https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83108
Bug ID: 83108
Summary: Unaligned 128bit dereference causes segfault with -O2
on x86-64
Product: gcc
Version: 7.2.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c
Assignee: unassigned at gcc dot gnu.org
Reporter: gcc-bugspam at coralbark dot net
Target Milestone: ---
Created attachment 42681
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42681&action=edit
Preprocessed Source file
If I have a __uint128_t *unalignedPtr
Which points to an 8 byte (but not 16 byte) location then
*unalignedPtr = 0;
causes a seg fault on my x86-64 machine using:
gcc (GCC) 7.2.1 20170915 (Red Hat 7.2.1-2)
I'm attaching the example source file and the preprocessed source.
I compile with (to see the problem):
gcc -O2 128alignchk.c -o 128alignchk
I've tried turning on individual optimisation to track down the problem but if
I compile with:
gcc -O1 -fthread-jumps -falign-functions -falign-jumps -falign-loops
-falign-labels -fcaller-saves -fcrossjumping -fcse-follow-jumps
-fcse-skip-blocks -fdelete-null-pointer-checks -fdevirtualize
-fdevirtualize-speculatively -fexpensive-optimizations -fgcse -fgcse-lm
-fhoist-adjacent-loads -finline-small-functions -findirect-inlining -fipa-cp
-fipa-bit-cp -fipa-vrp -fipa-sra -fipa-icf
-fisolate-erroneous-paths-dereference -flra-remat -foptimize-sibling-calls
-foptimize-strlen -fpartial-inlining -fpeephole2 -freorder-blocks-algorithm=stc
-freorder-blocks-and-partition -freorder-functions -frerun-cse-after-loop
-fsched-interblock -fsched-spec -fschedule-insns2 -fstore-merging
-fstrict-aliasing -ftree-builtin-call-dce -ftree-switch-conversion
-ftree-tail-merge -fcode-hoisting -ftree-pre -ftree-vrp -fipa-ra
-fstrict-overflow -fvect-cost-model=cheap 128alignchk.c -o 128alignchk
The resulting binary works fine (even though the output of -Q --help=optimizers
is the same as with -O2).
