linux-kernel - Re: [PATCH] scsi: fix the return value of scsi_logical_block

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <9c475f0e-d550-4ff7-9d6c-94dfc300908a@oracle.com>
Date: Wed, 7 Aug 2024 09:47:55 +0100
From: John Garry <john.g.garry@...cle.com>
To: Chaotian Jing <chaotian.jing@...iatek.com>,
        James.Bottomley@...senPartnership.com, martin.petersen@...cle.com
Cc: Matthias Brugger <matthias.bgg@...il.com>,
        AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>,
        linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org, srv_heupstream@...iatek.com
Subject: Re: [PATCH] scsi: fix the return value of scsi_logical_block_count

On 06/08/2024 08:26, Chaotian Jing wrote:
> scsi_logical_block_count() should return the block count of scsi device,

scsi command, not scsi device

> but the original code has a wrong implement.
> 
> Signed-off-by: Chaotian Jing <chaotian.jing@...iatek.com>
> ---
>   include/scsi/scsi_cmnd.h | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/scsi/scsi_cmnd.h b/include/scsi/scsi_cmnd.h
> index 45c40d200154..f0be0caa295a 100644
> --- a/include/scsi/scsi_cmnd.h
> +++ b/include/scsi/scsi_cmnd.h
> @@ -236,7 +236,7 @@ static inline unsigned int scsi_logical_block_count(struct scsi_cmnd *scmd)

I find it questionable why we have this in the core code, since it's 
only used in error handling logs for one driver. And obviously no one 
was paying attention to it.

>   {
>   	unsigned int shift = ilog2(scmd->device->sector_size) - SECTOR_SHIFT;
>   
> -	return blk_rq_bytes(scsi_cmd_to_rq(scmd)) >> shift;
> +	return blk_rq_sectors(scsi_cmd_to_rq(scmd)) >> shift;

blk_rq_sectors() converts from bytes to linux sectors, and then we shift 
it by diff in linux sectors and LBS.

To me, if this were used in fastpath - which it isn't - the following 
seems better:

return blk_rq_bytes(scsi_cmd_to_rq(scmd)) / scmd->device->sector_size;

>   }
>   
>   /*