On Tue, Jun 09, 2020 at 09:34:01AM -0700, H.J. Lu via Gcc-patches wrote:
> > > * gcc.target/i386/pr93492-3.c: Likewise.
> > > * gcc.target/i386/pr93492-5.c: Likewise.
These tests FAIL on i686-linux.
E.g. in the first one I see
.file "pr93492-3.c"
.text
.globl f10_endbr
.type f10_endbr, @function
f10_endbr:
.LFB0:
.cfi_startproc
endbr32
.section __patchable_function_entries,"aw",@progbits
.align 4
.long .LPFE1
.text
.LPFE1:
nop
1: call __fentry__
pushl %ebp
.cfi_def_cfa_offset 8
.cfi_offset 5, -8
movl %esp, %ebp
.cfi_def_cfa_register 5
popl %ebp
.cfi_restore 5
.cfi_def_cfa 4, 4
ret
.cfi_endproc
.LFE0:
.size f10_endbr, .-f10_endbr
so it doesn't match the scan regexp, because
call __fentry__
is not immediately followed by
ret
As -pg is incompatible with -fomit-frame-pointer, I don't see anything wrong
on that.
Another thing in the test is that I don't think you can rely on
.cfi_startproc actually being printed, you should add an effective target
that will either check __GCC_HAVE_DWARF2_CFI_ASM macro definition, or check
for presence of .cfi_startproc on some trivial function compiled with
-fasynchronous-unwind-tables.
Jakub
