All of lore.kernel.org
 help / color / mirror / Atom feed
From: ChenXiaoSong <chenxiaosong.chenxiaosong@linux.dev>
To: David Howells <dhowells@redhat.com>,
	Steve French <smfrench@gmail.com>,
	Namjae Jeon <linkinjeon@kernel.org>
Cc: pc@manguebit.org, ronniesahlberg@gmail.com,
	sprasad@microsoft.com, tom@talpey.com, bharathsm@microsoft.com,
	senozhatsky@chromium.org, linux-cifs@vger.kernel.org,
	ChenXiaoSong <chenxiaosong@kylinos.cn>
Subject: Re: [PATCH v6 4/5] smb/client: use bsearch() to find target in smb2_error_map_table
Date: Sat, 27 Dec 2025 10:18:47 +0800	[thread overview]
Message-ID: <f0740fbf-1142-42f8-b56f-937fa915a4bb@linux.dev> (raw)
In-Reply-To: <e64b6e2d-6ad8-43b0-bca3-fbeae76f6306@linux.dev>

Hi David,

Your second suggestion looks good. Should we change it to the following 
code? If you are okay with these changes, I can send the next version.

```
__inline_bsearch((const void *)(unsigned long)smb2_status, ...);

static __always_inline int cmp_smb2_status(const void *_key, const void 
*_pivot)
{
         __le32 key = (unsigned long)_key;
         const struct status_to_posix_error *pivot = _pivot;

         if (key < pivot->smb2_status)
                 return -1;
         if (key > pivot->smb2_status)
                 return 1;
         return 0;
}
```

Thanks,
ChenXiaoSong <chenxiaosong@kylinos.cn>

On 12/26/25 11:28 PM, ChenXiaoSong wrote:
> Hi David,
> 
> We cannot use "return b->smb2_status - a->smb2_status".
> 
> For example:
> 
>      a_status = 0xC0000005
>      b_status = 0x00000001
> 
> The subtraction is evaluated in an unsigned type:
> 
>      a_status - b_status = 0xC0000004 > 0
> 
> But the comparison function returns int, so the value is converted:
> 
>      (int)(a_status - b_status) < 0
> 
> Thanks,
> ChenXiaoSong.
> 
> On 12/26/25 8:06 PM, David Howells wrote:
>>
>> Actually...  It's probably sufficient to do:
>>
>>     static __always_inline
>>     int cmp_smb2_status(const void *_a, const void *_b)
>>     {
>>         const struct status_to_posix_error *a = _a, *b = _b;
>>
>>         return b->smb2_status - a->smb2_status;
>>     }
>>
>> as __inline_bsearch() only cares about the zeroness or sign of the return
>> value.  (Note the arguments to the subtraction might need to be flipped).
>>
>> It might even better just to cast the smb status you're looking for to 
>> the key
>> parameter:
>>
>>     __inline_bsearch((const void *)(long)status, ...);
>>
>> and then do this in the comparison function:
>>
>>     static __always_inline
>>     int cmp_smb2_status(const void *key, const void *_b)
>>     {
>>         const struct status_to_posix_error *b = _b;
>>         int status = (long)key;
>>
>>         return b->smb2_status - status;
>>     }
>>
>> as __inline_bsearch() doesn't attempt to dereference key.
>>
>> David
> 


  reply	other threads:[~2025-12-27  2:19 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-12-25  2:10 [PATCH v6 0/5] smb: improve search speed of SMB2 maperror chenxiaosong.chenxiaosong
2025-12-25  2:10 ` [PATCH v6 1/5] cifs: Label SMB2 statuses with errors chenxiaosong.chenxiaosong
2025-12-25  2:10 ` [PATCH v6 2/5] cifs: Autogenerate SMB2 error mapping table chenxiaosong.chenxiaosong
2025-12-27 10:53   ` ChenXiaoSong
2025-12-27 12:00     ` David Howells
2025-12-27 15:03       ` ChenXiaoSong
2025-12-27 15:50         ` David Howells
2025-12-27 16:05           ` ChenXiaoSong
2025-12-30  9:42             ` ChenXiaoSong
2025-12-30 21:08               ` David Howells
2025-12-30 21:14               ` David Howells
2025-12-30  3:55   ` ChenXiaoSong
2025-12-30  9:07     ` David Howells
2025-12-25  2:10 ` [PATCH v6 3/5] smb/client: check whether smb2_error_map_table is sorted in ascending order chenxiaosong.chenxiaosong
2025-12-25  2:10 ` [PATCH v6 4/5] smb/client: use bsearch() to find target in smb2_error_map_table chenxiaosong.chenxiaosong
2025-12-26 11:58   ` David Howells
2025-12-26 12:06   ` David Howells
2025-12-26 15:28     ` ChenXiaoSong
2025-12-27  2:18       ` ChenXiaoSong [this message]
2025-12-27  8:56         ` David Howells
2025-12-27  9:29           ` ChenXiaoSong
2025-12-27 11:59             ` David Howells
2025-12-27  8:53       ` David Howells
2025-12-25  2:10 ` [PATCH v6 5/5] smb/client: introduce KUnit test to check search result of smb2_error_map_table chenxiaosong.chenxiaosong
2025-12-28 15:07   ` kernel test robot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f0740fbf-1142-42f8-b56f-937fa915a4bb@linux.dev \
    --to=chenxiaosong.chenxiaosong@linux.dev \
    --cc=bharathsm@microsoft.com \
    --cc=chenxiaosong@kylinos.cn \
    --cc=dhowells@redhat.com \
    --cc=linkinjeon@kernel.org \
    --cc=linux-cifs@vger.kernel.org \
    --cc=pc@manguebit.org \
    --cc=ronniesahlberg@gmail.com \
    --cc=senozhatsky@chromium.org \
    --cc=smfrench@gmail.com \
    --cc=sprasad@microsoft.com \
    --cc=tom@talpey.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.