[Bug debugedit/33991] [5.3] test failures on aarch64-linux-gnu and hppa-linux-gnu

[mark at klomp dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=33991

Mark Wielaard <mark at klomp dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

--- Comment #5 from Mark Wielaard <mark at klomp dot org> ---
(In reply to Matthias Klose from comment #2)
> as I wrote, the test-summary files are dumped in the build log. Is this good
> enough?

O, right, sorry, I missed that the log file was accessible from the build page.

So for arm64 it is indeed what I suspected:
section [33] '.ARM.attributes' has unsupported type 1879048195
That should be fixed by elfutils patch:
https://sourceware.org/cgit/elfutils/patch/?id=b27adc5262e807f341ca0a4910ce04294144f79a

For hppa it will be harder to fix in elfutils/eu-elflint.
There is not hppa/parisc backend and eu-elflint doesn't know about any quirks.
It would probably best to add something like the following to the new tests to
skip them if eu-elflint doesn't even pass on the simple executable.

diff --git a/tests/find-debuginfo.at b/tests/find-debuginfo.at
index ba4beabffc09..8afb3058e00c 100644
--- a/tests/find-debuginfo.at
+++ b/tests/find-debuginfo.at
@@ -463,6 +463,8 @@ AT_CLEANUP
 AT_SETUP([find-debuginfo check-elf])
 AT_KEYWORDS([find-debuginfo] [check-elf])
 FIND_DEBUGINFO_PKG_BUILD_SETUP
+# Make sure eu-elflint --gnu work on the simple input file.
+AT_SKIP_IF([! eu-elflint --gnu ./subdir_build/bar | grep -q 'No errors'])
 # We need to set some environment variables for running find-debuginfo
 # normally set by rpmbuild.
 AT_CHECK([env RPM_BUILD_DIR=${PWD} \
@@ -480,6 +482,8 @@ AT_CLEANUP
 AT_SETUP([find-debuginfo check-elf debugdata])
 AT_KEYWORDS([find-debuginfo] [check-elf] [debugdata])
 FIND_DEBUGINFO_PKG_BUILD_SETUP
+# Make sure eu-elflint --gnu work on the simple input file.
+AT_SKIP_IF([! eu-elflint --gnu ./subdir_build/bar | grep -q 'No errors'])
 # We need to set some environment variables for running find-debuginfo
 # normally set by rpmbuild.
 #
@@ -501,6 +505,8 @@ AT_KEYWORDS([find-debuginfo] [check-elf] [debugdata]
[gdb-index])
 # Too new gdb with too old gdb
 AT_SKIP_IF([test "$GDB_ADD_INDEX_CHECKS_OK" = "no"])
 FIND_DEBUGINFO_PKG_BUILD_SETUP
+# Make sure eu-elflint --gnu work on the simple input file.
+AT_SKIP_IF([! eu-elflint --gnu ./subdir_build/bar | grep -q 'No errors'])
 # We need to set some environment variables for running find-debuginfo
 # normally set by rpmbuild.
 #
@@ -525,6 +531,8 @@ AT_SKIP_IF([test "$DWARF_5_DEBUGADDR" = "yes"])
 # Too new gdb with too old gdb
 AT_SKIP_IF([test "$GDB_ADD_INDEX_CHECKS_OK" = "no"])
 FIND_DEBUGINFO_PKG_BUILD_SETUP
+# Make sure eu-elflint --gnu work on the simple input file.
+AT_SKIP_IF([! eu-elflint --gnu ./subdir_build/bar | grep -q 'No errors'])
 # We need to set some environment variables for running find-debuginfo
 # normally set by rpmbuild.
 #

But even then things probably keep failing because of

  51: find-debuginfo.at:347 find-debuginfo sources debugdata gdb-index dwz jobs
      find-debuginfo sources debugdata gdb-index dwz jobs

+The futex facility returned an unexpected error code.
+
+
+Fatal signal: Aborted
+----- Backtrace -----
+0xfd2ff ???
+0x28eb33 ???
+0x78037387 ???
+0x79068c80 ???
+0x7900ede7 gsignal
+0x78ff6d9f abort
+---------------------
+A fatal error internal to GDB has been detected, further
+debugging is not possible.  GDB will now terminate.
+
+This is a bug, please report it.  For instructions, see:
+<https://www.gnu.org/software/gdb/bugs/>.
+
+Aborted
+gdb-add-index: gdb error generating index for
/<<PKGBUILDDIR>>/tests/testsuite.dir/51/subdir_build/bar
+*** ERROR:: GDB exited with exit status 134 during index generation

So that seems a bug in GDB or even glibc or the kernel?

-- 
You are receiving this mail because:
You are on the CC list for the bug.