Skip to content

Commit

Permalink
cfq: Give a chance for arming slice idle timer in case of group_idle
Browse files Browse the repository at this point in the history
In below scenario blkio cgroup does not work as per their assigned
weights :-
1. When the underlying device is nonrotational with a single HW queue
with depth of >= CFQ_HW_QUEUE_MIN
2. When the use case is forming two blkio cgroups cg1(weight 1000) &
cg2(wight 100) and two processes(file1 and file2) doing sync IO in
their respective blkio cgroups.

For above usecase result of fio (without this patch):-
file1: (groupid=0, jobs=1): err= 0: pid=685: Thu Jan  1 19:41:49 1970
  write: IOPS=1315, BW=41.1MiB/s (43.1MB/s)(1024MiB/24906msec)
<...>
file2: (groupid=0, jobs=1): err= 0: pid=686: Thu Jan  1 19:41:49 1970
  write: IOPS=1295, BW=40.5MiB/s (42.5MB/s)(1024MiB/25293msec)
<...>
// both the process BW is equal even though they belong to diff.
cgroups with weight of 1000(cg1) and 100(cg2)

In above case (for non rotational NCQ devices),
as soon as the request from cg1 is completed and even
though it is provided with higher set_slice=10, because of CFQ
algorithm when the driver tries to fetch the request, CFQ expires
this group without providing any idle time nor weight priority
and schedules another cfq group (in this case cg2).
And thus both cfq groups(cg1 & cg2) keep alternating to get the
disk time and hence loses the cgroup weight based scheduling.

Below patch gives a chance to cfq algorithm (cfq_arm_slice_timer)
to arm the slice timer in case group_idle is enabled.
In case if group_idle is also not required (including for nonrotational
NCQ drives), we need to explicitly set group_idle = 0 from sysfs for
such cases.

With this patch result of fio(for above usecase) :-
file1: (groupid=0, jobs=1): err= 0: pid=690: Thu Jan  1 00:06:08 1970
  write: IOPS=1706, BW=53.3MiB/s (55.9MB/s)(1024MiB/19197msec)
<..>
file2: (groupid=0, jobs=1): err= 0: pid=691: Thu Jan  1 00:06:08 1970
  write: IOPS=1043, BW=32.6MiB/s (34.2MB/s)(1024MiB/31401msec)
<..>
// In this processes BW is as per their respective cgroups weight.

Signed-off-by: Ritesh Harjani <riteshh@codeaurora.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
  • Loading branch information
Ritesh Harjani authored and axboe committed Aug 11, 2017
1 parent edaf942 commit b3193bc
Showing 1 changed file with 2 additions and 1 deletion.
3 changes: 2 additions & 1 deletion block/cfq-iosched.c
Original file line number Diff line number Diff line change
Expand Up @@ -2934,7 +2934,8 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
* for devices that support queuing, otherwise we still have a problem
* with sync vs async workloads.
*/
if (blk_queue_nonrot(cfqd->queue) && cfqd->hw_tag)
if (blk_queue_nonrot(cfqd->queue) && cfqd->hw_tag &&
!cfqd->cfq_group_idle)
return;

WARN_ON(!RB_EMPTY_ROOT(&cfqq->sort_list));
Expand Down

0 comments on commit b3193bc

Please sign in to comment.