task_work: cleanup notification modes
Subsystems (14)
PATCH 1/1 task_work: cleanup notification modes
A previous commit changed the notification mode from 0/1 to allowing 0/1/TWA_SIGNAL. Clean this up properly, and define a proper enum for the notification mode. Now we have:
- TWA_NONE. This is 0, same as before the original change, meaning no notification requested.
- TWA_RESUME. This is 1, same as before the original change, meaning that we use TIF_NOTIFY_RESUME.
- TWA_SIGNAL. This uses TIF_SIGPENDING/JOBCTL_TASK_WORK for the notification.
Clean up all the callers, switching their 0/1/false/true to using the appropriate TWA_* mode for notifications.
---
As promised, here's the task_work notification cleanup patch.
arch/x86/kernel/cpu/mce/core.c +1 -1
@@ -1277,7 +1277,7 @@ static void queue_task_work(struct mce *m, int kill_it)else current->mce_kill_me.func = kill_me_maybe; -task_work_add (current, ¤t->mce_kill_me, true); +task_work_add (current, ¤t->mce_kill_me,TWA_RESUME ); }/*
arch/x86/kernel/cpu/resctrl/rdtgroup.c +1 -1
@@ -561,7 +561,7 @@ static int __rdtgroup_move_task(struct task_struct *tsk,* callback has been invoked. */ atomic_inc (&rdtgrp->waitcount); - ret =task_work_add (tsk, &callback->work, true); + ret =task_work_add (tsk, &callback->work,TWA_RESUME );if (ret) {/* * Task is exiting. Drop the refcount and free the callback.
drivers/acpi/apei/ghes.c +1 -1
@@ -879,7 +879,7 @@ static void ghes_proc_in_irq(struct irq_work *irq_work) estatus_node->task_work.func = ghes_kick_task_work; estatus_node->task_work_cpu =smp_processor_id (); ret =task_work_add (current, &estatus_node->task_work, - true); +TWA_RESUME );if (ret) estatus_node->task_work.func =NULL ; }
drivers/android/binder.c +1 -1
@@ -2229,7 +2229,7 @@ static void binder_deferred_fd_close(int fd)__close_fd_get_file (fd, &twcb->file);if (twcb->file) {filp_close (twcb->file, current->files); -task_work_add (current, &twcb->twork, true); +task_work_add (current, &twcb->twork,TWA_RESUME ); }else {kfree (twcb); }
fs/file_table.c +1 -1
@@ -339,7 +339,7 @@ void fput_many(struct file *file, unsigned int refs)if (likely (!in_interrupt () && !(task->flags &PF_KTHREAD ))) {init_task_work (&file->f_u.fu_rcuhead, ____fput); -if (!task_work_add (task, &file->f_u.fu_rcuhead, true)) +if (!task_work_add (task, &file->f_u.fu_rcuhead,TWA_RESUME ))return ;/* * After this task has run exit_task_work(),
fs/io_uring.c +7 -6
@@ -1756,7 +1756,7 @@ static void __io_free_req(struct io_kiocb *req)struct task_struct *tsk; tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE ); } } }
@@ -1905,7 +1905,8 @@ static int io_req_task_work_add(struct io_kiocb *req, bool twa_signal_ok) {struct task_struct *tsk = req->task;struct io_ring_ctx *ctx = req->ctx; -int ret, notify; +enum task_work_notify_mode notify; +int ret;if (tsk->flags &PF_EXITING )return -ESRCH ;
@@ -1916,7 +1917,7 @@ static int io_req_task_work_add(struct io_kiocb *req, bool twa_signal_ok)* processing task_work. There's no reliable way to tell if TWA_RESUME * will do the job. */ - notify =0 ; + notify =TWA_NONE ;if (!(ctx->flags &IORING_SETUP_SQPOLL ) && twa_signal_ok) notify =TWA_SIGNAL ;
@@ -1985,7 +1986,7 @@ static void io_req_task_queue(struct io_kiocb *req)init_task_work (&req->task_work, io_req_task_cancel); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); } }
@@ -3199,7 +3200,7 @@ static int io_async_buf_func(struct wait_queue_entry *wait, unsigned mode,/* queue just for cancelation */ init_task_work (&req->task_work, io_req_task_cancel); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); }return 1 ;
@@ -4765,7 +4766,7 @@ static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,WRITE_ONCE (poll->canceled, true); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); }return 1 ;
fs/namespace.c +1 -1
@@ -1191,7 +1191,7 @@ static void mntput_no_expire(struct mount *mnt)struct task_struct *task = current;if (likely (!(task->flags &PF_KTHREAD ))) {init_task_work (&mnt->mnt_rcu, __cleanup_mnt); -if (!task_work_add (task, &mnt->mnt_rcu, true)) +if (!task_work_add (task, &mnt->mnt_rcu,TWA_RESUME ))return ; }if (llist_add (&mnt->mnt_llist, &delayed_mntput_list))
include/linux/task_work.h +8 -3
@@ -13,9 +13,14 @@ init_task_work(struct callback_head *twork, task_work_func_t func) twork->func = func; } -#define TWA_RESUME 1 -#define TWA_SIGNAL 2 -int task_work_add (struct task_struct *task,struct callback_head *twork,int ); +enum task_work_notify_mode { +TWA_NONE , +TWA_RESUME , +TWA_SIGNAL , +}; + +int task_work_add (struct task_struct *task,struct callback_head *twork, +enum task_work_notify_mode mode);struct callback_head *task_work_cancel (struct task_struct *,task_work_func_t );void task_work_run (void );
kernel/events/uprobes.c +1 -1
@@ -1823,7 +1823,7 @@ void uprobe_copy_process(struct task_struct *t, unsigned long flags) t->utask->dup_xol_addr = area->vaddr;init_task_work (&t->utask->dup_xol_work, dup_xol_work); -task_work_add (t, &t->utask->dup_xol_work, true); +task_work_add (t, &t->utask->dup_xol_work,TWA_RESUME ); }/*
kernel/irq/manage.c +1 -1
@@ -1162,7 +1162,7 @@ static int irq_thread(void *data) handler_fn = irq_thread_fn;init_task_work (&on_exit_work, irq_thread_dtor); -task_work_add (current, &on_exit_work, false); +task_work_add (current, &on_exit_work,TWA_NONE );irq_thread_check_affinity (desc, action);
kernel/sched/fair.c +1 -1
@@ -2928,7 +2928,7 @@ static void task_tick_numa(struct rq *rq, struct task_struct *curr) curr->node_stamp += period;if (!time_before (jiffies, curr->mm->numa_next_scan)) -task_work_add (curr, work, true); +task_work_add (curr, work,TWA_RESUME ); } }
kernel/task_work.c +5 -3
@@ -9,7 +9,7 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */* task_work_add - ask the @task to execute @work->func() * @task: the task which should run the callback * @work: the callback to run -* @notify: send the notification if true +* @notify: send chosen notification, if any * * Queue @work for task_work_run() below and notify the @task if @notify. * Fails if the @task is exiting/exited and thus it can't process this @work.
@@ -24,8 +24,8 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */* RETURNS: * 0 if succeeds or -ESRCH. */ -int -task_work_add (struct task_struct *task,struct callback_head *work,int notify) +int task_work_add (struct task_struct *task,struct callback_head *work, +enum task_work_notify_mode notify) {struct callback_head *head;unsigned long flags;
@@ -38,6 +38,8 @@ task_work_add(struct task_struct *task, struct callback_head *work, int notify) }while (cmpxchg (&task->task_works, head, work) != head);switch (notify) { +case TWA_NONE : +break ;case TWA_RESUME :set_notify_resume (task);break ;
security/keys/keyctl.c +1 -1
@@ -1693,7 +1693,7 @@ long keyctl_session_to_parent(void)/* the replacement session keyring is applied just prior to userspace * restarting */ - ret =task_work_add (parent, newwork, true); + ret =task_work_add (parent, newwork,TWA_RESUME );if (!ret) newwork =NULL ; unlock:
security/yama/yama_lsm.c +1 -1
@@ -99,7 +99,7 @@ static void report_access(const char *access, struct task_struct *target, info->access = access; info->target = target; info->agent = agent; -if (task_work_add (current, &info->work, true) ==0 ) +if (task_work_add (current, &info->work,TWA_RESUME ) ==0 )return ;/* success */ WARN (1 ,"report_access called from exiting task" );
Signature · 1 line
Jens Axboe
Quoted 1 line · Jens Axboe
A previous commit changed the notification mode from 0/1 to allowing
No. It changed it from boolean to an int.
There is a fundamental difference between 0/1 and false/true simply because it's a compiler implementation detail how to represent a boolean value.
Assume the following:
#define BAZ 0x08
task_work_add(tsk, &work, foo & BAZ);
So with a function signature of
task_work_add(*tsk, *work, bool signal);
the above 'foo & BAZ' becomes either true of false.
With the changed function signature of
task_work_add(*tsk, *work, int signal);
the above becomes the result of 'foo & BAZ' which means that this construct will not longer do the right thing.
It's pure luck that none of the usage sites relied on the boolean property of that argument.
Please spell it out correctly that converting a boolean argument to an integer argument is not equivalent.
kernel/task_work.c task_work_add(struct task_struct *task, struct callback_head *work, int notify) · line 45
Quoted 6 lines
kernel/task_work.c +2
-0switch (notify) { +case TWA_NONE : +break ;case TWA_RESUME :set_notify_resume (task);break ;
The enum will not prevent that either and what you really want to do is to have some form of debug warning if 'notify' is out of range, which would have been the right thing to do in the first place.
Quoted 2 lines
- * @notify: send the notification if true
+ * @notify: send chosen notification, if any
Is that really all you found to be wrong in that comment?
There is more and please take your time which will in turn save time and nerves on both ends.
Thanks,
tglx
Quoted 23 lines · Thomas Gleixner
Quoted 1 line · Jens Axboe
A previous commit changed the notification mode from 0/1 to allowing
No. It changed it from boolean to an int.
There is a fundamental difference between 0/1 and false/true simply because it's a compiler implementation detail how to represent a boolean value.
Assume the following:
#define BAZ
0x08task_work_add(tsk, &work, foo & BAZ);So with a function signature of
task_work_add(*tsk, *work, bool signal);the above 'foo & BAZ' becomes either true of false.
With the changed function signature of
Quoted 7 lines
task_work_add(*tsk, *work, int signal);the above becomes the result of 'foo & BAZ' which means that this construct will not longer do the right thing.
It's pure luck that none of the usage sites relied on the boolean property of that argument.
It wasn't pure luck, that was checked before that change was made. No users did anything funky, it was all false/true or 0/1.
Quoted 2 lines
Please spell it out correctly that converting a boolean argument to an integer argument is not equivalent.
Fixed up the commit message to be more descriptive.
Quoted 10 lines
kernel/task_work.c task_work_add(struct task_struct *task, struct callback_head *work, int notify) · line 45
Quoted 6 lines
kernel/task_work.c +2
-0switch (notify) { +case TWA_NONE : +break ;case TWA_RESUME :set_notify_resume (task);break ;The enum will not prevent that either and what you really want to do is to have some form of debug warning if 'notify' is out of range, which would have been the right thing to do in the first place.
I added a WARN_ON_ONCE() in the default case for that one.
Quoted 4 lines
Quoted 2 lines
- * @notify: send the notification if true
+ * @notify: send chosen notification, if anyIs that really all you found to be wrong in that comment?
There really is nothing wrong, but it's not very descriptive (wasn't before either). I've added a fuller description of the various TWA_* notification types now.
Signature · 1 line
Jens Axboe
Quoted 8 lines · Jens Axboe
Quoted 4 lines · Thomas Gleixner
Quoted 2 lines
- * @notify: send the notification if true
+ * @notify: send chosen notification, if anyIs that really all you found to be wrong in that comment?
There really is nothing wrong, but it's not very descriptive (wasn't before either).
* This is like the signal handler which runs in kernel mode, but it doesn't
* try to wake up the @task.
If find a lot of wrongs in that sentence in context of TWA_SIGNAL.
Agreed, it was hard to understand before that, but with TWA_SIGNAL it does not make sense at all.
Thanks,
tglx
Quoted 10 lines · Thomas Gleixner
Quoted 8 lines · Jens Axboe
Quoted 4 lines · Thomas Gleixner
Quoted 2 lines
- * @notify: send the notification if true
+ * @notify: send chosen notification, if anyIs that really all you found to be wrong in that comment?
There really is nothing wrong, but it's not very descriptive (wasn't before either).
Quoted 7 lines
* This is like the signal handler which runs in kernel mode, but it doesn't
* try to wake up the @task.If find a lot of wrongs in that sentence in context of
TWA_SIGNAL.Agreed, it was hard to understand before that, but with
TWA_SIGNALit does not make sense at all.
This is what I currently have:
/**
* task_work_add - ask the @task to execute @work->func()
* @task: the task which should run the callback
* @work: the callback to run
* @notify: how to notify the targeted task
*
* Queue @work for task_work_run() below and notify the @task if @notify
* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL work like signals, in that the
* it will interrupt the targeted task and run the task_work. @TWA_RESUME
* work is run only when the task exits the kernel and returns to user mode.
* Fails if the @task is exiting/exited and thus it can't process this @work.
* Otherwise @work->func() will be called when the @task returns from kernel
* mode or exits.
*
* Note: there is no ordering guarantee on works queued here.
*
* RETURNS:
* 0 if succeeds or -ESRCH.
*/
Signature · 1 line
Jens Axboe
Quoted 8 lines · Jens Axboe
/**
*task_work_add- ask the @task to execute @work->func()
* @task: the task which should run the callback
* @work: the callback to run
* @notify: how to notify the targeted task
*
* Queue @work fortask_work_run()below and notify the @task if @notify
* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL work like signals, in that the
s/the//
Quoted 5 lines
* it will interrupt the targeted task and run the
task_work. @TWA_RESUME
* work is run only when the task exits the kernel and returns to user mode.
* Fails if the @task is exiting/exited and thus it can't process this @work.
* Otherwise @work->func() will be called when the @task returns from kernel
* mode or exits.
Yes, that makes a lot more sense.
What's still lacking is a description of the return value and how to act upon it.
Most of the call sites ignore it, some are acting upon it but I can't make any sense of these actions:
fs/io_uring.c- notify = 0;fs/io_uring.c- if (!(ctx->flags & IORING_SETUP_SQPOLL) && twa_signal_ok)fs/io_uring.c- notify = TWA_SIGNAL;fs/io_uring.c-fs/io_uring.c: ret = task_work_add(tsk, &req->task_work, notify);fs/io_uring.c- if (!ret)fs/io_uring.c- wake_up_process(tsk);
???
fs/io_uring.c- if (unlikely(ret)) {fs/io_uring.c- struct task_struct *tsk;fs/io_uring.c-fs/io_uring.c- init_task_work(&req->task_work, io_req_task_cancel);fs/io_uring.c- tsk = io_wq_get_task(req->ctx->io_wq);fs/io_uring.c: task_work_add(tsk, &req->task_work, 0);fs/io_uring.c- wake_up_process(tsk);
yet more magic wakeup.
fs/io_uring.c-fs/io_uring.c- init_task_work(&req->task_work, io_req_task_submit);fs/io_uring.c- percpu_ref_get(&req->ctx->refs);fs/io_uring.c-fs/io_uring.c- /* submit ref gets dropped, acquire a new one */fs/io_uring.c- refcount_inc(&req->refs);fs/io_uring.c: ret = io_req_task_work_add(req, true);fs/io_uring.c- if (unlikely(ret)) {fs/io_uring.c- struct task_struct *tsk;fs/io_uring.c-fs/io_uring.c- /* queue just for cancelation */fs/io_uring.c- init_task_work(&req->task_work, io_req_task_cancel);fs/io_uring.c- tsk = io_wq_get_task(req->ctx->io_wq);fs/io_uring.c: task_work_add(tsk, &req->task_work, 0);fs/io_uring.c- wake_up_process(tsk);
Ditto. Why the heck is this wakeup making any sense? The initial task_work_add() within io_req_task_work_add() failed already ...
fs/io_uring.c: ret = io_req_task_work_add(req, twa_signal_ok);fs/io_uring.c- if (unlikely(ret)) {fs/io_uring.c- struct task_struct *tsk;fs/io_uring.c-fs/io_uring.c- WRITE_ONCE(poll->canceled, true);fs/io_uring.c- tsk = io_wq_get_task(req->ctx->io_wq);fs/io_uring.c: task_work_add(tsk, &req->task_work, 0);fs/io_uring.c- wake_up_process(tsk);
...
Thanks,
tglx
Quoted 11 lines · Thomas Gleixner
Quoted 8 lines · Jens Axboe
/**
*task_work_add- ask the @task to execute @work->func()
* @task: the task which should run the callback
* @work: the callback to run
* @notify: how to notify the targeted task
*
* Queue @work fortask_work_run()below and notify the @task if @notify
* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL work like signals, in that thes/the//
Thanks, good catch.
Quoted 10 lines
Quoted 5 lines
* it will interrupt the targeted task and run the
task_work. @TWA_RESUME
* work is run only when the task exits the kernel and returns to user mode.
* Fails if the @task is exiting/exited and thus it can't process this @work.
* Otherwise @work->func() will be called when the @task returns from kernel
* mode or exits.Yes, that makes a lot more sense.
What's still lacking is a description of the return value and how to act upon it.
That's really up to the caller. But we should add some explanation of that. Most callers use some alternative if the task is exiting, like using a work queue for example.
Quoted 1 line
Most of the call sites ignore it, some are acting upon it but I can't
If you know the task isn't exiting, then yeah you can ignore it. But seems a bit dicey...
Quoted 20 lines
make any sense of these actions:
fs/io_uring.c-notify = 0;fs/io_uring.c-if (!(ctx->flags &IORING_SETUP_SQPOLL) &&twa_signal_ok)fs/io_uring.c-notify =TWA_SIGNAL;fs/io_uring.c-fs/io_uring.c: ret =task_work_add(tsk, &req->task_work, notify);fs/io_uring.c-if (!ret)fs/io_uring.c-wake_up_process(tsk);???
fs/io_uring.c-if (unlikely(ret)) {fs/io_uring.c-structtask_struct*tsk;fs/io_uring.c-fs/io_uring.c-init_task_work(&req->task_work, io_req_task_cancel);fs/io_uring.c-tsk =io_wq_get_task(req->ctx->io_wq);fs/io_uring.c:task_work_add(tsk, &req->task_work, 0);fs/io_uring.c-wake_up_process(tsk);
Quoted 1 line
yet more magic wakeup.
It's not magic, but probably needs a comment... If we fail, that task is exiting. But we know we have our io-wq threads, so we use that as a fallback. Not really expected in the fast path.
Quoted 16 lines
fs/io_uring.c-fs/io_uring.c-init_task_work(&req->task_work, io_req_task_submit);fs/io_uring.c-percpu_ref_get(&req->ctx->refs);fs/io_uring.c-fs/io_uring.c-/* submit ref gets dropped, acquire a new one */fs/io_uring.c-refcount_inc(&req->refs);fs/io_uring.c: ret =io_req_task_work_add(req, true);fs/io_uring.c-if (unlikely(ret)) {fs/io_uring.c-structtask_struct*tsk;fs/io_uring.c-fs/io_uring.c-/* queue just for cancelation */fs/io_uring.c-init_task_work(&req->task_work, io_req_task_cancel);fs/io_uring.c-tsk =io_wq_get_task(req->ctx->io_wq);fs/io_uring.c:task_work_add(tsk, &req->task_work, 0);fs/io_uring.c-wake_up_process(tsk);
Quoted 2 lines
Ditto. Why the heck is this wakeup making any sense? The initial
task_work_add()withinio_req_task_work_add()failed already ...
Right, but we're using a new task for this. And that task is a kthread that we manage, hence no notification is needed outside of just waking it up.
Signature · 1 line
Jens Axboe
Quoted 3 lines · Jens Axboe
On 10/16/20 5:38 PM, Thomas Gleixner wrote:
If you know the task isn't exiting, then yeah you can ignore it. But seems a bit dicey...
Indeed.
Quoted 8 lines
Quoted 8 lines
fs/io_uring.c-if (unlikely(ret)) {fs/io_uring.c-structtask_struct*tsk;fs/io_uring.c-fs/io_uring.c-init_task_work(&req->task_work, io_req_task_cancel);fs/io_uring.c-tsk =io_wq_get_task(req->ctx->io_wq);fs/io_uring.c:task_work_add(tsk, &req->task_work, 0);fs/io_uring.c-wake_up_process(tsk);
Quoted 5 lines
Quoted 1 line
yet more magic wakeup.
It's not magic, but probably needs a comment... If we fail, that task is exiting. But we know we have our io-wq threads, so we use that as a fallback. Not really expected in the fast path.
I somehow misread it. So ignore me.
Thanks,
tglx
Quoted 15 lines · Thomas Gleixner
Quoted 3 lines · Jens Axboe
On 10/16/20 5:38 PM, Thomas Gleixner wrote:
If you know the task isn't exiting, then yeah you can ignore it. But seems a bit dicey...Indeed.
Quoted 8 lines
Quoted 8 lines
fs/io_uring.c-if (unlikely(ret)) {fs/io_uring.c-structtask_struct*tsk;fs/io_uring.c-fs/io_uring.c-init_task_work(&req->task_work, io_req_task_cancel);fs/io_uring.c-tsk =io_wq_get_task(req->ctx->io_wq);fs/io_uring.c:task_work_add(tsk, &req->task_work, 0);fs/io_uring.c-wake_up_process(tsk);
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
yet more magic wakeup.
It's not magic, but probably needs a comment... If we fail, that task is exiting. But we know we have our io-wq threads, so we use that as a fallback. Not really expected in the fast path.
I somehow misread it. So ignore me.
No worries, it probably needs a comment...
In any case, here's the current version. I added a comment on the failure case as well for task_work_add(), would be great if you could let me know if you're happy with it.
commit dc43ec64ca1374451853b53a81d5371c91ca221f
Author: Jens Axboe <[email protected]> Date: Fri Oct 16 09:02:26 2020 -0600
task_work: cleanup notification modes
A previous commit changed the notification mode from true/false to an int, allowing notify-no, notify-yes, or signal-notify. This was backwards compatible in the sense that any existing true/false user would translate to either 0 (on notification sent) or 1, the latter which mapped to TWA_RESUME. TWA_SIGNAL was assigned a value of 2.
Clean this up properly, and define a proper enum for the notification mode. Now we have:
- TWA_NONE. This is 0, same as before the original change, meaning no notification requested.
- TWA_RESUME. This is 1, same as before the original change, meaning that we use TIF_NOTIFY_RESUME.
- TWA_SIGNAL. This uses TIF_SIGPENDING/JOBCTL_TASK_WORK for the notification.
Clean up all the callers, switching their 0/1/false/true to using the appropriate TWA_* mode for notifications.
arch/x86/kernel/cpu/mce/core.c +1 -1
@@ -1277,7 +1277,7 @@ static void queue_task_work(struct mce *m, int kill_it)else current->mce_kill_me.func = kill_me_maybe; -task_work_add (current, ¤t->mce_kill_me, true); +task_work_add (current, ¤t->mce_kill_me,TWA_RESUME ); }/*
arch/x86/kernel/cpu/resctrl/rdtgroup.c +1 -1
@@ -561,7 +561,7 @@ static int __rdtgroup_move_task(struct task_struct *tsk,* callback has been invoked. */ atomic_inc (&rdtgrp->waitcount); - ret =task_work_add (tsk, &callback->work, true); + ret =task_work_add (tsk, &callback->work,TWA_RESUME );if (ret) {/* * Task is exiting. Drop the refcount and free the callback.
drivers/acpi/apei/ghes.c +1 -1
@@ -879,7 +879,7 @@ static void ghes_proc_in_irq(struct irq_work *irq_work) estatus_node->task_work.func = ghes_kick_task_work; estatus_node->task_work_cpu =smp_processor_id (); ret =task_work_add (current, &estatus_node->task_work, - true); +TWA_RESUME );if (ret) estatus_node->task_work.func =NULL ; }
drivers/android/binder.c +1 -1
@@ -2229,7 +2229,7 @@ static void binder_deferred_fd_close(int fd)__close_fd_get_file (fd, &twcb->file);if (twcb->file) {filp_close (twcb->file, current->files); -task_work_add (current, &twcb->twork, true); +task_work_add (current, &twcb->twork,TWA_RESUME ); }else {kfree (twcb); }
fs/file_table.c +1 -1
@@ -339,7 +339,7 @@ void fput_many(struct file *file, unsigned int refs)if (likely (!in_interrupt () && !(task->flags &PF_KTHREAD ))) {init_task_work (&file->f_u.fu_rcuhead, ____fput); -if (!task_work_add (task, &file->f_u.fu_rcuhead, true)) +if (!task_work_add (task, &file->f_u.fu_rcuhead,TWA_RESUME ))return ;/* * After this task has run exit_task_work(),
fs/io_uring.c +7 -6
@@ -1756,7 +1756,7 @@ static void __io_free_req(struct io_kiocb *req)struct task_struct *tsk; tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE ); } } }
@@ -1905,7 +1905,8 @@ static int io_req_task_work_add(struct io_kiocb *req, bool twa_signal_ok) {struct task_struct *tsk = req->task;struct io_ring_ctx *ctx = req->ctx; -int ret, notify; +enum task_work_notify_mode notify; +int ret;if (tsk->flags &PF_EXITING )return -ESRCH ;
@@ -1916,7 +1917,7 @@ static int io_req_task_work_add(struct io_kiocb *req, bool twa_signal_ok)* processing task_work. There's no reliable way to tell if TWA_RESUME * will do the job. */ - notify =0 ; + notify =TWA_NONE ;if (!(ctx->flags &IORING_SETUP_SQPOLL ) && twa_signal_ok) notify =TWA_SIGNAL ;
@@ -1985,7 +1986,7 @@ static void io_req_task_queue(struct io_kiocb *req)init_task_work (&req->task_work, io_req_task_cancel); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); } }
@@ -3199,7 +3200,7 @@ static int io_async_buf_func(struct wait_queue_entry *wait, unsigned mode,/* queue just for cancelation */ init_task_work (&req->task_work, io_req_task_cancel); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); }return 1 ;
@@ -4765,7 +4766,7 @@ static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,WRITE_ONCE (poll->canceled, true); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); }return 1 ;
fs/namespace.c +1 -1
@@ -1191,7 +1191,7 @@ static void mntput_no_expire(struct mount *mnt)struct task_struct *task = current;if (likely (!(task->flags &PF_KTHREAD ))) {init_task_work (&mnt->mnt_rcu, __cleanup_mnt); -if (!task_work_add (task, &mnt->mnt_rcu, true)) +if (!task_work_add (task, &mnt->mnt_rcu,TWA_RESUME ))return ; }if (llist_add (&mnt->mnt_llist, &delayed_mntput_list))
include/linux/task_work.h +8 -3
@@ -13,9 +13,14 @@ init_task_work(struct callback_head *twork, task_work_func_t func) twork->func = func; } -#define TWA_RESUME 1 -#define TWA_SIGNAL 2 -int task_work_add (struct task_struct *task,struct callback_head *twork,int ); +enum task_work_notify_mode { +TWA_NONE , +TWA_RESUME , +TWA_SIGNAL , +}; + +int task_work_add (struct task_struct *task,struct callback_head *twork, +enum task_work_notify_mode mode);struct callback_head *task_work_cancel (struct task_struct *,task_work_func_t );void task_work_run (void );
kernel/events/uprobes.c +1 -1
@@ -1823,7 +1823,7 @@ void uprobe_copy_process(struct task_struct *t, unsigned long flags) t->utask->dup_xol_addr = area->vaddr;init_task_work (&t->utask->dup_xol_work, dup_xol_work); -task_work_add (t, &t->utask->dup_xol_work, true); +task_work_add (t, &t->utask->dup_xol_work,TWA_RESUME ); }/*
kernel/irq/manage.c +1 -1
@@ -1162,7 +1162,7 @@ static int irq_thread(void *data) handler_fn = irq_thread_fn;init_task_work (&on_exit_work, irq_thread_dtor); -task_work_add (current, &on_exit_work, false); +task_work_add (current, &on_exit_work,TWA_NONE );irq_thread_check_affinity (desc, action);
kernel/sched/fair.c +1 -1
@@ -2928,7 +2928,7 @@ static void task_tick_numa(struct rq *rq, struct task_struct *curr) curr->node_stamp += period;if (!time_before (jiffies, curr->mm->numa_next_scan)) -task_work_add (curr, work, true); +task_work_add (curr, work,TWA_RESUME ); } }
kernel/task_work.c +17 -7
@@ -9,23 +9,28 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */* task_work_add - ask the @task to execute @work->func() * @task: the task which should run the callback * @work: the callback to run -* @notify: send the notification if true +* @notify: how to notify the targeted task * -* Queue @work for task_work_run() below and notify the @task if @notify. +* Queue @work for task_work_run() below and notify the @task if @notify +* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL work like signals, in that the +* it will interrupt the targeted task and run the task_work. @TWA_RESUME +* work is run only when the task exits the kernel and returns to user mode. * Fails if the @task is exiting/exited and thus it can't process this @work. * Otherwise @work->func() will be called when the @task returns from kernel * mode or exits. * -* This is like the signal handler which runs in kernel mode, but it doesn't -* try to wake up the @task. +* If the targeted task is exiting, then an error is returned and the work item +* is not queued. It's up to the caller to arrange for an alternative mechanism +* in tht case. * -* Note: there is no ordering guarantee on works queued here. +* Note: there is no ordering guarantee on works queued here. The task_work +* list is LIFO. * * RETURNS: * 0 if succeeds or -ESRCH. */ -int -task_work_add (struct task_struct *task,struct callback_head *work,int notify) +int task_work_add (struct task_struct *task,struct callback_head *work, +enum task_work_notify_mode notify) {struct callback_head *head;unsigned long flags;
@@ -38,6 +43,8 @@ task_work_add(struct task_struct *task, struct callback_head *work, int notify) }while (cmpxchg (&task->task_works, head, work) != head);switch (notify) { +case TWA_NONE : +break ;case TWA_RESUME :set_notify_resume (task);break ;
@@ -54,6 +61,9 @@ task_work_add(struct task_struct *task, struct callback_head *work, int notify)unlock_task_sighand (task, &flags); }break ; +default : +WARN_ON_ONCE (1 ); +break ; }return 0 ;
security/keys/keyctl.c +1 -1
@@ -1693,7 +1693,7 @@ long keyctl_session_to_parent(void)/* the replacement session keyring is applied just prior to userspace * restarting */ - ret =task_work_add (parent, newwork, true); + ret =task_work_add (parent, newwork,TWA_RESUME );if (!ret) newwork =NULL ; unlock:
security/yama/yama_lsm.c +1 -1
@@ -99,7 +99,7 @@ static void report_access(const char *access, struct task_struct *target, info->access = access; info->target = target; info->agent = agent; -if (task_work_add (current, &info->work, true) ==0 ) +if (task_work_add (current, &info->work,TWA_RESUME ) ==0 )return ;/* success */ WARN (1 ,"report_access called from exiting task" );
Signature · 1 line
Jens Axboe
Quoted 5 lines · Jens Axboe
kernel/task_work.c +0
-0
kernel/task_work.c static struct callback_head work_exited; /* all we need is ->next == NULL */ · line 15
Quoted 9 lines
kernel/task_work.c +3
-2@@ -9,23 +9,28 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */* task_work_add - ask the @task to execute @work->func() * @task: the task which should run the callback * @work: the callback to run -* @notify: send the notification if true +* @notify: how to notify the targeted task * -* Queue @work for task_work_run() below and notify the @task if @notify. +* Queue @work for task_work_run() below and notify the @task if @notify +* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL work like signals, in that the
s/work/works/
kernel/task_work.c static struct callback_head work_exited; /* all we need is ->next == NULL */ · line 17
Quoted 2 lines
kernel/task_work.c +2
-0+* it will interrupt the targeted task and run the task_work. @TWA_RESUME +* work is run only when the task exits the kernel and returns to user mode.
It's also run before entering a guest. At least on x86, but all other architectures should do the same.
kernel/task_work.c static struct callback_head work_exited; /* all we need is ->next == NULL */ · line 24
Quoted 9 lines
kernel/task_work.c +3
-2* Fails if the @task is exiting/exited and thus it can't process this @work. * Otherwise @work->func() will be called when the @task returns from kernel * mode or exits. * -* This is like the signal handler which runs in kernel mode, but it doesn't -* try to wake up the @task. +* If the targeted task is exiting, then an error is returned and the work item +* is not queued. It's up to the caller to arrange for an alternative mechanism +* in tht case.
s/tht/that/
Looks good otherwise.
Thanks,
tglx
Quoted 34 lines · Thomas Gleixner
Quoted 5 lines · Jens Axboe
kernel/task_work.c +0
-0kernel/task_work.c static struct callback_head work_exited; /* all we need is ->next == NULL */ · line 15
Quoted 9 lines
kernel/task_work.c +3
-2@@ -9,23 +9,28 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */* task_work_add - ask the @task to execute @work->func() * @task: the task which should run the callback * @work: the callback to run -* @notify: send the notification if true +* @notify: how to notify the targeted task * -* Queue @work for task_work_run() below and notify the @task if @notify. +* Queue @work for task_work_run() below and notify the @task if @notify +* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL work like signals, in that the s/work/works/
kernel/task_work.c static struct callback_head work_exited; /* all we need is ->next == NULL */ · line 17
Quoted 2 lines
kernel/task_work.c +2
-0+* it will interrupt the targeted task and run the task_work. @TWA_RESUME +* work is run only when the task exits the kernel and returns to user mode. It's also run before entering a guest. At least on x86, but all other architectures should do the same.
kernel/task_work.c static struct callback_head work_exited; /* all we need is ->next == NULL */ · line 24
Quoted 9 lines
kernel/task_work.c +3
-2* Fails if the @task is exiting/exited and thus it can't process this @work. * Otherwise @work->func() will be called when the @task returns from kernel * mode or exits. * -* This is like the signal handler which runs in kernel mode, but it doesn't -* try to wake up the @task. +* If the targeted task is exiting, then an error is returned and the work item +* is not queued. It's up to the caller to arrange for an alternative mechanism +* in tht case.
Quoted 3 lines
s/tht/that/
Looks good otherwise.
Thanks! Fixed the two typos, and also included the guest mode. If you're fine with it now, would be great to have your reviewed-by or similar.
commit 4799fc5b3bc3520a61d3afd714d11672e45c81cb
Author: Jens Axboe <[email protected]> Date: Fri Oct 16 09:02:26 2020 -0600
task_work: cleanup notification modes
A previous commit changed the notification mode from true/false to an int, allowing notify-no, notify-yes, or signal-notify. This was backwards compatible in the sense that any existing true/false user would translate to either 0 (on notification sent) or 1, the latter which mapped to TWA_RESUME. TWA_SIGNAL was assigned a value of 2.
Clean this up properly, and define a proper enum for the notification mode. Now we have:
- TWA_NONE. This is 0, same as before the original change, meaning no notification requested.
- TWA_RESUME. This is 1, same as before the original change, meaning that we use TIF_NOTIFY_RESUME.
- TWA_SIGNAL. This uses TIF_SIGPENDING/JOBCTL_TASK_WORK for the notification.
Clean up all the callers, switching their 0/1/false/true to using the appropriate TWA_* mode for notifications.
arch/x86/kernel/cpu/mce/core.c +1 -1
@@ -1277,7 +1277,7 @@ static void queue_task_work(struct mce *m, int kill_it)else current->mce_kill_me.func = kill_me_maybe; -task_work_add (current, ¤t->mce_kill_me, true); +task_work_add (current, ¤t->mce_kill_me,TWA_RESUME ); }/*
arch/x86/kernel/cpu/resctrl/rdtgroup.c +1 -1
@@ -561,7 +561,7 @@ static int __rdtgroup_move_task(struct task_struct *tsk,* callback has been invoked. */ atomic_inc (&rdtgrp->waitcount); - ret =task_work_add (tsk, &callback->work, true); + ret =task_work_add (tsk, &callback->work,TWA_RESUME );if (ret) {/* * Task is exiting. Drop the refcount and free the callback.
drivers/acpi/apei/ghes.c +1 -1
@@ -879,7 +879,7 @@ static void ghes_proc_in_irq(struct irq_work *irq_work) estatus_node->task_work.func = ghes_kick_task_work; estatus_node->task_work_cpu =smp_processor_id (); ret =task_work_add (current, &estatus_node->task_work, - true); +TWA_RESUME );if (ret) estatus_node->task_work.func =NULL ; }
drivers/android/binder.c +1 -1
@@ -2229,7 +2229,7 @@ static void binder_deferred_fd_close(int fd)__close_fd_get_file (fd, &twcb->file);if (twcb->file) {filp_close (twcb->file, current->files); -task_work_add (current, &twcb->twork, true); +task_work_add (current, &twcb->twork,TWA_RESUME ); }else {kfree (twcb); }
fs/file_table.c +1 -1
@@ -339,7 +339,7 @@ void fput_many(struct file *file, unsigned int refs)if (likely (!in_interrupt () && !(task->flags &PF_KTHREAD ))) {init_task_work (&file->f_u.fu_rcuhead, ____fput); -if (!task_work_add (task, &file->f_u.fu_rcuhead, true)) +if (!task_work_add (task, &file->f_u.fu_rcuhead,TWA_RESUME ))return ;/* * After this task has run exit_task_work(),
fs/io_uring.c +7 -6
@@ -1756,7 +1756,7 @@ static void __io_free_req(struct io_kiocb *req)struct task_struct *tsk; tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE ); } } }
@@ -1905,7 +1905,8 @@ static int io_req_task_work_add(struct io_kiocb *req, bool twa_signal_ok) {struct task_struct *tsk = req->task;struct io_ring_ctx *ctx = req->ctx; -int ret, notify; +enum task_work_notify_mode notify; +int ret;if (tsk->flags &PF_EXITING )return -ESRCH ;
@@ -1916,7 +1917,7 @@ static int io_req_task_work_add(struct io_kiocb *req, bool twa_signal_ok)* processing task_work. There's no reliable way to tell if TWA_RESUME * will do the job. */ - notify =0 ; + notify =TWA_NONE ;if (!(ctx->flags &IORING_SETUP_SQPOLL ) && twa_signal_ok) notify =TWA_SIGNAL ;
@@ -1985,7 +1986,7 @@ static void io_req_task_queue(struct io_kiocb *req)init_task_work (&req->task_work, io_req_task_cancel); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); } }
@@ -3199,7 +3200,7 @@ static int io_async_buf_func(struct wait_queue_entry *wait, unsigned mode,/* queue just for cancelation */ init_task_work (&req->task_work, io_req_task_cancel); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); }return 1 ;
@@ -4765,7 +4766,7 @@ static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,WRITE_ONCE (poll->canceled, true); tsk =io_wq_get_task (req->ctx->io_wq); -task_work_add (tsk, &req->task_work,0 ); +task_work_add (tsk, &req->task_work,TWA_NONE );wake_up_process (tsk); }return 1 ;
fs/namespace.c +1 -1
@@ -1191,7 +1191,7 @@ static void mntput_no_expire(struct mount *mnt)struct task_struct *task = current;if (likely (!(task->flags &PF_KTHREAD ))) {init_task_work (&mnt->mnt_rcu, __cleanup_mnt); -if (!task_work_add (task, &mnt->mnt_rcu, true)) +if (!task_work_add (task, &mnt->mnt_rcu,TWA_RESUME ))return ; }if (llist_add (&mnt->mnt_llist, &delayed_mntput_list))
include/linux/task_work.h +8 -3
@@ -13,9 +13,14 @@ init_task_work(struct callback_head *twork, task_work_func_t func) twork->func = func; } -#define TWA_RESUME 1 -#define TWA_SIGNAL 2 -int task_work_add (struct task_struct *task,struct callback_head *twork,int ); +enum task_work_notify_mode { +TWA_NONE , +TWA_RESUME , +TWA_SIGNAL , +}; + +int task_work_add (struct task_struct *task,struct callback_head *twork, +enum task_work_notify_mode mode);struct callback_head *task_work_cancel (struct task_struct *,task_work_func_t );void task_work_run (void );
kernel/events/uprobes.c +1 -1
@@ -1823,7 +1823,7 @@ void uprobe_copy_process(struct task_struct *t, unsigned long flags) t->utask->dup_xol_addr = area->vaddr;init_task_work (&t->utask->dup_xol_work, dup_xol_work); -task_work_add (t, &t->utask->dup_xol_work, true); +task_work_add (t, &t->utask->dup_xol_work,TWA_RESUME ); }/*
kernel/irq/manage.c +1 -1
@@ -1162,7 +1162,7 @@ static int irq_thread(void *data) handler_fn = irq_thread_fn;init_task_work (&on_exit_work, irq_thread_dtor); -task_work_add (current, &on_exit_work, false); +task_work_add (current, &on_exit_work,TWA_NONE );irq_thread_check_affinity (desc, action);
kernel/sched/fair.c +1 -1
@@ -2928,7 +2928,7 @@ static void task_tick_numa(struct rq *rq, struct task_struct *curr) curr->node_stamp += period;if (!time_before (jiffies, curr->mm->numa_next_scan)) -task_work_add (curr, work, true); +task_work_add (curr, work,TWA_RESUME ); } }
kernel/task_work.c +20 -10
@@ -9,23 +9,28 @@ static struct callback_head work_exited; /* all we need is ->next == NULL */* task_work_add - ask the @task to execute @work->func() * @task: the task which should run the callback * @work: the callback to run -* @notify: send the notification if true +* @notify: how to notify the targeted task * -* Queue @work for task_work_run() below and notify the @task if @notify. -* Fails if the @task is exiting/exited and thus it can't process this @work. -* Otherwise @work->func() will be called when the @task returns from kernel -* mode or exits. +* Queue @work for task_work_run() below and notify the @task if @notify +* is @TWA_RESUME or @TWA_SIGNAL. @TWA_SIGNAL works like signals, in that the +* it will interrupt the targeted task and run the task_work. @TWA_RESUME +* work is run only when the task exits the kernel and returns to user mode, +* or before entering guest mode. Fails if the @task is exiting/exited and thus +* it can't process this @work. Otherwise @work->func() will be called when the +* @task goes through one of the aforementioned transitions, or exits. * -* This is like the signal handler which runs in kernel mode, but it doesn't -* try to wake up the @task. +* If the targeted task is exiting, then an error is returned and the work item +* is not queued. It's up to the caller to arrange for an alternative mechanism +* in that case. * -* Note: there is no ordering guarantee on works queued here. +* Note: there is no ordering guarantee on works queued here. The task_work +* list is LIFO. * * RETURNS: * 0 if succeeds or -ESRCH. */ -int -task_work_add (struct task_struct *task,struct callback_head *work,int notify) +int task_work_add (struct task_struct *task,struct callback_head *work, +enum task_work_notify_mode notify) {struct callback_head *head;unsigned long flags;
@@ -38,6 +43,8 @@ task_work_add(struct task_struct *task, struct callback_head *work, int notify) }while (cmpxchg (&task->task_works, head, work) != head);switch (notify) { +case TWA_NONE : +break ;case TWA_RESUME :set_notify_resume (task);break ;
@@ -54,6 +61,9 @@ task_work_add(struct task_struct *task, struct callback_head *work, int notify)unlock_task_sighand (task, &flags); }break ; +default : +WARN_ON_ONCE (1 ); +break ; }return 0 ;
security/keys/keyctl.c +1 -1
@@ -1693,7 +1693,7 @@ long keyctl_session_to_parent(void)/* the replacement session keyring is applied just prior to userspace * restarting */ - ret =task_work_add (parent, newwork, true); + ret =task_work_add (parent, newwork,TWA_RESUME );if (!ret) newwork =NULL ; unlock:
security/yama/yama_lsm.c +1 -1
@@ -99,7 +99,7 @@ static void report_access(const char *access, struct task_struct *target, info->access = access; info->target = target; info->agent = agent; -if (task_work_add (current, &info->work, true) ==0 ) +if (task_work_add (current, &info->work,TWA_RESUME ) ==0 )return ;/* success */ WARN (1 ,"report_access called from exiting task" );
Signature · 1 line
Jens Axboe
Thomas Gleixner
Reviewed-byJens,
Quoted 4 lines · Jens Axboe
On 10/17/20 2:18 PM, Thomas Gleixner wrote:
Thanks! Fixed the two typos, and also included the guest mode. If you're fine with it now, would be great to have your reviewed-by or similar.
Sure. I assume you ship it to Linus, otherwise let me know and I'll pick it up.
Quoted 10 lines · Thomas Gleixner
Jens,
Quoted 4 lines · Jens Axboe
On 10/17/20 2:18 PM, Thomas Gleixner wrote:
Thanks! Fixed the two typos, and also included the guest mode. If you're fine with it now, would be great to have your reviewed-by or similar.
Sure. I assume you ship it to Linus, otherwise let me know and I'll pick it up.
I can, I have it bundled up with the TIF_NOTIFY_RESUME cleanup. Either way is fine with me, so if you're good with it, I'll ship it before -rc1.
Quoted 1 line
Reviewed-by Thomas Gleixner <[email protected]>
Thanks!
Signature · 1 line
Jens Axboe
Quoted 6 lines · Jens Axboe
Quoted 2 lines · Thomas Gleixner
Sure. I assume you ship it to Linus, otherwise let me know and I'll pick it up.
I can, I have it bundled up with the
TIF_NOTIFY_RESUMEcleanup. Either way is fine with me, so if you're good with it, I'll ship it before -rc1.
Works for me.