Re: [Patch 2/8] Sync block I/O and swapin delay collection

[Balbir Singh <balbir@in.ibm.com>]

On Tue, May 09, 2006 at 02:23:15PM +1000, Nick Piggin wrote:
> Balbir Singh wrote:
> 
> >On Mon, May 08, 2006 at 02:19:52PM -0700, Andrew Morton wrote:
> >
> >>Balbir Singh <balbir@in.ibm.com> wrote:
> >>
> >>>@@ -550,6 +550,12 @@ struct task_delay_info {
> >>>	 * Atomicity of updates to XXX_delay, XXX_count protected by
> >>>	 * single lock above (split into XXX_lock if contention is an issue).
> >>>	 */
> >>>+
> >>>+	struct timespec blkio_start, blkio_end;	/* Shared by blkio, swapin */
> >>>+	u64 blkio_delay;	/* wait for sync block io completion */
> >>>+	u64 swapin_delay;	/* wait for swapin block io completion */
> >>>+	u32 blkio_count;
> >>>+	u32 swapin_count;
> >>>
> >>These fields are a bit mystifying.
> >>
> >>In what units are blkio_delay and swapin_delay?
> >>
> >>What is the meaning behind blkio_count and swapin_count?
> >>
> >>Better comments needed, please.
> >>
> >
> >Will add more detailed comments and send them as updates.
> >
> 
> What kinds of usages will this stuff see? Will the CONFIG be usually 
> turned on,
> with some tasks occasionally using the statistics?
> 
> In which case, might it be better to make each delay collector in its 
> own data
> structure { .list; .start; .end; .delay; .count; .private; .name }, and 
> allocate
> them and hang them off the task structure when they're in use?
> 
> Or even put them in their own data structure (a small hash or something).
> 
> OTOH if they're often going to be in use by many tasks, then what you 
> have might
> be the best option.
> 
> Nick
> 
> Send instant messages to your online friends http://au.messenger.yahoo.com 

I expect/hope that the CONFIG will be turned on. There is a boot
option (called delayacct) to enable/disable the statistics collection.
Once turned on and enabled, all tasks will be filling in/using the statistics.


	Thanks,
	Balbir Singh,
	Linux Technology Center,
	IBM Software Labs
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/