| Message ID | alpine.LRH.2.02.2208070732160.30857@file01.intranet.prod.int.rdu2.redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v5] add barriers to buffer functions | expand |
On Sun, Aug 07, 2022 at 07:37:22AM -0400, Mikulas Patocka wrote: > @@ -135,6 +133,49 @@ BUFFER_FNS(Meta, meta) > BUFFER_FNS(Prio, prio) > BUFFER_FNS(Defer_Completion, defer_completion) > > +static __always_inline void set_buffer_uptodate(struct buffer_head *bh) > +{ > + /* > + * make it consistent with folio_mark_uptodate > + * pairs with smp_acquire__after_ctrl_dep in buffer_uptodate > + */ > + smp_wmb(); > + set_bit(BH_Uptodate, &bh->b_state); > +} > + > +static __always_inline void clear_buffer_uptodate(struct buffer_head *bh) > +{ > + clear_bit(BH_Uptodate, &bh->b_state); > +} > + > +static __always_inline int buffer_uptodate(const struct buffer_head *bh) > +{ > + bool ret = test_bit(BH_Uptodate, &bh->b_state); > + /* > + * make it consistent with folio_test_uptodate > + * pairs with smp_wmb in set_buffer_uptodate > + */ > + if (ret) > + smp_acquire__after_ctrl_dep(); > + return ret; > +} This all works for me. While we have the experts paying attention, would it be better to do return smp_load_acquire(&bh->b_state) & (1L << BH_Uptodate) > 0; > +static __always_inline void set_buffer_locked(struct buffer_head *bh) > +{ > + set_bit(BH_Lock, &bh->b_state); > +} > + > +static __always_inline int buffer_locked(const struct buffer_head *bh) > +{ > + bool ret = test_bit(BH_Lock, &bh->b_state); > + /* > + * pairs with smp_mb__after_atomic in unlock_buffer > + */ > + if (!ret) > + smp_acquire__after_ctrl_dep(); > + return ret; > +} Are there places that think that lock/unlock buffer implies a memory barrier?
On Sun, 7 Aug 2022, Matthew Wilcox wrote: > On Sun, Aug 07, 2022 at 07:37:22AM -0400, Mikulas Patocka wrote: > > @@ -135,6 +133,49 @@ BUFFER_FNS(Meta, meta) > > BUFFER_FNS(Prio, prio) > > BUFFER_FNS(Defer_Completion, defer_completion) > > > > +static __always_inline void set_buffer_uptodate(struct buffer_head *bh) > > +{ > > + /* > > + * make it consistent with folio_mark_uptodate > > + * pairs with smp_acquire__after_ctrl_dep in buffer_uptodate > > + */ > > + smp_wmb(); > > + set_bit(BH_Uptodate, &bh->b_state); > > +} > > + > > +static __always_inline void clear_buffer_uptodate(struct buffer_head *bh) > > +{ > > + clear_bit(BH_Uptodate, &bh->b_state); > > +} > > + > > +static __always_inline int buffer_uptodate(const struct buffer_head *bh) > > +{ > > + bool ret = test_bit(BH_Uptodate, &bh->b_state); > > + /* > > + * make it consistent with folio_test_uptodate > > + * pairs with smp_wmb in set_buffer_uptodate > > + */ > > + if (ret) > > + smp_acquire__after_ctrl_dep(); > > + return ret; > > +} > > This all works for me. While we have the experts paying attention, > would it be better to do > > return smp_load_acquire(&bh->b_state) & (1L << BH_Uptodate) > 0; Yes, it may be nicer. > > +static __always_inline void set_buffer_locked(struct buffer_head *bh) > > +{ > > + set_bit(BH_Lock, &bh->b_state); > > +} > > + > > +static __always_inline int buffer_locked(const struct buffer_head *bh) > > +{ > > + bool ret = test_bit(BH_Lock, &bh->b_state); > > + /* > > + * pairs with smp_mb__after_atomic in unlock_buffer > > + */ > > + if (!ret) > > + smp_acquire__after_ctrl_dep(); > > + return ret; > > +} > > Are there places that think that lock/unlock buffer implies a memory > barrier? There's this in fs/reiserfs: if (!buffer_dirty(bh) && !buffer_locked(bh)) { reiserfs_free_jh(bh); <--- this could be moved before buffer_locked if (buffer_locked((journal->j_header_bh))) { ... } journal->j_last_flush_trans_id = trans_id; journal->j_first_unflushed_offset = offset; jh = (struct reiserfs_journal_header *)(journal->j_header_bh->b_data); <--- this could be moved before buffer_locked Mikulas
On Mon, Aug 08, 2022 at 10:26:10AM -0400, Mikulas Patocka wrote: > On Sun, 7 Aug 2022, Matthew Wilcox wrote: > > > +static __always_inline void set_buffer_locked(struct buffer_head *bh) > > > +{ > > > + set_bit(BH_Lock, &bh->b_state); > > > +} > > > + > > > +static __always_inline int buffer_locked(const struct buffer_head *bh) > > > +{ > > > + bool ret = test_bit(BH_Lock, &bh->b_state); > > > + /* > > > + * pairs with smp_mb__after_atomic in unlock_buffer > > > + */ > > > + if (!ret) > > > + smp_acquire__after_ctrl_dep(); > > > + return ret; > > > +} > > > > Are there places that think that lock/unlock buffer implies a memory > > barrier? > > There's this in fs/reiserfs: > > if (!buffer_dirty(bh) && !buffer_locked(bh)) { > reiserfs_free_jh(bh); <--- this could be moved before buffer_locked It might be better to think of buffer_locked() as buffer_someone_has_exclusive_access(). I can't see the problem with moving the reads in reiserfs_free_jh() before the read of buffer_locked. > if (buffer_locked((journal->j_header_bh))) { > ... > } > journal->j_last_flush_trans_id = trans_id; > journal->j_first_unflushed_offset = offset; > jh = (struct reiserfs_journal_header *)(journal->j_header_bh->b_data); <--- this could be moved before buffer_locked I don't think b_data is going to be changed while someone else holds the buffer locked. That's initialised by set_bh_page(), which is an initialisation-time thing, before the BH is visible to any other thread.
On Mon, 8 Aug 2022, Matthew Wilcox wrote: > On Mon, Aug 08, 2022 at 10:26:10AM -0400, Mikulas Patocka wrote: > > On Sun, 7 Aug 2022, Matthew Wilcox wrote: > > > > +static __always_inline void set_buffer_locked(struct buffer_head *bh) > > > > +{ > > > > + set_bit(BH_Lock, &bh->b_state); > > > > +} > > > > + > > > > +static __always_inline int buffer_locked(const struct buffer_head *bh) > > > > +{ > > > > + bool ret = test_bit(BH_Lock, &bh->b_state); > > > > + /* > > > > + * pairs with smp_mb__after_atomic in unlock_buffer > > > > + */ > > > > + if (!ret) > > > > + smp_acquire__after_ctrl_dep(); > > > > + return ret; > > > > +} > > > > > > Are there places that think that lock/unlock buffer implies a memory > > > barrier? > > > > There's this in fs/reiserfs: > > > > if (!buffer_dirty(bh) && !buffer_locked(bh)) { > > reiserfs_free_jh(bh); <--- this could be moved before buffer_locked > > It might be better to think of buffer_locked() as > buffer_someone_has_exclusive_access(). I can't see the problem with > moving the reads in reiserfs_free_jh() before the read of buffer_locked. > > > if (buffer_locked((journal->j_header_bh))) { > > ... > > } > > journal->j_last_flush_trans_id = trans_id; > > journal->j_first_unflushed_offset = offset; > > jh = (struct reiserfs_journal_header *)(journal->j_header_bh->b_data); <--- this could be moved before buffer_locked > > I don't think b_data is going to be changed while someone else holds > the buffer locked. That's initialised by set_bh_page(), which is an > initialisation-time thing, before the BH is visible to any other thread. So, do you think that we don't need a barrier in buffer_locked()? There is also this (where the BUG_ON(!buffer_uptodate(bh)) saves it). if (buffer_locked(bh)) { int depth; PROC_INFO_INC(sb, scan_bitmap.wait); depth = reiserfs_write_unlock_nested(sb); __wait_on_buffer(bh); reiserfs_write_lock_nested(sb, depth); } BUG_ON(!buffer_uptodate(bh)); BUG_ON(atomic_read(&bh->b_count) == 0); if (info->free_count == UINT_MAX) reiserfs_cache_bitmap_metadata(sb, bh, info); <--- this could be moved before buffer_locked if there were no BUG_ONs Mikulas
On Mon, Aug 08, 2022 at 10:57:45AM -0400, Mikulas Patocka wrote: > > > On Mon, 8 Aug 2022, Matthew Wilcox wrote: > > > On Mon, Aug 08, 2022 at 10:26:10AM -0400, Mikulas Patocka wrote: > > > On Sun, 7 Aug 2022, Matthew Wilcox wrote: > > > > > +static __always_inline void set_buffer_locked(struct buffer_head *bh) > > > > > +{ > > > > > + set_bit(BH_Lock, &bh->b_state); > > > > > +} > > > > > + > > > > > +static __always_inline int buffer_locked(const struct buffer_head *bh) > > > > > +{ > > > > > + bool ret = test_bit(BH_Lock, &bh->b_state); > > > > > + /* > > > > > + * pairs with smp_mb__after_atomic in unlock_buffer > > > > > + */ > > > > > + if (!ret) > > > > > + smp_acquire__after_ctrl_dep(); > > > > > + return ret; > > > > > +} > > > > > > > > Are there places that think that lock/unlock buffer implies a memory > > > > barrier? > > > > > > There's this in fs/reiserfs: > > > > > > if (!buffer_dirty(bh) && !buffer_locked(bh)) { > > > reiserfs_free_jh(bh); <--- this could be moved before buffer_locked > > > > It might be better to think of buffer_locked() as > > buffer_someone_has_exclusive_access(). I can't see the problem with > > moving the reads in reiserfs_free_jh() before the read of buffer_locked. > > > > > if (buffer_locked((journal->j_header_bh))) { > > > ... > > > } > > > journal->j_last_flush_trans_id = trans_id; > > > journal->j_first_unflushed_offset = offset; > > > jh = (struct reiserfs_journal_header *)(journal->j_header_bh->b_data); <--- this could be moved before buffer_locked > > > > I don't think b_data is going to be changed while someone else holds > > the buffer locked. That's initialised by set_bh_page(), which is an > > initialisation-time thing, before the BH is visible to any other thread. > > So, do you think that we don't need a barrier in buffer_locked()? The question to ask here is "What prevents another call to buffer_locked() from returning false?" > There is also this (where the BUG_ON(!buffer_uptodate(bh)) saves it). > if (buffer_locked(bh)) { Right here, for example. If something prevents any change that might cause buffer_locked() to return false here, we don't need a barrier. If there is nothing preventing such a change, how is a barrier going to help? One way this code could be correct is if the above check is a heuristic, so that a false positive just consumes a bit more CPU and a false negative just delays this action. I must leave final judgment to those having better understanding of this code than do I. Thanx, Paul > int depth; > PROC_INFO_INC(sb, scan_bitmap.wait); > depth = reiserfs_write_unlock_nested(sb); > __wait_on_buffer(bh); > reiserfs_write_lock_nested(sb, depth); > } > BUG_ON(!buffer_uptodate(bh)); > BUG_ON(atomic_read(&bh->b_count) == 0); > > if (info->free_count == UINT_MAX) > reiserfs_cache_bitmap_metadata(sb, bh, info); <--- this could be moved before buffer_locked if there were no BUG_ONs > > Mikulas >
On Mon, Aug 08, 2022 at 10:57:45AM -0400, Mikulas Patocka wrote: > On Mon, 8 Aug 2022, Matthew Wilcox wrote: > > > On Mon, Aug 08, 2022 at 10:26:10AM -0400, Mikulas Patocka wrote: > > > On Sun, 7 Aug 2022, Matthew Wilcox wrote: > > > > > +static __always_inline void set_buffer_locked(struct buffer_head *bh) > > > > > +{ > > > > > + set_bit(BH_Lock, &bh->b_state); > > > > > +} > > > > > + > > > > > +static __always_inline int buffer_locked(const struct buffer_head *bh) > > > > > +{ > > > > > + bool ret = test_bit(BH_Lock, &bh->b_state); > > > > > + /* > > > > > + * pairs with smp_mb__after_atomic in unlock_buffer > > > > > + */ > > > > > + if (!ret) > > > > > + smp_acquire__after_ctrl_dep(); > > > > > + return ret; > > > > > +} > > > > > > > > Are there places that think that lock/unlock buffer implies a memory > > > > barrier? > > > > > > There's this in fs/reiserfs: > > > > > > if (!buffer_dirty(bh) && !buffer_locked(bh)) { > > > reiserfs_free_jh(bh); <--- this could be moved before buffer_locked > > > > It might be better to think of buffer_locked() as > > buffer_someone_has_exclusive_access(). I can't see the problem with > > moving the reads in reiserfs_free_jh() before the read of buffer_locked. > > > > > if (buffer_locked((journal->j_header_bh))) { > > > ... > > > } > > > journal->j_last_flush_trans_id = trans_id; > > > journal->j_first_unflushed_offset = offset; > > > jh = (struct reiserfs_journal_header *)(journal->j_header_bh->b_data); <--- this could be moved before buffer_locked > > > > I don't think b_data is going to be changed while someone else holds > > the buffer locked. That's initialised by set_bh_page(), which is an > > initialisation-time thing, before the BH is visible to any other thread. > > So, do you think that we don't need a barrier in buffer_locked()? That's my feeling. Of course, you might not be the only one confused, and if fs authors in general have made the mistake of thinking that buffer_locked is serialising, then it might be better to live up to that expectation. > There is also this (where the BUG_ON(!buffer_uptodate(bh)) saves it). > if (buffer_locked(bh)) { > int depth; > PROC_INFO_INC(sb, scan_bitmap.wait); > depth = reiserfs_write_unlock_nested(sb); > __wait_on_buffer(bh); > reiserfs_write_lock_nested(sb, depth); > } > BUG_ON(!buffer_uptodate(bh)); > BUG_ON(atomic_read(&bh->b_count) == 0); > > if (info->free_count == UINT_MAX) > reiserfs_cache_bitmap_metadata(sb, bh, info); <--- this could be moved before buffer_locked if there were no BUG_ONs It could be moved before buffer_locked(), but I don't see the harm in that. Look at how reiserfs_read_bitmap_block() gets the bh: bh = sb_bread(sb, block); __bread_gfp() has either already read the buffer (and it's uptodate), in which case it returns it. Or it calls __bread_slow() which will do the read and check uptodate before returning it. I wouldn't be surprised to find that this buffer_locked() check is actually dead code, but I have no desire to dive into reiserfs far enough to find out whether it's dead code or not.
Index: linux-2.6/include/linux/buffer_head.h =================================================================== --- linux-2.6.orig/include/linux/buffer_head.h +++ linux-2.6/include/linux/buffer_head.h @@ -117,10 +117,8 @@ static __always_inline int test_clear_bu * of the form "mark_buffer_foo()". These are higher-level functions which * do something in addition to setting a b_state bit. */ -BUFFER_FNS(Uptodate, uptodate) BUFFER_FNS(Dirty, dirty) TAS_BUFFER_FNS(Dirty, dirty) -BUFFER_FNS(Lock, locked) BUFFER_FNS(Req, req) TAS_BUFFER_FNS(Req, req) BUFFER_FNS(Mapped, mapped) @@ -135,6 +133,49 @@ BUFFER_FNS(Meta, meta) BUFFER_FNS(Prio, prio) BUFFER_FNS(Defer_Completion, defer_completion) +static __always_inline void set_buffer_uptodate(struct buffer_head *bh) +{ + /* + * make it consistent with folio_mark_uptodate + * pairs with smp_acquire__after_ctrl_dep in buffer_uptodate + */ + smp_wmb(); + set_bit(BH_Uptodate, &bh->b_state); +} + +static __always_inline void clear_buffer_uptodate(struct buffer_head *bh) +{ + clear_bit(BH_Uptodate, &bh->b_state); +} + +static __always_inline int buffer_uptodate(const struct buffer_head *bh) +{ + bool ret = test_bit(BH_Uptodate, &bh->b_state); + /* + * make it consistent with folio_test_uptodate + * pairs with smp_wmb in set_buffer_uptodate + */ + if (ret) + smp_acquire__after_ctrl_dep(); + return ret; +} + +static __always_inline void set_buffer_locked(struct buffer_head *bh) +{ + set_bit(BH_Lock, &bh->b_state); +} + +static __always_inline int buffer_locked(const struct buffer_head *bh) +{ + bool ret = test_bit(BH_Lock, &bh->b_state); + /* + * pairs with smp_mb__after_atomic in unlock_buffer + */ + if (!ret) + smp_acquire__after_ctrl_dep(); + return ret; +} + #define bh_offset(bh) ((unsigned long)(bh)->b_data & ~PAGE_MASK) /* If we *know* page->private refers to buffer_heads */