have an easy SMP scheduler to debug. This was especially important when all of the SMP modifications and support were new. A Simple Scheduler has a use case as a low resource alternative for small uniprocessor systems. But the SMP variant just doesn't seem to have a good use case. If you have an SMP system, the application is almost certain to have enough resources where the more complicated data structures used by the other schedulers wouldn't be a burden. The Deterministic Priority Scheduler uses ~3K for FIFO with 256 priorities. This should not be an issue for an SMP system.

This ticket is a proposal to remove this as no longer having a use case.

#3459 10 months ago fixed score Sebastian Huber Sebastian Huber 6 weeks ago
Summary

Rework initialization and interrupt stack support

Description

We need an initialization stack to run the sequential system initialization before multitasking is enabled. The system initialization is done with interrupts disabled.

We need an interrupt stack for interrupt processing. This helps to avoid a per thread stack overhead for interrupt processing. The size for interrupt stack is application dependent, e.g. maximum interrupt nest level, stack demands of interrupt handlers.

The initialization and interrupt stacks are needed for each processor in the system.

Since interrupts are disabled during the sequential system initialization we can re-use the interrupt stack for the initialization stack. This is important for low end targets, with very limited RAM sizes. We need the initialization stack before a proper C run-time environment is set up e.g. we cannot assume that the access to global data is available. The stack memory area begin and size should be available via global symbols (named addresses). On some BSPs, e.g. ARM, this is done via the linker command file.

It should be possible to set the stack size via the CONFIGURE_INTERRUPT_STACK_SIZE configuration option and not via some magic stuff in linker command files.

Many BSPs set the BSS area to zero during system initialization. Thus, the initialization stack must not be contained in the BSS area.

The interrupt stack implementation is currently controlled by the following CPU port defines:

/**
 * Does RTEMS manage a dedicated interrupt stack in software?
 *
 * If TRUE, then a stack is allocated in @ref _ISR_Handler_initialization.
 * If FALSE, nothing is done.
 *
 * If the CPU supports a dedicated interrupt stack in hardware,
 * then it is generally the responsibility of the BSP to allocate it
 * and set it up.
 *
 * If the CPU does not support a dedicated interrupt stack, then
 * the porter has two options: (1) execute interrupts on the
 * stack of the interrupted task, and (2) have RTEMS manage a dedicated
 * interrupt stack.
 *
 * If this is TRUE, @ref CPU_ALLOCATE_INTERRUPT_STACK should also be TRUE.
 *
 * Only one of @ref CPU_HAS_SOFTWARE_INTERRUPT_STACK and
 * @ref CPU_HAS_HARDWARE_INTERRUPT_STACK should be set to TRUE.  It is
 * possible that both are FALSE for a particular CPU.  Although it
 * is unclear what that would imply about the interrupt processing
 * procedure on that CPU.
 *
 * Port Specific Information:
 *
 * XXX document implementation including references if appropriate
 */
#define CPU_HAS_SOFTWARE_INTERRUPT_STACK FALSE

/**
 * Does this CPU have hardware support for a dedicated interrupt stack?
 *
 * If TRUE, then it must be installed during initialization.
 * If FALSE, then no installation is performed.
 *
 * If this is TRUE, @ref CPU_ALLOCATE_INTERRUPT_STACK should also be TRUE.
 *
 * Only one of @ref CPU_HAS_SOFTWARE_INTERRUPT_STACK and
 * @ref CPU_HAS_HARDWARE_INTERRUPT_STACK should be set to TRUE.  It is
 * possible that both are FALSE for a particular CPU.  Although it
 * is unclear what that would imply about the interrupt processing
 * procedure on that CPU.
 *
 * Port Specific Information:
 *
 * XXX document implementation including references if appropriate
 */
#define CPU_HAS_HARDWARE_INTERRUPT_STACK TRUE

/**
 * Does RTEMS allocate a dedicated interrupt stack in the Interrupt Manager?
 *
 * If TRUE, then the memory is allocated during initialization.
 * If FALSE, then the memory is allocated during initialization.
 *
 * This should be TRUE is CPU_HAS_SOFTWARE_INTERRUPT_STACK is TRUE.
 *
 * Port Specific Information:
 *
 * XXX document implementation including references if appropriate
 */
#define CPU_ALLOCATE_INTERRUPT_STACK TRUE

Do the following steps to unify and simplify the initialization and interrupt stack support.

  1. Add RTEMS_DECLARE_GLOBAL_SYMBOL() and RTEMS_DEFINE_GLOBAL_SYMBOL() macros to basedefs.h, to allow a global symbol definition via C code, e.g. in confdefs.h, to make the interrupt stack size available to the low level initialization code.
  1. Add a special input section ".rtemsstack" to the linker command files to allow a placement of the interrupt stacks. The BSPs can provide the optimal memory location for this section, e.g. on-chip RAM, tightly-coupled memory.

This makes the CPU_HAS_SOFTWARE_INTERRUPT_STACK and CPU_HAS_HARDWARE_INTERRUPT_STACK CPU port defines superfluous, since the low level initialization code has all information available via global symbols.

This makes the CPU_ALLOCATE_INTERRUPT_STACK CPU port define superfluous, since the interrupt stacks are allocated by confdefs.h for all architectures. There is no need for BSP-specific linker command file magic.

The optional _CPU_Interrupt_stack_setup() is still useful to customize the registration of the interrupt stack area in the per-CPU information.

#3475 9 months ago fixed score Sebastian Huber Sebastian Huber 9 months ago
Summary

Add RTEMS_PREDICT_TRUE() and RTEMS_PREDICT_FALSE() for static branch prediction hints

Description

Add macros to <rtems/score/basedefs.h> for the GNU extension builtin_expect(). Use RTEMS_PREDICT_TRUE() and RTEMS_PREDICT_FALSE() similar to the FreeBSD predict_true() and predict_false(). Alternatives are the Linux likely() and unlikely() or directly the GCC builtin_expect(), however, the FreeBSD names seem to be the most easy to understand.

#3496 9 months ago fixed score Sebastian Huber Sebastian Huber 8 months ago
Summary

Remove superfluous interrupt enable in _Thread_Dispatch_enable()

Description

The _Thread_Dispatch_enable() contains a superfluous interrupt enable. This bug had probably no effect since the interrupt enable is idempotent on all CPU ports.

RTEMS_INLINE_ROUTINE void _Thread_Dispatch_enable( Per_CPU_Control *cpu_self )
{
  uint32_t disable_level = cpu_self->thread_dispatch_disable_level;

  if ( disable_level == 1 ) {
    ISR_Level level;

    _ISR_Local_disable( level );

    if (
      cpu_self->dispatch_necessary
#if defined(RTEMS_SCORE_ROBUST_THREAD_DISPATCH)
        || !_ISR_Is_enabled( level )
#endif
    ) {
      _Thread_Do_dispatch( cpu_self, level ); <-- This function enabled interrupts
    } else {
      cpu_self->thread_dispatch_disable_level = 0;
      _Profiling_Thread_dispatch_enable( cpu_self, 0 );
    }

    _ISR_Local_enable( level ); <-- Here we enable it again
  } else {
    _Assert( disable_level > 0 );
    cpu_self->thread_dispatch_disable_level = disable_level - 1;
  }
}
#3507 8 months ago fixed score Sebastian Huber Sebastian Huber 2 months ago
Summary

Add flexible per-CPU data

Description

Add means to declare, define and get custom per-CPU data. The API should cover the APIs defined by the Linux <linux/percpu.h> and FreeBSD <sys/pcpu.h> header files.

#3508 8 months ago fixed score Sebastian Huber Sebastian Huber 8 months ago
Summary

Add support for thread to processor pinning

Description

FreeBSD started to use lock-free data structures (Concurrency Kit) with epoch based reclamation (EBR) in May 2018. The goal of this synchronization approach is to avoid atomic read-modify-write operations in the fast path. The algorithms need highly efficient access to per-processor data. This gives raise to add a new feature to RTEMS: thread to processor pinning. Thread pinning is orthogonal to thread processor affinity and overrules the processor affinity settings of a thread. It is intended for temporary use in short critical sections which allow preemption.

#3539 7 months ago fixed score Sebastian Huber Sebastian Huber 7 months ago
Summary

Remove CPU_PROVIDES_IDLE_THREAD_BODY

Description

Remove the CPU_PROVIDES_IDLE_THREAD_BODY option to avoid unnecessary conditional compilation.

#3603 5 months ago fixed score Sebastian Huber Sebastian Huber 5 months ago
Summary

Remove support for 16-bit object identifiers

Description

The RTEMS_USE_16_BIT_OBJECT define is not set by an RTEMS port. Remove support for 16-bit object identifiers. If someone really wants to use RTEMS on a 16-bit target, then it is better to use self-contained objects instead of playing around with object identifier optimizations.

#3621 5 months ago fixed score Sebastian Huber Sebastian Huber 4 months ago
Summary

Statically initialize object information structures

Description

Statically initialize the object information structures to make the configuration easier to review and simplify the debugging.

The workspace size estimate generated by <rtems/confdefs.h> looks currently like this:

  const rtems_configuration_table Configuration = { ( ( ( (ssize_t) ((((((1 +
0) != 0 ? 1 : 0) * ((Objects_Maximum) ((1 + 0) & ~0x80000000U))) *
(sizeof(Configuration_Thread_control))) != 0 ? 1 : 0) * ((((((1 + 0) != 0 ? 1 :
0) * ((Objects_Maximum) ((1 + 0) & ~0x80000000U))) *
(sizeof(Configuration_Thread_control)) + (2 * sizeof(uintptr_t) +
(sizeof(Heap_Protection_block_begin) + sizeof(Heap_Protection_block_end)))) +
((((sizeof(Heap_Block)) + (8) - 1) - ((sizeof(Heap_Block)) + (8) - 1) % (8))) -
1) - (((((1 + 0) != 0 ? 1 : 0) * ((Objects_Maximum) ((1 + 0) & ~0x80000000U)))
* (sizeof(Configuration_Thread_control)) + (2 * sizeof(uintptr_t) +
(sizeof(Heap_Protection_block_begin) + sizeof(Heap_Protection_block_end)))) +
((((sizeof(Heap_Block)) + (8) - 1) - ((sizeof(Heap_Block)) + (8) - 1) % (8)))

[more than 500 similar lines]

1) - (((sizeof(Configuration_Initial_Extensions) /
sizeof((Configuration_Initial_Extensions)[0])) *
sizeof(User_extensions_Switch_control) + (2 * sizeof(uintptr_t) +
(sizeof(Heap_Protection_block_begin) + sizeof(Heap_Protection_block_end)))) +
((((sizeof(Heap_Block)) + (8) - 1) - ((sizeof(Heap_Block)) + (8) - 1) % (8))) -
1) % ((((sizeof(Heap_Block)) + (8) - 1) - ((sizeof(Heap_Block)) + (8) - 1) %
(8)))))) + 0 + 0 + (0 * 1024) + ((((2 * sizeof(uintptr_t) +
(sizeof(Heap_Protection_block_begin) + sizeof(Heap_Protection_block_end)))) +
(8) - 1) - (((2 * sizeof(uintptr_t) + (sizeof(Heap_Protection_block_begin) +
sizeof(Heap_Protection_block_end)))) + (8) - 1) % (8)) ),

The object controls reside on the heap even for fixed object count configuration. Using a statically allocated array makes it easier to find the objects during debugging.

#3734 3 weeks ago fixed score Sebastian Huber Sebastian Huber 2 weeks ago
Summary

Add RTEMS_CONST attribute

Description

Add RTEMS_CONST attribute to make the compiler specific attribute ((const)) available.

#3088 21 months ago fixed shell likangbei chrisj@… 18 months ago
Summary

shell test in testsuites\samples\fileio many COMMANDs is Lost

Description

testsuites\samples\fileio test on atsam BSP, init.c: #define CONFIGURE_SHELL_COMMANDS_INIT #define CONFIGURE_SHELL_COMMANDS_ALL

but when press s -> start shell,I only find three COMMANDs(help,alias,time),Other COMMANDs is Lost。 I remember that the previous version was normal。

Sorry! My English is bad!

#3096 21 months ago fixed shell Chris Johns Chris Johns 18 months ago
Summary

Shell internal commands should be public.

Description

A few of the functions held in cpukit/libmisc/shell/internal.h are useful in building system. For example rtems_shell_register_monitor_commands() and rtems_shell_execute_cmd().

The shell commands are important and systems may provide other scripting mechanisms, for example sequences in YAML files. Providing public access lets users know the functions are supported.

#3720 7 weeks ago fixed shell Sebastian Huber Sebastian Huber 7 weeks ago
Summary

mfill shell command uses the wrong arguments for the memset()

#2377 4 years ago wontfix tool Sebastian Huber 18 months ago
Summary

rtems_waf: Tools without a version are not supported

Description

waf configure --prefix=/opt/rtems --rtems=/opt/rtems --rtems-tools=/opt/rtems --rtems-bsps=i386/pc686 Setting top to : /scratch/git-rtems-libbsd Setting out to : /scratch/git-rtems-libbsd/build Could not find any architectures (complete log in /scratch/git-rtems-libbsd-upstream/build/config.log)

#2464 3 years ago wontfix tool Sebastian Huber 18 months ago
Summary

RSB: Tool patches use the RTEMS version

Description

In rtems/conifg/rtems-urls.bset the tool patches are set to an RTEMS version dependent directory. This makes re-use of the general purpose files quite difficult.

#2527 3 years ago wontfix tool Joel Sherrill Chris Johns 18 months ago
Summary

Move pc386/tools/bin2boot to rtems-tools

Description

Per discussion with Chris. Begin to eliminate BSP specific tools.

#2867 2 years ago fixed tool Stavros Passas Chris Johns 18 months ago
Summary

Fix exclude rule in rtems-test-check

Description

rtems-test-check is responsible of checking the testsuite configuration of a given BSP and adapt it based on a given "command". Currently for the "exclude" command, we never exclude anything, the output of the script is always the whole list of the input tests.

This happens because the script always starts with output = $tests and appends on the list the tests that are not excluded, so the output is always all the tests.

#3061 22 months ago invalid tool Chris Johns chrisj@… 18 months ago
Summary

including 'unistd.h' in C++ does not build.

Description

Including unistd.h in a C++ program does not compile with the RSB for today:

$ /opt/work/rtems/4.12/bin/arm-rtems4.12-g++ -B/opt/work/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib -B/opt/work/si/rtems/4.12/arm-rtems4.12/xilinx_zynq_zc706/lib -specs bsp_specs -qrtems -march=armv7-a -mthumb -mfpu=neon -mfloat-abi=hard -mtune=cortex-a9 -g -O2 u.cpp                                                                                     
In file included from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/_pthreadtypes.h:24:0,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/types.h:239,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/unistd.h:12,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/unistd.h:4,
                 from u.cpp:6:
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h: In function 'void CPU_AND_S(size_t, cpu_set_t*, const cpu_set_t*, const cpu_set_t*)':
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:147:3: error: 'BIT_AND2' was not declared in this scope
   BIT_AND2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~~
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:147:3: note: suggested alternative: 'BIT_AND'
   BIT_AND2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~~
   BIT_AND
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h: In function 'void CPU_OR_S(size_t, cpu_set_t*, const cpu_set_t*, const cpu_set_t*)':
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:159:3: error: 'BIT_OR2' was not declared in this scope
   BIT_OR2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:159:3: note: suggested alternative: 'BIT_OR'
   BIT_OR2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~
   BIT_OR
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h: In function 'void CPU_XOR_S(size_t, cpu_set_t*, const cpu_set_t*, const cpu_set_t*)':
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:171:3: error: 'BIT_XOR2' was not declared in this scope
   BIT_XOR2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~~
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:171:3: note: suggested alternative: 'BIT_OR'
   BIT_XOR2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~~
   BIT_OR
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h: In function 'void CPU_NAND_S(size_t, cpu_set_t*, const cpu_set_t*, const cpu_set_t*)':
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:183:3: error: 'BIT_NAND2' was not declared in this scope
   BIT_NAND2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~~~
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:183:3: note: suggested alternative: 'BIT_NAND'
   BIT_NAND2(_cpu_set_bits(setsize), destset, srcset1, srcset2);
   ^~~~~~~~~
   BIT_NAND
In file included from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:46:0,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/_pthreadtypes.h:24,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/types.h:239,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/sys/unistd.h:12,
                 from /opt/work/rtems/4.12/arm-rtems4.12/include/unistd.h:4,
                 from u.cpp:6:
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h: In function 'int CPU_COUNT_S(size_t, const cpu_set_t*)':
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:194:10: error: '__bitcountl' was not declared in this scope
   return BIT_COUNT(_cpu_set_bits(setsize), set);
          ^
/opt/work/rtems/4.12/arm-rtems4.12/include/sys/cpuset.h:194:10: note: suggested alternative: '__count'
#3091 21 months ago fixed tool Joel Sherrill Sebastian Huber 18 months ago
Summary

Core Dump in powerpc-rtems4.12-ld

Description

This appears to have been introduced by the new binutils. Impacts qemuprep and qemuprep-altivec BSPs.

gmake[8]: Entering directory `/data/home/joel/rtems-work/rtems-testing/rtems/build-powerpc-qemuprep-rtems/powerpc-rtems4.12/c/qemuprep/lib/libbsp/powerpc/motorola_powerpc/qemu_fakerom' powerpc-rtems4.12-ld -o qemu_fakerom.bin qemu_fakerom.o qemu_fakeres.o --oformat binary -nostdlib -Ttext 0xfff00000 --section-start=.romentry=0xfffffffc gmake[8]: * [qemu_fakerom.bin] Segmentation fault

#3102 21 months ago fixed tool Chris Johns Chris Johns 18 months ago
Summary

rtems-exeinfo does not decode ARM static constructors.

Description

The sections for ARM are not the same as other architectures.

#3103 20 months ago fixed tool Joel Sherrill chrisj@… 18 months ago
Summary

rtems-tools on CentOS 7 Build Failure

Description

A build failure for rtems-tools on the master has been reported on CentOS 7. This is my notes as I try to reproduce it.

[joel@localhost rtems-tools]$ ./waf configure Setting top to : /home/joel/rtems-work/rtems-tools Setting out to : /home/joel/rtems-work/rtems-tools/build Version : 4.12.78bbe4c1a31d (4.12) Checking for program 'python' : /usr/bin/python Checking for python version >= 2.6.6 : 2.7.5 Checking for 'gcc' (C compiler) : /usr/bin/gcc Checking for 'g++' (C++ compiler) : /usr/bin/g++ Checking for header alloca.h : yes Checking for header fcntl.h : yes Checking for header process.h : not found Checking for header stdlib.h : yes Checking for header string.h : yes Checking for header strings.h : yes Checking for header sys/file.h : yes Checking for header sys/stat.h : yes Checking for header sys/time.h : yes Checking for header sys/types.h : yes Checking for header sys/wait.h : yes Checking for header unistd.h : yes Checking for header vfork.h : not found Checking for function getrusage : yes Checking for program 'm4' : /usr/bin/m4 Checking for header sys/wait.h : yes Checking for function kill : yes Checking for 'gcc' (C compiler) : /usr/bin/gcc Checking for 'g++' (C++ compiler) : /usr/bin/g++ Checking for 'g++' (C++ compiler) : /usr/bin/g++ Checking for function open64 : not found Checking for function stat64 : not found 'configure' finished successfully (0.786s) ========================================================= [joel@localhost rtems-tools]$ ./waf -j 1 --verbose Waf: Entering directory `/home/joel/rtems-work/rtems-tools/build' [ 88/151] Compiling rtemstoolkit/rld-process.cpp 11:58:16 runner ['/usr/bin/g++', '-pipe', '-g', '-O2', '-Wall', '-Wextra', '-pedantic', '-Irtemstoolkit', '-I../rtemstoolkit', '-Irtemstoolkit/elftoolchain/libelf', '-I../rtemstoolkit/elftoolchain/libelf', '-Irtemstoolkit/elftoolchain/common', '-I../rtemstoolkit/elftoolchain/common', '-Irtemstoolkit/libiberty', '-I../rtemstoolkit/libiberty', '-DHAVE_CONFIG_H=1', '-DRTEMS_VERSION="4.12"', '-DRTEMS_RELEASE="4.12.78bbe4c1a31d"', '-DFASTLZ_LEVEL=1', '../rtemstoolkit/rld-process.cpp', '-c', '-o/home/joel/rtems-work/rtems-tools/build/rtemstoolkit/rld-process.cpp.7.o'] In file included from ../rtemstoolkit/libiberty/libiberty.h:42:0,

from ../rtemstoolkit/rld-process.cpp:64:

../rtemstoolkit/libiberty/ansidecl.h:169:64: error: new declaration ‘char* basename(const char*)’

# define ATTRIBUTE_NONNULL(m) attribute ((nonnull (m)))

../rtemstoolkit/libiberty/libiberty.h:112:64: note: in expansion of macro ‘ATTRIBUTE_NONNULL’

extern char *basename (const char *) ATTRIBUTE_RETURNS_NONNULL ATTRIBUTE_NONNULL(1);

In file included from ../rtemstoolkit/rld-process.cpp:24:0: /usr/include/string.h:599:26: error: ambiguates old declaration ‘const char* basename(const char*)’

extern "C++" const char *basename (const char *filename)

In file included from ../rtemstoolkit/libiberty/libiberty.h:42:0,

from ../rtemstoolkit/rld-process.cpp:64:

../rtemstoolkit/libiberty/ansidecl.h:169:64: error: declaration of ‘int vasprintf(char, const char*, va_list_tag*)’ has a different exception specifier

# define ATTRIBUTE_NONNULL(m) attribute ((nonnull (m)))

../rtemstoolkit/libiberty/ansidecl.h:198:80: note: in expansion of macro ‘ATTRIBUTE_NONNULL’

#define ATTRIBUTE_PRINTF(m, n) attribute ((format (printf, m, n))) ATTRIBUTE_NONNULL(m)

../rtemstoolkit/libiberty/libiberty.h:651:55: note: in expansion of macro ‘ATTRIBUTE_PRINTF’

extern int vasprintf (char , const char *, va_list) ATTRIBUTE_PRINTF(2,0);

In file included from ../rtemstoolkit/rld-process.cpp:23:0: /usr/include/stdio.h:399:12: error: from previous declaration ‘int vasprintf(char, const char*, va_list_tag*) throw ()’

extern int vasprintf (char restrict ptr, const char *restrict f,

Waf: Leaving directory `/home/joel/rtems-work/rtems-tools/build' Build failed

-> task in 'rld' failed with exit status 1:

{task 23048432: cxx rld-process.cpp -> rld-process.cpp.7.o}

['/usr/bin/g++', '-pipe', '-g', '-O2', '-Wall', '-Wextra', '-pedantic', '-Irtemstoolkit', '-I../rtemstoolkit', '-Irtemstoolkit/elftoolchain/libelf', '-I../rtemstoolkit/elftoolchain/libelf', '-Irtemstoolkit/elftoolchain/common', '-I../rtemstoolkit/elftoolchain/common', '-Irtemstoolkit/libiberty', '-I../rtemstoolkit/libiberty', '-DHAVE_CONFIG_H=1', '-DRTEMS_VERSION="4.12"', '-DRTEMS_RELEASE="4.12.78bbe4c1a31d"', '-DFASTLZ_LEVEL=1', '../rtemstoolkit/rld-process.cpp', '-c', '-o/home/joel/rtems-work/rtems-tools/build/rtemstoolkit/rld-process.cpp.7.o'] ================================================================== Looking down into libiberty.h, I picked on basename()

/* HAVE_DECL_* is a three-state macro: undefined, 0 or 1. If it is

undefined, we haven't run the autoconf check so provide the declaration without arguments. If it is 0, we checked and failed to find the declaration so provide a fully prototyped one. If it is 1, we found it so don't provide any declaration at all. */

#if !HAVE_DECL_BASENAME

#if defined (GNU_LIBRARY )
defined (linux) \
defined (FreeBSD) defined (OpenBSD) defined (NetBSD) \
defined (CYGWIN) defined (CYGWIN32) defined (MINGW32) \
defined (DragonFly) defined (HAVE_DECL_BASENAME)

extern char *basename (const char *) ATTRIBUTE_RETURNS_NONNULL ATTRIBUTE_NONNULL(1); #else /* Do not allow basename to be used if there is no prototype seen. We

either need to use the above prototype or have one from autoconf which would result in HAVE_DECL_BASENAME being set. */

#define basename basename_cannot_be_used_without_a_prototype #endif #endif ============================================ The native CentOS 7 has this definition of basename:

# ifndef basename /* Return the file name within directory of FILENAME. We don't

declare the function if the `basename' macro is available (defined in <libgen.h>) which makes the XPG version of this function available. */

# ifdef CORRECT_ISO_CPP_STRING_H_PROTO extern "C++" char *basename (char *filename)

THROW asm ("basename") nonnull ((1));

extern "C++" const char *basename (const char *filename)

THROW asm ("basename") nonnull ((1));

# else extern char *basename (const char *filename) THROW nonnull ((1)); # endif # endif #endif

============================== I think we are getting the C++ prototype from string.h and a conflicting C prototype from libiberty.h

#3128 20 months ago fixed tool Chris Johns chrisj@… 18 months ago
Summary

RTEMS Tools corvar does not build on Windows.

Description

The following error has appeared on Windows:

In file included from ../rtemstoolkit/elftoolchain/libelf/gelf.h:34:0,
                 from ../rtemstoolkit/rld-elf-types.h:29,
                 from ../rtemstoolkit/rld.h:72,
                 from ../rtemstoolkit/rld-process.h:31,
                 from ../tester/covoar/ObjdumpProcessor.h:16,
                 from ../tester/covoar/DesiredSymbols.h:18,
                 from ../tester/covoar/app_common.h:6,
                 from ../tester/covoar/app_common.cc:40:
../rtemstoolkit/elftoolchain/libelf/libelf.h:33:23: fatal error: sys/queue.h: No such file or directory
 #include <sys/queue.h>
                       ^
compilation terminated.

Waf: Leaving directory `D:/opt/rtems/rsb.git/rtems/build/rtH/rtems-tools.git/build'
#3129 20 months ago fixed tool Chris Johns joel 17 months ago
Summary

RTEMS Tools covoar build fails on Windows

Description

The following warnings and errors are present so the RSB tools do not finish and install:

[ 97/150] Compiling linkers/rtems-syms.cpp
[ 98/150] Compiling linkers/rtems-rapper.cpp
[ 99/150] Compiling linkers/rtems-exeinfo.cpp
In file included from ../rtemstoolkit/rld-files.cpp:30:0:
../rtemstoolkit/rld-files.cpp: In destructor 'virtual rld::files::image::~image()':
../rtemstoolkit/rld.h:111:75: warning: throw will always call terminate() [-Wterminate]
     rld::error (_what, std::string (__FILE__) + ":" + to_string (__LINE__))
                                                                           ^
../rtemstoolkit/rld-files.cpp:256:15: note: in expansion of macro 'rld_error_at'
         throw rld_error_at ("references when destructing image");
               ^~~~~~~~~~~~
../rtemstoolkit/rld.h:111:75: note: in C++11 destructors default to noexcept
     rld::error (_what, std::string (__FILE__) + ":" + to_string (__LINE__))
                                                                           ^
../rtemstoolkit/rld-files.cpp:256:15: note: in expansion of macro 'rld_error_at'
         throw rld_error_at ("references when destructing image");
               ^~~~~~~~~~~~

[100/150] Compiling tester/covoar/app_common.cc
[101/150] Compiling tester/covoar/CoverageFactory.cc
[102/150] Compiling tester/covoar/CoverageMap.cc
[103/150] Compiling tester/covoar/CoverageMapBase.cc
[104/150] Compiling tester/covoar/CoverageRanges.cc
[105/150] Compiling tester/covoar/CoverageReaderBase.cc
[106/150] Compiling tester/covoar/CoverageReaderQEMU.cc
[107/150] Compiling tester/covoar/CoverageReaderRTEMS.cc
[108/150] Compiling tester/covoar/CoverageReaderSkyeye.cc
[109/150] Compiling tester/covoar/CoverageReaderTSIM.cc
[110/150] Compiling tester/covoar/CoverageWriterBase.cc
[111/150] Compiling tester/covoar/CoverageWriterRTEMS.cc
[112/150] Compiling tester/covoar/CoverageWriterSkyeye.cc
[113/150] Compiling tester/covoar/CoverageWriterTSIM.cc
[114/150] Compiling tester/covoar/DesiredSymbols.cc
[115/150] Compiling tester/covoar/ExecutableInfo.cc
[116/150] Compiling tester/covoar/Explanations.cc
[117/150] Compiling tester/covoar/GcovData.cc
[118/150] Compiling tester/covoar/GcovFunctionData.cc
[119/150] Compiling tester/covoar/ObjdumpProcessor.cc
../tester/covoar/DesiredSymbols.cc: In member function 'void Coverage::DesiredSymbols::determineSourceLines(Coverage::CoverageRanges*, Coverage::ExecutableInfo*)':
../tester/covoar/DesiredSymbols.cc:517:36: error: 'realpath' was not declared in this scope
       realpath( inputBuffer, rpath );
                                    ^

Waf: Leaving directory `D:/opt/rtems/rtems-tools.git/build'
Build failed
 -> task in 'ccovoar' failed with exit status 1 (run with -v to display more information)
#3190 19 months ago fixed tool Chris Johns Chris Johns 18 months ago
Summary

RTEMS Tester covoar does not link on MacOS

Description

The executables do not link on MacOS.

#3191 19 months ago fixed tool Chris Johns Joel Sherrill 16 months ago
Summary

RTEMS Tester covoar dies with no arguments.

Description

Running covoar terminate with an unhandled exception with no arguments on the command line:

$ ./build/tester/covoar/covoar 
error missing option: target -T
Usage: ./build/tester/covoar/covoar [-v] -T TARGET -f FORMAT [-E EXPLANATIONS] -1 EXECUTABLE coverage1 ... coverageN
--OR--
Usage: ./build/tester/covoar/covoar [-v] -T TARGET -f FORMAT [-E EXPLANATIONS] -e EXE_EXTENSION -c COVERAGEFILE_EXTENSION EXECUTABLE1 ... EXECUTABLE2

  -v                        - verbose at initialization
  -T TARGET                 - target name
  -f FORMAT                 - coverage file format (RTEMS, QEMU, TSIM or Skyeye)
  -E EXPLANATIONS           - name of file with explanations
  -s SYMBOLS_FILE           - name of file with symbols of interest
  -1 EXECUTABLE             - name of executable to get symbols from
  -e EXE_EXTENSION          - extension of the executables to analyze
  -c COVERAGEFILE_EXTENSION - extension of the coverage files to analyze
  -g GCNOS_LIST             - name of file with list of *.gcno files
  -p PROJECT_NAME           - name of the project
  -C ConfigurationFileName  - name of configuration file
  -O Output_Directory       - name of output directory (default=.
libc++abi.dylib: terminating with uncaught exception of type std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >
Abort trap: 6
#3278 15 months ago fixed tool Joel Sherrill Chris Johns 15 months ago
Summary

bsp-builder has incorrect print (%s in output)

Description

I don't think the tools have branches so only impacts master.

Notice the "run: %s:"

[1114/1565] powerpc/mpc5674fevb (profiling) Configuring run: %s: powerpc/mpc5674fevb.profiling\

/home/joel/rtems-work/rtems/configure --target=powerpc-rtems5\ --enable-rtemsbsp=mpc5674fevb --prefix=/home/joel/rtems-work/bsps\ --enable-profiling

#3374 13 months ago fixed tool Joel Sherrill Chris Johns 6 months ago
Summary

rtems-test does not honor --mail-from argument

Description

This is on the master but may apply to other branches.

$ /home/joel/rtems-work/rtems-toolstester/rtems-test --rtems-tools=/home/joel/rtems-work/tools/5 --rtems-bsp=erc32 --log=run.log --mail --mail-from=joel@… --mail-to=build@… ./sparc-rtems5/c/erc32/testsuites/samples/base_sp/base_sp.exe error: no valid from address for mail

The rtems-test command will work if you have a ~/.mailrc with something like this:

set from="Joel Sherrill <joel@…>"

#3376 13 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove cklength program

Description

The cklength program (tools/build/cklength.c) has no license information and is unused in the RTEMS build. General usability is questionable, for example a

awk 'length($0) > 80' < file

performs a similar task. Remove it.

#3377 13 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove eolstrip program

Description

The eolstrip program (tools/build/eolstrip.c) has no license information and is unused in the RTEMS build. General usability is questionable, for example a

sed -i 's/[[:space:]]*$//' file

performs a similar task. Remove it.

#3378 13 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove unhex program

Description

The unhex program (tools/build/unhex.c) has no license information and is unused in the RTEMS build. Users of HEX files should consider to use ELF instead. Remove it.

#3379 13 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove packhex program

Description

The packhex program ( tools/build/packhex.c) is exported to the standard RTEMS build infrastructure via the PACKHEX variable. It is used by some legacy BSPs. It as unclear license information:

/*****  P A C K H E X . C  ************************************************
 *
 *   Packhex is a hex-file compaction utility.  It attempts to concatenate
 *   hex records to produce more size-efficient packaging.
 *
 *   Limitations: Input files must be correctly formatted.  This utility
 *                is not robust enough to detect hex-record formatting
 *                errors.
 *
 *   Published:   May 1993 Embedded Systems Programming magazine
 *                "Creating Faster Hex Files"
 *
 *   URL:         ESP magazine: http://www.embedded.com
 *                Source Code:  ftp://ftp.mfi.com/pub/espmag/1993/pakhex.zip
 *
 *   Author:      Mark Gringrich
 *
 *   Compiler:    Microsoft C 6.0
 *                cl /F 1000 packhex.c
 *
 **************************************************************************/

Move it to rtems-tools.

#3380 13 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Move rtems-bin2c program to rtems-tools

Description

The rtems-bin2c program (tools/build/rtems-bin2c.c) is exported to the standard RTEMS build infrastructure via the BIN2C variable. Move it to rtems-tools.

#3395 12 months ago fixed tool Chris Johns Chris Johns 12 months ago
Summary

rtems-ld does not remove executable when there is an output error

Description

An error when outputting an executable does not clean up the file and leaves an incorrect format file.

This is happening with the beagle bone black BSP and test dl06.

#3396 12 months ago fixed tool Chris Johns Chris Johns 12 months ago
Summary

rtems-ld does not handle R_ARM_V4BX relocation records

Description

The R_ARM_V4BX does not have a symbol and this raised an error with dl06 with a ARMv7 instruction set when merging sections when creating a RAP image.

Ignore this relocation record.

#3401 12 months ago fixed tool Joel Sherrill chrisj@… 6 months ago
Summary

dl06: tms570* Mixed LSB/MSB Error

Description

ld-arm-tms570ls3137_hdk-rtems/arm-rtems5/c/tms570ls3137_hdk/testsuites/libtests' rtems-ld -r /home/joel/rtems-work/rtems-testing/rtems/build-arm-tms570ls3137_hdk-rtems/arm-rtems5/c/tms570ls3137_hdk -O rap -b dl06.pre -e rtems_main -s \

-o dl06.rap dl06-o1.o dl06-o2.o -lm

error: elf:check_file: /data/home/joel/rtems-work/tools/5/bin/../lib/gcc/arm-rtems5/7.3.0/../../../../arm-rtems5/lib/libc.a:lib_a-_Exit.o@23760: Mixed data types not allowed (LSB/MSB).

#3402 12 months ago fixed tool Joel Sherrill chrisj@… 6 months ago
Summary

dl06: mips hurricane Mixed Endian Error

Description

Also occurs on rbtx4925 and rbtx4938

rtems-ld -r /home/joel/rtems-work/rtems-testing/rtems/build-mips-hurricane-rtems/mips-rtems5/c/hurricane -O rap -b dl06.pre -e rtems_main -s \

-o dl06.rap dl06-o1.o dl06-o2.o -lm

error: elf:check_file: /data/home/joel/rtems-work/tools/5/bin/../lib/gcc/mips-rtems5/7.3.0/../../../../mips-rtems5/lib/libc.a:lib_a-_Exit.o@23298: Mixed data types not allowed (LSB/MSB).

#3407 12 months ago fixed tool Joel Sherrill 6 months ago
Summary

Move Gaisler.org and Gaisler.se hosted RSB patches to rtems.org

Description

Jiri has suggested that the patches used in the RSB that are hosted by him be moved to rtems.org and the RSB patches link be changed.

This impacts at least qemu.

#3417 12 months ago fixed tool Chris Johns Chris Johns 6 months ago
Summary

Add libdwarf to elftoolchain and provide a C++ wrapper

Description

Update the elftoolchain and add libdwarf.

Provide a C++ framework to create reusable access to libdwarf.

#3418 12 months ago fixed tool Sebastian Huber Sebastian Huber 12 months ago
Summary

Remove difftest and sorttimes test tools

Description

Evaluation of test results and report generation should move to somewhere else, e.g. the RTEMS tester.

#3443 11 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove shgen program

Description

Rename it to rtems-shgen.

#3444 11 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove nios2gen program

Description

Rename it to rtems-nios2gen

#3445 11 months ago fixed tool Sebastian Huber Sebastian Huber 10 months ago
Summary

Remove multigen script

Description

This script is unused and out dated.

#3446 11 months ago fixed tool Sebastian Huber Sebastian Huber