https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77608
Siddhesh Poyarekar <siddhesh at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |WAITING
--- Comment #10 from Siddhesh Poyarekar <siddhesh at gcc dot gnu.org> ---
I'm inclined to not specifically handle volatile offsets and leave it at the
fix that I just pushed, i.e. falling back to whole object size for
__builtin_object_size (p, {0,1}) if the offset is not a constant.
In general, handling trees with side-effects in the pass would be too much work
for what seems like very little gain; volatile offsets don't seem like a common
enough case to actually care about. I'll close this off in a couple of weeks
if there are no objections.
