mbox series

[v2,0/2] LoongArch: gas: Add LVZ and LBT instructions support

Message ID 20230630090416.3051425-1-mengqinggang@loongson.cn
Headers
Series LoongArch: gas: Add LVZ and LBT instructions support |

Message

mengqinggang June 30, 2023, 9:04 a.m. UTC 
  Changes from v1:
  Add LARCH_opts.ase_lvz and LARCH_opts.ase_lbt.

mengqinggang (2):
  LoongArch: gas: Add LVZ and LBT instructions support
  LoongArch: gas: Fix code style issues

 gas/config/tc-loongarch.c             |    4 +
 gas/testsuite/gas/loongarch/lvz-lbt.d |  191 ++
 gas/testsuite/gas/loongarch/lvz-lbt.s |  181 ++
 gas/testsuite/gas/loongarch/uleb128.d |   52 +-
 include/opcode/loongarch.h            |    6 +
 opcodes/loongarch-dis.c               |    2 +
 opcodes/loongarch-opc.c               | 3195 +++++++++++++------------
 7 files changed, 2105 insertions(+), 1526 deletions(-)
 create mode 100644 gas/testsuite/gas/loongarch/lvz-lbt.d
 create mode 100644 gas/testsuite/gas/loongarch/lvz-lbt.s

Comments

WANG Xuerui June 30, 2023, 9:15 a.m. UTC | #1 
On 2023/6/30 17:04, mengqinggang wrote:
> Changes from v1:
>    Add LARCH_opts.ase_lvz and LARCH_opts.ase_lbt.
> 
> mengqinggang (2):
>    LoongArch: gas: Add LVZ and LBT instructions support
>    LoongArch: gas: Fix code style issues
> 
>   gas/config/tc-loongarch.c             |    4 +
>   gas/testsuite/gas/loongarch/lvz-lbt.d |  191 ++
>   gas/testsuite/gas/loongarch/lvz-lbt.s |  181 ++
>   gas/testsuite/gas/loongarch/uleb128.d |   52 +-
>   include/opcode/loongarch.h            |    6 +
>   opcodes/loongarch-dis.c               |    2 +
>   opcodes/loongarch-opc.c               | 3195 +++++++++++++------------
>   7 files changed, 2105 insertions(+), 1526 deletions(-)
>   create mode 100644 gas/testsuite/gas/loongarch/lvz-lbt.d
>   create mode 100644 gas/testsuite/gas/loongarch/lvz-lbt.s
> 

Thanks for the patches, this now LGTM. (I have to fixup some offset 
operands' format strings and it's better done after these are applied.)

For the record: I got off-list confirmation that the "setx86*" and 
"setarm*" names won't get changed this time. While reasons were not 
given, I believe it's probably because the "setARCH*" ops are actually 
markers and do not actually touch arch-specific state, while the 
"ARCHset*" ones do; but without the manuals available that's just a wild 
guess. We'll have time to revisit anyway without compatibility concerns 
though, so I'm fine with the decision at the moment.