Skip to contentSkip to conversation

erofs: replace erofs_unzipd workqueue with per-cpu threads

PATCH In review Sandeep Dhavale 3 participants 6 messages Jan 6, 2023 – Feb 8, 2023 EROFS file system

lore
Share & tools
lore

Sandeep Dhavale

author reviewer
#0 lore raw

PATCH v4 1/1 erofs: replace erofs_unzipd workqueue with per-cpu threads

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

---
V3 -> V4
* Updated commit message with background information V2 -> V3
* Fix a warning Reported-by: kernel test robot <[email protected]> V1 -> V2
* Changed name of kthread_workers from z_erofs to erofs_worker
* Added kernel configuration to run kthread_workers at normal or high priority
* Added cpu hotplug support
* Added wrapped kthread_workers under worker_pool
* Added one unbound thread in a pool to handle a context where we already stopped per-cpu kthread worker
* Updated commit message

---
 fs/erofs/Kconfig |  11 +++
 fs/erofs/zdata.c | 201 +++++++++++++++++++++++++++++++++++++++++------
 fs/erofs/zdata.h |   4 +-
 3 files changed, 192 insertions(+), 24 deletions(-)
fs/erofs/Kconfig +11 -0
index 85490370e0ca..879f493c6641 100644
@@ -108,3 +108,14 @@ config EROFS_FS_ONDEMAND
 	  read support.
 
 	  If unsure, say N.
+
+config EROFS_FS_KTHREAD_HIPRI
+	bool "EROFS high priority percpu kthread workers"
+	depends on EROFS_FS
+	default n
+	help
+	  EROFS uses per cpu kthread workers pool to carry out async work.
+	  This permits EROFS to configure per cpu kthread workers to run
+	  at higher priority.
+
+	  If unsure, say N.
fs/erofs/zdata.c +178 -23

262 lines not on this page · raw message

fs/erofs/zdata.h +3 -1
index d98c95212985..808bbbf71b7b 100644
@@ -6,6 +6,8 @@
 #ifndef __EROFS_FS_ZDATA_H
 #define __EROFS_FS_ZDATA_H
 
+#include <linux/kthread.h>
+
 #include "internal.h"
 #include "tagptr.h"
 
@@ -107,7 +109,7 @@ struct z_erofs_decompressqueue {
 
 	union {
 		struct completion done;
-		struct work_struct work;
+		struct kthread_work work;
 	} u;
 
 	bool eio;
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
maintainer
#1 lore raw

Hi Sandeep,

Quoted 37 lines · Sandeep Dhavale

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

Quoted 14 lines

---
V3 -> V4
* Updated commit message with background information V2 -> V3
* Fix a warning Reported-by: kernel test robot <[email protected]> V1 -> V2
* Changed name of kthread_workers from z_erofs to erofs_worker
* Added kernel configuration to run kthread_workers at normal or high priority
* Added cpu hotplug support
* Added wrapped kthread_workers under worker_pool
* Added one unbound thread in a pool to handle a context where we already stopped per-cpu kthread worker
* Updated commit message

I've just modified your v4 patch based on erofs -dev branch with my previous suggestion [1], but I haven't tested it.

Could you help check if the updated patch looks good to you and test it on your side? If there are unexpected behaviors, please help update as well, thanks!

[1] https://lore.kernel.org/r/[email protected]/

Thanks,
Gao Xiang

From 2e87235abc745c0fef8e32abcd3a51546b4378ad Mon Sep 17 00:00:00 2001
From: Sandeep Dhavale <[email protected]>
Date: Mon, 6 Feb 2023 17:53:39 +0800
Subject: [PATCH] erofs: add per-cpu threads for decompression

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

---
 fs/erofs/Kconfig |  18 +++++
 fs/erofs/zdata.c | 190 ++++++++++++++++++++++++++++++++++++++++++-----
 2 files changed, 189 insertions(+), 19 deletions(-)
fs/erofs/Kconfig +18 -0
index 85490370e0ca..704fb59577e0 100644
@@ -108,3 +108,21 @@ config EROFS_FS_ONDEMAND
 	  read support.
 
 	  If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD
+	bool "EROFS per-cpu decompression kthread workers"
+	depends on EROFS_FS_ZIP
+	help
+	  Saying Y here enables per-CPU kthread workers pool to carry out
+	  async decompression for low latencies on some architectures.
+
+	  If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD_HIPRI
+	bool "EROFS high priority per-CPU kthread workers"
+	depends on EROFS_FS_ZIP && EROFS_FS_PCPU_KTHREAD
+	help
+	  This permits EROFS to configure per-CPU kthread workers to run
+	  at higher priority.
+
+	  If unsure, say N.
fs/erofs/zdata.c +171 -19

251 lines not on this page · raw message

Signature · 1 line
2.30.2
author reviewer
#2 lore raw
Quoted 56 lines · Gao Xiang

Hi Sandeep,

Quoted 37 lines · Sandeep Dhavale

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

Quoted 14 lines

---
V3 -> V4
* Updated commit message with background information V2 -> V3
* Fix a warning Reported-by: kernel test robot <[email protected]> V1 -> V2
* Changed name of kthread_workers from z_erofs to erofs_worker
* Added kernel configuration to run kthread_workers at normal or high priority
* Added cpu hotplug support
* Added wrapped kthread_workers under worker_pool
* Added one unbound thread in a pool to handle a context where we already stopped per-cpu kthread worker
* Updated commit message

Quoted 6 lines

I've just modified your v4 patch based on erofs -dev branch with my previous suggestion [1], but I haven't tested it.

Could you help check if the updated patch looks good to you and test it on your side? If there are unexpected behaviors, please help update as well, thanks!

Thanks Xiang, I was working on the same. I see that you have cleaned it up. I will test it and report/fix any problems.

Thanks,
Sandeep.

Quoted 346 lines

[1] https://lore.kernel.org/r/[email protected]/

Thanks,
Gao Xiang

From 2e87235abc745c0fef8e32abcd3a51546b4378ad Mon Sep 17 00:00:00 2001
From: Sandeep Dhavale <[email protected]>
Date: Mon, 6 Feb 2023 17:53:39 +0800
Subject: [PATCH] erofs: add per-cpu threads for decompression

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

---
 fs/erofs/Kconfig |  18 +++++
 fs/erofs/zdata.c | 190 ++++++++++++++++++++++++++++++++++++++++++-----
 2 files changed, 189 insertions(+), 19 deletions(-)
fs/erofs/Kconfig +18 -0
index 85490370e0ca..704fb59577e0 100644
@@ -108,3 +108,21 @@ config EROFS_FS_ONDEMAND
          read support.
 
          If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD
+       bool "EROFS per-cpu decompression kthread workers"
+       depends on EROFS_FS_ZIP
+       help
+         Saying Y here enables per-CPU kthread workers pool to carry out
+         async decompression for low latencies on some architectures.
+
+         If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD_HIPRI
+       bool "EROFS high priority per-CPU kthread workers"
+       depends on EROFS_FS_ZIP && EROFS_FS_PCPU_KTHREAD
+       help
+         This permits EROFS to configure per-CPU kthread workers to run
+         at higher priority.
+
+         If unsure, say N.
fs/erofs/zdata.c +171 -20
index 384f64292f73..73198f494a6a 100644
@@ -7,6 +7,8 @@
 #include "compress.h"
 #include <linux/prefetch.h>
 #include <linux/psi.h>
+#include <linux/slab.h>
+#include <linux/cpuhotplug.h>
 
 #include <trace/events/erofs.h>
 
@@ -109,6 +111,7 @@ struct z_erofs_decompressqueue {
        union {
                struct completion done;
                struct work_struct work;
+               struct kthread_work kthread_work;
        } u;
        bool eio, sync;
 };
@@ -341,24 +344,128 @@ static void z_erofs_free_pcluster(struct z_erofs_pcluster *pcl)
 
 static struct workqueue_struct *z_erofs_workqueue __read_mostly;
 
-void z_erofs_exit_zip_subsystem(void)
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+static struct kthread_worker __rcu **z_erofs_pcpu_workers;
+
+static void erofs_destroy_percpu_workers(void)
 {
-       destroy_workqueue(z_erofs_workqueue);
-       z_erofs_destroy_pcluster_pool();
+       struct kthread_worker *worker;
+       unsigned int cpu;
+
+       for_each_possible_cpu(cpu) {
+               worker = rcu_dereference_protected(
+                                       z_erofs_pcpu_workers[cpu], 1);
+               rcu_assign_pointer(z_erofs_pcpu_workers[cpu], NULL);
+               if (worker)
+                       kthread_destroy_worker(worker);
+       }
+       kfree(z_erofs_pcpu_workers);
 }
 
-static inline int z_erofs_init_workqueue(void)
+static struct kthread_worker *erofs_init_percpu_worker(int cpu)
 {
-       const unsigned int onlinecpus = num_possible_cpus();
+       struct kthread_worker *worker =
+               kthread_create_worker_on_cpu(cpu, 0, "erofs_worker/%u", cpu);
 
-       /*
-        * no need to spawn too many threads, limiting threads could minimum
-        * scheduling overhead, perhaps per-CPU threads should be better?
-        */
-       z_erofs_workqueue = alloc_workqueue("erofs_unzipd",
-                                           WQ_UNBOUND | WQ_HIGHPRI,
-                                           onlinecpus + onlinecpus / 4);
-       return z_erofs_workqueue ? 0 : -ENOMEM;
+       if (IS_ERR(worker))
+               return worker;
+       if (IS_ENABLED(CONFIG_EROFS_FS_PCPU_KTHREAD_HIPRI))
+               sched_set_fifo_low(worker->task);
+       else
+               sched_set_normal(worker->task, 0);
+       return worker;
+}
+
+static int erofs_init_percpu_workers(void)
+{
+       struct kthread_worker *worker;
+       unsigned int cpu;
+
+       z_erofs_pcpu_workers = kcalloc(num_possible_cpus(),
+                       sizeof(struct kthread_worker *), GFP_ATOMIC);
+       if (!z_erofs_pcpu_workers)
+               return -ENOMEM;
+
+       for_each_online_cpu(cpu) {      /* could miss cpu{off,on}line? */
+               worker = erofs_init_percpu_worker(cpu);
+               if (!IS_ERR(worker))
+                       rcu_assign_pointer(z_erofs_pcpu_workers[cpu], worker);
+       }
+       return 0;
+}
+#else
+static inline void erofs_destroy_percpu_workers(void) {}
+static inline int erofs_init_percpu_workers(void) { return 0; }
+#endif
+
+#if defined(CONFIG_HOTPLUG_CPU) && defined(EROFS_FS_PCPU_KTHREAD)
+static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
+static enum cpuhp_state erofs_cpuhp_state;
+
+static int erofs_cpu_online(unsigned int cpu)
+{
+       struct kthread_worker *worker, *old;
+
+       worker = erofs_init_percpu_worker(cpu);
+       if (IS_ERR(worker))
+               return ERR_PTR(worker);
+
+       spin_lock(&z_erofs_pcpu_worker_lock);
+       old = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
+                       lockdep_is_held(&z_erofs_pcpu_worker_lock));
+       if (!old)
+               rcu_assign_pointer(z_erofs_pcpu_workers[cpu], worker);
+       spin_unlock(&z_erofs_pcpu_worker_lock);
+       if (old)
+               kthread_destroy_worker(worker);
+       return 0;
+}
+
+static int erofs_cpu_offline(unsigned int cpu)
+{
+       struct kthread_worker *worker;
+
+       spin_lock(&z_erofs_pcpu_worker_lock);
+       worker = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
+                       lockdep_is_held(&z_erofs_pcpu_worker_lock));
+       rcu_assign_pointer(worker_pool.workers[cpu], NULL);
+       spin_unlock(&z_erofs_pcpu_worker_lock);
+
+       synchronize_rcu();
+       if (worker)
+               kthread_destroy_worker(worker);
+       return 0;
+}
+
+static int erofs_cpu_hotplug_init(void)
+{
+       int state;
+
+       state = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
+                       "fs/erofs:online", erofs_cpu_online, erofs_cpu_offline);
+       if (state < 0)
+               return state;
+
+       erofs_cpuhp_state = state;
+       return 0;
+}
+
+static void erofs_cpu_hotplug_destroy(void)
+{
+       if (erofs_cpuhp_state)
+               cpuhp_remove_state_nocalls(erofs_cpuhp_state);
+}
+#else /* !CONFIG_HOTPLUG_CPU || !CONFIG_EROFS_FS_PCPU_KTHREAD */
+static inline int erofs_cpu_hotplug_init(void) { return 0; }
+static inline void erofs_cpu_hotplug_destroy(void) {}
+#endif
+
+void z_erofs_exit_zip_subsystem(void)
+{
+       erofs_cpu_hotplug_destroy();
+       erofs_destroy_percpu_workers();
+       destroy_workqueue(z_erofs_workqueue);
+       z_erofs_destroy_pcluster_pool();
 }
 
 int __init z_erofs_init_zip_subsystem(void)
@@ -366,10 +473,29 @@ int __init z_erofs_init_zip_subsystem(void)
        int err = z_erofs_create_pcluster_pool();
 
        if (err)
-               return err;
-       err = z_erofs_init_workqueue();
+               goto out_error_pcluster_pool;
+
+       z_erofs_workqueue = alloc_workqueue("erofs_worker",
+                       WQ_UNBOUND | WQ_HIGHPRI, num_possible_cpus());
+       if (!z_erofs_workqueue)
+               goto out_error_workqueue_init;
+
+       err = erofs_init_percpu_workers();
        if (err)
-               z_erofs_destroy_pcluster_pool();
+               goto out_error_pcpu_worker;
+
+       err = erofs_cpu_hotplug_init();
+       if (err < 0)
+               goto out_error_cpuhp_init;
+       return err;
+
+out_error_cpuhp_init:
+       erofs_destroy_percpu_workers();
+out_error_pcpu_worker:
+       destroy_workqueue(z_erofs_workqueue);
+out_error_workqueue_init:
+       z_erofs_destroy_pcluster_pool();
+out_error_pcluster_pool:
        return err;
 }
 
@@ -1305,11 +1431,17 @@ static void z_erofs_decompressqueue_work(struct work_struct *work)
 
        DBG_BUGON(bgq->head == Z_EROFS_PCLUSTER_TAIL_CLOSED);
        z_erofs_decompress_queue(bgq, &pagepool);
-
        erofs_release_pages(&pagepool);
        kvfree(bgq);
 }
 
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+static void z_erofs_decompressqueue_kthread_work(struct kthread_work *work)
+{
+       z_erofs_decompressqueue_work((struct work_struct *)work);
+}
+#endif
+
 static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
                                       int bios)
 {
@@ -1324,9 +1456,24 @@ static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
 
        if (atomic_add_return(bios, &io->pending_bios))
                return;
-       /* Use workqueue and sync decompression for atomic contexts only */
+       /* Use (kthread_)work and sync decompression for atomic contexts only */
        if (in_atomic() || irqs_disabled()) {
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+               struct kthread_worker *worker;
+
+               rcu_read_lock();
+               worker = rcu_dereference(
+                               z_erofs_pcpu_workers[raw_smp_processor_id()]);
+               if (!worker) {
+                       INIT_WORK(&io->u.work, z_erofs_decompressqueue_work);
+                       queue_work(z_erofs_workqueue, &io->u.work);
+               } else {
+                       kthread_queue_work(worker, &io->u.kthread_work);
+               }
+               rcu_read_unlock();
+#else
                queue_work(z_erofs_workqueue, &io->u.work);
+#endif
                /* enable sync decompression for readahead */
                if (sbi->opt.sync_decompress == EROFS_SYNC_DECOMPRESS_AUTO)
                        sbi->opt.sync_decompress = EROFS_SYNC_DECOMPRESS_FORCE_ON;
@@ -1455,7 +1602,12 @@ static struct z_erofs_decompressqueue *jobqueue_init(struct super_block *sb,
                        *fg = true;
                        goto fg_out;
                }
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+               kthread_init_work(&q->u.kthread_work,
+                                 z_erofs_decompressqueue_kthread_work);
+#else
                INIT_WORK(&q->u.work, z_erofs_decompressqueue_work);
+#endif
        } else {
 fg_out:
                q = fgq;
@@ -1640,7 +1792,7 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
 
        /*
         * although background is preferred, no one is pending for submission.
-        * don't issue workqueue for decompression but drop it directly instead.
+        * don't issue decompression but drop it directly instead.
         */
        if (!*force_fg && !nr_bios) {
                kvfree(q[JQ_SUBMIT]);
--

2.30.2

#3 lore raw
Quoted 60 lines · Sandeep Dhavale
Quoted 59 lines · Gao Xiang

Hi Sandeep,

Quoted 37 lines · Sandeep Dhavale

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

Quoted 14 lines

---
V3 -> V4
* Updated commit message with background information V2 -> V3
* Fix a warning Reported-by: kernel test robot <[email protected]> V1 -> V2
* Changed name of kthread_workers from z_erofs to erofs_worker
* Added kernel configuration to run kthread_workers at normal or high priority
* Added cpu hotplug support
* Added wrapped kthread_workers under worker_pool
* Added one unbound thread in a pool to handle a context where we already stopped per-cpu kthread worker
* Updated commit message

I've just modified your v4 patch based on erofs -dev branch with my previous suggestion [1], but I haven't tested it.

Quoted 8 lines
Quoted 3 lines

Could you help check if the updated patch looks good to you and test it on your side? If there are unexpected behaviors, please help update as well, thanks!

Thanks Xiang, I was working on the same. I see that you have cleaned it up. I will test it and report/fix any problems.

Thanks,
Sandeep.

Thanks! Look forward to your test. BTW, we have < 2 weeks for 6.3, so I'd like to fix it this week so that we could catch 6.3 merge window.

I've fixed some cpu hotplug errors as below and added to a branch for 0day CI testing.

Thanks,
Gao Xiang

fs/erofs/zdata.c +3 -3
index 73198f494a6a..92a9e20948b0 100644
@@ -398,7 +398,7 @@ static inline void erofs_destroy_percpu_workers(void) {}
 static inline int erofs_init_percpu_workers(void) { return 0; }
 #endif
 
-#if defined(CONFIG_HOTPLUG_CPU) && defined(EROFS_FS_PCPU_KTHREAD)
+#if defined(CONFIG_HOTPLUG_CPU) && defined(CONFIG_EROFS_FS_PCPU_KTHREAD)
 static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
 static enum cpuhp_state erofs_cpuhp_state;
 
@@ -408,7 +408,7 @@ static int erofs_cpu_online(unsigned int cpu)
 
 	worker = erofs_init_percpu_worker(cpu);
 	if (IS_ERR(worker))
-		return ERR_PTR(worker);
+		return PTR_ERR(worker);
 
 	spin_lock(&z_erofs_pcpu_worker_lock);
 	old = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
@@ -428,7 +428,7 @@ static int erofs_cpu_offline(unsigned int cpu)
 	spin_lock(&z_erofs_pcpu_worker_lock);
 	worker = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
 			lockdep_is_held(&z_erofs_pcpu_worker_lock));
-	rcu_assign_pointer(worker_pool.workers[cpu], NULL);
+	rcu_assign_pointer(z_erofs_pcpu_workers[cpu], NULL);
 	spin_unlock(&z_erofs_pcpu_worker_lock);
 
 	synchronize_rcu();
Quoted 347 lines
Quoted 347 lines

[1] https://lore.kernel.org/r/[email protected]/

Thanks,
Gao Xiang

From 2e87235abc745c0fef8e32abcd3a51546b4378ad Mon Sep 17 00:00:00 2001
From: Sandeep Dhavale <[email protected]>
Date: Mon, 6 Feb 2023 17:53:39 +0800
Subject: [PATCH] erofs: add per-cpu threads for decompression

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

---
  fs/erofs/Kconfig |  18 +++++
  fs/erofs/zdata.c | 190 ++++++++++++++++++++++++++++++++++++++++++-----
  2 files changed, 189 insertions(+), 19 deletions(-)
fs/erofs/Kconfig +18 -0
index 85490370e0ca..704fb59577e0 100644
@@ -108,3 +108,21 @@ config EROFS_FS_ONDEMAND
           read support.
 
           If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD
+       bool "EROFS per-cpu decompression kthread workers"
+       depends on EROFS_FS_ZIP
+       help
+         Saying Y here enables per-CPU kthread workers pool to carry out
+         async decompression for low latencies on some architectures.
+
+         If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD_HIPRI
+       bool "EROFS high priority per-CPU kthread workers"
+       depends on EROFS_FS_ZIP && EROFS_FS_PCPU_KTHREAD
+       help
+         This permits EROFS to configure per-CPU kthread workers to run
+         at higher priority.
+
+         If unsure, say N.
fs/erofs/zdata.c +171 -20
index 384f64292f73..73198f494a6a 100644
@@ -7,6 +7,8 @@
  #include "compress.h"
  #include <linux/prefetch.h>
  #include <linux/psi.h>
+#include <linux/slab.h>
+#include <linux/cpuhotplug.h>
 
  #include <trace/events/erofs.h>
 
@@ -109,6 +111,7 @@ struct z_erofs_decompressqueue {
         union {
                 struct completion done;
                 struct work_struct work;
+               struct kthread_work kthread_work;
         } u;
         bool eio, sync;
  };
@@ -341,24 +344,128 @@ static void z_erofs_free_pcluster(struct z_erofs_pcluster *pcl)
 
  static struct workqueue_struct *z_erofs_workqueue __read_mostly;
 
-void z_erofs_exit_zip_subsystem(void)
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+static struct kthread_worker __rcu **z_erofs_pcpu_workers;
+
+static void erofs_destroy_percpu_workers(void)
  {
-       destroy_workqueue(z_erofs_workqueue);
-       z_erofs_destroy_pcluster_pool();
+       struct kthread_worker *worker;
+       unsigned int cpu;
+
+       for_each_possible_cpu(cpu) {
+               worker = rcu_dereference_protected(
+                                       z_erofs_pcpu_workers[cpu], 1);
+               rcu_assign_pointer(z_erofs_pcpu_workers[cpu], NULL);
+               if (worker)
+                       kthread_destroy_worker(worker);
+       }
+       kfree(z_erofs_pcpu_workers);
  }
 
-static inline int z_erofs_init_workqueue(void)
+static struct kthread_worker *erofs_init_percpu_worker(int cpu)
  {
-       const unsigned int onlinecpus = num_possible_cpus();
+       struct kthread_worker *worker =
+               kthread_create_worker_on_cpu(cpu, 0, "erofs_worker/%u", cpu);
 
-       /*
-        * no need to spawn too many threads, limiting threads could minimum
-        * scheduling overhead, perhaps per-CPU threads should be better?
-        */
-       z_erofs_workqueue = alloc_workqueue("erofs_unzipd",
-                                           WQ_UNBOUND | WQ_HIGHPRI,
-                                           onlinecpus + onlinecpus / 4);
-       return z_erofs_workqueue ? 0 : -ENOMEM;
+       if (IS_ERR(worker))
+               return worker;
+       if (IS_ENABLED(CONFIG_EROFS_FS_PCPU_KTHREAD_HIPRI))
+               sched_set_fifo_low(worker->task);
+       else
+               sched_set_normal(worker->task, 0);
+       return worker;
+}
+
+static int erofs_init_percpu_workers(void)
+{
+       struct kthread_worker *worker;
+       unsigned int cpu;
+
+       z_erofs_pcpu_workers = kcalloc(num_possible_cpus(),
+                       sizeof(struct kthread_worker *), GFP_ATOMIC);
+       if (!z_erofs_pcpu_workers)
+               return -ENOMEM;
+
+       for_each_online_cpu(cpu) {      /* could miss cpu{off,on}line? */
+               worker = erofs_init_percpu_worker(cpu);
+               if (!IS_ERR(worker))
+                       rcu_assign_pointer(z_erofs_pcpu_workers[cpu], worker);
+       }
+       return 0;
+}
+#else
+static inline void erofs_destroy_percpu_workers(void) {}
+static inline int erofs_init_percpu_workers(void) { return 0; }
+#endif
+
+#if defined(CONFIG_HOTPLUG_CPU) && defined(EROFS_FS_PCPU_KTHREAD)
+static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
+static enum cpuhp_state erofs_cpuhp_state;
+
+static int erofs_cpu_online(unsigned int cpu)
+{
+       struct kthread_worker *worker, *old;
+
+       worker = erofs_init_percpu_worker(cpu);
+       if (IS_ERR(worker))
+               return ERR_PTR(worker);
+
+       spin_lock(&z_erofs_pcpu_worker_lock);
+       old = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
+                       lockdep_is_held(&z_erofs_pcpu_worker_lock));
+       if (!old)
+               rcu_assign_pointer(z_erofs_pcpu_workers[cpu], worker);
+       spin_unlock(&z_erofs_pcpu_worker_lock);
+       if (old)
+               kthread_destroy_worker(worker);
+       return 0;
+}
+
+static int erofs_cpu_offline(unsigned int cpu)
+{
+       struct kthread_worker *worker;
+
+       spin_lock(&z_erofs_pcpu_worker_lock);
+       worker = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
+                       lockdep_is_held(&z_erofs_pcpu_worker_lock));
+       rcu_assign_pointer(worker_pool.workers[cpu], NULL);
+       spin_unlock(&z_erofs_pcpu_worker_lock);
+
+       synchronize_rcu();
+       if (worker)
+               kthread_destroy_worker(worker);
+       return 0;
+}
+
+static int erofs_cpu_hotplug_init(void)
+{
+       int state;
+
+       state = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
+                       "fs/erofs:online", erofs_cpu_online, erofs_cpu_offline);
+       if (state < 0)
+               return state;
+
+       erofs_cpuhp_state = state;
+       return 0;
+}
+
+static void erofs_cpu_hotplug_destroy(void)
+{
+       if (erofs_cpuhp_state)
+               cpuhp_remove_state_nocalls(erofs_cpuhp_state);
+}
+#else /* !CONFIG_HOTPLUG_CPU || !CONFIG_EROFS_FS_PCPU_KTHREAD */
+static inline int erofs_cpu_hotplug_init(void) { return 0; }
+static inline void erofs_cpu_hotplug_destroy(void) {}
+#endif
+
+void z_erofs_exit_zip_subsystem(void)
+{
+       erofs_cpu_hotplug_destroy();
+       erofs_destroy_percpu_workers();
+       destroy_workqueue(z_erofs_workqueue);
+       z_erofs_destroy_pcluster_pool();
  }
 
  int __init z_erofs_init_zip_subsystem(void)
@@ -366,10 +473,29 @@ int __init z_erofs_init_zip_subsystem(void)
         int err = z_erofs_create_pcluster_pool();
 
         if (err)
-               return err;
-       err = z_erofs_init_workqueue();
+               goto out_error_pcluster_pool;
+
+       z_erofs_workqueue = alloc_workqueue("erofs_worker",
+                       WQ_UNBOUND | WQ_HIGHPRI, num_possible_cpus());
+       if (!z_erofs_workqueue)
+               goto out_error_workqueue_init;
+
+       err = erofs_init_percpu_workers();
         if (err)
-               z_erofs_destroy_pcluster_pool();
+               goto out_error_pcpu_worker;
+
+       err = erofs_cpu_hotplug_init();
+       if (err < 0)
+               goto out_error_cpuhp_init;
+       return err;
+
+out_error_cpuhp_init:
+       erofs_destroy_percpu_workers();
+out_error_pcpu_worker:
+       destroy_workqueue(z_erofs_workqueue);
+out_error_workqueue_init:
+       z_erofs_destroy_pcluster_pool();
+out_error_pcluster_pool:
         return err;
  }
 
@@ -1305,11 +1431,17 @@ static void z_erofs_decompressqueue_work(struct work_struct *work)
 
         DBG_BUGON(bgq->head == Z_EROFS_PCLUSTER_TAIL_CLOSED);
         z_erofs_decompress_queue(bgq, &pagepool);
-
         erofs_release_pages(&pagepool);
         kvfree(bgq);
  }
 
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+static void z_erofs_decompressqueue_kthread_work(struct kthread_work *work)
+{
+       z_erofs_decompressqueue_work((struct work_struct *)work);
+}
+#endif
+
  static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
                                        int bios)
  {
@@ -1324,9 +1456,24 @@ static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
 
         if (atomic_add_return(bios, &io->pending_bios))
                 return;
-       /* Use workqueue and sync decompression for atomic contexts only */
+       /* Use (kthread_)work and sync decompression for atomic contexts only */
         if (in_atomic() || irqs_disabled()) {
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+               struct kthread_worker *worker;
+
+               rcu_read_lock();
+               worker = rcu_dereference(
+                               z_erofs_pcpu_workers[raw_smp_processor_id()]);
+               if (!worker) {
+                       INIT_WORK(&io->u.work, z_erofs_decompressqueue_work);
+                       queue_work(z_erofs_workqueue, &io->u.work);
+               } else {
+                       kthread_queue_work(worker, &io->u.kthread_work);
+               }
+               rcu_read_unlock();
+#else
                 queue_work(z_erofs_workqueue, &io->u.work);
+#endif
                 /* enable sync decompression for readahead */
                 if (sbi->opt.sync_decompress == EROFS_SYNC_DECOMPRESS_AUTO)
                         sbi->opt.sync_decompress = EROFS_SYNC_DECOMPRESS_FORCE_ON;
@@ -1455,7 +1602,12 @@ static struct z_erofs_decompressqueue *jobqueue_init(struct super_block *sb,
                         *fg = true;
                         goto fg_out;
                 }
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+               kthread_init_work(&q->u.kthread_work,
+                                 z_erofs_decompressqueue_kthread_work);
+#else
                 INIT_WORK(&q->u.work, z_erofs_decompressqueue_work);
+#endif
         } else {
  fg_out:
                 q = fgq;
@@ -1640,7 +1792,7 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
 
         /*
          * although background is preferred, no one is pending for submission.
-        * don't issue workqueue for decompression but drop it directly instead.
+        * don't issue decompression but drop it directly instead.
          */
         if (!*force_fg && !nr_bios) {
                 kvfree(q[JQ_SUBMIT]);
--

2.30.2

author reviewer
#4 lore raw
Quoted 73 lines · Gao Xiang
Quoted 60 lines · Sandeep Dhavale
Quoted 59 lines · Gao Xiang

Hi Sandeep,

Quoted 37 lines · Sandeep Dhavale

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

Quoted 14 lines

---
V3 -> V4
* Updated commit message with background information V2 -> V3
* Fix a warning Reported-by: kernel test robot <[email protected]> V1 -> V2
* Changed name of kthread_workers from z_erofs to erofs_worker
* Added kernel configuration to run kthread_workers at normal or high priority
* Added cpu hotplug support
* Added wrapped kthread_workers under worker_pool
* Added one unbound thread in a pool to handle a context where we already stopped per-cpu kthread worker
* Updated commit message

I've just modified your v4 patch based on erofs -dev branch with my previous suggestion [1], but I haven't tested it.

Quoted 8 lines
Quoted 3 lines

Could you help check if the updated patch looks good to you and test it on your side? If there are unexpected behaviors, please help update as well, thanks!

Thanks Xiang, I was working on the same. I see that you have cleaned it up. I will test it and report/fix any problems.

Thanks,
Sandeep.

Quoted 7 lines

Thanks! Look forward to your test. BTW, we have < 2 weeks for 6.3, so I'd like to fix it this week so that we could catch 6.3 merge window.

I've fixed some cpu hotplug errors as below and added to a branch for 0day CI testing.

Hi Xiang,
With this version of the patch I have tested
- Multiple device reboot test
- Cold App launch tests
- Cold App launch tests with cpu offline/online

All tests ran successfully and no issue was observed.

Thanks,
Sandeep.

Quoted 382 lines

Thanks,
Gao Xiang

fs/erofs/zdata.c +3 -3
index 73198f494a6a..92a9e20948b0 100644
@@ -398,7 +398,7 @@ static inline void erofs_destroy_percpu_workers(void) {}
  static inline int erofs_init_percpu_workers(void) { return 0; }
  #endif
 
-#if defined(CONFIG_HOTPLUG_CPU) && defined(EROFS_FS_PCPU_KTHREAD)
+#if defined(CONFIG_HOTPLUG_CPU) && defined(CONFIG_EROFS_FS_PCPU_KTHREAD)
  static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
  static enum cpuhp_state erofs_cpuhp_state;
 
@@ -408,7 +408,7 @@ static int erofs_cpu_online(unsigned int cpu)
 
        worker = erofs_init_percpu_worker(cpu);
        if (IS_ERR(worker))
-               return ERR_PTR(worker);
+               return PTR_ERR(worker);
 
        spin_lock(&z_erofs_pcpu_worker_lock);
        old = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
@@ -428,7 +428,7 @@ static int erofs_cpu_offline(unsigned int cpu)
        spin_lock(&z_erofs_pcpu_worker_lock);
        worker = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
                        lockdep_is_held(&z_erofs_pcpu_worker_lock));
-       rcu_assign_pointer(worker_pool.workers[cpu], NULL);
+       rcu_assign_pointer(z_erofs_pcpu_workers[cpu], NULL);
        spin_unlock(&z_erofs_pcpu_worker_lock);
 
        synchronize_rcu();
Quoted 348 lines
Quoted 348 lines

[1] https://lore.kernel.org/r/[email protected]/

Thanks,
Gao Xiang

From 2e87235abc745c0fef8e32abcd3a51546b4378ad Mon Sep 17 00:00:00 2001
From: Sandeep Dhavale <[email protected]>
Date: Mon, 6 Feb 2023 17:53:39 +0800
Subject: [PATCH] erofs: add per-cpu threads for decompression

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

---
  fs/erofs/Kconfig |  18 +++++
  fs/erofs/zdata.c | 190 ++++++++++++++++++++++++++++++++++++++++++-----
  2 files changed, 189 insertions(+), 19 deletions(-)
fs/erofs/Kconfig +18 -0
index 85490370e0ca..704fb59577e0 100644
@@ -108,3 +108,21 @@ config EROFS_FS_ONDEMAND
           read support.
 
           If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD
+       bool "EROFS per-cpu decompression kthread workers"
+       depends on EROFS_FS_ZIP
+       help
+         Saying Y here enables per-CPU kthread workers pool to carry out
+         async decompression for low latencies on some architectures.
+
+         If unsure, say N.
+
+config EROFS_FS_PCPU_KTHREAD_HIPRI
+       bool "EROFS high priority per-CPU kthread workers"
+       depends on EROFS_FS_ZIP && EROFS_FS_PCPU_KTHREAD
+       help
+         This permits EROFS to configure per-CPU kthread workers to run
+         at higher priority.
+
+         If unsure, say N.
fs/erofs/zdata.c +171 -20
index 384f64292f73..73198f494a6a 100644
@@ -7,6 +7,8 @@
  #include "compress.h"
  #include <linux/prefetch.h>
  #include <linux/psi.h>
+#include <linux/slab.h>
+#include <linux/cpuhotplug.h>
 
  #include <trace/events/erofs.h>
 
@@ -109,6 +111,7 @@ struct z_erofs_decompressqueue {
         union {
                 struct completion done;
                 struct work_struct work;
+               struct kthread_work kthread_work;
         } u;
         bool eio, sync;
  };
@@ -341,24 +344,128 @@ static void z_erofs_free_pcluster(struct z_erofs_pcluster *pcl)
 
  static struct workqueue_struct *z_erofs_workqueue __read_mostly;
 
-void z_erofs_exit_zip_subsystem(void)
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+static struct kthread_worker __rcu **z_erofs_pcpu_workers;
+
+static void erofs_destroy_percpu_workers(void)
  {
-       destroy_workqueue(z_erofs_workqueue);
-       z_erofs_destroy_pcluster_pool();
+       struct kthread_worker *worker;
+       unsigned int cpu;
+
+       for_each_possible_cpu(cpu) {
+               worker = rcu_dereference_protected(
+                                       z_erofs_pcpu_workers[cpu], 1);
+               rcu_assign_pointer(z_erofs_pcpu_workers[cpu], NULL);
+               if (worker)
+                       kthread_destroy_worker(worker);
+       }
+       kfree(z_erofs_pcpu_workers);
  }
 
-static inline int z_erofs_init_workqueue(void)
+static struct kthread_worker *erofs_init_percpu_worker(int cpu)
  {
-       const unsigned int onlinecpus = num_possible_cpus();
+       struct kthread_worker *worker =
+               kthread_create_worker_on_cpu(cpu, 0, "erofs_worker/%u", cpu);
 
-       /*
-        * no need to spawn too many threads, limiting threads could minimum
-        * scheduling overhead, perhaps per-CPU threads should be better?
-        */
-       z_erofs_workqueue = alloc_workqueue("erofs_unzipd",
-                                           WQ_UNBOUND | WQ_HIGHPRI,
-                                           onlinecpus + onlinecpus / 4);
-       return z_erofs_workqueue ? 0 : -ENOMEM;
+       if (IS_ERR(worker))
+               return worker;
+       if (IS_ENABLED(CONFIG_EROFS_FS_PCPU_KTHREAD_HIPRI))
+               sched_set_fifo_low(worker->task);
+       else
+               sched_set_normal(worker->task, 0);
+       return worker;
+}
+
+static int erofs_init_percpu_workers(void)
+{
+       struct kthread_worker *worker;
+       unsigned int cpu;
+
+       z_erofs_pcpu_workers = kcalloc(num_possible_cpus(),
+                       sizeof(struct kthread_worker *), GFP_ATOMIC);
+       if (!z_erofs_pcpu_workers)
+               return -ENOMEM;
+
+       for_each_online_cpu(cpu) {      /* could miss cpu{off,on}line? */
+               worker = erofs_init_percpu_worker(cpu);
+               if (!IS_ERR(worker))
+                       rcu_assign_pointer(z_erofs_pcpu_workers[cpu], worker);
+       }
+       return 0;
+}
+#else
+static inline void erofs_destroy_percpu_workers(void) {}
+static inline int erofs_init_percpu_workers(void) { return 0; }
+#endif
+
+#if defined(CONFIG_HOTPLUG_CPU) && defined(EROFS_FS_PCPU_KTHREAD)
+static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
+static enum cpuhp_state erofs_cpuhp_state;
+
+static int erofs_cpu_online(unsigned int cpu)
+{
+       struct kthread_worker *worker, *old;
+
+       worker = erofs_init_percpu_worker(cpu);
+       if (IS_ERR(worker))
+               return ERR_PTR(worker);
+
+       spin_lock(&z_erofs_pcpu_worker_lock);
+       old = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
+                       lockdep_is_held(&z_erofs_pcpu_worker_lock));
+       if (!old)
+               rcu_assign_pointer(z_erofs_pcpu_workers[cpu], worker);
+       spin_unlock(&z_erofs_pcpu_worker_lock);
+       if (old)
+               kthread_destroy_worker(worker);
+       return 0;
+}
+
+static int erofs_cpu_offline(unsigned int cpu)
+{
+       struct kthread_worker *worker;
+
+       spin_lock(&z_erofs_pcpu_worker_lock);
+       worker = rcu_dereference_protected(z_erofs_pcpu_workers[cpu],
+                       lockdep_is_held(&z_erofs_pcpu_worker_lock));
+       rcu_assign_pointer(worker_pool.workers[cpu], NULL);
+       spin_unlock(&z_erofs_pcpu_worker_lock);
+
+       synchronize_rcu();
+       if (worker)
+               kthread_destroy_worker(worker);
+       return 0;
+}
+
+static int erofs_cpu_hotplug_init(void)
+{
+       int state;
+
+       state = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
+                       "fs/erofs:online", erofs_cpu_online, erofs_cpu_offline);
+       if (state < 0)
+               return state;
+
+       erofs_cpuhp_state = state;
+       return 0;
+}
+
+static void erofs_cpu_hotplug_destroy(void)
+{
+       if (erofs_cpuhp_state)
+               cpuhp_remove_state_nocalls(erofs_cpuhp_state);
+}
+#else /* !CONFIG_HOTPLUG_CPU || !CONFIG_EROFS_FS_PCPU_KTHREAD */
+static inline int erofs_cpu_hotplug_init(void) { return 0; }
+static inline void erofs_cpu_hotplug_destroy(void) {}
+#endif
+
+void z_erofs_exit_zip_subsystem(void)
+{
+       erofs_cpu_hotplug_destroy();
+       erofs_destroy_percpu_workers();
+       destroy_workqueue(z_erofs_workqueue);
+       z_erofs_destroy_pcluster_pool();
  }
 
  int __init z_erofs_init_zip_subsystem(void)
@@ -366,10 +473,29 @@ int __init z_erofs_init_zip_subsystem(void)
         int err = z_erofs_create_pcluster_pool();
 
         if (err)
-               return err;
-       err = z_erofs_init_workqueue();
+               goto out_error_pcluster_pool;
+
+       z_erofs_workqueue = alloc_workqueue("erofs_worker",
+                       WQ_UNBOUND | WQ_HIGHPRI, num_possible_cpus());
+       if (!z_erofs_workqueue)
+               goto out_error_workqueue_init;
+
+       err = erofs_init_percpu_workers();
         if (err)
-               z_erofs_destroy_pcluster_pool();
+               goto out_error_pcpu_worker;
+
+       err = erofs_cpu_hotplug_init();
+       if (err < 0)
+               goto out_error_cpuhp_init;
+       return err;
+
+out_error_cpuhp_init:
+       erofs_destroy_percpu_workers();
+out_error_pcpu_worker:
+       destroy_workqueue(z_erofs_workqueue);
+out_error_workqueue_init:
+       z_erofs_destroy_pcluster_pool();
+out_error_pcluster_pool:
         return err;
  }
 
@@ -1305,11 +1431,17 @@ static void z_erofs_decompressqueue_work(struct work_struct *work)
 
         DBG_BUGON(bgq->head == Z_EROFS_PCLUSTER_TAIL_CLOSED);
         z_erofs_decompress_queue(bgq, &pagepool);
-
         erofs_release_pages(&pagepool);
         kvfree(bgq);
  }
 
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+static void z_erofs_decompressqueue_kthread_work(struct kthread_work *work)
+{
+       z_erofs_decompressqueue_work((struct work_struct *)work);
+}
+#endif
+
  static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
                                        int bios)
  {
@@ -1324,9 +1456,24 @@ static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
 
         if (atomic_add_return(bios, &io->pending_bios))
                 return;
-       /* Use workqueue and sync decompression for atomic contexts only */
+       /* Use (kthread_)work and sync decompression for atomic contexts only */
         if (in_atomic() || irqs_disabled()) {
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+               struct kthread_worker *worker;
+
+               rcu_read_lock();
+               worker = rcu_dereference(
+                               z_erofs_pcpu_workers[raw_smp_processor_id()]);
+               if (!worker) {
+                       INIT_WORK(&io->u.work, z_erofs_decompressqueue_work);
+                       queue_work(z_erofs_workqueue, &io->u.work);
+               } else {
+                       kthread_queue_work(worker, &io->u.kthread_work);
+               }
+               rcu_read_unlock();
+#else
                 queue_work(z_erofs_workqueue, &io->u.work);
+#endif
                 /* enable sync decompression for readahead */
                 if (sbi->opt.sync_decompress == EROFS_SYNC_DECOMPRESS_AUTO)
                         sbi->opt.sync_decompress = EROFS_SYNC_DECOMPRESS_FORCE_ON;
@@ -1455,7 +1602,12 @@ static struct z_erofs_decompressqueue *jobqueue_init(struct super_block *sb,
                         *fg = true;
                         goto fg_out;
                 }
+#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
+               kthread_init_work(&q->u.kthread_work,
+                                 z_erofs_decompressqueue_kthread_work);
+#else
                 INIT_WORK(&q->u.work, z_erofs_decompressqueue_work);
+#endif
         } else {
  fg_out:
                 q = fgq;
@@ -1640,7 +1792,7 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
 
         /*
          * although background is preferred, no one is pending for submission.
-        * don't issue workqueue for decompression but drop it directly instead.
+        * don't issue decompression but drop it directly instead.
          */
         if (!*force_fg && !nr_bios) {
                 kvfree(q[JQ_SUBMIT]);
--

2.30.2

#5 lore raw
Quoted 81 lines · Sandeep Dhavale
Quoted 73 lines · Gao Xiang
Quoted 60 lines · Sandeep Dhavale
Quoted 59 lines · Gao Xiang

Hi Sandeep,

Quoted 37 lines · Sandeep Dhavale

Using per-cpu thread pool we can reduce the scheduling latency compared to workqueue implementation. With this patch scheduling latency and variation is reduced as per-cpu threads are high priority kthread_workers.

The results were evaluated on arm64 Android devices running 5.10 kernel.

The table below shows resulting improvements of total scheduling latency for the same app launch benchmark runs with 50 iterations. Scheduling latency is the latency between when the task (workqueue kworker vs kthread_worker) became eligible to run to when it actually started running.

+-------------------------+-----------+----------------+---------+
|                         | workqueue | kthread_worker |  diff   |
+-------------------------+-----------+----------------+---------+
| Average (us)            |     15253 |           2914 | -80.89% |
| Median (us)             |     14001 |           2912 | -79.20% |
| Minimum (us)            |      3117 |           1027 | -67.05% |
| Maximum (us)            |     30170 |           3805 | -87.39% |
| Standard deviation (us) |      7166 |            359 |         |
+-------------------------+-----------+----------------+---------+

Background: Boot times and cold app launch benchmarks are very important to the android ecosystem as they directly translate to responsiveness from user point of view. While erofs provides a lot of important features like space savings, we saw some performance penalty in cold app launch benchmarks in few scenarios. Analysis showed that the significant variance was coming from the scheduling cost while decompression cost was more or less the same.

Having per-cpu thread pool we can see from the above table that this variation is reduced by ~80% on average. This problem was discussed at LPC 2022. Link to LPC 2022 slides and
talk at [1]

[1] https://lpc.events/event/16/contributions/1338/

Quoted 14 lines

---
V3 -> V4
* Updated commit message with background information V2 -> V3
* Fix a warning Reported-by: kernel test robot <[email protected]> V1 -> V2
* Changed name of kthread_workers from z_erofs to erofs_worker
* Added kernel configuration to run kthread_workers at normal or high priority
* Added cpu hotplug support
* Added wrapped kthread_workers under worker_pool
* Added one unbound thread in a pool to handle a context where we already stopped per-cpu kthread worker
* Updated commit message

I've just modified your v4 patch based on erofs -dev branch with my previous suggestion [1], but I haven't tested it.

Quoted 8 lines
Quoted 3 lines

Could you help check if the updated patch looks good to you and test it on your side? If there are unexpected behaviors, please help update as well, thanks!

Thanks Xiang, I was working on the same. I see that you have cleaned it up. I will test it and report/fix any problems.

Thanks,
Sandeep.

Quoted 7 lines

Thanks! Look forward to your test. BTW, we have < 2 weeks for 6.3, so I'd like to fix it this week so that we could catch 6.3 merge window.

I've fixed some cpu hotplug errors as below and added to a branch for 0day CI testing.

Quoted 7 lines

Hi Xiang,
With this version of the patch I have tested
- Multiple device reboot test
- Cold App launch tests
- Cold App launch tests with cpu offline/online

All tests ran successfully and no issue was observed.

Okay, thanks! I will resend & submit this version for -next now and test on my side if no other concerns.

Thanks,
Gao XIang