[Bug c/82118] New: Instruction `lsrsne` unknown
pmenzel+gcc at molgen dot mpg.de
gcc-bugzilla@gcc.gnu.org
Wed Sep 6 11:36:00 GMT 2017
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82118
Bug ID: 82118
Summary: Instruction `lsrsne` unknown
Product: gcc
Version: 7.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c
Assignee: unassigned at gcc dot gnu.org
Reporter: pmenzel+gcc at molgen dot mpg.de
Target Milestone: ---
>From the discussion of a change-set for coreboot [1], it turns out that GCC
does not understand valid instructions.
```
$ arm-linux-gnueabi-gcc-7 --version
arm-linux-gnueabi-gcc-7 (Debian 7.2.0-1) 7.2.0
Copyright (C) 2017 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
$ cat mov1.S
lsrsne r0, r0, #4
$ arm-linux-gnueabi-gcc-7 mov1.S -c -o mov1.gas
mov1.S: Assembler messages:
mov1.S:1: Error: bad instruction `lsrsne r0,r0,#4'
$ cat mov1.S
lsrsne r0, r0, #4
$ arm-linux-gnueabi-gcc-7 mov2.S -c -o mov2.gas
```
Julius Werner comments as below.
> FWIW, the ARMv7 A.R.M. says the S comes before the condition code, so this
> should've been MOVSNE instead of MOVNES anyway. It also doesn't explicitly
> allow a source operand shift for MOV, so LSRSNE should be the correct form. I
> bet GCC just has a few additional non-standard aliases, but I would be
> surprised if it can't recognize the official and most obvious notation
> (LSRSNE).
But there is an error in GCC.
[1] https://review.coreboot.org/21358/
More information about the Gcc-bugs
mailing list