linux-kernel - Re: [PATCH] scsi:libsas: Simplify sas_queue

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

Message-ID: <7fe583de-966e-1e86-3233-556886c943d1@kernel.org>
Date:   Mon, 31 Jul 2023 11:05:08 +0900
From:   Damien Le Moal <dlemoal@...nel.org>
To:     Jason Yan <yanaijie@...wei.com>,
        Wenchao Hao <haowenchao2@...wei.com>,
        John Garry <john.g.garry@...cle.com>,
        "James E . J . Bottomley" <jejb@...ux.ibm.com>,
        "Martin K . Petersen" <martin.petersen@...cle.com>,
        linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org
Cc:     louhongxiang@...wei.com
Subject: Re: [PATCH] scsi:libsas: Simplify sas_queue_reset and remove unused
 code

On 7/31/23 10:44, Jason Yan wrote:
> Hi Wenchao,
> 
> On 2023/7/29 18:24, Wenchao Hao wrote:
>> sas_queue_reset is always called with param "wait" set to 0, so
>> remove it from this function's param list. And remove unused
>> function sas_wait_eh.
> 
> I did not find any users passing '1' to sas_queue_reset() in the git 
> history. It seems it is never used. So It's ok to remove it, I guess.
> 
> Just one nit, there should be a blank between "scsi:" and "libsas" in 
> the subject:
> 
> scsi: libsas: Simplify sas_queue_reset and remove unused code
> 
> Otherwise looks good to me:
> Reviewed-by: Jason Yan <yanaijie@...wei.com>

Yeah, code wise, it looks good.
However, I am seeing issues with completions for commands that use command
duration limits. There are some unusual completions that needs special handling
with CDL (e.g. some aborted commands can be notified with a success and
sense-data-available-bit set. For these, we kick libata-EH but it seems that
this is not well working with libsas. So I wonder if this code may need to be
used for CDL... So let's please hold on before applying this. Let me check the
CDL issues I am seeing first.

-- 
Damien Le Moal
Western Digital Research