GNU bug report logs - #76180
[feature/igc] Remaining known tracing issues

Previous Next

Full log

Message #41 received at 76180 <at> debbugs.gnu.org (full text, mbox):

From: Pip Cet <pipcet <at> protonmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: eggert <at> cs.ucla.edu, 76180 <at> debbugs.gnu.org, mattiase <at> acm.org,
 gerd.moellmann <at> gmail.com, stefankangas <at> gmail.com, acorallo <at> gnu.org
Subject: Re: bug#76180: [feature/igc] Remaining known tracing issues
Date: Fri, 14 Feb 2025 12:57:49 +0000

"Eli Zaretskii" <eliz <at> gnu.org> writes:

>> Date: Fri, 14 Feb 2025 12:31:15 +0000
>> From: Pip Cet <pipcet <at> protonmail.com>
>> Cc: acorallo <at> gnu.org, stefankangas <at> gmail.com, 76180 <at> debbugs.gnu.org, gerd.moellmann <at> gmail.com, mattiase <at> acm.org, eggert <at> cs.ucla.edu
>>
>> "Eli Zaretskii" <eliz <at> gnu.org> writes:
>>
>> > Alternatively, maybe the nativecomp code could be modified not to use
>> > RBP?
>>
>> I'm not sure what you mean: the code is generated by libgccjit.
>
> Oh, I didn't realize that.  If we really have no control over this,
> then maybe a bug report against libgccjit is in order?

I think I misexpressed myself.  The problem is that libgccjit acts like
GCC, generating new native code which may or may not use %rbp depending
on the compiler options specified by comp.c.  Old versions of libgccjit
had no way of passing such compiler options, but current ones do.

The question is whether we're happy appending "-fno-omit-frame-pointer",
or whether we want to be ambitious and refuse to accept
"-fomit-frame-pointer".

I haven't tried this, but passing the wrong options to libgccjit might
make it assume the wrong ABI or hardware, so causing a crash for a
malicious user is probably not something we can avoid; the only question
is how we keep bug reports focussed so we don't have to keep this option
in mind for every single bug.

Pip

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.