[Bug target/88005] [7/8/9 Regression] GCC on hppa misoptimisation function pointer call
skrll at netbsd dot org
gcc-bugzilla@gcc.gnu.org
Tue Nov 13 15:21:00 GMT 2018
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88005
--- Comment #2 from Nick Hudson <skrll at netbsd dot org> ---
yes, it does.
Nick
nick@zoom:~/gcc$ ./inst/bin/hppa--linux-gcc -O2 -fno-delete-null-pointer-checks
-S -o pdc.s pdc.i
nick@zoom:~/gcc$ cat pdc.s
.LEVEL 1.1
.text
.align 4
.globl pdc_init
.type pdc_init, @function
pdc_init:
.PROC
.CALLINFO FRAME=64,CALLS,SAVE_RP
.ENTRY
stw %r2,-20(%r30)
addil LR'pdc-$global$,%r27
ldo 64(%r30),%r30
ldw 0(%r0),%r22
ldo RR'pdc-$global$(%r1),%r1
stw %r0,-52(%r30)
ldi 0,%r23
stw %r22,0(%r1)
addil LR'pdcbuf-$global$,%r27
ldi 2,%r25
ldo RR'pdcbuf-$global$(%r1),%r24
ldi 10,%r26
bb,>=,n %r22,30,.+12
ldw 2(%r22),%r19
ldw -2(%r22),%r22
ble 0(%sr4,%r22)
copy %r31,%r2
ldw -84(%r30),%r2
bv %r0(%r2)
ldo -64(%r30),%r30
.EXIT
.PROCEND
.size pdc_init, .-pdc_init
.comm pdcbuf,256,64
.comm pdc,4,4
.ident "GCC: (GNU) 9.0.0 20181112 (experimental)"
nick@zoom:~/gcc$
(In reply to Alexander Monakov from comment #1)
> Isn't simply because the code dereferences a null pointer when assigning to
> 'pdc'? Does it "work" as expected with -fno-delete-null-pointer-checks?
More information about the Gcc-bugs
mailing list