https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120429
Bug ID: 120429
Summary: pcmpeqd isn't used for all 1s in *movv2si_internal
Product: gcc
Version: 16.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: hjl.tools at gmail dot com
CC: liuhongt at gcc dot gnu.org, ubizjak at gmail dot com
Target Milestone: ---
[hjl@gnu-zen4-1 pr117839]$ cat dl-3.c
struct __pthread_mutex_s
{
int __lock;
unsigned int __count;
int __owner;
unsigned int __nusers;
int __kind;
short __spins;
short __elision;
void *p[2];
};
typedef union
{
struct __pthread_mutex_s __data;
char __size[40];
long int __align;
} pthread_mutex_t;
typedef struct { pthread_mutex_t mutex; } __rtld_lock_recursive_t;
void
foo (__rtld_lock_recursive_t *lock, int i)
{
lock[i] = (__rtld_lock_recursive_t) {{ { -1, -1, -1, -1, 1,
-1, -1, { ((void *)-1) , ((void *)-1) } } }};
}
[hjl@gnu-zen4-1 pr117839]$
/export/build/gnu/tools-build/gcc-debug/build-x86_64-linux/gcc/xgcc
-B/export/build/gnu/tools-build/gcc-debug/build-x86_64-linux/gcc/ -O2
-march=x86-64 -S dl-3.c -m32
[hjl@gnu-zen4-1 pr117839]$ cat dl-3.s
.file "dl-3.c"
.text
.p2align 4
.globl foo
.type foo, @function
foo:
.LFB0:
.cfi_startproc
movl 8(%esp), %eax
movl 4(%esp), %edx
movq .LC1, %xmm0 <<< pcmpeqd %xmm0, %xmm0 should be used instead
leal (%eax,%eax,4), %eax
leal (%edx,%eax,8), %eax
movl $-1, (%eax)
movl $-1, 4(%eax)
movl $-1, 8(%eax)
movl $-1, 12(%eax)
movl $1, 16(%eax)
movl $-1, 20(%eax)
movq %xmm0, 24(%eax)
ret
.cfi_endproc
.LFE0:
.size foo, .-foo
.section .rodata.cst8,"aM",@progbits,8
.align 8
.LC1:
.long -1
.long -1
.ident "GCC: (GNU) 16.0.0 20250524 (experimental)"
.section .note.GNU-stack,"",@progbits
[hjl@gnu-zen4-1 pr117839]$
