[lkml]   [2015]   [May]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 4/5] workqueue: don't expose workqueue_attrs to users
On 05/11/2015 10:59 PM, Tejun Heo wrote:
> On Mon, May 11, 2015 at 05:35:51PM +0800, Lai Jiangshan wrote:
>> workqueue_attrs is an internal-like structure and is exposed with
>> apply_workqueue_attrs() whose user has to investigate the structure
>> before use.
>> And the apply_workqueue_attrs() API is inconvenient with the structure.
>> The user (although there is no user yet currently) has to assemble
>> several LoC to use:
>> attrs = alloc_workqueue_attrs();
>> if (!attrs)
>> return;
>> attrs->nice = ...;
>> copy cpumask;
>> attrs->no_numa = ...;
>> apply_workqueue_attrs();
>> free_workqueue_attrs();
>> It is too elaborate. This patch changes apply_workqueue_attrs() API,
>> and one-line-code is enough to be called from user:
>> apply_workqueue_attrs(wq, cpumask, nice, numa);
>> This patch also reduces the code of workqueue.c, about -50 lines.
>> wq_sysfs_prep_attrs() is removed, wq_[nice|cpumask|numa]_store()
>> directly access to the ->unbound_attrs with the protection
>> of apply_wqattrs_lock();
>> This patch is also a preparation patch of next patch which
>> remove no_numa out from the structure workqueue_attrs which
>> requires apply_workqueue_attrs() has an argument to pass numa affinity.
> I'm not sure about this. Yeah, sure, it's a bit more lines of code
> but at the same time this'd allow us to make the public interface
> atomic too. What we prolly should do is changing the interface so
> that we do
> attrs = prepare_workqueue_attrs(gfp_mask); /* allocate, lock & copy */
> /* modify attrs as desired */
> commit_workqueue_attrs(attrs); /* apply, unlock and free */

I think the workqueue.c has too much complicated and rarely used APIs
and exposes too much in this way. No one can set the nice value
and the cpuallowed of a task atomically.

If the user want atomic-able, Her/he can just disable WQ_SYSFS
on its workqueue and maintain a copy of the cpumask, nice, numa values
under its own lock.

> Thanks.

 \ /
  Last update: 2015-05-12 04:41    [W:0.075 / U:11.156 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site