Re: [PATCH] fb: Rework locking to fix lock ordering on takeover

[Sasha Levin]

On Tue, Dec 25, 2012 at 9:41 PM, Cong Wang <xiyou.wangcong@xxxxxxxxx> wrote:
> On Wed, Dec 26, 2012 at 12:08 AM, Sasha Levin <levinsasha928@xxxxxxxxx> wrote:
>> On Tue, Dec 18, 2012 at 10:20 AM, Josh Boyer <jwboyer@xxxxxxxxx> wrote:
>>> On Wed, Nov 21, 2012 at 7:53 AM, Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> wrote:
>>>> On Wed, 21 Nov 2012 07:45:45 -0500
>>>> Josh Boyer <jwboyer@xxxxxxxxx> wrote:
>>>>
>>>>> On Fri, Nov 16, 2012 at 2:27 PM, Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> wrote:
>>>>> >
>>>>> > [The fb maintainer appears to be absent at the moment].
>>>>> >
>>>>> > This is needed to fix a pile of lockdep splats that now show up because console_lock()
>>>>> > is being properly audited. Hugh Dickins and Sasha Levin have tested it and both reports
>>>>> > all looks good. This is probably not the whole story - the entire fb layer has locking
>>>>> > confusion problems that were previously hidden but it seems to get the ones people hit
>>>>> > in testing. This hopefully explains a few of the weird fb hangs that have been floating
>>>>> > around forever.
>>>>> >
>>>>> > From: Alan Cox <alan@xxxxxxxxxxxxxxx>
>>>>> >
>>>>> > Adjust the console layer to allow a take over call where the caller already
>>>>> > holds the locks. Make the fb layer lock in order.
>>>>> >
>>>>> > This s partly a band aid, the fb layer is terminally confused about the
>>>>> > locking rules it uses for its notifiers it seems.
>>>>> >
>>>>> > Signed-off-by: Alan Cox <alan@xxxxxxxxxxxxxxx>
>>>>>
>>>>> Should this eventually get into the stable trees?
>>>>
>>>> Thats a question I'm not sure about at this point. I think the bug is
>>>> real but not caught by the lock checker in older trees but I've not
>>>> investigated.
>>>
>>> So... this patch seems to still be twisting in the wind.  It should
>>> probably be headed into 3.8 at this point, shouldn't it?
>>
>> Indeed it should. I'm seeing the original warnings in 3.8-rc1 and have
>> to carry this patch to avoid them.
>
> This patch can fix the following warning we saw?
> http://lkml.org/lkml/2012/12/22/53
>
> I will give it a try.

Yup, that's the same error I've reported couple of months ago.

It looks like the fb maintains are still absent, so it'll probably
need a different way to get upstream.


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