Re: [PATCH 2/2] cifs: Autogenerate SMB2 error mapping table

[ChenXiaoSong <chenxiaosong.chenxiaosong@linux.dev>]

Hi David,

Based on Steve's suggestions, should we consider only using the 
`gen_smb2_mapping` script to reorder the smb2_error_map_table[] array in 
client/smb2maperror.c?

This way, we can keep the code changes minimal (of course, we might have 
the opportunity for larger changes in the future).

Thanks,
ChenXiaoSong.

On 12/15/25 12:03, Steve French wrote:
> I would prefer not a large line of code change unless some perf benefit 
> but if alongside a perf improvement or bug fix then easier to justify
> 
> Thanks,
> 
> Steve
> 
> On Sun, Dec 14, 2025, 9:12 PM ChenXiaoSong 
> <chenxiaosong.chenxiaosong@linux.dev 
> <mailto:chenxiaosong.chenxiaosong@linux.dev>> wrote:
> 
>     Personally, I think using enum is a good idea.
> 
>     Thanks,
>     ChenXiaoSong.
> 
>     On 12/12/25 17:41, David Howells wrote:
>      > Also note that the format we end up going with for smb2status.h
>     is up for
>      > discussion.  My preferred idea is something along the lines of:
>      >
>      >       enum nt_status_codes {
>      >               STATUS_SUCCESS          = 0x00000000, // 0
>      >               STATUS_WAIT_0           = STATUS_SUCCESS,
>      >               STATUS_WAIT_1           = 0x00000001, // -EIO
>      >               STATUS_WAIT_2           = 0x00000002, // -EIO
>      >               STATUS_WAIT_3           = 0x00000003, // -EIO
>      >               ...
>      >       };
>      >
>      > and switching to using cpu-endian in the code.
>      >
>      > David
>