https://bugs.kde.org/show_bug.cgi?id=514094

--- Comment #15 from Mark Wielaard <[email protected]> ---
(In reply to Paul Floyd from comment #14)
> I think that, finally, this is OK now.

Yes, looks good. For the record this consisted of the following commits:

commit bf154d815a9fd7f4aaae97e31aa03cecf69448d4
Author: Paul Floyd <[email protected]>
Date:   Sat Jan 3 18:24:34 2026 +0100

    Solaris: set VG_(resolved_exename) in load_client()

    Haven't needed it yet, but I would like to try using it in the
    readlink syscall wrapper.

commit 5c0f5e604bc3a4e8822ea59c488c123af6284afd
Author: Paul Floyd <[email protected]>
Date:   Sat Jan 3 13:48:50 2026 +0100

    Bug 514094 - readlink("/proc/self/exe") overwrites buffer beyond its return 
value

    Used the reproducer as the basis for a test on Solaris and Linux.

commit bd9edb8fcd0a8692d865e08fab2a573a4cde4c16
Author: Paul Floyd <[email protected]>
Date:   Sat Jan 3 21:17:00 2026 +0100

    readlink[at] syswrap: limit copy to bufsiz when path is proc self exe

commit 987034c44105cdc2f6f8d84751135d23bd5c37b6
Author: Paul Floyd <[email protected]>
Date:   Sat Jan 3 22:16:58 2026 +0100

    Regtest: add missing readlinkat_self files

commit dae37ecd2692e0e5beba77c296c2648ebbf47637
Author: Paul Floyd <[email protected]>
Date:   Sat Jan 3 22:32:56 2026 +0100

    regtest: fix warning

    Added a nice const named variable then didn't use it

commit 294742a2d9f431fd2dcd73db161f67fb12ddd833
Author: Mark Wielaard <[email protected]>
Date:   Sat Jan 3 22:50:28 2026 +0100

    Fix bug514094,vgtest typo in none/tests/Makefile.am

commit 8d8023d107699c7c2d97acf2dcb77bae71c0b1cf
Author: Paul Floyd <[email protected]>
Date:   Sun Jan 4 09:34:54 2026 +0100

    syswrap readlink and linux readlinkat: check that buf is accessible for
proc
 self exe case

    Also update the t testcases to cover this.

commit 827a1b8c307a2eafa001788565e14af3445f2151
Author: Paul Floyd <[email protected]>
Date:   Sun Jan 4 10:17:15 2026 +0100

    solaris readlinkat: check that buf is accessible for proc self path a.out

I am going to try to backport this to the (not yet existing)
VALGRIND_3_26_BRANCH because it can be considered a regression.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to