Hi Kent,
Here is an update of few behaviours which caught my eye.
I have not tested any, so it is up to your judgement ;)
Alexander Gordeev (5):
percpu_ida: Fix data race on cpus_have_tags cpumask
percpu_ida: Move waking up waiters out of atomic contexts
percpu_ida: Optimize freeing tags when maximum cache size is 1
percpu_ida: Sanity check initialization parameters
percpu_ida: Allow variable maximum number of cached tags
block/blk-mq-tag.c | 9 +++++----
include/linux/percpu_ida.h | 5 +++--
lib/percpu_ida.c | 38 ++++++++++++++++++++++++++++----------
3 files changed, 36 insertions(+), 16 deletions(-)
--
1.7.7.6
--
Regards,
Alexander Gordeev
[email protected]
On Mon, 2013-10-28 at 11:04 +0100, Alexander Gordeev wrote:
> Hi Kent,
>
> Here is an update of few behaviours which caught my eye.
> I have not tested any, so it is up to your judgement ;)
>
> Alexander Gordeev (5):
> percpu_ida: Fix data race on cpus_have_tags cpumask
> percpu_ida: Move waking up waiters out of atomic contexts
> percpu_ida: Optimize freeing tags when maximum cache size is 1
> percpu_ida: Sanity check initialization parameters
> percpu_ida: Allow variable maximum number of cached tags
>
> block/blk-mq-tag.c | 9 +++++----
> include/linux/percpu_ida.h | 5 +++--
> lib/percpu_ida.c | 38 ++++++++++++++++++++++++++++----------
> 3 files changed, 36 insertions(+), 16 deletions(-)
>
Hi Alexander,
So this series is atop Shaohua's conversion blk-mq patch to use
percpu_ida, right..?
I'll wait for Kent's feedback on these, and end-up dropping the
blk-mq-tag part from Patch #5 for the moment..
--nab
On Mon, Oct 28, 2013 at 11:04:44AM +0100, Alexander Gordeev wrote:
> Hi Kent,
>
> Here is an update of few behaviours which caught my eye.
> I have not tested any, so it is up to your judgement ;)
>
> Alexander Gordeev (5):
> percpu_ida: Fix data race on cpus_have_tags cpumask
> percpu_ida: Move waking up waiters out of atomic contexts
> percpu_ida: Optimize freeing tags when maximum cache size is 1
> percpu_ida: Sanity check initialization parameters
> percpu_ida: Allow variable maximum number of cached tags
>
> block/blk-mq-tag.c | 9 +++++----
> include/linux/percpu_ida.h | 5 +++--
> lib/percpu_ida.c | 38 ++++++++++++++++++++++++++++----------
> 3 files changed, 36 insertions(+), 16 deletions(-)
What tree is this series based on? Patch 3 is based on something not in
mainline.
On Mon, Oct 28, 2013 at 02:23:52PM -0700, Kent Overstreet wrote:
> What tree is this series based on? Patch 3 is based on something not in
> mainline.
This is against the Jens's tree:
git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git
Sorry for missing to note right away.
--
Regards,
Alexander Gordeev
[email protected]
On Mon, Oct 28, 2013 at 02:29:06PM -0700, Nicholas A. Bellinger wrote:
Hi Nicholas,
> So this series is atop Shaohua's conversion blk-mq patch to use
> percpu_ida, right..?
Yep.
> I'll wait for Kent's feedback on these, and end-up dropping the
> blk-mq-tag part from Patch #5 for the moment..
Yeah, looking forward to try AHCI with percpu_ida.
> --nab
--
Regards,
Alexander Gordeev
[email protected]
On Mon, Oct 28, 2013 at 11:04:44AM +0100, Alexander Gordeev wrote:
> Here is an update of few behaviours which caught my eye.
> I have not tested any, so it is up to your judgement ;)
Hi Kent et al.
Can we go ahead with this series?
Thanks!
--
Regards,
Alexander Gordeev
[email protected]
On Tue, Jan 14, 2014 at 12:41:53PM +0100, Alexander Gordeev wrote:
> On Mon, Oct 28, 2013 at 11:04:44AM +0100, Alexander Gordeev wrote:
> > Here is an update of few behaviours which caught my eye.
> > I have not tested any, so it is up to your judgement ;)
>
> Hi Kent et al.
>
> Can we go ahead with this series?
I must've seen it at some point, but I can't find this patch series anywhere.
Can you resend it to me?