On 03/29/2016 03:20 PM, Christoph Hellwig wrote:
On Tue, Mar 29, 2016 at 02:51:37PM -0600, Jens Axboe wrote:
For Shaohua's patches, not binding per-device data to the home node makes a
lot of sense, though.
It does. But exposing the node estimation function just to work around
the fact that t
On Tue, Mar 29, 2016 at 02:51:37PM -0600, Jens Axboe wrote:
> For Shaohua's patches, not binding per-device data to the home node makes a
> lot of sense, though.
It does. But exposing the node estimation function just to work around
the fact that the way the cpu/node mapping is done is upside dow
On 03/29/2016 11:44 AM, Christoph Hellwig wrote:
On Tue, Mar 29, 2016 at 10:50:11AM -0600, Jens Axboe wrote:
This looks weird, shouldn't the cpu assignment be determined by block
core (blk-mq) because block core decides how to use the queue?
I agree, that belongs in the blk-mq proper, the driv
On Tue, Mar 29, 2016 at 10:50:11AM -0600, Jens Axboe wrote:
> >This looks weird, shouldn't the cpu assignment be determined by block
> >core (blk-mq) because block core decides how to use the queue?
>
> I agree, that belongs in the blk-mq proper, the driver should just follow
> the rules outlined,
On 03/29/2016 10:47 AM, Shaohua Li wrote:
On Tue, Mar 29, 2016 at 12:24:43AM -0700, Christoph Hellwig wrote:
On Fri, Mar 25, 2016 at 02:36:30PM -0700, Shaohua Li wrote:
we allocate most data structure in device's node, but some data
structures are not for DMA and mostly used by specific cpus/no
On Tue, Mar 29, 2016 at 12:24:43AM -0700, Christoph Hellwig wrote:
> On Fri, Mar 25, 2016 at 02:36:30PM -0700, Shaohua Li wrote:
> > we allocate most data structure in device's node, but some data
> > structures are not for DMA and mostly used by specific cpus/node which
> > could diff from device'
On Fri, Mar 25, 2016 at 02:36:30PM -0700, Shaohua Li wrote:
> we allocate most data structure in device's node, but some data
> structures are not for DMA and mostly used by specific cpus/node which
> could diff from device's node. Allocating such hot data in device's
> node doesn't make sense. Add
we allocate most data structure in device's node, but some data
structures are not for DMA and mostly used by specific cpus/node which
could diff from device's node. Allocating such hot data in device's
node doesn't make sense. Add an API to estimate hardware queue node.
This can be used before blk
8 matches
Mail list logo
