old-cross-binutils/ld/testsuite/ld-aarch64/eh-frame-foo.s
Nick Clifton 2836a869c4 Fix linker testsuite failures for Aarch64.
Resolves PR 17312

	* ld-aarch64/eh-frame-foo.s: Use 8-byte offset for X30.
	* ld-aarch64/eh-frame-bar.s: Likewise.
	* ld-aarch64/eh-frame.d: Updated to match current readelf output.
	* ld-elf/eh5.d: Allow any register names.
2014-08-26 15:06:49 +01:00

55 lines
1 KiB
ArmAsm

__longjmp:
.cfi_startproc
.cfi_def_cfa x0, 0
.cfi_offset x19, 16
.cfi_offset x20, 16
.cfi_offset x21, 16
.cfi_offset x22, 16
.cfi_offset x23, 24
.cfi_offset x24, 24
.cfi_offset x25, 24
.cfi_offset x26, 24
.cfi_offset x27, 24
.cfi_offset x28, 32
.cfi_offset x29, 32
.cfi_offset x30, 40
.cfi_offset d9, 8
/* This eh frame data differs from eh-frame-bar.s here. The eh
frame information is identical but changes at the end. The
initial identical section is long enough to overflow the
initial instruction buffer used in eh frame merging. This
checks that merging does something sane once the initial
instruction buffer overflows. */
.cfi_offset d10, 8
ldp x19, x20, [x0, #16]
ldp x21, x22, [x0, #16]
ldp x23, x24, [x0, #24]
ldp x25, x26, [x0, #24]
ldp x27, x28, [x0, #24]
ldp x29, x30, [x0, #32]
ldp d8, d9, [x0, #8]
ldp d10, d11, [x0, #8]
ldp d12, d13, [x0, #8]
ldp d14, d15, [x0, #8]
ldr x5, [x0, #48]
mov sp, x5
cmp x1, #0
mov x0, #1
csel x0, x1, x0, ne
br x30
.cfi_endproc
.global _start
_start: