Re: Review: Reduce in-core superblock lock contention near ENOSPC

[David Chinner <dgc@sgi.com>]

On Wed, Dec 06, 2006 at 08:43:47AM +0000, Lachlan McIlroy wrote:
> David Chinner wrote:
> >On Tue, Dec 05, 2006 at 12:46:46PM +0100, Klaus Strebel wrote:
> >
> >>Hi guys,
> >>
> >>just updated my CVS copy from oss.sgi.com ( the linux-2.6-xfs ) and 
> >>tried to compile ... but your patch failes to compile if HAVE_PERCPU_SB 
> >>is #ifndef'd :-(, the m_icsb_mutex is not in the struct see xfs_mount.h. 
> >>Make oldconfig didn't show HAVE_PERCPU_SB as option for .config, looks 
> >>like nobody tested on a single processor config ??
> >
> >
> >Sorry - my bad. The code did not change for UP, so I didn't think to
> >test it.  The patch below abstracts the icsb_mutex so that it
> >doesn't get directly referenced by code outside the per-cpu counter
> >code. Builds with and without HAVE_PERCPU_SB defined.
> >
> >I'll run a test cycle on it and get it fixed up.
> >
> >Cheers,
> >
> >Dave.
> 
> @@ -1803,6 +1803,7 @@ xfs_icsb_destroy_counters(
>  		unregister_hotcpu_notifier(&mp->m_icsb_notifier);
>  		free_percpu(mp->m_sb_cnts);
>  	}
> +	mutex_destroy(&mp->m_icsb_mutex);
>  }
> 
> Do you need to abstract the call to mutex_destroy too?

No. I didn't abstract the mutex_init and mutex_destroy calls because
they are in the init/destroy functions for the icsb subsystem and
those functions are #define'd out when HAVE_PERCPU_SB is not
defined.

> The rest of the change looks good.

Thanks, Lachlan.

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group