Message ID | 20201102194135.174806-5-preichl@redhat.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | xfs: Remove wrappers for some semaphores | expand |
On Mon, Nov 02, 2020 at 08:41:35PM +0100, Pavel Reichl wrote: > Remove mrlock_t as it does not provide any extra value over > rw_semaphores. Make i_lock and i_mmaplock native rw_semaphores and > replace mr*() functions with native rwsem calls. > > Release the lock in xfs_btree_split() just before the work-queue > executing xfs_btree_split_worker() is scheduled and make > xfs_btree_split_worker() to acquire the lock as a first thing and > release it just before returning from the function. This it done so the > ownership of the lock is transfered between kernel threads and thus > lockdep won't complain about lock being held by a different kernel > thread. > > Signed-off-by: Pavel Reichl <preichl@redhat.com> > Reviewed-by: Christoph Hellwig <hch@lst.de> > Reviewed-by: Brian Foster <bfoster@redhat.com> > > --- > fs/xfs/libxfs/xfs_btree.c | 16 ++++++++ > fs/xfs/mrlock.h | 78 --------------------------------------- > fs/xfs/xfs_inode.c | 52 ++++++++++++++------------ > fs/xfs/xfs_inode.h | 4 +- > fs/xfs/xfs_iops.c | 4 +- > fs/xfs/xfs_linux.h | 2 +- > fs/xfs/xfs_super.c | 6 +-- > 7 files changed, 51 insertions(+), 111 deletions(-) > delete mode 100644 fs/xfs/mrlock.h > > diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c > index 2d25bab68764..181d5797c97b 100644 > --- a/fs/xfs/libxfs/xfs_btree.c > +++ b/fs/xfs/libxfs/xfs_btree.c > @@ -2816,6 +2816,10 @@ xfs_btree_split_worker( > unsigned long pflags; > unsigned long new_pflags = PF_MEMALLOC_NOFS; > > + /* > + * Tranfer lock ownership to workqueue task. "Transfer", not "tranfer". I'll just fix that so you don't have to send this yet again. MR. LOCK is dead, long live MR. LOCK! Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> --D > + */ > + rwsem_acquire(&args->cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); > /* > * we are in a transaction context here, but may also be doing work > * in kswapd context, and hence we may need to inherit that state > @@ -2829,6 +2833,7 @@ xfs_btree_split_worker( > > args->result = __xfs_btree_split(args->cur, args->level, args->ptrp, > args->key, args->curp, args->stat); > + rwsem_release(&args->cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); > complete(args->done); > > current_restore_flags_nested(&pflags, new_pflags); > @@ -2863,8 +2868,19 @@ xfs_btree_split( > args.done = &done; > args.kswapd = current_is_kswapd(); > INIT_WORK_ONSTACK(&args.work, xfs_btree_split_worker); > + /* > + * Update lockdep's ownership information to reflect transfer of the > + * ilock from the current task to the worker. Otherwise assertions that > + * the lock is held (such as when logging the inode) might fail due to > + * incorrect task owner state. > + */ > + rwsem_release(&cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); > queue_work(xfs_alloc_wq, &args.work); > wait_for_completion(&done); > + /* > + * Tranfer lock ownership back to the thread. > + */ > + rwsem_acquire(&cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); > destroy_work_on_stack(&args.work); > return args.result; > } > diff --git a/fs/xfs/mrlock.h b/fs/xfs/mrlock.h > deleted file mode 100644 > index 79155eec341b..000000000000 > --- a/fs/xfs/mrlock.h > +++ /dev/null > @@ -1,78 +0,0 @@ > -// SPDX-License-Identifier: GPL-2.0 > -/* > - * Copyright (c) 2000-2006 Silicon Graphics, Inc. > - * All Rights Reserved. > - */ > -#ifndef __XFS_SUPPORT_MRLOCK_H__ > -#define __XFS_SUPPORT_MRLOCK_H__ > - > -#include <linux/rwsem.h> > - > -typedef struct { > - struct rw_semaphore mr_lock; > -#if defined(DEBUG) || defined(XFS_WARN) > - int mr_writer; > -#endif > -} mrlock_t; > - > -#if defined(DEBUG) || defined(XFS_WARN) > -#define mrinit(mrp, name) \ > - do { (mrp)->mr_writer = 0; init_rwsem(&(mrp)->mr_lock); } while (0) > -#else > -#define mrinit(mrp, name) \ > - do { init_rwsem(&(mrp)->mr_lock); } while (0) > -#endif > - > -#define mrlock_init(mrp, t,n,s) mrinit(mrp, n) > -#define mrfree(mrp) do { } while (0) > - > -static inline void mraccess_nested(mrlock_t *mrp, int subclass) > -{ > - down_read_nested(&mrp->mr_lock, subclass); > -} > - > -static inline void mrupdate_nested(mrlock_t *mrp, int subclass) > -{ > - down_write_nested(&mrp->mr_lock, subclass); > -#if defined(DEBUG) || defined(XFS_WARN) > - mrp->mr_writer = 1; > -#endif > -} > - > -static inline int mrtryaccess(mrlock_t *mrp) > -{ > - return down_read_trylock(&mrp->mr_lock); > -} > - > -static inline int mrtryupdate(mrlock_t *mrp) > -{ > - if (!down_write_trylock(&mrp->mr_lock)) > - return 0; > -#if defined(DEBUG) || defined(XFS_WARN) > - mrp->mr_writer = 1; > -#endif > - return 1; > -} > - > -static inline void mrunlock_excl(mrlock_t *mrp) > -{ > -#if defined(DEBUG) || defined(XFS_WARN) > - mrp->mr_writer = 0; > -#endif > - up_write(&mrp->mr_lock); > -} > - > -static inline void mrunlock_shared(mrlock_t *mrp) > -{ > - up_read(&mrp->mr_lock); > -} > - > -static inline void mrdemote(mrlock_t *mrp) > -{ > -#if defined(DEBUG) || defined(XFS_WARN) > - mrp->mr_writer = 0; > -#endif > - downgrade_write(&mrp->mr_lock); > -} > - > -#endif /* __XFS_SUPPORT_MRLOCK_H__ */ > diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c > index 16d481cf3793..43ecfcb63f99 100644 > --- a/fs/xfs/xfs_inode.c > +++ b/fs/xfs/xfs_inode.c > @@ -191,14 +191,15 @@ xfs_ilock( > } > > if (lock_flags & XFS_MMAPLOCK_EXCL) > - mrupdate_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); > + down_write_nested(&ip->i_mmaplock, > + XFS_MMAPLOCK_DEP(lock_flags)); > else if (lock_flags & XFS_MMAPLOCK_SHARED) > - mraccess_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); > + down_read_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); > > if (lock_flags & XFS_ILOCK_EXCL) > - mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); > + down_write_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); > else if (lock_flags & XFS_ILOCK_SHARED) > - mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); > + down_read_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); > } > > /* > @@ -242,27 +243,27 @@ xfs_ilock_nowait( > } > > if (lock_flags & XFS_MMAPLOCK_EXCL) { > - if (!mrtryupdate(&ip->i_mmaplock)) > + if (!down_write_trylock(&ip->i_mmaplock)) > goto out_undo_iolock; > } else if (lock_flags & XFS_MMAPLOCK_SHARED) { > - if (!mrtryaccess(&ip->i_mmaplock)) > + if (!down_read_trylock(&ip->i_mmaplock)) > goto out_undo_iolock; > } > > if (lock_flags & XFS_ILOCK_EXCL) { > - if (!mrtryupdate(&ip->i_lock)) > + if (!down_write_trylock(&ip->i_lock)) > goto out_undo_mmaplock; > } else if (lock_flags & XFS_ILOCK_SHARED) { > - if (!mrtryaccess(&ip->i_lock)) > + if (!down_read_trylock(&ip->i_lock)) > goto out_undo_mmaplock; > } > return 1; > > out_undo_mmaplock: > if (lock_flags & XFS_MMAPLOCK_EXCL) > - mrunlock_excl(&ip->i_mmaplock); > + up_write(&ip->i_mmaplock); > else if (lock_flags & XFS_MMAPLOCK_SHARED) > - mrunlock_shared(&ip->i_mmaplock); > + up_read(&ip->i_mmaplock); > out_undo_iolock: > if (lock_flags & XFS_IOLOCK_EXCL) > up_write(&VFS_I(ip)->i_rwsem); > @@ -309,14 +310,14 @@ xfs_iunlock( > up_read(&VFS_I(ip)->i_rwsem); > > if (lock_flags & XFS_MMAPLOCK_EXCL) > - mrunlock_excl(&ip->i_mmaplock); > + up_write(&ip->i_mmaplock); > else if (lock_flags & XFS_MMAPLOCK_SHARED) > - mrunlock_shared(&ip->i_mmaplock); > + up_read(&ip->i_mmaplock); > > if (lock_flags & XFS_ILOCK_EXCL) > - mrunlock_excl(&ip->i_lock); > + up_write(&ip->i_lock); > else if (lock_flags & XFS_ILOCK_SHARED) > - mrunlock_shared(&ip->i_lock); > + up_read(&ip->i_lock); > > trace_xfs_iunlock(ip, lock_flags, _RET_IP_); > } > @@ -335,9 +336,9 @@ xfs_ilock_demote( > ~(XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL)) == 0); > > if (lock_flags & XFS_ILOCK_EXCL) > - mrdemote(&ip->i_lock); > + downgrade_write(&ip->i_lock); > if (lock_flags & XFS_MMAPLOCK_EXCL) > - mrdemote(&ip->i_mmaplock); > + downgrade_write(&ip->i_mmaplock); > if (lock_flags & XFS_IOLOCK_EXCL) > downgrade_write(&VFS_I(ip)->i_rwsem); > > @@ -375,19 +376,22 @@ xfs_isilocked( > struct xfs_inode *ip, > uint lock_flags) > { > - if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) { > - if (!(lock_flags & XFS_ILOCK_SHARED)) > - return !!ip->i_lock.mr_writer; > - return rwsem_is_locked(&ip->i_lock.mr_lock); > + if (lock_flags & (XFS_ILOCK_EXCL | XFS_ILOCK_SHARED)) { > + ASSERT(!(lock_flags & ~(XFS_ILOCK_EXCL | XFS_ILOCK_SHARED))); > + return __xfs_rwsem_islocked(&ip->i_lock, lock_flags, > + XFS_ILOCK_FLAG_SHIFT); > } > > - if (lock_flags & (XFS_MMAPLOCK_EXCL|XFS_MMAPLOCK_SHARED)) { > - if (!(lock_flags & XFS_MMAPLOCK_SHARED)) > - return !!ip->i_mmaplock.mr_writer; > - return rwsem_is_locked(&ip->i_mmaplock.mr_lock); > + if (lock_flags & (XFS_MMAPLOCK_EXCL | XFS_MMAPLOCK_SHARED)) { > + ASSERT(!(lock_flags & > + ~(XFS_MMAPLOCK_EXCL | XFS_MMAPLOCK_SHARED))); > + return __xfs_rwsem_islocked(&ip->i_mmaplock, lock_flags, > + XFS_MMAPLOCK_FLAG_SHIFT); > } > > if (lock_flags & (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED)) { > + ASSERT(!(lock_flags & > + ~(XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED))); > return __xfs_rwsem_islocked(&VFS_I(ip)->i_rwsem, lock_flags, > XFS_IOLOCK_FLAG_SHIFT); > } > diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h > index 1392a9c452ae..66ceb127192e 100644 > --- a/fs/xfs/xfs_inode.h > +++ b/fs/xfs/xfs_inode.h > @@ -39,8 +39,8 @@ typedef struct xfs_inode { > > /* Transaction and locking information. */ > struct xfs_inode_log_item *i_itemp; /* logging information */ > - mrlock_t i_lock; /* inode lock */ > - mrlock_t i_mmaplock; /* inode mmap IO lock */ > + struct rw_semaphore i_lock; /* inode lock */ > + struct rw_semaphore i_mmaplock; /* inode mmap IO lock */ > atomic_t i_pincount; /* inode pin count */ > > /* > diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c > index 5e165456da68..8181f6785a7a 100644 > --- a/fs/xfs/xfs_iops.c > +++ b/fs/xfs/xfs_iops.c > @@ -1336,9 +1336,9 @@ xfs_setup_inode( > */ > lockdep_set_class(&inode->i_rwsem, > &inode->i_sb->s_type->i_mutex_dir_key); > - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class); > + lockdep_set_class(&ip->i_lock, &xfs_dir_ilock_class); > } else { > - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class); > + lockdep_set_class(&ip->i_lock, &xfs_nondir_ilock_class); > } > > /* > diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h > index 5b7a1e201559..64e28ec16cf7 100644 > --- a/fs/xfs/xfs_linux.h > +++ b/fs/xfs/xfs_linux.h > @@ -22,7 +22,6 @@ typedef __u32 xfs_nlink_t; > #include "xfs_types.h" > > #include "kmem.h" > -#include "mrlock.h" > > #include <linux/semaphore.h> > #include <linux/mm.h> > @@ -61,6 +60,7 @@ typedef __u32 xfs_nlink_t; > #include <linux/ratelimit.h> > #include <linux/rhashtable.h> > #include <linux/xattr.h> > +#include <linux/rwsem.h> > > #include <asm/page.h> > #include <asm/div64.h> > diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c > index e3e229e52512..380ba196b165 100644 > --- a/fs/xfs/xfs_super.c > +++ b/fs/xfs/xfs_super.c > @@ -708,10 +708,8 @@ xfs_fs_inode_init_once( > atomic_set(&ip->i_pincount, 0); > spin_lock_init(&ip->i_flags_lock); > > - mrlock_init(&ip->i_mmaplock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER, > - "xfsino", ip->i_ino); > - mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER, > - "xfsino", ip->i_ino); > + init_rwsem(&ip->i_mmaplock); > + init_rwsem(&ip->i_lock); > } > > /* > -- > 2.26.2 >
On Mon, Nov 02, 2020 at 08:41:35PM +0100, Pavel Reichl wrote: > Remove mrlock_t as it does not provide any extra value over > rw_semaphores. Make i_lock and i_mmaplock native rw_semaphores and > replace mr*() functions with native rwsem calls. > > Release the lock in xfs_btree_split() just before the work-queue > executing xfs_btree_split_worker() is scheduled and make > xfs_btree_split_worker() to acquire the lock as a first thing and > release it just before returning from the function. This it done so the > ownership of the lock is transfered between kernel threads and thus > lockdep won't complain about lock being held by a different kernel > thread. > > Signed-off-by: Pavel Reichl <preichl@redhat.com> > Reviewed-by: Christoph Hellwig <hch@lst.de> > Reviewed-by: Brian Foster <bfoster@redhat.com> > > --- > fs/xfs/libxfs/xfs_btree.c | 16 ++++++++ > fs/xfs/mrlock.h | 78 --------------------------------------- > fs/xfs/xfs_inode.c | 52 ++++++++++++++------------ > fs/xfs/xfs_inode.h | 4 +- > fs/xfs/xfs_iops.c | 4 +- > fs/xfs/xfs_linux.h | 2 +- > fs/xfs/xfs_super.c | 6 +-- > 7 files changed, 51 insertions(+), 111 deletions(-) > delete mode 100644 fs/xfs/mrlock.h > > diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c > index 2d25bab68764..181d5797c97b 100644 > --- a/fs/xfs/libxfs/xfs_btree.c > +++ b/fs/xfs/libxfs/xfs_btree.c > @@ -2816,6 +2816,10 @@ xfs_btree_split_worker( w unsigned long pflags; > unsigned long new_pflags = PF_MEMALLOC_NOFS; > > + /* > + * Tranfer lock ownership to workqueue task. > + */ > + rwsem_acquire(&args->cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); > /* > * we are in a transaction context here, but may also be doing work > * in kswapd context, and hence we may need to inherit that state > @@ -2829,6 +2833,7 @@ xfs_btree_split_worker( > > args->result = __xfs_btree_split(args->cur, args->level, args->ptrp, > args->key, args->curp, args->stat); > + rwsem_release(&args->cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); > complete(args->done); > > current_restore_flags_nested(&pflags, new_pflags); > @@ -2863,8 +2868,19 @@ xfs_btree_split( > args.done = &done; > args.kswapd = current_is_kswapd(); > INIT_WORK_ONSTACK(&args.work, xfs_btree_split_worker); > + /* > + * Update lockdep's ownership information to reflect transfer of the > + * ilock from the current task to the worker. Otherwise assertions that > + * the lock is held (such as when logging the inode) might fail due to > + * incorrect task owner state. > + */ > + rwsem_release(&cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); > queue_work(xfs_alloc_wq, &args.work); > wait_for_completion(&done); > + /* > + * Tranfer lock ownership back to the thread. > + */ > + rwsem_acquire(&cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); So I ran all this through fstests. On generic/324 on a fstests run with reasonable recent xfsprogs and all the features turned on (rmap in particular) I see the following lockdep report: ============================================ WARNING: possible recursive locking detected 5.10.0-rc1-djw #rc1 Not tainted -------------------------------------------- xfs_fsr/5438 is trying to acquire lock: ffff88801826b028 (&xfs_nondir_ilock_class){++++}-{3:3}, at: xfs_btree_make_block_unfull+0x19a/0x200 [xfs] but task is already holding lock: ffff88807fbc58a8 (&xfs_nondir_ilock_class){++++}-{3:3}, at: xfs_lock_two_inodes+0x116/0x3e0 [xfs] other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&xfs_nondir_ilock_class); lock(&xfs_nondir_ilock_class); *** DEADLOCK *** May be due to missing lock nesting notation 7 locks held by xfs_fsr/5438: #0: ffff88804166a430 (sb_writers#12){.+.+}-{0:0}, at: mnt_want_write_file+0x23/0x70 #1: ffff88801826b2d0 (&sb->s_type->i_mutex_key#13){++++}-{3:3}, at: lock_two_nondirectories+0x68/0x70 #2: ffff88807fbc5b50 (&sb->s_type->i_mutex_key#13/4){+.+.}-{3:3}, at: xfs_swap_extents+0x4e/0x8e0 [xfs] #3: ffff88807fbc5940 (&ip->i_mmaplock){+.+.}-{3:3}, at: xfs_ilock+0xfa/0x270 [xfs] #4: ffff88801826b0c0 (&ip->i_mmaplock){+.+.}-{3:3}, at: xfs_ilock_nowait+0x17f/0x310 [xfs] #5: ffff88804166a620 (sb_internal){.+.+}-{0:0}, at: xfs_trans_alloc+0x1b0/0x2a0 [xfs] #6: ffff88807fbc58a8 (&xfs_nondir_ilock_class){++++}-{3:3}, at: xfs_lock_two_inodes+0x116/0x3e0 [xfs] stack backtrace: CPU: 1 PID: 5438 Comm: xfs_fsr Not tainted 5.10.0-rc1-djw #rc1 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1 04/01/2014 Call Trace: dump_stack+0x7b/0x9b __lock_acquire.cold+0x208/0x2b9 lock_acquire+0xcd/0x430 ? xfs_btree_make_block_unfull+0x19a/0x200 [xfs] ? mark_held_locks+0x49/0x70 xfs_btree_split+0x1a2/0x1c0 [xfs] ? xfs_btree_make_block_unfull+0x19a/0x200 [xfs] ? xfs_btree_split+0x1c0/0x1c0 [xfs] xfs_btree_make_block_unfull+0x19a/0x200 [xfs] xfs_btree_insrec+0x448/0x9a0 [xfs] ? xfs_btree_lookup_get_block+0xee/0x1a0 [xfs] xfs_btree_insert+0xa0/0x1f0 [xfs] xfs_bmap_add_extent_hole_real+0x272/0xae0 [xfs] xfs_bmapi_remap+0x1d5/0x400 [xfs] xfs_bmap_finish_one+0x1ac/0x2a0 [xfs] xfs_bmap_update_finish_item+0x53/0xd0 [xfs] xfs_defer_finish_noroll+0x271/0xa90 [xfs] xfs_defer_finish+0x15/0xa0 [xfs] xfs_swap_extent_rmap+0x27c/0x810 [xfs] xfs_swap_extents+0x7d8/0x8e0 [xfs] xfs_ioc_swapext+0x131/0x150 [xfs] xfs_file_ioctl+0x291/0xca0 [xfs] __x64_sys_ioctl+0x87/0xa0 do_syscall_64+0x31/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9 RIP: 0033:0x7f73472ea50b Code: 0f 1e fa 48 8b 05 85 39 0d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 0f 1f 44 00 00 f3 0f 1e fa b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 55 39 0d 00 f7 d8 64 89 01 48 RSP: 002b:00007ffc7f707778 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 RAX: ffffffffffffffda RBX: 00007ffc7f7078e0 RCX: 00007f73472ea50b RDX: 00007ffc7f707780 RSI: 00000000c0c0586d RDI: 0000000000000005 RBP: 00007ffc7f7097c0 R08: 0000000000000005 R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: 00007ffc7f7097c0 R13: 00005567684f90c0 R14: 00000000000000f9 R15: 00000000000000f9 I think it's significant that xfs_lock_two_inodes figures in the lockdep report above. I suspect what's going on here is: xfs_swap_extents takes the ILOCK of the two inodes it's operating on. The first ILOCK gets tagged with subclass 0 and the second file's ILOCK gets subclass 1. While swapping extents, we decide that the second file requires a bmbt split and that we have to do this via workqueue. We tell lockdep to drop the current thread (pid 5438) from the lockdep map and kick off the worker. When the worker returns, however, we don't feed the subclass information to rwsem_acquire (the second parameter). Therefore, lockdep sees that pid 5438 currently holds one subclass 0 ILOCK (the first inode, which we didn't unlock) and complains that we're asking it to grab another subclass 0 ILOCK (the second inode). Unfortunately, I don't think you can solve this problem solely by passing the ILOCK subclass information to rwsem_acquire, either. generic/558 produces the following splat: ====================================================== WARNING: possible circular locking dependency detected 5.10.0-rc1-djw #rc1 Not tainted ------------------------------------------------------ rm/25035 is trying to acquire lock: ffff888009f5c970 (&xfs_dir_ilock_class){++++}-{3:3}, at: xfs_btree_make_block_unfull+0x19a/0x200 [xfs] but task is already holding lock: ffff88800af0c230 (&xfs_nondir_ilock_class/1){+.+.}-{3:3}, at: xfs_remove+0x180/0x510 [xfs] which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #1 (&xfs_nondir_ilock_class/1){+.+.}-{3:3}: down_write_nested+0x48/0x120 xfs_remove+0x180/0x510 [xfs] xfs_vn_unlink+0x57/0xa0 [xfs] vfs_unlink+0xf2/0x1e0 do_unlinkat+0x1a5/0x2e0 do_syscall_64+0x31/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9 -> #0 (&xfs_dir_ilock_class){++++}-{3:3}: __lock_acquire+0x1223/0x2200 lock_acquire+0xcd/0x430 xfs_btree_split+0x1a5/0x1c0 [xfs] xfs_btree_make_block_unfull+0x19a/0x200 [xfs] xfs_btree_insrec+0x448/0x9a0 [xfs] xfs_btree_insert+0xa0/0x1f0 [xfs] xfs_bmap_del_extent_real+0x6fc/0xc60 [xfs] __xfs_bunmapi+0x8ba/0x10b0 [xfs] xfs_bunmapi+0x19/0x30 [xfs] xfs_dir2_shrink_inode+0x94/0x2d0 [xfs] xfs_dir2_node_removename+0x87f/0x9e0 [xfs] xfs_dir_removename+0x1eb/0x2b0 [xfs] xfs_remove+0x426/0x510 [xfs] xfs_vn_unlink+0x57/0xa0 [xfs] vfs_unlink+0xf2/0x1e0 do_unlinkat+0x1a5/0x2e0 do_syscall_64+0x31/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9 other info that might help us debug this: Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&xfs_nondir_ilock_class/1); lock(&xfs_dir_ilock_class); lock(&xfs_nondir_ilock_class/1); lock(&xfs_dir_ilock_class); *** DEADLOCK *** 5 locks held by rm/25035: #0: ffff88803b9f8468 (sb_writers#12){.+.+}-{0:0}, at: mnt_want_write+0x1f/0x50 #1: ffff888009f5cc48 (&inode->i_sb->s_type->i_mutex_dir_key/1){+.+.}-{3:3}, at: do_unlinkat+0x13c/0x2e0 #2: ffff88800af0c508 (&sb->s_type->i_mutex_key#13){++++}-{3:3}, at: vfs_unlink+0x4c/0x1e0 #3: ffff88803b9f8688 (sb_internal){.+.+}-{0:0}, at: xfs_trans_alloc+0x1b0/0x2a0 [xfs] #4: ffff88800af0c230 (&xfs_nondir_ilock_class/1){+.+.}-{3:3}, at: xfs_remove+0x180/0x510 [xfs] stack backtrace: CPU: 1 PID: 25035 Comm: rm Not tainted 5.10.0-rc1-djw #rc1 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1 04/01/2014 Call Trace: dump_stack+0x7b/0x9b check_noncircular+0xf3/0x110 __lock_acquire+0x1223/0x2200 lock_acquire+0xcd/0x430 ? xfs_btree_make_block_unfull+0x19a/0x200 [xfs] ? mark_held_locks+0x45/0x70 xfs_btree_split+0x1a5/0x1c0 [xfs] ? xfs_btree_make_block_unfull+0x19a/0x200 [xfs] ? wait_for_completion+0xba/0x110 ? xfs_btree_split+0x1c0/0x1c0 [xfs] xfs_btree_make_block_unfull+0x19a/0x200 [xfs] xfs_btree_insrec+0x448/0x9a0 [xfs] xfs_btree_insert+0xa0/0x1f0 [xfs] ? xfs_btree_increment+0x95/0x590 [xfs] xfs_bmap_del_extent_real+0x6fc/0xc60 [xfs] __xfs_bunmapi+0x8ba/0x10b0 [xfs] xfs_bunmapi+0x19/0x30 [xfs] xfs_dir2_shrink_inode+0x94/0x2d0 [xfs] xfs_dir2_node_removename+0x87f/0x9e0 [xfs] xfs_dir_removename+0x1eb/0x2b0 [xfs] ? xfs_iunlink+0x169/0x310 [xfs] xfs_remove+0x426/0x510 [xfs] xfs_vn_unlink+0x57/0xa0 [xfs] vfs_unlink+0xf2/0x1e0 do_unlinkat+0x1a5/0x2e0 do_syscall_64+0x31/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9 RIP: 0033:0x7f892cca4e6b Just like in the first scenario, we hold two ILOCKs -- a directory, and a file that we're removing from the directory. The removal triggers the same bmbt split worker because we're shrinking the directory, and upon its completion we call rwsem_acquire to reset the lockdep maps. Unfortunately, in this case, we actually /are/ feeding the correct subclass information to rwsem_acquire. This time it's pointing out what it thinks is an inconsistency in our locking order: the first time we locked the directory and then the regular file inode, but now we hold the regular file inode and we're asking it for the directory ILOCK. (Note that we don't actually deadlock here because pid 25035 has maintained ownership of the directory ILOCK rwsem this whole time, but lockdep doesn't know that.) A crappy way to bypass this problem is the following garbage patch which disables lockdep chain checking since we never actually dropped any of the ILOCKs that are being complained about. Messing with low level lockdep internals seems sketchy to me, but so it goes. The patch also has the major flaw that it doesn't recapture the subclass information, but doing that is left as an exercise to the reader. ;) --D diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 9756a63e78f4..3146932de7fe 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -2820,7 +2820,7 @@ xfs_btree_split_worker( /* * Tranfer lock ownership to workqueue task. */ - rwsem_acquire(&args->cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); + lock_acquire(&args->cur->bc_ino.ip->i_lock.dep_map, 0, 0, 0, 0, NULL, _RET_IP_); /* * we are in a transaction context here, but may also be doing work * in kswapd context, and hence we may need to inherit that state @@ -2882,7 +2882,7 @@ xfs_btree_split( /* * Tranfer lock ownership back to the thread. */ - rwsem_acquire(&cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); + lock_acquire(&cur->bc_ino.ip->i_lock.dep_map, 0, 0, 0, 0, NULL, _RET_IP_); destroy_work_on_stack(&args.work); return args.result; }
On Tue, Nov 03, 2020 at 04:53:45PM -0800, Darrick J. Wong wrote: > On Mon, Nov 02, 2020 at 08:41:35PM +0100, Pavel Reichl wrote: > > Remove mrlock_t as it does not provide any extra value over > > rw_semaphores. Make i_lock and i_mmaplock native rw_semaphores and > > replace mr*() functions with native rwsem calls. > > > > Release the lock in xfs_btree_split() just before the work-queue > > executing xfs_btree_split_worker() is scheduled and make > > xfs_btree_split_worker() to acquire the lock as a first thing and > > release it just before returning from the function. This it done so the > > ownership of the lock is transfered between kernel threads and thus > > lockdep won't complain about lock being held by a different kernel > > thread. > > > > Signed-off-by: Pavel Reichl <preichl@redhat.com> > > Reviewed-by: Christoph Hellwig <hch@lst.de> > > Reviewed-by: Brian Foster <bfoster@redhat.com> > > > > --- > > fs/xfs/libxfs/xfs_btree.c | 16 ++++++++ > > fs/xfs/mrlock.h | 78 --------------------------------------- > > fs/xfs/xfs_inode.c | 52 ++++++++++++++------------ > > fs/xfs/xfs_inode.h | 4 +- > > fs/xfs/xfs_iops.c | 4 +- > > fs/xfs/xfs_linux.h | 2 +- > > fs/xfs/xfs_super.c | 6 +-- > > 7 files changed, 51 insertions(+), 111 deletions(-) > > delete mode 100644 fs/xfs/mrlock.h > > > > diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c > > index 2d25bab68764..181d5797c97b 100644 > > --- a/fs/xfs/libxfs/xfs_btree.c > > +++ b/fs/xfs/libxfs/xfs_btree.c > > @@ -2816,6 +2816,10 @@ xfs_btree_split_worker( > w unsigned long pflags; > > unsigned long new_pflags = PF_MEMALLOC_NOFS; > > > > + /* > > + * Tranfer lock ownership to workqueue task. > > + */ > > + rwsem_acquire(&args->cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); > > /* > > * we are in a transaction context here, but may also be doing work > > * in kswapd context, and hence we may need to inherit that state > > @@ -2829,6 +2833,7 @@ xfs_btree_split_worker( > > > > args->result = __xfs_btree_split(args->cur, args->level, args->ptrp, > > args->key, args->curp, args->stat); > > + rwsem_release(&args->cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); > > complete(args->done); > > > > current_restore_flags_nested(&pflags, new_pflags); > > @@ -2863,8 +2868,19 @@ xfs_btree_split( > > args.done = &done; > > args.kswapd = current_is_kswapd(); > > INIT_WORK_ONSTACK(&args.work, xfs_btree_split_worker); > > + /* > > + * Update lockdep's ownership information to reflect transfer of the > > + * ilock from the current task to the worker. Otherwise assertions that > > + * the lock is held (such as when logging the inode) might fail due to > > + * incorrect task owner state. > > + */ > > + rwsem_release(&cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); > > queue_work(xfs_alloc_wq, &args.work); > > wait_for_completion(&done); > > + /* > > + * Tranfer lock ownership back to the thread. > > + */ > > + rwsem_acquire(&cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); > > So I ran all this through fstests. On generic/324 on a fstests run with > reasonable recent xfsprogs and all the features turned on (rmap in > particular) I see the following lockdep report: Why, exactly do we need to transfer the ownership of the inode lock here? The code that runs in the workqueue does not attempt to lock the inode in any way, nor should it be trying to assert that the inode is locked because it's just doing btree work that -requires- the inode to be locked long before we get to these layers. If anything deep in the BMBT split code is trying to assert that the inode is locked, just remove those assertions as we've already run these asserts long before we hand this work off to another thread. [....] > Just like in the first scenario, we hold two ILOCKs -- a directory, and > a file that we're removing from the directory. The removal triggers the > same bmbt split worker because we're shrinking the directory, and upon > its completion we call rwsem_acquire to reset the lockdep maps. > > Unfortunately, in this case, we actually /are/ feeding the correct > subclass information to rwsem_acquire. This time it's pointing out what > it thinks is an inconsistency in our locking order: the first time we > locked the directory and then the regular file inode, but now we hold > the regular file inode and we're asking it for the directory ILOCK. > > (Note that we don't actually deadlock here because pid 25035 has > maintained ownership of the directory ILOCK rwsem this whole time, but > lockdep doesn't know that.) > > A crappy way to bypass this problem is the following garbage patch > which disables lockdep chain checking since we never actually dropped > any of the ILOCKs that are being complained about. Messing with low > level lockdep internals seems sketchy to me, but so it goes. > > The patch also has the major flaw that it doesn't recapture the subclass > information, but doing that is left as an exercise to the reader. ;) lockdep is a PITA when it comes to using semaphores as they were intended because lockdep is task context centric and semaphores are data object centric. i.e. with semaphores, the data object owns the lock, not the task context, and that creates a horrible impedence mismatch between the semaphore locking and lockdep tracking models. IOWs, lockdep has major issues with this, so the two options here are: 1. get rid of lockdep checks in the paths where we pass locked semaphores to other task contexts; or 2. add lockdep checks for "lockdep_assert_held_non_owner()" where it just checks that the lock is held read/write but without checking the owner matches. #2 matches how our existing inode locking assert checks work - we don't care who locked the inode, just that it is locked in the correct mode. This would get rid of the need to pass rwsem lockdep contexts between tasks. That, in turn, gets rid of all the nesting and annotation problems that this attmept to pass lockdep contexts to non-owner tasks introduces.... Cheers, Dave.
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 2d25bab68764..181d5797c97b 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -2816,6 +2816,10 @@ xfs_btree_split_worker( unsigned long pflags; unsigned long new_pflags = PF_MEMALLOC_NOFS; + /* + * Tranfer lock ownership to workqueue task. + */ + rwsem_acquire(&args->cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); /* * we are in a transaction context here, but may also be doing work * in kswapd context, and hence we may need to inherit that state @@ -2829,6 +2833,7 @@ xfs_btree_split_worker( args->result = __xfs_btree_split(args->cur, args->level, args->ptrp, args->key, args->curp, args->stat); + rwsem_release(&args->cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); complete(args->done); current_restore_flags_nested(&pflags, new_pflags); @@ -2863,8 +2868,19 @@ xfs_btree_split( args.done = &done; args.kswapd = current_is_kswapd(); INIT_WORK_ONSTACK(&args.work, xfs_btree_split_worker); + /* + * Update lockdep's ownership information to reflect transfer of the + * ilock from the current task to the worker. Otherwise assertions that + * the lock is held (such as when logging the inode) might fail due to + * incorrect task owner state. + */ + rwsem_release(&cur->bc_ino.ip->i_lock.dep_map, _THIS_IP_); queue_work(xfs_alloc_wq, &args.work); wait_for_completion(&done); + /* + * Tranfer lock ownership back to the thread. + */ + rwsem_acquire(&cur->bc_ino.ip->i_lock.dep_map, 0, 0, _RET_IP_); destroy_work_on_stack(&args.work); return args.result; } diff --git a/fs/xfs/mrlock.h b/fs/xfs/mrlock.h deleted file mode 100644 index 79155eec341b..000000000000 --- a/fs/xfs/mrlock.h +++ /dev/null @@ -1,78 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -/* - * Copyright (c) 2000-2006 Silicon Graphics, Inc. - * All Rights Reserved. - */ -#ifndef __XFS_SUPPORT_MRLOCK_H__ -#define __XFS_SUPPORT_MRLOCK_H__ - -#include <linux/rwsem.h> - -typedef struct { - struct rw_semaphore mr_lock; -#if defined(DEBUG) || defined(XFS_WARN) - int mr_writer; -#endif -} mrlock_t; - -#if defined(DEBUG) || defined(XFS_WARN) -#define mrinit(mrp, name) \ - do { (mrp)->mr_writer = 0; init_rwsem(&(mrp)->mr_lock); } while (0) -#else -#define mrinit(mrp, name) \ - do { init_rwsem(&(mrp)->mr_lock); } while (0) -#endif - -#define mrlock_init(mrp, t,n,s) mrinit(mrp, n) -#define mrfree(mrp) do { } while (0) - -static inline void mraccess_nested(mrlock_t *mrp, int subclass) -{ - down_read_nested(&mrp->mr_lock, subclass); -} - -static inline void mrupdate_nested(mrlock_t *mrp, int subclass) -{ - down_write_nested(&mrp->mr_lock, subclass); -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 1; -#endif -} - -static inline int mrtryaccess(mrlock_t *mrp) -{ - return down_read_trylock(&mrp->mr_lock); -} - -static inline int mrtryupdate(mrlock_t *mrp) -{ - if (!down_write_trylock(&mrp->mr_lock)) - return 0; -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 1; -#endif - return 1; -} - -static inline void mrunlock_excl(mrlock_t *mrp) -{ -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 0; -#endif - up_write(&mrp->mr_lock); -} - -static inline void mrunlock_shared(mrlock_t *mrp) -{ - up_read(&mrp->mr_lock); -} - -static inline void mrdemote(mrlock_t *mrp) -{ -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 0; -#endif - downgrade_write(&mrp->mr_lock); -} - -#endif /* __XFS_SUPPORT_MRLOCK_H__ */ diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 16d481cf3793..43ecfcb63f99 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -191,14 +191,15 @@ xfs_ilock( } if (lock_flags & XFS_MMAPLOCK_EXCL) - mrupdate_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); + down_write_nested(&ip->i_mmaplock, + XFS_MMAPLOCK_DEP(lock_flags)); else if (lock_flags & XFS_MMAPLOCK_SHARED) - mraccess_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); + down_read_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); if (lock_flags & XFS_ILOCK_EXCL) - mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); + down_write_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); else if (lock_flags & XFS_ILOCK_SHARED) - mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); + down_read_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); } /* @@ -242,27 +243,27 @@ xfs_ilock_nowait( } if (lock_flags & XFS_MMAPLOCK_EXCL) { - if (!mrtryupdate(&ip->i_mmaplock)) + if (!down_write_trylock(&ip->i_mmaplock)) goto out_undo_iolock; } else if (lock_flags & XFS_MMAPLOCK_SHARED) { - if (!mrtryaccess(&ip->i_mmaplock)) + if (!down_read_trylock(&ip->i_mmaplock)) goto out_undo_iolock; } if (lock_flags & XFS_ILOCK_EXCL) { - if (!mrtryupdate(&ip->i_lock)) + if (!down_write_trylock(&ip->i_lock)) goto out_undo_mmaplock; } else if (lock_flags & XFS_ILOCK_SHARED) { - if (!mrtryaccess(&ip->i_lock)) + if (!down_read_trylock(&ip->i_lock)) goto out_undo_mmaplock; } return 1; out_undo_mmaplock: if (lock_flags & XFS_MMAPLOCK_EXCL) - mrunlock_excl(&ip->i_mmaplock); + up_write(&ip->i_mmaplock); else if (lock_flags & XFS_MMAPLOCK_SHARED) - mrunlock_shared(&ip->i_mmaplock); + up_read(&ip->i_mmaplock); out_undo_iolock: if (lock_flags & XFS_IOLOCK_EXCL) up_write(&VFS_I(ip)->i_rwsem); @@ -309,14 +310,14 @@ xfs_iunlock( up_read(&VFS_I(ip)->i_rwsem); if (lock_flags & XFS_MMAPLOCK_EXCL) - mrunlock_excl(&ip->i_mmaplock); + up_write(&ip->i_mmaplock); else if (lock_flags & XFS_MMAPLOCK_SHARED) - mrunlock_shared(&ip->i_mmaplock); + up_read(&ip->i_mmaplock); if (lock_flags & XFS_ILOCK_EXCL) - mrunlock_excl(&ip->i_lock); + up_write(&ip->i_lock); else if (lock_flags & XFS_ILOCK_SHARED) - mrunlock_shared(&ip->i_lock); + up_read(&ip->i_lock); trace_xfs_iunlock(ip, lock_flags, _RET_IP_); } @@ -335,9 +336,9 @@ xfs_ilock_demote( ~(XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL)) == 0); if (lock_flags & XFS_ILOCK_EXCL) - mrdemote(&ip->i_lock); + downgrade_write(&ip->i_lock); if (lock_flags & XFS_MMAPLOCK_EXCL) - mrdemote(&ip->i_mmaplock); + downgrade_write(&ip->i_mmaplock); if (lock_flags & XFS_IOLOCK_EXCL) downgrade_write(&VFS_I(ip)->i_rwsem); @@ -375,19 +376,22 @@ xfs_isilocked( struct xfs_inode *ip, uint lock_flags) { - if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) { - if (!(lock_flags & XFS_ILOCK_SHARED)) - return !!ip->i_lock.mr_writer; - return rwsem_is_locked(&ip->i_lock.mr_lock); + if (lock_flags & (XFS_ILOCK_EXCL | XFS_ILOCK_SHARED)) { + ASSERT(!(lock_flags & ~(XFS_ILOCK_EXCL | XFS_ILOCK_SHARED))); + return __xfs_rwsem_islocked(&ip->i_lock, lock_flags, + XFS_ILOCK_FLAG_SHIFT); } - if (lock_flags & (XFS_MMAPLOCK_EXCL|XFS_MMAPLOCK_SHARED)) { - if (!(lock_flags & XFS_MMAPLOCK_SHARED)) - return !!ip->i_mmaplock.mr_writer; - return rwsem_is_locked(&ip->i_mmaplock.mr_lock); + if (lock_flags & (XFS_MMAPLOCK_EXCL | XFS_MMAPLOCK_SHARED)) { + ASSERT(!(lock_flags & + ~(XFS_MMAPLOCK_EXCL | XFS_MMAPLOCK_SHARED))); + return __xfs_rwsem_islocked(&ip->i_mmaplock, lock_flags, + XFS_MMAPLOCK_FLAG_SHIFT); } if (lock_flags & (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED)) { + ASSERT(!(lock_flags & + ~(XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED))); return __xfs_rwsem_islocked(&VFS_I(ip)->i_rwsem, lock_flags, XFS_IOLOCK_FLAG_SHIFT); } diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 1392a9c452ae..66ceb127192e 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -39,8 +39,8 @@ typedef struct xfs_inode { /* Transaction and locking information. */ struct xfs_inode_log_item *i_itemp; /* logging information */ - mrlock_t i_lock; /* inode lock */ - mrlock_t i_mmaplock; /* inode mmap IO lock */ + struct rw_semaphore i_lock; /* inode lock */ + struct rw_semaphore i_mmaplock; /* inode mmap IO lock */ atomic_t i_pincount; /* inode pin count */ /* diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 5e165456da68..8181f6785a7a 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -1336,9 +1336,9 @@ xfs_setup_inode( */ lockdep_set_class(&inode->i_rwsem, &inode->i_sb->s_type->i_mutex_dir_key); - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class); + lockdep_set_class(&ip->i_lock, &xfs_dir_ilock_class); } else { - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class); + lockdep_set_class(&ip->i_lock, &xfs_nondir_ilock_class); } /* diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h index 5b7a1e201559..64e28ec16cf7 100644 --- a/fs/xfs/xfs_linux.h +++ b/fs/xfs/xfs_linux.h @@ -22,7 +22,6 @@ typedef __u32 xfs_nlink_t; #include "xfs_types.h" #include "kmem.h" -#include "mrlock.h" #include <linux/semaphore.h> #include <linux/mm.h> @@ -61,6 +60,7 @@ typedef __u32 xfs_nlink_t; #include <linux/ratelimit.h> #include <linux/rhashtable.h> #include <linux/xattr.h> +#include <linux/rwsem.h> #include <asm/page.h> #include <asm/div64.h> diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index e3e229e52512..380ba196b165 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -708,10 +708,8 @@ xfs_fs_inode_init_once( atomic_set(&ip->i_pincount, 0); spin_lock_init(&ip->i_flags_lock); - mrlock_init(&ip->i_mmaplock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER, - "xfsino", ip->i_ino); - mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER, - "xfsino", ip->i_ino); + init_rwsem(&ip->i_mmaplock); + init_rwsem(&ip->i_lock); } /*