Re: [PATCH v2 06/15] migration: Yield bitmap_mutex properly when sending/sleeping
Peter Xu wrote: > On Wed, Oct 12, 2022 at 01:51:07PM -0400, Peter Xu wrote: >> On Wed, Oct 12, 2022 at 05:43:53PM +0100, Dr. David Alan Gilbert wrote: >> > * Peter Xu (pet...@redhat.com) wrote: >> > > Don't take the bitmap mutex when sending pages, or when being throttled >> > > by >> > > migration_rate_limit() (which is a bit tricky to call it here in ram >> > > code, >> > > but seems still helpful). >> > > >> > > It prepares for the possibility of concurrently sending pages in >1 >> > > threads >> > > using the function ram_save_host_page() because all threads may need the >> > > bitmap_mutex to operate on bitmaps, so that either sendmsg() or any kind >> > > of >> > > qemu_sem_wait() blocking for one thread will not block the other from >> > > progressing. >> > > >> > > Signed-off-by: Peter Xu >> > >> > Reviewed-by: Dr. David Alan Gilbert >> > >> > although a comment above the reclaration of ram_save_host_pages saying >> > it can drop the lock would be veyr good. >> >> Let me add that. Thanks, > > A fixup to this patch attached to touch up the comment for > ram_save_host_page(). Reviewed-by: Juan Quintela
Re: [PATCH v2 06/15] migration: Yield bitmap_mutex properly when sending/sleeping
* Peter Xu (pet...@redhat.com) wrote: > On Wed, Oct 12, 2022 at 01:51:07PM -0400, Peter Xu wrote: > > On Wed, Oct 12, 2022 at 05:43:53PM +0100, Dr. David Alan Gilbert wrote: > > > * Peter Xu (pet...@redhat.com) wrote: > > > > Don't take the bitmap mutex when sending pages, or when being throttled > > > > by > > > > migration_rate_limit() (which is a bit tricky to call it here in ram > > > > code, > > > > but seems still helpful). > > > > > > > > It prepares for the possibility of concurrently sending pages in >1 > > > > threads > > > > using the function ram_save_host_page() because all threads may need the > > > > bitmap_mutex to operate on bitmaps, so that either sendmsg() or any > > > > kind of > > > > qemu_sem_wait() blocking for one thread will not block the other from > > > > progressing. > > > > > > > > Signed-off-by: Peter Xu > > > > > > Reviewed-by: Dr. David Alan Gilbert > > > > > > although a comment above the reclaration of ram_save_host_pages saying > > > it can drop the lock would be veyr good. > > > > Let me add that. Thanks, > > A fixup to this patch attached to touch up the comment for > ram_save_host_page(). Yep, that's right (I don't think we have any formal annotation for locks) Reviewed-by: Dr. David Alan Gilbert > -- > Peter Xu > From dcc3adce062df7216851890d49f7d2b1fa2e84a4 Mon Sep 17 00:00:00 2001 > From: Peter Xu > Date: Thu, 13 Oct 2022 12:18:04 -0400 > Subject: [PATCH] fixup! migration: Yield bitmap_mutex properly when > sending/sleeping > Content-type: text/plain > > Signed-off-by: Peter Xu > --- > migration/ram.c | 5 + > 1 file changed, 5 insertions(+) > > diff --git a/migration/ram.c b/migration/ram.c > index 538667b974..b311ece48c 100644 > --- a/migration/ram.c > +++ b/migration/ram.c > @@ -2407,9 +2407,14 @@ out: > * a host page in which case the remainder of the hostpage is sent. > * Only dirty target pages are sent. Note that the host page size may > * be a huge page for this block. > + * > * The saving stops at the boundary of the used_length of the block > * if the RAMBlock isn't a multiple of the host page size. > * > + * The caller must be with ram_state.bitmap_mutex held to call this > + * function. Note that this function can temporarily release the lock, but > + * when the function is returned it'll make sure the lock is still held. > + * > * Returns the number of pages written or negative on error > * > * @rs: current RAM state > -- > 2.37.3 > -- Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
Re: [PATCH v2 06/15] migration: Yield bitmap_mutex properly when sending/sleeping
On Wed, Oct 12, 2022 at 01:51:07PM -0400, Peter Xu wrote: > On Wed, Oct 12, 2022 at 05:43:53PM +0100, Dr. David Alan Gilbert wrote: > > * Peter Xu (pet...@redhat.com) wrote: > > > Don't take the bitmap mutex when sending pages, or when being throttled by > > > migration_rate_limit() (which is a bit tricky to call it here in ram code, > > > but seems still helpful). > > > > > > It prepares for the possibility of concurrently sending pages in >1 > > > threads > > > using the function ram_save_host_page() because all threads may need the > > > bitmap_mutex to operate on bitmaps, so that either sendmsg() or any kind > > > of > > > qemu_sem_wait() blocking for one thread will not block the other from > > > progressing. > > > > > > Signed-off-by: Peter Xu > > > > Reviewed-by: Dr. David Alan Gilbert > > > > although a comment above the reclaration of ram_save_host_pages saying > > it can drop the lock would be veyr good. > > Let me add that. Thanks, A fixup to this patch attached to touch up the comment for ram_save_host_page(). -- Peter Xu >From dcc3adce062df7216851890d49f7d2b1fa2e84a4 Mon Sep 17 00:00:00 2001 From: Peter Xu Date: Thu, 13 Oct 2022 12:18:04 -0400 Subject: [PATCH] fixup! migration: Yield bitmap_mutex properly when sending/sleeping Content-type: text/plain Signed-off-by: Peter Xu --- migration/ram.c | 5 + 1 file changed, 5 insertions(+) diff --git a/migration/ram.c b/migration/ram.c index 538667b974..b311ece48c 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -2407,9 +2407,14 @@ out: * a host page in which case the remainder of the hostpage is sent. * Only dirty target pages are sent. Note that the host page size may * be a huge page for this block. + * * The saving stops at the boundary of the used_length of the block * if the RAMBlock isn't a multiple of the host page size. * + * The caller must be with ram_state.bitmap_mutex held to call this + * function. Note that this function can temporarily release the lock, but + * when the function is returned it'll make sure the lock is still held. + * * Returns the number of pages written or negative on error * * @rs: current RAM state -- 2.37.3
Re: [PATCH v2 06/15] migration: Yield bitmap_mutex properly when sending/sleeping
On Wed, Oct 12, 2022 at 05:43:53PM +0100, Dr. David Alan Gilbert wrote: > * Peter Xu (pet...@redhat.com) wrote: > > Don't take the bitmap mutex when sending pages, or when being throttled by > > migration_rate_limit() (which is a bit tricky to call it here in ram code, > > but seems still helpful). > > > > It prepares for the possibility of concurrently sending pages in >1 threads > > using the function ram_save_host_page() because all threads may need the > > bitmap_mutex to operate on bitmaps, so that either sendmsg() or any kind of > > qemu_sem_wait() blocking for one thread will not block the other from > > progressing. > > > > Signed-off-by: Peter Xu > > Reviewed-by: Dr. David Alan Gilbert > > although a comment above the reclaration of ram_save_host_pages saying > it can drop the lock would be veyr good. Let me add that. Thanks, -- Peter Xu
Re: [PATCH v2 06/15] migration: Yield bitmap_mutex properly when sending/sleeping
* Peter Xu (pet...@redhat.com) wrote: > Don't take the bitmap mutex when sending pages, or when being throttled by > migration_rate_limit() (which is a bit tricky to call it here in ram code, > but seems still helpful). > > It prepares for the possibility of concurrently sending pages in >1 threads > using the function ram_save_host_page() because all threads may need the > bitmap_mutex to operate on bitmaps, so that either sendmsg() or any kind of > qemu_sem_wait() blocking for one thread will not block the other from > progressing. > > Signed-off-by: Peter Xu Reviewed-by: Dr. David Alan Gilbert although a comment above the reclaration of ram_save_host_pages saying it can drop the lock would be veyr good. Dave > --- > migration/ram.c | 41 ++--- > 1 file changed, 30 insertions(+), 11 deletions(-) > > diff --git a/migration/ram.c b/migration/ram.c > index b9ac2d6921..578ad8d70a 100644 > --- a/migration/ram.c > +++ b/migration/ram.c > @@ -2462,6 +2462,7 @@ static void postcopy_preempt_reset_channel(RAMState *rs) > */ > static int ram_save_host_page(RAMState *rs, PageSearchStatus *pss) > { > +bool page_dirty, preempt_active = postcopy_preempt_active(); > int tmppages, pages = 0; > size_t pagesize_bits = > qemu_ram_pagesize(pss->block) >> TARGET_PAGE_BITS; > @@ -2485,22 +2486,40 @@ static int ram_save_host_page(RAMState *rs, > PageSearchStatus *pss) > break; > } > > -/* Check the pages is dirty and if it is send it */ > -if (migration_bitmap_clear_dirty(rs, pss->block, pss->page)) { > -tmppages = ram_save_target_page(rs, pss); > -if (tmppages < 0) { > -return tmppages; > -} > +page_dirty = migration_bitmap_clear_dirty(rs, pss->block, pss->page); > > -pages += tmppages; > +/* Check the pages is dirty and if it is send it */ > +if (page_dirty) { > /* > - * Allow rate limiting to happen in the middle of huge pages if > - * something is sent in the current iteration. > + * Properly yield the lock only in postcopy preempt mode > + * because both migration thread and rp-return thread can > + * operate on the bitmaps. > */ > -if (pagesize_bits > 1 && tmppages > 0) { > -migration_rate_limit(); > +if (preempt_active) { > +qemu_mutex_unlock(>bitmap_mutex); > +} > +tmppages = ram_save_target_page(rs, pss); > +if (tmppages >= 0) { > +pages += tmppages; > +/* > + * Allow rate limiting to happen in the middle of huge pages > if > + * something is sent in the current iteration. > + */ > +if (pagesize_bits > 1 && tmppages > 0) { > +migration_rate_limit(); > +} > } > +if (preempt_active) { > +qemu_mutex_lock(>bitmap_mutex); > +} > +} else { > +tmppages = 0; > } > + > +if (tmppages < 0) { > +return tmppages; > +} > + > pss->page = migration_bitmap_find_dirty(rs, pss->block, pss->page); > } while ((pss->page < hostpage_boundary) && > offset_in_ramblock(pss->block, > -- > 2.37.3 > -- Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
[PATCH v2 06/15] migration: Yield bitmap_mutex properly when sending/sleeping
Don't take the bitmap mutex when sending pages, or when being throttled by migration_rate_limit() (which is a bit tricky to call it here in ram code, but seems still helpful). It prepares for the possibility of concurrently sending pages in >1 threads using the function ram_save_host_page() because all threads may need the bitmap_mutex to operate on bitmaps, so that either sendmsg() or any kind of qemu_sem_wait() blocking for one thread will not block the other from progressing. Signed-off-by: Peter Xu --- migration/ram.c | 41 ++--- 1 file changed, 30 insertions(+), 11 deletions(-) diff --git a/migration/ram.c b/migration/ram.c index b9ac2d6921..578ad8d70a 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -2462,6 +2462,7 @@ static void postcopy_preempt_reset_channel(RAMState *rs) */ static int ram_save_host_page(RAMState *rs, PageSearchStatus *pss) { +bool page_dirty, preempt_active = postcopy_preempt_active(); int tmppages, pages = 0; size_t pagesize_bits = qemu_ram_pagesize(pss->block) >> TARGET_PAGE_BITS; @@ -2485,22 +2486,40 @@ static int ram_save_host_page(RAMState *rs, PageSearchStatus *pss) break; } -/* Check the pages is dirty and if it is send it */ -if (migration_bitmap_clear_dirty(rs, pss->block, pss->page)) { -tmppages = ram_save_target_page(rs, pss); -if (tmppages < 0) { -return tmppages; -} +page_dirty = migration_bitmap_clear_dirty(rs, pss->block, pss->page); -pages += tmppages; +/* Check the pages is dirty and if it is send it */ +if (page_dirty) { /* - * Allow rate limiting to happen in the middle of huge pages if - * something is sent in the current iteration. + * Properly yield the lock only in postcopy preempt mode + * because both migration thread and rp-return thread can + * operate on the bitmaps. */ -if (pagesize_bits > 1 && tmppages > 0) { -migration_rate_limit(); +if (preempt_active) { +qemu_mutex_unlock(>bitmap_mutex); +} +tmppages = ram_save_target_page(rs, pss); +if (tmppages >= 0) { +pages += tmppages; +/* + * Allow rate limiting to happen in the middle of huge pages if + * something is sent in the current iteration. + */ +if (pagesize_bits > 1 && tmppages > 0) { +migration_rate_limit(); +} } +if (preempt_active) { +qemu_mutex_lock(>bitmap_mutex); +} +} else { +tmppages = 0; } + +if (tmppages < 0) { +return tmppages; +} + pss->page = migration_bitmap_find_dirty(rs, pss->block, pss->page); } while ((pss->page < hostpage_boundary) && offset_in_ramblock(pss->block, -- 2.37.3