https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127420
Bug ID: 127420
Summary: Emit UD2 after indirect jump
Product: gcc
Version: 17.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: hjl.tools at gmail dot com
Target Milestone: ---
Target: x86
The Intel optimization manual says:
| Assembly/Compiler Coding Rule 13. (M impact, L generality) When
| indirect branches are present, try to put the most likely target of
| an indirect branch immediately following the indirect
| branch. Alternatively, if indirect branches are common but they
| cannot be predicted by branch prediction hardware, then follow the
| indirect branch with a UD2 instruction, which will stop the
| processor from decoding down the fall-through path.
[hjl@gnu-tgl-3 tmp]$ cat x.c
extern void (*foo) (void);
extern void bar (void);
void
func1 (void)
{
foo ();
}
void
func2 (void)
{
bar ();
}
[hjl@gnu-tgl-3 tmp]$ gcc -O2 -S -fPIC -fno-plt x.c
[hjl@gnu-tgl-3 tmp]$ cat x.s
.file "x.c"
.text
.p2align 4
.globl func1
.type func1, @function
func1:
.LFB0:
.cfi_startproc
movq foo@GOTPCREL(%rip), %rax
jmp *(%rax)
.cfi_endproc
.LFE0:
.size func1, .-func1
.p2align 4
.globl func2
.type func2, @function
func2:
.LFB1:
.cfi_startproc
jmp *bar@GOTPCREL(%rip)
.cfi_endproc
.LFE1:
.size func2, .-func2
.ident "GCC: (GNU) 16.2.1 20260819 (Red Hat 16.2.1-2)"
.section .note.GNU-stack,"",@progbits
[hjl@gnu-tgl-3 tmp]$
We should generate UD2 after indirect jump if not optimizing for space.
