| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
From Linux 80f23935cadb ("powerpc: Convert cmp to cmpd in idle enter sequence"):
PowerPC's "cmp" instruction has four operands. Normally people write
"cmpw" or "cmpd" for the second cmp operand 0 or 1. But, frequently
people forget, and write "cmp" with just three operands.
With older binutils this is silently accepted as if this was "cmpw",
while often "cmpd" is wanted. With newer binutils GAS will complain
about this for 64-bit code. For 32-bit code it still silently assumes
"cmpw" is what is meant.
In this instance the code comes directly from ISA v2.07, including the
cmp, but cmpd is correct. Backport to stable so that new toolchains can
build old kernels.
This is change is a noop with the existing toolchain. We change from
implicitly generating the cmpwi to explicitly stating in for
compatibility with newer toolchains.
With gcc 4.9.3, binutils 2.25.2:
$ cat asm-test.S
.text
.global test
cmpi 0, 8, 1
$ powerpc64-linux-gcc -c asm-test.S
$ objdump -d asm-test.o
0000000000000000 <.text>:
2c 08 00 01 cmpwi r8,1
Old compiler, updated instruction:
$ cat asm-test.S
.text
.global test
cmpwi 0, 8, 1
$ powerpc64-linux-gcc -c asm-test.S
$ objdump -d asm-test.o
0000000000000000 <.text>:
2c 08 00 01 cmpwi r8,1
And then the new toolchain (gcc 6.3.0, binutils 2.28) with the updated
asm:
$ cat asm-test.S
.text
.global test
cmpwi 0, 8, 1
$ powerpc64-linux-gnu-gcc -c asm-test.S
$ objdump -d asm-test.o
0000000000000000 <.text>:
2c 08 00 01 cmpwi r8,1
Change-Id: If981e20a578ec98ede68a31eee2888c27d5c3d10
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/42259
Tested-by: Jenkins Server <pfd-jenkins+hostboot@us.ibm.com>
Tested-by: FSP CI Jenkins <fsp-CI-jenkins+hostboot@us.ibm.com>
Reviewed-by: STEWART E. SMITH <stewart@linux.vnet.ibm.com>
Reviewed-by: Daniel M. Crowell <dcrowell@us.ibm.com>
|
