| Message ID | 1304335365-13892-1-git-send-email-lczerner@redhat.com |
|---|---|
| State | Superseded, archived |
| Headers | show |
On 5/2/11 6:22 AM, Lukas Czerner wrote: > In order to make lazyinit eat approx. 10% of io bandwidth at max, we are > sleeping between zeroing each single inode table. For that purpose we > are using timer which wakes up thread when it expires. It is set via > add_timer() and this may cause troubles in the case that thread has been > woken up earlier and in next iteration we call add_timer() on still > running timer hence hitting BUG_ON in add_timer(). We could fix that by > using mod_timer() instead however we can use > schedule_timeout_interruptible() for waiting and hence simplifying > things a lot. > > This commit exchange the old "waiting mechanism" with simple > schedule_timeout_interruptible(), setting the time to sleep. Hence we do > not longer need li_wait_daemon waiting queue and others, so get rid of > it. > > This solves Red Hat bug 699708. > > Signed-off-by: Lukas Czerner <lczerner@redhat.com> > --- > fs/ext4/ext4.h | 4 ---- > fs/ext4/super.c | 26 +++----------------------- > 2 files changed, 3 insertions(+), 27 deletions(-) > > diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h > index 4daaf2b..1e37c09 100644 > --- a/fs/ext4/ext4.h > +++ b/fs/ext4/ext4.h > @@ -1590,12 +1590,8 @@ void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr, > */ > struct ext4_lazy_init { > unsigned long li_state; > - > - wait_queue_head_t li_wait_daemon; > wait_queue_head_t li_wait_task; > - struct timer_list li_timer; > struct task_struct *li_task; > - > struct list_head li_request_list; > struct mutex li_list_mtx; > }; > diff --git a/fs/ext4/super.c b/fs/ext4/super.c > index 8553dfb..daf3300 100644 > --- a/fs/ext4/super.c > +++ b/fs/ext4/super.c > @@ -2659,12 +2659,6 @@ static void print_daily_error_info(unsigned long arg) > mod_timer(&sbi->s_err_report, jiffies + 24*60*60*HZ); /* Once a day */ > } > > -static void ext4_lazyinode_timeout(unsigned long data) > -{ > - struct task_struct *p = (struct task_struct *)data; > - wake_up_process(p); > -} > - > /* Find next suitable group and run ext4_init_inode_table */ > static int ext4_run_li_request(struct ext4_li_request *elr) > { > @@ -2712,7 +2706,7 @@ static int ext4_run_li_request(struct ext4_li_request *elr) > > /* > * Remove lr_request from the list_request and free the > - * request tructure. Should be called with li_list_mtx held > + * request structure. Should be called with li_list_mtx held > */ > static void ext4_remove_li_request(struct ext4_li_request *elr) > { > @@ -2757,13 +2751,9 @@ static int ext4_lazyinit_thread(void *arg) > struct list_head *pos, *n; > struct ext4_li_request *elr; > unsigned long next_wakeup; > - DEFINE_WAIT(wait); > > BUG_ON(NULL == eli); > > - eli->li_timer.data = (unsigned long)current; > - eli->li_timer.function = ext4_lazyinode_timeout; > - > eli->li_task = current; > wake_up(&eli->li_wait_task); > > @@ -2803,13 +2793,8 @@ cont_thread: > continue; > } > > - eli->li_timer.expires = next_wakeup; > - add_timer(&eli->li_timer); > - prepare_to_wait(&eli->li_wait_daemon, &wait, > - TASK_INTERRUPTIBLE); > - if (time_before(jiffies, next_wakeup)) > - schedule(); > - finish_wait(&eli->li_wait_daemon, &wait); > + schedule_timeout_interruptible(next_wakeup - jiffies); Jeff asked whether this can wrap. I think it can? Do we need to do the trick of delta = (long)next_wakeup - (long)jiffies to make sure we don't schedule out forever if jiffies has wrapped and next_wakeup has not? Or maybe I'm behind on my jiffy-math.... :( -Eric > + > if (kthread_should_stop()) { > ext4_clear_request_list(); > goto exit_thread; > @@ -2833,12 +2818,10 @@ exit_thread: > goto cont_thread; > } > mutex_unlock(&eli->li_list_mtx); > - del_timer_sync(&ext4_li_info->li_timer); > eli->li_task = NULL; > wake_up(&eli->li_wait_task); > > kfree(ext4_li_info); > - ext4_lazyinit_task = NULL; > ext4_li_info = NULL; > mutex_unlock(&ext4_li_mtx); > > @@ -2866,7 +2849,6 @@ static int ext4_run_lazyinit_thread(void) > if (IS_ERR(ext4_lazyinit_task)) { > int err = PTR_ERR(ext4_lazyinit_task); > ext4_clear_request_list(); > - del_timer_sync(&ext4_li_info->li_timer); > kfree(ext4_li_info); > ext4_li_info = NULL; > printk(KERN_CRIT "EXT4: error %d creating inode table " > @@ -2915,9 +2897,7 @@ static int ext4_li_info_new(void) > INIT_LIST_HEAD(&eli->li_request_list); > mutex_init(&eli->li_list_mtx); > > - init_waitqueue_head(&eli->li_wait_daemon); > init_waitqueue_head(&eli->li_wait_task); > - init_timer(&eli->li_timer); > eli->li_state |= EXT4_LAZYINIT_QUIT; > > ext4_li_info = eli; -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, 4 May 2011, Eric Sandeen wrote: > On 5/2/11 6:22 AM, Lukas Czerner wrote: > > In order to make lazyinit eat approx. 10% of io bandwidth at max, we are > > sleeping between zeroing each single inode table. For that purpose we > > are using timer which wakes up thread when it expires. It is set via > > add_timer() and this may cause troubles in the case that thread has been > > woken up earlier and in next iteration we call add_timer() on still > > running timer hence hitting BUG_ON in add_timer(). We could fix that by > > using mod_timer() instead however we can use > > schedule_timeout_interruptible() for waiting and hence simplifying > > things a lot. > > > > This commit exchange the old "waiting mechanism" with simple > > schedule_timeout_interruptible(), setting the time to sleep. Hence we do > > not longer need li_wait_daemon waiting queue and others, so get rid of > > it. > > > > This solves Red Hat bug 699708. > > > > Signed-off-by: Lukas Czerner <lczerner@redhat.com> > > --- > > fs/ext4/ext4.h | 4 ---- > > fs/ext4/super.c | 26 +++----------------------- > > 2 files changed, 3 insertions(+), 27 deletions(-) > > > > diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h > > index 4daaf2b..1e37c09 100644 > > --- a/fs/ext4/ext4.h > > +++ b/fs/ext4/ext4.h > > @@ -1590,12 +1590,8 @@ void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr, > > */ > > struct ext4_lazy_init { > > unsigned long li_state; > > - > > - wait_queue_head_t li_wait_daemon; > > wait_queue_head_t li_wait_task; > > - struct timer_list li_timer; > > struct task_struct *li_task; > > - > > struct list_head li_request_list; > > struct mutex li_list_mtx; > > }; > > diff --git a/fs/ext4/super.c b/fs/ext4/super.c > > index 8553dfb..daf3300 100644 > > --- a/fs/ext4/super.c > > +++ b/fs/ext4/super.c > > @@ -2659,12 +2659,6 @@ static void print_daily_error_info(unsigned long arg) > > mod_timer(&sbi->s_err_report, jiffies + 24*60*60*HZ); /* Once a day */ > > } > > > > -static void ext4_lazyinode_timeout(unsigned long data) > > -{ > > - struct task_struct *p = (struct task_struct *)data; > > - wake_up_process(p); > > -} > > - > > /* Find next suitable group and run ext4_init_inode_table */ > > static int ext4_run_li_request(struct ext4_li_request *elr) > > { > > @@ -2712,7 +2706,7 @@ static int ext4_run_li_request(struct ext4_li_request *elr) > > > > /* > > * Remove lr_request from the list_request and free the > > - * request tructure. Should be called with li_list_mtx held > > + * request structure. Should be called with li_list_mtx held > > */ > > static void ext4_remove_li_request(struct ext4_li_request *elr) > > { > > @@ -2757,13 +2751,9 @@ static int ext4_lazyinit_thread(void *arg) > > struct list_head *pos, *n; > > struct ext4_li_request *elr; > > unsigned long next_wakeup; > > - DEFINE_WAIT(wait); > > > > BUG_ON(NULL == eli); > > > > - eli->li_timer.data = (unsigned long)current; > > - eli->li_timer.function = ext4_lazyinode_timeout; > > - > > eli->li_task = current; > > wake_up(&eli->li_wait_task); > > > > @@ -2803,13 +2793,8 @@ cont_thread: > > continue; > > } > > > > - eli->li_timer.expires = next_wakeup; > > - add_timer(&eli->li_timer); > > - prepare_to_wait(&eli->li_wait_daemon, &wait, > > - TASK_INTERRUPTIBLE); > > - if (time_before(jiffies, next_wakeup)) > > - schedule(); > > - finish_wait(&eli->li_wait_daemon, &wait); > > + schedule_timeout_interruptible(next_wakeup - jiffies); > > Jeff asked whether this can wrap. I think it can? > > Do we need to do the trick of delta = (long)next_wakeup - (long)jiffies to make sure we don't schedule out forever if jiffies has wrapped and next_wakeup has not? Or maybe I'm behind on my jiffy-math.... :( > > -Eric Ok, you're right. Since jiffies variable is volatile it can change any time, hence it can wrap just after the if (time_before(jiffies, next_wakeup)) condition which is pretty tight window, but it is there. So what about this ? ... unsigned long next_wakeup, current; ... current = jiffies; if ((time_after_eq(current, next_wakeup)) || (MAX_JIFFY_OFFSET == next_wakeup)) { cond_resched(); continue; } schedule_timeout_interruptible(next_wakeup - current); ... I think that this is the best solution which does not require additional condition, which the "(long)" trick would require anyway (since you can end up with negative value in the case that jiffies wrapped). Thanks! -Lukas > > > + > > if (kthread_should_stop()) { > > ext4_clear_request_list(); > > goto exit_thread; > > @@ -2833,12 +2818,10 @@ exit_thread: > > goto cont_thread; > > } > > mutex_unlock(&eli->li_list_mtx); > > - del_timer_sync(&ext4_li_info->li_timer); > > eli->li_task = NULL; > > wake_up(&eli->li_wait_task); > > > > kfree(ext4_li_info); > > - ext4_lazyinit_task = NULL; > > ext4_li_info = NULL; > > mutex_unlock(&ext4_li_mtx); > > > > @@ -2866,7 +2849,6 @@ static int ext4_run_lazyinit_thread(void) > > if (IS_ERR(ext4_lazyinit_task)) { > > int err = PTR_ERR(ext4_lazyinit_task); > > ext4_clear_request_list(); > > - del_timer_sync(&ext4_li_info->li_timer); > > kfree(ext4_li_info); > > ext4_li_info = NULL; > > printk(KERN_CRIT "EXT4: error %d creating inode table " > > @@ -2915,9 +2897,7 @@ static int ext4_li_info_new(void) > > INIT_LIST_HEAD(&eli->li_request_list); > > mutex_init(&eli->li_list_mtx); > > > > - init_waitqueue_head(&eli->li_wait_daemon); > > init_waitqueue_head(&eli->li_wait_task); > > - init_timer(&eli->li_timer); > > eli->li_state |= EXT4_LAZYINIT_QUIT; > > > > ext4_li_info = eli; > >
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 4daaf2b..1e37c09 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1590,12 +1590,8 @@ void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr, */ struct ext4_lazy_init { unsigned long li_state; - - wait_queue_head_t li_wait_daemon; wait_queue_head_t li_wait_task; - struct timer_list li_timer; struct task_struct *li_task; - struct list_head li_request_list; struct mutex li_list_mtx; }; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 8553dfb..daf3300 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2659,12 +2659,6 @@ static void print_daily_error_info(unsigned long arg) mod_timer(&sbi->s_err_report, jiffies + 24*60*60*HZ); /* Once a day */ } -static void ext4_lazyinode_timeout(unsigned long data) -{ - struct task_struct *p = (struct task_struct *)data; - wake_up_process(p); -} - /* Find next suitable group and run ext4_init_inode_table */ static int ext4_run_li_request(struct ext4_li_request *elr) { @@ -2712,7 +2706,7 @@ static int ext4_run_li_request(struct ext4_li_request *elr) /* * Remove lr_request from the list_request and free the - * request tructure. Should be called with li_list_mtx held + * request structure. Should be called with li_list_mtx held */ static void ext4_remove_li_request(struct ext4_li_request *elr) { @@ -2757,13 +2751,9 @@ static int ext4_lazyinit_thread(void *arg) struct list_head *pos, *n; struct ext4_li_request *elr; unsigned long next_wakeup; - DEFINE_WAIT(wait); BUG_ON(NULL == eli); - eli->li_timer.data = (unsigned long)current; - eli->li_timer.function = ext4_lazyinode_timeout; - eli->li_task = current; wake_up(&eli->li_wait_task); @@ -2803,13 +2793,8 @@ cont_thread: continue; } - eli->li_timer.expires = next_wakeup; - add_timer(&eli->li_timer); - prepare_to_wait(&eli->li_wait_daemon, &wait, - TASK_INTERRUPTIBLE); - if (time_before(jiffies, next_wakeup)) - schedule(); - finish_wait(&eli->li_wait_daemon, &wait); + schedule_timeout_interruptible(next_wakeup - jiffies); + if (kthread_should_stop()) { ext4_clear_request_list(); goto exit_thread; @@ -2833,12 +2818,10 @@ exit_thread: goto cont_thread; } mutex_unlock(&eli->li_list_mtx); - del_timer_sync(&ext4_li_info->li_timer); eli->li_task = NULL; wake_up(&eli->li_wait_task); kfree(ext4_li_info); - ext4_lazyinit_task = NULL; ext4_li_info = NULL; mutex_unlock(&ext4_li_mtx); @@ -2866,7 +2849,6 @@ static int ext4_run_lazyinit_thread(void) if (IS_ERR(ext4_lazyinit_task)) { int err = PTR_ERR(ext4_lazyinit_task); ext4_clear_request_list(); - del_timer_sync(&ext4_li_info->li_timer); kfree(ext4_li_info); ext4_li_info = NULL; printk(KERN_CRIT "EXT4: error %d creating inode table " @@ -2915,9 +2897,7 @@ static int ext4_li_info_new(void) INIT_LIST_HEAD(&eli->li_request_list); mutex_init(&eli->li_list_mtx); - init_waitqueue_head(&eli->li_wait_daemon); init_waitqueue_head(&eli->li_wait_task); - init_timer(&eli->li_timer); eli->li_state |= EXT4_LAZYINIT_QUIT; ext4_li_info = eli;
In order to make lazyinit eat approx. 10% of io bandwidth at max, we are sleeping between zeroing each single inode table. For that purpose we are using timer which wakes up thread when it expires. It is set via add_timer() and this may cause troubles in the case that thread has been woken up earlier and in next iteration we call add_timer() on still running timer hence hitting BUG_ON in add_timer(). We could fix that by using mod_timer() instead however we can use schedule_timeout_interruptible() for waiting and hence simplifying things a lot. This commit exchange the old "waiting mechanism" with simple schedule_timeout_interruptible(), setting the time to sleep. Hence we do not longer need li_wait_daemon waiting queue and others, so get rid of it. This solves Red Hat bug 699708. Signed-off-by: Lukas Czerner <lczerner@redhat.com> --- fs/ext4/ext4.h | 4 ---- fs/ext4/super.c | 26 +++----------------------- 2 files changed, 3 insertions(+), 27 deletions(-)