Message ID | 20231115121839.12556-1-xiaobing.li@samsung.com (mailing list archive) |
---|---|
State | New |
Headers | show |
Series | [v3] io_uring: Statistics of the true utilization of sq threads. | expand |
On 11/15/23 5:18 AM, Xiaobing Li wrote: > v3: > 1.Since the sq thread has a while(1) structure, during this process, > there may be a lot of time that is not processing IO but does not > exceed the timeout period, therefore, the sqpoll thread will keep > running and will keep occupying the CPU. Obviously, the CPU is wasted at > this time;Our goal is to count the part of the time that the sqpoll > thread actually processes IO, so as to reflect the part of the CPU it > uses to process IO, which can be used to help improve the actual > utilization of the CPU in the future. > > 2."work_time" in the code represents the sum of the jiffies count of the > sq thread actually processing IO, that is, how many milliseconds it > actually takes to process IO. "total_time" represents the total time > that the sq thread has elapsed from the beginning of the loop to the > current time point, that is, how many milliseconds it has spent in > total. > The output "SqBusy" represents the percentage of time utilization that > the sq thread actually uses to process IO. > > 3.The task_pid value in the io_sq_data structure should be assigned > after the sq thread is created, otherwise the pid of its parent > process will be recorded. > > 4.After many tests, we do not need to obtain ctx->uring_lock in advance > when obtaining ctx->sq_data. We can avoid null pointer references by > judging that ctx is not null. This is mixed in with my patch, please just base it on top of the v4 I'll send out now rather than integrate it. And then you'd need that rcu free patch as well on top of that, then your changes as a separate patch.
Hi Xiaobing, kernel test robot noticed the following build warnings: https://git-scm.com/docs/git-format-patch#_base_tree_information] url: https://github.com/intel-lab-lkp/linux/commits/Xiaobing-Li/io_uring-Statistics-of-the-true-utilization-of-sq-threads/20231115-211954 base: linus/master patch link: https://lore.kernel.org/r/20231115121839.12556-1-xiaobing.li%40samsung.com patch subject: [PATCH v3] io_uring: Statistics of the true utilization of sq threads. config: x86_64-randconfig-161-20231115 (https://download.01.org/0day-ci/archive/20231116/202311160629.h4GrebJh-lkp@intel.com/config) compiler: gcc-12 (Debian 12.2.0-14) 12.2.0 reproduce: (https://download.01.org/0day-ci/archive/20231116/202311160629.h4GrebJh-lkp@intel.com/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp@intel.com> | Reported-by: Dan Carpenter <error27@gmail.com> | Closes: https://lore.kernel.org/r/202311160629.h4GrebJh-lkp@intel.com/ smatch warnings: io_uring/fdinfo.c:138 io_uring_show_fdinfo() warn: variable dereferenced before check 'ctx' (see line 57) io_uring/fdinfo.c:144 io_uring_show_fdinfo() error: we previously assumed 'sq' could be null (see line 141) vim +/ctx +138 io_uring/fdinfo.c 3aaf22b62a9270 Jens Axboe 2023-07-10 53 __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *f) a4ad4f748ea962 Jens Axboe 2022-05-25 54 { 3aaf22b62a9270 Jens Axboe 2023-07-10 55 struct io_ring_ctx *ctx = f->private_data; a4ad4f748ea962 Jens Axboe 2022-05-25 56 struct io_overflow_cqe *ocqe; a4ad4f748ea962 Jens Axboe 2022-05-25 @57 struct io_rings *r = ctx->rings; ^^^^^^^^^^ Tons of unchecked dereferences so ctx can't be NULL. a4ad4f748ea962 Jens Axboe 2022-05-25 58 unsigned int sq_mask = ctx->sq_entries - 1, cq_mask = ctx->cq_entries - 1; a4ad4f748ea962 Jens Axboe 2022-05-25 59 unsigned int sq_head = READ_ONCE(r->sq.head); a4ad4f748ea962 Jens Axboe 2022-05-25 60 unsigned int sq_tail = READ_ONCE(r->sq.tail); a4ad4f748ea962 Jens Axboe 2022-05-25 61 unsigned int cq_head = READ_ONCE(r->cq.head); a4ad4f748ea962 Jens Axboe 2022-05-25 62 unsigned int cq_tail = READ_ONCE(r->cq.tail); a4ad4f748ea962 Jens Axboe 2022-05-25 63 unsigned int cq_shift = 0; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 64 unsigned int sq_shift = 0; a4ad4f748ea962 Jens Axboe 2022-05-25 65 unsigned int sq_entries, cq_entries; 7644b1a1c9a7ae Jens Axboe 2023-10-21 66 int sq_pid = -1, sq_cpu = -1; 5b1b61674371b7 Xiaobing Li 2023-11-15 67 int sq_busy = 0; a4ad4f748ea962 Jens Axboe 2022-05-25 68 bool has_lock; a4ad4f748ea962 Jens Axboe 2022-05-25 69 unsigned int i; a4ad4f748ea962 Jens Axboe 2022-05-25 70 4f731705cc1f15 Jens Axboe 2022-09-11 71 if (ctx->flags & IORING_SETUP_CQE32) a4ad4f748ea962 Jens Axboe 2022-05-25 72 cq_shift = 1; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 73 if (ctx->flags & IORING_SETUP_SQE128) 3b8fdd1dc35e39 Jens Axboe 2022-09-11 74 sq_shift = 1; a4ad4f748ea962 Jens Axboe 2022-05-25 75 a4ad4f748ea962 Jens Axboe 2022-05-25 76 /* a4ad4f748ea962 Jens Axboe 2022-05-25 77 * we may get imprecise sqe and cqe info if uring is actively running a4ad4f748ea962 Jens Axboe 2022-05-25 78 * since we get cached_sq_head and cached_cq_tail without uring_lock a4ad4f748ea962 Jens Axboe 2022-05-25 79 * and sq_tail and cq_head are changed by userspace. But it's ok since a4ad4f748ea962 Jens Axboe 2022-05-25 80 * we usually use these info when it is stuck. a4ad4f748ea962 Jens Axboe 2022-05-25 81 */ a4ad4f748ea962 Jens Axboe 2022-05-25 82 seq_printf(m, "SqMask:\t0x%x\n", sq_mask); a4ad4f748ea962 Jens Axboe 2022-05-25 83 seq_printf(m, "SqHead:\t%u\n", sq_head); a4ad4f748ea962 Jens Axboe 2022-05-25 84 seq_printf(m, "SqTail:\t%u\n", sq_tail); a4ad4f748ea962 Jens Axboe 2022-05-25 85 seq_printf(m, "CachedSqHead:\t%u\n", ctx->cached_sq_head); a4ad4f748ea962 Jens Axboe 2022-05-25 86 seq_printf(m, "CqMask:\t0x%x\n", cq_mask); a4ad4f748ea962 Jens Axboe 2022-05-25 87 seq_printf(m, "CqHead:\t%u\n", cq_head); a4ad4f748ea962 Jens Axboe 2022-05-25 88 seq_printf(m, "CqTail:\t%u\n", cq_tail); a4ad4f748ea962 Jens Axboe 2022-05-25 89 seq_printf(m, "CachedCqTail:\t%u\n", ctx->cached_cq_tail); 3b8fdd1dc35e39 Jens Axboe 2022-09-11 90 seq_printf(m, "SQEs:\t%u\n", sq_tail - sq_head); a4ad4f748ea962 Jens Axboe 2022-05-25 91 sq_entries = min(sq_tail - sq_head, ctx->sq_entries); a4ad4f748ea962 Jens Axboe 2022-05-25 92 for (i = 0; i < sq_entries; i++) { a4ad4f748ea962 Jens Axboe 2022-05-25 93 unsigned int entry = i + sq_head; a4ad4f748ea962 Jens Axboe 2022-05-25 94 struct io_uring_sqe *sqe; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 95 unsigned int sq_idx; a4ad4f748ea962 Jens Axboe 2022-05-25 96 32f5dea040ee6e Jens Axboe 2023-09-01 97 if (ctx->flags & IORING_SETUP_NO_SQARRAY) 32f5dea040ee6e Jens Axboe 2023-09-01 98 break; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 99 sq_idx = READ_ONCE(ctx->sq_array[entry & sq_mask]); a4ad4f748ea962 Jens Axboe 2022-05-25 100 if (sq_idx > sq_mask) a4ad4f748ea962 Jens Axboe 2022-05-25 101 continue; 00927931cb630b Pavel Begunkov 2022-10-11 102 sqe = &ctx->sq_sqes[sq_idx << sq_shift]; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 103 seq_printf(m, "%5u: opcode:%s, fd:%d, flags:%x, off:%llu, " 3b8fdd1dc35e39 Jens Axboe 2022-09-11 104 "addr:0x%llx, rw_flags:0x%x, buf_index:%d " 3b8fdd1dc35e39 Jens Axboe 2022-09-11 105 "user_data:%llu", 3b8fdd1dc35e39 Jens Axboe 2022-09-11 106 sq_idx, io_uring_get_opcode(sqe->opcode), sqe->fd, 3b8fdd1dc35e39 Jens Axboe 2022-09-11 107 sqe->flags, (unsigned long long) sqe->off, 3b8fdd1dc35e39 Jens Axboe 2022-09-11 108 (unsigned long long) sqe->addr, sqe->rw_flags, 3b8fdd1dc35e39 Jens Axboe 2022-09-11 109 sqe->buf_index, sqe->user_data); 3b8fdd1dc35e39 Jens Axboe 2022-09-11 110 if (sq_shift) { 3b8fdd1dc35e39 Jens Axboe 2022-09-11 111 u64 *sqeb = (void *) (sqe + 1); 3b8fdd1dc35e39 Jens Axboe 2022-09-11 112 int size = sizeof(struct io_uring_sqe) / sizeof(u64); 3b8fdd1dc35e39 Jens Axboe 2022-09-11 113 int j; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 114 3b8fdd1dc35e39 Jens Axboe 2022-09-11 115 for (j = 0; j < size; j++) { 3b8fdd1dc35e39 Jens Axboe 2022-09-11 116 seq_printf(m, ", e%d:0x%llx", j, 3b8fdd1dc35e39 Jens Axboe 2022-09-11 117 (unsigned long long) *sqeb); 3b8fdd1dc35e39 Jens Axboe 2022-09-11 118 sqeb++; 3b8fdd1dc35e39 Jens Axboe 2022-09-11 119 } 3b8fdd1dc35e39 Jens Axboe 2022-09-11 120 } 3b8fdd1dc35e39 Jens Axboe 2022-09-11 121 seq_printf(m, "\n"); a4ad4f748ea962 Jens Axboe 2022-05-25 122 } a4ad4f748ea962 Jens Axboe 2022-05-25 123 seq_printf(m, "CQEs:\t%u\n", cq_tail - cq_head); a4ad4f748ea962 Jens Axboe 2022-05-25 124 cq_entries = min(cq_tail - cq_head, ctx->cq_entries); a4ad4f748ea962 Jens Axboe 2022-05-25 125 for (i = 0; i < cq_entries; i++) { a4ad4f748ea962 Jens Axboe 2022-05-25 126 unsigned int entry = i + cq_head; a4ad4f748ea962 Jens Axboe 2022-05-25 127 struct io_uring_cqe *cqe = &r->cqes[(entry & cq_mask) << cq_shift]; a4ad4f748ea962 Jens Axboe 2022-05-25 128 4f731705cc1f15 Jens Axboe 2022-09-11 129 seq_printf(m, "%5u: user_data:%llu, res:%d, flag:%x", a4ad4f748ea962 Jens Axboe 2022-05-25 130 entry & cq_mask, cqe->user_data, cqe->res, a4ad4f748ea962 Jens Axboe 2022-05-25 131 cqe->flags); 4f731705cc1f15 Jens Axboe 2022-09-11 132 if (cq_shift) 4f731705cc1f15 Jens Axboe 2022-09-11 133 seq_printf(m, ", extra1:%llu, extra2:%llu\n", 4f731705cc1f15 Jens Axboe 2022-09-11 134 cqe->big_cqe[0], cqe->big_cqe[1]); 4f731705cc1f15 Jens Axboe 2022-09-11 135 seq_printf(m, "\n"); a4ad4f748ea962 Jens Axboe 2022-05-25 136 } a4ad4f748ea962 Jens Axboe 2022-05-25 137 5b1b61674371b7 Xiaobing Li 2023-11-15 @138 if (ctx && (ctx->flags & IORING_SETUP_SQPOLL)) { ^^^ Delete this check. 5b1b61674371b7 Xiaobing Li 2023-11-15 139 struct io_sq_data *sq = ctx->sq_data; 5b1b61674371b7 Xiaobing Li 2023-11-15 140 5b1b61674371b7 Xiaobing Li 2023-11-15 @141 if (sq && sq->total_time != 0) ^^ sq can be NULL? 5b1b61674371b7 Xiaobing Li 2023-11-15 142 sq_busy = (int)(sq->work_time * 100 / sq->total_time); 5b1b61674371b7 Xiaobing Li 2023-11-15 143 5b1b61674371b7 Xiaobing Li 2023-11-15 @144 sq_pid = sq->task_pid; ^^^^ Unchecked dereference.
diff --git a/io_uring/fdinfo.c b/io_uring/fdinfo.c index f04a43044d91..b9e2e339140d 100644 --- a/io_uring/fdinfo.c +++ b/io_uring/fdinfo.c @@ -64,6 +64,7 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *f) unsigned int sq_shift = 0; unsigned int sq_entries, cq_entries; int sq_pid = -1, sq_cpu = -1; + int sq_busy = 0; bool has_lock; unsigned int i; @@ -134,6 +135,21 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *f) seq_printf(m, "\n"); } + if (ctx && (ctx->flags & IORING_SETUP_SQPOLL)) { + struct io_sq_data *sq = ctx->sq_data; + + if (sq && sq->total_time != 0) + sq_busy = (int)(sq->work_time * 100 / sq->total_time); + + sq_pid = sq->task_pid; + sq_cpu = sq->sq_cpu; + } + + seq_printf(m, "SqThread:\t%d\n", sq_pid); + seq_printf(m, "SqThreadCpu:\t%d\n", sq_cpu); + seq_printf(m, "SqBusy:\t%d%%\n", sq_busy); + seq_printf(m, "UserFiles:\t%u\n", ctx->nr_user_files); + /* * Avoid ABBA deadlock between the seq lock and the io_uring mutex, * since fdinfo case grabs it in the opposite direction of normal use @@ -142,21 +158,6 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *f) */ has_lock = mutex_trylock(&ctx->uring_lock); - if (has_lock && (ctx->flags & IORING_SETUP_SQPOLL)) { - struct io_sq_data *sq = ctx->sq_data; - - if (mutex_trylock(&sq->lock)) { - if (sq->thread) { - sq_pid = task_pid_nr(sq->thread); - sq_cpu = task_cpu(sq->thread); - } - mutex_unlock(&sq->lock); - } - } - - seq_printf(m, "SqThread:\t%d\n", sq_pid); - seq_printf(m, "SqThreadCpu:\t%d\n", sq_cpu); - seq_printf(m, "UserFiles:\t%u\n", ctx->nr_user_files); for (i = 0; has_lock && i < ctx->nr_user_files; i++) { struct file *f = io_file_from_index(&ctx->file_table, i); diff --git a/io_uring/sqpoll.c b/io_uring/sqpoll.c index bd6c2c7959a5..dc093adc1ce5 100644 --- a/io_uring/sqpoll.c +++ b/io_uring/sqpoll.c @@ -224,17 +224,21 @@ static int io_sq_thread(void *data) struct io_ring_ctx *ctx; unsigned long timeout = 0; char buf[TASK_COMM_LEN]; + unsigned long sq_start, sq_work_begin, sq_work_end; DEFINE_WAIT(wait); snprintf(buf, sizeof(buf), "iou-sqp-%d", sqd->task_pid); set_task_comm(current, buf); - if (sqd->sq_cpu != -1) + if (sqd->sq_cpu != -1) { set_cpus_allowed_ptr(current, cpumask_of(sqd->sq_cpu)); - else + } else { set_cpus_allowed_ptr(current, cpu_online_mask); + sqd->sq_cpu = raw_smp_processor_id(); + } mutex_lock(&sqd->lock); + sq_start = jiffies; while (1) { bool cap_entries, sqt_spin = false; @@ -245,6 +249,7 @@ static int io_sq_thread(void *data) } cap_entries = !list_is_singular(&sqd->ctx_list); + sq_work_begin = jiffies; list_for_each_entry(ctx, &sqd->ctx_list, sqd_list) { int ret = __io_sq_thread(ctx, cap_entries); @@ -254,6 +259,11 @@ static int io_sq_thread(void *data) if (io_run_task_work()) sqt_spin = true; + sq_work_end = jiffies; + sqd->total_time = sq_work_end - sq_start; + if (sqt_spin == true) + sqd->work_time += sq_work_end - sq_work_begin; + if (sqt_spin || !time_after(jiffies, timeout)) { if (sqt_spin) timeout = jiffies + sqd->sq_thread_idle; @@ -261,6 +271,7 @@ static int io_sq_thread(void *data) mutex_unlock(&sqd->lock); cond_resched(); mutex_lock(&sqd->lock); + sqd->sq_cpu = raw_smp_processor_id(); } continue; } @@ -294,6 +305,7 @@ static int io_sq_thread(void *data) mutex_unlock(&sqd->lock); schedule(); mutex_lock(&sqd->lock); + sqd->sq_cpu = raw_smp_processor_id(); } list_for_each_entry(ctx, &sqd->ctx_list, sqd_list) atomic_andnot(IORING_SQ_NEED_WAKEUP, @@ -395,14 +407,14 @@ __cold int io_sq_offload_create(struct io_ring_ctx *ctx, sqd->sq_cpu = -1; } - sqd->task_pid = current->pid; - sqd->task_tgid = current->tgid; + sqd->task_tgid = current->pid; tsk = create_io_thread(io_sq_thread, sqd, NUMA_NO_NODE); if (IS_ERR(tsk)) { ret = PTR_ERR(tsk); goto err_sqpoll; } + sqd->task_pid = task_pid_nr(tsk); sqd->thread = tsk; ret = io_uring_alloc_task_context(tsk, ctx); wake_up_new_task(tsk); diff --git a/io_uring/sqpoll.h b/io_uring/sqpoll.h index 8df37e8c9149..fd6fa9587843 100644 --- a/io_uring/sqpoll.h +++ b/io_uring/sqpoll.h @@ -16,6 +16,8 @@ struct io_sq_data { pid_t task_pid; pid_t task_tgid; + unsigned long work_time; + unsigned long total_time; unsigned long state; struct completion exited; };
v3: 1.Since the sq thread has a while(1) structure, during this process, there may be a lot of time that is not processing IO but does not exceed the timeout period, therefore, the sqpoll thread will keep running and will keep occupying the CPU. Obviously, the CPU is wasted at this time;Our goal is to count the part of the time that the sqpoll thread actually processes IO, so as to reflect the part of the CPU it uses to process IO, which can be used to help improve the actual utilization of the CPU in the future. 2."work_time" in the code represents the sum of the jiffies count of the sq thread actually processing IO, that is, how many milliseconds it actually takes to process IO. "total_time" represents the total time that the sq thread has elapsed from the beginning of the loop to the current time point, that is, how many milliseconds it has spent in total. The output "SqBusy" represents the percentage of time utilization that the sq thread actually uses to process IO. 3.The task_pid value in the io_sq_data structure should be assigned after the sq thread is created, otherwise the pid of its parent process will be recorded. 4.After many tests, we do not need to obtain ctx->uring_lock in advance when obtaining ctx->sq_data. We can avoid null pointer references by judging that ctx is not null. Signed-off-by: Xiaobing Li <xiaobing.li@samsung.com> The test results are as follows: Every 0.5s: cat /proc/281126/fdinfo/6 | grep Sq SqMask: 0x3 SqHead: 1168417 SqTail: 1168418 CachedSqHead: 1168418 SqThread: 281126 SqThreadCpu: 55 SqBusy: 96% --- io_uring/fdinfo.c | 31 ++++++++++++++++--------------- io_uring/sqpoll.c | 20 ++++++++++++++++---- io_uring/sqpoll.h | 2 ++ 3 files changed, 34 insertions(+), 19 deletions(-)