[cfe-dev] [llvm-dev] Issue with the function hash used as PGO index
Xinliang David Li via cfe-dev
cfe-dev at lists.llvm.org
Fri May 15 11:14:20 PDT 2020
FYI, there is also IR PGO that computes function hash using CFG. To turn it
on, use -fprofile-generate and -fprofile-use options.
On Thu, May 14, 2020 at 11:59 PM Serge Guelton via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
> Hi Folks,
> This review
> fixes a relatively serious bug in the computation of the function hash
> used to
> compute index in profile files in clang. Basically the bug was allowing
> hash collision of two function that only differ by their last
> statement(s). This
> could lead to segfaukts like this one .
> I'm happy with the patch (if you want to review it, please jump in) but
> I'm also
> worried that it changes the hash of *most* functions, thus invalidating all
> profile files generated with previous clang revision.
> So if it lands, people will find themselves with a lot of obsolete profile
> that need to be regenerated... I think I'd at least state that clearly on
> mlist, due to the potential large impact.
> - Serge
>  https://bugzilla.redhat.com/show_bug.cgi?id=1827282
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cfe-dev