| Message ID | 20230421025046.4008499-13-tj@kernel.org (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | None | expand |
On Thu, Apr 20, 2023 at 04:50:36PM -1000, Tejun Heo wrote: > BACKGROUND > ========== > > When multiple work items are queued to a workqueue, their execution order > doesn't match the queueing order. They may get executed in any order and > simultaneously. When fully serialized execution - one by one in the queueing > order - is needed, an ordered workqueue should be used which can be created > with alloc_ordered_workqueue(). > > However, alloc_ordered_workqueue() was a later addition. Before it, an > ordered workqueue could be obtained by creating an UNBOUND workqueue with > @max_active==1. This originally was an implementation side-effect which was > broken by 4c16bd327c74 ("workqueue: restore WQ_UNBOUND/max_active==1 to be > ordered"). Because there were users that depended on the ordered execution, > 5c0338c68706 ("workqueue: restore WQ_UNBOUND/max_active==1 to be ordered") > made workqueue allocation path to implicitly promote UNBOUND workqueues w/ > @max_active==1 to ordered workqueues. > > While this has worked okay, overloading the UNBOUND allocation interface > this way creates other issues. It's difficult to tell whether a given > workqueue actually needs to be ordered and users that legitimately want a > min concurrency level wq unexpectedly gets an ordered one instead. With > planned UNBOUND workqueue updates to improve execution locality and more > prevalence of chiplet designs which can benefit from such improvements, this > isn't a state we wanna be in forever. > > This patch series audits all callsites that create an UNBOUND workqueue w/ > @max_active==1 and converts them to alloc_ordered_workqueue() as necessary. > > WHAT TO LOOK FOR > ================ > > The conversions are from > > alloc_workqueue(WQ_UNBOUND | flags, 1, args..) > > to > > alloc_ordered_workqueue(flags, args...) > > which don't cause any functional changes. If you know that fully ordered > execution is not ncessary, please let me know. I'll drop the conversion and > instead add a comment noting the fact to reduce confusion while conversion > is in progress. > > If you aren't fully sure, it's completely fine to let the conversion > through. The behavior will stay exactly the same and we can always > reconsider later. > > As there are follow-up workqueue core changes, I'd really appreciate if the > patch can be routed through the workqueue tree w/ your acks. Thanks. > > Signed-off-by: Tejun Heo <tj@kernel.org> > Cc: "James E.J. Bottomley" <jejb@linux.ibm.com> > Cc: "Martin K. Petersen" <martin.petersen@oracle.com> > Cc: linux-scsi@vger.kernel.org > --- > drivers/scsi/NCR5380.c | 5 ++--- > drivers/scsi/hosts.c | 12 ++++++------ > drivers/scsi/libiscsi.c | 5 ++--- > 3 files changed, 10 insertions(+), 12 deletions(-) > The conversions look good to me. Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Applied to wq/for-6.5-cleanup-ordered. Thanks.
On Mon, May 08, 2023 at 01:57:30PM -1000, Tejun Heo wrote:
> Applied to wq/for-6.5-cleanup-ordered.
Oops, strike that. All scsi core workqueues have WQ_SYSFS set which means
that their max_active could be adjusted upwards through sysfs. The shouldn't
be ordered workqueues. This only leaves NCR5380 the only remaining
conversion candidate; however, that one only uses a single work item, so the
better thing to do there is using the default @max_active instead.
I'm dropping this patch and will add a patch for NCR5380 in the next round.
Thanks.
diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index ca85bddb582b..b18dd4591492 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -415,9 +415,8 @@ static int NCR5380_init(struct Scsi_Host *instance, int flags) hostdata->flags = flags; INIT_WORK(&hostdata->main_task, NCR5380_main); - hostdata->work_q = alloc_workqueue("ncr5380_%d", - WQ_UNBOUND | WQ_MEM_RECLAIM, - 1, instance->host_no); + hostdata->work_q = alloc_ordered_workqueue("ncr5380_%d", + WQ_MEM_RECLAIM, instance->host_no); if (!hostdata->work_q) return -ENOMEM; diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c index 9b6fbbe15d92..30bf9f49ca6c 100644 --- a/drivers/scsi/hosts.c +++ b/drivers/scsi/hosts.c @@ -294,9 +294,9 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev, if (shost->transportt->create_work_queue) { snprintf(shost->work_q_name, sizeof(shost->work_q_name), "scsi_wq_%d", shost->host_no); - shost->work_q = alloc_workqueue("%s", - WQ_SYSFS | __WQ_LEGACY | WQ_MEM_RECLAIM | WQ_UNBOUND, - 1, shost->work_q_name); + shost->work_q = alloc_ordered_workqueue("%s", + WQ_SYSFS | __WQ_LEGACY | WQ_MEM_RECLAIM, + shost->work_q_name); if (!shost->work_q) { error = -EINVAL; @@ -510,9 +510,9 @@ struct Scsi_Host *scsi_host_alloc(struct scsi_host_template *sht, int privsize) goto fail; } - shost->tmf_work_q = alloc_workqueue("scsi_tmf_%d", - WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, - 1, shost->host_no); + shost->tmf_work_q = alloc_ordered_workqueue("scsi_tmf_%d", + WQ_MEM_RECLAIM | WQ_SYSFS, + shost->host_no); if (!shost->tmf_work_q) { shost_printk(KERN_WARNING, shost, "failed to create tmf workq\n"); diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c index 127f3d7f19dc..d0eba590dc69 100644 --- a/drivers/scsi/libiscsi.c +++ b/drivers/scsi/libiscsi.c @@ -2907,9 +2907,8 @@ struct Scsi_Host *iscsi_host_alloc(struct scsi_host_template *sht, ihost = shost_priv(shost); if (xmit_can_sleep) { - ihost->workq = alloc_workqueue("iscsi_q_%d", - WQ_SYSFS | __WQ_LEGACY | WQ_MEM_RECLAIM | WQ_UNBOUND, - 1, shost->host_no); + ihost->workq = alloc_ordered_workqueue("iscsi_q_%d", + WQ_SYSFS | __WQ_LEGACY | WQ_MEM_RECLAIM, shost->host_no); if (!ihost->workq) goto free_host; }
BACKGROUND ========== When multiple work items are queued to a workqueue, their execution order doesn't match the queueing order. They may get executed in any order and simultaneously. When fully serialized execution - one by one in the queueing order - is needed, an ordered workqueue should be used which can be created with alloc_ordered_workqueue(). However, alloc_ordered_workqueue() was a later addition. Before it, an ordered workqueue could be obtained by creating an UNBOUND workqueue with @max_active==1. This originally was an implementation side-effect which was broken by 4c16bd327c74 ("workqueue: restore WQ_UNBOUND/max_active==1 to be ordered"). Because there were users that depended on the ordered execution, 5c0338c68706 ("workqueue: restore WQ_UNBOUND/max_active==1 to be ordered") made workqueue allocation path to implicitly promote UNBOUND workqueues w/ @max_active==1 to ordered workqueues. While this has worked okay, overloading the UNBOUND allocation interface this way creates other issues. It's difficult to tell whether a given workqueue actually needs to be ordered and users that legitimately want a min concurrency level wq unexpectedly gets an ordered one instead. With planned UNBOUND workqueue updates to improve execution locality and more prevalence of chiplet designs which can benefit from such improvements, this isn't a state we wanna be in forever. This patch series audits all callsites that create an UNBOUND workqueue w/ @max_active==1 and converts them to alloc_ordered_workqueue() as necessary. WHAT TO LOOK FOR ================ The conversions are from alloc_workqueue(WQ_UNBOUND | flags, 1, args..) to alloc_ordered_workqueue(flags, args...) which don't cause any functional changes. If you know that fully ordered execution is not ncessary, please let me know. I'll drop the conversion and instead add a comment noting the fact to reduce confusion while conversion is in progress. If you aren't fully sure, it's completely fine to let the conversion through. The behavior will stay exactly the same and we can always reconsider later. As there are follow-up workqueue core changes, I'd really appreciate if the patch can be routed through the workqueue tree w/ your acks. Thanks. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: "James E.J. Bottomley" <jejb@linux.ibm.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: linux-scsi@vger.kernel.org --- drivers/scsi/NCR5380.c | 5 ++--- drivers/scsi/hosts.c | 12 ++++++------ drivers/scsi/libiscsi.c | 5 ++--- 3 files changed, 10 insertions(+), 12 deletions(-)