| Message ID | 8f4ff428-e158-0df5-cf54-ae3cdea7ad1f@kernel.dk (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Delegated to: | Mike Snitzer |
| Headers | show |
On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: > On 06/29/2017 10:00 AM, Jens Axboe wrote: >> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>> >>>>> This has been done before, but I've never really liked it. The reason is >>>>> that it means that reading the part stat inflight count now has to >>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>> How many CPUs were configured? When I last tested this a few years ago >>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>> latencies), it was a net loss. >>>> >>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>> 96 cores, and dual numa nodes) too, the performance can be >>>> recovered basically if per numa-node counter is introduced and >>>> used in this case, but the patch was never posted out. >>>> If anyone is interested in that, I can rebase the patch on current >>>> block tree and post out. I guess the performance issue might be >>>> related with system cache coherency implementation more or less. >>>> This issue on ARM64 can be observed with the following userspace >>>> atomic counting test too: >>>> >>>> http://kernel.ubuntu.com/~ming/test/cache/ >>> >>> How well did the per-node thing work? Doesn't seem to me like it would >>> go far enough. And per CPU is too much. One potential improvement would >>> be to change the part_stat_read() to just loop online CPUs, instead of >>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>> ensure the stats are sane. Often there's a huge difference between >>> NR_CPUS configured and what the system has. As Brian states, RH ships >>> with 2048, while I doubt a lot of customers actually run that... >>> >>> Outside of coming up with a more clever data structure that is fully >>> CPU topology aware, one thing that could work is just having X cache >>> line separated read/write inflight counters per node, where X is some >>> suitable value (like 4). That prevents us from having cross node >>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>> provide a nice balance between cost of incrementing the inflight >>> counting, and the cost of looping for reading it. >>> >>> And that brings me to the next part... >>> >>>>> I do agree that we should do something about it, and it's one of those >>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>> up. It's just not great as it currently stands, but I don't think per >>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>> counter. >>>> >>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>> 'in_flight' counter. I thought about it before, but never got a >>>> perfect solution, and looks it is a bit hard, :-) >>> >>> The tags are already a bit spread out, so it's worth a shot. That would >>> remove the need to do anything in the inc/dec path, as the tags already >>> do that. The inlight count could be easily retrieved with >>> sbitmap_weight(). The only issue here is that we need separate read and >>> write counters, and the weight would obviously only get us the total >>> count. But we can have a slower path for that, just iterate the tags and >>> count them. The fast path only cares about total count. >>> >>> Let me try that out real quick. >> >> Well, that only works for whole disk stats, of course... There's no way >> around iterating the tags and checking for this to truly work. > > Totally untested proof of concept for using the tags for this. I based > this on top of Brian's patch, so it includes his patch plus the > _double() stuff I did which is no longer really needed. > > > diff --git a/block/bio.c b/block/bio.c > index 9cf98b29588a..ec99d9ba0f33 100644 > --- a/block/bio.c > +++ b/block/bio.c > @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, > part_round_stats(cpu, part); > part_stat_inc(cpu, part, ios[rw]); > part_stat_add(cpu, part, sectors[rw], sectors); > - part_inc_in_flight(part, rw); > + part_inc_in_flight(cpu, part, rw); > > part_stat_unlock(); > } > @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, > > part_stat_add(cpu, part, ticks[rw], duration); > part_round_stats(cpu, part); > - part_dec_in_flight(part, rw); > + part_dec_in_flight(cpu, part, rw); > > part_stat_unlock(); > } > diff --git a/block/blk-core.c b/block/blk-core.c > index af393d5a9680..6ab2efbe940b 100644 > --- a/block/blk-core.c > +++ b/block/blk-core.c > @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) > > part_stat_inc(cpu, part, ios[rw]); > part_stat_add(cpu, part, ticks[rw], duration); > - part_round_stats(cpu, part); > - part_dec_in_flight(part, rw); > + > + if (req->q->mq_ops) > + part_round_stats_mq(req->q, cpu, part); > + else { > + part_round_stats(cpu, part); > + part_dec_in_flight(cpu, part, rw); > + } > > hd_struct_put(part); > part_stat_unlock(); > @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) > part = &rq->rq_disk->part0; > hd_struct_get(part); > } > - part_round_stats(cpu, part); > - part_inc_in_flight(part, rw); > + if (rq->q->mq_ops) > + part_round_stats_mq(rq->q, cpu, part); > + else { > + part_round_stats(cpu, part); > + part_inc_in_flight(cpu, part, rw); > + } > rq->part = part; > } > > diff --git a/block/blk-merge.c b/block/blk-merge.c > index 99038830fb42..3b5eb2d4b964 100644 > --- a/block/blk-merge.c > +++ b/block/blk-merge.c > @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) > part = req->part; > > part_round_stats(cpu, part); > - part_dec_in_flight(part, rq_data_dir(req)); > + part_dec_in_flight(cpu, part, rq_data_dir(req)); > > hd_struct_put(part); > part_stat_unlock(); > diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c > index d0be72ccb091..a7b897740c47 100644 > --- a/block/blk-mq-tag.c > +++ b/block/blk-mq-tag.c > @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) > bitnr += tags->nr_reserved_tags; > rq = tags->rqs[bitnr]; > > - if (rq->q == hctx->queue) > + if (rq && rq->q == hctx->queue) > iter_data->fn(hctx, rq, iter_data->data, reserved); > return true; > } > diff --git a/block/blk-mq.c b/block/blk-mq.c > index 05dfa3f270ae..cad4d2c26285 100644 > --- a/block/blk-mq.c > +++ b/block/blk-mq.c > @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); > static void blk_mq_poll_stats_start(struct request_queue *q); > static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); > > +struct mq_inflight { > + struct hd_struct *part; > + unsigned int inflight; > +}; > + > +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, > + struct request *rq, void *priv, > + bool reserved) > +{ > + struct mq_inflight *mi = priv; > + > + if (rq->part == mi->part && > + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) > + mi->inflight++; > +} > + > +unsigned long part_in_flight_mq(struct request_queue *q, > + struct hd_struct *part) > +{ > + struct mq_inflight mi = { .part = part, .inflight = 0 }; > + > + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); > + return mi.inflight; > +} Compared with the totally percpu approach, this way might help 1:M or N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to each CPU(especially there are huge hw queues on a big system), :-( Thanks, Ming Lei -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel
On 06/29/2017 07:20 PM, Ming Lei wrote: > On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: >> On 06/29/2017 10:00 AM, Jens Axboe wrote: >>> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>>> >>>>>> This has been done before, but I've never really liked it. The reason is >>>>>> that it means that reading the part stat inflight count now has to >>>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>>> How many CPUs were configured? When I last tested this a few years ago >>>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>>> latencies), it was a net loss. >>>>> >>>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>>> 96 cores, and dual numa nodes) too, the performance can be >>>>> recovered basically if per numa-node counter is introduced and >>>>> used in this case, but the patch was never posted out. >>>>> If anyone is interested in that, I can rebase the patch on current >>>>> block tree and post out. I guess the performance issue might be >>>>> related with system cache coherency implementation more or less. >>>>> This issue on ARM64 can be observed with the following userspace >>>>> atomic counting test too: >>>>> >>>>> http://kernel.ubuntu.com/~ming/test/cache/ >>>> >>>> How well did the per-node thing work? Doesn't seem to me like it would >>>> go far enough. And per CPU is too much. One potential improvement would >>>> be to change the part_stat_read() to just loop online CPUs, instead of >>>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>>> ensure the stats are sane. Often there's a huge difference between >>>> NR_CPUS configured and what the system has. As Brian states, RH ships >>>> with 2048, while I doubt a lot of customers actually run that... >>>> >>>> Outside of coming up with a more clever data structure that is fully >>>> CPU topology aware, one thing that could work is just having X cache >>>> line separated read/write inflight counters per node, where X is some >>>> suitable value (like 4). That prevents us from having cross node >>>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>>> provide a nice balance between cost of incrementing the inflight >>>> counting, and the cost of looping for reading it. >>>> >>>> And that brings me to the next part... >>>> >>>>>> I do agree that we should do something about it, and it's one of those >>>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>>> up. It's just not great as it currently stands, but I don't think per >>>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>>> counter. >>>>> >>>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>>> 'in_flight' counter. I thought about it before, but never got a >>>>> perfect solution, and looks it is a bit hard, :-) >>>> >>>> The tags are already a bit spread out, so it's worth a shot. That would >>>> remove the need to do anything in the inc/dec path, as the tags already >>>> do that. The inlight count could be easily retrieved with >>>> sbitmap_weight(). The only issue here is that we need separate read and >>>> write counters, and the weight would obviously only get us the total >>>> count. But we can have a slower path for that, just iterate the tags and >>>> count them. The fast path only cares about total count. >>>> >>>> Let me try that out real quick. >>> >>> Well, that only works for whole disk stats, of course... There's no way >>> around iterating the tags and checking for this to truly work. >> >> Totally untested proof of concept for using the tags for this. I based >> this on top of Brian's patch, so it includes his patch plus the >> _double() stuff I did which is no longer really needed. >> >> >> diff --git a/block/bio.c b/block/bio.c >> index 9cf98b29588a..ec99d9ba0f33 100644 >> --- a/block/bio.c >> +++ b/block/bio.c >> @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, >> part_round_stats(cpu, part); >> part_stat_inc(cpu, part, ios[rw]); >> part_stat_add(cpu, part, sectors[rw], sectors); >> - part_inc_in_flight(part, rw); >> + part_inc_in_flight(cpu, part, rw); >> >> part_stat_unlock(); >> } >> @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, >> >> part_stat_add(cpu, part, ticks[rw], duration); >> part_round_stats(cpu, part); >> - part_dec_in_flight(part, rw); >> + part_dec_in_flight(cpu, part, rw); >> >> part_stat_unlock(); >> } >> diff --git a/block/blk-core.c b/block/blk-core.c >> index af393d5a9680..6ab2efbe940b 100644 >> --- a/block/blk-core.c >> +++ b/block/blk-core.c >> @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) >> >> part_stat_inc(cpu, part, ios[rw]); >> part_stat_add(cpu, part, ticks[rw], duration); >> - part_round_stats(cpu, part); >> - part_dec_in_flight(part, rw); >> + >> + if (req->q->mq_ops) >> + part_round_stats_mq(req->q, cpu, part); >> + else { >> + part_round_stats(cpu, part); >> + part_dec_in_flight(cpu, part, rw); >> + } >> >> hd_struct_put(part); >> part_stat_unlock(); >> @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) >> part = &rq->rq_disk->part0; >> hd_struct_get(part); >> } >> - part_round_stats(cpu, part); >> - part_inc_in_flight(part, rw); >> + if (rq->q->mq_ops) >> + part_round_stats_mq(rq->q, cpu, part); >> + else { >> + part_round_stats(cpu, part); >> + part_inc_in_flight(cpu, part, rw); >> + } >> rq->part = part; >> } >> >> diff --git a/block/blk-merge.c b/block/blk-merge.c >> index 99038830fb42..3b5eb2d4b964 100644 >> --- a/block/blk-merge.c >> +++ b/block/blk-merge.c >> @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) >> part = req->part; >> >> part_round_stats(cpu, part); >> - part_dec_in_flight(part, rq_data_dir(req)); >> + part_dec_in_flight(cpu, part, rq_data_dir(req)); >> >> hd_struct_put(part); >> part_stat_unlock(); >> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c >> index d0be72ccb091..a7b897740c47 100644 >> --- a/block/blk-mq-tag.c >> +++ b/block/blk-mq-tag.c >> @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) >> bitnr += tags->nr_reserved_tags; >> rq = tags->rqs[bitnr]; >> >> - if (rq->q == hctx->queue) >> + if (rq && rq->q == hctx->queue) >> iter_data->fn(hctx, rq, iter_data->data, reserved); >> return true; >> } >> diff --git a/block/blk-mq.c b/block/blk-mq.c >> index 05dfa3f270ae..cad4d2c26285 100644 >> --- a/block/blk-mq.c >> +++ b/block/blk-mq.c >> @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); >> static void blk_mq_poll_stats_start(struct request_queue *q); >> static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); >> >> +struct mq_inflight { >> + struct hd_struct *part; >> + unsigned int inflight; >> +}; >> + >> +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, >> + struct request *rq, void *priv, >> + bool reserved) >> +{ >> + struct mq_inflight *mi = priv; >> + >> + if (rq->part == mi->part && >> + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) >> + mi->inflight++; >> +} >> + >> +unsigned long part_in_flight_mq(struct request_queue *q, >> + struct hd_struct *part) >> +{ >> + struct mq_inflight mi = { .part = part, .inflight = 0 }; >> + >> + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); >> + return mi.inflight; >> +} > > Compared with the totally percpu approach, this way might help 1:M or > N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to > each CPU(especially there are huge hw queues on a big system), :-( Not disagreeing with that, without having some mechanism to only loop queues that have pending requests. That would be similar to the ctx_map for sw to hw queues. But I don't think that would be worthwhile doing, I like your pnode approach better. However, I'm still not fully convinced that one per node is enough to get the scalability we need. Would be great if Brian could re-test with your updated patch, so we know how it works for him at least.
On 06/29/2017 09:17 PM, Jens Axboe wrote: > On 06/29/2017 07:20 PM, Ming Lei wrote: >> On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: >>> On 06/29/2017 10:00 AM, Jens Axboe wrote: >>>> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>>>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>>>> >>>>>>> This has been done before, but I've never really liked it. The reason is >>>>>>> that it means that reading the part stat inflight count now has to >>>>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>>>> How many CPUs were configured? When I last tested this a few years ago >>>>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>>>> latencies), it was a net loss. >>>>>> >>>>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>>>> 96 cores, and dual numa nodes) too, the performance can be >>>>>> recovered basically if per numa-node counter is introduced and >>>>>> used in this case, but the patch was never posted out. >>>>>> If anyone is interested in that, I can rebase the patch on current >>>>>> block tree and post out. I guess the performance issue might be >>>>>> related with system cache coherency implementation more or less. >>>>>> This issue on ARM64 can be observed with the following userspace >>>>>> atomic counting test too: >>>>>> >>>>>> http://kernel.ubuntu.com/~ming/test/cache/ >>>>> >>>>> How well did the per-node thing work? Doesn't seem to me like it would >>>>> go far enough. And per CPU is too much. One potential improvement would >>>>> be to change the part_stat_read() to just loop online CPUs, instead of >>>>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>>>> ensure the stats are sane. Often there's a huge difference between >>>>> NR_CPUS configured and what the system has. As Brian states, RH ships >>>>> with 2048, while I doubt a lot of customers actually run that... >>>>> >>>>> Outside of coming up with a more clever data structure that is fully >>>>> CPU topology aware, one thing that could work is just having X cache >>>>> line separated read/write inflight counters per node, where X is some >>>>> suitable value (like 4). That prevents us from having cross node >>>>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>>>> provide a nice balance between cost of incrementing the inflight >>>>> counting, and the cost of looping for reading it. >>>>> >>>>> And that brings me to the next part... >>>>> >>>>>>> I do agree that we should do something about it, and it's one of those >>>>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>>>> up. It's just not great as it currently stands, but I don't think per >>>>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>>>> counter. >>>>>> >>>>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>>>> 'in_flight' counter. I thought about it before, but never got a >>>>>> perfect solution, and looks it is a bit hard, :-) >>>>> >>>>> The tags are already a bit spread out, so it's worth a shot. That would >>>>> remove the need to do anything in the inc/dec path, as the tags already >>>>> do that. The inlight count could be easily retrieved with >>>>> sbitmap_weight(). The only issue here is that we need separate read and >>>>> write counters, and the weight would obviously only get us the total >>>>> count. But we can have a slower path for that, just iterate the tags and >>>>> count them. The fast path only cares about total count. >>>>> >>>>> Let me try that out real quick. >>>> >>>> Well, that only works for whole disk stats, of course... There's no way >>>> around iterating the tags and checking for this to truly work. >>> >>> Totally untested proof of concept for using the tags for this. I based >>> this on top of Brian's patch, so it includes his patch plus the >>> _double() stuff I did which is no longer really needed. >>> >>> >>> diff --git a/block/bio.c b/block/bio.c >>> index 9cf98b29588a..ec99d9ba0f33 100644 >>> --- a/block/bio.c >>> +++ b/block/bio.c >>> @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, >>> part_round_stats(cpu, part); >>> part_stat_inc(cpu, part, ios[rw]); >>> part_stat_add(cpu, part, sectors[rw], sectors); >>> - part_inc_in_flight(part, rw); >>> + part_inc_in_flight(cpu, part, rw); >>> >>> part_stat_unlock(); >>> } >>> @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, >>> >>> part_stat_add(cpu, part, ticks[rw], duration); >>> part_round_stats(cpu, part); >>> - part_dec_in_flight(part, rw); >>> + part_dec_in_flight(cpu, part, rw); >>> >>> part_stat_unlock(); >>> } >>> diff --git a/block/blk-core.c b/block/blk-core.c >>> index af393d5a9680..6ab2efbe940b 100644 >>> --- a/block/blk-core.c >>> +++ b/block/blk-core.c >>> @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) >>> >>> part_stat_inc(cpu, part, ios[rw]); >>> part_stat_add(cpu, part, ticks[rw], duration); >>> - part_round_stats(cpu, part); >>> - part_dec_in_flight(part, rw); >>> + >>> + if (req->q->mq_ops) >>> + part_round_stats_mq(req->q, cpu, part); >>> + else { >>> + part_round_stats(cpu, part); >>> + part_dec_in_flight(cpu, part, rw); >>> + } >>> >>> hd_struct_put(part); >>> part_stat_unlock(); >>> @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) >>> part = &rq->rq_disk->part0; >>> hd_struct_get(part); >>> } >>> - part_round_stats(cpu, part); >>> - part_inc_in_flight(part, rw); >>> + if (rq->q->mq_ops) >>> + part_round_stats_mq(rq->q, cpu, part); >>> + else { >>> + part_round_stats(cpu, part); >>> + part_inc_in_flight(cpu, part, rw); >>> + } >>> rq->part = part; >>> } >>> >>> diff --git a/block/blk-merge.c b/block/blk-merge.c >>> index 99038830fb42..3b5eb2d4b964 100644 >>> --- a/block/blk-merge.c >>> +++ b/block/blk-merge.c >>> @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) >>> part = req->part; >>> >>> part_round_stats(cpu, part); >>> - part_dec_in_flight(part, rq_data_dir(req)); >>> + part_dec_in_flight(cpu, part, rq_data_dir(req)); >>> >>> hd_struct_put(part); >>> part_stat_unlock(); >>> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c >>> index d0be72ccb091..a7b897740c47 100644 >>> --- a/block/blk-mq-tag.c >>> +++ b/block/blk-mq-tag.c >>> @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) >>> bitnr += tags->nr_reserved_tags; >>> rq = tags->rqs[bitnr]; >>> >>> - if (rq->q == hctx->queue) >>> + if (rq && rq->q == hctx->queue) >>> iter_data->fn(hctx, rq, iter_data->data, reserved); >>> return true; >>> } >>> diff --git a/block/blk-mq.c b/block/blk-mq.c >>> index 05dfa3f270ae..cad4d2c26285 100644 >>> --- a/block/blk-mq.c >>> +++ b/block/blk-mq.c >>> @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); >>> static void blk_mq_poll_stats_start(struct request_queue *q); >>> static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); >>> >>> +struct mq_inflight { >>> + struct hd_struct *part; >>> + unsigned int inflight; >>> +}; >>> + >>> +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, >>> + struct request *rq, void *priv, >>> + bool reserved) >>> +{ >>> + struct mq_inflight *mi = priv; >>> + >>> + if (rq->part == mi->part && >>> + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) >>> + mi->inflight++; >>> +} >>> + >>> +unsigned long part_in_flight_mq(struct request_queue *q, >>> + struct hd_struct *part) >>> +{ >>> + struct mq_inflight mi = { .part = part, .inflight = 0 }; >>> + >>> + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); >>> + return mi.inflight; >>> +} >> >> Compared with the totally percpu approach, this way might help 1:M or >> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >> each CPU(especially there are huge hw queues on a big system), :-( > > Not disagreeing with that, without having some mechanism to only > loop queues that have pending requests. That would be similar to the > ctx_map for sw to hw queues. But I don't think that would be worthwhile > doing, I like your pnode approach better. However, I'm still not fully > convinced that one per node is enough to get the scalability we need. > > Would be great if Brian could re-test with your updated patch, so we > know how it works for him at least. I'll try running with both approaches today and see how they compare. Thanks!!! Brian
On 06/30/2017 07:05 AM, Brian King wrote: > On 06/29/2017 09:17 PM, Jens Axboe wrote: >> On 06/29/2017 07:20 PM, Ming Lei wrote: >>> On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>> On 06/29/2017 10:00 AM, Jens Axboe wrote: >>>>> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>>>>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>>>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>>>>> >>>>>>>> This has been done before, but I've never really liked it. The reason is >>>>>>>> that it means that reading the part stat inflight count now has to >>>>>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>>>>> How many CPUs were configured? When I last tested this a few years ago >>>>>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>>>>> latencies), it was a net loss. >>>>>>> >>>>>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>>>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>>>>> 96 cores, and dual numa nodes) too, the performance can be >>>>>>> recovered basically if per numa-node counter is introduced and >>>>>>> used in this case, but the patch was never posted out. >>>>>>> If anyone is interested in that, I can rebase the patch on current >>>>>>> block tree and post out. I guess the performance issue might be >>>>>>> related with system cache coherency implementation more or less. >>>>>>> This issue on ARM64 can be observed with the following userspace >>>>>>> atomic counting test too: >>>>>>> >>>>>>> http://kernel.ubuntu.com/~ming/test/cache/ >>>>>> >>>>>> How well did the per-node thing work? Doesn't seem to me like it would >>>>>> go far enough. And per CPU is too much. One potential improvement would >>>>>> be to change the part_stat_read() to just loop online CPUs, instead of >>>>>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>>>>> ensure the stats are sane. Often there's a huge difference between >>>>>> NR_CPUS configured and what the system has. As Brian states, RH ships >>>>>> with 2048, while I doubt a lot of customers actually run that... >>>>>> >>>>>> Outside of coming up with a more clever data structure that is fully >>>>>> CPU topology aware, one thing that could work is just having X cache >>>>>> line separated read/write inflight counters per node, where X is some >>>>>> suitable value (like 4). That prevents us from having cross node >>>>>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>>>>> provide a nice balance between cost of incrementing the inflight >>>>>> counting, and the cost of looping for reading it. >>>>>> >>>>>> And that brings me to the next part... >>>>>> >>>>>>>> I do agree that we should do something about it, and it's one of those >>>>>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>>>>> up. It's just not great as it currently stands, but I don't think per >>>>>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>>>>> counter. >>>>>>> >>>>>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>>>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>>>>> 'in_flight' counter. I thought about it before, but never got a >>>>>>> perfect solution, and looks it is a bit hard, :-) >>>>>> >>>>>> The tags are already a bit spread out, so it's worth a shot. That would >>>>>> remove the need to do anything in the inc/dec path, as the tags already >>>>>> do that. The inlight count could be easily retrieved with >>>>>> sbitmap_weight(). The only issue here is that we need separate read and >>>>>> write counters, and the weight would obviously only get us the total >>>>>> count. But we can have a slower path for that, just iterate the tags and >>>>>> count them. The fast path only cares about total count. >>>>>> >>>>>> Let me try that out real quick. >>>>> >>>>> Well, that only works for whole disk stats, of course... There's no way >>>>> around iterating the tags and checking for this to truly work. >>>> >>>> Totally untested proof of concept for using the tags for this. I based >>>> this on top of Brian's patch, so it includes his patch plus the >>>> _double() stuff I did which is no longer really needed. >>>> >>>> >>>> diff --git a/block/bio.c b/block/bio.c >>>> index 9cf98b29588a..ec99d9ba0f33 100644 >>>> --- a/block/bio.c >>>> +++ b/block/bio.c >>>> @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, >>>> part_round_stats(cpu, part); >>>> part_stat_inc(cpu, part, ios[rw]); >>>> part_stat_add(cpu, part, sectors[rw], sectors); >>>> - part_inc_in_flight(part, rw); >>>> + part_inc_in_flight(cpu, part, rw); >>>> >>>> part_stat_unlock(); >>>> } >>>> @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, >>>> >>>> part_stat_add(cpu, part, ticks[rw], duration); >>>> part_round_stats(cpu, part); >>>> - part_dec_in_flight(part, rw); >>>> + part_dec_in_flight(cpu, part, rw); >>>> >>>> part_stat_unlock(); >>>> } >>>> diff --git a/block/blk-core.c b/block/blk-core.c >>>> index af393d5a9680..6ab2efbe940b 100644 >>>> --- a/block/blk-core.c >>>> +++ b/block/blk-core.c >>>> @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) >>>> >>>> part_stat_inc(cpu, part, ios[rw]); >>>> part_stat_add(cpu, part, ticks[rw], duration); >>>> - part_round_stats(cpu, part); >>>> - part_dec_in_flight(part, rw); >>>> + >>>> + if (req->q->mq_ops) >>>> + part_round_stats_mq(req->q, cpu, part); >>>> + else { >>>> + part_round_stats(cpu, part); >>>> + part_dec_in_flight(cpu, part, rw); >>>> + } >>>> >>>> hd_struct_put(part); >>>> part_stat_unlock(); >>>> @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) >>>> part = &rq->rq_disk->part0; >>>> hd_struct_get(part); >>>> } >>>> - part_round_stats(cpu, part); >>>> - part_inc_in_flight(part, rw); >>>> + if (rq->q->mq_ops) >>>> + part_round_stats_mq(rq->q, cpu, part); >>>> + else { >>>> + part_round_stats(cpu, part); >>>> + part_inc_in_flight(cpu, part, rw); >>>> + } >>>> rq->part = part; >>>> } >>>> >>>> diff --git a/block/blk-merge.c b/block/blk-merge.c >>>> index 99038830fb42..3b5eb2d4b964 100644 >>>> --- a/block/blk-merge.c >>>> +++ b/block/blk-merge.c >>>> @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) >>>> part = req->part; >>>> >>>> part_round_stats(cpu, part); >>>> - part_dec_in_flight(part, rq_data_dir(req)); >>>> + part_dec_in_flight(cpu, part, rq_data_dir(req)); >>>> >>>> hd_struct_put(part); >>>> part_stat_unlock(); >>>> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c >>>> index d0be72ccb091..a7b897740c47 100644 >>>> --- a/block/blk-mq-tag.c >>>> +++ b/block/blk-mq-tag.c >>>> @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) >>>> bitnr += tags->nr_reserved_tags; >>>> rq = tags->rqs[bitnr]; >>>> >>>> - if (rq->q == hctx->queue) >>>> + if (rq && rq->q == hctx->queue) >>>> iter_data->fn(hctx, rq, iter_data->data, reserved); >>>> return true; >>>> } >>>> diff --git a/block/blk-mq.c b/block/blk-mq.c >>>> index 05dfa3f270ae..cad4d2c26285 100644 >>>> --- a/block/blk-mq.c >>>> +++ b/block/blk-mq.c >>>> @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); >>>> static void blk_mq_poll_stats_start(struct request_queue *q); >>>> static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); >>>> >>>> +struct mq_inflight { >>>> + struct hd_struct *part; >>>> + unsigned int inflight; >>>> +}; >>>> + >>>> +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, >>>> + struct request *rq, void *priv, >>>> + bool reserved) >>>> +{ >>>> + struct mq_inflight *mi = priv; >>>> + >>>> + if (rq->part == mi->part && >>>> + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) >>>> + mi->inflight++; >>>> +} >>>> + >>>> +unsigned long part_in_flight_mq(struct request_queue *q, >>>> + struct hd_struct *part) >>>> +{ >>>> + struct mq_inflight mi = { .part = part, .inflight = 0 }; >>>> + >>>> + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); >>>> + return mi.inflight; >>>> +} >>> >>> Compared with the totally percpu approach, this way might help 1:M or >>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>> each CPU(especially there are huge hw queues on a big system), :-( >> >> Not disagreeing with that, without having some mechanism to only >> loop queues that have pending requests. That would be similar to the >> ctx_map for sw to hw queues. But I don't think that would be worthwhile >> doing, I like your pnode approach better. However, I'm still not fully >> convinced that one per node is enough to get the scalability we need. >> >> Would be great if Brian could re-test with your updated patch, so we >> know how it works for him at least. > > I'll try running with both approaches today and see how they compare. Focus on Ming's, a variant of that is the most likely path forward, imho. It'd be great to do a quick run on mine as well, just to establish how it compares to mainline, though.
On 06/30/2017 09:08 AM, Jens Axboe wrote: >>>> Compared with the totally percpu approach, this way might help 1:M or >>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>> each CPU(especially there are huge hw queues on a big system), :-( >>> >>> Not disagreeing with that, without having some mechanism to only >>> loop queues that have pending requests. That would be similar to the >>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>> doing, I like your pnode approach better. However, I'm still not fully >>> convinced that one per node is enough to get the scalability we need. >>> >>> Would be great if Brian could re-test with your updated patch, so we >>> know how it works for him at least. >> >> I'll try running with both approaches today and see how they compare. > > Focus on Ming's, a variant of that is the most likely path forward, > imho. It'd be great to do a quick run on mine as well, just to establish > how it compares to mainline, though. On my initial runs, the one from you Jens, appears to perform a bit better, although both are a huge improvement from what I was seeing before. I ran 4k random reads using fio to nullblk in two configurations on my 20 core system with 4 NUMA nodes and 4-way SMT, so 80 logical CPUs. I ran both 80 threads to a single null_blk as well as 80 threads to 80 null_block devices, so one thread per null_blk. This is what I saw on this machine: Using the Per node atomic change from Ming Lei 1 null_blk, 80 threads iops=9376.5K 80 null_blk, 1 thread iops=9523.5K Using the alternate patch from Jens using the tags 1 null_blk, 80 threads iops=9725.8K 80 null_blk, 1 thread iops=9569.4K Its interesting that with this change the single device, 80 threads scenario actually got better than the 80 null_blk scenario. I'll try on a larger machine as well. I've got a 32 core machine I can try this on too. Next week I can work with our performance team on running this on a system with a bunch of nvme devices so we can then test the disk partition case as well and see if there is any noticeable overhead. Thanks, Brian
Hi Bian, On Sat, Jul 1, 2017 at 2:33 AM, Brian King <brking@linux.vnet.ibm.com> wrote: > On 06/30/2017 09:08 AM, Jens Axboe wrote: >>>>> Compared with the totally percpu approach, this way might help 1:M or >>>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>>> each CPU(especially there are huge hw queues on a big system), :-( >>>> >>>> Not disagreeing with that, without having some mechanism to only >>>> loop queues that have pending requests. That would be similar to the >>>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>>> doing, I like your pnode approach better. However, I'm still not fully >>>> convinced that one per node is enough to get the scalability we need. >>>> >>>> Would be great if Brian could re-test with your updated patch, so we >>>> know how it works for him at least. >>> >>> I'll try running with both approaches today and see how they compare. >> >> Focus on Ming's, a variant of that is the most likely path forward, >> imho. It'd be great to do a quick run on mine as well, just to establish >> how it compares to mainline, though. > > On my initial runs, the one from you Jens, appears to perform a bit better, although > both are a huge improvement from what I was seeing before. > > I ran 4k random reads using fio to nullblk in two configurations on my 20 core > system with 4 NUMA nodes and 4-way SMT, so 80 logical CPUs. I ran both 80 threads > to a single null_blk as well as 80 threads to 80 null_block devices, so one thread Could you share what the '80 null_block devices' is? It means you create 80 null_blk devices? Or you create one null_blk and make its hw queue number as 80 via module parameter of ''submit_queues"? I guess we should focus on multi-queue case since it is the normal way of NVMe. > per null_blk. This is what I saw on this machine: > > Using the Per node atomic change from Ming Lei > 1 null_blk, 80 threads > iops=9376.5K > > 80 null_blk, 1 thread > iops=9523.5K > > > Using the alternate patch from Jens using the tags > 1 null_blk, 80 threads > iops=9725.8K > > 80 null_blk, 1 thread > iops=9569.4K If 1 thread means single fio job, looks the number is too too high, that means one random IO can complete in about 0.1us(100ns) on single CPU, not sure if it is possible, :-) Thanks, Ming Lei -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel
On 06/30/2017 05:23 PM, Ming Lei wrote: > Hi Bian, > > On Sat, Jul 1, 2017 at 2:33 AM, Brian King <brking@linux.vnet.ibm.com> wrote: >> On 06/30/2017 09:08 AM, Jens Axboe wrote: >>>>>> Compared with the totally percpu approach, this way might help 1:M or >>>>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>>>> each CPU(especially there are huge hw queues on a big system), :-( >>>>> >>>>> Not disagreeing with that, without having some mechanism to only >>>>> loop queues that have pending requests. That would be similar to the >>>>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>>>> doing, I like your pnode approach better. However, I'm still not fully >>>>> convinced that one per node is enough to get the scalability we need. >>>>> >>>>> Would be great if Brian could re-test with your updated patch, so we >>>>> know how it works for him at least. >>>> >>>> I'll try running with both approaches today and see how they compare. >>> >>> Focus on Ming's, a variant of that is the most likely path forward, >>> imho. It'd be great to do a quick run on mine as well, just to establish >>> how it compares to mainline, though. >> >> On my initial runs, the one from you Jens, appears to perform a bit better, although >> both are a huge improvement from what I was seeing before. >> >> I ran 4k random reads using fio to nullblk in two configurations on my 20 core >> system with 4 NUMA nodes and 4-way SMT, so 80 logical CPUs. I ran both 80 threads >> to a single null_blk as well as 80 threads to 80 null_block devices, so one thread > > Could you share what the '80 null_block devices' is? It means you > create 80 null_blk > devices? Or you create one null_blk and make its hw queue number as 80 > via module > parameter of ''submit_queues"? That's a valid question, was going to ask that too. But I assumed that Brian used submit_queues to set as many queues as he has logical CPUs in the system. > > I guess we should focus on multi-queue case since it is the normal way of NVMe. > >> per null_blk. This is what I saw on this machine: >> >> Using the Per node atomic change from Ming Lei >> 1 null_blk, 80 threads >> iops=9376.5K >> >> 80 null_blk, 1 thread >> iops=9523.5K >> >> >> Using the alternate patch from Jens using the tags >> 1 null_blk, 80 threads >> iops=9725.8K >> >> 80 null_blk, 1 thread >> iops=9569.4K > > If 1 thread means single fio job, looks the number is too too high, that means > one random IO can complete in about 0.1us(100ns) on single CPU, not sure if it > is possible, :-) It means either 1 null_blk device, 80 threads running IO to it. Or 80 null_blk devices, each with a thread running IO to it. See above, he details that it's 80 threads on 80 devices for that case.
On 06/30/2017 06:26 PM, Jens Axboe wrote: > On 06/30/2017 05:23 PM, Ming Lei wrote: >> Hi Bian, >> >> On Sat, Jul 1, 2017 at 2:33 AM, Brian King <brking@linux.vnet.ibm.com> wrote: >>> On 06/30/2017 09:08 AM, Jens Axboe wrote: >>>>>>> Compared with the totally percpu approach, this way might help 1:M or >>>>>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>>>>> each CPU(especially there are huge hw queues on a big system), :-( >>>>>> >>>>>> Not disagreeing with that, without having some mechanism to only >>>>>> loop queues that have pending requests. That would be similar to the >>>>>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>>>>> doing, I like your pnode approach better. However, I'm still not fully >>>>>> convinced that one per node is enough to get the scalability we need. >>>>>> >>>>>> Would be great if Brian could re-test with your updated patch, so we >>>>>> know how it works for him at least. >>>>> >>>>> I'll try running with both approaches today and see how they compare. >>>> >>>> Focus on Ming's, a variant of that is the most likely path forward, >>>> imho. It'd be great to do a quick run on mine as well, just to establish >>>> how it compares to mainline, though. >>> >>> On my initial runs, the one from you Jens, appears to perform a bit better, although >>> both are a huge improvement from what I was seeing before. >>> >>> I ran 4k random reads using fio to nullblk in two configurations on my 20 core >>> system with 4 NUMA nodes and 4-way SMT, so 80 logical CPUs. I ran both 80 threads >>> to a single null_blk as well as 80 threads to 80 null_block devices, so one thread >> >> Could you share what the '80 null_block devices' is? It means you >> create 80 null_blk >> devices? Or you create one null_blk and make its hw queue number as 80 >> via module >> parameter of ''submit_queues"? > > That's a valid question, was going to ask that too. But I assumed that Brian > used submit_queues to set as many queues as he has logical CPUs in the system. >> >> I guess we should focus on multi-queue case since it is the normal way of NVMe. >> >>> per null_blk. This is what I saw on this machine: >>> >>> Using the Per node atomic change from Ming Lei >>> 1 null_blk, 80 threads >>> iops=9376.5K >>> >>> 80 null_blk, 1 thread >>> iops=9523.5K >>> >>> >>> Using the alternate patch from Jens using the tags >>> 1 null_blk, 80 threads >>> iops=9725.8K >>> >>> 80 null_blk, 1 thread >>> iops=9569.4K >> >> If 1 thread means single fio job, looks the number is too too high, that means >> one random IO can complete in about 0.1us(100ns) on single CPU, not sure if it >> is possible, :-) > > It means either 1 null_blk device, 80 threads running IO to it. Or 80 null_blk > devices, each with a thread running IO to it. See above, he details that it's > 80 threads on 80 devices for that case. Right. So the two modes I'm running in are: 1. 80 null_blk devices, each with one submit_queue, with one fio job per null_blk device, so 80 threads total. 80 logical CPUs 2. 1 null_blk device, with 80 submit_queues, 80 fio jobs, 80 logical CPUs. In theory, the two should result in similar numbers. Here are the commands and fio configurations: Scenario #1 modprobe null_blk submit_queues=80 nr_devices=1 irqmode=0 FIO config: [global] buffered=0 invalidate=1 bs=4k iodepth=64 numjobs=80 group_reporting=1 rw=randrw rwmixread=100 rwmixwrite=0 ioengine=libaio runtime=60 time_based [job1] filename=/dev/nullb0 Scenario #2 modprobe null_blk submit_queues=1 nr_devices=80 irqmode=0 FIO config [global] buffered=0 invalidate=1 bs=4k iodepth=64 numjobs=1 group_reporting=1 rw=randrw rwmixread=100 rwmixwrite=0 ioengine=libaio runtime=60 time_based [job1] filename=/dev/nullb0 [job2] filename=/dev/nullb1 [job3] filename=/dev/nullb2 [job4] filename=/dev/nullb3 [job5] filename=/dev/nullb4 [job6] filename=/dev/nullb5 [job7] filename=/dev/nullb6 [job8] filename=/dev/nullb7 [job9] filename=/dev/nullb8 [job10] filename=/dev/nullb9 [job11] filename=/dev/nullb10 [job12] filename=/dev/nullb11 [job13] filename=/dev/nullb12 [job14] filename=/dev/nullb13 [job15] filename=/dev/nullb14 [job16] filename=/dev/nullb15 [job17] filename=/dev/nullb16 [job18] filename=/dev/nullb17 [job19] filename=/dev/nullb18 [job20] filename=/dev/nullb19 [job21] filename=/dev/nullb20 [job22] filename=/dev/nullb21 [job23] filename=/dev/nullb22 [job24] filename=/dev/nullb23 [job25] filename=/dev/nullb24 [job26] filename=/dev/nullb25 [job27] filename=/dev/nullb26 [job28] filename=/dev/nullb27 [job29] filename=/dev/nullb28 [job30] filename=/dev/nullb29 [job31] filename=/dev/nullb30 [job32] filename=/dev/nullb31 [job33] filename=/dev/nullb32 [job34] filename=/dev/nullb33 [job35] filename=/dev/nullb34 [job36] filename=/dev/nullb35 [job37] filename=/dev/nullb36 [job38] filename=/dev/nullb37 [job39] filename=/dev/nullb38 [job40] filename=/dev/nullb39 [job41] filename=/dev/nullb40 [job42] filename=/dev/nullb41 [job43] filename=/dev/nullb42 [job44] filename=/dev/nullb43 [job45] filename=/dev/nullb44 [job46] filename=/dev/nullb45 [job47] filename=/dev/nullb46 [job48] filename=/dev/nullb47 [job49] filename=/dev/nullb48 [job50] filename=/dev/nullb49 [job51] filename=/dev/nullb50 [job52] filename=/dev/nullb51 [job53] filename=/dev/nullb52 [job54] filename=/dev/nullb53 [job55] filename=/dev/nullb54 [job56] filename=/dev/nullb55 [job57] filename=/dev/nullb56 [job58] filename=/dev/nullb57 [job59] filename=/dev/nullb58 [job60] filename=/dev/nullb59 [job61] filename=/dev/nullb60 [job62] filename=/dev/nullb61 [job63] filename=/dev/nullb62 [job64] filename=/dev/nullb63 [job65] filename=/dev/nullb64 [job66] filename=/dev/nullb65 [job67] filename=/dev/nullb66 [job68] filename=/dev/nullb67 [job69] filename=/dev/nullb68 [job70] filename=/dev/nullb69 [job71] filename=/dev/nullb70 [job72] filename=/dev/nullb71 [job73] filename=/dev/nullb72 [job74] filename=/dev/nullb73 [job75] filename=/dev/nullb74 [job76] filename=/dev/nullb75 [job77] filename=/dev/nullb76 [job78] filename=/dev/nullb77 [job79] filename=/dev/nullb78 [job80] filename=/dev/nullb79 -Brian
On 06/30/2017 08:08 AM, Jens Axboe wrote: > On 06/30/2017 07:05 AM, Brian King wrote: >> On 06/29/2017 09:17 PM, Jens Axboe wrote: >>> On 06/29/2017 07:20 PM, Ming Lei wrote: >>>> On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>> On 06/29/2017 10:00 AM, Jens Axboe wrote: >>>>>> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>>>>>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>>>>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>>>>>> >>>>>>>>> This has been done before, but I've never really liked it. The reason is >>>>>>>>> that it means that reading the part stat inflight count now has to >>>>>>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>>>>>> How many CPUs were configured? When I last tested this a few years ago >>>>>>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>>>>>> latencies), it was a net loss. >>>>>>>> >>>>>>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>>>>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>>>>>> 96 cores, and dual numa nodes) too, the performance can be >>>>>>>> recovered basically if per numa-node counter is introduced and >>>>>>>> used in this case, but the patch was never posted out. >>>>>>>> If anyone is interested in that, I can rebase the patch on current >>>>>>>> block tree and post out. I guess the performance issue might be >>>>>>>> related with system cache coherency implementation more or less. >>>>>>>> This issue on ARM64 can be observed with the following userspace >>>>>>>> atomic counting test too: >>>>>>>> >>>>>>>> http://kernel.ubuntu.com/~ming/test/cache/ >>>>>>> >>>>>>> How well did the per-node thing work? Doesn't seem to me like it would >>>>>>> go far enough. And per CPU is too much. One potential improvement would >>>>>>> be to change the part_stat_read() to just loop online CPUs, instead of >>>>>>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>>>>>> ensure the stats are sane. Often there's a huge difference between >>>>>>> NR_CPUS configured and what the system has. As Brian states, RH ships >>>>>>> with 2048, while I doubt a lot of customers actually run that... >>>>>>> >>>>>>> Outside of coming up with a more clever data structure that is fully >>>>>>> CPU topology aware, one thing that could work is just having X cache >>>>>>> line separated read/write inflight counters per node, where X is some >>>>>>> suitable value (like 4). That prevents us from having cross node >>>>>>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>>>>>> provide a nice balance between cost of incrementing the inflight >>>>>>> counting, and the cost of looping for reading it. >>>>>>> >>>>>>> And that brings me to the next part... >>>>>>> >>>>>>>>> I do agree that we should do something about it, and it's one of those >>>>>>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>>>>>> up. It's just not great as it currently stands, but I don't think per >>>>>>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>>>>>> counter. >>>>>>>> >>>>>>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>>>>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>>>>>> 'in_flight' counter. I thought about it before, but never got a >>>>>>>> perfect solution, and looks it is a bit hard, :-) >>>>>>> >>>>>>> The tags are already a bit spread out, so it's worth a shot. That would >>>>>>> remove the need to do anything in the inc/dec path, as the tags already >>>>>>> do that. The inlight count could be easily retrieved with >>>>>>> sbitmap_weight(). The only issue here is that we need separate read and >>>>>>> write counters, and the weight would obviously only get us the total >>>>>>> count. But we can have a slower path for that, just iterate the tags and >>>>>>> count them. The fast path only cares about total count. >>>>>>> >>>>>>> Let me try that out real quick. >>>>>> >>>>>> Well, that only works for whole disk stats, of course... There's no way >>>>>> around iterating the tags and checking for this to truly work. >>>>> >>>>> Totally untested proof of concept for using the tags for this. I based >>>>> this on top of Brian's patch, so it includes his patch plus the >>>>> _double() stuff I did which is no longer really needed. >>>>> >>>>> >>>>> diff --git a/block/bio.c b/block/bio.c >>>>> index 9cf98b29588a..ec99d9ba0f33 100644 >>>>> --- a/block/bio.c >>>>> +++ b/block/bio.c >>>>> @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, >>>>> part_round_stats(cpu, part); >>>>> part_stat_inc(cpu, part, ios[rw]); >>>>> part_stat_add(cpu, part, sectors[rw], sectors); >>>>> - part_inc_in_flight(part, rw); >>>>> + part_inc_in_flight(cpu, part, rw); >>>>> >>>>> part_stat_unlock(); >>>>> } >>>>> @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, >>>>> >>>>> part_stat_add(cpu, part, ticks[rw], duration); >>>>> part_round_stats(cpu, part); >>>>> - part_dec_in_flight(part, rw); >>>>> + part_dec_in_flight(cpu, part, rw); >>>>> >>>>> part_stat_unlock(); >>>>> } >>>>> diff --git a/block/blk-core.c b/block/blk-core.c >>>>> index af393d5a9680..6ab2efbe940b 100644 >>>>> --- a/block/blk-core.c >>>>> +++ b/block/blk-core.c >>>>> @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) >>>>> >>>>> part_stat_inc(cpu, part, ios[rw]); >>>>> part_stat_add(cpu, part, ticks[rw], duration); >>>>> - part_round_stats(cpu, part); >>>>> - part_dec_in_flight(part, rw); >>>>> + >>>>> + if (req->q->mq_ops) >>>>> + part_round_stats_mq(req->q, cpu, part); >>>>> + else { >>>>> + part_round_stats(cpu, part); >>>>> + part_dec_in_flight(cpu, part, rw); >>>>> + } >>>>> >>>>> hd_struct_put(part); >>>>> part_stat_unlock(); >>>>> @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) >>>>> part = &rq->rq_disk->part0; >>>>> hd_struct_get(part); >>>>> } >>>>> - part_round_stats(cpu, part); >>>>> - part_inc_in_flight(part, rw); >>>>> + if (rq->q->mq_ops) >>>>> + part_round_stats_mq(rq->q, cpu, part); >>>>> + else { >>>>> + part_round_stats(cpu, part); >>>>> + part_inc_in_flight(cpu, part, rw); >>>>> + } >>>>> rq->part = part; >>>>> } >>>>> >>>>> diff --git a/block/blk-merge.c b/block/blk-merge.c >>>>> index 99038830fb42..3b5eb2d4b964 100644 >>>>> --- a/block/blk-merge.c >>>>> +++ b/block/blk-merge.c >>>>> @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) >>>>> part = req->part; >>>>> >>>>> part_round_stats(cpu, part); >>>>> - part_dec_in_flight(part, rq_data_dir(req)); >>>>> + part_dec_in_flight(cpu, part, rq_data_dir(req)); >>>>> >>>>> hd_struct_put(part); >>>>> part_stat_unlock(); >>>>> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c >>>>> index d0be72ccb091..a7b897740c47 100644 >>>>> --- a/block/blk-mq-tag.c >>>>> +++ b/block/blk-mq-tag.c >>>>> @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) >>>>> bitnr += tags->nr_reserved_tags; >>>>> rq = tags->rqs[bitnr]; >>>>> >>>>> - if (rq->q == hctx->queue) >>>>> + if (rq && rq->q == hctx->queue) >>>>> iter_data->fn(hctx, rq, iter_data->data, reserved); >>>>> return true; >>>>> } >>>>> diff --git a/block/blk-mq.c b/block/blk-mq.c >>>>> index 05dfa3f270ae..cad4d2c26285 100644 >>>>> --- a/block/blk-mq.c >>>>> +++ b/block/blk-mq.c >>>>> @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); >>>>> static void blk_mq_poll_stats_start(struct request_queue *q); >>>>> static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); >>>>> >>>>> +struct mq_inflight { >>>>> + struct hd_struct *part; >>>>> + unsigned int inflight; >>>>> +}; >>>>> + >>>>> +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, >>>>> + struct request *rq, void *priv, >>>>> + bool reserved) >>>>> +{ >>>>> + struct mq_inflight *mi = priv; >>>>> + >>>>> + if (rq->part == mi->part && >>>>> + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) >>>>> + mi->inflight++; >>>>> +} >>>>> + >>>>> +unsigned long part_in_flight_mq(struct request_queue *q, >>>>> + struct hd_struct *part) >>>>> +{ >>>>> + struct mq_inflight mi = { .part = part, .inflight = 0 }; >>>>> + >>>>> + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); >>>>> + return mi.inflight; >>>>> +} >>>> >>>> Compared with the totally percpu approach, this way might help 1:M or >>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>> each CPU(especially there are huge hw queues on a big system), :-( >>> >>> Not disagreeing with that, without having some mechanism to only >>> loop queues that have pending requests. That would be similar to the >>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>> doing, I like your pnode approach better. However, I'm still not fully >>> convinced that one per node is enough to get the scalability we need. >>> >>> Would be great if Brian could re-test with your updated patch, so we >>> know how it works for him at least. >> >> I'll try running with both approaches today and see how they compare. > > Focus on Ming's, a variant of that is the most likely path forward, > imho. It'd be great to do a quick run on mine as well, just to establish > how it compares to mainline, though. Here's a cleaned up series: http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight (it's against mainline for now, I will update it to be against for-4.13/block in a rebase). One optimization on top of this I want to do is to only iterate once, even for a partition - pass in both parts, and increment two different counts. If we collapse the two part time stamps, then that's doable, and it means we only have to iterate once. Essentially this series makes the inc/dec a noop, since we don't have to do anything. The reading is basically no worse than a cpu online iteration, since we never have more queues than online CPUs. That's an improvement over per-cpu for-each-possible loops. For a lot of cases, it's much less, since we have fewer queues than CPUs. I'll need an hour or two to hone this a bit more, but then it would be great if you can retest. I'll send out an email when that's done, it'll be some time over this weekend.
On 06/30/2017 10:17 PM, Jens Axboe wrote: > On 06/30/2017 08:08 AM, Jens Axboe wrote: >> On 06/30/2017 07:05 AM, Brian King wrote: >>> On 06/29/2017 09:17 PM, Jens Axboe wrote: >>>> On 06/29/2017 07:20 PM, Ming Lei wrote: >>>>> On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>> On 06/29/2017 10:00 AM, Jens Axboe wrote: >>>>>>> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>>>>>>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>>>>>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>>>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>>>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>>>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>>>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>>>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>>>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>>>>>>> >>>>>>>>>> This has been done before, but I've never really liked it. The reason is >>>>>>>>>> that it means that reading the part stat inflight count now has to >>>>>>>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>>>>>>> How many CPUs were configured? When I last tested this a few years ago >>>>>>>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>>>>>>> latencies), it was a net loss. >>>>>>>>> >>>>>>>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>>>>>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>>>>>>> 96 cores, and dual numa nodes) too, the performance can be >>>>>>>>> recovered basically if per numa-node counter is introduced and >>>>>>>>> used in this case, but the patch was never posted out. >>>>>>>>> If anyone is interested in that, I can rebase the patch on current >>>>>>>>> block tree and post out. I guess the performance issue might be >>>>>>>>> related with system cache coherency implementation more or less. >>>>>>>>> This issue on ARM64 can be observed with the following userspace >>>>>>>>> atomic counting test too: >>>>>>>>> >>>>>>>>> http://kernel.ubuntu.com/~ming/test/cache/ >>>>>>>> >>>>>>>> How well did the per-node thing work? Doesn't seem to me like it would >>>>>>>> go far enough. And per CPU is too much. One potential improvement would >>>>>>>> be to change the part_stat_read() to just loop online CPUs, instead of >>>>>>>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>>>>>>> ensure the stats are sane. Often there's a huge difference between >>>>>>>> NR_CPUS configured and what the system has. As Brian states, RH ships >>>>>>>> with 2048, while I doubt a lot of customers actually run that... >>>>>>>> >>>>>>>> Outside of coming up with a more clever data structure that is fully >>>>>>>> CPU topology aware, one thing that could work is just having X cache >>>>>>>> line separated read/write inflight counters per node, where X is some >>>>>>>> suitable value (like 4). That prevents us from having cross node >>>>>>>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>>>>>>> provide a nice balance between cost of incrementing the inflight >>>>>>>> counting, and the cost of looping for reading it. >>>>>>>> >>>>>>>> And that brings me to the next part... >>>>>>>> >>>>>>>>>> I do agree that we should do something about it, and it's one of those >>>>>>>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>>>>>>> up. It's just not great as it currently stands, but I don't think per >>>>>>>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>>>>>>> counter. >>>>>>>>> >>>>>>>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>>>>>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>>>>>>> 'in_flight' counter. I thought about it before, but never got a >>>>>>>>> perfect solution, and looks it is a bit hard, :-) >>>>>>>> >>>>>>>> The tags are already a bit spread out, so it's worth a shot. That would >>>>>>>> remove the need to do anything in the inc/dec path, as the tags already >>>>>>>> do that. The inlight count could be easily retrieved with >>>>>>>> sbitmap_weight(). The only issue here is that we need separate read and >>>>>>>> write counters, and the weight would obviously only get us the total >>>>>>>> count. But we can have a slower path for that, just iterate the tags and >>>>>>>> count them. The fast path only cares about total count. >>>>>>>> >>>>>>>> Let me try that out real quick. >>>>>>> >>>>>>> Well, that only works for whole disk stats, of course... There's no way >>>>>>> around iterating the tags and checking for this to truly work. >>>>>> >>>>>> Totally untested proof of concept for using the tags for this. I based >>>>>> this on top of Brian's patch, so it includes his patch plus the >>>>>> _double() stuff I did which is no longer really needed. >>>>>> >>>>>> >>>>>> diff --git a/block/bio.c b/block/bio.c >>>>>> index 9cf98b29588a..ec99d9ba0f33 100644 >>>>>> --- a/block/bio.c >>>>>> +++ b/block/bio.c >>>>>> @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, >>>>>> part_round_stats(cpu, part); >>>>>> part_stat_inc(cpu, part, ios[rw]); >>>>>> part_stat_add(cpu, part, sectors[rw], sectors); >>>>>> - part_inc_in_flight(part, rw); >>>>>> + part_inc_in_flight(cpu, part, rw); >>>>>> >>>>>> part_stat_unlock(); >>>>>> } >>>>>> @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, >>>>>> >>>>>> part_stat_add(cpu, part, ticks[rw], duration); >>>>>> part_round_stats(cpu, part); >>>>>> - part_dec_in_flight(part, rw); >>>>>> + part_dec_in_flight(cpu, part, rw); >>>>>> >>>>>> part_stat_unlock(); >>>>>> } >>>>>> diff --git a/block/blk-core.c b/block/blk-core.c >>>>>> index af393d5a9680..6ab2efbe940b 100644 >>>>>> --- a/block/blk-core.c >>>>>> +++ b/block/blk-core.c >>>>>> @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) >>>>>> >>>>>> part_stat_inc(cpu, part, ios[rw]); >>>>>> part_stat_add(cpu, part, ticks[rw], duration); >>>>>> - part_round_stats(cpu, part); >>>>>> - part_dec_in_flight(part, rw); >>>>>> + >>>>>> + if (req->q->mq_ops) >>>>>> + part_round_stats_mq(req->q, cpu, part); >>>>>> + else { >>>>>> + part_round_stats(cpu, part); >>>>>> + part_dec_in_flight(cpu, part, rw); >>>>>> + } >>>>>> >>>>>> hd_struct_put(part); >>>>>> part_stat_unlock(); >>>>>> @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) >>>>>> part = &rq->rq_disk->part0; >>>>>> hd_struct_get(part); >>>>>> } >>>>>> - part_round_stats(cpu, part); >>>>>> - part_inc_in_flight(part, rw); >>>>>> + if (rq->q->mq_ops) >>>>>> + part_round_stats_mq(rq->q, cpu, part); >>>>>> + else { >>>>>> + part_round_stats(cpu, part); >>>>>> + part_inc_in_flight(cpu, part, rw); >>>>>> + } >>>>>> rq->part = part; >>>>>> } >>>>>> >>>>>> diff --git a/block/blk-merge.c b/block/blk-merge.c >>>>>> index 99038830fb42..3b5eb2d4b964 100644 >>>>>> --- a/block/blk-merge.c >>>>>> +++ b/block/blk-merge.c >>>>>> @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) >>>>>> part = req->part; >>>>>> >>>>>> part_round_stats(cpu, part); >>>>>> - part_dec_in_flight(part, rq_data_dir(req)); >>>>>> + part_dec_in_flight(cpu, part, rq_data_dir(req)); >>>>>> >>>>>> hd_struct_put(part); >>>>>> part_stat_unlock(); >>>>>> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c >>>>>> index d0be72ccb091..a7b897740c47 100644 >>>>>> --- a/block/blk-mq-tag.c >>>>>> +++ b/block/blk-mq-tag.c >>>>>> @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) >>>>>> bitnr += tags->nr_reserved_tags; >>>>>> rq = tags->rqs[bitnr]; >>>>>> >>>>>> - if (rq->q == hctx->queue) >>>>>> + if (rq && rq->q == hctx->queue) >>>>>> iter_data->fn(hctx, rq, iter_data->data, reserved); >>>>>> return true; >>>>>> } >>>>>> diff --git a/block/blk-mq.c b/block/blk-mq.c >>>>>> index 05dfa3f270ae..cad4d2c26285 100644 >>>>>> --- a/block/blk-mq.c >>>>>> +++ b/block/blk-mq.c >>>>>> @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); >>>>>> static void blk_mq_poll_stats_start(struct request_queue *q); >>>>>> static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); >>>>>> >>>>>> +struct mq_inflight { >>>>>> + struct hd_struct *part; >>>>>> + unsigned int inflight; >>>>>> +}; >>>>>> + >>>>>> +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, >>>>>> + struct request *rq, void *priv, >>>>>> + bool reserved) >>>>>> +{ >>>>>> + struct mq_inflight *mi = priv; >>>>>> + >>>>>> + if (rq->part == mi->part && >>>>>> + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) >>>>>> + mi->inflight++; >>>>>> +} >>>>>> + >>>>>> +unsigned long part_in_flight_mq(struct request_queue *q, >>>>>> + struct hd_struct *part) >>>>>> +{ >>>>>> + struct mq_inflight mi = { .part = part, .inflight = 0 }; >>>>>> + >>>>>> + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); >>>>>> + return mi.inflight; >>>>>> +} >>>>> >>>>> Compared with the totally percpu approach, this way might help 1:M or >>>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>>> each CPU(especially there are huge hw queues on a big system), :-( >>>> >>>> Not disagreeing with that, without having some mechanism to only >>>> loop queues that have pending requests. That would be similar to the >>>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>>> doing, I like your pnode approach better. However, I'm still not fully >>>> convinced that one per node is enough to get the scalability we need. >>>> >>>> Would be great if Brian could re-test with your updated patch, so we >>>> know how it works for him at least. >>> >>> I'll try running with both approaches today and see how they compare. >> >> Focus on Ming's, a variant of that is the most likely path forward, >> imho. It'd be great to do a quick run on mine as well, just to establish >> how it compares to mainline, though. > > Here's a cleaned up series: > > http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight > > (it's against mainline for now, I will update it to be against > for-4.13/block in a rebase). > > One optimization on top of this I want to do is to only iterate once, > even for a partition - pass in both parts, and increment two different > counts. If we collapse the two part time stamps, then that's doable, and > it means we only have to iterate once. > > Essentially this series makes the inc/dec a noop, since we don't have to > do anything. The reading is basically no worse than a cpu online > iteration, since we never have more queues than online CPUs. That's an > improvement over per-cpu for-each-possible loops. For a lot of cases, > it's much less, since we have fewer queues than CPUs. I'll need an hour > or two to hone this a bit more, but then it would be great if you can > retest. I'll send out an email when that's done, it'll be some time over > this weekend. Did the double-read with one iteration change, it was pretty trivial: http://git.kernel.dk/cgit/linux-block/commit/?h=mq-inflight&id=b841804f826df072f706ae86d0eb533342f0297a And updated the branch here: http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight to include that, and be based on top of for-4.13/block. If you prefer just pulling a branch, pull: git://git.kernel.dk/linux-block mq-inflight
On 06/30/2017 10:59 PM, Jens Axboe wrote: > On 06/30/2017 10:17 PM, Jens Axboe wrote: >> On 06/30/2017 08:08 AM, Jens Axboe wrote: >>> On 06/30/2017 07:05 AM, Brian King wrote: >>>> On 06/29/2017 09:17 PM, Jens Axboe wrote: >>>>> On 06/29/2017 07:20 PM, Ming Lei wrote: >>>>>> On Fri, Jun 30, 2017 at 2:42 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>>> On 06/29/2017 10:00 AM, Jens Axboe wrote: >>>>>>>> On 06/29/2017 09:58 AM, Jens Axboe wrote: >>>>>>>>> On 06/29/2017 02:40 AM, Ming Lei wrote: >>>>>>>>>> On Thu, Jun 29, 2017 at 5:49 AM, Jens Axboe <axboe@kernel.dk> wrote: >>>>>>>>>>> On 06/28/2017 03:12 PM, Brian King wrote: >>>>>>>>>>>> This patch converts the in_flight counter in struct hd_struct from a >>>>>>>>>>>> pair of atomics to a pair of percpu counters. This eliminates a couple >>>>>>>>>>>> of atomics from the hot path. When running this on a Power system, to >>>>>>>>>>>> a single null_blk device with 80 submission queues, irq mode 0, with >>>>>>>>>>>> 80 fio jobs, I saw IOPs go from 1.5M IO/s to 11.4 IO/s. >>>>>>>>>>> >>>>>>>>>>> This has been done before, but I've never really liked it. The reason is >>>>>>>>>>> that it means that reading the part stat inflight count now has to >>>>>>>>>>> iterate over every possible CPU. Did you use partitions in your testing? >>>>>>>>>>> How many CPUs were configured? When I last tested this a few years ago >>>>>>>>>>> on even a quad core nehalem (which is notoriously shitty for cross-node >>>>>>>>>>> latencies), it was a net loss. >>>>>>>>>> >>>>>>>>>> One year ago, I saw null_blk's IOPS can be decreased to 10% >>>>>>>>>> of non-RQF_IO_STAT on a dual socket ARM64(each CPU has >>>>>>>>>> 96 cores, and dual numa nodes) too, the performance can be >>>>>>>>>> recovered basically if per numa-node counter is introduced and >>>>>>>>>> used in this case, but the patch was never posted out. >>>>>>>>>> If anyone is interested in that, I can rebase the patch on current >>>>>>>>>> block tree and post out. I guess the performance issue might be >>>>>>>>>> related with system cache coherency implementation more or less. >>>>>>>>>> This issue on ARM64 can be observed with the following userspace >>>>>>>>>> atomic counting test too: >>>>>>>>>> >>>>>>>>>> http://kernel.ubuntu.com/~ming/test/cache/ >>>>>>>>> >>>>>>>>> How well did the per-node thing work? Doesn't seem to me like it would >>>>>>>>> go far enough. And per CPU is too much. One potential improvement would >>>>>>>>> be to change the part_stat_read() to just loop online CPUs, instead of >>>>>>>>> all possible CPUs. When CPUs go on/offline, use that as the slow path to >>>>>>>>> ensure the stats are sane. Often there's a huge difference between >>>>>>>>> NR_CPUS configured and what the system has. As Brian states, RH ships >>>>>>>>> with 2048, while I doubt a lot of customers actually run that... >>>>>>>>> >>>>>>>>> Outside of coming up with a more clever data structure that is fully >>>>>>>>> CPU topology aware, one thing that could work is just having X cache >>>>>>>>> line separated read/write inflight counters per node, where X is some >>>>>>>>> suitable value (like 4). That prevents us from having cross node >>>>>>>>> traffic, and it also keeps the cross cpu traffic fairly low. That should >>>>>>>>> provide a nice balance between cost of incrementing the inflight >>>>>>>>> counting, and the cost of looping for reading it. >>>>>>>>> >>>>>>>>> And that brings me to the next part... >>>>>>>>> >>>>>>>>>>> I do agree that we should do something about it, and it's one of those >>>>>>>>>>> items I've highlighted in talks about blk-mq on pending issues to fix >>>>>>>>>>> up. It's just not great as it currently stands, but I don't think per >>>>>>>>>>> CPU counters is the right way to fix it, at least not for the inflight >>>>>>>>>>> counter. >>>>>>>>>> >>>>>>>>>> Yeah, it won't be a issue for non-mq path, and for blk-mq path, maybe >>>>>>>>>> we can use some blk-mq knowledge(tagset?) to figure out the >>>>>>>>>> 'in_flight' counter. I thought about it before, but never got a >>>>>>>>>> perfect solution, and looks it is a bit hard, :-) >>>>>>>>> >>>>>>>>> The tags are already a bit spread out, so it's worth a shot. That would >>>>>>>>> remove the need to do anything in the inc/dec path, as the tags already >>>>>>>>> do that. The inlight count could be easily retrieved with >>>>>>>>> sbitmap_weight(). The only issue here is that we need separate read and >>>>>>>>> write counters, and the weight would obviously only get us the total >>>>>>>>> count. But we can have a slower path for that, just iterate the tags and >>>>>>>>> count them. The fast path only cares about total count. >>>>>>>>> >>>>>>>>> Let me try that out real quick. >>>>>>>> >>>>>>>> Well, that only works for whole disk stats, of course... There's no way >>>>>>>> around iterating the tags and checking for this to truly work. >>>>>>> >>>>>>> Totally untested proof of concept for using the tags for this. I based >>>>>>> this on top of Brian's patch, so it includes his patch plus the >>>>>>> _double() stuff I did which is no longer really needed. >>>>>>> >>>>>>> >>>>>>> diff --git a/block/bio.c b/block/bio.c >>>>>>> index 9cf98b29588a..ec99d9ba0f33 100644 >>>>>>> --- a/block/bio.c >>>>>>> +++ b/block/bio.c >>>>>>> @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, >>>>>>> part_round_stats(cpu, part); >>>>>>> part_stat_inc(cpu, part, ios[rw]); >>>>>>> part_stat_add(cpu, part, sectors[rw], sectors); >>>>>>> - part_inc_in_flight(part, rw); >>>>>>> + part_inc_in_flight(cpu, part, rw); >>>>>>> >>>>>>> part_stat_unlock(); >>>>>>> } >>>>>>> @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, >>>>>>> >>>>>>> part_stat_add(cpu, part, ticks[rw], duration); >>>>>>> part_round_stats(cpu, part); >>>>>>> - part_dec_in_flight(part, rw); >>>>>>> + part_dec_in_flight(cpu, part, rw); >>>>>>> >>>>>>> part_stat_unlock(); >>>>>>> } >>>>>>> diff --git a/block/blk-core.c b/block/blk-core.c >>>>>>> index af393d5a9680..6ab2efbe940b 100644 >>>>>>> --- a/block/blk-core.c >>>>>>> +++ b/block/blk-core.c >>>>>>> @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) >>>>>>> >>>>>>> part_stat_inc(cpu, part, ios[rw]); >>>>>>> part_stat_add(cpu, part, ticks[rw], duration); >>>>>>> - part_round_stats(cpu, part); >>>>>>> - part_dec_in_flight(part, rw); >>>>>>> + >>>>>>> + if (req->q->mq_ops) >>>>>>> + part_round_stats_mq(req->q, cpu, part); >>>>>>> + else { >>>>>>> + part_round_stats(cpu, part); >>>>>>> + part_dec_in_flight(cpu, part, rw); >>>>>>> + } >>>>>>> >>>>>>> hd_struct_put(part); >>>>>>> part_stat_unlock(); >>>>>>> @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) >>>>>>> part = &rq->rq_disk->part0; >>>>>>> hd_struct_get(part); >>>>>>> } >>>>>>> - part_round_stats(cpu, part); >>>>>>> - part_inc_in_flight(part, rw); >>>>>>> + if (rq->q->mq_ops) >>>>>>> + part_round_stats_mq(rq->q, cpu, part); >>>>>>> + else { >>>>>>> + part_round_stats(cpu, part); >>>>>>> + part_inc_in_flight(cpu, part, rw); >>>>>>> + } >>>>>>> rq->part = part; >>>>>>> } >>>>>>> >>>>>>> diff --git a/block/blk-merge.c b/block/blk-merge.c >>>>>>> index 99038830fb42..3b5eb2d4b964 100644 >>>>>>> --- a/block/blk-merge.c >>>>>>> +++ b/block/blk-merge.c >>>>>>> @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) >>>>>>> part = req->part; >>>>>>> >>>>>>> part_round_stats(cpu, part); >>>>>>> - part_dec_in_flight(part, rq_data_dir(req)); >>>>>>> + part_dec_in_flight(cpu, part, rq_data_dir(req)); >>>>>>> >>>>>>> hd_struct_put(part); >>>>>>> part_stat_unlock(); >>>>>>> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c >>>>>>> index d0be72ccb091..a7b897740c47 100644 >>>>>>> --- a/block/blk-mq-tag.c >>>>>>> +++ b/block/blk-mq-tag.c >>>>>>> @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) >>>>>>> bitnr += tags->nr_reserved_tags; >>>>>>> rq = tags->rqs[bitnr]; >>>>>>> >>>>>>> - if (rq->q == hctx->queue) >>>>>>> + if (rq && rq->q == hctx->queue) >>>>>>> iter_data->fn(hctx, rq, iter_data->data, reserved); >>>>>>> return true; >>>>>>> } >>>>>>> diff --git a/block/blk-mq.c b/block/blk-mq.c >>>>>>> index 05dfa3f270ae..cad4d2c26285 100644 >>>>>>> --- a/block/blk-mq.c >>>>>>> +++ b/block/blk-mq.c >>>>>>> @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); >>>>>>> static void blk_mq_poll_stats_start(struct request_queue *q); >>>>>>> static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); >>>>>>> >>>>>>> +struct mq_inflight { >>>>>>> + struct hd_struct *part; >>>>>>> + unsigned int inflight; >>>>>>> +}; >>>>>>> + >>>>>>> +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, >>>>>>> + struct request *rq, void *priv, >>>>>>> + bool reserved) >>>>>>> +{ >>>>>>> + struct mq_inflight *mi = priv; >>>>>>> + >>>>>>> + if (rq->part == mi->part && >>>>>>> + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) >>>>>>> + mi->inflight++; >>>>>>> +} >>>>>>> + >>>>>>> +unsigned long part_in_flight_mq(struct request_queue *q, >>>>>>> + struct hd_struct *part) >>>>>>> +{ >>>>>>> + struct mq_inflight mi = { .part = part, .inflight = 0 }; >>>>>>> + >>>>>>> + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); >>>>>>> + return mi.inflight; >>>>>>> +} >>>>>> >>>>>> Compared with the totally percpu approach, this way might help 1:M or >>>>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>>>> each CPU(especially there are huge hw queues on a big system), :-( >>>>> >>>>> Not disagreeing with that, without having some mechanism to only >>>>> loop queues that have pending requests. That would be similar to the >>>>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>>>> doing, I like your pnode approach better. However, I'm still not fully >>>>> convinced that one per node is enough to get the scalability we need. >>>>> >>>>> Would be great if Brian could re-test with your updated patch, so we >>>>> know how it works for him at least. >>>> >>>> I'll try running with both approaches today and see how they compare. >>> >>> Focus on Ming's, a variant of that is the most likely path forward, >>> imho. It'd be great to do a quick run on mine as well, just to establish >>> how it compares to mainline, though. >> >> Here's a cleaned up series: >> >> http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight >> >> (it's against mainline for now, I will update it to be against >> for-4.13/block in a rebase). >> >> One optimization on top of this I want to do is to only iterate once, >> even for a partition - pass in both parts, and increment two different >> counts. If we collapse the two part time stamps, then that's doable, and >> it means we only have to iterate once. >> >> Essentially this series makes the inc/dec a noop, since we don't have to >> do anything. The reading is basically no worse than a cpu online >> iteration, since we never have more queues than online CPUs. That's an >> improvement over per-cpu for-each-possible loops. For a lot of cases, >> it's much less, since we have fewer queues than CPUs. I'll need an hour >> or two to hone this a bit more, but then it would be great if you can >> retest. I'll send out an email when that's done, it'll be some time over >> this weekend. > > Did the double-read with one iteration change, it was pretty trivial: > > http://git.kernel.dk/cgit/linux-block/commit/?h=mq-inflight&id=b841804f826df072f706ae86d0eb533342f0297a Now: http://git.kernel.dk/cgit/linux-block/commit/?h=mq-inflight&id=87f73ef2b9edb6834001df8f7cb48c7a116e8cd3 > And updated the branch here: > > http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight > > to include that, and be based on top of for-4.13/block. If you prefer just > pulling a branch, pull: > > git://git.kernel.dk/linux-block mq-inflight I've tested it, and made a few adjustments and fixes. The branches are the same, just rebased. Works fine for me, even with partitions now. Let me know how it works for you.
On Sat, Jul 1, 2017 at 10:18 AM, Brian King <brking@linux.vnet.ibm.com> wrote: > On 06/30/2017 06:26 PM, Jens Axboe wrote: >> On 06/30/2017 05:23 PM, Ming Lei wrote: >>> Hi Bian, >>> >>> On Sat, Jul 1, 2017 at 2:33 AM, Brian King <brking@linux.vnet.ibm.com> wrote: >>>> On 06/30/2017 09:08 AM, Jens Axboe wrote: >>>>>>>> Compared with the totally percpu approach, this way might help 1:M or >>>>>>>> N:M mapping, but won't help 1:1 map(NVMe), when hctx is mapped to >>>>>>>> each CPU(especially there are huge hw queues on a big system), :-( >>>>>>> >>>>>>> Not disagreeing with that, without having some mechanism to only >>>>>>> loop queues that have pending requests. That would be similar to the >>>>>>> ctx_map for sw to hw queues. But I don't think that would be worthwhile >>>>>>> doing, I like your pnode approach better. However, I'm still not fully >>>>>>> convinced that one per node is enough to get the scalability we need. >>>>>>> >>>>>>> Would be great if Brian could re-test with your updated patch, so we >>>>>>> know how it works for him at least. >>>>>> >>>>>> I'll try running with both approaches today and see how they compare. >>>>> >>>>> Focus on Ming's, a variant of that is the most likely path forward, >>>>> imho. It'd be great to do a quick run on mine as well, just to establish >>>>> how it compares to mainline, though. >>>> >>>> On my initial runs, the one from you Jens, appears to perform a bit better, although >>>> both are a huge improvement from what I was seeing before. >>>> >>>> I ran 4k random reads using fio to nullblk in two configurations on my 20 core >>>> system with 4 NUMA nodes and 4-way SMT, so 80 logical CPUs. I ran both 80 threads >>>> to a single null_blk as well as 80 threads to 80 null_block devices, so one thread >>> >>> Could you share what the '80 null_block devices' is? It means you >>> create 80 null_blk >>> devices? Or you create one null_blk and make its hw queue number as 80 >>> via module >>> parameter of ''submit_queues"? >> >> That's a valid question, was going to ask that too. But I assumed that Brian >> used submit_queues to set as many queues as he has logical CPUs in the system. >>> >>> I guess we should focus on multi-queue case since it is the normal way of NVMe. >>> >>>> per null_blk. This is what I saw on this machine: >>>> >>>> Using the Per node atomic change from Ming Lei >>>> 1 null_blk, 80 threads >>>> iops=9376.5K >>>> >>>> 80 null_blk, 1 thread >>>> iops=9523.5K >>>> >>>> >>>> Using the alternate patch from Jens using the tags >>>> 1 null_blk, 80 threads >>>> iops=9725.8K >>>> >>>> 80 null_blk, 1 thread >>>> iops=9569.4K >>> >>> If 1 thread means single fio job, looks the number is too too high, that means >>> one random IO can complete in about 0.1us(100ns) on single CPU, not sure if it >>> is possible, :-) >> >> It means either 1 null_blk device, 80 threads running IO to it. Or 80 null_blk >> devices, each with a thread running IO to it. See above, he details that it's >> 80 threads on 80 devices for that case. > > Right. So the two modes I'm running in are: > > 1. 80 null_blk devices, each with one submit_queue, with one fio job per null_blk device, > so 80 threads total. 80 logical CPUs > 2. 1 null_blk device, with 80 submit_queues, 80 fio jobs, 80 logical CPUs. > > In theory, the two should result in similar numbers. > > Here are the commands and fio configurations: > > Scenario #1 > modprobe null_blk submit_queues=80 nr_devices=1 irqmode=0 > > FIO config: > [global] > buffered=0 > invalidate=1 > bs=4k > iodepth=64 > numjobs=80 > group_reporting=1 > rw=randrw > rwmixread=100 > rwmixwrite=0 > ioengine=libaio > runtime=60 > time_based > > [job1] > filename=/dev/nullb0 > > > Scenario #2 > modprobe null_blk submit_queues=1 nr_devices=80 irqmode=0 > > FIO config > [global] > buffered=0 > invalidate=1 > bs=4k > iodepth=64 > numjobs=1 > group_reporting=1 > rw=randrw > rwmixread=100 > rwmixwrite=0 > ioengine=libaio > runtime=60 > time_based > > [job1] > filename=/dev/nullb0 > [job2] > filename=/dev/nullb1 > [job3] > filename=/dev/nullb2 > [job4] > filename=/dev/nullb3 > [job5] > filename=/dev/nullb4 > [job6] > filename=/dev/nullb5 > [job7] > filename=/dev/nullb6 > [job8] > filename=/dev/nullb7 > [job9] > filename=/dev/nullb8 > [job10] > filename=/dev/nullb9 > [job11] > filename=/dev/nullb10 > [job12] > filename=/dev/nullb11 > [job13] > filename=/dev/nullb12 > [job14] > filename=/dev/nullb13 > [job15] > filename=/dev/nullb14 > [job16] > filename=/dev/nullb15 > [job17] > filename=/dev/nullb16 > [job18] > filename=/dev/nullb17 > [job19] > filename=/dev/nullb18 > [job20] > filename=/dev/nullb19 > [job21] > filename=/dev/nullb20 > [job22] > filename=/dev/nullb21 > [job23] > filename=/dev/nullb22 > [job24] > filename=/dev/nullb23 > [job25] > filename=/dev/nullb24 > [job26] > filename=/dev/nullb25 > [job27] > filename=/dev/nullb26 > [job28] > filename=/dev/nullb27 > [job29] > filename=/dev/nullb28 > [job30] > filename=/dev/nullb29 > [job31] > filename=/dev/nullb30 > [job32] > filename=/dev/nullb31 > [job33] > filename=/dev/nullb32 > [job34] > filename=/dev/nullb33 > [job35] > filename=/dev/nullb34 > [job36] > filename=/dev/nullb35 > [job37] > filename=/dev/nullb36 > [job38] > filename=/dev/nullb37 > [job39] > filename=/dev/nullb38 > [job40] > filename=/dev/nullb39 > [job41] > filename=/dev/nullb40 > [job42] > filename=/dev/nullb41 > [job43] > filename=/dev/nullb42 > [job44] > filename=/dev/nullb43 > [job45] > filename=/dev/nullb44 > [job46] > filename=/dev/nullb45 > [job47] > filename=/dev/nullb46 > [job48] > filename=/dev/nullb47 > [job49] > filename=/dev/nullb48 > [job50] > filename=/dev/nullb49 > [job51] > filename=/dev/nullb50 > [job52] > filename=/dev/nullb51 > [job53] > filename=/dev/nullb52 > [job54] > filename=/dev/nullb53 > [job55] > filename=/dev/nullb54 > [job56] > filename=/dev/nullb55 > [job57] > filename=/dev/nullb56 > [job58] > filename=/dev/nullb57 > [job59] > filename=/dev/nullb58 > [job60] > filename=/dev/nullb59 > [job61] > filename=/dev/nullb60 > [job62] > filename=/dev/nullb61 > [job63] > filename=/dev/nullb62 > [job64] > filename=/dev/nullb63 > [job65] > filename=/dev/nullb64 > [job66] > filename=/dev/nullb65 > [job67] > filename=/dev/nullb66 > [job68] > filename=/dev/nullb67 > [job69] > filename=/dev/nullb68 > [job70] > filename=/dev/nullb69 > [job71] > filename=/dev/nullb70 > [job72] > filename=/dev/nullb71 > [job73] > filename=/dev/nullb72 > [job74] > filename=/dev/nullb73 > [job75] > filename=/dev/nullb74 > [job76] > filename=/dev/nullb75 > [job77] > filename=/dev/nullb76 > [job78] > filename=/dev/nullb77 > [job79] > filename=/dev/nullb78 > [job80] > filename=/dev/nullb79 IMO it should be more reasonable to use single null_blk with 80 queues via setting submit_queues as 80 than simply 80 null_blks. So suggest to switch to test 80 queues in your future test. thanks, Ming Lei -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel
On 07/01/2017 11:43 AM, Jens Axboe wrote: > Now: > > http://git.kernel.dk/cgit/linux-block/commit/?h=mq-inflight&id=87f73ef2b9edb6834001df8f7cb48c7a116e8cd3 > >> And updated the branch here: >> >> http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight >> >> to include that, and be based on top of for-4.13/block. If you prefer just >> pulling a branch, pull: >> >> git://git.kernel.dk/linux-block mq-inflight > > I've tested it, and made a few adjustments and fixes. The branches are > the same, just rebased. Works fine for me, even with partitions now. > > Let me know how it works for you. This is looking very nice. This is what I'm seeing now: 1 null_blk device / 80 submit_queues / fio numjobs=80 9.4M IOPs 80 null_blk devices / 1 submit_queue per null_blk / fio numjobs=1 (1 per null_blk) 9.3M IOPs 1 null_blk device / 160 submit_queues / fio numjobs=160 9.8M IOPs 160 null_blk devices / 1 submit_queue per null_blk / fio numjobs=1 (1 per null_blk) 8.9M IOPs What is queued in mq-inflight looks to resolve the issue I've been seeing. Thanks!!! Brian
On 07/03/2017 08:20 PM, Ming Lei wrote: >> Right. So the two modes I'm running in are: >> >> 1. 80 null_blk devices, each with one submit_queue, with one fio job per null_blk device, >> so 80 threads total. 80 logical CPUs >> 2. 1 null_blk device, with 80 submit_queues, 80 fio jobs, 80 logical CPUs. >> >> In theory, the two should result in similar numbers. >> >> Here are the commands and fio configurations: >> >> Scenario #1 >> modprobe null_blk submit_queues=80 nr_devices=1 irqmode=0 >> >> FIO config: >> [global] >> buffered=0 >> invalidate=1 >> bs=4k >> iodepth=64 >> numjobs=80 >> group_reporting=1 >> rw=randrw >> rwmixread=100 >> rwmixwrite=0 >> ioengine=libaio >> runtime=60 >> time_based >> >> [job1] >> filename=/dev/nullb0 > > IMO it should be more reasonable to use single null_blk with 80 queues > via setting submit_queues as 80 than simply 80 null_blks. So suggest to switch > to test 80 queues in your future test. That should be what my Scenario #1 covers. Thanks, Brian
On 07/04/2017 02:55 PM, Brian King wrote: > On 07/01/2017 11:43 AM, Jens Axboe wrote: >> Now: >> >> http://git.kernel.dk/cgit/linux-block/commit/?h=mq-inflight&id=87f73ef2b9edb6834001df8f7cb48c7a116e8cd3 >> >>> And updated the branch here: >>> >>> http://git.kernel.dk/cgit/linux-block/log/?h=mq-inflight >>> >>> to include that, and be based on top of for-4.13/block. If you prefer just >>> pulling a branch, pull: >>> >>> git://git.kernel.dk/linux-block mq-inflight >> >> I've tested it, and made a few adjustments and fixes. The branches are >> the same, just rebased. Works fine for me, even with partitions now. >> >> Let me know how it works for you. > > This is looking very nice. This is what I'm seeing now: > > 1 null_blk device / 80 submit_queues / fio numjobs=80 > 9.4M IOPs > > 80 null_blk devices / 1 submit_queue per null_blk / fio numjobs=1 (1 per null_blk) > 9.3M IOPs > > 1 null_blk device / 160 submit_queues / fio numjobs=160 > 9.8M IOPs > > 160 null_blk devices / 1 submit_queue per null_blk / fio numjobs=1 (1 per null_blk) > 8.9M IOPs > > What is queued in mq-inflight looks to resolve the issue I've been seeing. Good to hear! I'll test some corner cases and post a "real" version for review. Thanks for all your testing.
diff --git a/block/bio.c b/block/bio.c index 9cf98b29588a..ec99d9ba0f33 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1737,7 +1737,7 @@ void generic_start_io_acct(int rw, unsigned long sectors, part_round_stats(cpu, part); part_stat_inc(cpu, part, ios[rw]); part_stat_add(cpu, part, sectors[rw], sectors); - part_inc_in_flight(part, rw); + part_inc_in_flight(cpu, part, rw); part_stat_unlock(); } @@ -1751,7 +1751,7 @@ void generic_end_io_acct(int rw, struct hd_struct *part, part_stat_add(cpu, part, ticks[rw], duration); part_round_stats(cpu, part); - part_dec_in_flight(part, rw); + part_dec_in_flight(cpu, part, rw); part_stat_unlock(); } diff --git a/block/blk-core.c b/block/blk-core.c index af393d5a9680..6ab2efbe940b 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -2434,8 +2434,13 @@ void blk_account_io_done(struct request *req) part_stat_inc(cpu, part, ios[rw]); part_stat_add(cpu, part, ticks[rw], duration); - part_round_stats(cpu, part); - part_dec_in_flight(part, rw); + + if (req->q->mq_ops) + part_round_stats_mq(req->q, cpu, part); + else { + part_round_stats(cpu, part); + part_dec_in_flight(cpu, part, rw); + } hd_struct_put(part); part_stat_unlock(); @@ -2492,8 +2497,12 @@ void blk_account_io_start(struct request *rq, bool new_io) part = &rq->rq_disk->part0; hd_struct_get(part); } - part_round_stats(cpu, part); - part_inc_in_flight(part, rw); + if (rq->q->mq_ops) + part_round_stats_mq(rq->q, cpu, part); + else { + part_round_stats(cpu, part); + part_inc_in_flight(cpu, part, rw); + } rq->part = part; } diff --git a/block/blk-merge.c b/block/blk-merge.c index 99038830fb42..3b5eb2d4b964 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -634,7 +634,7 @@ static void blk_account_io_merge(struct request *req) part = req->part; part_round_stats(cpu, part); - part_dec_in_flight(part, rq_data_dir(req)); + part_dec_in_flight(cpu, part, rq_data_dir(req)); hd_struct_put(part); part_stat_unlock(); diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c index d0be72ccb091..a7b897740c47 100644 --- a/block/blk-mq-tag.c +++ b/block/blk-mq-tag.c @@ -214,7 +214,7 @@ static bool bt_iter(struct sbitmap *bitmap, unsigned int bitnr, void *data) bitnr += tags->nr_reserved_tags; rq = tags->rqs[bitnr]; - if (rq->q == hctx->queue) + if (rq && rq->q == hctx->queue) iter_data->fn(hctx, rq, iter_data->data, reserved); return true; } diff --git a/block/blk-mq.c b/block/blk-mq.c index 05dfa3f270ae..cad4d2c26285 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -43,6 +43,58 @@ static LIST_HEAD(all_q_list); static void blk_mq_poll_stats_start(struct request_queue *q); static void blk_mq_poll_stats_fn(struct blk_stat_callback *cb); +struct mq_inflight { + struct hd_struct *part; + unsigned int inflight; +}; + +static void blk_mq_check_inflight(struct blk_mq_hw_ctx *hctx, + struct request *rq, void *priv, + bool reserved) +{ + struct mq_inflight *mi = priv; + + if (rq->part == mi->part && + test_bit(REQ_ATOM_STARTED, &rq->atomic_flags)) + mi->inflight++; +} + +unsigned long part_in_flight_mq(struct request_queue *q, + struct hd_struct *part) +{ + struct mq_inflight mi = { .part = part, .inflight = 0 }; + + blk_mq_queue_tag_busy_iter(q, blk_mq_check_inflight, &mi); + return mi.inflight; +} + +static void __part_round_stats_mq(struct request_queue *q, int cpu, + struct hd_struct *part, unsigned long now) +{ + unsigned long inflight; + + if (now == part->stamp) + return; + + inflight = part_in_flight_mq(q, part); + if (inflight) { + __part_stat_add(cpu, part, time_in_queue, + inflight * (now - part->stamp)); + __part_stat_add(cpu, part, io_ticks, (now - part->stamp)); + } + part->stamp = now; +} + +void part_round_stats_mq(struct request_queue *q, int cpu, + struct hd_struct *part) +{ + unsigned long now = jiffies; + + if (part->partno) + __part_round_stats_mq(q, cpu, &part_to_disk(part)->part0, now); + __part_round_stats_mq(q, cpu, part, now); +} + static int blk_mq_poll_stats_bkt(const struct request *rq) { int ddir, bytes, bucket; diff --git a/block/blk.h b/block/blk.h index 01ebb8185f6b..4803289467ac 100644 --- a/block/blk.h +++ b/block/blk.h @@ -264,6 +264,11 @@ static inline void req_set_nomerge(struct request_queue *q, struct request *req) q->last_merge = NULL; } +extern void part_round_stats_mq(struct request_queue *q, int cpu, + struct hd_struct *part); +extern unsigned long part_in_flight_mq(struct request_queue *q, + struct hd_struct *part); + /* * Internal io_context interface */ diff --git a/block/genhd.c b/block/genhd.c index 7f520fa25d16..8ec19773ce68 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -1204,6 +1204,7 @@ static int diskstats_show(struct seq_file *seqf, void *v) struct disk_part_iter piter; struct hd_struct *hd; char buf[BDEVNAME_SIZE]; + unsigned long inflight; int cpu; /* @@ -1217,10 +1218,17 @@ static int diskstats_show(struct seq_file *seqf, void *v) disk_part_iter_init(&piter, gp, DISK_PITER_INCL_EMPTY_PART0); while ((hd = disk_part_iter_next(&piter))) { cpu = part_stat_lock(); - part_round_stats(cpu, hd); + if (gp->queue->mq_ops) + part_round_stats_mq(gp->queue, cpu, hd); + else + part_round_stats(cpu, hd); part_stat_unlock(); + if (gp->queue->mq_ops) + inflight = part_in_flight_mq(gp->queue, hd); + else + inflight = part_in_flight(hd); seq_printf(seqf, "%4d %7d %s %lu %lu %lu " - "%u %lu %lu %lu %u %u %u %u\n", + "%u %lu %lu %lu %u %lu %u %u\n", MAJOR(part_devt(hd)), MINOR(part_devt(hd)), disk_name(gp, hd->partno, buf), part_stat_read(hd, ios[READ]), @@ -1231,7 +1239,7 @@ static int diskstats_show(struct seq_file *seqf, void *v) part_stat_read(hd, merges[WRITE]), part_stat_read(hd, sectors[WRITE]), jiffies_to_msecs(part_stat_read(hd, ticks[WRITE])), - part_in_flight(hd), + inflight, jiffies_to_msecs(part_stat_read(hd, io_ticks)), jiffies_to_msecs(part_stat_read(hd, time_in_queue)) ); diff --git a/block/partition-generic.c b/block/partition-generic.c index c5ec8246e25e..94aa92c3c010 100644 --- a/block/partition-generic.c +++ b/block/partition-generic.c @@ -19,6 +19,7 @@ #include <linux/blktrace_api.h> #include "partitions/check.h" +#include "blk.h" #ifdef CONFIG_BLK_DEV_MD extern void md_autodetect_dev(dev_t dev); @@ -111,16 +112,27 @@ static ssize_t part_discard_alignment_show(struct device *dev, ssize_t part_stat_show(struct device *dev, struct device_attribute *attr, char *buf) { + struct gendisk *disk = dev_to_disk(dev); struct hd_struct *p = dev_to_part(dev); + unsigned long inflight; int cpu; cpu = part_stat_lock(); - part_round_stats(cpu, p); + if (disk->queue->mq_ops) + part_round_stats_mq(disk->queue, cpu, p); + else + part_round_stats(cpu, p); part_stat_unlock(); + + if (disk->queue->mq_ops) + inflight = part_in_flight_mq(disk->queue, p); + else + inflight = part_in_flight(p); + return sprintf(buf, "%8lu %8lu %8llu %8u " "%8lu %8lu %8llu %8u " - "%8u %8u %8u" + "%8lu %8u %8u" "\n", part_stat_read(p, ios[READ]), part_stat_read(p, merges[READ]), @@ -130,7 +142,7 @@ ssize_t part_stat_show(struct device *dev, part_stat_read(p, merges[WRITE]), (unsigned long long)part_stat_read(p, sectors[WRITE]), jiffies_to_msecs(part_stat_read(p, ticks[WRITE])), - part_in_flight(p), + inflight, jiffies_to_msecs(part_stat_read(p, io_ticks)), jiffies_to_msecs(part_stat_read(p, time_in_queue))); } @@ -140,8 +152,8 @@ ssize_t part_inflight_show(struct device *dev, { struct hd_struct *p = dev_to_part(dev); - return sprintf(buf, "%8u %8u\n", atomic_read(&p->in_flight[0]), - atomic_read(&p->in_flight[1])); + return sprintf(buf, "%8lu %8lu\n", part_stat_read(p, in_flight[0]), + part_stat_read(p, in_flight[1])); } #ifdef CONFIG_FAIL_MAKE_REQUEST diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 402946035308..1034abffd10d 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -517,9 +517,9 @@ static void start_io_acct(struct dm_io *io) cpu = part_stat_lock(); part_round_stats(cpu, &dm_disk(md)->part0); + part_inc_in_flight(cpu, &dm_disk(md)->part0, rw); + atomic_inc(&md->pending[rw]); part_stat_unlock(); - atomic_set(&dm_disk(md)->part0.in_flight[rw], - atomic_inc_return(&md->pending[rw])); if (unlikely(dm_stats_used(&md->stats))) dm_stats_account_io(&md->stats, bio_data_dir(bio), @@ -532,7 +532,7 @@ static void end_io_acct(struct dm_io *io) struct mapped_device *md = io->md; struct bio *bio = io->bio; unsigned long duration = jiffies - io->start_time; - int pending; + int pending, cpu; int rw = bio_data_dir(bio); generic_end_io_acct(rw, &dm_disk(md)->part0, io->start_time); @@ -546,9 +546,11 @@ static void end_io_acct(struct dm_io *io) * After this is decremented the bio must not be touched if it is * a flush. */ + cpu = part_stat_lock(); pending = atomic_dec_return(&md->pending[rw]); - atomic_set(&dm_disk(md)->part0.in_flight[rw], pending); + part_dec_in_flight(cpu, &dm_disk(md)->part0, rw); pending += atomic_read(&md->pending[rw^0x1]); + part_stat_unlock(); /* nudge anyone waiting on suspend queue */ if (!pending) diff --git a/include/linux/genhd.h b/include/linux/genhd.h index e619fae2f037..d050a509bdd4 100644 --- a/include/linux/genhd.h +++ b/include/linux/genhd.h @@ -87,6 +87,7 @@ struct disk_stats { unsigned long ticks[2]; unsigned long io_ticks; unsigned long time_in_queue; + unsigned long in_flight[2]; }; #define PARTITION_META_INFO_VOLNAMELTH 64 @@ -120,7 +121,6 @@ struct hd_struct { int make_it_fail; #endif unsigned long stamp; - atomic_t in_flight[2]; #ifdef CONFIG_SMP struct disk_stats __percpu *dkstats; #else @@ -292,6 +292,17 @@ extern struct hd_struct *disk_map_sector_rcu(struct gendisk *disk, #define __part_stat_add(cpu, part, field, addnd) \ (per_cpu_ptr((part)->dkstats, (cpu))->field += (addnd)) +#define part_stat_read_double(part, field1, field2) \ +({ \ + typeof((part)->dkstats->field1) res = 0; \ + unsigned int _cpu; \ + for_each_possible_cpu(_cpu) { \ + res += per_cpu_ptr((part)->dkstats, _cpu)->field1; \ + res += per_cpu_ptr((part)->dkstats, _cpu)->field2; \ + } \ + res; \ +}) + #define part_stat_read(part, field) \ ({ \ typeof((part)->dkstats->field) res = 0; \ @@ -362,23 +373,23 @@ static inline void free_part_stats(struct hd_struct *part) #define part_stat_sub(cpu, gendiskp, field, subnd) \ part_stat_add(cpu, gendiskp, field, -subnd) -static inline void part_inc_in_flight(struct hd_struct *part, int rw) +static inline void part_inc_in_flight(int cpu, struct hd_struct *part, int rw) { - atomic_inc(&part->in_flight[rw]); + part_stat_inc(cpu, part, in_flight[rw]); if (part->partno) - atomic_inc(&part_to_disk(part)->part0.in_flight[rw]); + part_stat_inc(cpu, &part_to_disk(part)->part0, in_flight[rw]); } -static inline void part_dec_in_flight(struct hd_struct *part, int rw) +static inline void part_dec_in_flight(int cpu, struct hd_struct *part, int rw) { - atomic_dec(&part->in_flight[rw]); + part_stat_dec(cpu, part, in_flight[rw]); if (part->partno) - atomic_dec(&part_to_disk(part)->part0.in_flight[rw]); + part_stat_dec(cpu, &part_to_disk(part)->part0, in_flight[rw]); } -static inline int part_in_flight(struct hd_struct *part) +static inline unsigned long part_in_flight(struct hd_struct *part) { - return atomic_read(&part->in_flight[0]) + atomic_read(&part->in_flight[1]); + return part_stat_read_double(part, in_flight[0], in_flight[1]); } static inline struct partition_meta_info *alloc_part_info(struct gendisk *disk)