Notice: We have migrated to GitLab launching 2024-05-01 see here: https://gitlab.rtems.org/

Changes between Version 1 and Version 2 of Debugging/sis


Ignore:
Timestamp:
07/31/17 13:45:07 (7 years ago)
Author:
Jiri Gaisler
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Debugging/sis

    v1 v2  
    11
    2 [[TOC(sis, depth=3)]]
     2[[TOC(Debugging/sis, depth=2)]]
    33
    44= SIS SPARC Simulator =
     
    6868}}}
    6969
    70 The above trace shows a data trap (trap type 0x9) at address 0x4002565c (ld  [ %g2 ], %i3). The load address in %g2 is 0x40FFFCD0, which is not a valid Leon2 address. To see which code generates this trap, we will need some symbolic information. Since SIS does not support debug symbols, we must turn to gdb with built-in sis:\
     70The above trace shows a data trap (trap type 0x9) at address 0x4002565c (ld  [ %g2 ], %i3). The load address in %g2 is 0x40FFFCD0, which is not a valid Leon2 address. To see which code generates this trap, we will need some symbolic information. Since SIS does not support debug symbols, we must turn to gdb with built-in sis:
    7171
    7272=== Debugging with SIS in gdb ===
     
    105105    at ../../../../../../gcc-7.1.0/newlib/libc/string/memcpy.c:87
    10610687      ../../../../../../gcc-7.1.0/newlib/libc/string/memcpy.c: No such file or directory.
     107
     108(gdb) bt
     109#0  memcpy (dst0=0x40fffcd0, src0=0xfffff000, len0=32)
     110    at ../../../../../../gcc-7.1.0/newlib/libc/string/memcpy.c:87
     111#1  0x4001487c in ambapp_scan2 (abus=0x40039320 <ambapp_plb>, ioarea=4293918720,
     112    memfunc=0x400255f0 <memcpy>, parent=0x0, root=0x40039320 <ambapp_plb>)
     113    at /home/jiri/ibm/src/rtems/rtems.git/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/amba/ambapp.c:177
     114#2  0x4001429c in amba_initialize ()
     115    at /home/jiri/ibm/src/rtems/rtems.git/c/src/lib/libbsp/sparc/leon3/amba/amba.c:99
     116#3  0x4001d148 in rtems_initialize_executive ()
     117    at /home/jiri/ibm/src/rtems/rtems.git/c/src/../../cpukit/sapi/src/exinit.c:138
     118#4  0x400134d4 in boot_card (cmdline=0x0)
     119    at /home/jiri/ibm/src/rtems/rtems.git/c/src/lib/libbsp/sparc/leon3/../../shared/bootcard.c:78
     120#5  0x400010c8 in zerobss ()
     121    at /home/jiri/ibm/src/rtems/rtems.git/c/src/lib/libbsp/sparc/leon3/../../sparc/shared/start/start.S:372
     122
    107123(gdb) up
    108124#1  0x4001487c in ambapp_scan2 (abus=0x40039320 <ambapp_plb>, ioarea=4293918720,
     
    124140}}}
    125141
    126 We can not see that the data trap occurs in memcpy, called by ambapp_scan2. The cause to the trap is rather trivial, we are trying to run a leon3 binary with sis configured for leon2! Running the binary with sis in leon3 mode will give us the desired result:
     142We can now see that the data trap occurs in memcpy, called by ambapp_scan2. The cause to the trap is rather trivial, we are trying to run a leon3 binary with sis configured for leon2! Running the binary with sis in leon3 mode will give us the desired result:
    127143
    128144{{{
     
    145161LOCAL: Hey I'm in base class constructor number 6 for 0x4003da68.
    146162LOCAL: Hey I'm in base class constructor number 7 for 0x4003da74.
    147 LOCAL: Hey I'm in derived class constructor number 8 for 0x4003da74.
    148 IO Stream not tested
    149 LOCAL: Hey I'm in derived class destructor number 8 for 4003DA74.
    150 Derived class - Instantiation order 8
    151 LOCAL: Hey I'm in base class destructor number 7 for 4003DA74.
    152 Derived class - Instantiation order 8
    153 LOCAL: Hey I'm in base class destructor number 6 for 4003DA68.
    154 Derived class - Instantiation order 6
    155 LOCAL: Hey I'm in base class destructor number 5 for 4003DA5C.
    156 Derived class - Instantiation order 5
    157 LOCAL: Hey I'm in base class destructor number 4 for 4003DA50.
    158 Derived class - Instantiation order 5
    159 *** TESTING C++ EXCEPTIONS ***
    160 
    161 foo_function() catch block called:
    162    < foo_function() throw this exception  >
    163 Success catching a char * exception
    164 foo_function() re-throwing execption...
    165 throw std::runtime: caught: thrown std::runtime object
    166 throw an instance based exception
    167 RtemsException raised=File:/home/jiri/ibm/src/rtems/rtems.git/c/src/../../testsuites/samples/cdtest/main.cc, Line:227, Error=55
    168 Success catching RtemsException...
    169 RtemsException ---> File:/home/jiri/ibm/src/rtems/rtems.git/c/src/../../testsuites/samples/cdtest/main.cc, Line:227, Error=55
    170 Exceptions are working properly.
    171 Global Dtors should be called after this line....
    172 GLOBAL: Hey I'm in derived class destructor number 3 for 400389FC.
    173 Derived class - Instantiation order 3
     163.
     164.
     165.
     166.
    174167GLOBAL: Hey I'm in base class destructor number 2 for 400389FC.
    175168Derived class - Instantiation order 3
     
    181174}}}
    182175
    183 To run the binary again, the gdb load command must be given followed by run.
    184 
    185 === Running test suites with SIS and rtems-tester ===
     176To run the binary again, the gdb '''load''' command must be given followed by '''run'''.
     177
     178=== Running test suites with SIS and rtems-test ===
    186179
    187180SIS can run the RTEMS test suites for erc32, leon2 and leon3 bsp in an automatic way using the rtems-test script. The test suites are available in the RTEMS build or install directory if RTEMS was configured with --enable-tests. Below is how to run the tests for leon3, but it works equally well erc32 and leon2. The tests will run in about 2 - 3 minutes on a high-end desktop.
     
    248241}}}
    249242
     243Link to DDD image: https://gaisler.org/gdb/doc/ddd1.png