linux-kernel - Re: [linux-pm] libata, hddtemp + s2ram is racy

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: <Pine.LNX.4.44L0.1005100958430.1626-100000@iolanthe.rowland.org>
Date:	Mon, 10 May 2010 10:04:20 -0400 (EDT)
From:	Alan Stern <stern@...land.harvard.edu>
To:	Bruno Prémont <bonbons@...ux-vserver.org>
cc:	linux-ide@...r.kernel.org, <linux-pm@...ts.linux-foundation.org>,
	<linux-kernel@...r.kernel.org>
Subject: Re: [linux-pm] libata, hddtemp + s2ram is racy

On Fri, 7 May 2010, Bruno [UTF-8] PrÃ©mont wrote:

> Hi,
> 
> On a SMP system I've hit a race condition between suspend and hddtemp
> checking disk's temperature.
> 
> System details:
> - Dual-core AMD Turion CPU
> - 00:12.0 SATA controller [0106]: ATI Technologies Inc SB600 Non-Raid-5 SATA [1002:4380]
> - hddtemp-0.3_beta15 (Gentoo package app-admin/hddtemp-0.3_beta15-r3)
> - Linus' tree shortly after v2.6.34-rc6, at commit
>   be1066bbcd443a65df312fdecea7e4959adedb45 with some drm updates on
>   top of it.
> 
> 
> It looked like hddtemp had sent the SMART request to disk right before
> suspend and during suspend process ata2 did complain without aborting
> suspend (see below).
> After resume access to that disk was dead-locked (any further
> attempt to suspend timed-out freezing hddtemp and any access attempt
> towards that disk did put userspace tasks in uninterruptible state and
> caused soft-raid to mark the disk failed).
> 
> Is suspend not waiting on SG_IO ioctls to complete (at ata host level)?

If you use the libata driver then the ATA I/O is handled by the SCSI 
midlayer.  The SCSI midlayer waits for the device's command queue to 
drain completely before initiating a suspend.

Furthermore, the hddtemp process has to get frozen before the suspend 
can begin, and I believe a process cannot be frozen while it is waiting 
for an SG_IO to complete.

At any rate, you should try posting this bug report on the linux-scsi 
mailing list.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/