[This is a CI regression report that wasn't sent due to infra problem.]
Hi Florian,
Your glibc patch fixes a lot of failures in GDB testsuite, but also creates a few new FAILs -- are these expected?
Thanks!
=== Dear contributor,
Our automatic CI has detected problems related to your patch(es). Please find some details below.
In master-aarch64, after: | commit glibc-2.41.9000-556-g620f0730f31 | Author: Florian Weimer fweimer@redhat.com | Date: Mon Jul 28 14:16:52 2025 +0200 | | elf: Compile _dl_debug_state separately (bug 33224) | | This ensures that the compiler will not inline it, so that | debuggers which do not use the Systemtap probes can reliably | set a breakpoint on it. | ... 3 lines of the commit log omitted.
Produces 8 regressions 1564 fixes: | | regressions.sum: | Running gdb:gdb.dap/cxx-exception.exp ... | FAIL: gdb.dap/cxx-exception.exp: exceptions in log file | Running gdb:gdb.dap/ptrref.exp ... | FAIL: gdb.dap/ptrref.exp: exceptions in log file | Running gdb:gdb.threads/process-dies-while-detaching.exp ... | ... and 7 more | | fixes.sum: | Running gdb:gdb.arch/aarch64-sve-sigunwind.exp ... | FAIL: gdb.arch/aarch64-sve-sigunwind.exp: $\z2.b.u contents were correctly unwound | FAIL: gdb.arch/aarch64-sve-sigunwind.exp: size of $\z3.b.u was correctly unwound | FAIL: gdb.arch/aarch64-sve-sigunwind.exp: $\z12.b.u contents were correctly unwound | FAIL: gdb.arch/aarch64-sve-sigunwind.exp: $\z22.b.u contents were correctly unwound | ... and 1676 more
Used configuration : *CI config* tcwg_gnu_native_check_gdb master-aarch64 *configure and test flags:* none, autodetected on aarch64-unknown-linux-gnu
We track this bug report under https://linaro.atlassian.net/browse/GNU-692. Please let us know if you have a fix.
If you have any questions regarding this report, please ask on linaro-toolchain@lists.linaro.org mailing list.
-----------------8<--------------------------8<--------------------------8<--------------------------
The information below contains the details of the failures, and the ways to reproduce a debug environment:
You can find the failure logs in *.log.1.xz files in * https://ci.linaro.org/job/tcwg_gnu_native_check_gdb--master-aarch64-build/16... The full lists of regressions and improvements as well as configure and make commands are in * https://ci.linaro.org/job/tcwg_gnu_native_check_gdb--master-aarch64-build/16... The list of [ignored] baseline and flaky failures are in * https://ci.linaro.org/job/tcwg_gnu_native_check_gdb--master-aarch64-build/16...
Current build : https://ci.linaro.org/job/tcwg_gnu_native_check_gdb--master-aarch64-build/16... Reference build : https://ci.linaro.org/job/tcwg_gnu_native_check_gdb--master-aarch64-build/16...
Instruction to reproduce the build : https://gitlab.com/Linaro/tcwg/ci/interesting-commits/-/raw/master/glibc/sha...
Full commit : https://sourceware.org/git/?p=glibc.git%3Ba=commitdiff%3Bh=620f0730f311635cd...
-- Maxim Kuvyrkov https://www.linaro.org
* Maxim Kuvyrkov:
Produces 8 regressions 1564 fixes: | | regressions.sum: | Running gdb:gdb.dap/cxx-exception.exp ... | FAIL: gdb.dap/cxx-exception.exp: exceptions in log file
Traceback (most recent call last): File "/home/tcwg-buildslave/workspace/tcwg_gnu_2/abe/builds/destdir/aarch64-unknown-linux-gnu/libc/lib/libstdc++.so.6.0.34-gdb.py", line 69, in <module> from libstdcxx.v6 import register_libstdcxx_printers ModuleNotFoundError: No module named 'libstdcxx'
This may just be because the toolchain has not been ported to Debian's multi-arch paths.
It's possible that the previous failures obscured this problem, so it gets reported now?
Thanks, Florian
Hello Florian and Maxim,
Florian Weimer fweimer@redhat.com writes:
- Maxim Kuvyrkov:
Produces 8 regressions 1564 fixes: | | regressions.sum: | Running gdb:gdb.dap/cxx-exception.exp ... | FAIL: gdb.dap/cxx-exception.exp: exceptions in log file
Traceback (most recent call last): File "/home/tcwg-buildslave/workspace/tcwg_gnu_2/abe/builds/destdir/aarch64-unknown-linux-gnu/libc/lib/libstdc++.so.6.0.34-gdb.py", line 69, in <module> from libstdcxx.v6 import register_libstdcxx_printers ModuleNotFoundError: No module named 'libstdcxx'
This may just be because the toolchain has not been ported to Debian's multi-arch paths.
It's possible that the previous failures obscured this problem, so it gets reported now?
This patch restores the testsuite successes and failures from before your "elf: Restore support for _r_debug interpositions and copy relocations" patch series, so everything is good. Thanks for the fix!
The ModuleNotFoundError error is preexisting. It's a known problem due to how we install glibc in our CI scripts.
The high number of fixes is because the patch series mentioned above had broken the ld.so breakpoint hook used by GDB to detect loading of shared libraries (which affected many testcases) and this patch fixed it. We currently build glibc without system tap probes support. I will change that.
linaro-toolchain@lists.linaro.org
-
Florian Weimer -
Maxim Kuvyrkov -
Thiago Jung Bauermann