Re: [PATCH 00/17] fs, btrfs refcount conversions

From: David Sterba <dsterba@suse.cz>
To: Qu Wenruo <quwenruo@cn.fujitsu.com>
Cc: "Reshetova, Elena" <elena.reshetova@intel.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-fsdevel@vger.kernel.org" <linux-fsdevel@vger.kernel.org>,
	"linux-btrfs@vger.kernel.org" <linux-btrfs@vger.kernel.org>,
	"peterz@infradead.org" <peterz@infradead.org>,
	"gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
	"jbacik@fb.com" <jbacik@fb.com>, "clm@fb.com" <clm@fb.com>,
	"dsterba@suse.com" <dsterba@suse.com>
Subject: Re: [PATCH 00/17] fs, btrfs refcount conversions
Date: Thu, 9 Mar 2017 16:29:29 +0100	[thread overview]
Message-ID: <20170309152929.GG14605@twin.jikos.cz> (raw)
In-Reply-To: <2b9ad52c-34fa-1c54-0374-01d443530f63@cn.fujitsu.com>

On Tue, Mar 07, 2017 at 03:49:52PM +0800, Qu Wenruo wrote:
> >>> If the patches pass all tests on your side, could you please take them in and
> >> propagate further?
> >>> I will continue with other kernel subsystems.
> >>
> >> The patchset itself looks like a common cleanup, while I did encounter
> >> several cases (almost all scrub tests) causing kernel warning due to
> >> underflow.
> >
> > Oh, could you please send me the warning outputs? I can hopefully analyze and fix them.
> 
> Attached. Which is the generated by running btrfs/070 test case.
> And I canceled the case almost instantly, so output is not much, but 
> still contains enough info.
> 
> Both refcount_inc() and refcount_sub_and_test() are causing warning.
> 
> So now I'm not sure which is the cause, btrfs or bad use of refcount?

We we do atomic_inc to get the first reference after initialization in
scrub_pages, instead of atomic_set (or an equivalent):

2266                 spage = kzalloc(sizeof(*spage), GFP_KERNEL);
2267                 if (!spage) {
...
2274                 }
...
2276                 scrub_page_get(spage);

so the references are 0 and refcount_inc will catch that, the fix is simple.

The refcount_sub_and_test reports seem to catch a bug in refcounting, I'm
analyzing it right now.