LinuxLists.cc - [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

by tip-bot2 for Alexey Makhalov

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Sun, Sep 04, 2022 at 12:09:37PM +0200, Ingo Molnar wrote:

> BTW., we should probably mark/document all PF_ holes with a PF__RESERVED
> kind of scheme? Something simple, like:
>
> #define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */
> #define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */
> + #define PF__RESERVED_04000 0x00004000 /* Unused */
> #define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */
> + #define PF__RESERVED_10000 0x00010000 /* Unused */
> #define PF_KSWAPD 0x00020000 /* I am kswapd */
> #define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */
> #define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */

How's this then, it immediately shows how holey it is :-)

--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1722,7 +1722,9 @@ extern struct pid *cad_pid;
#define PF_MEMALLOC 0x00000800 /* Allocating memory */
#define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */
#define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */
+#define PF__HOLE__00004000 0x00004000 /* A HOLE */
#define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */
+#define PF__HOLE__00010000 0x00010000 /* A HOLE */
#define PF_KSWAPD 0x00020000 /* I am kswapd */
#define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */
#define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */
@@ -1730,9 +1732,14 @@ extern struct pid *cad_pid;
* I am cleaning dirty pages from some other bdi. */
#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
#define PF_RANDOMIZE 0x00400000 /* Randomize virtual address space */
+#define PF__HOLE__00800000 0x00800000 /* A HOLE */
+#define PF__HOLE__01000000 0x01000000 /* A HOLE */
+#define PF__HOLE__02000000 0x02000000 /* A HOLE */
#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_mask */
#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
#define PF_MEMALLOC_PIN 0x10000000 /* Allocation context constrained to zones which allow long term pinning. */
+#define PF__HOLE__20000000 0x20000000 /* A HOLE */
+#define PF__HOLE__40000000 0x40000000 /* A HOLE */
#define PF_SUSPEND_TASK 0x80000000 /* This thread called freeze_processes() and should not be frozen */

/*

2022-09-07 07:44:31

by Ingo Molnar

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

* Peter Zijlstra <[email protected]> wrote:

> On Sun, Sep 04, 2022 at 12:09:37PM +0200, Ingo Molnar wrote:
>
> > BTW., we should probably mark/document all PF_ holes with a PF__RESERVED
> > kind of scheme? Something simple, like:
> >
> > #define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */
> > #define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */
> > + #define PF__RESERVED_04000 0x00004000 /* Unused */
> > #define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */
> > + #define PF__RESERVED_10000 0x00010000 /* Unused */
> > #define PF_KSWAPD 0x00020000 /* I am kswapd */
> > #define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */
> > #define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */
>
>
> How's this then, it immediately shows how holey it is :-)
>
> --- a/include/linux/sched.h
> +++ b/include/linux/sched.h
> @@ -1722,7 +1722,9 @@ extern struct pid *cad_pid;
> #define PF_MEMALLOC 0x00000800 /* Allocating memory */
> #define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */
> #define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */
> +#define PF__HOLE__00004000 0x00004000 /* A HOLE */
> #define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */
> +#define PF__HOLE__00010000 0x00010000 /* A HOLE */
> #define PF_KSWAPD 0x00020000 /* I am kswapd */
> #define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */
> #define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */
> @@ -1730,9 +1732,14 @@ extern struct pid *cad_pid;
> * I am cleaning dirty pages from some other bdi. */
> #define PF_KTHREAD 0x00200000 /* I am a kernel thread */
> #define PF_RANDOMIZE 0x00400000 /* Randomize virtual address space */
> +#define PF__HOLE__00800000 0x00800000 /* A HOLE */
> +#define PF__HOLE__01000000 0x01000000 /* A HOLE */
> +#define PF__HOLE__02000000 0x02000000 /* A HOLE */
> #define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_mask */
> #define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
> #define PF_MEMALLOC_PIN 0x10000000 /* Allocation context constrained to zones which allow long term pinning. */
> +#define PF__HOLE__20000000 0x20000000 /* A HOLE */
> +#define PF__HOLE__40000000 0x40000000 /* A HOLE */
> #define PF_SUSPEND_TASK 0x80000000 /* This thread called freeze_processes() and should not be frozen */

LGTM - OTOH this looks quite a bit more cluttery than I imagined it in my
head. :-/ So I'd leave out the comment part at minimum. With that:

Reviewed-by: Ingo Molnar <[email protected]>

Thanks,

Ingo

2022-09-09 09:23:04

[permalink] [raw]

Subject: [tip: sched/core] freezer,sched: Rewrite core freezer logic

The following commit has been merged into the sched/core branch of tip:

Commit-ID: f5d39b020809146cc28e6e73369bf8065e0310aa
Gitweb: https://git.kernel.org/tip/f5d39b020809146cc28e6e73369bf8065e0310aa
Author: Peter Zijlstra <[email protected]>
AuthorDate: Mon, 22 Aug 2022 13:18:22 +02:00
Committer: Peter Zijlstra <[email protected]>
CommitterDate: Wed, 07 Sep 2022 21:53:50 +02:00

freezer,sched: Rewrite core freezer logic

Rewrite the core freezer to behave better wrt thawing and be simpler
in general.

By replacing PF_FROZEN with TASK_FROZEN, a special block state, it is
ensured frozen tasks stay frozen until thawed and don't randomly wake
up early, as is currently possible.

As such, it does away with PF_FROZEN and PF_FREEZER_SKIP, freeing up
two PF_flags (yay!).

Specifically; the current scheme works a little like:

freezer_do_not_count();
schedule();
freezer_count();

And either the task is blocked, or it lands in try_to_freezer()
through freezer_count(). Now, when it is blocked, the freezer
considers it frozen and continues.

However, on thawing, once pm_freezing is cleared, freezer_count()
stops working, and any random/spurious wakeup will let a task run
before its time.

That is, thawing tries to thaw things in explicit order; kernel
threads and workqueues before doing bringing SMP back before userspace
etc.. However due to the above mentioned races it is entirely possible
for userspace tasks to thaw (by accident) before SMP is back.

This can be a fatal problem in asymmetric ISA architectures (eg ARMv9)
where the userspace task requires a special CPU to run.

As said; replace this with a special task state TASK_FROZEN and add
the following state transitions:

TASK_FREEZABLE -> TASK_FROZEN
__TASK_STOPPED -> TASK_FROZEN
__TASK_TRACED -> TASK_FROZEN

The new TASK_FREEZABLE can be set on any state part of TASK_NORMAL
(IOW. TASK_INTERRUPTIBLE and TASK_UNINTERRUPTIBLE) -- any such state
is already required to deal with spurious wakeups and the freezer
causes one such when thawing the task (since the original state is
lost).

The special __TASK_{STOPPED,TRACED} states *can* be restored since
their canonical state is in ->jobctl.

With this, frozen tasks need an explicit TASK_FROZEN wakeup and are
free of undue (early / spurious) wakeups.

Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Reviewed-by: Ingo Molnar <[email protected]>
Acked-by: Rafael J. Wysocki <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
drivers/android/binder.c | 4 +-
drivers/media/pci/pt3/pt3.c | 4 +-
fs/cifs/inode.c | 4 +-
fs/cifs/transport.c | 5 +-
fs/coredump.c | 5 +-
fs/nfs/file.c | 3 +-
fs/nfs/inode.c | 12 +--
fs/nfs/nfs3proc.c | 3 +-
fs/nfs/nfs4proc.c | 14 +-
fs/nfs/nfs4state.c | 3 +-
fs/nfs/pnfs.c | 4 +-
fs/xfs/xfs_trans_ail.c | 8 +-
include/linux/freezer.h | 245 +-------------------------------
include/linux/sched.h | 13 +-
include/linux/sunrpc/sched.h | 7 +-
include/linux/wait.h | 12 +-
kernel/cgroup/legacy_freezer.c | 23 +--
kernel/exit.c | 4 +-
kernel/fork.c | 5 +-
kernel/freezer.c | 133 ++++++++++++-----
kernel/futex/waitwake.c | 8 +-
kernel/hung_task.c | 4 +-
kernel/power/main.c | 6 +-
kernel/power/process.c | 10 +-
kernel/ptrace.c | 2 +-
kernel/sched/core.c | 2 +-
kernel/signal.c | 14 +-
kernel/time/hrtimer.c | 4 +-
kernel/umh.c | 20 +--
mm/khugepaged.c | 4 +-
net/sunrpc/sched.c | 12 +--
net/unix/af_unix.c | 8 +-
32 files changed, 210 insertions(+), 395 deletions(-)

diff --git a/drivers/android/binder.c b/drivers/android/binder.c
index c964d7c..50197ef 100644
--- a/drivers/android/binder.c
+++ b/drivers/android/binder.c
@@ -4247,10 +4247,9 @@ static int binder_wait_for_work(struct binder_thread *thread,
struct binder_proc *proc = thread->proc;
int ret = 0;

- freezer_do_not_count();
binder_inner_proc_lock(proc);
for (;;) {
- prepare_to_wait(&thread->wait, &wait, TASK_INTERRUPTIBLE);
+ prepare_to_wait(&thread->wait, &wait, TASK_INTERRUPTIBLE|TASK_FREEZABLE);
if (binder_has_work_ilocked(thread, do_proc_work))
break;
if (do_proc_work)
@@ -4267,7 +4266,6 @@ static int binder_wait_for_work(struct binder_thread *thread,
}
finish_wait(&thread->wait, &wait);
binder_inner_proc_unlock(proc);
- freezer_count();

return ret;
}
diff --git a/drivers/media/pci/pt3/pt3.c b/drivers/media/pci/pt3/pt3.c
index 0d51bdf..f6deac8 100644
--- a/drivers/media/pci/pt3/pt3.c
+++ b/drivers/media/pci/pt3/pt3.c
@@ -445,8 +445,8 @@ static int pt3_fetch_thread(void *data)
pt3_proc_dma(adap);

delay = ktime_set(0, PT3_FETCH_DELAY * NSEC_PER_MSEC);
- set_current_state(TASK_UNINTERRUPTIBLE);
- freezable_schedule_hrtimeout_range(&delay,
+ set_current_state(TASK_UNINTERRUPTIBLE|TASK_FREEZABLE);
+ schedule_hrtimeout_range(&delay,
PT3_FETCH_DELAY_DELTA * NSEC_PER_MSEC,
HRTIMER_MODE_REL);
}
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index eeeaba3..54e6bfc 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -2326,7 +2326,7 @@ cifs_invalidate_mapping(struct inode *inode)
static int
cifs_wait_bit_killable(struct wait_bit_key *key, int mode)
{
- freezable_schedule_unsafe();
+ schedule();
if (signal_pending_state(mode, current))
return -ERESTARTSYS;
return 0;
@@ -2344,7 +2344,7 @@ cifs_revalidate_mapping(struct inode *inode)
return 0;

rc = wait_on_bit_lock_action(flags, CIFS_INO_LOCK, cifs_wait_bit_killable,
- TASK_KILLABLE);
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
if (rc)
return rc;

diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c
index de7aece..eda09fb 100644
--- a/fs/cifs/transport.c
+++ b/fs/cifs/transport.c
@@ -757,8 +757,9 @@ wait_for_response(struct TCP_Server_Info *server, struct mid_q_entry *midQ)
{
int error;

- error = wait_event_freezekillable_unsafe(server->response_q,
- midQ->mid_state != MID_REQUEST_SUBMITTED);
+ error = wait_event_state(server->response_q,
+ midQ->mid_state != MID_REQUEST_SUBMITTED,
+ (TASK_KILLABLE|TASK_FREEZABLE_UNSAFE));
if (error < 0)
return -ERESTARTSYS;

diff --git a/fs/coredump.c b/fs/coredump.c
index 3d9054c..9ce59f2 100644
--- a/fs/coredump.c
+++ b/fs/coredump.c
@@ -402,9 +402,8 @@ static int coredump_wait(int exit_code, struct core_state *core_state)
if (core_waiters > 0) {
struct core_thread *ptr;

- freezer_do_not_count();
- wait_for_completion(&core_state->startup);
- freezer_count();
+ wait_for_completion_state(&core_state->startup,
+ TASK_UNINTERRUPTIBLE|TASK_FREEZABLE);
/*
* Wait for all the threads to become inactive, so that
* all the thread context (extended register state, like
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index d2bcd48..e8301a3 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -570,7 +570,8 @@ static vm_fault_t nfs_vm_page_mkwrite(struct vm_fault *vmf)
}

wait_on_bit_action(&NFS_I(inode)->flags, NFS_INO_INVALIDATING,
- nfs_wait_bit_killable, TASK_KILLABLE);
+ nfs_wait_bit_killable,
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);

lock_page(page);
mapping = page_file_mapping(page);
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index b4e46b0..a6058f5 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -72,18 +72,13 @@ nfs_fattr_to_ino_t(struct nfs_fattr *fattr)
return nfs_fileid_to_ino_t(fattr->fileid);
}

-static int nfs_wait_killable(int mode)
+int nfs_wait_bit_killable(struct wait_bit_key *key, int mode)
{
- freezable_schedule_unsafe();
+ schedule();
if (signal_pending_state(mode, current))
return -ERESTARTSYS;
return 0;
}
-
-int nfs_wait_bit_killable(struct wait_bit_key *key, int mode)
-{
- return nfs_wait_killable(mode);
-}
EXPORT_SYMBOL_GPL(nfs_wait_bit_killable);

/**
@@ -1331,7 +1326,8 @@ int nfs_clear_invalid_mapping(struct address_space *mapping)
*/
for (;;) {
ret = wait_on_bit_action(bitlock, NFS_INO_INVALIDATING,
- nfs_wait_bit_killable, TASK_KILLABLE);
+ nfs_wait_bit_killable,
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
if (ret)
goto out;
spin_lock(&inode->i_lock);
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index 1597eef..2e75796 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -36,7 +36,8 @@ nfs3_rpc_wrapper(struct rpc_clnt *clnt, struct rpc_message *msg, int flags)
res = rpc_call_sync(clnt, msg, flags);
if (res != -EJUKEBOX)
break;
- freezable_schedule_timeout_killable_unsafe(NFS_JUKEBOX_RETRY_TIME);
+ __set_current_state(TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
+ schedule_timeout(NFS_JUKEBOX_RETRY_TIME);
res = -ERESTARTSYS;
} while (!fatal_signal_pending(current));
return res;
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 3ed14a2..4553803 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -416,8 +416,8 @@ static int nfs4_delay_killable(long *timeout)
{
might_sleep();

- freezable_schedule_timeout_killable_unsafe(
- nfs4_update_delay(timeout));
+ __set_current_state(TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
+ schedule_timeout(nfs4_update_delay(timeout));
if (!__fatal_signal_pending(current))
return 0;
return -EINTR;
@@ -427,7 +427,8 @@ static int nfs4_delay_interruptible(long *timeout)
{
might_sleep();

- freezable_schedule_timeout_interruptible_unsafe(nfs4_update_delay(timeout));
+ __set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE_UNSAFE);
+ schedule_timeout(nfs4_update_delay(timeout));
if (!signal_pending(current))
return 0;
return __fatal_signal_pending(current) ? -EINTR :-ERESTARTSYS;
@@ -7406,7 +7407,8 @@ nfs4_retry_setlk_simple(struct nfs4_state *state, int cmd,
status = nfs4_proc_setlk(state, cmd, request);
if ((status != -EAGAIN) || IS_SETLK(cmd))
break;
- freezable_schedule_timeout_interruptible(timeout);
+ __set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
+ schedule_timeout(timeout);
timeout *= 2;
timeout = min_t(unsigned long, NFS4_LOCK_MAXTIMEOUT, timeout);
status = -ERESTARTSYS;
@@ -7474,10 +7476,8 @@ nfs4_retry_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
break;

status = -ERESTARTSYS;
- freezer_do_not_count();
- wait_woken(&waiter.wait, TASK_INTERRUPTIBLE,
+ wait_woken(&waiter.wait, TASK_INTERRUPTIBLE|TASK_FREEZABLE,
NFS4_LOCK_MAXTIMEOUT);
- freezer_count();
} while (!signalled());

remove_wait_queue(q, &waiter.wait);
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 9bab3e9..7e185f7 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -1314,7 +1314,8 @@ int nfs4_wait_clnt_recover(struct nfs_client *clp)

refcount_inc(&clp->cl_count);
res = wait_on_bit_action(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING,
- nfs_wait_bit_killable, TASK_KILLABLE);
+ nfs_wait_bit_killable,
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
if (res)
goto out;
if (clp->cl_cons_state < 0)
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 41a9b6b..70706db 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1908,7 +1908,7 @@ static int pnfs_prepare_to_retry_layoutget(struct pnfs_layout_hdr *lo)
pnfs_layoutcommit_inode(lo->plh_inode, false);
return wait_on_bit_action(&lo->plh_flags, NFS_LAYOUT_RETURN,
nfs_wait_bit_killable,
- TASK_KILLABLE);
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
}

static void nfs_layoutget_begin(struct pnfs_layout_hdr *lo)
@@ -3193,7 +3193,7 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
status = wait_on_bit_lock_action(&nfsi->flags,
NFS_INO_LAYOUTCOMMITTING,
nfs_wait_bit_killable,
- TASK_KILLABLE);
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
if (status)
goto out;
}
diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index d3a97a0..16fbf2a 100644
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -602,9 +602,9 @@ xfsaild(

while (1) {
if (tout && tout <= 20)
- set_current_state(TASK_KILLABLE);
+ set_current_state(TASK_KILLABLE|TASK_FREEZABLE);
else
- set_current_state(TASK_INTERRUPTIBLE);
+ set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);

/*
* Check kthread_should_stop() after we set the task state to
@@ -653,14 +653,14 @@ xfsaild(
ailp->ail_target == ailp->ail_target_prev &&
list_empty(&ailp->ail_buf_list)) {
spin_unlock(&ailp->ail_lock);
- freezable_schedule();
+ schedule();
tout = 0;
continue;
}
spin_unlock(&ailp->ail_lock);

if (tout)
- freezable_schedule_timeout(msecs_to_jiffies(tout));
+ schedule_timeout(msecs_to_jiffies(tout));

__set_current_state(TASK_RUNNING);

diff --git a/include/linux/freezer.h b/include/linux/freezer.h
index 0621c5f..b303472 100644
--- a/include/linux/freezer.h
+++ b/include/linux/freezer.h
@@ -8,9 +8,11 @@
#include <linux/sched.h>
#include <linux/wait.h>
#include <linux/atomic.h>
+#include <linux/jump_label.h>

#ifdef CONFIG_FREEZER
-extern atomic_t system_freezing_cnt; /* nr of freezing conds in effect */
+DECLARE_STATIC_KEY_FALSE(freezer_active);
+
extern bool pm_freezing; /* PM freezing in effect */
extern bool pm_nosig_freezing; /* PM nosig freezing in effect */

@@ -22,10 +24,7 @@ extern unsigned int freeze_timeout_msecs;
/*
* Check if a process has been frozen
*/
-static inline bool frozen(struct task_struct *p)
-{
- return p->flags & PF_FROZEN;
-}
+extern bool frozen(struct task_struct *p);

extern bool freezing_slow_path(struct task_struct *p);

@@ -34,9 +33,10 @@ extern bool freezing_slow_path(struct task_struct *p);
*/
static inline bool freezing(struct task_struct *p)
{
- if (likely(!atomic_read(&system_freezing_cnt)))
- return false;
- return freezing_slow_path(p);
+ if (static_branch_unlikely(&freezer_active))
+ return freezing_slow_path(p);
+
+ return false;
}

/* Takes and releases task alloc lock using task_lock() */
@@ -48,23 +48,14 @@ extern int freeze_kernel_threads(void);
extern void thaw_processes(void);
extern void thaw_kernel_threads(void);

-/*
- * DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION
- * If try_to_freeze causes a lockdep warning it means the caller may deadlock
- */
-static inline bool try_to_freeze_unsafe(void)
+static inline bool try_to_freeze(void)
{
might_sleep();
if (likely(!freezing(current)))
return false;
- return __refrigerator(false);
-}
-
-static inline bool try_to_freeze(void)
-{
if (!(current->flags & PF_NOFREEZE))
debug_check_no_locks_held();
- return try_to_freeze_unsafe();
+ return __refrigerator(false);
}

extern bool freeze_task(struct task_struct *p);
@@ -79,195 +70,6 @@ static inline bool cgroup_freezing(struct task_struct *task)
}
#endif /* !CONFIG_CGROUP_FREEZER */

-/*
- * The PF_FREEZER_SKIP flag should be set by a vfork parent right before it
- * calls wait_for_completion(&vfork) and reset right after it returns from this
- * function. Next, the parent should call try_to_freeze() to freeze itself
- * appropriately in case the child has exited before the freezing of tasks is
- * complete. However, we don't want kernel threads to be frozen in unexpected
- * places, so we allow them to block freeze_processes() instead or to set
- * PF_NOFREEZE if needed. Fortunately, in the ____call_usermodehelper() case the
- * parent won't really block freeze_processes(), since ____call_usermodehelper()
- * (the child) does a little before exec/exit and it can't be frozen before
- * waking up the parent.
- */
-
-
-/**
- * freezer_do_not_count - tell freezer to ignore %current
- *
- * Tell freezers to ignore the current task when determining whether the
- * target frozen state is reached. IOW, the current task will be
- * considered frozen enough by freezers.
- *
- * The caller shouldn't do anything which isn't allowed for a frozen task
- * until freezer_cont() is called. Usually, freezer[_do_not]_count() pair
- * wrap a scheduling operation and nothing much else.
- */
-static inline void freezer_do_not_count(void)
-{
- current->flags |= PF_FREEZER_SKIP;
-}
-
-/**
- * freezer_count - tell freezer to stop ignoring %current
- *
- * Undo freezer_do_not_count(). It tells freezers that %current should be
- * considered again and tries to freeze if freezing condition is already in
- * effect.
- */
-static inline void freezer_count(void)
-{
- current->flags &= ~PF_FREEZER_SKIP;
- /*
- * If freezing is in progress, the following paired with smp_mb()
- * in freezer_should_skip() ensures that either we see %true
- * freezing() or freezer_should_skip() sees !PF_FREEZER_SKIP.
- */
- smp_mb();
- try_to_freeze();
-}
-
-/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-static inline void freezer_count_unsafe(void)
-{
- current->flags &= ~PF_FREEZER_SKIP;
- smp_mb();
- try_to_freeze_unsafe();
-}
-
-/**
- * freezer_should_skip - whether to skip a task when determining frozen
- * state is reached
- * @p: task in quesion
- *
- * This function is used by freezers after establishing %true freezing() to
- * test whether a task should be skipped when determining the target frozen
- * state is reached. IOW, if this function returns %true, @p is considered
- * frozen enough.
- */
-static inline bool freezer_should_skip(struct task_struct *p)
-{
- /*
- * The following smp_mb() paired with the one in freezer_count()
- * ensures that either freezer_count() sees %true freezing() or we
- * see cleared %PF_FREEZER_SKIP and return %false. This makes it
- * impossible for a task to slip frozen state testing after
- * clearing %PF_FREEZER_SKIP.
- */
- smp_mb();
- return p->flags & PF_FREEZER_SKIP;
-}
-
-/*
- * These functions are intended to be used whenever you want allow a sleeping
- * task to be frozen. Note that neither return any clear indication of
- * whether a freeze event happened while in this function.
- */
-
-/* Like schedule(), but should not block the freezer. */
-static inline void freezable_schedule(void)
-{
- freezer_do_not_count();
- schedule();
- freezer_count();
-}
-
-/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-static inline void freezable_schedule_unsafe(void)
-{
- freezer_do_not_count();
- schedule();
- freezer_count_unsafe();
-}
-
-/*
- * Like schedule_timeout(), but should not block the freezer. Do not
- * call this with locks held.
- */
-static inline long freezable_schedule_timeout(long timeout)
-{
- long __retval;
- freezer_do_not_count();
- __retval = schedule_timeout(timeout);
- freezer_count();
- return __retval;
-}
-
-/*
- * Like schedule_timeout_interruptible(), but should not block the freezer. Do not
- * call this with locks held.
- */
-static inline long freezable_schedule_timeout_interruptible(long timeout)
-{
- long __retval;
- freezer_do_not_count();
- __retval = schedule_timeout_interruptible(timeout);
- freezer_count();
- return __retval;
-}
-
-/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-static inline long freezable_schedule_timeout_interruptible_unsafe(long timeout)
-{
- long __retval;
-
- freezer_do_not_count();
- __retval = schedule_timeout_interruptible(timeout);
- freezer_count_unsafe();
- return __retval;
-}
-
-/* Like schedule_timeout_killable(), but should not block the freezer. */
-static inline long freezable_schedule_timeout_killable(long timeout)
-{
- long __retval;
- freezer_do_not_count();
- __retval = schedule_timeout_killable(timeout);
- freezer_count();
- return __retval;
-}
-
-/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-static inline long freezable_schedule_timeout_killable_unsafe(long timeout)
-{
- long __retval;
- freezer_do_not_count();
- __retval = schedule_timeout_killable(timeout);
- freezer_count_unsafe();
- return __retval;
-}
-
-/*
- * Like schedule_hrtimeout_range(), but should not block the freezer. Do not
- * call this with locks held.
- */
-static inline int freezable_schedule_hrtimeout_range(ktime_t *expires,
- u64 delta, const enum hrtimer_mode mode)
-{
- int __retval;
- freezer_do_not_count();
- __retval = schedule_hrtimeout_range(expires, delta, mode);
- freezer_count();
- return __retval;
-}
-
-/*
- * Freezer-friendly wrappers around wait_event_interruptible(),
- * wait_event_killable() and wait_event_interruptible_timeout(), originally
- * defined in <linux/wait.h>
- */
-
-/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-#define wait_event_freezekillable_unsafe(wq, condition) \
-({ \
- int __retval; \
- freezer_do_not_count(); \
- __retval = wait_event_killable(wq, (condition)); \
- freezer_count_unsafe(); \
- __retval; \
-})
-
#else /* !CONFIG_FREEZER */
static inline bool frozen(struct task_struct *p) { return false; }
static inline bool freezing(struct task_struct *p) { return false; }
@@ -281,35 +83,8 @@ static inline void thaw_kernel_threads(void) {}

static inline bool try_to_freeze(void) { return false; }

-static inline void freezer_do_not_count(void) {}
-static inline void freezer_count(void) {}
-static inline int freezer_should_skip(struct task_struct *p) { return 0; }
static inline void set_freezable(void) {}

-#define freezable_schedule() schedule()
-
-#define freezable_schedule_unsafe() schedule()
-
-#define freezable_schedule_timeout(timeout) schedule_timeout(timeout)
-
-#define freezable_schedule_timeout_interruptible(timeout) \
- schedule_timeout_interruptible(timeout)
-
-#define freezable_schedule_timeout_interruptible_unsafe(timeout) \
- schedule_timeout_interruptible(timeout)
-
-#define freezable_schedule_timeout_killable(timeout) \
- schedule_timeout_killable(timeout)
-
-#define freezable_schedule_timeout_killable_unsafe(timeout) \
- schedule_timeout_killable(timeout)
-
-#define freezable_schedule_hrtimeout_range(expires, delta, mode) \
- schedule_hrtimeout_range(expires, delta, mode)
-
-#define wait_event_freezekillable_unsafe(wq, condition) \
- wait_event_killable(wq, condition)
-
#endif /* !CONFIG_FREEZER */

#endif /* FREEZER_H_INCLUDED */
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 9fc23a2..4c91efd 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -97,12 +97,19 @@ struct task_group;
#define TASK_WAKING 0x00000200
#define TASK_NOLOAD 0x00000400
#define TASK_NEW 0x00000800
-/* RT specific auxilliary flag to mark RT lock waiters */
#define TASK_RTLOCK_WAIT 0x00001000
-#define TASK_STATE_MAX 0x00002000
+#define TASK_FREEZABLE 0x00002000
+#define __TASK_FREEZABLE_UNSAFE (0x00004000 * IS_ENABLED(CONFIG_LOCKDEP))
+#define TASK_FROZEN 0x00008000
+#define TASK_STATE_MAX 0x00010000

#define TASK_ANY (TASK_STATE_MAX-1)

+/*
+ * DO NOT ADD ANY NEW USERS !
+ */
+#define TASK_FREEZABLE_UNSAFE (TASK_FREEZABLE | __TASK_FREEZABLE_UNSAFE)
+
/* Convenience macros for the sake of set_current_state: */
#define TASK_KILLABLE (TASK_WAKEKILL | TASK_UNINTERRUPTIBLE)
#define TASK_STOPPED (TASK_WAKEKILL | __TASK_STOPPED)
@@ -1716,7 +1723,6 @@ extern struct pid *cad_pid;
#define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */
#define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */
#define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */
-#define PF_FROZEN 0x00010000 /* Frozen for system suspend */
#define PF_KSWAPD 0x00020000 /* I am kswapd */
#define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */
#define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */
@@ -1727,7 +1733,6 @@ extern struct pid *cad_pid;
#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_mask */
#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
#define PF_MEMALLOC_PIN 0x10000000 /* Allocation context constrained to zones which allow long term pinning. */
-#define PF_FREEZER_SKIP 0x40000000 /* Freezer should not count it as freezable */
#define PF_SUSPEND_TASK 0x80000000 /* This thread called freeze_processes() and should not be frozen */

/*
diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h
index acc6264..baeca2f 100644
--- a/include/linux/sunrpc/sched.h
+++ b/include/linux/sunrpc/sched.h
@@ -252,7 +252,7 @@ int rpc_malloc(struct rpc_task *);
void rpc_free(struct rpc_task *);
int rpciod_up(void);
void rpciod_down(void);
-int __rpc_wait_for_completion_task(struct rpc_task *task, wait_bit_action_f *);
+int rpc_wait_for_completion_task(struct rpc_task *task);
#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
struct net;
void rpc_show_tasks(struct net *);
@@ -264,11 +264,6 @@ extern struct workqueue_struct *xprtiod_workqueue;
void rpc_prepare_task(struct rpc_task *task);
gfp_t rpc_task_gfp_mask(void);

-static inline int rpc_wait_for_completion_task(struct rpc_task *task)
-{
- return __rpc_wait_for_completion_task(task, NULL);
-}
-
#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) || IS_ENABLED(CONFIG_TRACEPOINTS)
static inline const char * rpc_qname(const struct rpc_wait_queue *q)
{
diff --git a/include/linux/wait.h b/include/linux/wait.h
index b926eb9..14ad8a0 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -361,8 +361,8 @@ do { \
} while (0)

#define __wait_event_freezable(wq_head, condition) \
- ___wait_event(wq_head, condition, TASK_INTERRUPTIBLE, 0, 0, \
- freezable_schedule())
+ ___wait_event(wq_head, condition, (TASK_INTERRUPTIBLE|TASK_FREEZABLE), \
+ 0, 0, schedule())

/**
* wait_event_freezable - sleep (or freeze) until a condition gets true
@@ -420,8 +420,8 @@ do { \

#define __wait_event_freezable_timeout(wq_head, condition, timeout) \
___wait_event(wq_head, ___wait_cond_timeout(condition), \
- TASK_INTERRUPTIBLE, 0, timeout, \
- __ret = freezable_schedule_timeout(__ret))
+ (TASK_INTERRUPTIBLE|TASK_FREEZABLE), 0, timeout, \
+ __ret = schedule_timeout(__ret))

/*
* like wait_event_timeout() -- except it uses TASK_INTERRUPTIBLE to avoid
@@ -642,8 +642,8 @@ do { \

#define __wait_event_freezable_exclusive(wq, condition) \
- ___wait_event(wq, condition, TASK_INTERRUPTIBLE, 1, 0, \
- freezable_schedule())
+ ___wait_event(wq, condition, (TASK_INTERRUPTIBLE|TASK_FREEZABLE), 1, 0,\
+ schedule())

#define wait_event_freezable_exclusive(wq, condition) \
({ \
diff --git a/kernel/cgroup/legacy_freezer.c b/kernel/cgroup/legacy_freezer.c
index 0823679..1b6b218 100644
--- a/kernel/cgroup/legacy_freezer.c
+++ b/kernel/cgroup/legacy_freezer.c
@@ -113,7 +113,7 @@ static int freezer_css_online(struct cgroup_subsys_state *css)

if (parent && (parent->state & CGROUP_FREEZING)) {
freezer->state |= CGROUP_FREEZING_PARENT | CGROUP_FROZEN;
- atomic_inc(&system_freezing_cnt);
+ static_branch_inc(&freezer_active);
}

mutex_unlock(&freezer_mutex);
@@ -134,7 +134,7 @@ static void freezer_css_offline(struct cgroup_subsys_state *css)
mutex_lock(&freezer_mutex);

if (freezer->state & CGROUP_FREEZING)
- atomic_dec(&system_freezing_cnt);
+ static_branch_dec(&freezer_active);

freezer->state = 0;

@@ -179,6 +179,7 @@ static void freezer_attach(struct cgroup_taskset *tset)
__thaw_task(task);
} else {
freeze_task(task);
+
/* clear FROZEN and propagate upwards */
while (freezer && (freezer->state & CGROUP_FROZEN)) {
freezer->state &= ~CGROUP_FROZEN;
@@ -271,16 +272,8 @@ static void update_if_frozen(struct cgroup_subsys_state *css)
css_task_iter_start(css, 0, &it);

while ((task = css_task_iter_next(&it))) {
- if (freezing(task)) {
- /*
- * freezer_should_skip() indicates that the task
- * should be skipped when determining freezing
- * completion. Consider it frozen in addition to
- * the usual frozen condition.
- */
- if (!frozen(task) && !freezer_should_skip(task))
- goto out_iter_end;
- }
+ if (freezing(task) && !frozen(task))
+ goto out_iter_end;
}

freezer->state |= CGROUP_FROZEN;
@@ -357,7 +350,7 @@ static void freezer_apply_state(struct freezer *freezer, bool freeze,

if (freeze) {
if (!(freezer->state & CGROUP_FREEZING))
- atomic_inc(&system_freezing_cnt);
+ static_branch_inc(&freezer_active);
freezer->state |= state;
freeze_cgroup(freezer);
} else {
@@ -366,9 +359,9 @@ static void freezer_apply_state(struct freezer *freezer, bool freeze,
freezer->state &= ~state;

if (!(freezer->state & CGROUP_FREEZING)) {
- if (was_freezing)
- atomic_dec(&system_freezing_cnt);
freezer->state &= ~CGROUP_FROZEN;
+ if (was_freezing)
+ static_branch_dec(&freezer_active);
unfreeze_cgroup(freezer);
}
}
diff --git a/kernel/exit.c b/kernel/exit.c
index 84021b2..ba683f8 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -374,10 +374,10 @@ static void coredump_task_exit(struct task_struct *tsk)
complete(&core_state->startup);

for (;;) {
- set_current_state(TASK_UNINTERRUPTIBLE);
+ set_current_state(TASK_UNINTERRUPTIBLE|TASK_FREEZABLE);
if (!self.task) /* see coredump_finish() */
break;
- freezable_schedule();
+ schedule();
}
__set_current_state(TASK_RUNNING);
}
diff --git a/kernel/fork.c b/kernel/fork.c
index 90c85b1..a1de2f5 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1420,13 +1420,12 @@ static void complete_vfork_done(struct task_struct *tsk)
static int wait_for_vfork_done(struct task_struct *child,
struct completion *vfork)
{
+ unsigned int state = TASK_UNINTERRUPTIBLE|TASK_KILLABLE|TASK_FREEZABLE;
int killed;

- freezer_do_not_count();
cgroup_enter_frozen();
- killed = wait_for_completion_killable(vfork);
+ killed = wait_for_completion_state(vfork, state);
cgroup_leave_frozen(false);
- freezer_count();

if (killed) {
task_lock(child);
diff --git a/kernel/freezer.c b/kernel/freezer.c
index 45ab36f..4fad0e6 100644
--- a/kernel/freezer.c
+++ b/kernel/freezer.c
@@ -13,10 +13,11 @@
#include <linux/kthread.h>

/* total number of freezing conditions in effect */
-atomic_t system_freezing_cnt = ATOMIC_INIT(0);
-EXPORT_SYMBOL(system_freezing_cnt);
+DEFINE_STATIC_KEY_FALSE(freezer_active);
+EXPORT_SYMBOL(freezer_active);

-/* indicate whether PM freezing is in effect, protected by
+/*
+ * indicate whether PM freezing is in effect, protected by
* system_transition_mutex
*/
bool pm_freezing;
@@ -29,7 +30,7 @@ static DEFINE_SPINLOCK(freezer_lock);
* freezing_slow_path - slow path for testing whether a task needs to be frozen
* @p: task to be tested
*
- * This function is called by freezing() if system_freezing_cnt isn't zero
+ * This function is called by freezing() if freezer_active isn't zero
* and tests whether @p needs to enter and stay in frozen state. Can be
* called under any context. The freezers are responsible for ensuring the
* target tasks see the updated state.
@@ -52,41 +53,40 @@ bool freezing_slow_path(struct task_struct *p)
}
EXPORT_SYMBOL(freezing_slow_path);

+bool frozen(struct task_struct *p)
+{
+ return READ_ONCE(p->__state) & TASK_FROZEN;
+}
+
/* Refrigerator is place where frozen processes are stored :-). */
bool __refrigerator(bool check_kthr_stop)
{
- /* Hmm, should we be allowed to suspend when there are realtime
- processes around? */
+ unsigned int state = get_current_state();
bool was_frozen = false;
- unsigned int save = get_current_state();

pr_debug("%s entered refrigerator\n", current->comm);

+ WARN_ON_ONCE(state && !(state & TASK_NORMAL));
+
for (;;) {
- set_current_state(TASK_UNINTERRUPTIBLE);
+ bool freeze;
+
+ set_current_state(TASK_FROZEN);

spin_lock_irq(&freezer_lock);
- current->flags |= PF_FROZEN;
- if (!freezing(current) ||
- (check_kthr_stop && kthread_should_stop()))
- current->flags &= ~PF_FROZEN;
+ freeze = freezing(current) && !(check_kthr_stop && kthread_should_stop());
spin_unlock_irq(&freezer_lock);

- if (!(current->flags & PF_FROZEN))
+ if (!freeze)
break;
+
was_frozen = true;
schedule();
}
+ __set_current_state(TASK_RUNNING);

pr_debug("%s left refrigerator\n", current->comm);

- /*
- * Restore saved task state before returning. The mb'd version
- * needs to be used; otherwise, it might silently break
- * synchronization which depends on ordered task state change.
- */
- set_current_state(save);
-
return was_frozen;
}
EXPORT_SYMBOL(__refrigerator);
@@ -101,6 +101,44 @@ static void fake_signal_wake_up(struct task_struct *p)
}
}

+static int __set_task_frozen(struct task_struct *p, void *arg)
+{
+ unsigned int state = READ_ONCE(p->__state);
+
+ if (p->on_rq)
+ return 0;
+
+ if (p != current && task_curr(p))
+ return 0;
+
+ if (!(state & (TASK_FREEZABLE | __TASK_STOPPED | __TASK_TRACED)))
+ return 0;
+
+ /*
+ * Only TASK_NORMAL can be augmented with TASK_FREEZABLE, since they
+ * can suffer spurious wakeups.
+ */
+ if (state & TASK_FREEZABLE)
+ WARN_ON_ONCE(!(state & TASK_NORMAL));
+
+#ifdef CONFIG_LOCKDEP
+ /*
+ * It's dangerous to freeze with locks held; there be dragons there.
+ */
+ if (!(state & __TASK_FREEZABLE_UNSAFE))
+ WARN_ON_ONCE(debug_locks && p->lockdep_depth);
+#endif
+
+ WRITE_ONCE(p->__state, TASK_FROZEN);
+ return TASK_FROZEN;
+}
+
+static bool __freeze_task(struct task_struct *p)
+{
+ /* TASK_FREEZABLE|TASK_STOPPED|TASK_TRACED -> TASK_FROZEN */
+ return task_call_func(p, __set_task_frozen, NULL);
+}
+
/**
* freeze_task - send a freeze request to given task
* @p: task to send the request to
@@ -116,20 +154,8 @@ bool freeze_task(struct task_struct *p)
{
unsigned long flags;

- /*
- * This check can race with freezer_do_not_count, but worst case that
- * will result in an extra wakeup being sent to the task. It does not
- * race with freezer_count(), the barriers in freezer_count() and
- * freezer_should_skip() ensure that either freezer_count() sees
- * freezing == true in try_to_freeze() and freezes, or
- * freezer_should_skip() sees !PF_FREEZE_SKIP and freezes the task
- * normally.
- */
- if (freezer_should_skip(p))
- return false;
-
spin_lock_irqsave(&freezer_lock, flags);
- if (!freezing(p) || frozen(p)) {
+ if (!freezing(p) || frozen(p) || __freeze_task(p)) {
spin_unlock_irqrestore(&freezer_lock, flags);
return false;
}
@@ -137,19 +163,52 @@ bool freeze_task(struct task_struct *p)
if (!(p->flags & PF_KTHREAD))
fake_signal_wake_up(p);
else
- wake_up_state(p, TASK_INTERRUPTIBLE);
+ wake_up_state(p, TASK_NORMAL);

spin_unlock_irqrestore(&freezer_lock, flags);
return true;
}

+/*
+ * The special task states (TASK_STOPPED, TASK_TRACED) keep their canonical
+ * state in p->jobctl. If either of them got a wakeup that was missed because
+ * TASK_FROZEN, then their canonical state reflects that and the below will
+ * refuse to restore the special state and instead issue the wakeup.
+ */
+static int __set_task_special(struct task_struct *p, void *arg)
+{
+ unsigned int state = 0;
+
+ if (p->jobctl & JOBCTL_TRACED)
+ state = TASK_TRACED;
+
+ else if (p->jobctl & JOBCTL_STOPPED)
+ state = TASK_STOPPED;
+
+ if (state)
+ WRITE_ONCE(p->__state, state);
+
+ return state;
+}
+
void __thaw_task(struct task_struct *p)
{
- unsigned long flags;
+ unsigned long flags, flags2;

spin_lock_irqsave(&freezer_lock, flags);
- if (frozen(p))
- wake_up_process(p);
+ if (WARN_ON_ONCE(freezing(p)))
+ goto unlock;
+
+ if (lock_task_sighand(p, &flags2)) {
+ /* TASK_FROZEN -> TASK_{STOPPED,TRACED} */
+ bool ret = task_call_func(p, __set_task_special, NULL);
+ unlock_task_sighand(p, &flags2);
+ if (ret)
+ goto unlock;
+ }
+
+ wake_up_state(p, TASK_FROZEN);
+unlock:
spin_unlock_irqrestore(&freezer_lock, flags);
}

diff --git a/kernel/futex/waitwake.c b/kernel/futex/waitwake.c
index 4ce0923..ba01b94 100644
--- a/kernel/futex/waitwake.c
+++ b/kernel/futex/waitwake.c
@@ -334,7 +334,7 @@ void futex_wait_queue(struct futex_hash_bucket *hb, struct futex_q *q,
* futex_queue() calls spin_unlock() upon completion, both serializing
* access to the hash list and forcing another memory barrier.
*/
- set_current_state(TASK_INTERRUPTIBLE);
+ set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
futex_queue(q, hb);

/* Arm the timer */
@@ -352,7 +352,7 @@ void futex_wait_queue(struct futex_hash_bucket *hb, struct futex_q *q,
* is no timeout, or if it has yet to expire.
*/
if (!timeout || timeout->task)
- freezable_schedule();
+ schedule();
}
__set_current_state(TASK_RUNNING);
}
@@ -430,7 +430,7 @@ retry:
return ret;
}

- set_current_state(TASK_INTERRUPTIBLE);
+ set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);

for (i = 0; i < count; i++) {
u32 __user *uaddr = (u32 __user *)(unsigned long)vs[i].w.uaddr;
@@ -504,7 +504,7 @@ static void futex_sleep_multiple(struct futex_vector *vs, unsigned int count,
return;
}

- freezable_schedule();
+ schedule();
}

/**
diff --git a/kernel/hung_task.c b/kernel/hung_task.c
index bb2354f..f1321c0 100644
--- a/kernel/hung_task.c
+++ b/kernel/hung_task.c
@@ -95,8 +95,8 @@ static void check_hung_task(struct task_struct *t, unsigned long timeout)
* Ensure the task is not frozen.
* Also, skip vfork and any other user process that freezer should skip.
*/
- if (unlikely(t->flags & (PF_FROZEN | PF_FREEZER_SKIP)))
- return;
+ if (unlikely(READ_ONCE(t->__state) & (TASK_FREEZABLE | TASK_FROZEN)))
+ return;

/*
* When a freshly created task is scheduled once, changes its state to
diff --git a/kernel/power/main.c b/kernel/power/main.c
index dae3d17..31ec4a9 100644
--- a/kernel/power/main.c
+++ b/kernel/power/main.c
@@ -24,7 +24,7 @@
unsigned int lock_system_sleep(void)
{
unsigned int flags = current->flags;
- current->flags |= PF_FREEZER_SKIP;
+ current->flags |= PF_NOFREEZE;
mutex_lock(&system_transition_mutex);
return flags;
}
@@ -48,8 +48,8 @@ void unlock_system_sleep(unsigned int flags)
* Which means, if we use try_to_freeze() here, it would make them
* enter the refrigerator, thus causing hibernation to lockup.
*/
- if (!(flags & PF_FREEZER_SKIP))
- current->flags &= ~PF_FREEZER_SKIP;
+ if (!(flags & PF_NOFREEZE))
+ current->flags &= ~PF_NOFREEZE;
mutex_unlock(&system_transition_mutex);
}
EXPORT_SYMBOL_GPL(unlock_system_sleep);
diff --git a/kernel/power/process.c b/kernel/power/process.c
index 3068601..ddd9988 100644
--- a/kernel/power/process.c
+++ b/kernel/power/process.c
@@ -50,8 +50,7 @@ static int try_to_freeze_tasks(bool user_only)
if (p == current || !freeze_task(p))
continue;

- if (!freezer_should_skip(p))
- todo++;
+ todo++;
}
read_unlock(&tasklist_lock);

@@ -96,8 +95,7 @@ static int try_to_freeze_tasks(bool user_only)
if (!wakeup || pm_debug_messages_on) {
read_lock(&tasklist_lock);
for_each_process_thread(g, p) {
- if (p != current && !freezer_should_skip(p)
- && freezing(p) && !frozen(p))
+ if (p != current && freezing(p) && !frozen(p))
sched_show_task(p);
}
read_unlock(&tasklist_lock);
@@ -129,7 +127,7 @@ int freeze_processes(void)
current->flags |= PF_SUSPEND_TASK;

if (!pm_freezing)
- atomic_inc(&system_freezing_cnt);
+ static_branch_inc(&freezer_active);

pm_wakeup_clear(0);
pr_info("Freezing user space processes ... ");
@@ -190,7 +188,7 @@ void thaw_processes(void)

trace_suspend_resume(TPS("thaw_processes"), 0, true);
if (pm_freezing)
- atomic_dec(&system_freezing_cnt);
+ static_branch_dec(&freezer_active);
pm_freezing = false;
pm_nosig_freezing = false;

diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index 1893d90..5448219 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -269,7 +269,7 @@ static int ptrace_check_attach(struct task_struct *child, bool ignore_state)
read_unlock(&tasklist_lock);

if (!ret && !ignore_state &&
- WARN_ON_ONCE(!wait_task_inactive(child, __TASK_TRACED)))
+ WARN_ON_ONCE(!wait_task_inactive(child, __TASK_TRACED|TASK_FROZEN)))
ret = -ESRCH;

return ret;
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index ea26526..2b85d1b 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -6428,7 +6428,7 @@ static void __sched notrace __schedule(unsigned int sched_mode)
prev->sched_contributes_to_load =
(prev_state & TASK_UNINTERRUPTIBLE) &&
!(prev_state & TASK_NOLOAD) &&
- !(prev->flags & PF_FROZEN);
+ !(prev_state & TASK_FROZEN);

if (prev->sched_contributes_to_load)
rq->nr_uninterruptible++;
diff --git a/kernel/signal.c b/kernel/signal.c
index 6f86fda..c2e90d6 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -2304,7 +2304,7 @@ static int ptrace_stop(int exit_code, int why, unsigned long message,
read_unlock(&tasklist_lock);
cgroup_enter_frozen();
preempt_enable_no_resched();
- freezable_schedule();
+ schedule();
cgroup_leave_frozen(true);

/*
@@ -2473,7 +2473,7 @@ static bool do_signal_stop(int signr)

/* Now we don't run again until woken by SIGCONT or SIGKILL */
cgroup_enter_frozen();
- freezable_schedule();
+ schedule();
return true;
} else {
/*
@@ -2548,11 +2548,11 @@ static void do_freezer_trap(void)
* immediately (if there is a non-fatal signal pending), and
* put the task into sleep.
*/
- __set_current_state(TASK_INTERRUPTIBLE);
+ __set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
clear_thread_flag(TIF_SIGPENDING);
spin_unlock_irq(&current->sighand->siglock);
cgroup_enter_frozen();
- freezable_schedule();
+ schedule();
}

static int ptrace_signal(int signr, kernel_siginfo_t *info, enum pid_type type)
@@ -3600,9 +3600,9 @@ static int do_sigtimedwait(const sigset_t *which, kernel_siginfo_t *info,
recalc_sigpending();
spin_unlock_irq(&tsk->sighand->siglock);

- __set_current_state(TASK_INTERRUPTIBLE);
- ret = freezable_schedule_hrtimeout_range(to, tsk->timer_slack_ns,
- HRTIMER_MODE_REL);
+ __set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
+ ret = schedule_hrtimeout_range(to, tsk->timer_slack_ns,
+ HRTIMER_MODE_REL);
spin_lock_irq(&tsk->sighand->siglock);
__set_task_blocked(tsk, &tsk->real_blocked);
sigemptyset(&tsk->real_blocked);
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index 23af5ec..3ae661a 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -2037,11 +2037,11 @@ static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mod
struct restart_block *restart;

do {
- set_current_state(TASK_INTERRUPTIBLE);
+ set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
hrtimer_sleeper_start_expires(t, mode);

if (likely(t->task))
- freezable_schedule();
+ schedule();

hrtimer_cancel(&t->timer);
mode = HRTIMER_MODE_ABS;
diff --git a/kernel/umh.c b/kernel/umh.c
index 8945eaf..8506315 100644
--- a/kernel/umh.c
+++ b/kernel/umh.c
@@ -404,6 +404,7 @@ EXPORT_SYMBOL(call_usermodehelper_setup);
*/
int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait)
{
+ unsigned int state = TASK_UNINTERRUPTIBLE;
DECLARE_COMPLETION_ONSTACK(done);
int retval = 0;

@@ -437,25 +438,22 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int wait)
if (wait == UMH_NO_WAIT) /* task has freed sub_info */
goto unlock;

+ if (wait & UMH_KILLABLE)
+ state |= TASK_KILLABLE;
+
if (wait & UMH_FREEZABLE)
- freezer_do_not_count();
+ state |= TASK_FREEZABLE;

- if (wait & UMH_KILLABLE) {
- retval = wait_for_completion_killable(&done);
- if (!retval)
- goto wait_done;
+ retval = wait_for_completion_state(&done, state);
+ if (!retval)
+ goto wait_done;

+ if (wait & UMH_KILLABLE) {
/* umh_complete() will see NULL and free sub_info */
if (xchg(&sub_info->complete, NULL))
goto unlock;
- /* fallthrough, umh_complete() was already called */
}

- wait_for_completion(&done);
-
- if (wait & UMH_FREEZABLE)
- freezer_count();
-
wait_done:
retval = sub_info->retval;
out:
diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 01f7178..77162fb 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -730,8 +730,8 @@ static void khugepaged_alloc_sleep(void)
DEFINE_WAIT(wait);

add_wait_queue(&khugepaged_wait, &wait);
- freezable_schedule_timeout_interruptible(
- msecs_to_jiffies(khugepaged_alloc_sleep_millisecs));
+ __set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
+ schedule_timeout(msecs_to_jiffies(khugepaged_alloc_sleep_millisecs));
remove_wait_queue(&khugepaged_wait, &wait);
}

diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 25b9221..46cbf15 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -269,7 +269,7 @@ EXPORT_SYMBOL_GPL(rpc_destroy_wait_queue);

static int rpc_wait_bit_killable(struct wait_bit_key *key, int mode)
{
- freezable_schedule_unsafe();
+ schedule();
if (signal_pending_state(mode, current))
return -ERESTARTSYS;
return 0;
@@ -333,14 +333,12 @@ static int rpc_complete_task(struct rpc_task *task)
* to enforce taking of the wq->lock and hence avoid races with
* rpc_complete_task().
*/
-int __rpc_wait_for_completion_task(struct rpc_task *task, wait_bit_action_f *action)
+int rpc_wait_for_completion_task(struct rpc_task *task)
{
- if (action == NULL)
- action = rpc_wait_bit_killable;
return out_of_line_wait_on_bit(&task->tk_runstate, RPC_TASK_ACTIVE,
- action, TASK_KILLABLE);
+ rpc_wait_bit_killable, TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
}
-EXPORT_SYMBOL_GPL(__rpc_wait_for_completion_task);
+EXPORT_SYMBOL_GPL(rpc_wait_for_completion_task);

/*
* Make an RPC task runnable.
@@ -964,7 +962,7 @@ static void __rpc_execute(struct rpc_task *task)
trace_rpc_task_sync_sleep(task, task->tk_action);
status = out_of_line_wait_on_bit(&task->tk_runstate,
RPC_TASK_QUEUED, rpc_wait_bit_killable,
- TASK_KILLABLE);
+ TASK_KILLABLE|TASK_FREEZABLE);
if (status < 0) {
/*
* When a sync task receives a signal, it exits with
diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index bf338b7..dda9eb1 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -2543,13 +2543,14 @@ static long unix_stream_data_wait(struct sock *sk, long timeo,
struct sk_buff *last, unsigned int last_len,
bool freezable)
{
+ unsigned int state = TASK_INTERRUPTIBLE | freezable * TASK_FREEZABLE;
struct sk_buff *tail;
DEFINE_WAIT(wait);

unix_state_lock(sk);

for (;;) {
- prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
+ prepare_to_wait(sk_sleep(sk), &wait, state);

tail = skb_peek_tail(&sk->sk_receive_queue);
if (tail != last ||
@@ -2562,10 +2563,7 @@ static long unix_stream_data_wait(struct sock *sk, long timeo,

sk_set_bit(SOCKWQ_ASYNC_WAITDATA, sk);
unix_state_unlock(sk);
- if (freezable)
- timeo = freezable_schedule_timeout(timeo);
- else
- timeo = schedule_timeout(timeo);
+ timeo = schedule_timeout(timeo);
unix_state_lock(sk);

if (sock_flag(sk, SOCK_DEAD))

2022-09-09 09:49:45

by tip-bot2 for Alexey Makhalov

[permalink] [raw]

Subject: [tip: sched/core] sched: Show PF_flag holes

The following commit has been merged into the sched/core branch of tip:

Commit-ID: fb04563d1cae6f361892b4a339ad92100b1eb0d0
Gitweb: https://git.kernel.org/tip/fb04563d1cae6f361892b4a339ad92100b1eb0d0
Author: Peter Zijlstra <[email protected]>
AuthorDate: Tue, 06 Sep 2022 13:27:32 +02:00
Committer: Peter Zijlstra <[email protected]>
CommitterDate: Wed, 07 Sep 2022 21:53:51 +02:00

sched: Show PF_flag holes

Put placeholders in PF_flag so we can see how holey it is.

Suggested-by: Ingo Molnar <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Reviewed-by: Ingo Molnar <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
---
include/linux/sched.h | 7 +++++++
1 file changed, 7 insertions(+)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 4c91efd..15e3bd9 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1722,7 +1722,9 @@ extern struct pid *cad_pid;
#define PF_MEMALLOC 0x00000800 /* Allocating memory */
#define PF_NPROC_EXCEEDED 0x00001000 /* set_user() noticed that RLIMIT_NPROC was exceeded */
#define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */
+#define PF__HOLE__00004000 0x00004000
#define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */
+#define PF__HOLE__00010000 0x00010000
#define PF_KSWAPD 0x00020000 /* I am kswapd */
#define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */
#define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */
@@ -1730,9 +1732,14 @@ extern struct pid *cad_pid;
* I am cleaning dirty pages from some other bdi. */
#define PF_KTHREAD 0x00200000 /* I am a kernel thread */
#define PF_RANDOMIZE 0x00400000 /* Randomize virtual address space */
+#define PF__HOLE__00800000 0x00800000
+#define PF__HOLE__01000000 0x01000000
+#define PF__HOLE__02000000 0x02000000
#define PF_NO_SETAFFINITY 0x04000000 /* Userland is not allowed to meddle with cpus_mask */
#define PF_MCE_EARLY 0x08000000 /* Early kill for mce process policy */
#define PF_MEMALLOC_PIN 0x10000000 /* Allocation context constrained to zones which allow long term pinning. */
+#define PF__HOLE__20000000 0x20000000
+#define PF__HOLE__40000000 0x40000000
#define PF_SUSPEND_TASK 0x80000000 /* This thread called freeze_processes() and should not be frozen */

/*

2022-09-23 07:40:01

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Peter,

as a heads-up. This commit (bisected and verified) triggers a
regression in our KVM on s390x CI. The symptom is that a specific
testcase (start a guest with next kernel and a poky ramdisk,
then ssh via vsock into the guest and run the reboot command) now
takes much longer (300 instead of 20 seconds). From a first look
it seems that the sshd takes very long to end during shutdown
but I have not looked into that yet.
Any quick idea?

Christian

2022-09-23 08:06:50

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 23.09.22 um 09:21 schrieb Christian Borntraeger:
> Peter,
>
> as a heads-up. This commit (bisected and verified) triggers a
> regression in our KVM on s390x CI. The symptom is that a specific
> testcase (start a guest with next kernel and a poky ramdisk,
> then ssh via vsock into the guest and run the reboot command) now
> takes much longer (300 instead of 20 seconds). From a first look
> it seems that the sshd takes very long to end during shutdown
> but I have not looked into that yet.
> Any quick idea?
>
> Christian

the sshd seems to hang in virtio-serial (not vsock).

PID: 237 TASK: 81d1a100 CPU: 1 COMMAND: "sshd"
LOWCORE INFO:
-psw : 0x0404e00180000000 0x0000000131ceb136
-function : __list_add_valid at 131ceb136
-prefix : 0x00410000
-cpu timer: 0x7fffffd3ec4f33d4
-clock cmp: 0x2639f08006283e00
-general registers:
0x00000008dcea2dce 0x00000001387d44b8
0x0000000081d1a228 0x00000001387d44b8
0x00000001387d44b8 0x00000001387d44b8
0x00000001387d3800 0x00000001387d3700
0x0000000081d1a100 0x00000001387d44b8
0x00000001387d44b8 0x0000000081d1a228
0x0000000081d1a100 0x0000000081d1a100
0x0000000131608b32 0x00000380004b7aa8
-access registers:
0x000003ff 0x8fff5870 0000000000 0000000000
0000000000 0000000000 0000000000 0000000000
0000000000 0000000000 0000000000 0000000000
0000000000 0000000000 0000000000 0000000000
-control registers:
0x00a0000014966a10 0x0000000133348007
0x00000000028c6140 000000000000000000
0x000000000000ffff 0x00000000028c6140
0x0000000033000000 0x0000000081f001c7
0x0000000000008000 000000000000000000
000000000000000000 000000000000000000
000000000000000000 0x0000000133348007
0x00000000db000000 0x00000000028c6000
-floating point registers:
0x000003ffb82a9761 0x0000006400000000
0x000003ffb82a345c 000000000000000000
0x0000000000007fff 0x000003ffe22fe000
000000000000000000 0x000003ffe22fa51c
0x000003ffb81889c0 000000000000000000
0x000002aa3ce2b470 000000000000000000
000000000000000000 000000000000000000
000000000000000000 000000000000000000

#0 [380004b7b00] pick_next_task at 1315f2088
#1 [380004b7b98] __schedule at 13215e954
#2 [380004b7c08] schedule at 13215eeea
#3 [380004b7c38] wait_port_writable at 3ff80149b2e [virtio_console]
#4 [380004b7cc0] port_fops_write at 3ff8014a282 [virtio_console]
#5 [380004b7d40] vfs_write at 131889e3c
#6 [380004b7e00] ksys_write at 13188a2e8
#7 [380004b7e50] __do_syscall at 13215761c
#8 [380004b7e98] system_call at 132166332
PSW: 0705000180000000 000003ff8f8f3a2a (user space)
GPRS: 0000000000000015 000003ff00000000 ffffffffffffffda 000002aa02fc68a0
0000000000000015 0000000000000015 0000000000000000 000002aa02fc68a0
0000000000000004 000003ff8f8f3a08 0000000000000015 0000000000000000
000003ff8ffa9f58 0000000000000000 000002aa02365b20 000003ffdf4798d0

2022-09-26 09:16:22

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 23.09.22 um 09:53 schrieb Christian Borntraeger:
> Am 23.09.22 um 09:21 schrieb Christian Borntraeger:
>> Peter,
>>
>> as a heads-up. This commit (bisected and verified) triggers a
>> regression in our KVM on s390x CI. The symptom is that a specific
>> testcase (start a guest with next kernel and a poky ramdisk,
>> then ssh via vsock into the guest and run the reboot command) now
>> takes much longer (300 instead of 20 seconds). From a first look
>> it seems that the sshd takes very long to end during shutdown
>> but I have not looked into that yet.
>> Any quick idea?
>>
>> Christian
>
> the sshd seems to hang in virtio-serial (not vsock).

FWIW, sshd does not seem to hang, instead it seems to busy loop in
wait_port_writable calling into the scheduler over and over again.

>
> PID: 237      TASK: 81d1a100          CPU: 1    COMMAND: "sshd"
> LOWCORE INFO:
> -psw      : 0x0404e00180000000 0x0000000131ceb136
> -function : __list_add_valid at 131ceb136
> -prefix   : 0x00410000
> -cpu timer: 0x7fffffd3ec4f33d4
> -clock cmp: 0x2639f08006283e00
> -general registers:
>      0x00000008dcea2dce 0x00000001387d44b8
>      0x0000000081d1a228 0x00000001387d44b8
>      0x00000001387d44b8 0x00000001387d44b8
>      0x00000001387d3800 0x00000001387d3700
>      0x0000000081d1a100 0x00000001387d44b8
>      0x00000001387d44b8 0x0000000081d1a228
>      0x0000000081d1a100 0x0000000081d1a100
>      0x0000000131608b32 0x00000380004b7aa8
> -access registers:
>      0x000003ff 0x8fff5870 0000000000 0000000000
>      0000000000 0000000000 0000000000 0000000000
>      0000000000 0000000000 0000000000 0000000000
>      0000000000 0000000000 0000000000 0000000000
> -control registers:
>      0x00a0000014966a10 0x0000000133348007
>      0x00000000028c6140 000000000000000000
>      0x000000000000ffff 0x00000000028c6140
>      0x0000000033000000 0x0000000081f001c7
>      0x0000000000008000 000000000000000000
>      000000000000000000 000000000000000000
>      000000000000000000 0x0000000133348007
>      0x00000000db000000 0x00000000028c6000
> -floating point registers:
>      0x000003ffb82a9761 0x0000006400000000
>      0x000003ffb82a345c 000000000000000000
>      0x0000000000007fff 0x000003ffe22fe000
>      000000000000000000 0x000003ffe22fa51c
>      0x000003ffb81889c0 000000000000000000
>      0x000002aa3ce2b470 000000000000000000
>      000000000000000000 000000000000000000
>      000000000000000000 000000000000000000
>
> #0 [380004b7b00] pick_next_task at 1315f2088
> #1 [380004b7b98] __schedule at 13215e954
> #2 [380004b7c08] schedule at 13215eeea
> #3 [380004b7c38] wait_port_writable at 3ff80149b2e [virtio_console]
> #4 [380004b7cc0] port_fops_write at 3ff8014a282 [virtio_console]
> #5 [380004b7d40] vfs_write at 131889e3c
> #6 [380004b7e00] ksys_write at 13188a2e8
> #7 [380004b7e50] __do_syscall at 13215761c
> #8 [380004b7e98] system_call at 132166332
> PSW: 0705000180000000 000003ff8f8f3a2a (user space)
> GPRS: 0000000000000015 000003ff00000000 ffffffffffffffda 000002aa02fc68a0
>        0000000000000015 0000000000000015 0000000000000000 000002aa02fc68a0
>        0000000000000004 000003ff8f8f3a08 0000000000000015 0000000000000000
>        000003ff8ffa9f58 0000000000000000 000002aa02365b20 000003ffdf4798d0

2022-09-26 12:29:48

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 26.09.22 um 10:06 schrieb Christian Borntraeger:
>
>
> Am 23.09.22 um 09:53 schrieb Christian Borntraeger:
>> Am 23.09.22 um 09:21 schrieb Christian Borntraeger:
>>> Peter,
>>>
>>> as a heads-up. This commit (bisected and verified) triggers a
>>> regression in our KVM on s390x CI. The symptom is that a specific
>>> testcase (start a guest with next kernel and a poky ramdisk,
>>> then ssh via vsock into the guest and run the reboot command) now
>>> takes much longer (300 instead of 20 seconds). From a first look
>>> it seems that the sshd takes very long to end during shutdown
>>> but I have not looked into that yet.
>>> Any quick idea?
>>>
>>> Christian
>>
>> the sshd seems to hang in virtio-serial (not vsock).
>
> FWIW, sshd does not seem to hang, instead it seems to busy loop in
> wait_port_writable calling into the scheduler over and over again.

-#define TASK_FREEZABLE 0x00002000
+#define TASK_FREEZABLE 0x00000000

"Fixes" the issue. Just have to find out which of users is responsible.

2022-09-26 14:22:43

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Mon, Sep 26, 2022 at 12:55:21PM +0200, Christian Borntraeger wrote:
>
>
> Am 26.09.22 um 10:06 schrieb Christian Borntraeger:
> >
> >
> > Am 23.09.22 um 09:53 schrieb Christian Borntraeger:
> > > Am 23.09.22 um 09:21 schrieb Christian Borntraeger:
> > > > Peter,
> > > >
> > > > as a heads-up. This commit (bisected and verified) triggers a
> > > > regression in our KVM on s390x CI. The symptom is that a specific
> > > > testcase (start a guest with next kernel and a poky ramdisk,
> > > > then ssh via vsock into the guest and run the reboot command) now
> > > > takes much longer (300 instead of 20 seconds). From a first look
> > > > it seems that the sshd takes very long to end during shutdown
> > > > but I have not looked into that yet.
> > > > Any quick idea?
> > > >
> > > > Christian
> > >
> > > the sshd seems to hang in virtio-serial (not vsock).
> >
> > FWIW, sshd does not seem to hang, instead it seems to busy loop in
> > wait_port_writable calling into the scheduler over and over again.
>
> -#define TASK_FREEZABLE 0x00002000
> +#define TASK_FREEZABLE 0x00000000
>
> "Fixes" the issue. Just have to find out which of users is responsible.

Since it's not the wait_port_writable() one -- we already tested that by
virtue of 's/wait_event_freezable/wait_event/' there, it must be on the
producing side of that port. But I'm having a wee bit of trouble
following that code.

Is there a task stuck in FROZEN state? -- then again, I thought you said
there was no actual suspend involved, so that should not be it either.

I'm curious though -- how far does it get into the scheduler? It should
call schedule() with __state == TASK_INTERRUPTIBLE|TASK_FREEZABLE, which
is quite sufficient to get it off the runqueue, who then puts it back?
Or is it bailing early in the wait_event loop?

2022-09-26 14:56:08

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 26.09.22 um 12:55 schrieb Christian Borntraeger:
>
>
> Am 26.09.22 um 10:06 schrieb Christian Borntraeger:
>>
>>
>> Am 23.09.22 um 09:53 schrieb Christian Borntraeger:
>>> Am 23.09.22 um 09:21 schrieb Christian Borntraeger:
>>>> Peter,
>>>>
>>>> as a heads-up. This commit (bisected and verified) triggers a
>>>> regression in our KVM on s390x CI. The symptom is that a specific
>>>> testcase (start a guest with next kernel and a poky ramdisk,
>>>> then ssh via vsock into the guest and run the reboot command) now
>>>> takes much longer (300 instead of 20 seconds). From a first look
>>>> it seems that the sshd takes very long to end during shutdown
>>>> but I have not looked into that yet.
>>>> Any quick idea?
>>>>
>>>> Christian
>>>
>>> the sshd seems to hang in virtio-serial (not vsock).
>>
>> FWIW, sshd does not seem to hang, instead it seems to busy loop in
>> wait_port_writable calling into the scheduler over and over again.
>
> -#define TASK_FREEZABLE 0x00002000
> +#define TASK_FREEZABLE 0x00000000
>
> "Fixes" the issue. Just have to find out which of users is responsible.

So it seems that my initial test was not good enough.

diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c
index 9fa3c76a267f..e93df4f735fe 100644
--- a/drivers/char/virtio_console.c
+++ b/drivers/char/virtio_console.c
@@ -790,7 +790,7 @@ static int wait_port_writable(struct port *port, bool nonblock)
if (nonblock)
return -EAGAIN;

- ret = wait_event_freezable(port->waitqueue,
+ ret = wait_event_interruptible(port->waitqueue,
!will_write_block(port));
if (ret < 0)
return ret;

Does fix the problem.
My initial test was the following

--- a/drivers/char/virtio_console.c
+++ b/drivers/char/virtio_console.c
@@ -790,10 +790,8 @@ static int wait_port_writable(struct port *port, bool nonblock)
if (nonblock)
return -EAGAIN;

- ret = wait_event_freezable(port->waitqueue,
+ wait_event(port->waitqueue,
!will_write_block(port));
- if (ret < 0)
- return ret;
}
/* Port got hot-unplugged. */
if (!port->guest_connected)

and obviously it did not provide an exit path.

2022-09-26 15:11:56

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Mon, Sep 26, 2022 at 03:23:10PM +0200, Christian Borntraeger wrote:
> Am 26.09.22 um 14:55 schrieb Peter Zijlstra:
>
> > Could you please test with something like the below on? I can boot that
> > with KVM, but obviously I didn't suffer any weirdness to begin with :/
> >
> > ---
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index 4e6a6417211f..ef9ccfc3a8c0 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -4051,6 +4051,8 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> > unsigned long flags;
> > int cpu, success = 0;
> > + WARN_ON_ONCE(state & TASK_FREEZABLE);
> > +
> > preempt_disable();
> > if (p == current) {
> > /*
>
> Does not seem to trigger.

Moo -- quite the puzzle this :/ I'll go stare at it more then.

2022-09-26 15:49:34

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 26.09.22 um 15:37 schrieb Peter Zijlstra:
> On Mon, Sep 26, 2022 at 03:23:10PM +0200, Christian Borntraeger wrote:
>> Am 26.09.22 um 14:55 schrieb Peter Zijlstra:
>>
>>> Could you please test with something like the below on? I can boot that
>>> with KVM, but obviously I didn't suffer any weirdness to begin with :/
>>>
>>> ---
>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>> index 4e6a6417211f..ef9ccfc3a8c0 100644
>>> --- a/kernel/sched/core.c
>>> +++ b/kernel/sched/core.c
>>> @@ -4051,6 +4051,8 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
>>> unsigned long flags;
>>> int cpu, success = 0;
>>> + WARN_ON_ONCE(state & TASK_FREEZABLE);
>>> +
>>> preempt_disable();
>>> if (p == current) {
>>> /*
>>
>> Does not seem to trigger.
>
> Moo -- quite the puzzle this :/ I'll go stare at it more then.

In the end this is about the end of the sshd process (shutting it down).
I can also trigger the problem by sending a SIGTERM so its not about
the shutdown itself.
Pofiling the guest I see scheduler functions like sched_clock, pick_next_entity,
update_min_vruntime and so on with 100% system time.

2022-09-26 15:52:33

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Mon, Sep 26, 2022 at 02:32:24PM +0200, Christian Borntraeger wrote:
> diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c
> index 9fa3c76a267f..e93df4f735fe 100644
> --- a/drivers/char/virtio_console.c
> +++ b/drivers/char/virtio_console.c
> @@ -790,7 +790,7 @@ static int wait_port_writable(struct port *port, bool nonblock)
> if (nonblock)
> return -EAGAIN;
> - ret = wait_event_freezable(port->waitqueue,
> + ret = wait_event_interruptible(port->waitqueue,
> !will_write_block(port));
> if (ret < 0)
> return ret;
>
> Does fix the problem.

It's almost as if someone does try_to_wake_up(.state = TASK_FREEZABLE)
-- which would be quite insane.

Could you please test with something like the below on? I can boot that
with KVM, but obviously I didn't suffer any weirdness to begin with :/

---
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 4e6a6417211f..ef9ccfc3a8c0 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4051,6 +4051,8 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
unsigned long flags;
int cpu, success = 0;

+ WARN_ON_ONCE(state & TASK_FREEZABLE);
+
preempt_disable();
if (p == current) {
/*

2022-09-26 15:53:54

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 26.09.22 um 14:55 schrieb Peter Zijlstra:

> Could you please test with something like the below on? I can boot that
> with KVM, but obviously I didn't suffer any weirdness to begin with :/
>
> ---
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 4e6a6417211f..ef9ccfc3a8c0 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -4051,6 +4051,8 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> unsigned long flags;
> int cpu, success = 0;
>
> + WARN_ON_ONCE(state & TASK_FREEZABLE);
> +
> preempt_disable();
> if (p == current) {
> /*

Does not seem to trigger.

2022-09-26 18:02:59

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 26.09.22 um 15:37 schrieb Peter Zijlstra:
> On Mon, Sep 26, 2022 at 03:23:10PM +0200, Christian Borntraeger wrote:
>> Am 26.09.22 um 14:55 schrieb Peter Zijlstra:
>>
>>> Could you please test with something like the below on? I can boot that
>>> with KVM, but obviously I didn't suffer any weirdness to begin with :/
>>>
>>> ---
>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>> index 4e6a6417211f..ef9ccfc3a8c0 100644
>>> --- a/kernel/sched/core.c
>>> +++ b/kernel/sched/core.c
>>> @@ -4051,6 +4051,8 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
>>> unsigned long flags;
>>> int cpu, success = 0;
>>> + WARN_ON_ONCE(state & TASK_FREEZABLE);
>>> +
>>> preempt_disable();
>>> if (p == current) {
>>> /*
>>
>> Does not seem to trigger.
>
> Moo -- quite the puzzle this :/ I'll go stare at it more then.

Hmm,

#define ___wait_is_interruptible(state) \
(!__builtin_constant_p(state) || \
state == TASK_INTERRUPTIBLE || state == TASK_KILLABLE) \

That would not trigger when state is also TASK_FREEZABLE, no?

2022-09-26 18:30:45

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Mon, Sep 26, 2022 at 05:49:16PM +0200, Christian Borntraeger wrote:

> Hmm,
>
> #define ___wait_is_interruptible(state) \
> (!__builtin_constant_p(state) || \
> state == TASK_INTERRUPTIBLE || state == TASK_KILLABLE) \
>
> That would not trigger when state is also TASK_FREEZABLE, no?

Spot on!

signal_pending_state() writes that as:

state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)

which is the correct form.

diff --git a/include/linux/wait.h b/include/linux/wait.h
index 14ad8a0e9fac..7f5a51aae0a7 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -281,7 +281,7 @@ static inline void wake_up_pollfree(struct wait_queue_head *wq_head)

#define ___wait_is_interruptible(state) \
(!__builtin_constant_p(state) || \
- state == TASK_INTERRUPTIBLE || state == TASK_KILLABLE) \
+ (state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))

extern void init_wait_entry(struct wait_queue_entry *wq_entry, int flags);

Let me go git-grep some to see if there's more similar fail.

2022-09-26 19:18:52

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Mon, Sep 26, 2022 at 08:06:46PM +0200, Peter Zijlstra wrote:

> Let me go git-grep some to see if there's more similar fail.

I've ended up with the below...

---
include/linux/wait.h | 2 +-
kernel/hung_task.c | 8 ++++++--
kernel/sched/core.c | 2 +-
3 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/include/linux/wait.h b/include/linux/wait.h
index 14ad8a0e9fac..7f5a51aae0a7 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -281,7 +281,7 @@ static inline void wake_up_pollfree(struct wait_queue_head *wq_head)

#define ___wait_is_interruptible(state) \
(!__builtin_constant_p(state) || \
- state == TASK_INTERRUPTIBLE || state == TASK_KILLABLE) \
+ (state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))

extern void init_wait_entry(struct wait_queue_entry *wq_entry, int flags);

diff --git a/kernel/hung_task.c b/kernel/hung_task.c
index f1321c03c32a..4a8a713fd67b 100644
--- a/kernel/hung_task.c
+++ b/kernel/hung_task.c
@@ -191,6 +191,8 @@ static void check_hung_uninterruptible_tasks(unsigned long timeout)
hung_task_show_lock = false;
rcu_read_lock();
for_each_process_thread(g, t) {
+ unsigned int state;
+
if (!max_count--)
goto unlock;
if (time_after(jiffies, last_break + HUNG_TASK_LOCK_BREAK)) {
@@ -198,8 +200,10 @@ static void check_hung_uninterruptible_tasks(unsigned long timeout)
goto unlock;
last_break = jiffies;
}
- /* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
- if (READ_ONCE(t->__state) == TASK_UNINTERRUPTIBLE)
+ /* skip the TASK_KILLABLE tasks -- these can be killed */
+ state == READ_ONCE(t->__state);
+ if ((state & TASK_UNINTERRUPTIBLE) &&
+ !(state & TASK_WAKEKILL))
check_hung_task(t, timeout);
}
unlock:
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 1095917ed048..12ee5b98e2c4 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -8885,7 +8885,7 @@ state_filter_match(unsigned long state_filter, struct task_struct *p)
* When looking for TASK_UNINTERRUPTIBLE skip TASK_IDLE (allows
* TASK_KILLABLE).
*/
- if (state_filter == TASK_UNINTERRUPTIBLE && state == TASK_IDLE)
+ if (state_filter == TASK_UNINTERRUPTIBLE && state & TASK_NOLOAD)
return false;

return true;

2022-09-27 05:54:26

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 26.09.22 um 20:22 schrieb Peter Zijlstra:
> On Mon, Sep 26, 2022 at 08:06:46PM +0200, Peter Zijlstra wrote:
>
>> Let me go git-grep some to see if there's more similar fail.
>
> I've ended up with the below...

Tested-by: Christian Borntraeger <[email protected]>

Kind of scary that nobody else has reported any regression. I guess the freezable variant is just not used widely.
>
> ---
> include/linux/wait.h | 2 +-
> kernel/hung_task.c | 8 ++++++--
> kernel/sched/core.c | 2 +-
> 3 files changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/include/linux/wait.h b/include/linux/wait.h
> index 14ad8a0e9fac..7f5a51aae0a7 100644
> --- a/include/linux/wait.h
> +++ b/include/linux/wait.h
> @@ -281,7 +281,7 @@ static inline void wake_up_pollfree(struct wait_queue_head *wq_head)
>
> #define ___wait_is_interruptible(state) \
> (!__builtin_constant_p(state) || \
> - state == TASK_INTERRUPTIBLE || state == TASK_KILLABLE) \
> + (state & (TASK_INTERRUPTIBLE | TASK_WAKEKILL)))
>
> extern void init_wait_entry(struct wait_queue_entry *wq_entry, int flags);
>
> diff --git a/kernel/hung_task.c b/kernel/hung_task.c
> index f1321c03c32a..4a8a713fd67b 100644
> --- a/kernel/hung_task.c
> +++ b/kernel/hung_task.c
> @@ -191,6 +191,8 @@ static void check_hung_uninterruptible_tasks(unsigned long timeout)
> hung_task_show_lock = false;
> rcu_read_lock();
> for_each_process_thread(g, t) {
> + unsigned int state;
> +
> if (!max_count--)
> goto unlock;
> if (time_after(jiffies, last_break + HUNG_TASK_LOCK_BREAK)) {
> @@ -198,8 +200,10 @@ static void check_hung_uninterruptible_tasks(unsigned long timeout)
> goto unlock;
> last_break = jiffies;
> }
> - /* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
> - if (READ_ONCE(t->__state) == TASK_UNINTERRUPTIBLE)
> + /* skip the TASK_KILLABLE tasks -- these can be killed */
> + state == READ_ONCE(t->__state);
> + if ((state & TASK_UNINTERRUPTIBLE) &&
> + !(state & TASK_WAKEKILL))
> check_hung_task(t, timeout);
> }
> unlock:
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 1095917ed048..12ee5b98e2c4 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -8885,7 +8885,7 @@ state_filter_match(unsigned long state_filter, struct task_struct *p)
> * When looking for TASK_UNINTERRUPTIBLE skip TASK_IDLE (allows
> * TASK_KILLABLE).
> */
> - if (state_filter == TASK_UNINTERRUPTIBLE && state == TASK_IDLE)
> + if (state_filter == TASK_UNINTERRUPTIBLE && state & TASK_NOLOAD)
> return false;
>
> return true;

2022-09-28 06:08:56

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

Am 27.09.22 um 07:35 schrieb Christian Borntraeger:
>
>
> Am 26.09.22 um 20:22 schrieb Peter Zijlstra:
>> On Mon, Sep 26, 2022 at 08:06:46PM +0200, Peter Zijlstra wrote:
>>
>>> Let me go git-grep some to see if there's more similar fail.
>>
>> I've ended up with the below...
>
> Tested-by: Christian Borntraeger <[email protected]>
>
> Kind of scary that nobody else has reported any regression. I guess the freezable variant is just not used widely.

Will you queue this fix for next soon?

2022-10-21 17:27:51

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Mon, Aug 22, 2022 at 01:18:22PM +0200, Peter Zijlstra wrote:
> +#ifdef CONFIG_LOCKDEP
> + /*
> + * It's dangerous to freeze with locks held; there be dragons there.
> + */
> + if (!(state & __TASK_FREEZABLE_UNSAFE))
> + WARN_ON_ONCE(debug_locks && p->lockdep_depth);
> +#endif

We now seem to be hitting this sporadically in the intel gfx CI.

I've spotted it on two machines so far:
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12270/shard-tglb7/igt@gem_ctx_isolation@[email protected]
https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109950v1/shard-snb5/igt@kms_flip@[email protected]

Here's the full splat. Looks a bit funny since the
WARN()->printk()->console_lock() itself trips lockdep:

<6>[ 59.998117] PM: suspend entry (s2idle)
<6>[ 59.999878] Filesystems sync: 0.001 seconds
<6>[ 60.000881] Freezing user space processes ...
<4>[ 60.001059] ------------[ cut here ]------------
<4>[ 60.001071] ======================================================
<4>[ 60.001071] WARNING: possible circular locking dependency detected
<4>[ 60.001072] 6.1.0-rc1-CI_DRM_12270-ga9d18ead9885+ #1 Not tainted
<4>[ 60.001073] ------------------------------------------------------
<4>[ 60.001073] rtcwake/1152 is trying to acquire lock:
<4>[ 60.001074] ffffffff82735198 ((console_sem).lock){..-.}-{2:2}, at: down_trylock+0xa/0x30
<4>[ 60.001082]
but task is already holding lock:
<4>[ 60.001082] ffff888111a708e0 (&p->pi_lock){-.-.}-{2:2}, at: task_call_func+0x34/0xe0
<4>[ 60.001088]
which lock already depends on the new lock.

<4>[ 60.001089]
the existing dependency chain (in reverse order) is:
<4>[ 60.001089]
-> #1 (&p->pi_lock){-.-.}-{2:2}:
<4>[ 60.001091] lock_acquire+0xd3/0x310
<4>[ 60.001094] _raw_spin_lock_irqsave+0x33/0x50
<4>[ 60.001097] try_to_wake_up+0x6b/0x610
<4>[ 60.001098] up+0x3b/0x50
<4>[ 60.001099] __up_console_sem+0x5c/0x70
<4>[ 60.001102] console_unlock+0x1bc/0x1d0
<4>[ 60.001104] do_con_write+0x654/0xa20
<4>[ 60.001108] con_write+0xa/0x20
<4>[ 60.001110] do_output_char+0x119/0x1e0
<4>[ 60.001113] n_tty_write+0x20f/0x490
<4>[ 60.001114] file_tty_write.isra.29+0x17d/0x320
<4>[ 60.001117] do_iter_readv_writev+0xdb/0x140
<4>[ 60.001120] do_iter_write+0x6c/0x1a0
<4>[ 60.001121] vfs_writev+0x97/0x290
<4>[ 60.001123] do_writev+0x63/0x110
<4>[ 60.001125] do_syscall_64+0x37/0x90
<4>[ 60.001128] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 60.001130]
-> #0 ((console_sem).lock){..-.}-{2:2}:
<4>[ 60.001131] validate_chain+0xb3d/0x2000
<4>[ 60.001132] __lock_acquire+0x5a4/0xb70
<4>[ 60.001133] lock_acquire+0xd3/0x310
<4>[ 60.001134] _raw_spin_lock_irqsave+0x33/0x50
<4>[ 60.001136] down_trylock+0xa/0x30
<4>[ 60.001137] __down_trylock_console_sem+0x25/0xb0
<4>[ 60.001139] console_trylock+0xe/0x70
<4>[ 60.001140] vprintk_emit+0x13c/0x380
<4>[ 60.001142] _printk+0x53/0x6e
<4>[ 60.001145] report_bug.cold.2+0x10/0x52
<4>[ 60.001147] handle_bug+0x3f/0x70
<4>[ 60.001148] exc_invalid_op+0x13/0x60
<4>[ 60.001150] asm_exc_invalid_op+0x16/0x20
<4>[ 60.001152] __set_task_frozen+0x58/0x80
<4>[ 60.001156] task_call_func+0xc2/0xe0
<4>[ 60.001157] freeze_task+0x84/0xe0
<4>[ 60.001159] try_to_freeze_tasks+0xac/0x260
<4>[ 60.001160] freeze_processes+0x56/0xb0
<4>[ 60.001162] pm_suspend.cold.7+0x1d9/0x31c
<4>[ 60.001164] state_store+0x7b/0xe0
<4>[ 60.001165] kernfs_fop_write_iter+0x121/0x1c0
<4>[ 60.001169] vfs_write+0x34c/0x4e0
<4>[ 60.001170] ksys_write+0x57/0xd0
<4>[ 60.001172] do_syscall_64+0x37/0x90
<4>[ 60.001174] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 60.001176]
other info that might help us debug this:

<4>[ 60.001176] Possible unsafe locking scenario:

<4>[ 60.001176] CPU0 CPU1
<4>[ 60.001176] ---- ----
<4>[ 60.001177] lock(&p->pi_lock);
<4>[ 60.001177] lock((console_sem).lock);
<4>[ 60.001178] lock(&p->pi_lock);
<4>[ 60.001179] lock((console_sem).lock);
<4>[ 60.001179]
*** DEADLOCK ***

<4>[ 60.001180] 7 locks held by rtcwake/1152:
<4>[ 60.001180] #0: ffff888105e99430 (sb_writers#5){.+.+}-{0:0}, at: ksys_write+0x57/0xd0
<4>[ 60.001184] #1: ffff88810a048288 (&of->mutex){+.+.}-{3:3}, at: kernfs_fop_write_iter+0xee/0x1c0
<4>[ 60.001191] #2: ffff888100c58538 (kn->active#155){.+.+}-{0:0}, at: kernfs_fop_write_iter+0xf7/0x1c0
<4>[ 60.001194] #3: ffffffff8264db08 (system_transition_mutex){+.+.}-{3:3}, at: pm_suspend.cold.7+0xfa/0x31c
<4>[ 60.001197] #4: ffffffff82606098 (tasklist_lock){.+.+}-{2:2}, at: try_to_freeze_tasks+0x63/0x260
<4>[ 60.001201] #5: ffffffff8273aed8 (freezer_lock){....}-{2:2}, at: freeze_task+0x27/0xe0
<4>[ 60.001204] #6: ffff888111a708e0 (&p->pi_lock){-.-.}-{2:2}, at: task_call_func+0x34/0xe0
<4>[ 60.001207]
stack backtrace:
<4>[ 60.001207] CPU: 2 PID: 1152 Comm: rtcwake Not tainted 6.1.0-rc1-CI_DRM_12270-ga9d18ead9885+ #1
<4>[ 60.001210] Hardware name: Intel Corporation Tiger Lake Client Platform/TigerLake U DDR4 SODIMM RVP, BIOS TGLSFWI1.R00.3197.A00.2005110542 05/11/2020
<4>[ 60.001211] Call Trace:
<4>[ 60.001211] <TASK>
<4>[ 60.001212] dump_stack_lvl+0x56/0x7f
<4>[ 60.001215] check_noncircular+0x132/0x150
<4>[ 60.001217] validate_chain+0xb3d/0x2000
<4>[ 60.001220] __lock_acquire+0x5a4/0xb70
<4>[ 60.001222] lock_acquire+0xd3/0x310
<4>[ 60.001223] ? down_trylock+0xa/0x30
<4>[ 60.001226] ? vprintk_emit+0x13c/0x380
<4>[ 60.001228] _raw_spin_lock_irqsave+0x33/0x50
<4>[ 60.001230] ? down_trylock+0xa/0x30
<4>[ 60.001231] down_trylock+0xa/0x30
<4>[ 60.001233] __down_trylock_console_sem+0x25/0xb0
<4>[ 60.001234] console_trylock+0xe/0x70
<4>[ 60.001235] vprintk_emit+0x13c/0x380
<4>[ 60.001237] _printk+0x53/0x6e
<4>[ 60.001240] ? __set_task_frozen+0x58/0x80
<4>[ 60.001241] report_bug.cold.2+0x10/0x52
<4>[ 60.001244] handle_bug+0x3f/0x70
<4>[ 60.001245] exc_invalid_op+0x13/0x60
<4>[ 60.001247] asm_exc_invalid_op+0x16/0x20
<4>[ 60.001250] RIP: 0010:__set_task_frozen+0x58/0x80
<4>[ 60.001252] Code: f7 c5 00 20 00 00 74 06 40 f6 c5 03 74 3a 81 e5 00 40 00 00 75 16 8b 15 a2 b9 71 01 85 d2 74 0c 8b 83 60 09 00 00 85 c0 74 02 <0f> 0b c7 43 18 00 80 00 00 b8 00 80 00 00 5b 5d c3 cc cc cc cc 31
<4>[ 60.001254] RSP: 0018:ffffc9000335fcf0 EFLAGS: 00010002
<4>[ 60.001255] RAX: 0000000000000001 RBX: ffff888111a70040 RCX: 0000000000000000
<4>[ 60.001256] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff888111a70040
<4>[ 60.001257] RBP: 0000000000000000 R08: 0000000000000001 R09: 00000000fffffffe
<4>[ 60.001258] R10: 0000000001e6f6b9 R11: 00000000934a4c67 R12: 0000000000000246
<4>[ 60.001259] R13: ffffffff811653e0 R14: 0000000000000000 R15: ffff888111a70040
<4>[ 60.001260] ? __set_task_special+0x40/0x40
<4>[ 60.001263] task_call_func+0xc2/0xe0
<4>[ 60.001265] freeze_task+0x84/0xe0
<4>[ 60.001267] try_to_freeze_tasks+0xac/0x260
<4>[ 60.001270] freeze_processes+0x56/0xb0
<4>[ 60.001272] pm_suspend.cold.7+0x1d9/0x31c
<4>[ 60.001274] state_store+0x7b/0xe0
<4>[ 60.001276] kernfs_fop_write_iter+0x121/0x1c0
<4>[ 60.001278] vfs_write+0x34c/0x4e0
<4>[ 60.001281] ksys_write+0x57/0xd0
<4>[ 60.001284] do_syscall_64+0x37/0x90
<4>[ 60.001285] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 60.001288] RIP: 0033:0x7fb4705521e7
<4>[ 60.001289] Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 24
<4>[ 60.001290] RSP: 002b:00007ffe3efac3d8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
<4>[ 60.001291] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007fb4705521e7
<4>[ 60.001292] RDX: 0000000000000004 RSI: 0000559af7969590 RDI: 000000000000000b
<4>[ 60.001293] RBP: 0000559af7969590 R08: 0000000000000000 R09: 0000000000000004
<4>[ 60.001293] R10: 0000559af60922a6 R11: 0000000000000246 R12: 0000000000000004
<4>[ 60.001294] R13: 0000559af7967540 R14: 00007fb47062e4a0 R15: 00007fb47062d8a0
<4>[ 60.001296] </TASK>
<4>[ 60.001634] WARNING: CPU: 2 PID: 1152 at kernel/freezer.c:129 __set_task_frozen+0x58/0x80
<4>[ 60.001641] Modules linked in: fuse snd_hda_codec_hdmi i915 x86_pkg_temp_thermal coretemp mei_pxp mei_hdcp kvm_intel wmi_bmof snd_hda_intel kvm snd_intel_dspcfg prime_numbers snd_hda_codec cdc_ether ttm e1000e irqbypass snd_hwdep crct10dif_pclmul usbnet drm_buddy crc32_pclmul mii ghash_clmulni_intel snd_hda_core drm_display_helper ptp i2c_i801 pps_core mei_me drm_kms_helper i2c_smbus snd_pcm syscopyarea mei sysfillrect sysimgblt intel_lpss_pci fb_sys_fops video wmi
<4>[ 60.001717] CPU: 2 PID: 1152 Comm: rtcwake Not tainted 6.1.0-rc1-CI_DRM_12270-ga9d18ead9885+ #1
<4>[ 60.001723] Hardware name: Intel Corporation Tiger Lake Client Platform/TigerLake U DDR4 SODIMM RVP, BIOS TGLSFWI1.R00.3197.A00.2005110542 05/11/2020
<4>[ 60.001729] RIP: 0010:__set_task_frozen+0x58/0x80
<4>[ 60.001735] Code: f7 c5 00 20 00 00 74 06 40 f6 c5 03 74 3a 81 e5 00 40 00 00 75 16 8b 15 a2 b9 71 01 85 d2 74 0c 8b 83 60 09 00 00 85 c0 74 02 <0f> 0b c7 43 18 00 80 00 00 b8 00 80 00 00 5b 5d c3 cc cc cc cc 31
<4>[ 60.001744] RSP: 0018:ffffc9000335fcf0 EFLAGS: 00010002
<4>[ 60.001747] RAX: 0000000000000001 RBX: ffff888111a70040 RCX: 0000000000000000
<4>[ 60.001751] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff888111a70040
<4>[ 60.001757] RBP: 0000000000000000 R08: 0000000000000001 R09: 00000000fffffffe
<4>[ 60.001763] R10: 0000000001e6f6b9 R11: 00000000934a4c67 R12: 0000000000000246
<4>[ 60.001769] R13: ffffffff811653e0 R14: 0000000000000000 R15: ffff888111a70040
<4>[ 60.001776] FS: 00007fb47043e740(0000) GS:ffff8884a0300000(0000) knlGS:0000000000000000
<4>[ 60.001784] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
<4>[ 60.001789] CR2: 00007f3903c603d8 CR3: 000000010a25a003 CR4: 0000000000770ee0
<4>[ 60.001795] PKRU: 55555554
<4>[ 60.001798] Call Trace:
<4>[ 60.001801] <TASK>
<4>[ 60.001804] task_call_func+0xc2/0xe0
<4>[ 60.001809] freeze_task+0x84/0xe0
<4>[ 60.001815] try_to_freeze_tasks+0xac/0x260
<4>[ 60.001821] freeze_processes+0x56/0xb0
<4>[ 60.001826] pm_suspend.cold.7+0x1d9/0x31c
<4>[ 60.001832] state_store+0x7b/0xe0
<4>[ 60.001837] kernfs_fop_write_iter+0x121/0x1c0
<4>[ 60.001843] vfs_write+0x34c/0x4e0
<4>[ 60.001850] ksys_write+0x57/0xd0
<4>[ 60.001855] do_syscall_64+0x37/0x90
<4>[ 60.001860] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 60.001866] RIP: 0033:0x7fb4705521e7
<4>[ 60.001870] Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 24
<4>[ 60.001884] RSP: 002b:00007ffe3efac3d8 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
<4>[ 60.001892] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007fb4705521e7
<4>[ 60.001898] RDX: 0000000000000004 RSI: 0000559af7969590 RDI: 000000000000000b
<4>[ 60.001904] RBP: 0000559af7969590 R08: 0000000000000000 R09: 0000000000000004
<4>[ 60.001910] R10: 0000559af60922a6 R11: 0000000000000246 R12: 0000000000000004
<4>[ 60.001917] R13: 0000559af7967540 R14: 00007fb47062e4a0 R15: 00007fb47062d8a0
<4>[ 60.001925] </TASK>
<4>[ 60.001928] irq event stamp: 8712
<4>[ 60.001931] hardirqs last enabled at (8711): [<ffffffff81b73784>] _raw_spin_unlock_irqrestore+0x54/0x70
<4>[ 60.001941] hardirqs last disabled at (8712): [<ffffffff81b7352b>] _raw_spin_lock_irqsave+0x4b/0x50
<4>[ 60.001950] softirqs last enabled at (8348): [<ffffffff81e0031e>] __do_softirq+0x31e/0x48a
<4>[ 60.001957] softirqs last disabled at (8341): [<ffffffff810c1b08>] irq_exit_rcu+0xb8/0xe0
<4>[ 60.001969] ---[ end trace 0000000000000000 ]---
<4>[ 60.003326] (elapsed 0.002 seconds) done.
<6>[ 60.003332] OOM killer disabled.
<6>[ 60.003334] Freezing remaining freezable tasks ... (elapsed 0.006 seconds) done.
<6>[ 60.010062] printk: Suspending console(s) (use no_console_suspend to debug)
<6>[ 60.041543] e1000e: EEE TX LPI TIMER: 00000011
<6>[ 60.368938] ACPI: EC: interrupt blocked

--
Ville Syrj?l?
Intel

2022-10-25 05:00:56

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Fri, Oct 21, 2022 at 08:22:41PM +0300, Ville Syrj?l? wrote:
> On Mon, Aug 22, 2022 at 01:18:22PM +0200, Peter Zijlstra wrote:
> > +#ifdef CONFIG_LOCKDEP
> > + /*
> > + * It's dangerous to freeze with locks held; there be dragons there.
> > + */
> > + if (!(state & __TASK_FREEZABLE_UNSAFE))
> > + WARN_ON_ONCE(debug_locks && p->lockdep_depth);
> > +#endif
>
> We now seem to be hitting this sporadically in the intel gfx CI.
>
> I've spotted it on two machines so far:
> https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12270/shard-tglb7/igt@gem_ctx_isolation@[email protected]
> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109950v1/shard-snb5/igt@kms_flip@[email protected]

Sadly no luck in reproducing this locally so far. In the meantime
I added the following patch into our topic/core-for-CI branch in
the hopes of CI stumbling on it again and dumping a bit more data:

--- a/kernel/freezer.c
+++ b/kernel/freezer.c
@@ -125,8 +125,16 @@ static int __set_task_frozen(struct task_struct *p, void *arg)
/*
* It's dangerous to freeze with locks held; there be dragons there.
*/
- if (!(state & __TASK_FREEZABLE_UNSAFE))
- WARN_ON_ONCE(debug_locks && p->lockdep_depth);
+ if (!(state & __TASK_FREEZABLE_UNSAFE)) {
+ static bool warned = false;
+
+ if (!warned && debug_locks && p->lockdep_depth) {
+ debug_show_held_locks(p);
+ WARN(1, "%s/%d holding locks while freezing\n",
+ p->comm, task_pid_nr(p));
+ warned = true;
+ }
+ }
#endif

WRITE_ONCE(p->__state, TASK_FROZEN);
--
2.37.4

--
Ville Syrj?l?
Intel

2022-10-25 11:01:14

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Tue, Oct 25, 2022 at 07:52:07AM +0300, Ville Syrj?l? wrote:
> On Fri, Oct 21, 2022 at 08:22:41PM +0300, Ville Syrj?l? wrote:
> > On Mon, Aug 22, 2022 at 01:18:22PM +0200, Peter Zijlstra wrote:
> > > +#ifdef CONFIG_LOCKDEP
> > > + /*
> > > + * It's dangerous to freeze with locks held; there be dragons there.
> > > + */
> > > + if (!(state & __TASK_FREEZABLE_UNSAFE))
> > > + WARN_ON_ONCE(debug_locks && p->lockdep_depth);
> > > +#endif
> >
> > We now seem to be hitting this sporadically in the intel gfx CI.
> >
> > I've spotted it on two machines so far:
> > https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12270/shard-tglb7/igt@gem_ctx_isolation@[email protected]
> > https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109950v1/shard-snb5/igt@kms_flip@[email protected]
>
> Sadly no luck in reproducing this locally so far. In the meantime
> I added the following patch into our topic/core-for-CI branch in
> the hopes of CI stumbling on it again and dumping a bit more data:
>
> --- a/kernel/freezer.c
> +++ b/kernel/freezer.c
> @@ -125,8 +125,16 @@ static int __set_task_frozen(struct task_struct *p, void *arg)
> /*
> * It's dangerous to freeze with locks held; there be dragons there.
> */
> - if (!(state & __TASK_FREEZABLE_UNSAFE))
> - WARN_ON_ONCE(debug_locks && p->lockdep_depth);
> + if (!(state & __TASK_FREEZABLE_UNSAFE)) {
> + static bool warned = false;
> +
> + if (!warned && debug_locks && p->lockdep_depth) {
> + debug_show_held_locks(p);
> + WARN(1, "%s/%d holding locks while freezing\n",
> + p->comm, task_pid_nr(p));
> + warned = true;
> + }
> + }
> #endif
>
> WRITE_ONCE(p->__state, TASK_FROZEN);

That seems reasonable. But note that this constraint isn't new; the
previous freezer had much the same constraint but perhaps it wasn't
triggered for mysterious raisins. see the previous
try_to_freeze_unsafe() function.

2022-10-26 10:40:52

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Tue, Oct 25, 2022 at 12:49:13PM +0200, Peter Zijlstra wrote:
> On Tue, Oct 25, 2022 at 07:52:07AM +0300, Ville Syrj?l? wrote:
> > On Fri, Oct 21, 2022 at 08:22:41PM +0300, Ville Syrj?l? wrote:
> > > On Mon, Aug 22, 2022 at 01:18:22PM +0200, Peter Zijlstra wrote:
> > > > +#ifdef CONFIG_LOCKDEP
> > > > + /*
> > > > + * It's dangerous to freeze with locks held; there be dragons there.
> > > > + */
> > > > + if (!(state & __TASK_FREEZABLE_UNSAFE))
> > > > + WARN_ON_ONCE(debug_locks && p->lockdep_depth);
> > > > +#endif
> > >
> > > We now seem to be hitting this sporadically in the intel gfx CI.
> > >
> > > I've spotted it on two machines so far:
> > > https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12270/shard-tglb7/igt@gem_ctx_isolation@[email protected]
> > > https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_109950v1/shard-snb5/igt@kms_flip@[email protected]
> >
> > Sadly no luck in reproducing this locally so far. In the meantime
> > I added the following patch into our topic/core-for-CI branch in
> > the hopes of CI stumbling on it again and dumping a bit more data:
> >
> > --- a/kernel/freezer.c
> > +++ b/kernel/freezer.c
> > @@ -125,8 +125,16 @@ static int __set_task_frozen(struct task_struct *p, void *arg)
> > /*
> > * It's dangerous to freeze with locks held; there be dragons there.
> > */
> > - if (!(state & __TASK_FREEZABLE_UNSAFE))
> > - WARN_ON_ONCE(debug_locks && p->lockdep_depth);
> > + if (!(state & __TASK_FREEZABLE_UNSAFE)) {
> > + static bool warned = false;
> > +
> > + if (!warned && debug_locks && p->lockdep_depth) {
> > + debug_show_held_locks(p);
> > + WARN(1, "%s/%d holding locks while freezing\n",
> > + p->comm, task_pid_nr(p));
> > + warned = true;
> > + }
> > + }
> > #endif
> >
> > WRITE_ONCE(p->__state, TASK_FROZEN);
>
> That seems reasonable. But note that this constraint isn't new; the
> previous freezer had much the same constraint but perhaps it wasn't
> triggered for mysterious raisins. see the previous
> try_to_freeze_unsafe() function.

Looks like we caught one with the extra debugs now.

Short form looks to be this:
<4>[ 355.437846] 1 lock held by rs:main Q:Reg/359:
<4>[ 355.438418] #0: ffff88844693b758 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x1b/0x30
<4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing

Based on a quick google that process seems to be some rsyslog thing.

Here's the full splat with the console_lock mess included:
<6>[ 355.437502] Freezing user space processes ...
<4>[ 355.437846] 1 lock held by rs:main Q:Reg/359:

<4>[ 355.437865] ======================================================
<4>[ 355.437866] WARNING: possible circular locking dependency detected
<4>[ 355.437867] 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1 Tainted: G U
<4>[ 355.437870] ------------------------------------------------------
<4>[ 355.437871] rtcwake/6211 is trying to acquire lock:
<4>[ 355.437872] ffffffff82735198 ((console_sem).lock){-.-.}-{2:2}, at: down_trylock+0xa/0x30
<4>[ 355.437883]
but task is already holding lock:
<4>[ 355.437885] ffff88810d0908e0 (&p->pi_lock){-.-.}-{2:2}, at: task_call_func+0x34/0xe0
<4>[ 355.437893]
which lock already depends on the new lock.

<4>[ 355.437894]
the existing dependency chain (in reverse order) is:
<4>[ 355.437895]
-> #1 (&p->pi_lock){-.-.}-{2:2}:
<4>[ 355.437899] lock_acquire+0xd3/0x310
<4>[ 355.437903] _raw_spin_lock_irqsave+0x33/0x50
<4>[ 355.437907] try_to_wake_up+0x6b/0x610
<4>[ 355.437911] up+0x3b/0x50
<4>[ 355.437914] __up_console_sem+0x5c/0x70
<4>[ 355.437917] console_unlock+0x1bc/0x1d0
<4>[ 355.437920] con_font_op+0x2e2/0x3a0
<4>[ 355.437925] vt_ioctl+0x4f5/0x13b0
<4>[ 355.437930] tty_ioctl+0x233/0x8e0
<4>[ 355.437934] __x64_sys_ioctl+0x71/0xb0
<4>[ 355.437938] do_syscall_64+0x3a/0x90
<4>[ 355.437943] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 355.437948]
-> #0 ((console_sem).lock){-.-.}-{2:2}:
<4>[ 355.437952] validate_chain+0xb3d/0x2000
<4>[ 355.437955] __lock_acquire+0x5a4/0xb70
<4>[ 355.437958] lock_acquire+0xd3/0x310
<4>[ 355.437960] _raw_spin_lock_irqsave+0x33/0x50
<4>[ 355.437965] down_trylock+0xa/0x30
<4>[ 355.437968] __down_trylock_console_sem+0x25/0xb0
<4>[ 355.437971] console_trylock+0xe/0x70
<4>[ 355.437974] vprintk_emit+0x13c/0x380
<4>[ 355.437977] _printk+0x53/0x6e
<4>[ 355.437981] lockdep_print_held_locks+0x5c/0xab
<4>[ 355.437985] __set_task_frozen+0x6d/0xb0
<4>[ 355.437989] task_call_func+0xc4/0xe0
<4>[ 355.437993] freeze_task+0x84/0xe0
<4>[ 355.437997] try_to_freeze_tasks+0xac/0x260
<4>[ 355.438001] freeze_processes+0x56/0xb0
<4>[ 355.438005] pm_suspend.cold.7+0x1d9/0x31c
<4>[ 355.438008] state_store+0x7b/0xe0
<4>[ 355.438012] kernfs_fop_write_iter+0x124/0x1c0
<4>[ 355.438016] vfs_write+0x34f/0x4e0
<4>[ 355.438021] ksys_write+0x57/0xd0
<4>[ 355.438025] do_syscall_64+0x3a/0x90
<4>[ 355.438029] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 355.438034]
other info that might help us debug this:

<4>[ 355.438035] Possible unsafe locking scenario:

<4>[ 355.438036] CPU0 CPU1
<4>[ 355.438037] ---- ----
<4>[ 355.438037] lock(&p->pi_lock);
<4>[ 355.438040] lock((console_sem).lock);
<4>[ 355.438042] lock(&p->pi_lock);
<4>[ 355.438044] lock((console_sem).lock);
<4>[ 355.438046]
*** DEADLOCK ***

<4>[ 355.438047] 7 locks held by rtcwake/6211:
<4>[ 355.438049] #0: ffff888104d11430 (sb_writers#5){.+.+}-{0:0}, at: ksys_write+0x57/0xd0
<4>[ 355.438058] #1: ffff88810e6bac88 (&of->mutex){+.+.}-{3:3}, at: kernfs_fop_write_iter+0xee/0x1c0
<4>[ 355.438066] #2: ffff8881001c0538 (kn->active#167){.+.+}-{0:0}, at: kernfs_fop_write_iter+0xf7/0x1c0
<4>[ 355.438074] #3: ffffffff8264db08 (system_transition_mutex){+.+.}-{3:3}, at: pm_suspend.cold.7+0xfa/0x31c
<4>[ 355.438082] #4: ffffffff82606098 (tasklist_lock){.+.+}-{2:2}, at: try_to_freeze_tasks+0x63/0x260
<4>[ 355.438090] #5: ffffffff8273aed8 (freezer_lock){....}-{2:2}, at: freeze_task+0x27/0xe0
<4>[ 355.438098] #6: ffff88810d0908e0 (&p->pi_lock){-.-.}-{2:2}, at: task_call_func+0x34/0xe0
<4>[ 355.438105]
stack backtrace:
<4>[ 355.438107] CPU: 0 PID: 6211 Comm: rtcwake Tainted: G U 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1
<4>[ 355.438110] Hardware name: /NUC5i7RYB, BIOS RYBDWi35.86A.0385.2020.0519.1558 05/19/2020
<4>[ 355.438112] Call Trace:
<4>[ 355.438114] <TASK>
<4>[ 355.438116] dump_stack_lvl+0x56/0x7f
<4>[ 355.438121] check_noncircular+0x132/0x150
<4>[ 355.438125] ? validate_chain+0x247/0x2000
<4>[ 355.438131] validate_chain+0xb3d/0x2000
<4>[ 355.438138] __lock_acquire+0x5a4/0xb70
<4>[ 355.438144] lock_acquire+0xd3/0x310
<4>[ 355.438147] ? down_trylock+0xa/0x30
<4>[ 355.438154] ? vprintk_emit+0x13c/0x380
<4>[ 355.438158] _raw_spin_lock_irqsave+0x33/0x50
<4>[ 355.438163] ? down_trylock+0xa/0x30
<4>[ 355.438167] down_trylock+0xa/0x30
<4>[ 355.438171] __down_trylock_console_sem+0x25/0xb0
<4>[ 355.438175] console_trylock+0xe/0x70
<4>[ 355.438178] vprintk_emit+0x13c/0x380
<4>[ 355.438183] ? __set_task_special+0x40/0x40
<4>[ 355.438187] _printk+0x53/0x6e
<4>[ 355.438195] lockdep_print_held_locks+0x5c/0xab
<4>[ 355.438199] ? __set_task_special+0x40/0x40
<4>[ 355.438203] __set_task_frozen+0x6d/0xb0
<4>[ 355.438208] task_call_func+0xc4/0xe0
<4>[ 355.438214] freeze_task+0x84/0xe0
<4>[ 355.438219] try_to_freeze_tasks+0xac/0x260
<4>[ 355.438226] freeze_processes+0x56/0xb0
<4>[ 355.438230] pm_suspend.cold.7+0x1d9/0x31c
<4>[ 355.438235] state_store+0x7b/0xe0
<4>[ 355.438241] kernfs_fop_write_iter+0x124/0x1c0
<4>[ 355.438247] vfs_write+0x34f/0x4e0
<4>[ 355.438255] ksys_write+0x57/0xd0
<4>[ 355.438261] do_syscall_64+0x3a/0x90
<4>[ 355.438266] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 355.438271] RIP: 0033:0x7fcfa44d80a7
<4>[ 355.438275] Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 24
<4>[ 355.438278] RSP: 002b:00007ffd72160e28 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
<4>[ 355.438281] RAX: ffffffffffffffda RBX: 0000000000000007 RCX: 00007fcfa44d80a7
<4>[ 355.438284] RDX: 0000000000000007 RSI: 000055dd45bf4590 RDI: 000000000000000b
<4>[ 355.438286] RBP: 000055dd45bf4590 R08: 0000000000000000 R09: 0000000000000007
<4>[ 355.438288] R10: 000055dd441d22a6 R11: 0000000000000246 R12: 0000000000000007
<4>[ 355.438290] R13: 000055dd45bf2540 R14: 00007fcfa45b34a0 R15: 00007fcfa45b28a0
<4>[ 355.438298] </TASK>
<4>[ 355.438418] #0: ffff88844693b758 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x1b/0x30
<4>[ 355.438429] ------------[ cut here ]------------
<4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
<4>[ 355.438439] WARNING: CPU: 0 PID: 6211 at kernel/freezer.c:134 __set_task_frozen+0x86/0xb0
<4>[ 355.438447] Modules linked in: snd_hda_intel i915 mei_hdcp mei_pxp drm_display_helper drm_kms_helper vgem drm_shmem_helper snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core snd_pcm prime_numbers ttm drm_buddy syscopyarea sysfillrect sysimgblt fb_sys_fops fuse x86_pkg_temp_thermal coretemp kvm_intel btusb btrtl btbcm btintel kvm irqbypass bluetooth crct10dif_pclmul crc32_pclmul ecdh_generic ghash_clmulni_intel ecc e1000e mei_me i2c_i801 ptp mei i2c_smbus pps_core lpc_ich video wmi [last unloaded: drm_kms_helper]
<4>[ 355.438521] CPU: 0 PID: 6211 Comm: rtcwake Tainted: G U 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1
<4>[ 355.438526] Hardware name: /NUC5i7RYB, BIOS RYBDWi35.86A.0385.2020.0519.1558 05/19/2020
<4>[ 355.438530] RIP: 0010:__set_task_frozen+0x86/0xb0
<4>[ 355.438536] Code: 83 60 09 00 00 85 c0 74 2a 48 89 df e8 ac 02 9b 00 8b 93 38 05 00 00 48 8d b3 48 07 00 00 48 c7 c7 a0 62 2b 82 e8 ee c1 9a 00 <0f> 0b c6 05 51 75 e3 02 01 c7 43 18 00 80 00 00 b8 00 80 00 00 5b
<4>[ 355.438541] RSP: 0018:ffffc900012cbcf0 EFLAGS: 00010086
<4>[ 355.438546] RAX: 0000000000000000 RBX: ffff88810d090040 RCX: 0000000000000004
<4>[ 355.438550] RDX: 0000000000000004 RSI: 00000000fffff5de RDI: 00000000ffffffff
<4>[ 355.438553] RBP: 0000000000000000 R08: 0000000000000000 R09: c0000000fffff5de
<4>[ 355.438557] R10: 00000000002335f8 R11: ffffc900012cbb88 R12: 0000000000000246
<4>[ 355.438561] R13: ffffffff81165430 R14: 0000000000000000 R15: ffff88810d090040
<4>[ 355.438565] FS: 00007fcfa43c7740(0000) GS:ffff888446800000(0000) knlGS:0000000000000000
<4>[ 355.438569] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
<4>[ 355.438582] CR2: 00007fceb380f6b8 CR3: 0000000117c5c004 CR4: 00000000003706f0
<4>[ 355.438586] Call Trace:
<4>[ 355.438589] <TASK>
<4>[ 355.438592] task_call_func+0xc4/0xe0
<4>[ 355.438600] freeze_task+0x84/0xe0
<4>[ 355.438607] try_to_freeze_tasks+0xac/0x260
<4>[ 355.438616] freeze_processes+0x56/0xb0
<4>[ 355.438622] pm_suspend.cold.7+0x1d9/0x31c
<4>[ 355.438629] state_store+0x7b/0xe0
<4>[ 355.438637] kernfs_fop_write_iter+0x124/0x1c0
<4>[ 355.438644] vfs_write+0x34f/0x4e0
<4>[ 355.438655] ksys_write+0x57/0xd0
<4>[ 355.438663] do_syscall_64+0x3a/0x90
<4>[ 355.438670] entry_SYSCALL_64_after_hwframe+0x63/0xcd
<4>[ 355.438676] RIP: 0033:0x7fcfa44d80a7
<4>[ 355.438681] Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 24
<4>[ 355.438685] RSP: 002b:00007ffd72160e28 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
<4>[ 355.438690] RAX: ffffffffffffffda RBX: 0000000000000007 RCX: 00007fcfa44d80a7
<4>[ 355.438695] RDX: 0000000000000007 RSI: 000055dd45bf4590 RDI: 000000000000000b
<4>[ 355.438698] RBP: 000055dd45bf4590 R08: 0000000000000000 R09: 0000000000000007
<4>[ 355.438702] R10: 000055dd441d22a6 R11: 0000000000000246 R12: 0000000000000007
<4>[ 355.438706] R13: 000055dd45bf2540 R14: 00007fcfa45b34a0 R15: 00007fcfa45b28a0
<4>[ 355.438716] </TASK>
<4>[ 355.438718] irq event stamp: 7462
<4>[ 355.438721] hardirqs last enabled at (7461): [<ffffffff81b73764>] _raw_spin_unlock_irqrestore+0x54/0x70
<4>[ 355.438729] hardirqs last disabled at (7462): [<ffffffff81b7350b>] _raw_spin_lock_irqsave+0x4b/0x50
<4>[ 355.438736] softirqs last enabled at (7322): [<ffffffff81e0031e>] __do_softirq+0x31e/0x48a
<4>[ 355.438742] softirqs last disabled at (7313): [<ffffffff810c1b58>] irq_exit_rcu+0xb8/0xe0
<4>[ 355.438749] ---[ end trace 0000000000000000 ]---
<4>[ 355.440204] (elapsed 0.002 seconds) done.
<6>[ 355.440210] OOM killer disabled.
<6>[ 355.440212] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.

--
Ville Syrj?l?
Intel

2022-10-26 11:53:35

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Wed, Oct 26, 2022 at 01:32:31PM +0300, Ville Syrj?l? wrote:
> Short form looks to be this:
> <4>[ 355.437846] 1 lock held by rs:main Q:Reg/359:
> <4>[ 355.438418] #0: ffff88844693b758 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x1b/0x30
> <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing

> <4>[ 355.438429] ------------[ cut here ]------------
> <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
> <4>[ 355.438439] WARNING: CPU: 0 PID: 6211 at kernel/freezer.c:134 __set_task_frozen+0x86/0xb0
> <4>[ 355.438447] Modules linked in: snd_hda_intel i915 mei_hdcp mei_pxp drm_display_helper drm_kms_helper vgem drm_shmem_helper snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core snd_pcm prime_numbers ttm drm_buddy syscopyarea sysfillrect sysimgblt fb_sys_fops fuse x86_pkg_temp_thermal coretemp kvm_intel btusb btrtl btbcm btintel kvm irqbypass bluetooth crct10dif_pclmul crc32_pclmul ecdh_generic ghash_clmulni_intel ecc e1000e mei_me i2c_i801 ptp mei i2c_smbus pps_core lpc_ich video wmi [last unloaded: drm_kms_helper]
> <4>[ 355.438521] CPU: 0 PID: 6211 Comm: rtcwake Tainted: G U 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1
> <4>[ 355.438526] Hardware name: /NUC5i7RYB, BIOS RYBDWi35.86A.0385.2020.0519.1558 05/19/2020
> <4>[ 355.438530] RIP: 0010:__set_task_frozen+0x86/0xb0
> <4>[ 355.438536] Code: 83 60 09 00 00 85 c0 74 2a 48 89 df e8 ac 02 9b 00 8b 93 38 05 00 00 48 8d b3 48 07 00 00 48 c7 c7 a0 62 2b 82 e8 ee c1 9a 00 <0f> 0b c6 05 51 75 e3 02 01 c7 43 18 00 80 00 00 b8 00 80 00 00 5b
> <4>[ 355.438541] RSP: 0018:ffffc900012cbcf0 EFLAGS: 00010086
> <4>[ 355.438546] RAX: 0000000000000000 RBX: ffff88810d090040 RCX: 0000000000000004
> <4>[ 355.438550] RDX: 0000000000000004 RSI: 00000000fffff5de RDI: 00000000ffffffff
> <4>[ 355.438553] RBP: 0000000000000000 R08: 0000000000000000 R09: c0000000fffff5de
> <4>[ 355.438557] R10: 00000000002335f8 R11: ffffc900012cbb88 R12: 0000000000000246
> <4>[ 355.438561] R13: ffffffff81165430 R14: 0000000000000000 R15: ffff88810d090040
> <4>[ 355.438565] FS: 00007fcfa43c7740(0000) GS:ffff888446800000(0000) knlGS:0000000000000000
> <4>[ 355.438569] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> <4>[ 355.438582] CR2: 00007fceb380f6b8 CR3: 0000000117c5c004 CR4: 00000000003706f0
> <4>[ 355.438586] Call Trace:
> <4>[ 355.438589] <TASK>
> <4>[ 355.438592] task_call_func+0xc4/0xe0
> <4>[ 355.438600] freeze_task+0x84/0xe0
> <4>[ 355.438607] try_to_freeze_tasks+0xac/0x260
> <4>[ 355.438616] freeze_processes+0x56/0xb0
> <4>[ 355.438622] pm_suspend.cold.7+0x1d9/0x31c
> <4>[ 355.438629] state_store+0x7b/0xe0
> <4>[ 355.438637] kernfs_fop_write_iter+0x124/0x1c0
> <4>[ 355.438644] vfs_write+0x34f/0x4e0
> <4>[ 355.438655] ksys_write+0x57/0xd0
> <4>[ 355.438663] do_syscall_64+0x3a/0x90
> <4>[ 355.438670] entry_SYSCALL_64_after_hwframe+0x63/0xcd

Oh I think I see what's going on.

It's a very narrow race between schedule() and task_call_func().

CPU0 CPU1

__schedule()
rq_lock();
prev_state = READ_ONCE(prev->__state);
if (... && prev_state) {
deactivate_tasl(rq, prev, ...)
prev->on_rq = 0;

task_call_func()
raw_spin_lock_irqsave(p->pi_lock);
state = READ_ONCE(p->__state);
smp_rmb();
if (... || p->on_rq) // false!!!
rq = __task_rq_lock()

ret = func();

next = pick_next_task();
rq = context_switch(prev, next)
prepare_lock_switch()
spin_release(&__rq_lockp(rq)->dep_map...)

So while the task is on it's way out, it still holds rq->lock for a
little while, and right then task_call_func() comes in and figures it
doesn't need rq->lock anymore (because the task is already dequeued --
but still running there) and then the __set_task_frozen() thing observes
it's holding rq->lock and yells murder.

Could you please give the below a spin?

---
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index cb2aa2b54c7a..f519f44cd4c7 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4200,6 +4200,37 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
return success;
}

+static bool __task_needs_rq_lock(struct task_struct *p)
+{
+ unsigned int state = READ_ONCE(p->__state);
+
+ /*
+ * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
+ * the task is blocked. Make sure to check @state since ttwu() can drop
+ * locks at the end, see ttwu_queue_wakelist().
+ */
+ if (state == TASK_RUNNING || state == TASK_WAKING)
+ return true;
+
+ /*
+ * Ensure we load p->on_rq after p->__state, otherwise it would be
+ * possible to, falsely, observe p->on_rq == 0.
+ *
+ * See try_to_wake_up() for a longer comment.
+ */
+ smp_rmb();
+ if (p->on_rq)
+ return true;
+
+#ifdef CONFIG_SMP
+ smp_rmb();
+ if (p->on_cpu)
+ return true;
+#endif
+
+ return false;
+}
+
/**
* task_call_func - Invoke a function on task in fixed state
* @p: Process for which the function is to be invoked, can be @current.
@@ -4217,28 +4248,12 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
int task_call_func(struct task_struct *p, task_call_f func, void *arg)
{
struct rq *rq = NULL;
- unsigned int state;
struct rq_flags rf;
int ret;

raw_spin_lock_irqsave(&p->pi_lock, rf.flags);

- state = READ_ONCE(p->__state);
-
- /*
- * Ensure we load p->on_rq after p->__state, otherwise it would be
- * possible to, falsely, observe p->on_rq == 0.
- *
- * See try_to_wake_up() for a longer comment.
- */
- smp_rmb();
-
- /*
- * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
- * the task is blocked. Make sure to check @state since ttwu() can drop
- * locks at the end, see ttwu_queue_wakelist().
- */
- if (state == TASK_RUNNING || state == TASK_WAKING || p->on_rq)
+ if (__task_needs_rq_lock(p))
rq = __task_rq_lock(p, &rf);

/*

2022-10-26 12:57:55

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

2022-10-26 13:17:29

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Wed, Oct 26, 2022 at 02:12:02PM +0200, Peter Zijlstra wrote:
> On Wed, Oct 26, 2022 at 01:43:00PM +0200, Peter Zijlstra wrote:
> > On Wed, Oct 26, 2022 at 01:32:31PM +0300, Ville Syrj?l? wrote:
> > > Short form looks to be this:
> > > <4>[ 355.437846] 1 lock held by rs:main Q:Reg/359:
> > > <4>[ 355.438418] #0: ffff88844693b758 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x1b/0x30
> > > <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
> >
> > > <4>[ 355.438429] ------------[ cut here ]------------
> > > <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
> > > <4>[ 355.438439] WARNING: CPU: 0 PID: 6211 at kernel/freezer.c:134 __set_task_frozen+0x86/0xb0
> > > <4>[ 355.438447] Modules linked in: snd_hda_intel i915 mei_hdcp mei_pxp drm_display_helper drm_kms_helper vgem drm_shmem_helper snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core snd_pcm prime_numbers ttm drm_buddy syscopyarea sysfillrect sysimgblt fb_sys_fops fuse x86_pkg_temp_thermal coretemp kvm_intel btusb btrtl btbcm btintel kvm irqbypass bluetooth crct10dif_pclmul crc32_pclmul ecdh_generic ghash_clmulni_intel ecc e1000e mei_me i2c_i801 ptp mei i2c_smbus pps_core lpc_ich video wmi [last unloaded: drm_kms_helper]
> > > <4>[ 355.438521] CPU: 0 PID: 6211 Comm: rtcwake Tainted: G U 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1
> > > <4>[ 355.438526] Hardware name: /NUC5i7RYB, BIOS RYBDWi35.86A.0385.2020.0519.1558 05/19/2020
> > > <4>[ 355.438530] RIP: 0010:__set_task_frozen+0x86/0xb0
> > > <4>[ 355.438536] Code: 83 60 09 00 00 85 c0 74 2a 48 89 df e8 ac 02 9b 00 8b 93 38 05 00 00 48 8d b3 48 07 00 00 48 c7 c7 a0 62 2b 82 e8 ee c1 9a 00 <0f> 0b c6 05 51 75 e3 02 01 c7 43 18 00 80 00 00 b8 00 80 00 00 5b
> > > <4>[ 355.438541] RSP: 0018:ffffc900012cbcf0 EFLAGS: 00010086
> > > <4>[ 355.438546] RAX: 0000000000000000 RBX: ffff88810d090040 RCX: 0000000000000004
> > > <4>[ 355.438550] RDX: 0000000000000004 RSI: 00000000fffff5de RDI: 00000000ffffffff
> > > <4>[ 355.438553] RBP: 0000000000000000 R08: 0000000000000000 R09: c0000000fffff5de
> > > <4>[ 355.438557] R10: 00000000002335f8 R11: ffffc900012cbb88 R12: 0000000000000246
> > > <4>[ 355.438561] R13: ffffffff81165430 R14: 0000000000000000 R15: ffff88810d090040
> > > <4>[ 355.438565] FS: 00007fcfa43c7740(0000) GS:ffff888446800000(0000) knlGS:0000000000000000
> > > <4>[ 355.438569] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > > <4>[ 355.438582] CR2: 00007fceb380f6b8 CR3: 0000000117c5c004 CR4: 00000000003706f0
> > > <4>[ 355.438586] Call Trace:
> > > <4>[ 355.438589] <TASK>
> > > <4>[ 355.438592] task_call_func+0xc4/0xe0
> > > <4>[ 355.438600] freeze_task+0x84/0xe0
> > > <4>[ 355.438607] try_to_freeze_tasks+0xac/0x260
> > > <4>[ 355.438616] freeze_processes+0x56/0xb0
> > > <4>[ 355.438622] pm_suspend.cold.7+0x1d9/0x31c
> > > <4>[ 355.438629] state_store+0x7b/0xe0
> > > <4>[ 355.438637] kernfs_fop_write_iter+0x124/0x1c0
> > > <4>[ 355.438644] vfs_write+0x34f/0x4e0
> > > <4>[ 355.438655] ksys_write+0x57/0xd0
> > > <4>[ 355.438663] do_syscall_64+0x3a/0x90
> > > <4>[ 355.438670] entry_SYSCALL_64_after_hwframe+0x63/0xcd
> >
> > Oh I think I see what's going on.
> >
> > It's a very narrow race between schedule() and task_call_func().
> >
> > CPU0 CPU1
> >
> > __schedule()
> > rq_lock();
> > prev_state = READ_ONCE(prev->__state);
> > if (... && prev_state) {
> > deactivate_tasl(rq, prev, ...)
> > prev->on_rq = 0;
> >
> > task_call_func()
> > raw_spin_lock_irqsave(p->pi_lock);
> > state = READ_ONCE(p->__state);
> > smp_rmb();
> > if (... || p->on_rq) // false!!!
> > rq = __task_rq_lock()
> >
> > ret = func();
> >
> > next = pick_next_task();
> > rq = context_switch(prev, next)
> > prepare_lock_switch()
> > spin_release(&__rq_lockp(rq)->dep_map...)
> >
> >
> >
> > So while the task is on it's way out, it still holds rq->lock for a
> > little while, and right then task_call_func() comes in and figures it
> > doesn't need rq->lock anymore (because the task is already dequeued --
> > but still running there) and then the __set_task_frozen() thing observes
> > it's holding rq->lock and yells murder.
> >
> > Could you please give the below a spin?
>
> Urgh.. that'll narrow the race more, but won't solve it, that
> prepare_lock_switch() is after we clear ->on_cpu.
>
> Let me ponder this a wee bit more..

Oh, n/m, I got myself confused, it's prepare_lock_switch() that releases
the lock (from lockdep's pov) and that *IS* before finish_task() which
clears ->on_cpu.

So all well, please test the patch.

2022-10-27 06:30:35

by Chen Yu

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On 2022-10-26 at 13:43:00 +0200, Peter Zijlstra wrote:
> On Wed, Oct 26, 2022 at 01:32:31PM +0300, Ville Syrj?l? wrote:
> > Short form looks to be this:
> > <4>[ 355.437846] 1 lock held by rs:main Q:Reg/359:
> > <4>[ 355.438418] #0: ffff88844693b758 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x1b/0x30
> > <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
>
> > <4>[ 355.438429] ------------[ cut here ]------------
> > <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
> > <4>[ 355.438439] WARNING: CPU: 0 PID: 6211 at kernel/freezer.c:134 __set_task_frozen+0x86/0xb0
> > <4>[ 355.438447] Modules linked in: snd_hda_intel i915 mei_hdcp mei_pxp drm_display_helper drm_kms_helper vgem drm_shmem_helper snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core snd_pcm prime_numbers ttm drm_buddy syscopyarea sysfillrect sysimgblt fb_sys_fops fuse x86_pkg_temp_thermal coretemp kvm_intel btusb btrtl btbcm btintel kvm irqbypass bluetooth crct10dif_pclmul crc32_pclmul ecdh_generic ghash_clmulni_intel ecc e1000e mei_me i2c_i801 ptp mei i2c_smbus pps_core lpc_ich video wmi [last unloaded: drm_kms_helper]
> > <4>[ 355.438521] CPU: 0 PID: 6211 Comm: rtcwake Tainted: G U 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1
> > <4>[ 355.438526] Hardware name: /NUC5i7RYB, BIOS RYBDWi35.86A.0385.2020.0519.1558 05/19/2020
> > <4>[ 355.438530] RIP: 0010:__set_task_frozen+0x86/0xb0
> > <4>[ 355.438536] Code: 83 60 09 00 00 85 c0 74 2a 48 89 df e8 ac 02 9b 00 8b 93 38 05 00 00 48 8d b3 48 07 00 00 48 c7 c7 a0 62 2b 82 e8 ee c1 9a 00 <0f> 0b c6 05 51 75 e3 02 01 c7 43 18 00 80 00 00 b8 00 80 00 00 5b
> > <4>[ 355.438541] RSP: 0018:ffffc900012cbcf0 EFLAGS: 00010086
> > <4>[ 355.438546] RAX: 0000000000000000 RBX: ffff88810d090040 RCX: 0000000000000004
> > <4>[ 355.438550] RDX: 0000000000000004 RSI: 00000000fffff5de RDI: 00000000ffffffff
> > <4>[ 355.438553] RBP: 0000000000000000 R08: 0000000000000000 R09: c0000000fffff5de
> > <4>[ 355.438557] R10: 00000000002335f8 R11: ffffc900012cbb88 R12: 0000000000000246
> > <4>[ 355.438561] R13: ffffffff81165430 R14: 0000000000000000 R15: ffff88810d090040
> > <4>[ 355.438565] FS: 00007fcfa43c7740(0000) GS:ffff888446800000(0000) knlGS:0000000000000000
> > <4>[ 355.438569] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > <4>[ 355.438582] CR2: 00007fceb380f6b8 CR3: 0000000117c5c004 CR4: 00000000003706f0
> > <4>[ 355.438586] Call Trace:
> > <4>[ 355.438589] <TASK>
> > <4>[ 355.438592] task_call_func+0xc4/0xe0
> > <4>[ 355.438600] freeze_task+0x84/0xe0
> > <4>[ 355.438607] try_to_freeze_tasks+0xac/0x260
> > <4>[ 355.438616] freeze_processes+0x56/0xb0
> > <4>[ 355.438622] pm_suspend.cold.7+0x1d9/0x31c
> > <4>[ 355.438629] state_store+0x7b/0xe0
> > <4>[ 355.438637] kernfs_fop_write_iter+0x124/0x1c0
> > <4>[ 355.438644] vfs_write+0x34f/0x4e0
> > <4>[ 355.438655] ksys_write+0x57/0xd0
> > <4>[ 355.438663] do_syscall_64+0x3a/0x90
> > <4>[ 355.438670] entry_SYSCALL_64_after_hwframe+0x63/0xcd
>
> Oh I think I see what's going on.
>
> It's a very narrow race between schedule() and task_call_func().
>
> CPU0 CPU1
>
> __schedule()
> rq_lock();
> prev_state = READ_ONCE(prev->__state);
> if (... && prev_state) {
> deactivate_tasl(rq, prev, ...)
> prev->on_rq = 0;
>
> task_call_func()
> raw_spin_lock_irqsave(p->pi_lock);
> state = READ_ONCE(p->__state);
> smp_rmb();
> if (... || p->on_rq) // false!!!
> rq = __task_rq_lock()
>
> ret = func();
>
> next = pick_next_task();
> rq = context_switch(prev, next)
> prepare_lock_switch()
> spin_release(&__rq_lockp(rq)->dep_map...)
>
>
>
> So while the task is on it's way out, it still holds rq->lock for a
> little while, and right then task_call_func() comes in and figures it
> doesn't need rq->lock anymore (because the task is already dequeued --
> but still running there) and then the __set_task_frozen() thing observes
> it's holding rq->lock and yells murder.
>
> Could you please give the below a spin?
>
> ---
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index cb2aa2b54c7a..f519f44cd4c7 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -4200,6 +4200,37 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> return success;
> }
>
> +static bool __task_needs_rq_lock(struct task_struct *p)
> +{
> + unsigned int state = READ_ONCE(p->__state);
> +
> + /*
> + * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
> + * the task is blocked. Make sure to check @state since ttwu() can drop
> + * locks at the end, see ttwu_queue_wakelist().
> + */
> + if (state == TASK_RUNNING || state == TASK_WAKING)
> + return true;
> +
> + /*
> + * Ensure we load p->on_rq after p->__state, otherwise it would be
> + * possible to, falsely, observe p->on_rq == 0.
> + *
> + * See try_to_wake_up() for a longer comment.
> + */
> + smp_rmb();
> + if (p->on_rq)
> + return true;
> +
> +#ifdef CONFIG_SMP
> + smp_rmb();
> + if (p->on_cpu)
> + return true;
> +#endif
Should we also add p->on_cpu check to return 0 in __set_task_frozen()?
Otherwise it might still warn that p is holding the lock?

thanks,
Chenyu

2022-10-27 07:49:10

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Thu, Oct 27, 2022 at 01:58:09PM +0800, Chen Yu wrote:

> > It's a very narrow race between schedule() and task_call_func().
> >
> > CPU0 CPU1
> >
> > __schedule()
> > rq_lock();
> > prev_state = READ_ONCE(prev->__state);
> > if (... && prev_state) {
> > deactivate_tasl(rq, prev, ...)
> > prev->on_rq = 0;
> >
> > task_call_func()
> > raw_spin_lock_irqsave(p->pi_lock);
> > state = READ_ONCE(p->__state);
> > smp_rmb();
> > if (... || p->on_rq) // false!!!
> > rq = __task_rq_lock()
> >
> > ret = func();
> >
> > next = pick_next_task();
> > rq = context_switch(prev, next)
> > prepare_lock_switch()
> > spin_release(&__rq_lockp(rq)->dep_map...)
> >
> >
> >
> > So while the task is on it's way out, it still holds rq->lock for a
> > little while, and right then task_call_func() comes in and figures it
> > doesn't need rq->lock anymore (because the task is already dequeued --
> > but still running there) and then the __set_task_frozen() thing observes
> > it's holding rq->lock and yells murder.
> >
> > Could you please give the below a spin?
> >
> > ---
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index cb2aa2b54c7a..f519f44cd4c7 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -4200,6 +4200,37 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> > return success;
> > }
> >
> > +static bool __task_needs_rq_lock(struct task_struct *p)
> > +{
> > + unsigned int state = READ_ONCE(p->__state);
> > +
> > + /*
> > + * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
> > + * the task is blocked. Make sure to check @state since ttwu() can drop
> > + * locks at the end, see ttwu_queue_wakelist().
> > + */
> > + if (state == TASK_RUNNING || state == TASK_WAKING)
> > + return true;
> > +
> > + /*
> > + * Ensure we load p->on_rq after p->__state, otherwise it would be
> > + * possible to, falsely, observe p->on_rq == 0.
> > + *
> > + * See try_to_wake_up() for a longer comment.
> > + */
> > + smp_rmb();
> > + if (p->on_rq)
> > + return true;
> > +
> > +#ifdef CONFIG_SMP
> > + smp_rmb();
> > + if (p->on_cpu)
> > + return true;
> > +#endif
> Should we also add p->on_cpu check to return 0 in __set_task_frozen()?
> Otherwise it might still warn that p is holding the lock?

With this, I don't think __set_task_frozen() should ever see
'p->on_cpu && !p->on_rq'. By forcing task_call_func() to acquire
rq->lock that window is closed. That is, this window only exits in
__schedule() while it holds rq->lock, since we're now serializing
against that, we should no longer observe it.

2022-10-27 14:01:50

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Wed, Oct 26, 2022 at 01:43:00PM +0200, Peter Zijlstra wrote:
> On Wed, Oct 26, 2022 at 01:32:31PM +0300, Ville Syrj?l? wrote:
> > Short form looks to be this:
> > <4>[ 355.437846] 1 lock held by rs:main Q:Reg/359:
> > <4>[ 355.438418] #0: ffff88844693b758 (&rq->__lock){-.-.}-{2:2}, at: raw_spin_rq_lock_nested+0x1b/0x30
> > <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
>
> > <4>[ 355.438429] ------------[ cut here ]------------
> > <4>[ 355.438432] rs:main Q:Reg/359 holding locks while freezing
> > <4>[ 355.438439] WARNING: CPU: 0 PID: 6211 at kernel/freezer.c:134 __set_task_frozen+0x86/0xb0
> > <4>[ 355.438447] Modules linked in: snd_hda_intel i915 mei_hdcp mei_pxp drm_display_helper drm_kms_helper vgem drm_shmem_helper snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_intel_dspcfg snd_hda_codec snd_hwdep snd_hda_core snd_pcm prime_numbers ttm drm_buddy syscopyarea sysfillrect sysimgblt fb_sys_fops fuse x86_pkg_temp_thermal coretemp kvm_intel btusb btrtl btbcm btintel kvm irqbypass bluetooth crct10dif_pclmul crc32_pclmul ecdh_generic ghash_clmulni_intel ecc e1000e mei_me i2c_i801 ptp mei i2c_smbus pps_core lpc_ich video wmi [last unloaded: drm_kms_helper]
> > <4>[ 355.438521] CPU: 0 PID: 6211 Comm: rtcwake Tainted: G U 6.1.0-rc2-CI_DRM_12295-g3844a56a0922+ #1
> > <4>[ 355.438526] Hardware name: /NUC5i7RYB, BIOS RYBDWi35.86A.0385.2020.0519.1558 05/19/2020
> > <4>[ 355.438530] RIP: 0010:__set_task_frozen+0x86/0xb0
> > <4>[ 355.438536] Code: 83 60 09 00 00 85 c0 74 2a 48 89 df e8 ac 02 9b 00 8b 93 38 05 00 00 48 8d b3 48 07 00 00 48 c7 c7 a0 62 2b 82 e8 ee c1 9a 00 <0f> 0b c6 05 51 75 e3 02 01 c7 43 18 00 80 00 00 b8 00 80 00 00 5b
> > <4>[ 355.438541] RSP: 0018:ffffc900012cbcf0 EFLAGS: 00010086
> > <4>[ 355.438546] RAX: 0000000000000000 RBX: ffff88810d090040 RCX: 0000000000000004
> > <4>[ 355.438550] RDX: 0000000000000004 RSI: 00000000fffff5de RDI: 00000000ffffffff
> > <4>[ 355.438553] RBP: 0000000000000000 R08: 0000000000000000 R09: c0000000fffff5de
> > <4>[ 355.438557] R10: 00000000002335f8 R11: ffffc900012cbb88 R12: 0000000000000246
> > <4>[ 355.438561] R13: ffffffff81165430 R14: 0000000000000000 R15: ffff88810d090040
> > <4>[ 355.438565] FS: 00007fcfa43c7740(0000) GS:ffff888446800000(0000) knlGS:0000000000000000
> > <4>[ 355.438569] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > <4>[ 355.438582] CR2: 00007fceb380f6b8 CR3: 0000000117c5c004 CR4: 00000000003706f0
> > <4>[ 355.438586] Call Trace:
> > <4>[ 355.438589] <TASK>
> > <4>[ 355.438592] task_call_func+0xc4/0xe0
> > <4>[ 355.438600] freeze_task+0x84/0xe0
> > <4>[ 355.438607] try_to_freeze_tasks+0xac/0x260
> > <4>[ 355.438616] freeze_processes+0x56/0xb0
> > <4>[ 355.438622] pm_suspend.cold.7+0x1d9/0x31c
> > <4>[ 355.438629] state_store+0x7b/0xe0
> > <4>[ 355.438637] kernfs_fop_write_iter+0x124/0x1c0
> > <4>[ 355.438644] vfs_write+0x34f/0x4e0
> > <4>[ 355.438655] ksys_write+0x57/0xd0
> > <4>[ 355.438663] do_syscall_64+0x3a/0x90
> > <4>[ 355.438670] entry_SYSCALL_64_after_hwframe+0x63/0xcd
>
> Oh I think I see what's going on.
>
> It's a very narrow race between schedule() and task_call_func().
>
> CPU0 CPU1
>
> __schedule()
> rq_lock();
> prev_state = READ_ONCE(prev->__state);
> if (... && prev_state) {
> deactivate_tasl(rq, prev, ...)
> prev->on_rq = 0;
>
> task_call_func()
> raw_spin_lock_irqsave(p->pi_lock);
> state = READ_ONCE(p->__state);
> smp_rmb();
> if (... || p->on_rq) // false!!!
> rq = __task_rq_lock()
>
> ret = func();
>
> next = pick_next_task();
> rq = context_switch(prev, next)
> prepare_lock_switch()
> spin_release(&__rq_lockp(rq)->dep_map...)
>
>
>
> So while the task is on it's way out, it still holds rq->lock for a
> little while, and right then task_call_func() comes in and figures it
> doesn't need rq->lock anymore (because the task is already dequeued --
> but still running there) and then the __set_task_frozen() thing observes
> it's holding rq->lock and yells murder.
>
> Could you please give the below a spin?

Thanks. I've added this to our CI branch. I'll try to keep and eye
on it in the coming days and let you know if anything still trips.
And I'll report back maybe ~middle of next week if we haven't caught
anything by then.

>
> ---
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index cb2aa2b54c7a..f519f44cd4c7 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -4200,6 +4200,37 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> return success;
> }
>
> +static bool __task_needs_rq_lock(struct task_struct *p)
> +{
> + unsigned int state = READ_ONCE(p->__state);
> +
> + /*
> + * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
> + * the task is blocked. Make sure to check @state since ttwu() can drop
> + * locks at the end, see ttwu_queue_wakelist().
> + */
> + if (state == TASK_RUNNING || state == TASK_WAKING)
> + return true;
> +
> + /*
> + * Ensure we load p->on_rq after p->__state, otherwise it would be
> + * possible to, falsely, observe p->on_rq == 0.
> + *
> + * See try_to_wake_up() for a longer comment.
> + */
> + smp_rmb();
> + if (p->on_rq)
> + return true;
> +
> +#ifdef CONFIG_SMP
> + smp_rmb();
> + if (p->on_cpu)
> + return true;
> +#endif
> +
> + return false;
> +}
> +
> /**
> * task_call_func - Invoke a function on task in fixed state
> * @p: Process for which the function is to be invoked, can be @current.
> @@ -4217,28 +4248,12 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> int task_call_func(struct task_struct *p, task_call_f func, void *arg)
> {
> struct rq *rq = NULL;
> - unsigned int state;
> struct rq_flags rf;
> int ret;
>
> raw_spin_lock_irqsave(&p->pi_lock, rf.flags);
>
> - state = READ_ONCE(p->__state);
> -
> - /*
> - * Ensure we load p->on_rq after p->__state, otherwise it would be
> - * possible to, falsely, observe p->on_rq == 0.
> - *
> - * See try_to_wake_up() for a longer comment.
> - */
> - smp_rmb();
> -
> - /*
> - * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
> - * the task is blocked. Make sure to check @state since ttwu() can drop
> - * locks at the end, see ttwu_queue_wakelist().
> - */
> - if (state == TASK_RUNNING || state == TASK_WAKING || p->on_rq)
> + if (__task_needs_rq_lock(p))
> rq = __task_rq_lock(p, &rf);
>
> /*

--
Ville Syrj?l?
Intel

2022-10-27 16:58:37

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Thu, Oct 27, 2022 at 04:09:01PM +0300, Ville Syrj?l? wrote:
> On Wed, Oct 26, 2022 at 01:43:00PM +0200, Peter Zijlstra wrote:

> > Could you please give the below a spin?
>
> Thanks. I've added this to our CI branch. I'll try to keep and eye
> on it in the coming days and let you know if anything still trips.
> And I'll report back maybe ~middle of next week if we haven't caught
> anything by then.

Thanks!

2022-11-02 17:16:12

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Thu, Oct 27, 2022 at 06:53:23PM +0200, Peter Zijlstra wrote:
> On Thu, Oct 27, 2022 at 04:09:01PM +0300, Ville Syrj?l? wrote:
> > On Wed, Oct 26, 2022 at 01:43:00PM +0200, Peter Zijlstra wrote:
>
> > > Could you please give the below a spin?
> >
> > Thanks. I've added this to our CI branch. I'll try to keep and eye
> > on it in the coming days and let you know if anything still trips.
> > And I'll report back maybe ~middle of next week if we haven't caught
> > anything by then.
>
> Thanks!

Looks like we haven't caught anything since I put the patch in.
So the fix seems good.

--
Ville Syrj?l?
Intel

2022-11-02 22:52:36

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Wed, Nov 02, 2022 at 06:57:51PM +0200, Ville Syrj?l? wrote:
> On Thu, Oct 27, 2022 at 06:53:23PM +0200, Peter Zijlstra wrote:
> > On Thu, Oct 27, 2022 at 04:09:01PM +0300, Ville Syrj?l? wrote:
> > > On Wed, Oct 26, 2022 at 01:43:00PM +0200, Peter Zijlstra wrote:
> >
> > > > Could you please give the below a spin?
> > >
> > > Thanks. I've added this to our CI branch. I'll try to keep and eye
> > > on it in the coming days and let you know if anything still trips.
> > > And I'll report back maybe ~middle of next week if we haven't caught
> > > anything by then.
> >
> > Thanks!
>
> Looks like we haven't caught anything since I put the patch in.
> So the fix seems good.

While writing up the Changelog, it occured to me it might be possible to
fix another way, could I bother you to also run the below patch for a
bit?

---
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index cb2aa2b54c7a..daff72f00385 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4200,6 +4200,40 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
return success;
}

+static bool __task_needs_rq_lock(struct task_struct *p)
+{
+ unsigned int state = READ_ONCE(p->__state);
+
+ /*
+ * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
+ * the task is blocked. Make sure to check @state since ttwu() can drop
+ * locks at the end, see ttwu_queue_wakelist().
+ */
+ if (state == TASK_RUNNING || state == TASK_WAKING)
+ return true;
+
+ /*
+ * Ensure we load p->on_rq after p->__state, otherwise it would be
+ * possible to, falsely, observe p->on_rq == 0.
+ *
+ * See try_to_wake_up() for a longer comment.
+ */
+ smp_rmb();
+ if (p->on_rq)
+ return true;
+
+#ifdef CONFIG_SMP
+ /*
+ * Ensure the task has finished __schedule() and will not be referenced
+ * anymore. Again, see try_to_wake_up() for a longer comment.
+ */
+ smp_rmb();
+ smp_cond_load_acquire(&p->on_cpu, !VAL);
+#endif
+
+ return false;
+}
+
/**
* task_call_func - Invoke a function on task in fixed state
* @p: Process for which the function is to be invoked, can be @current.
@@ -4217,28 +4251,12 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
int task_call_func(struct task_struct *p, task_call_f func, void *arg)
{
struct rq *rq = NULL;
- unsigned int state;
struct rq_flags rf;
int ret;

raw_spin_lock_irqsave(&p->pi_lock, rf.flags);

- state = READ_ONCE(p->__state);
-
- /*
- * Ensure we load p->on_rq after p->__state, otherwise it would be
- * possible to, falsely, observe p->on_rq == 0.
- *
- * See try_to_wake_up() for a longer comment.
- */
- smp_rmb();
-
- /*
- * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
- * the task is blocked. Make sure to check @state since ttwu() can drop
- * locks at the end, see ttwu_queue_wakelist().
- */
- if (state == TASK_RUNNING || state == TASK_WAKING || p->on_rq)
+ if (__task_needs_rq_lock(p))
rq = __task_rq_lock(p, &rf);

/*

2022-11-07 12:06:57

[permalink] [raw]

Subject: Re: [PATCH v3 6/6] freezer,sched: Rewrite core freezer logic

On Wed, Nov 02, 2022 at 11:16:48PM +0100, Peter Zijlstra wrote:
> On Wed, Nov 02, 2022 at 06:57:51PM +0200, Ville Syrj?l? wrote:
> > On Thu, Oct 27, 2022 at 06:53:23PM +0200, Peter Zijlstra wrote:
> > > On Thu, Oct 27, 2022 at 04:09:01PM +0300, Ville Syrj?l? wrote:
> > > > On Wed, Oct 26, 2022 at 01:43:00PM +0200, Peter Zijlstra wrote:
> > >
> > > > > Could you please give the below a spin?
> > > >
> > > > Thanks. I've added this to our CI branch. I'll try to keep and eye
> > > > on it in the coming days and let you know if anything still trips.
> > > > And I'll report back maybe ~middle of next week if we haven't caught
> > > > anything by then.
> > >
> > > Thanks!
> >
> > Looks like we haven't caught anything since I put the patch in.
> > So the fix seems good.
>
> While writing up the Changelog, it occured to me it might be possible to
> fix another way, could I bother you to also run the below patch for a
> bit?

I swapped in the new patch to the CI branch. I'll check back
after a few days.

>
> ---
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index cb2aa2b54c7a..daff72f00385 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -4200,6 +4200,40 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> return success;
> }
>
> +static bool __task_needs_rq_lock(struct task_struct *p)
> +{
> + unsigned int state = READ_ONCE(p->__state);
> +
> + /*
> + * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
> + * the task is blocked. Make sure to check @state since ttwu() can drop
> + * locks at the end, see ttwu_queue_wakelist().
> + */
> + if (state == TASK_RUNNING || state == TASK_WAKING)
> + return true;
> +
> + /*
> + * Ensure we load p->on_rq after p->__state, otherwise it would be
> + * possible to, falsely, observe p->on_rq == 0.
> + *
> + * See try_to_wake_up() for a longer comment.
> + */
> + smp_rmb();
> + if (p->on_rq)
> + return true;
> +
> +#ifdef CONFIG_SMP
> + /*
> + * Ensure the task has finished __schedule() and will not be referenced
> + * anymore. Again, see try_to_wake_up() for a longer comment.
> + */
> + smp_rmb();
> + smp_cond_load_acquire(&p->on_cpu, !VAL);
> +#endif
> +
> + return false;
> +}
> +
> /**
> * task_call_func - Invoke a function on task in fixed state
> * @p: Process for which the function is to be invoked, can be @current.
> @@ -4217,28 +4251,12 @@ try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags)
> int task_call_func(struct task_struct *p, task_call_f func, void *arg)
> {
> struct rq *rq = NULL;
> - unsigned int state;
> struct rq_flags rf;
> int ret;
>
> raw_spin_lock_irqsave(&p->pi_lock, rf.flags);
>
> - state = READ_ONCE(p->__state);
> -
> - /*
> - * Ensure we load p->on_rq after p->__state, otherwise it would be
> - * possible to, falsely, observe p->on_rq == 0.
> - *
> - * See try_to_wake_up() for a longer comment.
> - */
> - smp_rmb();
> -
> - /*
> - * Since pi->lock blocks try_to_wake_up(), we don't need rq->lock when
> - * the task is blocked. Make sure to check @state since ttwu() can drop
> - * locks at the end, see ttwu_queue_wakelist().
> - */
> - if (state == TASK_RUNNING || state == TASK_WAKING || p->on_rq)
> + if (__task_needs_rq_lock(p))
> rq = __task_rq_lock(p, &rf);
>
> /*

--
Ville Syrj?l?
Intel

2022-11-10 20:39:04