Re: kernel BUG at mm/page-writeback.c:2241 [ BUG_ON(PageWriteback(page); ]

From: Jens Axboe <axboe@kernel.dk>
To: Matthew Wilcox <willy@infradead.org>, Jan Kara <jack@suse.cz>
Cc: Qian Cai <cai@lca.pw>, Christoph Hellwig <hch@infradead.org>,
	"Darrick J. Wong" <darrick.wong@oracle.com>,
	linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: kernel BUG at mm/page-writeback.c:2241 [ BUG_ON(PageWriteback(page); ]
Date: Mon, 26 Oct 2020 07:55:44 -0600	[thread overview]
Message-ID: <d06d3d2a-7032-91da-35fa-a9dee4440a14@kernel.dk> (raw)
In-Reply-To: <20201026131353.GP20115@casper.infradead.org>

On 10/26/20 7:13 AM, Matthew Wilcox wrote:
> On Mon, Oct 26, 2020 at 10:49:48AM +0100, Jan Kara wrote:
>> On Thu 22-10-20 01:49:06, Matthew Wilcox wrote:
>>> On Wed, Oct 21, 2020 at 08:30:18PM -0400, Qian Cai wrote:
>>>> Today's linux-next starts to trigger this wondering if anyone has any clue.
>>>
>>> I've seen that occasionally too.  I changed that BUG_ON to VM_BUG_ON_PAGE
>>> to try to get a clue about it.  Good to know it's not the THP patches
>>> since they aren't in linux-next.
>>>
>>> I don't understand how it can happen.  We have the page locked, and then we do:
>>>
>>>                         if (PageWriteback(page)) {
>>>                                 if (wbc->sync_mode != WB_SYNC_NONE)
>>>                                         wait_on_page_writeback(page);
>>>                                 else
>>>                                         goto continue_unlock;
>>>                         }
>>>
>>>                         VM_BUG_ON_PAGE(PageWriteback(page), page);
>>>
>>> Nobody should be able to put this page under writeback while we have it
>>> locked ... right?  The page can be redirtied by the code that's supposed
>>> to be writing it back, but I don't see how anyone can make PageWriteback
>>> true while we're holding the page lock.
>>
>> FWIW here's very similar report for ext4 [1] and I strongly suspect this
>> started happening after Linus' rewrite of the page bit waiting logic. Linus
>> thinks it's preexisting bug which just got exposed by his changes (which is
>> possible). I've been searching a culprit for some time but so far I failed.
>> It's good to know it isn't ext4 specific so we should be searching in the
>> generic code ;). So far I was concentrating more on ext4 bits...
>>
>> 								Honza
>>
>> [1] https://lore.kernel.org/lkml/000000000000d3a33205add2f7b2@google.com/
> 
> Oh good, I was wondering if it was an XFS bug ;-)
> 
> I hope Qian gets it to reproduce soon with the assert because that will
> tell us whether it's a spurious wakeup or someone calling SetPageWriteback
> without holding the page lock.

I've tried to reproduce this as well, to no avail. Qian, could you perhaps
detail the setup? What kind of storage, kernel config, compiler, etc.

-- 
Jens Axboe