Clarification on CVE-2025-1176 Fix Backport – Bug 32636 in binutils-2_44 and other older Branch

[Dora, Sunil Kumar]

Dear Binutils Maintainers,

I'm writing regarding Bug 32636<https://sourceware.org/bugzilla/show_bug.cgi?id=32636>, which concerns a heap-buffer-overflow in ld triggered in _bfd_elf_gc_mark_rsec (bfd/elflink.c). I see that the issue has been resolved in the master branch via commit f9978defb6fab0bd8583942d97c112b0932ac814<https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=f9978defb6fab0bd8583942d97c112b0932ac814>.

I attempted to backport this fix to the binutils-2_44 branch and 2_42 as well, but unfortunately the issue still reproduces on my end when using the same proof-of-concept input. I’ve double-checked the patch application, and the build completes without errors, yet AddressSanitizer continues to report the same heap-buffer-overflow.

Could you please advise:

Whether additional patches or supporting changes are needed beyond
that commit to fully address the issue in 2.44?

If this CVE fix has already been backported to the 2.44 branch, and
I may have missed something?

I'd greatly appreciate any insights or guidance you could offer. Please let me know if you would like me to share more details, such as my backported patch, testing output, or build configuration.

Thank you for your time and for all the work you do maintaining Binutils.

My config :
CFLAGS="-fsanitize=address -DFORTIFY_SOURCE=2 -fstack-protector-all -fno-omit-frame-pointer -g -Wno-error" LDFLAGS=" -fsanitize=address -ldl -lutil" ../binutils-gdb/configure --disable-shared --disable-gdb --disable-libdecnumber --disable-readline --disable-sim --prefix=<Custom Path>

Best regards,
Sunil Dora

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://sourceware.org/pipermail/binutils/attachments/20250414/8bd520ec/attachment.htm>