Re: [PATCH resend] fs/proc: Move kfree outside pde_unload_lock

[Nathan Zimmer <nzimmer@sgi.com>]

On 04/05/2013 12:36 PM, Al Viro wrote:
> On Fri, Apr 05, 2013 at 12:05:26PM -0500, Nathan Zimmer wrote:
>> On 04/04/2013 03:44 PM, Al Viro wrote:
>>> On Thu, Apr 04, 2013 at 12:12:05PM -0500, Nathan Zimmer wrote:
>>>
>>>> Ok I am cloning the tree now.
>>>> It does look like the patches would conflict.
>>>> I'll run some tests and take a deeper look.
>>> FWIW, I've just pushed there a tentative patch that switches to hopefully
>>> saner locking (head should be at cb673c115c1f99d3480471ca5d8cb3f89a1e3bee).
>>> Is that more or less what you want wrt spinlock contention?
>>>
>>> One note: for any given pde_opener, close_pdeo() can be called at most
>>> by two threads - final fput() and remove_proc_entry() resp.  I think
>>> the use of completion + flag is safe there; pde->pde_unload_lock
>>> should serialize the critical areas.
>> Something isn't quite right.  I keep getting hung during boot.
>> dracut: Mounted root filesystem /dev/sda8
>> dracut: Switching root
>>
>> I'll try to get some more info on a smaller box.
> Umm...  Try to add WARN_ON(1) in entry_rundown(), just to see what's
> getting hit (don't bother with entry name, stack trace will be enough).
That didn't produce anything.  I'll run some bisections over the weekend 
and see what I can sort out.