linux-kernel - Re: [PATCH] ata: Disable queued TRIM for Samsung 860 SSDs

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: <5059c002-2452-f427-76da-7b08a02748dc@0882a8b5-c6c3-11e9-b005-00805fc181fe>
Date:   Wed, 15 Jul 2020 22:12:43 +0100
From:   Simon Arlott <simon@...iron.net>
To:     "Martin K. Petersen" <martin.petersen@...cle.com>
Cc:     Jens Axboe <axboe@...nel.dk>, linux-ide@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Park Ju Hyung <qkrwngud825@...il.com>,
        Tejun Heo <tj@...nel.org>
Subject: Re: [PATCH] ata: Disable queued TRIM for Samsung 860 SSDs

On 15/07/2020 21:53, Martin K. Petersen wrote:
>> Despite the unsubstantiated claim from Samsung that "the improved
>> queued trim enhances Linux compatibility" this does not appear to be
>> true, even on Intel SATA controllers:
> 
> I am aware of several people using 860 drives with queued TRIM. And I
> haven't heard any complaints outside of the bug you referenced.
> 
> Also, I have tested both 860 2.5" Pro and 860 mSATA EVO on a few
> different systems in my lab without any problems. See:
> 
>     https://lore.kernel.org/stable/yq1h87du82d.fsf@oracle.com/T/

I've just checked and it happens on both my SATA 860s:
  860 EVO 2TB (RVT04B6Q) on Intel Z170
  860 PRO 1TB (RVM02B6Q) on Intel H170

> I really wish we had some more data to work with :(

Is there a reliable way of reproducing this?

I have a Marvell 88SE9235 that I could try with the EVO.

> Lacking a proper heuristic I guess we don't have any choice to disable
> the feature. But that's sad news for the people who currently don't have
> problems since their performance will inevitably suffer.

Samsung need to identify what the problem is before claiming that their
Linux support is better, especially if specific chipsets are known to
have issues...

-- 
Simon Arlott