Many failures on LEON3 with SMP disabled

[Joel Sherrill]

There are approximately ~100 failures, timeouts, etc on the LEON3 BSP. See this thread for some discussion where Jiri notes it is broken on his checkout from December:

​https://lists.rtems.org/pipermail/devel/2017-March/017277.html

Passed: 458
Failed: 20
Timeouts: 73
Invalid: 3

---

Total: 554

Failures:

cdtest.exe
spintrcritical20.exe
dl05.exe
spintrcritical01.exe
spintrcritical04.exe
spintrcritical10.exe
spintrcritical22.exe
sp69.exe
spintrcritical21.exe
sp11.exe
spintrcritical16.exe
spintrcritical23.exe
psxfile01.exe
spintrcritical05.exe
spintrcritical02.exe
spintrcritical08.exe
psxgetrusage01.exe
spcpucounter01.exe
spintrcritical03.exe
psxtimes01.exe

Timeouts:

nsecs.exe
sptask_err02.exe
spprivenv01.exe
psxkey03.exe
psxsignal01.exe
psx06.exe
psx10.exe
sp04.exe
mrfs_fstime.exe
ticker.exe
psxmsgq03.exe
psxkey09.exe
psx07.exe
sptimerserver01.exe
psxusleep.exe
psxstack02.exe
psxkey07.exe
psxkey10.exe
stackchk.exe
sp01.exe
fileio.exe
spsimplesched01.exe
sp03.exe
psxcond01.exe
sp65.exe
sp62.exe
psx11.exe
psx12.exe
psx02.exe
imfs_fstime.exe
crypt01.exe
psxstack01.exe
spcbssched01.exe
termios.exe
mimfs_fstime.exe
psxsignal02.exe
psx08.exe
top.exe
psxrwlock01.exe
sp22.exe
psxsignal04.exe
psxkey04.exe
mouse01.exe
sp24.exe
psx04.exe
spedfsched01.exe
uid01.exe
mdosfs_fstime.exe
psx16.exe
psxaio03.exe
sp19.exe
psxtime.exe
psx09.exe
psxkey06.exe
psxclock.exe
cpuuse.exe
psx05.exe
sp66.exe
psxsignal03.exe
capture.exe
sp30.exe
psxcleanup.exe
psxcancel.exe
jffs2_fstime.exe
psxsignal06.exe
spstdthreads01.exe
psxbarrier01.exe
sp31.exe
sp73.exe
psxualarm.exe
spfifo03.exe
psxtimer01.exe
monitor.exe

Invalid:

cxx_iostream.exe
spinternalerror01.exe
sptimecounter01.exe

[Chris Johns]

Replying to Joel Sherrill:

There are approximately ~100 failures, timeouts, etc on the LEON3 BSP. See this thread for some discussion where Jiri notes it is broken on his checkout from December:

​https://lists.rtems.org/pipermail/devel/2017-March/017277.html

Passed: 458
Failed: 20
Timeouts: 73
Invalid: 3

---

Total: 554

What are the results with SMP enabled?

[Daniel Hellstrom]

I have not managed to reproduce so many failures on GR712RC, GR740 or TSIM. Below are the failures I get from the run yesterday on non-SMP RTEMS configuration at GR712RC LEON3. The five fstests below have been discussed before and I believe they are related to the same faulty newlib configure used when the toolchain was built. Left is 4 tests which we will have to analyse.

I'm would argue that this should not block the 4.12 branching?

[usr1] #######################################################
[usr1] # RTEMS TESTSUITE FAILURE SUMMARY
[usr1] #
[usr1] # Result Test ExecRes? ConsoleRes? ExitCode1 ExitCode2
[usr1] # FAIL: ./fstests/imfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/jffs2_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/mdosfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/mimfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/mrfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./libtests/tar01 OK FAIL 5 0
[usr1] # FAIL: ./psxtests/psxsem01 OK FAIL 5 11
[usr1] # FAIL: ./sptests/spinternalerror01 OK N/A -559038737 1611526157
[usr1] # FAIL: ./sptests/sptimecounter01 OK N/A 5 0
[usr1] #
[usr1] # SUMMARY
[usr1] # Tests failing: 9
[usr1] # Tests successful: 534
[usr1] #
[usr1] #######################################################

[Sebastian Huber]

Which tool chain did you use?

Why are there fsscandir01 failures?

[Daniel Hellstrom]

It happens on both GCC-6 and GCC-7 series. It is our own build, I think it is the configuration of it that is an issue.

[usr1] # FAIL: ./fstests/imfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/jffs2_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/mdosfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/mimfs_fsscandir01 OK FAIL 5 0
[usr1] # FAIL: ./fstests/mrfs_fsscandir01 OK FAIL 5 0

all fail with similar error message:

[term1] * BEGIN OF TEST FSSCANDIR ROOT IMFS *
[term1] Initializing filesystem ROOT IMFS
[term1] scandir: Success
[term1] /opt/rtems-4.12/src/rtems-4.12/c/src/../../testsuites/fstests/imfs_fsscandir01/../fsscandir01/init.c: 47 MAXNAMLEN == NAME_MAX

[Sebastian Huber]

It would be quite nice if your RCC is next to identical to the tool chain produced by the RSB. This issue was fixed in January 2017 in the RTEMS master.

[Sebastian Huber]

Milestone renamed