[dm-devel] [PATCH 1/1] improve the performance of dm-log-userspace

Brassow Jonathan jbrassow at redhat.com
Mon Jan 13 01:18:23 UTC 2014


I've cleaned up the userspace and kernel code a bit.  I'll test them tomorrow.  When things work out, I'll check-in the userspace code and resubmit the kernel code here.

 brassow

On Jan 5, 2014, at 11:06 PM, Dongmao Zhang wrote:

> hi,brassow,
>  How is going about the userspace cmirrord's patch testing? Is there anything I could help?
> 
> Dongmao Zhang.
> 
> 
> 于 2013年12月11日 16:01, Dongmao Zhang 写道:
>> hi, brassow
>> 
>> Yes, I had submited a patch lvm-devel.
>> 
>> http://www.redhat.com/archives/lvm-devel/2013-November/msg00093.html.
>> 
>> please review.
>> 
>> I should have add a patch version indicator.
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 于 2013年12月11日 09:17, Brassow Jonathan 写道:
>>> Looks good. Did you happen to post the accompanying userspace patches?
>>> If you pass those along, I will review that also and test them
>>> together. Once that's done, I will ACK.
>>> 
>>> brassow
>>> 
>>> On Dec 3, 2013, at 11:35 PM, dongmao zhang wrote:
>>> 
>>>> modified workqueue calls on Hayakawa's comments
>>>> 
>>>> In the cluster evironment, cluster write has poor performance.
>>> 
>>> </snip>
>>> 
>>>> spinlock_t flush_lock;
>>>> struct list_head mark_list;
>>>> struct list_head clear_list;
>>>> +
>>>> + /*work queue for flush clear region*/
>>> 
>>> Use spaces after /* and before */
>>> 
>>>> + struct workqueue_struct *dmlog_wq;
>>>> + struct delayed_work flush_log_work;
>>>> + atomic_t sched_flush;
>>>> + uint32_t integrated_flush;
>>>> };
>>> 
>>> </snip>
>>> 
>>>> devices_rdata = kzalloc(devices_rdata_size, GFP_KERNEL);
>>>> if (!devices_rdata) {
>>>> DMERR("Failed to allocate memory for device information");
>>>> @@ -246,6 +268,20 @@ static int userspace_ctr(struct dm_dirty_log
>>>> *log, struct dm_target *ti,
>>>> DMERR("Failed to register %s with device-mapper",
>>>> devices_rdata);
>>>> }
>>>> +
>>>> +
>>> 
>>> Extra newline...
>>> 
>>> The rest looks good. (Cosmetic issues like this wouldn't prevent my
>>> ACK though.)
>>> 
>>> --
>>> dm-devel mailing list
>>> dm-devel at redhat.com
>>> https://www.redhat.com/mailman/listinfo/dm-devel
>>> 
>> 
>> --
>> dm-devel mailing list
>> dm-devel at redhat.com
>> https://www.redhat.com/mailman/listinfo/dm-devel
> 





More information about the dm-devel mailing list