mbox series

[BUGFIX/IMPROVEMENT,0/4] bfq: two fixes and one improvement related to I/O control

Message ID 20180816165118.17402-1-paolo.valente@linaro.org
Headers show
Series bfq: two fixes and one improvement related to I/O control | expand

Message

Paolo Valente Aug. 16, 2018, 4:51 p.m. UTC 
Hi Jens,

while working a little bit on cgroups I/O control, I found two nasty
bugs in bfq. They break bandwidth control in simple configurations
with one-process groups. These bugs are fixed by the first two patches
in this series.

These fixes improved I/O control so much, that I could reduce the
write overcharge factor, used by bfq to counter write-induced
issues. This reduction is performed by the third patch.

The fourth patch contains a little code improvement I made in a
function that has to do with I/O control.

I hope we are still in time for 4.19.

Thanks,
Paolo

Paolo Valente (4):
  block, bfq: readd missing reset of parent-entity service
  block, bfq: always update the budget of an entity when needed
  block, bfq: reduce write overcharge
  block, bfq: improve code of bfq_bfqq_charge_time

 block/bfq-iosched.c | 54 +++++++++++++++++++++++++++++++++++++++--------------
 block/bfq-wf2q.c    | 22 +++++++++++-----------
 2 files changed, 51 insertions(+), 25 deletions(-)

--
2.16.1

Comments

Jens Axboe Aug. 16, 2018, 7:10 p.m. UTC | #1 
On 8/16/18 10:51 AM, Paolo Valente wrote:
> Hi Jens,

> 

> while working a little bit on cgroups I/O control, I found two nasty

> bugs in bfq. They break bandwidth control in simple configurations

> with one-process groups. These bugs are fixed by the first two patches

> in this series.

> 

> These fixes improved I/O control so much, that I could reduce the

> write overcharge factor, used by bfq to counter write-induced

> issues. This reduction is performed by the third patch.

> 

> The fourth patch contains a little code improvement I made in a

> function that has to do with I/O control.

> 

> I hope we are still in time for 4.19.


Patches look fine to me. You are in fact several weeks late for 4.19, I
need to have anything that's going in to 4.19 by -rc6 or -rc7 time
(depending on whether we have an -rc8 or not). This is to ensure that it
gets plenty of time in linux-next as well. I've queued up patches this
time (since they were small), but please send them in in due time next
time.

-- 
Jens Axboe
Paolo Valente Aug. 17, 2018, 4:21 p.m. UTC | #2 
> Il giorno 16 ago 2018, alle ore 21:10, Jens Axboe <axboe@kernel.dk> ha scritto:

> 

> On 8/16/18 10:51 AM, Paolo Valente wrote:

>> Hi Jens,

>> 

>> while working a little bit on cgroups I/O control, I found two nasty

>> bugs in bfq. They break bandwidth control in simple configurations

>> with one-process groups. These bugs are fixed by the first two patches

>> in this series.

>> 

>> These fixes improved I/O control so much, that I could reduce the

>> write overcharge factor, used by bfq to counter write-induced

>> issues. This reduction is performed by the third patch.

>> 

>> The fourth patch contains a little code improvement I made in a

>> function that has to do with I/O control.

>> 

>> I hope we are still in time for 4.19.

> 

> Patches look fine to me. You are in fact several weeks late for 4.19, I

> need to have anything that's going in to 4.19 by -rc6 or -rc7 time

> (depending on whether we have an -rc8 or not). This is to ensure that it

> gets plenty of time in linux-next as well. I've queued up patches this

> time (since they were small), but please send them in in due time next

> time.

> 


Thanks for your patience. Next time I'll respect due time.

Paolo

> -- 

> Jens Axboe

>