Re: Kernel rwlock design, Multicore and IGMP

[Chris Metcalf <cmetcalf@tilera.com>]

On 11/12/2010 2:13 AM, Américo Wang wrote:
> On Fri, Nov 12, 2010 at 11:32:59AM +0800, Cypher Wu wrote:
>> On Thu, Nov 11, 2010 at 11:23 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
>>> Le jeudi 11 novembre 2010 à 21:49 +0800, Cypher Wu a écrit :
>>>> I'm using TILEPro and its rwlock in kernel is a liitle different than
>>>> other platforms. It have a priority for write lock that when tried it
>>>> will block the following read lock even if read lock is hold by
>>>> others. Its code can be read in Linux Kernel 2.6.36 in
>>>> arch/tile/lib/spinlock_32.c.
>>>
>>> This seems a bug to me.
>>> [...]
>>>
>> It seems not a problem that read_lock() can be nested or not since
>> rwlock doesn't have 'owner', it's just that should we give
>> write_lock() a priority than read_lock() since if there have a lot
>> read_lock()s then they'll starve write_lock().
>> We should work out a well defined behavior so all the
>> platform-dependent raw_rwlock has to design under that principle.
> 
> It is a known weakness of rwlock, it is designed like that. :)

Exactly.  The tile rwlock correctly allows recursively reacquiring the read
lock.  But it does give priority to writers, for the (unfortunately
incorrect) reasons Cypher Wu outlined above, e.g.:

- Core A takes a read lock
- Core B tries for a write lock and blocks new read locks
- Core A tries for a (recursive) read lock and blocks

Core A and B are now deadlocked.

The solution is actually to simplify the tile rwlock implementation so that
both readers and writers contend fairly for the lock.

I'll post a patch in the next day or two for tile.

-- 
Chris Metcalf, Tilera Corp.
http://www.tilera.com