diff options
| author | Roger Sayle <roger@nextmovesoftware.com> | 2022-03-26 08:10:27 -1000 |
|---|---|---|
| committer | Roger Sayle <roger@nextmovesoftware.com> | 2022-03-26 08:10:27 -1000 |
| commit | 41d1f11f5f693a2a06c65c9467a28dfeb02aed85 (patch) | |
| tree | 7c80f1a2348d793c1f3d046d37cfdd5d89ad9b6e /gcc/cp/constraint.cc | |
| parent | 6459e6537632bc06e04e6011ca7fb6488f0e8e7d (diff) | |
| download | gcc-41d1f11f5f693a2a06c65c9467a28dfeb02aed85.zip gcc-41d1f11f5f693a2a06c65c9467a28dfeb02aed85.tar.gz gcc-41d1f11f5f693a2a06c65c9467a28dfeb02aed85.tar.bz2 | |
PR middle-end/104885: Fix ICE with large stack frame on powerpc64.
My recent testcase for PR c++/84964.C stress tests the middle-end by
attempting to pass a UINT_MAX sized structure on the stack. Although
my fix to PR84964 avoids the ICE after sorry on x86_64 and similar
targets, a related issue still exists on powerpc64 (and similar
ACCUMULATE_OUTGOING_ARGS/ARGS_GROW_DOWNWARD targets) which don't
issue a "sorry, unimplemented" message, but instead ICE elsewhere.
After attempting several alternate fixes, the simplest solution is
to just defensively check in mark_stack_region_used that the upper
bound of the region lies within the allocated stack_usage_map
array, which is of size highest_outgoing_arg_in_use. When this isn't
the case, the code now follows the same path as for variable sized
regions, and uses stack_usage_watermark rather than a map.
2022-03-26 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
PR middle-end/104885
* calls.cc (mark_stack_region_used): Check that the region
is within the allocated size of stack_usage_map.
Diffstat (limited to 'gcc/cp/constraint.cc')
0 files changed, 0 insertions, 0 deletions