linux-kernel - Re: [PATCH] scsi: sd: add runtime pm to open / release

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: <20200811134851.GB331864@rowland.harvard.edu>
Date:   Tue, 11 Aug 2020 09:48:51 -0400
From:   Alan Stern <stern@...land.harvard.edu>
To:     Martin Kepplinger <martin.kepplinger@...i.sm>
Cc:     James Bottomley <James.Bottomley@...senpartnership.com>,
        Bart Van Assche <bvanassche@....org>,
        Can Guo <cang@...eaurora.org>, martin.petersen@...cle.com,
        linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org,
        kernel@...i.sm
Subject: Re: [PATCH] scsi: sd: add runtime pm to open / release

On Tue, Aug 11, 2020 at 09:55:54AM +0200, Martin Kepplinger wrote:
> On 10.08.20 16:13, Alan Stern wrote:
> > This may not matter...  but it's worth pointing out that 
> > expecting_media_change doesn't get cleared if ++scmd->retries > 
> > scmd->allowed.
> 
> absolutely worth pointing out and I'm not yet sure about that one.
> 
> > 
> > It also doesn't get cleared in cases where the device _doesn't_ 
> > report a Unit Attention.
> 
> true. but don't we set the flag for a future UA we don't yet know of? If
> we would want to clear it outside of a UA, I think we'd need to keep
> track of a suspend/resume cycle and if we see that we *had* successfully
> "done requests" after resuming, we could clear it...

The point is that expecting_media_change should apply only to the _next_ 
command.  It should be cleared after _every_ command.  You do not want 
to leave it hanging around -- if you do then you will miss real media 
changes.

There's a potential issue when a bunch of commands get sent and 
completed all at once, but hopefully it won't matter here.

Alan Stern