linux-kernel - Re: Bug in 2.6.29 ide-cd: Kernel freeze: bisected + unacceptable workaround

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: <9ea470500903310706i23e2225fo42ff17f14c277594@mail.gmail.com>
Date:	Tue, 31 Mar 2009 16:06:11 +0200
From:	Borislav Petkov <petkovbb@...glemail.com>
To:	Michael Roth <mroth@...sie.de>
Cc:	Alan Cox <alan@...rguk.ukuu.org.uk>,
	Bartlomiej Zolnierkiewicz <bzolnier@...il.com>,
	linux-kernel@...r.kernel.org
Subject: Re: Bug in 2.6.29 ide-cd: Kernel freeze: bisected + unacceptable 
	workaround

Hi,

On Tue, Mar 31, 2009 at 2:35 PM, Michael Roth <mroth@...sie.de> wrote:
>>> take it ok. Also, if I remember correctly, the original ide-cd behavior
>>> was to issue the command and _then_ start DMA so we're back to that. I
>>> guess now we should be concerned whether the other ATAPI devices can
>>> handle the reverse situation where you first issue a command and _then_
>>> start DMA.. Hmm...
>>
>> Most devices don't care but you do need to follow the exact ATAPI spec.
>> If you search way back through the archives you'll find long discussions
>> on this during either 2.4 or 2.2 when HP hit problems caused by us
>> turning DMA on too early.
>
> AFAIK, the old bevavior was to start DMA _after_ issuing the command _only_
> for CD-drives and for the remaining ide devices the other way around,
> why not just restore the old behavior and do the DMA after command only
> for CD-drives?

because common atapi code should adhere to the spec, as Alan said, no matter
whether it is a cd or a floppy or [you name it]-drive.

@Alan: I guess you refer to
http://lkml.indiana.edu/hypermail/linux/kernel/0305.1/2241.html, for example.
I think I'll be on the safe side when I turn on DMA _after_
writing the ATAPI command packet into the data register - it is the correct
order of steps according to the ATA spec. So, if there are devices which still
choke on that, then they're simply buggy and those will hopefully be a small
number so that a blacklist with a flag as a workaround would be more or
less ok for them - we do that for other b0rked devices anyway.

-- 
Regards/Gruss,
Boris
--
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/