From 0c38a2512df272b14ef4238b476a2e4f70da1479 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 25 Aug 2011 07:17:01 +0000 Subject: [PATCH 01/69] xfs: don't serialise direct IO reads on page cache checks There is no need to grab the i_mutex of the IO lock in exclusive mode if we don't need to invalidate the page cache. Taking these locks on every direct IO effective serialises them as taking the IO lock in exclusive mode has to wait for all shared holders to drop the lock. That only happens when IO is complete, so effective it prevents dispatch of concurrent direct IO reads to the same inode. Fix this by taking the IO lock shared to check the page cache state, and only then drop it and take the IO lock exclusively if there is work to be done. Hence for the normal direct IO case, no exclusive locking will occur. Signed-off-by: Dave Chinner Tested-by: Joern Engel Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_file.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 7f7b42469ea7..8fd4a0708d30 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -317,7 +317,19 @@ xfs_file_aio_read( if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; - if (unlikely(ioflags & IO_ISDIRECT)) { + /* + * Locking is a bit tricky here. If we take an exclusive lock + * for direct IO, we effectively serialise all new concurrent + * read IO to this file and block it behind IO that is currently in + * progress because IO in progress holds the IO lock shared. We only + * need to hold the lock exclusive to blow away the page cache, so + * only take lock exclusively if the page cache needs invalidation. + * This allows the normal direct IO case of no page cache pages to + * proceeed concurrently without serialisation. + */ + xfs_rw_ilock(ip, XFS_IOLOCK_SHARED); + if ((ioflags & IO_ISDIRECT) && inode->i_mapping->nrpages) { + xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED); xfs_rw_ilock(ip, XFS_IOLOCK_EXCL); if (inode->i_mapping->nrpages) { @@ -330,8 +342,7 @@ xfs_file_aio_read( } } xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL); - } else - xfs_rw_ilock(ip, XFS_IOLOCK_SHARED); + } trace_xfs_file_read(ip, size, iocb->ki_pos, ioflags); From 7271d243f9d1b4106289e4cf876c8b1203de59ab Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 25 Aug 2011 07:17:02 +0000 Subject: [PATCH 02/69] xfs: don't serialise adjacent concurrent direct IO appending writes For append write workloads, extending the file requires a certain amount of exclusive locking to be done up front to ensure sanity in things like ensuring that we've zeroed any allocated regions between the old EOF and the start of the new IO. For single threads, this typically isn't a problem, and for large IOs we don't serialise enough for it to be a problem for two threads on really fast block devices. However for smaller IO and larger thread counts we have a problem. Take 4 concurrent sequential, single block sized and aligned IOs. After the first IO is submitted but before it completes, we end up with this state: IO 1 IO 2 IO 3 IO 4 +-------+-------+-------+-------+ ^ ^ | | | | | | | \- ip->i_new_size \- ip->i_size And the IO is done without exclusive locking because offset <= ip->i_size. When we submit IO 2, we see offset > ip->i_size, and grab the IO lock exclusive, because there is a chance we need to do EOF zeroing. However, there is already an IO in progress that avoids the need for IO zeroing because offset <= ip->i_new_size. hence we could avoid holding the IO lock exlcusive for this. Hence after submission of the second IO, we'd end up this state: IO 1 IO 2 IO 3 IO 4 +-------+-------+-------+-------+ ^ ^ | | | | | | | \- ip->i_new_size \- ip->i_size There is no need to grab the i_mutex of the IO lock in exclusive mode if we don't need to invalidate the page cache. Taking these locks on every direct IO effective serialises them as taking the IO lock in exclusive mode has to wait for all shared holders to drop the lock. That only happens when IO is complete, so effective it prevents dispatch of concurrent direct IO writes to the same inode. And so you can see that for the third concurrent IO, we'd avoid exclusive locking for the same reason we avoided the exclusive lock for the second IO. Fixing this is a bit more complex than that, because we need to hold a write-submission local value of ip->i_new_size to that clearing the value is only done if no other thread has updated it before our IO completes..... Signed-off-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_file.c | 68 ++++++++++++++++++++++++++++++++++++----------- 1 file changed, 52 insertions(+), 16 deletions(-) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 8fd4a0708d30..cbbac5cc9c26 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -418,11 +418,13 @@ xfs_aio_write_isize_update( */ STATIC void xfs_aio_write_newsize_update( - struct xfs_inode *ip) + struct xfs_inode *ip, + xfs_fsize_t new_size) { - if (ip->i_new_size) { + if (new_size == ip->i_new_size) { xfs_rw_ilock(ip, XFS_ILOCK_EXCL); - ip->i_new_size = 0; + if (new_size == ip->i_new_size) + ip->i_new_size = 0; if (ip->i_d.di_size > ip->i_size) ip->i_d.di_size = ip->i_size; xfs_rw_iunlock(ip, XFS_ILOCK_EXCL); @@ -473,7 +475,7 @@ xfs_file_splice_write( ret = generic_file_splice_write(pipe, outfilp, ppos, count, flags); xfs_aio_write_isize_update(inode, ppos, ret); - xfs_aio_write_newsize_update(ip); + xfs_aio_write_newsize_update(ip, new_size); xfs_iunlock(ip, XFS_IOLOCK_EXCL); return ret; } @@ -670,6 +672,7 @@ xfs_file_aio_write_checks( struct file *file, loff_t *pos, size_t *count, + xfs_fsize_t *new_sizep, int *iolock) { struct inode *inode = file->f_mapping->host; @@ -677,6 +680,8 @@ xfs_file_aio_write_checks( xfs_fsize_t new_size; int error = 0; + *new_sizep = 0; +restart: error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode)); if (error) { xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock); @@ -684,20 +689,41 @@ xfs_file_aio_write_checks( return error; } - new_size = *pos + *count; - if (new_size > ip->i_size) - ip->i_new_size = new_size; - if (likely(!(file->f_mode & FMODE_NOCMTIME))) file_update_time(file); /* * If the offset is beyond the size of the file, we need to zero any * blocks that fall between the existing EOF and the start of this - * write. + * write. There is no need to issue zeroing if another in-flght IO ends + * at or before this one If zeronig is needed and we are currently + * holding the iolock shared, we need to update it to exclusive which + * involves dropping all locks and relocking to maintain correct locking + * order. If we do this, restart the function to ensure all checks and + * values are still valid. */ - if (*pos > ip->i_size) + if ((ip->i_new_size && *pos > ip->i_new_size) || + (!ip->i_new_size && *pos > ip->i_size)) { + if (*iolock == XFS_IOLOCK_SHARED) { + xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock); + *iolock = XFS_IOLOCK_EXCL; + xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock); + goto restart; + } error = -xfs_zero_eof(ip, *pos, ip->i_size); + } + + /* + * If this IO extends beyond EOF, we may need to update ip->i_new_size. + * We have already zeroed space beyond EOF (if necessary). Only update + * ip->i_new_size if this IO ends beyond any other in-flight writes. + */ + new_size = *pos + *count; + if (new_size > ip->i_size) { + if (new_size > ip->i_new_size) + ip->i_new_size = new_size; + *new_sizep = new_size; + } xfs_rw_iunlock(ip, XFS_ILOCK_EXCL); if (error) @@ -744,6 +770,7 @@ xfs_file_dio_aio_write( unsigned long nr_segs, loff_t pos, size_t ocount, + xfs_fsize_t *new_size, int *iolock) { struct file *file = iocb->ki_filp; @@ -764,13 +791,20 @@ xfs_file_dio_aio_write( if ((pos & mp->m_blockmask) || ((pos + count) & mp->m_blockmask)) unaligned_io = 1; - if (unaligned_io || mapping->nrpages || pos > ip->i_size) + /* + * We don't need to take an exclusive lock unless there page cache needs + * to be invalidated or unaligned IO is being executed. We don't need to + * consider the EOF extension case here because + * xfs_file_aio_write_checks() will relock the inode as necessary for + * EOF zeroing cases and fill out the new inode size as appropriate. + */ + if (unaligned_io || mapping->nrpages) *iolock = XFS_IOLOCK_EXCL; else *iolock = XFS_IOLOCK_SHARED; xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock); - ret = xfs_file_aio_write_checks(file, &pos, &count, iolock); + ret = xfs_file_aio_write_checks(file, &pos, &count, new_size, iolock); if (ret) return ret; @@ -809,6 +843,7 @@ xfs_file_buffered_aio_write( unsigned long nr_segs, loff_t pos, size_t ocount, + xfs_fsize_t *new_size, int *iolock) { struct file *file = iocb->ki_filp; @@ -822,7 +857,7 @@ xfs_file_buffered_aio_write( *iolock = XFS_IOLOCK_EXCL; xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock); - ret = xfs_file_aio_write_checks(file, &pos, &count, iolock); + ret = xfs_file_aio_write_checks(file, &pos, &count, new_size, iolock); if (ret) return ret; @@ -862,6 +897,7 @@ xfs_file_aio_write( ssize_t ret; int iolock; size_t ocount = 0; + xfs_fsize_t new_size = 0; XFS_STATS_INC(xs_write_calls); @@ -881,10 +917,10 @@ xfs_file_aio_write( if (unlikely(file->f_flags & O_DIRECT)) ret = xfs_file_dio_aio_write(iocb, iovp, nr_segs, pos, - ocount, &iolock); + ocount, &new_size, &iolock); else ret = xfs_file_buffered_aio_write(iocb, iovp, nr_segs, pos, - ocount, &iolock); + ocount, &new_size, &iolock); xfs_aio_write_isize_update(inode, &iocb->ki_pos, ret); @@ -905,7 +941,7 @@ xfs_file_aio_write( } out_unlock: - xfs_aio_write_newsize_update(ip); + xfs_aio_write_newsize_update(ip, new_size); xfs_rw_iunlock(ip, iolock); return ret; } From 375ec69d2ef6e0797f19f5823e36e249765c3d41 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:03 +0000 Subject: [PATCH 03/69] xfs: remove delwri buffer handling from xfs_buf_iorequest We cannot ever reach xfs_buf_iorequest for a buffer with XBF_DELWRI set, given that all write handlers make sure that the buffer is remove from the delwri queue before, and we never do reads with the XBF_DELWRI flag set (which the code would not handle correctly anyway). Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index c57836dc778f..2e71a26da22e 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1275,15 +1275,10 @@ xfs_buf_iorequest( { trace_xfs_buf_iorequest(bp, _RET_IP_); - if (bp->b_flags & XBF_DELWRI) { - xfs_buf_delwri_queue(bp, 1); - return 0; - } + ASSERT(!(bp->b_flags & XBF_DELWRI)); - if (bp->b_flags & XBF_WRITE) { + if (bp->b_flags & XBF_WRITE) xfs_buf_wait_unpin(bp); - } - xfs_buf_hold(bp); /* Set the count to 1 initially, this will stop an I/O From 527cfdf19dd538a5a9e46b9bed0f30a38c28438d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:04 +0000 Subject: [PATCH 04/69] xfs: remove the unlock argument to xfs_buf_delwri_queue We can just unlock the buffer in the caller, and the decrement of b_hold would also be needed in the !unlock, we just never hit that case currently given that the caller handles that case. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 2e71a26da22e..04689dbbcbba 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -43,7 +43,7 @@ static kmem_zone_t *xfs_buf_zone; STATIC int xfsbufd(void *); -STATIC void xfs_buf_delwri_queue(xfs_buf_t *, int); +STATIC void xfs_buf_delwri_queue(xfs_buf_t *); static struct workqueue_struct *xfslogd_workqueue; struct workqueue_struct *xfsdatad_workqueue; @@ -940,7 +940,7 @@ xfs_buf_unlock( if ((bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)) == XBF_DELWRI) { atomic_inc(&bp->b_hold); bp->b_flags |= XBF_ASYNC; - xfs_buf_delwri_queue(bp, 0); + xfs_buf_delwri_queue(bp); } XB_CLEAR_OWNER(bp); @@ -1049,7 +1049,8 @@ xfs_bdwrite( bp->b_flags &= ~XBF_READ; bp->b_flags |= (XBF_DELWRI | XBF_ASYNC); - xfs_buf_delwri_queue(bp, 1); + xfs_buf_delwri_queue(bp); + xfs_buf_unlock(bp); } /* @@ -1562,8 +1563,7 @@ xfs_alloc_buftarg( */ STATIC void xfs_buf_delwri_queue( - xfs_buf_t *bp, - int unlock) + xfs_buf_t *bp) { struct list_head *dwq = &bp->b_target->bt_delwrite_queue; spinlock_t *dwlk = &bp->b_target->bt_delwrite_lock; @@ -1576,8 +1576,7 @@ xfs_buf_delwri_queue( /* If already in the queue, dequeue and place at tail */ if (!list_empty(&bp->b_list)) { ASSERT(bp->b_flags & _XBF_DELWRI_Q); - if (unlock) - atomic_dec(&bp->b_hold); + atomic_dec(&bp->b_hold); list_del(&bp->b_list); } @@ -1590,9 +1589,6 @@ xfs_buf_delwri_queue( list_add_tail(&bp->b_list, dwq); bp->b_queuetime = jiffies; spin_unlock(dwlk); - - if (unlock) - xfs_buf_unlock(bp); } void From 5a8ee6bafdd0ab8555adceac8b2cec539a552a1f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:05 +0000 Subject: [PATCH 05/69] xfs: move more delwri setup into xfs_buf_delwri_queue Do not transfer a reference held by the caller to the buffer on the list, or decrement it in xfs_buf_delwri_queue, but instead grab a new reference if needed, and let the caller drop its own reference. Also move setting of the XBF_DELWRI and XBF_ASYNC flags into xfs_buf_delwri_queue, and only do it if needed. Note that for now xfs_buf_unlock already has XBF_DELWRI, but that will change in the following patches. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 31 ++++++++++++------------------- 1 file changed, 12 insertions(+), 19 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 04689dbbcbba..86c0945053c9 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -937,11 +937,8 @@ void xfs_buf_unlock( struct xfs_buf *bp) { - if ((bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)) == XBF_DELWRI) { - atomic_inc(&bp->b_hold); - bp->b_flags |= XBF_ASYNC; + if ((bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)) == XBF_DELWRI) xfs_buf_delwri_queue(bp); - } XB_CLEAR_OWNER(bp); up(&bp->b_sema); @@ -1046,11 +1043,8 @@ xfs_bdwrite( { trace_xfs_buf_bdwrite(bp, _RET_IP_); - bp->b_flags &= ~XBF_READ; - bp->b_flags |= (XBF_DELWRI | XBF_ASYNC); - xfs_buf_delwri_queue(bp); - xfs_buf_unlock(bp); + xfs_buf_relse(bp); } /* @@ -1570,23 +1564,22 @@ xfs_buf_delwri_queue( trace_xfs_buf_delwri_queue(bp, _RET_IP_); - ASSERT((bp->b_flags&(XBF_DELWRI|XBF_ASYNC)) == (XBF_DELWRI|XBF_ASYNC)); + ASSERT(!(bp->b_flags & XBF_READ)); spin_lock(dwlk); - /* If already in the queue, dequeue and place at tail */ if (!list_empty(&bp->b_list)) { + /* if already in the queue, move it to the tail */ ASSERT(bp->b_flags & _XBF_DELWRI_Q); - atomic_dec(&bp->b_hold); - list_del(&bp->b_list); - } - - if (list_empty(dwq)) { + list_move_tail(&bp->b_list, dwq); + } else { /* start xfsbufd as it is about to have something to do */ - wake_up_process(bp->b_target->bt_task); - } + if (list_empty(dwq)) + wake_up_process(bp->b_target->bt_task); - bp->b_flags |= _XBF_DELWRI_Q; - list_add_tail(&bp->b_list, dwq); + atomic_inc(&bp->b_hold); + bp->b_flags |= XBF_DELWRI | _XBF_DELWRI_Q | XBF_ASYNC; + list_add_tail(&bp->b_list, dwq); + } bp->b_queuetime = jiffies; spin_unlock(dwlk); } From 61551f1ee536289084a4a8f1c4f187e2f371c440 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:06 +0000 Subject: [PATCH 06/69] xfs: call xfs_buf_delwri_queue directly Unify the ways we add buffers to the delwri queue by always calling xfs_buf_delwri_queue directly. The xfs_bdwrite functions is removed and opencoded in its callers, and the two places setting XBF_DELWRI while a buffer is locked and expecting xfs_buf_unlock to pick it up are converted to call xfs_buf_delwri_queue directly, too. Also replace the XFS_BUF_UNDELAYWRITE macro with direct calls to xfs_buf_delwri_dequeue to make the explicit queuing/dequeuing more obvious. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 2 +- fs/xfs/xfs_buf.c | 21 +++------------------ fs/xfs/xfs_buf.h | 8 +++----- fs/xfs/xfs_buf_item.c | 4 ++-- fs/xfs/xfs_dquot.c | 6 ++++-- fs/xfs/xfs_inode.c | 6 ++++-- fs/xfs/xfs_log_recover.c | 9 ++++++--- fs/xfs/xfs_mount.c | 2 +- fs/xfs/xfs_qm.c | 3 ++- fs/xfs/xfs_rw.c | 2 +- fs/xfs/xfs_trace.h | 1 - fs/xfs/xfs_trans_buf.c | 5 +++-- 12 files changed, 30 insertions(+), 39 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 160bcdc34a6e..b097fd5a2b3f 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2189,7 +2189,7 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt, XBF_TRYLOCK); if (bp) { XFS_BUF_STALE(bp); - XFS_BUF_UNDELAYWRITE(bp); + xfs_buf_delwri_dequeue(bp); xfs_buf_relse(bp); bp = NULL; } diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 86c0945053c9..309eca75fad4 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -43,7 +43,6 @@ static kmem_zone_t *xfs_buf_zone; STATIC int xfsbufd(void *); -STATIC void xfs_buf_delwri_queue(xfs_buf_t *); static struct workqueue_struct *xfslogd_workqueue; struct workqueue_struct *xfsdatad_workqueue; @@ -937,9 +936,6 @@ void xfs_buf_unlock( struct xfs_buf *bp) { - if ((bp->b_flags & (XBF_DELWRI|_XBF_DELWRI_Q)) == XBF_DELWRI) - xfs_buf_delwri_queue(bp); - XB_CLEAR_OWNER(bp); up(&bp->b_sema); @@ -1036,17 +1032,6 @@ xfs_bwrite( return error; } -void -xfs_bdwrite( - void *mp, - struct xfs_buf *bp) -{ - trace_xfs_buf_bdwrite(bp, _RET_IP_); - - xfs_buf_delwri_queue(bp); - xfs_buf_relse(bp); -} - /* * Called when we want to stop a buffer from getting written or read. * We attach the EIO error, muck with its flags, and call xfs_buf_ioend @@ -1069,7 +1054,7 @@ xfs_bioerror( * We're calling xfs_buf_ioend, so delete XBF_DONE flag. */ XFS_BUF_UNREAD(bp); - XFS_BUF_UNDELAYWRITE(bp); + xfs_buf_delwri_dequeue(bp); XFS_BUF_UNDONE(bp); XFS_BUF_STALE(bp); @@ -1098,7 +1083,7 @@ xfs_bioerror_relse( * change that interface. */ XFS_BUF_UNREAD(bp); - XFS_BUF_UNDELAYWRITE(bp); + xfs_buf_delwri_dequeue(bp); XFS_BUF_DONE(bp); XFS_BUF_STALE(bp); bp->b_iodone = NULL; @@ -1555,7 +1540,7 @@ xfs_alloc_buftarg( /* * Delayed write buffer handling */ -STATIC void +void xfs_buf_delwri_queue( xfs_buf_t *bp) { diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 620972b8094d..f1a8933becb6 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -198,7 +198,6 @@ extern void xfs_buf_unlock(xfs_buf_t *); /* Buffer Read and Write Routines */ extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp); -extern void xfs_bdwrite(void *mp, xfs_buf_t *bp); extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); extern int xfs_bdstrat_cb(struct xfs_buf *); @@ -221,8 +220,9 @@ static inline int xfs_buf_geterror(xfs_buf_t *bp) extern xfs_caddr_t xfs_buf_offset(xfs_buf_t *, size_t); /* Delayed Write Buffer Routines */ -extern void xfs_buf_delwri_dequeue(xfs_buf_t *); -extern void xfs_buf_delwri_promote(xfs_buf_t *); +extern void xfs_buf_delwri_queue(struct xfs_buf *); +extern void xfs_buf_delwri_dequeue(struct xfs_buf *); +extern void xfs_buf_delwri_promote(struct xfs_buf *); /* Buffer Daemon Setup Routines */ extern int xfs_buf_init(void); @@ -251,8 +251,6 @@ void xfs_buf_stale(struct xfs_buf *bp); XFS_BUF_DONE(bp); \ } while (0) -#define XFS_BUF_DELAYWRITE(bp) ((bp)->b_flags |= XBF_DELWRI) -#define XFS_BUF_UNDELAYWRITE(bp) xfs_buf_delwri_dequeue(bp) #define XFS_BUF_ISDELAYWRITE(bp) ((bp)->b_flags & XBF_DELWRI) #define XFS_BUF_DONE(bp) ((bp)->b_flags |= XBF_DONE) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index cac2ecfa6746..3243083d8693 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -992,7 +992,7 @@ xfs_buf_iodone_callbacks( xfs_buf_ioerror(bp, 0); /* errno of 0 unsets the flag */ if (!XFS_BUF_ISSTALE(bp)) { - XFS_BUF_DELAYWRITE(bp); + xfs_buf_delwri_queue(bp); XFS_BUF_DONE(bp); } ASSERT(bp->b_iodone != NULL); @@ -1007,7 +1007,7 @@ xfs_buf_iodone_callbacks( */ XFS_BUF_STALE(bp); XFS_BUF_DONE(bp); - XFS_BUF_UNDELAYWRITE(bp); + xfs_buf_delwri_dequeue(bp); trace_xfs_buf_error_relse(bp, _RET_IP_); diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index db62959bed13..0f78dd46415c 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -1243,8 +1243,10 @@ xfs_qm_dqflush( if (flags & SYNC_WAIT) error = xfs_bwrite(mp, bp); - else - xfs_bdwrite(mp, bp); + else { + xfs_buf_delwri_queue(bp); + xfs_buf_relse(bp); + } trace_xfs_dqflush_done(dqp); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 0239a7c7c886..f8fe1c66d420 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2598,8 +2598,10 @@ xfs_iflush( if (flags & SYNC_WAIT) error = xfs_bwrite(mp, bp); - else - xfs_bdwrite(mp, bp); + else { + xfs_buf_delwri_queue(bp); + xfs_buf_relse(bp); + } return error; corrupt_out: diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index a199dbcee7d8..22946949bf5e 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2176,7 +2176,8 @@ xlog_recover_buffer_pass2( } else { ASSERT(bp->b_target->bt_mount == mp); bp->b_iodone = xlog_recover_iodone; - xfs_bdwrite(mp, bp); + xfs_buf_delwri_queue(bp); + xfs_buf_relse(bp); } return (error); @@ -2439,7 +2440,8 @@ xlog_recover_inode_pass2( write_inode_buffer: ASSERT(bp->b_target->bt_mount == mp); bp->b_iodone = xlog_recover_iodone; - xfs_bdwrite(mp, bp); + xfs_buf_delwri_queue(bp); + xfs_buf_relse(bp); error: if (need_free) kmem_free(in_f); @@ -2561,7 +2563,8 @@ xlog_recover_dquot_pass2( ASSERT(dq_f->qlf_size == 2); ASSERT(bp->b_target->bt_mount == mp); bp->b_iodone = xlog_recover_iodone; - xfs_bdwrite(mp, bp); + xfs_buf_delwri_queue(bp); + xfs_buf_relse(bp); return (0); } diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 0081657ad985..a957e437bee1 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1612,7 +1612,7 @@ xfs_unmountfs_writesb(xfs_mount_t *mp) XFS_BUF_UNDONE(sbp); XFS_BUF_UNREAD(sbp); - XFS_BUF_UNDELAYWRITE(sbp); + xfs_buf_delwri_dequeue(sbp); XFS_BUF_WRITE(sbp); XFS_BUF_UNASYNC(sbp); ASSERT(sbp->b_target == mp->m_ddev_targp); diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 9a0aa76facdf..f51bef885e6d 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1296,7 +1296,8 @@ xfs_qm_dqiter_bufs( break; xfs_qm_reset_dqcounts(mp, bp, firstid, type); - xfs_bdwrite(mp, bp); + xfs_buf_delwri_queue(bp); + xfs_buf_relse(bp); /* * goto the next block. */ diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index c96a8a05ac03..99823c3b9aca 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -149,7 +149,7 @@ xfs_read_buf( } if (bp) { XFS_BUF_UNDONE(bp); - XFS_BUF_UNDELAYWRITE(bp); + xfs_buf_delwri_dequeue(bp); XFS_BUF_STALE(bp); /* * brelse clears B_ERROR and b_error diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 690fc7a7bd72..bb5e660e0fab 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -320,7 +320,6 @@ DEFINE_BUF_EVENT(xfs_buf_rele); DEFINE_BUF_EVENT(xfs_buf_iodone); DEFINE_BUF_EVENT(xfs_buf_iorequest); DEFINE_BUF_EVENT(xfs_buf_bawrite); -DEFINE_BUF_EVENT(xfs_buf_bdwrite); DEFINE_BUF_EVENT(xfs_buf_lock); DEFINE_BUF_EVENT(xfs_buf_lock_done); DEFINE_BUF_EVENT(xfs_buf_trylock); diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 137e2b9e2948..5e5196a269dd 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -643,13 +643,14 @@ xfs_trans_log_buf(xfs_trans_t *tp, * inside the b_bdstrat callback so that this won't get written to * disk. */ - XFS_BUF_DELAYWRITE(bp); XFS_BUF_DONE(bp); ASSERT(atomic_read(&bip->bli_refcount) > 0); bp->b_iodone = xfs_buf_iodone_callbacks; bip->bli_item.li_cb = xfs_buf_iodone; + xfs_buf_delwri_queue(bp); + trace_xfs_trans_log_buf(bip); /* @@ -738,7 +739,7 @@ xfs_trans_binval( * We set the stale bit in the buffer as well since we're getting * rid of it. */ - XFS_BUF_UNDELAYWRITE(bp); + xfs_buf_delwri_dequeue(bp); XFS_BUF_STALE(bp); bip->bli_flags |= XFS_BLI_STALE; bip->bli_flags &= ~(XFS_BLI_INODE_BUF | XFS_BLI_LOGGED | XFS_BLI_DIRTY); From c2b006c1da1602551def200e4661535f02b82488 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:07 +0000 Subject: [PATCH 07/69] xfs: let xfs_bwrite callers handle the xfs_buf_relse Remove the xfs_buf_relse from xfs_bwrite and let the caller handle it to mirror the delwri and read paths. Also remove the mount pointer passed to xfs_bwrite, which is superflous now that we have a mount pointer in the buftarg. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 7 ++++--- fs/xfs/xfs_buf.c | 8 ++++---- fs/xfs/xfs_buf.h | 2 +- fs/xfs/xfs_dquot.c | 8 ++++---- fs/xfs/xfs_fsops.c | 40 ++++++++++++++++++++++------------------ fs/xfs/xfs_inode.c | 8 ++++---- fs/xfs/xfs_log_recover.c | 11 +++++++---- fs/xfs/xfs_sync.c | 6 ++++-- 8 files changed, 50 insertions(+), 40 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index b097fd5a2b3f..8f0f65833500 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2128,9 +2128,10 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) xfs_buf_iomove(bp, 0, tmp, src, XBRW_WRITE); if (tmp < XFS_BUF_SIZE(bp)) xfs_buf_zero(bp, tmp, XFS_BUF_SIZE(bp) - tmp); - if ((error = xfs_bwrite(mp, bp))) {/* GROT: NOTE: synchronous write */ - return (error); - } + error = xfs_bwrite(bp); /* GROT: NOTE: synchronous write */ + xfs_buf_relse(bp); + if (error) + return error; src += tmp; valuelen -= tmp; diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 309eca75fad4..63dbeb9efc49 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1014,7 +1014,6 @@ xfs_buf_ioerror( int xfs_bwrite( - struct xfs_mount *mp, struct xfs_buf *bp) { int error; @@ -1026,9 +1025,10 @@ xfs_bwrite( xfs_bdstrat_cb(bp); error = xfs_buf_iowait(bp); - if (error) - xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR); - xfs_buf_relse(bp); + if (error) { + xfs_force_shutdown(bp->b_target->bt_mount, + SHUTDOWN_META_IO_ERROR); + } return error; } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index f1a8933becb6..3f543ed3009b 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -197,7 +197,7 @@ extern void xfs_buf_unlock(xfs_buf_t *); ((bp)->b_sema.count <= 0) /* Buffer Read and Write Routines */ -extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp); +extern int xfs_bwrite(struct xfs_buf *bp); extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); extern int xfs_bdstrat_cb(struct xfs_buf *); diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 0f78dd46415c..3e2ccaedc51e 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -1242,11 +1242,11 @@ xfs_qm_dqflush( } if (flags & SYNC_WAIT) - error = xfs_bwrite(mp, bp); - else { + error = xfs_bwrite(bp); + else xfs_buf_delwri_queue(bp); - xfs_buf_relse(bp); - } + + xfs_buf_relse(bp); trace_xfs_dqflush_done(dqp); diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 9153d2c77caf..e023f940a3dd 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -216,10 +216,11 @@ xfs_growfs_data_private( tmpsize = agsize - XFS_PREALLOC_BLOCKS(mp); agf->agf_freeblks = cpu_to_be32(tmpsize); agf->agf_longest = cpu_to_be32(tmpsize); - error = xfs_bwrite(mp, bp); - if (error) { + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) goto error0; - } + /* * AG inode header block */ @@ -240,10 +241,11 @@ xfs_growfs_data_private( agi->agi_dirino = cpu_to_be32(NULLAGINO); for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++) agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO); - error = xfs_bwrite(mp, bp); - if (error) { + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) goto error0; - } + /* * BNO btree root block */ @@ -262,10 +264,11 @@ xfs_growfs_data_private( arec->ar_startblock = cpu_to_be32(XFS_PREALLOC_BLOCKS(mp)); arec->ar_blockcount = cpu_to_be32( agsize - be32_to_cpu(arec->ar_startblock)); - error = xfs_bwrite(mp, bp); - if (error) { + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) goto error0; - } + /* * CNT btree root block */ @@ -285,10 +288,11 @@ xfs_growfs_data_private( arec->ar_blockcount = cpu_to_be32( agsize - be32_to_cpu(arec->ar_startblock)); nfree += be32_to_cpu(arec->ar_blockcount); - error = xfs_bwrite(mp, bp); - if (error) { + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) goto error0; - } + /* * INO btree root block */ @@ -303,10 +307,10 @@ xfs_growfs_data_private( block->bb_numrecs = 0; block->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK); block->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK); - error = xfs_bwrite(mp, bp); - if (error) { + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) goto error0; - } } xfs_trans_agblocks_delta(tp, nfree); /* @@ -396,9 +400,9 @@ xfs_growfs_data_private( * just issue a warning and continue. The real work is * already done and committed. */ - if (!(error = xfs_bwrite(mp, bp))) { - continue; - } else { + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) { xfs_warn(mp, "write error %d updating secondary superblock for ag %d", error, agno); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index f8fe1c66d420..7f237ba3c292 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2597,11 +2597,11 @@ xfs_iflush( goto cluster_corrupt_out; if (flags & SYNC_WAIT) - error = xfs_bwrite(mp, bp); - else { + error = xfs_bwrite(bp); + else xfs_buf_delwri_queue(bp); - xfs_buf_relse(bp); - } + + xfs_buf_relse(bp); return error; corrupt_out: diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 22946949bf5e..be173852b2ca 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -268,9 +268,12 @@ xlog_bwrite( xfs_buf_lock(bp); XFS_BUF_SET_COUNT(bp, BBTOB(nbblks)); - if ((error = xfs_bwrite(log->l_mp, bp))) + error = xfs_bwrite(bp); + if (error) { xfs_ioerror_alert("xlog_bwrite", log->l_mp, bp, XFS_BUF_ADDR(bp)); + } + xfs_buf_relse(bp); return error; } @@ -2172,15 +2175,15 @@ xlog_recover_buffer_pass2( (XFS_BUF_COUNT(bp) != MAX(log->l_mp->m_sb.sb_blocksize, (__uint32_t)XFS_INODE_CLUSTER_SIZE(log->l_mp)))) { XFS_BUF_STALE(bp); - error = xfs_bwrite(mp, bp); + error = xfs_bwrite(bp); } else { ASSERT(bp->b_target->bt_mount == mp); bp->b_iodone = xlog_recover_iodone; xfs_buf_delwri_queue(bp); - xfs_buf_relse(bp); } - return (error); + xfs_buf_relse(bp); + return error; } STATIC int diff --git a/fs/xfs/xfs_sync.c b/fs/xfs/xfs_sync.c index 4604f90f86a3..90cc197e0433 100644 --- a/fs/xfs/xfs_sync.c +++ b/fs/xfs/xfs_sync.c @@ -322,6 +322,7 @@ xfs_sync_fsdata( struct xfs_mount *mp) { struct xfs_buf *bp; + int error; /* * If the buffer is pinned then push on the log so we won't get stuck @@ -334,8 +335,9 @@ xfs_sync_fsdata( bp = xfs_getsb(mp, 0); if (xfs_buf_ispinned(bp)) xfs_log_force(mp, 0); - - return xfs_bwrite(mp, bp); + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + return error; } /* From c4e1c098ee8a72ea563a697a2b175868be86fdc9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:08 +0000 Subject: [PATCH 08/69] xfs: use the "delwri" terminology consistently And also remove the strange local lock and delwri list pointers in a few functions. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 43 +++++++++++++++++++------------------------ fs/xfs/xfs_buf.h | 4 ++-- 2 files changed, 21 insertions(+), 26 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 63dbeb9efc49..d3c2b58d7d70 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1489,12 +1489,12 @@ xfs_setsize_buftarg( } STATIC int -xfs_alloc_delwrite_queue( +xfs_alloc_delwri_queue( xfs_buftarg_t *btp, const char *fsname) { - INIT_LIST_HEAD(&btp->bt_delwrite_queue); - spin_lock_init(&btp->bt_delwrite_lock); + INIT_LIST_HEAD(&btp->bt_delwri_queue); + spin_lock_init(&btp->bt_delwri_lock); btp->bt_flags = 0; btp->bt_task = kthread_run(xfsbufd, btp, "xfsbufd/%s", fsname); if (IS_ERR(btp->bt_task)) @@ -1524,7 +1524,7 @@ xfs_alloc_buftarg( spin_lock_init(&btp->bt_lru_lock); if (xfs_setsize_buftarg_early(btp, bdev)) goto error; - if (xfs_alloc_delwrite_queue(btp, fsname)) + if (xfs_alloc_delwri_queue(btp, fsname)) goto error; btp->bt_shrinker.shrink = xfs_buftarg_shrink; btp->bt_shrinker.seeks = DEFAULT_SEEKS; @@ -1544,46 +1544,44 @@ void xfs_buf_delwri_queue( xfs_buf_t *bp) { - struct list_head *dwq = &bp->b_target->bt_delwrite_queue; - spinlock_t *dwlk = &bp->b_target->bt_delwrite_lock; + struct xfs_buftarg *btp = bp->b_target; trace_xfs_buf_delwri_queue(bp, _RET_IP_); ASSERT(!(bp->b_flags & XBF_READ)); - spin_lock(dwlk); + spin_lock(&btp->bt_delwri_lock); if (!list_empty(&bp->b_list)) { /* if already in the queue, move it to the tail */ ASSERT(bp->b_flags & _XBF_DELWRI_Q); - list_move_tail(&bp->b_list, dwq); + list_move_tail(&bp->b_list, &btp->bt_delwri_queue); } else { /* start xfsbufd as it is about to have something to do */ - if (list_empty(dwq)) + if (list_empty(&btp->bt_delwri_queue)) wake_up_process(bp->b_target->bt_task); atomic_inc(&bp->b_hold); bp->b_flags |= XBF_DELWRI | _XBF_DELWRI_Q | XBF_ASYNC; - list_add_tail(&bp->b_list, dwq); + list_add_tail(&bp->b_list, &btp->bt_delwri_queue); } bp->b_queuetime = jiffies; - spin_unlock(dwlk); + spin_unlock(&btp->bt_delwri_lock); } void xfs_buf_delwri_dequeue( xfs_buf_t *bp) { - spinlock_t *dwlk = &bp->b_target->bt_delwrite_lock; int dequeued = 0; - spin_lock(dwlk); + spin_lock(&bp->b_target->bt_delwri_lock); if ((bp->b_flags & XBF_DELWRI) && !list_empty(&bp->b_list)) { ASSERT(bp->b_flags & _XBF_DELWRI_Q); list_del_init(&bp->b_list); dequeued = 1; } bp->b_flags &= ~(XBF_DELWRI|_XBF_DELWRI_Q); - spin_unlock(dwlk); + spin_unlock(&bp->b_target->bt_delwri_lock); if (dequeued) xfs_buf_rele(bp); @@ -1615,9 +1613,9 @@ xfs_buf_delwri_promote( if (bp->b_queuetime < jiffies - age) return; bp->b_queuetime = jiffies - age; - spin_lock(&btp->bt_delwrite_lock); - list_move(&bp->b_list, &btp->bt_delwrite_queue); - spin_unlock(&btp->bt_delwrite_lock); + spin_lock(&btp->bt_delwri_lock); + list_move(&bp->b_list, &btp->bt_delwri_queue); + spin_unlock(&btp->bt_delwri_lock); } STATIC void @@ -1638,15 +1636,13 @@ xfs_buf_delwri_split( unsigned long age) { xfs_buf_t *bp, *n; - struct list_head *dwq = &target->bt_delwrite_queue; - spinlock_t *dwlk = &target->bt_delwrite_lock; int skipped = 0; int force; force = test_and_clear_bit(XBT_FORCE_FLUSH, &target->bt_flags); INIT_LIST_HEAD(list); - spin_lock(dwlk); - list_for_each_entry_safe(bp, n, dwq, b_list) { + spin_lock(&target->bt_delwri_lock); + list_for_each_entry_safe(bp, n, &target->bt_delwri_queue, b_list) { ASSERT(bp->b_flags & XBF_DELWRI); if (!xfs_buf_ispinned(bp) && xfs_buf_trylock(bp)) { @@ -1663,10 +1659,9 @@ xfs_buf_delwri_split( } else skipped++; } - spin_unlock(dwlk); + spin_unlock(&target->bt_delwri_lock); return skipped; - } /* @@ -1716,7 +1711,7 @@ xfsbufd( } /* sleep for a long time if there is nothing to do. */ - if (list_empty(&target->bt_delwrite_queue)) + if (list_empty(&target->bt_delwri_queue)) tout = MAX_SCHEDULE_TIMEOUT; schedule_timeout_interruptible(tout); diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 3f543ed3009b..c23826685d3d 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -105,8 +105,8 @@ typedef struct xfs_buftarg { /* per device delwri queue */ struct task_struct *bt_task; - struct list_head bt_delwrite_queue; - spinlock_t bt_delwrite_lock; + struct list_head bt_delwri_queue; + spinlock_t bt_delwri_lock; unsigned long bt_flags; /* LRU control structures */ From 398d25ef23b10ce75424e0336a8d059dda1dbc8d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:09 +0000 Subject: [PATCH 09/69] xfs: remove dead ENODEV handling in xfs_destroy_ioend No driver returns ENODEV from it bio completion handler, not has this ever been documented. Remove the dead code dealing with it. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_aops.c | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 8c37dde4c521..d91564404abf 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -122,17 +122,6 @@ xfs_destroy_ioend( bh->b_end_io(bh, !ioend->io_error); } - /* - * Volume managers supporting multiple paths can send back ENODEV - * when the final path disappears. In this case continuing to fill - * the page cache with dirty data which cannot be written out is - * evil, so prevent that. - */ - if (unlikely(ioend->io_error == -ENODEV)) { - xfs_do_force_shutdown(ip->i_mount, SHUTDOWN_DEVICE_REQ, - __FILE__, __LINE__); - } - xfs_ioend_wake(ip); mempool_free(ioend, xfs_ioend_pool); } From c859cdd1da008b3825555be3242908088a3de366 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:10 +0000 Subject: [PATCH 10/69] xfs: defer AIO/DIO completions We really shouldn't complete AIO or DIO requests until we have finished the unwritten extent conversion and size update. This means fsync never has to pick up any ioends as all work has been completed when signalling I/O completion. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_aops.c | 26 +++++++++----------------- fs/xfs/xfs_aops.h | 1 + 2 files changed, 10 insertions(+), 17 deletions(-) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index d91564404abf..10660c364105 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -122,6 +122,11 @@ xfs_destroy_ioend( bh->b_end_io(bh, !ioend->io_error); } + if (ioend->io_iocb) { + if (ioend->io_isasync) + aio_complete(ioend->io_iocb, ioend->io_result, 0); + inode_dio_done(ioend->io_inode); + } xfs_ioend_wake(ip); mempool_free(ioend, xfs_ioend_pool); } @@ -236,8 +241,6 @@ xfs_end_io( /* ensure we don't spin on blocked ioends */ delay(1); } else { - if (ioend->io_iocb) - aio_complete(ioend->io_iocb, ioend->io_result, 0); xfs_destroy_ioend(ioend); } } @@ -274,6 +277,7 @@ xfs_alloc_ioend( * all the I/O from calling the completion routine too early. */ atomic_set(&ioend->io_remaining, 1); + ioend->io_isasync = 0; ioend->io_error = 0; ioend->io_list = NULL; ioend->io_type = type; @@ -1289,7 +1293,6 @@ xfs_end_io_direct_write( bool is_async) { struct xfs_ioend *ioend = iocb->private; - struct inode *inode = ioend->io_inode; /* * blockdev_direct_IO can return an error even after the I/O @@ -1300,28 +1303,17 @@ xfs_end_io_direct_write( ioend->io_offset = offset; ioend->io_size = size; + ioend->io_iocb = iocb; + ioend->io_result = ret; if (private && size > 0) ioend->io_type = IO_UNWRITTEN; if (is_async) { - /* - * If we are converting an unwritten extent we need to delay - * the AIO completion until after the unwrittent extent - * conversion has completed, otherwise do it ASAP. - */ - if (ioend->io_type == IO_UNWRITTEN) { - ioend->io_iocb = iocb; - ioend->io_result = ret; - } else { - aio_complete(iocb, ret, 0); - } + ioend->io_isasync = 1; xfs_finish_ioend(ioend); } else { xfs_finish_ioend_sync(ioend); } - - /* XXX: probably should move into the real I/O completion handler */ - inode_dio_done(inode); } STATIC ssize_t diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index 71f721e1a71f..ce3dcb50762e 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -47,6 +47,7 @@ typedef struct xfs_ioend { unsigned int io_type; /* delalloc / unwritten */ int io_error; /* I/O error code */ atomic_t io_remaining; /* hold count */ + unsigned int io_isasync : 1; /* needs aio_complete */ struct inode *io_inode; /* file being written to */ struct buffer_head *io_buffer_head;/* buffer linked list head */ struct buffer_head *io_buffer_tail;/* buffer linked list tail */ From fc0063c4474599b7a066ba76b90902abe21bc675 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:11 +0000 Subject: [PATCH 11/69] xfs: reduce ioend latency There is no reason to queue up ioends for processing in user context unless we actually need it. Just complete ioends that do not convert unwritten extents or need a size update from the end_io context. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_aops.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 10660c364105..e1ff0770784e 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -149,6 +149,15 @@ xfs_ioend_new_eof( return isize > ip->i_d.di_size ? isize : 0; } +/* + * Fast and loose check if this write could update the on-disk inode size. + */ +static inline bool xfs_ioend_is_append(struct xfs_ioend *ioend) +{ + return ioend->io_offset + ioend->io_size > + XFS_I(ioend->io_inode)->i_d.di_size; +} + /* * Update on-disk file size now that data has been written to disk. The * current in-memory file size is i_size. If a write is beyond eof i_new_size @@ -186,6 +195,9 @@ xfs_setfilesize( /* * Schedule IO completion handling on the final put of an ioend. + * + * If there is no work to do we might as well call it a day and free the + * ioend right now. */ STATIC void xfs_finish_ioend( @@ -194,8 +206,10 @@ xfs_finish_ioend( if (atomic_dec_and_test(&ioend->io_remaining)) { if (ioend->io_type == IO_UNWRITTEN) queue_work(xfsconvertd_workqueue, &ioend->io_work); - else + else if (xfs_ioend_is_append(ioend)) queue_work(xfsdatad_workqueue, &ioend->io_work); + else + xfs_destroy_ioend(ioend); } } From 2b3ffd7eb7b4392e3657c5046b055ca9f1f7cf5e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:12 +0000 Subject: [PATCH 12/69] xfs: wait for I/O completion when writing out pages in xfs_setattr_size The current code relies on the xfs_ioend_wait call later on to make sure all I/O actually has completed. The xfs_ioend_wait call will go away soon, so prepare for that by using the waiting filemap function. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_iops.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 673704fab748..32aca87bde5e 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -833,8 +833,8 @@ xfs_setattr_size( * care about here. */ if (ip->i_size != ip->i_d.di_size && iattr->ia_size > ip->i_d.di_size) { - error = xfs_flush_pages(ip, ip->i_d.di_size, iattr->ia_size, - XBF_ASYNC, FI_NONE); + error = xfs_flush_pages(ip, ip->i_d.di_size, iattr->ia_size, 0, + FI_NONE); if (error) goto out_unlock; } From 4a06fd262dbeb70a2c315f7259e063efa493fe3d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 23 Aug 2011 08:28:13 +0000 Subject: [PATCH 13/69] xfs: remove i_iocount We now have an i_dio_count filed and surrounding infrastructure to wait for direct I/O completion instead of i_icount, and we have never needed to iocount waits for buffered I/O given that we only set the page uptodate after finishing all required work. Thus remove i_iocount, and replace the actually needed waits with calls to inode_dio_wait. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_aops.c | 39 +-------------------------------------- fs/xfs/xfs_aops.h | 3 --- fs/xfs/xfs_file.c | 8 ++------ fs/xfs/xfs_iget.c | 2 -- fs/xfs/xfs_inode.h | 1 - fs/xfs/xfs_iops.c | 4 ++-- fs/xfs/xfs_super.c | 7 +------ fs/xfs/xfs_sync.c | 8 ++------ fs/xfs/xfs_vnodeops.c | 4 +--- 9 files changed, 9 insertions(+), 67 deletions(-) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index e1ff0770784e..46bba3e0af47 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -38,40 +38,6 @@ #include #include - -/* - * Prime number of hash buckets since address is used as the key. - */ -#define NVSYNC 37 -#define to_ioend_wq(v) (&xfs_ioend_wq[((unsigned long)v) % NVSYNC]) -static wait_queue_head_t xfs_ioend_wq[NVSYNC]; - -void __init -xfs_ioend_init(void) -{ - int i; - - for (i = 0; i < NVSYNC; i++) - init_waitqueue_head(&xfs_ioend_wq[i]); -} - -void -xfs_ioend_wait( - xfs_inode_t *ip) -{ - wait_queue_head_t *wq = to_ioend_wq(ip); - - wait_event(*wq, (atomic_read(&ip->i_iocount) == 0)); -} - -STATIC void -xfs_ioend_wake( - xfs_inode_t *ip) -{ - if (atomic_dec_and_test(&ip->i_iocount)) - wake_up(to_ioend_wq(ip)); -} - void xfs_count_page_state( struct page *page, @@ -115,7 +81,6 @@ xfs_destroy_ioend( xfs_ioend_t *ioend) { struct buffer_head *bh, *next; - struct xfs_inode *ip = XFS_I(ioend->io_inode); for (bh = ioend->io_buffer_head; bh; bh = next) { next = bh->b_private; @@ -127,7 +92,7 @@ xfs_destroy_ioend( aio_complete(ioend->io_iocb, ioend->io_result, 0); inode_dio_done(ioend->io_inode); } - xfs_ioend_wake(ip); + mempool_free(ioend, xfs_ioend_pool); } @@ -298,7 +263,6 @@ xfs_alloc_ioend( ioend->io_inode = inode; ioend->io_buffer_head = NULL; ioend->io_buffer_tail = NULL; - atomic_inc(&XFS_I(ioend->io_inode)->i_iocount); ioend->io_offset = 0; ioend->io_size = 0; ioend->io_iocb = NULL; @@ -558,7 +522,6 @@ xfs_cancel_ioend( unlock_buffer(bh); } while ((bh = next_bh) != NULL); - xfs_ioend_wake(XFS_I(ioend->io_inode)); mempool_free(ioend, xfs_ioend_pool); } while ((ioend = next) != NULL); } diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index ce3dcb50762e..116dd5c37034 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -61,9 +61,6 @@ typedef struct xfs_ioend { extern const struct address_space_operations xfs_address_space_operations; extern int xfs_get_blocks(struct inode *, sector_t, struct buffer_head *, int); -extern void xfs_ioend_init(void); -extern void xfs_ioend_wait(struct xfs_inode *); - extern void xfs_count_page_state(struct page *, int *, int *); #endif /* __XFS_AOPS_H__ */ diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index cbbac5cc9c26..ee63c4fb3639 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -149,10 +149,6 @@ xfs_file_fsync( xfs_iflags_clear(ip, XFS_ITRUNCATED); - xfs_ilock(ip, XFS_IOLOCK_SHARED); - xfs_ioend_wait(ip); - xfs_iunlock(ip, XFS_IOLOCK_SHARED); - if (mp->m_flags & XFS_MOUNT_BARRIER) { /* * If we have an RT and/or log subvolume we need to make sure @@ -758,7 +754,7 @@ xfs_file_aio_write_checks( * the dio layer. To avoid the problem with aio, we also need to wait for * outstanding IOs to complete so that unwritten extent conversion is completed * before we try to map the overlapping block. This is currently implemented by - * hitting it with a big hammer (i.e. xfs_ioend_wait()). + * hitting it with a big hammer (i.e. inode_dio_wait()). * * Returns with locks held indicated by @iolock and errors indicated by * negative return values. @@ -821,7 +817,7 @@ xfs_file_dio_aio_write( * otherwise demote the lock if we had to flush cached pages */ if (unaligned_io) - xfs_ioend_wait(ip); + inode_dio_wait(inode); else if (*iolock == XFS_IOLOCK_EXCL) { xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL); *iolock = XFS_IOLOCK_SHARED; diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c index 7759812c1bbe..0fa98b1c70ea 100644 --- a/fs/xfs/xfs_iget.c +++ b/fs/xfs/xfs_iget.c @@ -75,7 +75,6 @@ xfs_inode_alloc( return NULL; } - ASSERT(atomic_read(&ip->i_iocount) == 0); ASSERT(atomic_read(&ip->i_pincount) == 0); ASSERT(!spin_is_locked(&ip->i_flags_lock)); ASSERT(completion_done(&ip->i_flush)); @@ -150,7 +149,6 @@ xfs_inode_free( } /* asserts to verify all state is correct here */ - ASSERT(atomic_read(&ip->i_iocount) == 0); ASSERT(atomic_read(&ip->i_pincount) == 0); ASSERT(!spin_is_locked(&ip->i_flags_lock)); ASSERT(completion_done(&ip->i_flush)); diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 2380a4bcbece..760140d1dd66 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -257,7 +257,6 @@ typedef struct xfs_inode { xfs_fsize_t i_size; /* in-memory size */ xfs_fsize_t i_new_size; /* size when write completes */ - atomic_t i_iocount; /* outstanding I/O count */ /* VFS inode */ struct inode i_vnode; /* embedded VFS inode */ diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 32aca87bde5e..e041e917c1d9 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -840,9 +840,9 @@ xfs_setattr_size( } /* - * Wait for all I/O to complete. + * Wait for all direct I/O to complete. */ - xfs_ioend_wait(ip); + inode_dio_wait(inode); error = -block_truncate_page(inode->i_mapping, iattr->ia_size, xfs_get_blocks); diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 2366c54cc4fa..54d5e102ffe1 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -796,8 +796,6 @@ xfs_fs_destroy_inode( if (is_bad_inode(inode)) goto out_reclaim; - xfs_ioend_wait(ip); - ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0); /* @@ -837,7 +835,6 @@ xfs_fs_inode_init_once( inode_init_once(VFS_I(ip)); /* xfs inode */ - atomic_set(&ip->i_iocount, 0); atomic_set(&ip->i_pincount, 0); spin_lock_init(&ip->i_flags_lock); init_waitqueue_head(&ip->i_ipin_wait); @@ -914,9 +911,8 @@ xfs_fs_write_inode( * of forcing it all the way to stable storage using a * synchronous transaction we let the log force inside the * ->sync_fs call do that for thus, which reduces the number - * of synchronous log foces dramatically. + * of synchronous log forces dramatically. */ - xfs_ioend_wait(ip); error = xfs_log_inode(ip); if (error) goto out; @@ -1681,7 +1677,6 @@ init_xfs_fs(void) printk(KERN_INFO XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled\n"); - xfs_ioend_init(); xfs_dir_startup(); error = xfs_init_zones(); diff --git a/fs/xfs/xfs_sync.c b/fs/xfs/xfs_sync.c index 90cc197e0433..bf2b38c21caa 100644 --- a/fs/xfs/xfs_sync.c +++ b/fs/xfs/xfs_sync.c @@ -227,21 +227,17 @@ xfs_sync_inode_data( int error = 0; if (!mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) - goto out_wait; + return 0; if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED)) { if (flags & SYNC_TRYLOCK) - goto out_wait; + return 0; xfs_ilock(ip, XFS_IOLOCK_SHARED); } error = xfs_flush_pages(ip, 0, -1, (flags & SYNC_WAIT) ? 0 : XBF_ASYNC, FI_NONE); xfs_iunlock(ip, XFS_IOLOCK_SHARED); - - out_wait: - if (flags & SYNC_WAIT) - xfs_ioend_wait(ip); return error; } diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 51fc429527bc..c2ff0fc86567 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -647,8 +647,6 @@ xfs_inactive( if (truncate) { xfs_ilock(ip, XFS_IOLOCK_EXCL); - xfs_ioend_wait(ip); - error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, @@ -2076,7 +2074,7 @@ xfs_free_file_space( if (need_iolock) { xfs_ilock(ip, XFS_IOLOCK_EXCL); /* wait for the completion of any pending DIOs */ - xfs_ioend_wait(ip); + inode_dio_wait(VFS_I(ip)); } rounding = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE); From 859f57ca00805e6c482eef1a7ab073097d02c8ca Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sat, 27 Aug 2011 14:45:11 +0000 Subject: [PATCH 14/69] xfs: avoid synchronous transactions when deleting attr blocks Currently xfs_attr_inactive causes a synchronous transactions if we are removing a file that has any extents allocated to the attribute fork, and thus makes XFS extremely slow at removing files with out of line extended attributes. The code looks a like a relict from the days before the busy extent list, but with the busy extent list we avoid reusing data and attr extents that have been freed but not commited yet, so this code is just as superflous as the synchronous transactions for data blocks. Signed-off-by: Christoph Hellwig Reported-by: Bernd Schubert Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 12 ------------ fs/xfs/xfs_bmap.c | 10 +--------- 2 files changed, 1 insertion(+), 21 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 8f0f65833500..3dd5c9c374cb 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -823,18 +823,6 @@ xfs_attr_inactive(xfs_inode_t *dp) if (error) goto out; - /* - * Signal synchronous inactive transactions unless this is a - * synchronous mount filesystem in which case we know that we're here - * because we've been called out of xfs_inactive which means that the - * last reference is gone and the unlink transaction has already hit - * the disk so async inactive transactions are safe. - */ - if (!(mp->m_flags & XFS_MOUNT_WSYNC)) { - if (dp->i_d.di_anextents > 0) - xfs_trans_set_sync(trans); - } - error = xfs_itruncate_extents(&trans, dp, XFS_ATTR_FORK, 0); if (error) goto out; diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 452a291383ab..a6075c0f8456 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -3782,19 +3782,11 @@ xfs_bmap_compute_maxlevels( * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi * caller. Frees all the extents that need freeing, which must be done * last due to locking considerations. We never free any extents in - * the first transaction. This is to allow the caller to make the first - * transaction a synchronous one so that the pointers to the data being - * broken in this transaction will be permanent before the data is actually - * freed. This is necessary to prevent blocks from being reallocated - * and written to before the free and reallocation are actually permanent. - * We do not just make the first transaction synchronous here, because - * there are more efficient ways to gain the same protection in some cases - * (see the file truncation code). + * the first transaction. * * Return 1 if the given transaction was committed and a new one * started, and 0 otherwise in the committed parameter. */ -/*ARGSUSED*/ int /* error */ xfs_bmap_finish( xfs_trans_t **tp, /* transaction pointer addr */ From c58cb165bd44de8aaee9755a144136ae743be116 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sat, 27 Aug 2011 14:42:53 +0000 Subject: [PATCH 15/69] xfs: avoid direct I/O write vs buffered I/O race Currently a buffered reader or writer can add pages to the pagecache while we are waiting for the iolock in xfs_file_dio_aio_write. Prevent this by re-checking mapping->nrpages after we got the iolock, and if nessecary upgrade the lock to exclusive mode. To simplify this a bit only take the ilock inside of xfs_file_aio_write_checks. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_file.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index ee63c4fb3639..06fe97e56e48 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -676,6 +676,7 @@ xfs_file_aio_write_checks( xfs_fsize_t new_size; int error = 0; + xfs_rw_ilock(ip, XFS_ILOCK_EXCL); *new_sizep = 0; restart: error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode)); @@ -798,14 +799,24 @@ xfs_file_dio_aio_write( *iolock = XFS_IOLOCK_EXCL; else *iolock = XFS_IOLOCK_SHARED; - xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock); + xfs_rw_ilock(ip, *iolock); + + /* + * Recheck if there are cached pages that need invalidate after we got + * the iolock to protect against other threads adding new pages while + * we were waiting for the iolock. + */ + if (mapping->nrpages && *iolock == XFS_IOLOCK_SHARED) { + xfs_rw_iunlock(ip, *iolock); + *iolock = XFS_IOLOCK_EXCL; + xfs_rw_ilock(ip, *iolock); + } ret = xfs_file_aio_write_checks(file, &pos, &count, new_size, iolock); if (ret) return ret; if (mapping->nrpages) { - WARN_ON(*iolock != XFS_IOLOCK_EXCL); ret = -xfs_flushinval_pages(ip, (pos & PAGE_CACHE_MASK), -1, FI_REMAPF_LOCKED); if (ret) @@ -851,7 +862,7 @@ xfs_file_buffered_aio_write( size_t count = ocount; *iolock = XFS_IOLOCK_EXCL; - xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock); + xfs_rw_ilock(ip, *iolock); ret = xfs_file_aio_write_checks(file, &pos, &count, new_size, iolock); if (ret) From 04f658ee229f60dbb9a0dc2f3d6871b12b758051 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 24 Aug 2011 05:59:25 +0000 Subject: [PATCH 16/69] xfs: improve ioend error handling Return unwritten extent conversion errors to aio_complete. Skip both unwritten extent conversion and size updates if we had an I/O error or the filesystem has been shut down. Return -EIO to the aio/buffer completion handlers in case of a forced shutdown. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_aops.c | 27 +++++++++++++++++---------- 1 file changed, 17 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 46bba3e0af47..22aadf667862 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -88,8 +88,10 @@ xfs_destroy_ioend( } if (ioend->io_iocb) { - if (ioend->io_isasync) - aio_complete(ioend->io_iocb, ioend->io_result, 0); + if (ioend->io_isasync) { + aio_complete(ioend->io_iocb, ioend->io_error ? + ioend->io_error : ioend->io_result, 0); + } inode_dio_done(ioend->io_inode); } @@ -141,9 +143,6 @@ xfs_setfilesize( xfs_inode_t *ip = XFS_I(ioend->io_inode); xfs_fsize_t isize; - if (unlikely(ioend->io_error)) - return 0; - if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) return EAGAIN; @@ -189,17 +188,24 @@ xfs_end_io( struct xfs_inode *ip = XFS_I(ioend->io_inode); int error = 0; + if (XFS_FORCED_SHUTDOWN(ip->i_mount)) { + error = -EIO; + goto done; + } + if (ioend->io_error) + goto done; + /* * For unwritten extents we need to issue transactions to convert a * range to normal written extens after the data I/O has finished. */ - if (ioend->io_type == IO_UNWRITTEN && - likely(!ioend->io_error && !XFS_FORCED_SHUTDOWN(ip->i_mount))) { - + if (ioend->io_type == IO_UNWRITTEN) { error = xfs_iomap_write_unwritten(ip, ioend->io_offset, ioend->io_size); - if (error) - ioend->io_error = error; + if (error) { + ioend->io_error = -error; + goto done; + } } /* @@ -209,6 +215,7 @@ xfs_end_io( error = xfs_setfilesize(ioend); ASSERT(!error || error == EAGAIN); +done: /* * If we didn't complete processing of the ioend, requeue it to the * tail of the workqueue for another attempt later. Otherwise destroy From b522950f0ab8551f2ef56c210ebd50e6c6396601 Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Wed, 7 Sep 2011 19:37:54 +0000 Subject: [PATCH 17/69] xfs: Check the return value of xfs_buf_get() Check the return value of xfs_buf_get() and fail appropriately. Signed-off-by: Chandra Seetharaman Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 4 ++-- fs/xfs/xfs_fsops.c | 20 ++++++++++++++++++++ 2 files changed, 22 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 3dd5c9c374cb..c7dab0c0bdda 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2109,8 +2109,8 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) bp = xfs_buf_get(mp->m_ddev_targp, dblkno, blkcnt, XBF_LOCK | XBF_DONT_BLOCK); - ASSERT(!xfs_buf_geterror(bp)); - + if (!bp) + return ENOMEM; tmp = (valuelen < XFS_BUF_SIZE(bp)) ? valuelen : XFS_BUF_SIZE(bp); xfs_buf_iomove(bp, 0, tmp, src, XBRW_WRITE); diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index e023f940a3dd..1c6fdeb702ff 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -194,6 +194,10 @@ xfs_growfs_data_private( bp = xfs_buf_get(mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), XBF_LOCK | XBF_MAPPED); + if (!bp) { + error = ENOMEM; + goto error0; + } agf = XFS_BUF_TO_AGF(bp); memset(agf, 0, mp->m_sb.sb_sectsize); agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC); @@ -227,6 +231,10 @@ xfs_growfs_data_private( bp = xfs_buf_get(mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), XBF_LOCK | XBF_MAPPED); + if (!bp) { + error = ENOMEM; + goto error0; + } agi = XFS_BUF_TO_AGI(bp); memset(agi, 0, mp->m_sb.sb_sectsize); agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC); @@ -253,6 +261,10 @@ xfs_growfs_data_private( XFS_AGB_TO_DADDR(mp, agno, XFS_BNO_BLOCK(mp)), BTOBB(mp->m_sb.sb_blocksize), XBF_LOCK | XBF_MAPPED); + if (!bp) { + error = ENOMEM; + goto error0; + } block = XFS_BUF_TO_BLOCK(bp); memset(block, 0, mp->m_sb.sb_blocksize); block->bb_magic = cpu_to_be32(XFS_ABTB_MAGIC); @@ -276,6 +288,10 @@ xfs_growfs_data_private( XFS_AGB_TO_DADDR(mp, agno, XFS_CNT_BLOCK(mp)), BTOBB(mp->m_sb.sb_blocksize), XBF_LOCK | XBF_MAPPED); + if (!bp) { + error = ENOMEM; + goto error0; + } block = XFS_BUF_TO_BLOCK(bp); memset(block, 0, mp->m_sb.sb_blocksize); block->bb_magic = cpu_to_be32(XFS_ABTC_MAGIC); @@ -300,6 +316,10 @@ xfs_growfs_data_private( XFS_AGB_TO_DADDR(mp, agno, XFS_IBT_BLOCK(mp)), BTOBB(mp->m_sb.sb_blocksize), XBF_LOCK | XBF_MAPPED); + if (!bp) { + error = ENOMEM; + goto error0; + } block = XFS_BUF_TO_BLOCK(bp); memset(block, 0, mp->m_sb.sb_blocksize); block->bb_magic = cpu_to_be32(XFS_IBT_MAGIC); From 2a30f36d9069b0646dcdd73def5fd7ab674bffd6 Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Tue, 20 Sep 2011 13:56:55 +0000 Subject: [PATCH 18/69] xfs: Check the return value of xfs_trans_get_buf() Check the return value of xfs_trans_get_buf() and fail appropriately. Signed-off-by: Chandra Seetharaman Signed-off-by: Alex Elder --- fs/xfs/xfs_attr_leaf.c | 2 ++ fs/xfs/xfs_btree.c | 3 ++- fs/xfs/xfs_dquot.c | 5 ++++- fs/xfs/xfs_ialloc.c | 13 ++++++++----- fs/xfs/xfs_inode.c | 9 ++++++--- fs/xfs/xfs_vnodeops.c | 9 ++++++++- 6 files changed, 30 insertions(+), 11 deletions(-) diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 8fad9602542b..58c3add07b65 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -2948,6 +2948,8 @@ xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp, bp = xfs_trans_get_buf(*trans, dp->i_mount->m_ddev_targp, dblkno, dblkcnt, XBF_LOCK); + if (!bp) + return ENOMEM; xfs_trans_binval(*trans, bp); /* * Roll to next transaction. diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c index 2b9fd385e27d..28cc0199dc1f 100644 --- a/fs/xfs/xfs_btree.c +++ b/fs/xfs/xfs_btree.c @@ -970,7 +970,8 @@ xfs_btree_get_buf_block( *bpp = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d, mp->m_bsize, flags); - ASSERT(!xfs_buf_geterror(*bpp)); + if (!*bpp) + return ENOMEM; *block = XFS_BUF_TO_BLOCK(*bpp); return 0; diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 3e2ccaedc51e..0c5fe66ce92b 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -402,8 +402,11 @@ xfs_qm_dqalloc( dqp->q_blkno, mp->m_quotainfo->qi_dqchunklen, 0); - if (!bp || (error = xfs_buf_geterror(bp))) + + error = xfs_buf_geterror(bp); + if (error) goto error1; + /* * Make a chunk of dquots out of this buffer and log * the entire thing. diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 9f24ec28283b..207e0b0c0730 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -150,7 +150,7 @@ xfs_check_agi_freecount( /* * Initialise a new set of inodes. */ -STATIC void +STATIC int xfs_ialloc_inode_init( struct xfs_mount *mp, struct xfs_trans *tp, @@ -202,8 +202,8 @@ xfs_ialloc_inode_init( fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize * blks_per_cluster, XBF_LOCK); - ASSERT(!xfs_buf_geterror(fbuf)); - + if (!fbuf) + return ENOMEM; /* * Initialize all inodes in this buffer and then log them. * @@ -225,6 +225,7 @@ xfs_ialloc_inode_init( } xfs_trans_inode_alloc_buf(tp, fbuf); } + return 0; } /* @@ -369,9 +370,11 @@ xfs_ialloc_ag_alloc( * rather than a linear progression to prevent the next generation * number from being easily guessable. */ - xfs_ialloc_inode_init(args.mp, tp, agno, args.agbno, args.len, - random32()); + error = xfs_ialloc_inode_init(args.mp, tp, agno, args.agbno, + args.len, random32()); + if (error) + return error; /* * Convert the results. */ diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 7f237ba3c292..d689253fdfda 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1644,7 +1644,7 @@ xfs_iunlink_remove( * inodes that are in memory - they all must be marked stale and attached to * the cluster buffer. */ -STATIC void +STATIC int xfs_ifree_cluster( xfs_inode_t *free_ip, xfs_trans_t *tp, @@ -1690,6 +1690,8 @@ xfs_ifree_cluster( mp->m_bsize * blks_per_cluster, XBF_LOCK); + if (!bp) + return ENOMEM; /* * Walk the inodes already attached to the buffer and mark them * stale. These will all have the flush locks held, so an @@ -1799,6 +1801,7 @@ xfs_ifree_cluster( } xfs_perag_put(pag); + return 0; } /* @@ -1878,10 +1881,10 @@ xfs_ifree( dip->di_mode = 0; if (delete) { - xfs_ifree_cluster(ip, tp, first_ino); + error = xfs_ifree_cluster(ip, tp, first_ino); } - return 0; + return error; } /* diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index c2ff0fc86567..0d1caec873a1 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -308,6 +308,10 @@ xfs_inactive_symlink_rmt( bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, mval[i].br_startblock), XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0); + if (!bp) { + error = ENOMEM; + goto error1; + } xfs_trans_binval(tp, bp); } /* @@ -1648,7 +1652,10 @@ xfs_symlink( byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount); bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, BTOBB(byte_cnt), 0); - ASSERT(!xfs_buf_geterror(bp)); + if (!bp) { + error = ENOMEM; + goto error2; + } if (pathlen < byte_cnt) { byte_cnt = pathlen; } From eabbaf118239d0d4188298b52751040f3b4cc28f Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Thu, 8 Sep 2011 20:18:50 +0000 Subject: [PATCH 19/69] xfs: Fix the incorrect comment in the header of _xfs_buf_find Fix the incorrect comment in the header of the function _xfs_buf_find(). Signed-off-by: Chandra Seetharaman Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index d3c2b58d7d70..e3af85095ddb 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -415,10 +415,7 @@ _xfs_buf_map_pages( /* * Look up, and creates if absent, a lockable buffer for * a given range of an inode. The buffer is returned - * locked. If other overlapping buffers exist, they are - * released before the new buffer is created and locked, - * which may imply that this call will block until those buffers - * are unlocked. No I/O is implied by this call. + * locked. No I/O is implied by this call. */ xfs_buf_t * _xfs_buf_find( From ed32201e65e15f3e6955cb84cbb544b08f81e5a5 Mon Sep 17 00:00:00 2001 From: Mitsuo Hayasaka Date: Sat, 17 Sep 2011 13:38:38 +0000 Subject: [PATCH 20/69] xfs: Return -EIO when xfs_vn_getattr() failed An attribute of inode can be fetched via xfs_vn_getattr() in XFS. Currently it returns EIO, not negative value, when it failed. As a result, the system call returns not negative value even though an error occured. The stat(2), ls and mv commands cannot handle this error and do not work correctly. This patch fixes this bug, and returns -EIO, not EIO when an error is detected in xfs_vn_getattr(). Signed-off-by: Mitsuo Hayasaka Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_iops.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index e041e917c1d9..e6b3e7620888 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -465,7 +465,7 @@ xfs_vn_getattr( trace_xfs_getattr(ip); if (XFS_FORCED_SHUTDOWN(mp)) - return XFS_ERROR(EIO); + return -XFS_ERROR(EIO); stat->size = XFS_ISIZE(ip); stat->dev = inode->i_sb->s_dev; From e7455e02e5effcdd49bb28e7dfface2d3473de52 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:41 +0000 Subject: [PATCH 21/69] xfs: remove the first extent special case in xfs_bmap_add_extent Both xfs_bmap_add_extent_hole_delay and xfs_bmap_add_extent_hole_real already contain code to handle the case where there is no extent to merge with, which is effectively the same as the code duplicated here. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 19 +------------------ 1 file changed, 1 insertion(+), 18 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index a6075c0f8456..927e5454a43c 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -465,27 +465,10 @@ xfs_bmap_add_extent( ASSERT(*idx >= 0); ASSERT(*idx <= nextents); - /* - * This is the first extent added to a new/empty file. - * Special case this one, so other routines get to assume there are - * already extents in the list. - */ - if (nextents == 0) { - xfs_iext_insert(ip, *idx, 1, new, - whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0); - - ASSERT(cur == NULL); - - if (!isnullstartblock(new->br_startblock)) { - XFS_IFORK_NEXT_SET(ip, whichfork, 1); - logflags = XFS_ILOG_CORE | xfs_ilog_fext(whichfork); - } else - logflags = 0; - } /* * Any kind of new delayed allocation goes here. */ - else if (isnullstartblock(new->br_startblock)) { + if (isnullstartblock(new->br_startblock)) { if (cur) ASSERT((cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL) == 0); From b9b984d7846e37c57e5b3f8cd883ad45e8ebc2cf Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:42 +0000 Subject: [PATCH 22/69] xfs: remove impossible to read code in xfs_bmap_add_extent_delay_real We already have the worst case blocks reserved, so xfs_icsb_modify_counters won't fail in xfs_bmap_add_extent_delay_real. In fact we've had an assert to catch this case since day and it never triggered. So remove the code to try smaller reservations, and just return the error for that case in addition to keeping the assert. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 35 ++++++++--------------------------- 1 file changed, 8 insertions(+), 27 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 927e5454a43c..ebcd38b14830 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -1045,34 +1045,15 @@ xfs_bmap_add_extent_delay_real( temp2 = xfs_bmap_worst_indlen(ip, temp2); diff = (int)(temp + temp2 - startblockval(PREV.br_startblock) - (cur ? cur->bc_private.b.allocated : 0)); - if (diff > 0 && - xfs_icsb_modify_counters(ip->i_mount, XFS_SBS_FDBLOCKS, - -((int64_t)diff), 0)) { - /* - * Ick gross gag me with a spoon. - */ - ASSERT(0); /* want to see if this ever happens! */ - while (diff > 0) { - if (temp) { - temp--; - diff--; - if (!diff || - !xfs_icsb_modify_counters(ip->i_mount, - XFS_SBS_FDBLOCKS, - -((int64_t)diff), 0)) - break; - } - if (temp2) { - temp2--; - diff--; - if (!diff || - !xfs_icsb_modify_counters(ip->i_mount, - XFS_SBS_FDBLOCKS, - -((int64_t)diff), 0)) - break; - } - } + if (diff > 0) { + error = xfs_icsb_modify_counters(ip->i_mount, + XFS_SBS_FDBLOCKS, + -((int64_t)diff), 0); + ASSERT(!error); + if (error) + goto done; } + ep = xfs_iext_get_ext(ifp, *idx); xfs_bmbt_set_startblock(ep, nullstartblock((int)temp)); trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); From ecee76ba9d91fdcbdff933ca1bd41465ca4c4fdb Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:43 +0000 Subject: [PATCH 23/69] xfs: remove the nextents variable in xfs_bmapi Instead of using a local variable that needs to updated when we modify the extent map just check ifp->if_bytes directly where we use it. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index ebcd38b14830..05d38adbb85e 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4294,7 +4294,6 @@ xfs_bmapi( xfs_mount_t *mp; /* xfs mount structure */ int n; /* current extent index */ int nallocs; /* number of extents alloc'd */ - xfs_extnum_t nextents; /* number of extents in file */ xfs_fileoff_t obno; /* old block number (offset) */ xfs_bmbt_irec_t prev; /* previous file extent record */ int tmp_logflags; /* temp flags holder */ @@ -4380,7 +4379,6 @@ xfs_bmapi( goto error0; ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); n = 0; end = bno + len; obno = bno; @@ -4622,7 +4620,6 @@ xfs_bmapi( if (error) goto error0; ep = xfs_iext_get_ext(ifp, lastx); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); xfs_bmbt_get_all(ep, &got); ASSERT(got.br_startoff <= aoff); ASSERT(got.br_startoff + got.br_blockcount >= @@ -4723,7 +4720,6 @@ xfs_bmapi( if (error) goto error0; ep = xfs_iext_get_ext(ifp, lastx); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); xfs_bmbt_get_all(ep, &got); /* * We may have combined previously unwritten @@ -4781,7 +4777,7 @@ xfs_bmapi( * Else go on to the next record. */ prev = got; - if (++lastx < nextents) { + if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) { ep = xfs_iext_get_ext(ifp, lastx); xfs_bmbt_get_all(ep, &got); } else { From aef9a89586fc8475bf0333b8736d5aa8aa6f4897 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:44 +0000 Subject: [PATCH 24/69] xfs: factor extent map manipulations out of xfs_bmapi To further improve the readability of xfs_bmapi(), factor the pure extent map manipulations out into separate functions. This removes large blocks of logic from the xfs_bmapi() code loop and makes it easier to see the operational logic flow for xfs_bmapi(). Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 181 +++++++++++++++++++++++++++------------------- 1 file changed, 107 insertions(+), 74 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 05d38adbb85e..59f63949d3bc 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4248,6 +4248,107 @@ xfs_bmap_validate_ret( #endif /* DEBUG */ +/* + * Trim the returned map to the required bounds + */ +STATIC void +xfs_bmapi_trim_map( + struct xfs_bmbt_irec *mval, + struct xfs_bmbt_irec *got, + xfs_fileoff_t *bno, + xfs_filblks_t len, + xfs_fileoff_t obno, + xfs_fileoff_t end, + int n, + int flags) +{ + if ((flags & XFS_BMAPI_ENTIRE) || + got->br_startoff + got->br_blockcount <= obno) { + *mval = *got; + if (isnullstartblock(got->br_startblock)) + mval->br_startblock = DELAYSTARTBLOCK; + return; + } + + if (obno > *bno) + *bno = obno; + ASSERT((*bno >= obno) || (n == 0)); + ASSERT(*bno < end); + mval->br_startoff = *bno; + if (isnullstartblock(got->br_startblock)) + mval->br_startblock = DELAYSTARTBLOCK; + else + mval->br_startblock = got->br_startblock + + (*bno - got->br_startoff); + /* + * Return the minimum of what we got and what we asked for for + * the length. We can use the len variable here because it is + * modified below and we could have been there before coming + * here if the first part of the allocation didn't overlap what + * was asked for. + */ + mval->br_blockcount = XFS_FILBLKS_MIN(end - *bno, + got->br_blockcount - (*bno - got->br_startoff)); + mval->br_state = got->br_state; + ASSERT(mval->br_blockcount <= len); + return; +} + +/* + * Update and validate the extent map to return + */ +STATIC void +xfs_bmapi_update_map( + struct xfs_bmbt_irec **map, + xfs_fileoff_t *bno, + xfs_filblks_t *len, + xfs_fileoff_t obno, + xfs_fileoff_t end, + int *n, + int flags) +{ + xfs_bmbt_irec_t *mval = *map; + + ASSERT((flags & XFS_BMAPI_ENTIRE) || + ((mval->br_startoff + mval->br_blockcount) <= end)); + ASSERT((flags & XFS_BMAPI_ENTIRE) || (mval->br_blockcount <= *len) || + (mval->br_startoff < obno)); + + *bno = mval->br_startoff + mval->br_blockcount; + *len = end - *bno; + if (*n > 0 && mval->br_startoff == mval[-1].br_startoff) { + /* update previous map with new information */ + ASSERT(mval->br_startblock == mval[-1].br_startblock); + ASSERT(mval->br_blockcount > mval[-1].br_blockcount); + ASSERT(mval->br_state == mval[-1].br_state); + mval[-1].br_blockcount = mval->br_blockcount; + mval[-1].br_state = mval->br_state; + } else if (*n > 0 && mval->br_startblock != DELAYSTARTBLOCK && + mval[-1].br_startblock != DELAYSTARTBLOCK && + mval[-1].br_startblock != HOLESTARTBLOCK && + mval->br_startblock == mval[-1].br_startblock + + mval[-1].br_blockcount && + ((flags & XFS_BMAPI_IGSTATE) || + mval[-1].br_state == mval->br_state)) { + ASSERT(mval->br_startoff == + mval[-1].br_startoff + mval[-1].br_blockcount); + mval[-1].br_blockcount += mval->br_blockcount; + } else if (*n > 0 && + mval->br_startblock == DELAYSTARTBLOCK && + mval[-1].br_startblock == DELAYSTARTBLOCK && + mval->br_startoff == + mval[-1].br_startoff + mval[-1].br_blockcount) { + mval[-1].br_blockcount += mval->br_blockcount; + mval[-1].br_state = mval->br_state; + } else if (!((*n == 0) && + ((mval->br_startoff + mval->br_blockcount) <= + obno))) { + mval++; + (*n)++; + } + *map = mval; +} + /* * Map file blocks to filesystem blocks. * File range is given by the bno/len pair. @@ -4650,44 +4751,9 @@ xfs_bmapi( n++; continue; } - /* - * Then deal with the allocated space we found. - */ - ASSERT(ep != NULL); - if (!(flags & XFS_BMAPI_ENTIRE) && - (got.br_startoff + got.br_blockcount > obno)) { - if (obno > bno) - bno = obno; - ASSERT((bno >= obno) || (n == 0)); - ASSERT(bno < end); - mval->br_startoff = bno; - if (isnullstartblock(got.br_startblock)) { - ASSERT(!wr || (flags & XFS_BMAPI_DELAY)); - mval->br_startblock = DELAYSTARTBLOCK; - } else - mval->br_startblock = - got.br_startblock + - (bno - got.br_startoff); - /* - * Return the minimum of what we got and what we - * asked for for the length. We can use the len - * variable here because it is modified below - * and we could have been there before coming - * here if the first part of the allocation - * didn't overlap what was asked for. - */ - mval->br_blockcount = - XFS_FILBLKS_MIN(end - bno, got.br_blockcount - - (bno - got.br_startoff)); - mval->br_state = got.br_state; - ASSERT(mval->br_blockcount <= len); - } else { - *mval = got; - if (isnullstartblock(mval->br_startblock)) { - ASSERT(!wr || (flags & XFS_BMAPI_DELAY)); - mval->br_startblock = DELAYSTARTBLOCK; - } - } + + /* Deal with the allocated space we found. */ + xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); /* * Check if writing previously allocated but @@ -4730,42 +4796,9 @@ xfs_bmapi( continue; } - ASSERT((flags & XFS_BMAPI_ENTIRE) || - ((mval->br_startoff + mval->br_blockcount) <= end)); - ASSERT((flags & XFS_BMAPI_ENTIRE) || - (mval->br_blockcount <= len) || - (mval->br_startoff < obno)); - bno = mval->br_startoff + mval->br_blockcount; - len = end - bno; - if (n > 0 && mval->br_startoff == mval[-1].br_startoff) { - ASSERT(mval->br_startblock == mval[-1].br_startblock); - ASSERT(mval->br_blockcount > mval[-1].br_blockcount); - ASSERT(mval->br_state == mval[-1].br_state); - mval[-1].br_blockcount = mval->br_blockcount; - mval[-1].br_state = mval->br_state; - } else if (n > 0 && mval->br_startblock != DELAYSTARTBLOCK && - mval[-1].br_startblock != DELAYSTARTBLOCK && - mval[-1].br_startblock != HOLESTARTBLOCK && - mval->br_startblock == - mval[-1].br_startblock + mval[-1].br_blockcount && - ((flags & XFS_BMAPI_IGSTATE) || - mval[-1].br_state == mval->br_state)) { - ASSERT(mval->br_startoff == - mval[-1].br_startoff + mval[-1].br_blockcount); - mval[-1].br_blockcount += mval->br_blockcount; - } else if (n > 0 && - mval->br_startblock == DELAYSTARTBLOCK && - mval[-1].br_startblock == DELAYSTARTBLOCK && - mval->br_startoff == - mval[-1].br_startoff + mval[-1].br_blockcount) { - mval[-1].br_blockcount += mval->br_blockcount; - mval[-1].br_state = mval->br_state; - } else if (!((n == 0) && - ((mval->br_startoff + mval->br_blockcount) <= - obno))) { - mval++; - n++; - } + /* update the extent map to return */ + xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags); + /* * If we're done, stop now. Stop when we've allocated * XFS_BMAP_MAX_NMAP extents no matter what. Otherwise From 5c8ed2021ff291f5e399a9b43c4f699b2fc58fbb Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:45 +0000 Subject: [PATCH 25/69] xfs: introduce xfs_bmapi_read() xfs_bmapi() currently handles both extent map reading and allocation. As a result, the code is littered with "if (wr)" branches to conditionally do allocation operations if required. This makes the code much harder to follow and causes significant indent issues with the code. Given that read mapping is much simpler than allocation, we can split out read mapping from xfs_bmapi() and reuse the logic that we have already factored out do do all the hard work of handling the extent map manipulations. The results in a much simpler function for the common extent read operations, and will allow the allocation code to be simplified in another commit. Once xfs_bmapi_read() is implemented, convert all the callers of xfs_bmapi() that are only reading extents to use the new function. Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_aops.c | 8 ++-- fs/xfs/xfs_attr.c | 30 +++++------- fs/xfs/xfs_attr_leaf.c | 5 +- fs/xfs/xfs_bmap.c | 104 ++++++++++++++++++++++++++++++++++++++--- fs/xfs/xfs_bmap.h | 4 ++ fs/xfs/xfs_da_btree.c | 9 ++-- fs/xfs/xfs_dir2_leaf.c | 6 +-- fs/xfs/xfs_dquot.c | 5 +- fs/xfs/xfs_file.c | 10 ++-- fs/xfs/xfs_inode.c | 12 ++--- fs/xfs/xfs_iomap.c | 4 +- fs/xfs/xfs_qm.c | 7 +-- fs/xfs/xfs_vnodeops.c | 24 +++++----- 13 files changed, 151 insertions(+), 77 deletions(-) diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 22aadf667862..11b2aad982d4 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -315,8 +315,8 @@ xfs_map_blocks( count = mp->m_maxioffset - offset; end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count); offset_fsb = XFS_B_TO_FSBT(mp, offset); - error = xfs_bmapi(NULL, ip, offset_fsb, end_fsb - offset_fsb, - bmapi_flags, NULL, 0, imap, &nimaps, NULL); + error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb, + imap, &nimaps, bmapi_flags); xfs_iunlock(ip, XFS_ILOCK_SHARED); if (error) @@ -1138,8 +1138,8 @@ __xfs_get_blocks( end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + size); offset_fsb = XFS_B_TO_FSBT(mp, offset); - error = xfs_bmapi(NULL, ip, offset_fsb, end_fsb - offset_fsb, - XFS_BMAPI_ENTIRE, NULL, 0, &imap, &nimaps, NULL); + error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb, + &imap, &nimaps, XFS_BMAPI_ENTIRE); if (error) goto out_unlock; diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index c7dab0c0bdda..41ef02b7185a 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -1963,10 +1963,9 @@ xfs_attr_rmtval_get(xfs_da_args_t *args) lblkno = args->rmtblkno; while (valuelen > 0) { nmap = ATTR_RMTVALUE_MAPSIZE; - error = xfs_bmapi(args->trans, args->dp, (xfs_fileoff_t)lblkno, - args->rmtblkcnt, - XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA, - NULL, 0, map, &nmap, NULL); + error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno, + args->rmtblkcnt, map, &nmap, + XFS_BMAPI_ATTRFORK); if (error) return(error); ASSERT(nmap >= 1); @@ -2092,14 +2091,11 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) */ xfs_bmap_init(args->flist, args->firstblock); nmap = 1; - error = xfs_bmapi(NULL, dp, (xfs_fileoff_t)lblkno, - args->rmtblkcnt, - XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA, - args->firstblock, 0, &map, &nmap, - NULL); - if (error) { + error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno, + args->rmtblkcnt, &map, &nmap, + XFS_BMAPI_ATTRFORK); + if (error) return(error); - } ASSERT(nmap == 1); ASSERT((map.br_startblock != DELAYSTARTBLOCK) && (map.br_startblock != HOLESTARTBLOCK)); @@ -2155,16 +2151,12 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) /* * Try to remember where we decided to put the value. */ - xfs_bmap_init(args->flist, args->firstblock); nmap = 1; - error = xfs_bmapi(NULL, args->dp, (xfs_fileoff_t)lblkno, - args->rmtblkcnt, - XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA, - args->firstblock, 0, &map, &nmap, - args->flist); - if (error) { + error = xfs_bmapi_read(args->dp, (xfs_fileoff_t)lblkno, + args->rmtblkcnt, &map, &nmap, + XFS_BMAPI_ATTRFORK); + if (error) return(error); - } ASSERT(nmap == 1); ASSERT((map.br_startblock != DELAYSTARTBLOCK) && (map.br_startblock != HOLESTARTBLOCK)); diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 58c3add07b65..d4906e7c9787 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -2926,9 +2926,8 @@ xfs_attr_leaf_freextent(xfs_trans_t **trans, xfs_inode_t *dp, * Try to remember where we decided to put the value. */ nmap = 1; - error = xfs_bmapi(*trans, dp, (xfs_fileoff_t)tblkno, tblkcnt, - XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA, - NULL, 0, &map, &nmap, NULL); + error = xfs_bmapi_read(dp, (xfs_fileoff_t)tblkno, tblkcnt, + &map, &nmap, XFS_BMAPI_ATTRFORK); if (error) { return(error); } diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 59f63949d3bc..fc5acf0b921e 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4349,6 +4349,98 @@ xfs_bmapi_update_map( *map = mval; } +/* + * Map file blocks to filesystem blocks without allocation. + */ +int +xfs_bmapi_read( + struct xfs_inode *ip, + xfs_fileoff_t bno, + xfs_filblks_t len, + struct xfs_bmbt_irec *mval, + int *nmap, + int flags) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp; + struct xfs_bmbt_irec got; + struct xfs_bmbt_irec prev; + xfs_fileoff_t obno; + xfs_fileoff_t end; + xfs_extnum_t lastx; + int error; + int eof; + int n = 0; + int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? + XFS_ATTR_FORK : XFS_DATA_FORK; + + ASSERT(*nmap >= 1); + ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE| + XFS_BMAPI_IGSTATE))); + + if (unlikely(XFS_TEST_ERROR( + (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE), + mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) { + XFS_ERROR_REPORT("xfs_bmapi_read", XFS_ERRLEVEL_LOW, mp); + return XFS_ERROR(EFSCORRUPTED); + } + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + XFS_STATS_INC(xs_blk_mapr); + + ifp = XFS_IFORK_PTR(ip, whichfork); + ASSERT(ifp->if_ext_max == + XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t)); + + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + error = xfs_iread_extents(NULL, ip, whichfork); + if (error) + return error; + } + + xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev); + end = bno + len; + obno = bno; + + while (bno < end && n < *nmap) { + /* Reading past eof, act as though there's a hole up to end. */ + if (eof) + got.br_startoff = end; + if (got.br_startoff > bno) { + /* Reading in a hole. */ + mval->br_startoff = bno; + mval->br_startblock = HOLESTARTBLOCK; + mval->br_blockcount = + XFS_FILBLKS_MIN(len, got.br_startoff - bno); + mval->br_state = XFS_EXT_NORM; + bno += mval->br_blockcount; + len -= mval->br_blockcount; + mval++; + n++; + continue; + } + + /* set up the extent map to return. */ + xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); + xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags); + + /* If we're done, stop now. */ + if (bno >= end || n >= *nmap) + break; + + /* Else go on to the next record. */ + if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got); + else + eof = 1; + } + *nmap = n; + return 0; +} + /* * Map file blocks to filesystem blocks. * File range is given by the bno/len pair. @@ -5490,10 +5582,9 @@ xfs_getbmap( do { nmap = (nexleft > subnex) ? subnex : nexleft; - error = xfs_bmapi(NULL, ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset), - XFS_BB_TO_FSB(mp, bmv->bmv_length), - bmapi_flags, NULL, 0, map, &nmap, - NULL); + error = xfs_bmapi_read(ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset), + XFS_BB_TO_FSB(mp, bmv->bmv_length), + map, &nmap, bmapi_flags); if (error) goto out_free_map; ASSERT(nmap <= subnex); @@ -6084,9 +6175,8 @@ xfs_bmap_punch_delalloc_range( * trying to remove a real extent (which requires a * transaction) or a hole, which is probably a bad idea... */ - error = xfs_bmapi(NULL, ip, start_fsb, 1, - XFS_BMAPI_ENTIRE, NULL, 0, &imap, - &nimaps, NULL); + error = xfs_bmapi_read(ip, start_fsb, 1, &imap, &nimaps, + XFS_BMAPI_ENTIRE); if (error) { /* something screwed, just bail */ diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index c62234bde053..16257c6eba71 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -294,6 +294,10 @@ xfs_bmapi( int *nmap, /* i/o: mval size/count */ xfs_bmap_free_t *flist); /* i/o: list extents to free */ +int xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno, + xfs_filblks_t len, struct xfs_bmbt_irec *mval, + int *nmap, int flags); + /* * Map file blocks to filesystem blocks, simple version. * One block only, read-only. diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index ee9d5427fcd4..44dfa39099ef 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -1995,11 +1995,10 @@ xfs_da_do_buf( } else { mapp = kmem_alloc(sizeof(*mapp) * nfsb, KM_SLEEP); nmap = nfsb; - if ((error = xfs_bmapi(trans, dp, (xfs_fileoff_t)bno, - nfsb, - XFS_BMAPI_METADATA | - xfs_bmapi_aflag(whichfork), - NULL, 0, mapp, &nmap, NULL))) + error = xfs_bmapi_read(dp, (xfs_fileoff_t)bno, nfsb, + mapp, &nmap, + xfs_bmapi_aflag(whichfork)); + if (error) goto exit0; } } else { diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c index ca2386d82cdf..66e108f561a3 100644 --- a/fs/xfs/xfs_dir2_leaf.c +++ b/fs/xfs/xfs_dir2_leaf.c @@ -888,12 +888,10 @@ xfs_dir2_leaf_getdents( * we already have in the table. */ nmap = map_size - map_valid; - error = xfs_bmapi(NULL, dp, - map_off, + error = xfs_bmapi_read(dp, map_off, xfs_dir2_byte_to_da(mp, XFS_DIR2_LEAF_OFFSET) - map_off, - XFS_BMAPI_METADATA, NULL, 0, - &map[map_valid], &nmap, NULL); + &map[map_valid], &nmap, 0); /* * Don't know if we should ignore this or * try to return an error. diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 0c5fe66ce92b..c377961657ee 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -488,9 +488,8 @@ xfs_qm_dqtobp( /* * Find the block map; no allocations yet */ - error = xfs_bmapi(NULL, quotip, dqp->q_fileoffset, - XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA, - NULL, 0, &map, &nmaps, NULL); + error = xfs_bmapi_read(quotip, dqp->q_fileoffset, + XFS_DQUOT_CLUSTER_SIZE_FSB, &map, &nmaps, 0); xfs_iunlock(quotip, XFS_ILOCK_SHARED); if (error) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 06fe97e56e48..558543c146b3 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -509,11 +509,9 @@ xfs_zero_last_block( last_fsb = XFS_B_TO_FSBT(mp, isize); nimaps = 1; - error = xfs_bmapi(NULL, ip, last_fsb, 1, 0, NULL, 0, &imap, - &nimaps, NULL); - if (error) { + error = xfs_bmapi_read(ip, last_fsb, 1, &imap, &nimaps, 0); + if (error) return error; - } ASSERT(nimaps > 0); /* * If the block underlying isize is just a hole, then there @@ -604,8 +602,8 @@ xfs_zero_eof( while (start_zero_fsb <= end_zero_fsb) { nimaps = 1; zero_count_fsb = end_zero_fsb - start_zero_fsb + 1; - error = xfs_bmapi(NULL, ip, start_zero_fsb, zero_count_fsb, - 0, NULL, 0, &imap, &nimaps, NULL); + error = xfs_bmapi_read(ip, start_zero_fsb, zero_count_fsb, + &imap, &nimaps, 0); if (error) { ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL)); return error; diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index d689253fdfda..f690d3a10b34 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1187,6 +1187,7 @@ xfs_isize_check( xfs_fileoff_t map_first; int nimaps; xfs_bmbt_irec_t imaps[2]; + int error; if (!S_ISREG(ip->i_d.di_mode)) return; @@ -1203,13 +1204,12 @@ xfs_isize_check( * The filesystem could be shutting down, so bmapi may return * an error. */ - if (xfs_bmapi(NULL, ip, map_first, + error = xfs_bmapi_read(ip, map_first, (XFS_B_TO_FSB(mp, - (xfs_ufsize_t)XFS_MAXIOFFSET(mp)) - - map_first), - XFS_BMAPI_ENTIRE, NULL, 0, imaps, &nimaps, - NULL)) - return; + (xfs_ufsize_t)XFS_MAXIOFFSET(mp)) - map_first), + imaps, &nimaps, XFS_BMAPI_ENTIRE); + if (error) + return; ASSERT(nimaps == 1); ASSERT(imaps[0].br_startblock == HOLESTARTBLOCK); } diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 091d82b94c4d..544f053860f1 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -300,8 +300,8 @@ xfs_iomap_eof_want_preallocate( while (count_fsb > 0) { imaps = nimaps; firstblock = NULLFSBLOCK; - error = xfs_bmapi(NULL, ip, start_fsb, count_fsb, 0, - &firstblock, 0, imap, &imaps, NULL); + error = xfs_bmapi_read(ip, start_fsb, count_fsb, imap, &imaps, + 0); if (error) return error; for (n = 0; n < imaps; n++) { diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index f51bef885e6d..ddaf97a57ec6 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1347,11 +1347,8 @@ xfs_qm_dqiterate( * the inode is never added to the transaction. */ xfs_ilock(qip, XFS_ILOCK_SHARED); - error = xfs_bmapi(NULL, qip, lblkno, - maxlblkcnt - lblkno, - XFS_BMAPI_METADATA, - NULL, - 0, map, &nmaps, NULL); + error = xfs_bmapi_read(qip, lblkno, maxlblkcnt - lblkno, + map, &nmaps, 0); xfs_iunlock(qip, XFS_ILOCK_SHARED); if (error) break; diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 0d1caec873a1..63874a87b378 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -72,8 +72,8 @@ xfs_readlink_bmap( xfs_buf_t *bp; int error = 0; - error = xfs_bmapi(NULL, ip, 0, XFS_B_TO_FSB(mp, pathlen), 0, NULL, 0, - mval, &nmaps, NULL); + error = xfs_bmapi_read(ip, 0, XFS_B_TO_FSB(mp, pathlen), mval, &nmaps, + 0); if (error) goto out; @@ -178,8 +178,7 @@ xfs_free_eofblocks( nimaps = 1; xfs_ilock(ip, XFS_ILOCK_SHARED); - error = xfs_bmapi(NULL, ip, end_fsb, map_len, 0, - NULL, 0, &imap, &nimaps, NULL); + error = xfs_bmapi_read(ip, end_fsb, map_len, &imap, &nimaps, 0); xfs_iunlock(ip, XFS_ILOCK_SHARED); if (!error && (nimaps != 0) && @@ -297,9 +296,9 @@ xfs_inactive_symlink_rmt( done = 0; xfs_bmap_init(&free_list, &first_block); nmaps = ARRAY_SIZE(mval); - if ((error = xfs_bmapi(tp, ip, 0, XFS_B_TO_FSB(mp, size), - XFS_BMAPI_METADATA, &first_block, 0, mval, &nmaps, - &free_list))) + error = xfs_bmapi_read(ip, 0, XFS_B_TO_FSB(mp, size), + mval, &nmaps, 0); + if (error) goto error0; /* * Invalidate the block(s). @@ -1981,8 +1980,7 @@ xfs_zero_remaining_bytes( for (offset = startoff; offset <= endoff; offset = lastoffset + 1) { offset_fsb = XFS_B_TO_FSBT(mp, offset); nimap = 1; - error = xfs_bmapi(NULL, ip, offset_fsb, 1, 0, - NULL, 0, &imap, &nimap, NULL); + error = xfs_bmapi_read(ip, offset_fsb, 1, &imap, &nimap, 0); if (error || nimap < 1) break; ASSERT(imap.br_blockcount >= 1); @@ -2101,8 +2099,8 @@ xfs_free_file_space( */ if (rt && !xfs_sb_version_hasextflgbit(&mp->m_sb)) { nimap = 1; - error = xfs_bmapi(NULL, ip, startoffset_fsb, - 1, 0, NULL, 0, &imap, &nimap, NULL); + error = xfs_bmapi_read(ip, startoffset_fsb, 1, + &imap, &nimap, 0); if (error) goto out_unlock_iolock; ASSERT(nimap == 0 || nimap == 1); @@ -2116,8 +2114,8 @@ xfs_free_file_space( startoffset_fsb += mp->m_sb.sb_rextsize - mod; } nimap = 1; - error = xfs_bmapi(NULL, ip, endoffset_fsb - 1, - 1, 0, NULL, 0, &imap, &nimap, NULL); + error = xfs_bmapi_read(ip, endoffset_fsb - 1, 1, + &imap, &nimap, 0); if (error) goto out_unlock_iolock; ASSERT(nimap == 0 || nimap == 1); From 5b777ad517ee75d3bb8d67c142d808822e46601b Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:46 +0000 Subject: [PATCH 26/69] xfs: remove xfs_bmapi_single() Now we have xfs_bmapi_read, there is no need for xfs_bmapi_single(). Change the remaining caller over and kill the function. Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 52 ------------------------------------------- fs/xfs/xfs_bmap.h | 15 ------------- fs/xfs/xfs_da_btree.c | 32 +++++++------------------- fs/xfs/xfs_rtalloc.c | 32 +++++++++----------------- 4 files changed, 19 insertions(+), 112 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index fc5acf0b921e..fd06bf17d09c 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4968,58 +4968,6 @@ xfs_bmapi( return error; } -/* - * Map file blocks to filesystem blocks, simple version. - * One block (extent) only, read-only. - * For flags, only the XFS_BMAPI_ATTRFORK flag is examined. - * For the other flag values, the effect is as if XFS_BMAPI_METADATA - * was set and all the others were clear. - */ -int /* error */ -xfs_bmapi_single( - xfs_trans_t *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode */ - int whichfork, /* data or attr fork */ - xfs_fsblock_t *fsb, /* output: mapped block */ - xfs_fileoff_t bno) /* starting file offs. mapped */ -{ - int eof; /* we've hit the end of extents */ - int error; /* error return */ - xfs_bmbt_irec_t got; /* current file extent record */ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_extnum_t lastx; /* last useful extent number */ - xfs_bmbt_irec_t prev; /* previous file extent record */ - - ifp = XFS_IFORK_PTR(ip, whichfork); - if (unlikely( - XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE && - XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)) { - XFS_ERROR_REPORT("xfs_bmapi_single", XFS_ERRLEVEL_LOW, - ip->i_mount); - return XFS_ERROR(EFSCORRUPTED); - } - if (XFS_FORCED_SHUTDOWN(ip->i_mount)) - return XFS_ERROR(EIO); - XFS_STATS_INC(xs_blk_mapr); - if (!(ifp->if_flags & XFS_IFEXTENTS) && - (error = xfs_iread_extents(tp, ip, whichfork))) - return error; - (void)xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, - &prev); - /* - * Reading past eof, act as though there's a hole - * up to end. - */ - if (eof || got.br_startoff > bno) { - *fsb = NULLFSBLOCK; - return 0; - } - ASSERT(!isnullstartblock(got.br_startblock)); - ASSERT(bno < got.br_startoff + got.br_blockcount); - *fsb = got.br_startblock + (bno - got.br_startoff); - return 0; -} - /* * Unmap (remove) blocks from a file. * If nexts is nonzero then the number of extents to remove is limited to diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 16257c6eba71..01693390df21 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -298,21 +298,6 @@ int xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, struct xfs_bmbt_irec *mval, int *nmap, int flags); -/* - * Map file blocks to filesystem blocks, simple version. - * One block only, read-only. - * For flags, only the XFS_BMAPI_ATTRFORK flag is examined. - * For the other flag values, the effect is as if XFS_BMAPI_METADATA - * was set and all the others were clear. - */ -int /* error */ -xfs_bmapi_single( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - int whichfork, /* data or attr fork */ - xfs_fsblock_t *fsb, /* output: mapped block */ - xfs_fileoff_t bno); /* starting file offs. mapped */ - /* * Unmap (remove) blocks from a file. * If nexts is nonzero then the number of extents to remove is limited to diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 44dfa39099ef..70a5f580e5ed 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -1975,32 +1975,16 @@ xfs_da_do_buf( /* * Optimize the one-block case. */ - if (nfsb == 1) { - xfs_fsblock_t fsb; - - if ((error = - xfs_bmapi_single(trans, dp, whichfork, &fsb, - (xfs_fileoff_t)bno))) { - return error; - } + if (nfsb == 1) mapp = ↦ - if (fsb == NULLFSBLOCK) { - nmap = 0; - } else { - map.br_startblock = fsb; - map.br_startoff = (xfs_fileoff_t)bno; - map.br_blockcount = 1; - nmap = 1; - } - } else { + else mapp = kmem_alloc(sizeof(*mapp) * nfsb, KM_SLEEP); - nmap = nfsb; - error = xfs_bmapi_read(dp, (xfs_fileoff_t)bno, nfsb, - mapp, &nmap, - xfs_bmapi_aflag(whichfork)); - if (error) - goto exit0; - } + + nmap = nfsb; + error = xfs_bmapi_read(dp, (xfs_fileoff_t)bno, nfsb, mapp, + &nmap, xfs_bmapi_aflag(whichfork)); + if (error) + goto exit0; } else { map.br_startblock = XFS_DADDR_TO_FSB(mp, mappedbno); map.br_startoff = (xfs_fileoff_t)bno; diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index 35561a511b57..e5f40c6460b2 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -856,33 +856,23 @@ xfs_rtbuf_get( xfs_buf_t **bpp) /* output: buffer for the block */ { xfs_buf_t *bp; /* block buffer, result */ - xfs_daddr_t d; /* disk addr of block */ - int error; /* error value */ - xfs_fsblock_t fsb; /* fs block number for block */ xfs_inode_t *ip; /* bitmap or summary inode */ + xfs_bmbt_irec_t map; + int nmap; + int error; /* error value */ ip = issum ? mp->m_rsumip : mp->m_rbmip; - /* - * Map from the file offset (block) and inode number to the - * file system block. - */ - error = xfs_bmapi_single(tp, ip, XFS_DATA_FORK, &fsb, block); - if (error) { + + error = xfs_bmapi_read(ip, block, 1, &map, &nmap, XFS_DATA_FORK); + if (error) return error; - } - ASSERT(fsb != NULLFSBLOCK); - /* - * Convert to disk address for buffer cache. - */ - d = XFS_FSB_TO_DADDR(mp, fsb); - /* - * Read the buffer. - */ - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d, + + ASSERT(map.br_startblock != NULLFSBLOCK); + error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, + XFS_FSB_TO_DADDR(mp, map.br_startblock), mp->m_bsize, 0, &bp); - if (error) { + if (error) return error; - } ASSERT(!xfs_buf_geterror(bp)); *bpp = bp; return 0; From b64dfe4e180ab5047c59bcbe379538eb23be4d8e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:47 +0000 Subject: [PATCH 27/69] xfs: factor delalloc reservations out of xfs_bmapi Move the reservation of delayed allocations, and addition of delalloc regions to the extent trees into a new helper function. For now this adds some twisted goto logic to xfs_bmapi, but that will be cleaned up in the following patches. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 202 +++++++++++++++++++++++++++------------------- 1 file changed, 118 insertions(+), 84 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index fd06bf17d09c..cd01783a89d7 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4441,6 +4441,120 @@ xfs_bmapi_read( return 0; } +STATIC int +xfs_bmapi_reserve_delalloc( + struct xfs_inode *ip, + xfs_fileoff_t aoff, + xfs_filblks_t len, + struct xfs_bmbt_irec *got, + struct xfs_bmbt_irec *prev, + xfs_extnum_t *lastx, + int eof) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + xfs_extlen_t alen; + xfs_extlen_t indlen; + xfs_fsblock_t firstblock = NULLFSBLOCK; + struct xfs_btree_cur *cur = NULL; + int tmp_logflags = 0; + char rt = XFS_IS_REALTIME_INODE(ip); + xfs_extlen_t extsz; + int error; + + alen = XFS_FILBLKS_MIN(len, MAXEXTLEN); + if (!eof) + alen = XFS_FILBLKS_MIN(alen, got->br_startoff - aoff); + + /* Figure out the extent size, adjust alen */ + extsz = xfs_get_extsz_hint(ip); + if (extsz) { + /* + * Make sure we don't exceed a single extent length when we + * align the extent by reducing length we are going to + * allocate by the maximum amount extent size aligment may + * require. + */ + alen = XFS_FILBLKS_MIN(len, MAXEXTLEN - (2 * extsz - 1)); + error = xfs_bmap_extsize_align(mp, got, prev, extsz, rt, eof, + 1, 0, &aoff, &alen); + ASSERT(!error); + } + + if (rt) + extsz = alen / mp->m_sb.sb_rextsize; + + /* + * Make a transaction-less quota reservation for delayed allocation + * blocks. This number gets adjusted later. We return if we haven't + * allocated blocks already inside this loop. + */ + error = xfs_trans_reserve_quota_nblks(NULL, ip, (long)alen, 0, + rt ? XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS); + if (error) + return error; + + /* + * Split changing sb for alen and indlen since they could be coming + * from different places. + */ + indlen = (xfs_extlen_t)xfs_bmap_worst_indlen(ip, alen); + ASSERT(indlen > 0); + + if (rt) { + error = xfs_mod_incore_sb(mp, XFS_SBS_FREXTENTS, + -((int64_t)extsz), 0); + } else { + error = xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS, + -((int64_t)alen), 0); + } + + if (error) + goto out_unreserve_quota; + + error = xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS, + -((int64_t)indlen), 0); + if (error) + goto out_unreserve_blocks; + + + ip->i_delayed_blks += alen; + + got->br_startoff = aoff; + got->br_startblock = nullstartblock(indlen); + got->br_blockcount = alen; + got->br_state = XFS_EXT_NORM; + + error = xfs_bmap_add_extent(NULL, ip, lastx, &cur, got, &firstblock, + NULL, &tmp_logflags, XFS_DATA_FORK); + ASSERT(!error); + ASSERT(!tmp_logflags); + ASSERT(!cur); + + /* + * Update our extent pointer, given that xfs_bmap_add_extent might + * have merged it into one of the neighbouring ones. + */ + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), got); + + ASSERT(got->br_startoff <= aoff); + ASSERT(got->br_startoff + got->br_blockcount >= aoff + alen); + ASSERT(isnullstartblock(got->br_startblock)); + ASSERT(got->br_state == XFS_EXT_NORM); + return 0; + +out_unreserve_blocks: + if (rt) + xfs_mod_incore_sb(mp, XFS_SBS_FREXTENTS, extsz, 0); + else + xfs_icsb_modify_counters(mp, XFS_SBS_FDBLOCKS, alen, 0); +out_unreserve_quota: + if (XFS_IS_QUOTA_ON(mp)) + xfs_trans_unreserve_quota_nblks(NULL, ip, alen, 0, rt ? + XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS); + return error; +} + /* * Map file blocks to filesystem blocks. * File range is given by the bno/len pair. @@ -4479,7 +4593,6 @@ xfs_bmapi( int error; /* error return */ xfs_bmbt_irec_t got; /* current file extent record */ xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_extlen_t indlen; /* indirect blocks length */ xfs_extnum_t lastx; /* last useful extent number */ int logflags; /* flags for transaction logging */ xfs_extlen_t minleft; /* min blocks left after allocation */ @@ -4615,43 +4728,8 @@ xfs_bmapi( } minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; if (flags & XFS_BMAPI_DELAY) { - xfs_extlen_t extsz; - - /* Figure out the extent size, adjust alen */ - extsz = xfs_get_extsz_hint(ip); - if (extsz) { - /* - * make sure we don't exceed a single - * extent length when we align the - * extent by reducing length we are - * going to allocate by the maximum - * amount extent size aligment may - * require. - */ - alen = XFS_FILBLKS_MIN(len, - MAXEXTLEN - (2 * extsz - 1)); - error = xfs_bmap_extsize_align(mp, - &got, &prev, extsz, - rt, eof, - flags&XFS_BMAPI_DELAY, - flags&XFS_BMAPI_CONVERT, - &aoff, &alen); - ASSERT(!error); - } - - if (rt) - extsz = alen / mp->m_sb.sb_rextsize; - - /* - * Make a transaction-less quota reservation for - * delayed allocation blocks. This number gets - * adjusted later. We return if we haven't - * allocated blocks already inside this loop. - */ - error = xfs_trans_reserve_quota_nblks( - NULL, ip, (long)alen, 0, - rt ? XFS_QMOPT_RES_RTBLKS : - XFS_QMOPT_RES_REGBLKS); + error = xfs_bmapi_reserve_delalloc(ip, bno, len, &got, + &prev, &lastx, eof); if (error) { if (n == 0) { *nmap = 0; @@ -4661,51 +4739,7 @@ xfs_bmapi( break; } - /* - * Split changing sb for alen and indlen since - * they could be coming from different places. - */ - indlen = (xfs_extlen_t) - xfs_bmap_worst_indlen(ip, alen); - ASSERT(indlen > 0); - - if (rt) { - error = xfs_mod_incore_sb(mp, - XFS_SBS_FREXTENTS, - -((int64_t)extsz), 0); - } else { - error = xfs_icsb_modify_counters(mp, - XFS_SBS_FDBLOCKS, - -((int64_t)alen), 0); - } - if (!error) { - error = xfs_icsb_modify_counters(mp, - XFS_SBS_FDBLOCKS, - -((int64_t)indlen), 0); - if (error && rt) - xfs_mod_incore_sb(mp, - XFS_SBS_FREXTENTS, - (int64_t)extsz, 0); - else if (error) - xfs_icsb_modify_counters(mp, - XFS_SBS_FDBLOCKS, - (int64_t)alen, 0); - } - - if (error) { - if (XFS_IS_QUOTA_ON(mp)) - /* unreserve the blocks now */ - (void) - xfs_trans_unreserve_quota_nblks( - NULL, ip, - (long)alen, 0, rt ? - XFS_QMOPT_RES_RTBLKS : - XFS_QMOPT_RES_REGBLKS); - break; - } - - ip->i_delayed_blks += alen; - abno = nullstartblock(indlen); + goto trim_extent; } else { /* * If first time, allocate and fill in @@ -4843,7 +4877,7 @@ xfs_bmapi( n++; continue; } - +trim_extent: /* Deal with the allocated space we found. */ xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); From 4403280aa5c00c6074f2dc23e1cfc11f2bfb0032 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:48 +0000 Subject: [PATCH 28/69] xfs: introduce xfs_bmapi_delay() Delalloc reservations are much simpler than allocations, so give them a separate bmapi-level interface. Using the previously added xfs_bmapi_reserve_delalloc we get a function that is only minimally more complicated than xfs_bmapi_read, which is far from the complexity in xfs_bmapi. Also remove the XFS_BMAPI_DELAY code after switching over the only user to xfs_bmapi_delay. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 115 +++++++++++++++++++++++++++++++++++---------- fs/xfs/xfs_bmap.h | 5 +- fs/xfs/xfs_iomap.c | 9 +--- 3 files changed, 94 insertions(+), 35 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index cd01783a89d7..6f3276974555 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4238,7 +4238,7 @@ xfs_bmap_validate_ret( ASSERT(i == 0 || mval[i - 1].br_startoff + mval[i - 1].br_blockcount == mval[i].br_startoff); - if ((flags & XFS_BMAPI_WRITE) && !(flags & XFS_BMAPI_DELAY)) + if (flags & XFS_BMAPI_WRITE) ASSERT(mval[i].br_startblock != DELAYSTARTBLOCK && mval[i].br_startblock != HOLESTARTBLOCK); ASSERT(mval[i].br_state == XFS_EXT_NORM || @@ -4555,6 +4555,90 @@ xfs_bmapi_reserve_delalloc( return error; } +/* + * Map file blocks to filesystem blocks, adding delayed allocations as needed. + */ +int +xfs_bmapi_delay( + struct xfs_inode *ip, /* incore inode */ + xfs_fileoff_t bno, /* starting file offs. mapped */ + xfs_filblks_t len, /* length to map in file */ + struct xfs_bmbt_irec *mval, /* output: map values */ + int *nmap, /* i/o: mval size/count */ + int flags) /* XFS_BMAPI_... */ +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + struct xfs_bmbt_irec got; /* current file extent record */ + struct xfs_bmbt_irec prev; /* previous file extent record */ + xfs_fileoff_t obno; /* old block number (offset) */ + xfs_fileoff_t end; /* end of mapped file region */ + xfs_extnum_t lastx; /* last useful extent number */ + int eof; /* we've hit the end of extents */ + int n = 0; /* current extent index */ + int error = 0; + + ASSERT(*nmap >= 1); + ASSERT(*nmap <= XFS_BMAP_MAX_NMAP); + ASSERT(!(flags & ~XFS_BMAPI_ENTIRE)); + + if (unlikely(XFS_TEST_ERROR( + (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_BTREE), + mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) { + XFS_ERROR_REPORT("xfs_bmapi_delay", XFS_ERRLEVEL_LOW, mp); + return XFS_ERROR(EFSCORRUPTED); + } + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + XFS_STATS_INC(xs_blk_mapw); + + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK); + if (error) + return error; + } + + xfs_bmap_search_extents(ip, bno, XFS_DATA_FORK, &eof, &lastx, &got, &prev); + end = bno + len; + obno = bno; + + while (bno < end && n < *nmap) { + if (eof || got.br_startoff > bno) { + error = xfs_bmapi_reserve_delalloc(ip, bno, len, &got, + &prev, &lastx, eof); + if (error) { + if (n == 0) { + *nmap = 0; + return error; + } + break; + } + } + + /* set up the extent map to return. */ + xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); + xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags); + + /* If we're done, stop now. */ + if (bno >= end || n >= *nmap) + break; + + /* Else go on to the next record. */ + prev = got; + if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got); + else + eof = 1; + } + + *nmap = n; + return 0; +} + + /* * Map file blocks to filesystem blocks. * File range is given by the bno/len pair. @@ -4663,7 +4747,6 @@ xfs_bmapi( */ if ((flags & XFS_BMAPI_IGSTATE) && wr) /* if writing unwritten space */ wr = 0; /* no allocations are allowed */ - ASSERT(wr || !(flags & XFS_BMAPI_DELAY)); logflags = 0; nallocs = 0; cur = NULL; @@ -4698,8 +4781,7 @@ xfs_bmapi( if (eof && !wr) got.br_startoff = end; inhole = eof || got.br_startoff > bno; - wasdelay = wr && !inhole && !(flags & XFS_BMAPI_DELAY) && - isnullstartblock(got.br_startblock); + wasdelay = wr && !inhole && isnullstartblock(got.br_startblock); /* * First, deal with the hole before the allocated space * that we found, if any. @@ -4727,20 +4809,7 @@ xfs_bmapi( aoff = bno; } minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; - if (flags & XFS_BMAPI_DELAY) { - error = xfs_bmapi_reserve_delalloc(ip, bno, len, &got, - &prev, &lastx, eof); - if (error) { - if (n == 0) { - *nmap = 0; - ASSERT(cur == NULL); - return error; - } - break; - } - - goto trim_extent; - } else { + { /* * If first time, allocate and fill in * once-only bma fields. @@ -4851,14 +4920,8 @@ xfs_bmapi( ASSERT(got.br_startoff <= aoff); ASSERT(got.br_startoff + got.br_blockcount >= aoff + alen); -#ifdef DEBUG - if (flags & XFS_BMAPI_DELAY) { - ASSERT(isnullstartblock(got.br_startblock)); - ASSERT(startblockval(got.br_startblock) > 0); - } ASSERT(got.br_state == XFS_EXT_NORM || got.br_state == XFS_EXT_UNWRITTEN); -#endif /* * Fall down into the found allocated space case. */ @@ -4877,7 +4940,7 @@ xfs_bmapi( n++; continue; } -trim_extent: + /* Deal with the allocated space we found. */ xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); @@ -4887,7 +4950,7 @@ xfs_bmapi( */ if (wr && ((mval->br_state == XFS_EXT_UNWRITTEN && - ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_DELAY)) == 0)) || + ((flags & XFS_BMAPI_PREALLOC) == 0)) || (mval->br_state == XFS_EXT_NORM && ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_CONVERT)) == (XFS_BMAPI_PREALLOC|XFS_BMAPI_CONVERT))))) { diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 01693390df21..c70c19c7f1fa 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -65,7 +65,6 @@ typedef struct xfs_bmap_free * Flags for xfs_bmapi */ #define XFS_BMAPI_WRITE 0x001 /* write operation: allocate space */ -#define XFS_BMAPI_DELAY 0x002 /* delayed write operation */ #define XFS_BMAPI_ENTIRE 0x004 /* return entire extent, not trimmed */ #define XFS_BMAPI_METADATA 0x008 /* mapping metadata not user data */ #define XFS_BMAPI_ATTRFORK 0x010 /* use attribute fork not data */ @@ -82,7 +81,6 @@ typedef struct xfs_bmap_free #define XFS_BMAPI_FLAGS \ { XFS_BMAPI_WRITE, "WRITE" }, \ - { XFS_BMAPI_DELAY, "DELAY" }, \ { XFS_BMAPI_ENTIRE, "ENTIRE" }, \ { XFS_BMAPI_METADATA, "METADATA" }, \ { XFS_BMAPI_ATTRFORK, "ATTRFORK" }, \ @@ -297,6 +295,9 @@ xfs_bmapi( int xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, struct xfs_bmbt_irec *mval, int *nmap, int flags); +int xfs_bmapi_delay(struct xfs_inode *ip, xfs_fileoff_t bno, + xfs_filblks_t len, struct xfs_bmbt_irec *mval, + int *nmap, int flags); /* * Unmap (remove) blocks from a file. diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 544f053860f1..681ba34c9233 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -381,7 +381,6 @@ xfs_iomap_write_delay( xfs_fileoff_t last_fsb; xfs_off_t aligned_offset; xfs_fileoff_t ioalign; - xfs_fsblock_t firstblock; xfs_extlen_t extsz; int nimaps; xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS]; @@ -425,12 +424,8 @@ xfs_iomap_write_delay( } nimaps = XFS_WRITE_IMAPS; - firstblock = NULLFSBLOCK; - error = xfs_bmapi(NULL, ip, offset_fsb, - (xfs_filblks_t)(last_fsb - offset_fsb), - XFS_BMAPI_DELAY | XFS_BMAPI_WRITE | - XFS_BMAPI_ENTIRE, &firstblock, 1, imap, - &nimaps, NULL); + error = xfs_bmapi_delay(ip, offset_fsb, last_fsb - offset_fsb, + imap, &nimaps, XFS_BMAPI_ENTIRE); switch (error) { case 0: case ENOSPC: From 1fd044d9c6735e669f0db025f18023e56a608130 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:49 +0000 Subject: [PATCH 29/69] xfs: do not use xfs_bmap_add_extent for adding delalloc extents We can just call xfs_bmap_add_extent_hole_delay directly to add a delayed allocated regions to the extent tree, instead of going through all the complexities of xfs_bmap_add_extent that aren't needed for this simple case. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 52 ++++++++--------------------------------------- 1 file changed, 9 insertions(+), 43 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 6f3276974555..c9492e23447c 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -100,17 +100,6 @@ xfs_bmap_add_extent_delay_real( xfs_bmap_free_t *flist, /* list of extents to be freed */ int *logflagsp); /* inode logging flags */ -/* - * Called by xfs_bmap_add_extent to handle cases converting a hole - * to a delayed allocation. - */ -STATIC int /* error */ -xfs_bmap_add_extent_hole_delay( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - int *logflagsp); /* inode logging flags */ - /* * Called by xfs_bmap_add_extent to handle cases converting a hole * to a real allocation. @@ -431,8 +420,7 @@ xfs_bmap_add_attrfork_local( } /* - * Called by xfs_bmapi to update file extent records and the btree - * after allocating space (or doing a delayed allocation). + * Update file extent records and the btree after allocating space. */ STATIC int /* error */ xfs_bmap_add_extent( @@ -464,21 +452,12 @@ xfs_bmap_add_extent( ASSERT(*idx >= 0); ASSERT(*idx <= nextents); + ASSERT(!isnullstartblock(new->br_startblock)); - /* - * Any kind of new delayed allocation goes here. - */ - if (isnullstartblock(new->br_startblock)) { - if (cur) - ASSERT((cur->bc_private.b.flags & - XFS_BTCUR_BPRV_WASDEL) == 0); - error = xfs_bmap_add_extent_hole_delay(ip, idx, new, - &logflags); - } /* * Real allocation off the end of the file. */ - else if (*idx == nextents) { + if (*idx == nextents) { if (cur) ASSERT((cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL) == 0); @@ -1581,16 +1560,13 @@ xfs_bmap_add_extent_unwritten_real( } /* - * Called by xfs_bmap_add_extent to handle cases converting a hole - * to a delayed allocation. + * Convert a hole to a delayed allocation. */ -/*ARGSUSED*/ -STATIC int /* error */ +STATIC void xfs_bmap_add_extent_hole_delay( xfs_inode_t *ip, /* incore inode pointer */ xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - int *logflagsp) /* inode logging flags */ + xfs_bmbt_irec_t *new) /* new data to add to file extents */ { xfs_ifork_t *ifp; /* inode fork pointer */ xfs_bmbt_irec_t left; /* left neighbor extent entry */ @@ -1725,8 +1701,6 @@ xfs_bmap_add_extent_hole_delay( * Nothing to do for disk quota accounting here. */ } - *logflagsp = 0; - return 0; } /* @@ -4455,9 +4429,6 @@ xfs_bmapi_reserve_delalloc( struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); xfs_extlen_t alen; xfs_extlen_t indlen; - xfs_fsblock_t firstblock = NULLFSBLOCK; - struct xfs_btree_cur *cur = NULL; - int tmp_logflags = 0; char rt = XFS_IS_REALTIME_INODE(ip); xfs_extlen_t extsz; int error; @@ -4524,16 +4495,11 @@ xfs_bmapi_reserve_delalloc( got->br_startblock = nullstartblock(indlen); got->br_blockcount = alen; got->br_state = XFS_EXT_NORM; - - error = xfs_bmap_add_extent(NULL, ip, lastx, &cur, got, &firstblock, - NULL, &tmp_logflags, XFS_DATA_FORK); - ASSERT(!error); - ASSERT(!tmp_logflags); - ASSERT(!cur); + xfs_bmap_add_extent_hole_delay(ip, lastx, got); /* - * Update our extent pointer, given that xfs_bmap_add_extent might - * have merged it into one of the neighbouring ones. + * Update our extent pointer, given that xfs_bmap_add_extent_hole_delay + * might have merged it into one of the neighbouring ones. */ xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), got); From 7e47a4efde33aa3f0cb901e086a75751c2269f04 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:50 +0000 Subject: [PATCH 30/69] xfs: factor extent allocation out of xfs_bmapi To further improve the readability of xfs_bmapi(), factor the extent allocation out into a separate function. This removes a large block of logic from the xfs_bmapi() code loop and makes it easier to see the operational logic flow for xfs_bmapi(). Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 302 +++++++++++++++++++++++++--------------------- 1 file changed, 162 insertions(+), 140 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index c9492e23447c..311cbc1d64c5 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4605,6 +4605,147 @@ xfs_bmapi_delay( } +STATIC int +xfs_bmapi_allocate( + struct xfs_bmalloca *bma, + xfs_extnum_t *lastx, + struct xfs_btree_cur **cur, + xfs_fsblock_t *firstblock, + struct xfs_bmap_free *flist, + int flags, + int *nallocs, + int *logflags) +{ + struct xfs_mount *mp = bma->ip->i_mount; + int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? + XFS_ATTR_FORK : XFS_DATA_FORK; + struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); + xfs_fsblock_t abno; + xfs_extlen_t alen; + xfs_fileoff_t aoff; + int error; + int rt; + + rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(bma->ip); + + /* + * For the wasdelay case, we could also just allocate the stuff asked + * for in this bmap call but that wouldn't be as good. + */ + if (bma->wasdel) { + alen = (xfs_extlen_t)bma->gotp->br_blockcount; + aoff = bma->gotp->br_startoff; + if (*lastx != NULLEXTNUM && *lastx) { + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx - 1), + bma->prevp); + } + } else { + alen = (xfs_extlen_t)XFS_FILBLKS_MIN(bma->alen, MAXEXTLEN); + if (!bma->eof) + alen = (xfs_extlen_t)XFS_FILBLKS_MIN(alen, + bma->gotp->br_startoff - bma->off); + aoff = bma->off; + } + + /* + * Indicate if this is the first user data in the file, or just any + * user data. + */ + if (!(flags & XFS_BMAPI_METADATA)) { + bma->userdata = (aoff == 0) ? + XFS_ALLOC_INITIAL_USER_DATA : XFS_ALLOC_USERDATA; + } + + /* + * Fill in changeable bma fields. + */ + bma->alen = alen; + bma->off = aoff; + bma->firstblock = *firstblock; + bma->minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; + bma->low = flist->xbf_low; + bma->aeof = 0; + + /* + * Only want to do the alignment at the eof if it is userdata and + * allocation length is larger than a stripe unit. + */ + if (mp->m_dalign && alen >= mp->m_dalign && + !(flags & XFS_BMAPI_METADATA) && whichfork == XFS_DATA_FORK) { + error = xfs_bmap_isaeof(bma->ip, aoff, whichfork, &bma->aeof); + if (error) + return error; + } + + error = xfs_bmap_alloc(bma); + if (error) + return error; + + /* + * Copy out result fields. + */ + abno = bma->rval; + flist->xbf_low = bma->low; + alen = bma->alen; + aoff = bma->off; + ASSERT(*firstblock == NULLFSBLOCK || + XFS_FSB_TO_AGNO(mp, *firstblock) == + XFS_FSB_TO_AGNO(mp, bma->firstblock) || + (flist->xbf_low && + XFS_FSB_TO_AGNO(mp, *firstblock) < + XFS_FSB_TO_AGNO(mp, bma->firstblock))); + *firstblock = bma->firstblock; + if (*cur) + (*cur)->bc_private.b.firstblock = *firstblock; + if (abno == NULLFSBLOCK) + return 0; + if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { + (*cur) = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork); + (*cur)->bc_private.b.firstblock = *firstblock; + (*cur)->bc_private.b.flist = flist; + } + /* + * Bump the number of extents we've allocated + * in this call. + */ + (*nallocs)++; + + if (*cur) + (*cur)->bc_private.b.flags = + bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0; + + bma->gotp->br_startoff = aoff; + bma->gotp->br_startblock = abno; + bma->gotp->br_blockcount = alen; + bma->gotp->br_state = XFS_EXT_NORM; + + /* + * A wasdelay extent has been initialized, so shouldn't be flagged + * as unwritten. + */ + if (!bma->wasdel && (flags & XFS_BMAPI_PREALLOC) && + xfs_sb_version_hasextflgbit(&mp->m_sb)) + bma->gotp->br_state = XFS_EXT_UNWRITTEN; + + error = xfs_bmap_add_extent(bma->tp, bma->ip, lastx, cur, bma->gotp, + firstblock, flist, logflags, whichfork); + if (error) + return error; + + /* + * Update our extent pointer, given that xfs_bmap_add_extent might + * have merged it into one of the neighbouring ones. + */ + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), bma->gotp); + + ASSERT(bma->gotp->br_startoff <= aoff); + ASSERT(bma->gotp->br_startoff + bma->gotp->br_blockcount >= + aoff + alen); + ASSERT(bma->gotp->br_state == XFS_EXT_NORM || + bma->gotp->br_state == XFS_EXT_UNWRITTEN); + return 0; +} + /* * Map file blocks to filesystem blocks. * File range is given by the bno/len pair. @@ -4632,9 +4773,6 @@ xfs_bmapi( int *nmap, /* i/o: mval size/count */ xfs_bmap_free_t *flist) /* i/o: list extents to free */ { - xfs_fsblock_t abno; /* allocated block number */ - xfs_extlen_t alen; /* allocated extent length */ - xfs_fileoff_t aoff; /* allocated file offset */ xfs_bmalloca_t bma = { 0 }; /* args for xfs_bmap_alloc */ xfs_btree_cur_t *cur; /* bmap btree cursor */ xfs_fileoff_t end; /* end of mapped file region */ @@ -4646,7 +4784,6 @@ xfs_bmapi( xfs_extnum_t lastx; /* last useful extent number */ int logflags; /* flags for transaction logging */ xfs_extlen_t minleft; /* min blocks left after allocation */ - xfs_extlen_t minlen; /* min allocation size */ xfs_mount_t *mp; /* xfs mount structure */ int n; /* current extent index */ int nallocs; /* number of extents alloc'd */ @@ -4737,7 +4874,13 @@ xfs_bmapi( n = 0; end = bno + len; obno = bno; - bma.ip = NULL; + + bma.tp = tp; + bma.ip = ip; + bma.prevp = &prev; + bma.gotp = &got; + bma.total = total; + bma.userdata = 0; while (bno < end && n < *nmap) { /* @@ -4753,144 +4896,23 @@ xfs_bmapi( * that we found, if any. */ if (wr && (inhole || wasdelay)) { - /* - * For the wasdelay case, we could also just - * allocate the stuff asked for in this bmap call - * but that wouldn't be as good. - */ - if (wasdelay) { - alen = (xfs_extlen_t)got.br_blockcount; - aoff = got.br_startoff; - if (lastx != NULLEXTNUM && lastx) { - ep = xfs_iext_get_ext(ifp, lastx - 1); - xfs_bmbt_get_all(ep, &prev); - } - } else { - alen = (xfs_extlen_t) - XFS_FILBLKS_MIN(len, MAXEXTLEN); - if (!eof) - alen = (xfs_extlen_t) - XFS_FILBLKS_MIN(alen, - got.br_startoff - bno); - aoff = bno; - } - minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; - { - /* - * If first time, allocate and fill in - * once-only bma fields. - */ - if (bma.ip == NULL) { - bma.tp = tp; - bma.ip = ip; - bma.prevp = &prev; - bma.gotp = &got; - bma.total = total; - bma.userdata = 0; - } - /* Indicate if this is the first user data - * in the file, or just any user data. - */ - if (!(flags & XFS_BMAPI_METADATA)) { - bma.userdata = (aoff == 0) ? - XFS_ALLOC_INITIAL_USER_DATA : - XFS_ALLOC_USERDATA; - } - /* - * Fill in changeable bma fields. - */ - bma.eof = eof; - bma.firstblock = *firstblock; - bma.alen = alen; - bma.off = aoff; - bma.conv = !!(flags & XFS_BMAPI_CONVERT); - bma.wasdel = wasdelay; - bma.minlen = minlen; - bma.low = flist->xbf_low; - bma.minleft = minleft; - /* - * Only want to do the alignment at the - * eof if it is userdata and allocation length - * is larger than a stripe unit. - */ - if (mp->m_dalign && alen >= mp->m_dalign && - (!(flags & XFS_BMAPI_METADATA)) && - (whichfork == XFS_DATA_FORK)) { - if ((error = xfs_bmap_isaeof(ip, aoff, - whichfork, &bma.aeof))) - goto error0; - } else - bma.aeof = 0; - /* - * Call allocator. - */ - if ((error = xfs_bmap_alloc(&bma))) - goto error0; - /* - * Copy out result fields. - */ - abno = bma.rval; - if ((flist->xbf_low = bma.low)) - minleft = 0; - alen = bma.alen; - aoff = bma.off; - ASSERT(*firstblock == NULLFSBLOCK || - XFS_FSB_TO_AGNO(mp, *firstblock) == - XFS_FSB_TO_AGNO(mp, bma.firstblock) || - (flist->xbf_low && - XFS_FSB_TO_AGNO(mp, *firstblock) < - XFS_FSB_TO_AGNO(mp, bma.firstblock))); - *firstblock = bma.firstblock; - if (cur) - cur->bc_private.b.firstblock = - *firstblock; - if (abno == NULLFSBLOCK) - break; - if ((ifp->if_flags & XFS_IFBROOT) && !cur) { - cur = xfs_bmbt_init_cursor(mp, tp, - ip, whichfork); - cur->bc_private.b.firstblock = - *firstblock; - cur->bc_private.b.flist = flist; - } - /* - * Bump the number of extents we've allocated - * in this call. - */ - nallocs++; - } - if (cur) - cur->bc_private.b.flags = - wasdelay ? XFS_BTCUR_BPRV_WASDEL : 0; - got.br_startoff = aoff; - got.br_startblock = abno; - got.br_blockcount = alen; - got.br_state = XFS_EXT_NORM; /* assume normal */ - /* - * Determine state of extent, and the filesystem. - * A wasdelay extent has been initialized, so - * shouldn't be flagged as unwritten. - */ - if (wr && xfs_sb_version_hasextflgbit(&mp->m_sb)) { - if (!wasdelay && (flags & XFS_BMAPI_PREALLOC)) - got.br_state = XFS_EXT_UNWRITTEN; - } - error = xfs_bmap_add_extent(tp, ip, &lastx, &cur, &got, - firstblock, flist, &tmp_logflags, - whichfork); + bma.eof = eof; + bma.conv = !!(flags & XFS_BMAPI_CONVERT); + bma.wasdel = wasdelay; + bma.alen = len; + bma.off = bno; + bma.minleft = minleft; + + error = xfs_bmapi_allocate(&bma, &lastx, &cur, + firstblock, flist, flags, &nallocs, + &tmp_logflags); logflags |= tmp_logflags; if (error) goto error0; - ep = xfs_iext_get_ext(ifp, lastx); - xfs_bmbt_get_all(ep, &got); - ASSERT(got.br_startoff <= aoff); - ASSERT(got.br_startoff + got.br_blockcount >= - aoff + alen); - ASSERT(got.br_state == XFS_EXT_NORM || - got.br_state == XFS_EXT_UNWRITTEN); - /* - * Fall down into the found allocated space case. - */ + if (flist && flist->xbf_low) + minleft = 0; + if (bma.rval == NULLFSBLOCK) + break; } else if (inhole) { /* * Reading in a hole. From b447fe5a05cbd01c4bf7fe2fa41cb9e99ce7e58e Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:51 +0000 Subject: [PATCH 31/69] xfs: factor unwritten extent map manipulations out of xfs_bmapi To further improve the readability of xfs_bmapi(), factor the unwritten extent conversion out into a separate function. This removes large block of logic from the xfs_bmapi() code loop and makes it easier to see the operational logic flow for xfs_bmapi(). Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 107 ++++++++++++++++++++++++++++++---------------- 1 file changed, 70 insertions(+), 37 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 311cbc1d64c5..c2e49fd18bfd 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4089,7 +4089,6 @@ xfs_bmap_read_extents( xfs_extnum_t num_recs; xfs_extnum_t start; - num_recs = xfs_btree_get_numrecs(block); if (unlikely(i + num_recs > room)) { ASSERT(i + num_recs <= room); @@ -4746,6 +4745,69 @@ xfs_bmapi_allocate( return 0; } +STATIC int +xfs_bmapi_convert_unwritten( + struct xfs_bmalloca *bma, + struct xfs_bmbt_irec *mval, + xfs_filblks_t len, + xfs_extnum_t *lastx, + struct xfs_btree_cur **cur, + xfs_fsblock_t *firstblock, + struct xfs_bmap_free *flist, + int flags, + int *logflags) +{ + int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? + XFS_ATTR_FORK : XFS_DATA_FORK; + struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); + int error; + + *logflags = 0; + + /* check if we need to do unwritten->real conversion */ + if (mval->br_state == XFS_EXT_UNWRITTEN && + (flags & XFS_BMAPI_PREALLOC)) + return 0; + + /* check if we need to do real->unwritten conversion */ + if (mval->br_state == XFS_EXT_NORM && + (flags & (XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT)) != + (XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT)) + return 0; + + /* + * Modify (by adding) the state flag, if writing. + */ + ASSERT(mval->br_blockcount <= len); + if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { + *cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp, + bma->ip, whichfork); + (*cur)->bc_private.b.firstblock = *firstblock; + (*cur)->bc_private.b.flist = flist; + } + mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) + ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; + + error = xfs_bmap_add_extent(bma->tp, bma->ip, lastx, cur, mval, + firstblock, flist, logflags, whichfork); + if (error) + return error; + + /* + * Update our extent pointer, given that xfs_bmap_add_extent might + * have merged it into one of the neighbouring ones. + */ + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), bma->gotp); + + /* + * We may have combined previously unwritten space with written space, + * so generate another request. + */ + if (mval->br_blockcount < len) + return EAGAIN; + return 0; +} + /* * Map file blocks to filesystem blocks. * File range is given by the bno/len pair. @@ -4932,45 +4994,16 @@ xfs_bmapi( /* Deal with the allocated space we found. */ xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); - /* - * Check if writing previously allocated but - * unwritten extents. - */ - if (wr && - ((mval->br_state == XFS_EXT_UNWRITTEN && - ((flags & XFS_BMAPI_PREALLOC) == 0)) || - (mval->br_state == XFS_EXT_NORM && - ((flags & (XFS_BMAPI_PREALLOC|XFS_BMAPI_CONVERT)) == - (XFS_BMAPI_PREALLOC|XFS_BMAPI_CONVERT))))) { - /* - * Modify (by adding) the state flag, if writing. - */ - ASSERT(mval->br_blockcount <= len); - if ((ifp->if_flags & XFS_IFBROOT) && !cur) { - cur = xfs_bmbt_init_cursor(mp, - tp, ip, whichfork); - cur->bc_private.b.firstblock = - *firstblock; - cur->bc_private.b.flist = flist; - } - mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) - ? XFS_EXT_NORM - : XFS_EXT_UNWRITTEN; - error = xfs_bmap_add_extent(tp, ip, &lastx, &cur, mval, - firstblock, flist, &tmp_logflags, - whichfork); + /* Execute unwritten extent conversion if necessary */ + if (wr) { + error = xfs_bmapi_convert_unwritten(&bma, mval, len, + &lastx, &cur, firstblock, flist, flags, + &tmp_logflags); logflags |= tmp_logflags; + if (error == EAGAIN) + continue; if (error) goto error0; - ep = xfs_iext_get_ext(ifp, lastx); - xfs_bmbt_get_all(ep, &got); - /* - * We may have combined previously unwritten - * space with written space, so generate - * another request. - */ - if (mval->br_blockcount < len) - continue; } /* update the extent map to return */ From c0dc7828af6952643219292be29e482ef74cb261 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:52 +0000 Subject: [PATCH 32/69] xfs: rename xfs_bmapi to xfs_bmapi_write Now that all the read-only users of xfs_bmapi have been converted to use xfs_bmapi_read(), we can remove all the read-only handling cases from xfs_bmapi(). Once this is done, rename xfs_bmapi to xfs_bmapi_write to reflect the fact it is for allocation only. This enables us to kill the XFS_BMAPI_WRITE flag as well. Also clean up xfs_bmapi_write to the style used in the newly added xfs_bmapi_read/delay functions. Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 5 +- fs/xfs/xfs_bmap.c | 225 ++++++++++++++++++------------------------ fs/xfs/xfs_bmap.h | 50 +++------- fs/xfs/xfs_da_btree.c | 10 +- fs/xfs/xfs_dquot.c | 12 +-- fs/xfs/xfs_iomap.c | 20 ++-- fs/xfs/xfs_rtalloc.c | 6 +- fs/xfs/xfs_vnodeops.c | 19 ++-- 8 files changed, 135 insertions(+), 212 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 41ef02b7185a..5484766938f9 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2039,10 +2039,9 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) */ xfs_bmap_init(args->flist, args->firstblock); nmap = 1; - error = xfs_bmapi(args->trans, dp, (xfs_fileoff_t)lblkno, + error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno, blkcnt, - XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA | - XFS_BMAPI_WRITE, + XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA, args->firstblock, args->total, &map, &nmap, args->flist); if (!error) { diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index c2e49fd18bfd..595cc6311937 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4211,9 +4211,8 @@ xfs_bmap_validate_ret( ASSERT(i == 0 || mval[i - 1].br_startoff + mval[i - 1].br_blockcount == mval[i].br_startoff); - if (flags & XFS_BMAPI_WRITE) - ASSERT(mval[i].br_startblock != DELAYSTARTBLOCK && - mval[i].br_startblock != HOLESTARTBLOCK); + ASSERT(mval[i].br_startblock != DELAYSTARTBLOCK && + mval[i].br_startblock != HOLESTARTBLOCK); ASSERT(mval[i].br_state == XFS_EXT_NORM || mval[i].br_state == XFS_EXT_UNWRITTEN); } @@ -4809,60 +4808,57 @@ xfs_bmapi_convert_unwritten( } /* - * Map file blocks to filesystem blocks. - * File range is given by the bno/len pair. - * Adds blocks to file if a write ("flags & XFS_BMAPI_WRITE" set) - * into a hole or past eof. - * Only allocates blocks from a single allocation group, - * to avoid locking problems. + * Map file blocks to filesystem blocks, and allocate blocks or convert the + * extent state if necessary. Details behaviour is controlled by the flags + * parameter. Only allocates blocks from a single allocation group, to avoid + * locking problems. + * * The returned value in "firstblock" from the first call in a transaction * must be remembered and presented to subsequent calls in "firstblock". * An upper bound for the number of blocks to be allocated is supplied to * the first call in "total"; if no allocation group has that many free * blocks then the call will fail (return NULLFSBLOCK in "firstblock"). */ -int /* error */ -xfs_bmapi( - xfs_trans_t *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode */ - xfs_fileoff_t bno, /* starting file offs. mapped */ - xfs_filblks_t len, /* length to map in file */ - int flags, /* XFS_BMAPI_... */ - xfs_fsblock_t *firstblock, /* first allocated block - controls a.g. for allocs */ - xfs_extlen_t total, /* total blocks needed */ - xfs_bmbt_irec_t *mval, /* output: map values */ - int *nmap, /* i/o: mval size/count */ - xfs_bmap_free_t *flist) /* i/o: list extents to free */ +int +xfs_bmapi_write( + struct xfs_trans *tp, /* transaction pointer */ + struct xfs_inode *ip, /* incore inode */ + xfs_fileoff_t bno, /* starting file offs. mapped */ + xfs_filblks_t len, /* length to map in file */ + int flags, /* XFS_BMAPI_... */ + xfs_fsblock_t *firstblock, /* first allocated block + controls a.g. for allocs */ + xfs_extlen_t total, /* total blocks needed */ + struct xfs_bmbt_irec *mval, /* output: map values */ + int *nmap, /* i/o: mval size/count */ + struct xfs_bmap_free *flist) /* i/o: list extents to free */ { - xfs_bmalloca_t bma = { 0 }; /* args for xfs_bmap_alloc */ - xfs_btree_cur_t *cur; /* bmap btree cursor */ - xfs_fileoff_t end; /* end of mapped file region */ - int eof; /* we've hit the end of extents */ - xfs_bmbt_rec_host_t *ep; /* extent record pointer */ - int error; /* error return */ - xfs_bmbt_irec_t got; /* current file extent record */ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_extnum_t lastx; /* last useful extent number */ - int logflags; /* flags for transaction logging */ - xfs_extlen_t minleft; /* min blocks left after allocation */ - xfs_mount_t *mp; /* xfs mount structure */ - int n; /* current extent index */ - int nallocs; /* number of extents alloc'd */ - xfs_fileoff_t obno; /* old block number (offset) */ - xfs_bmbt_irec_t prev; /* previous file extent record */ - int tmp_logflags; /* temp flags holder */ - int whichfork; /* data or attr fork */ - char inhole; /* current location is hole in file */ - char wasdelay; /* old extent was delayed */ - char wr; /* this is a write request */ - char rt; /* this is a realtime file */ + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp; + struct xfs_bmalloca bma = { 0 }; /* args for xfs_bmap_alloc */ + struct xfs_btree_cur *cur; /* bmap btree cursor */ + xfs_fileoff_t end; /* end of mapped file region */ + int eof; /* after the end of extents */ + int error; /* error return */ + struct xfs_bmbt_irec got; /* current file extent record */ + xfs_extnum_t lastx; /* last useful extent number */ + int logflags; /* flags for transaction logging */ + xfs_extlen_t minleft; /* min blocks left after allocation */ + int n; /* current extent index */ + int nallocs; /* number of extents alloc'd */ + xfs_fileoff_t obno; /* old block number (offset) */ + struct xfs_bmbt_irec prev; /* previous file extent record */ + int tmp_logflags; /* temp flags holder */ + int whichfork; /* data or attr fork */ + char inhole; /* current location is hole in file */ + char wasdelay; /* old extent was delayed */ + #ifdef DEBUG - xfs_fileoff_t orig_bno; /* original block number value */ - int orig_flags; /* original flags arg value */ - xfs_filblks_t orig_len; /* original value of len arg */ - xfs_bmbt_irec_t *orig_mval; /* original value of mval */ - int orig_nmap; /* original value of *nmap */ + xfs_fileoff_t orig_bno; /* original block number value */ + int orig_flags; /* original flags arg value */ + xfs_filblks_t orig_len; /* original value of len arg */ + struct xfs_bmbt_irec *orig_mval; /* original value of mval */ + int orig_nmap; /* original value of *nmap */ orig_bno = bno; orig_len = len; @@ -4870,69 +4866,60 @@ xfs_bmapi( orig_mval = mval; orig_nmap = *nmap; #endif + ASSERT(*nmap >= 1); - ASSERT(*nmap <= XFS_BMAP_MAX_NMAP || !(flags & XFS_BMAPI_WRITE)); + ASSERT(*nmap <= XFS_BMAP_MAX_NMAP); + ASSERT(!(flags & XFS_BMAPI_IGSTATE)); + ASSERT(tp != NULL); + whichfork = (flags & XFS_BMAPI_ATTRFORK) ? XFS_ATTR_FORK : XFS_DATA_FORK; - mp = ip->i_mount; + if (unlikely(XFS_TEST_ERROR( (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS && XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE && XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL), mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) { - XFS_ERROR_REPORT("xfs_bmapi", XFS_ERRLEVEL_LOW, mp); + XFS_ERROR_REPORT("xfs_bmapi_write", XFS_ERRLEVEL_LOW, mp); return XFS_ERROR(EFSCORRUPTED); } + if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); - rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip); + ifp = XFS_IFORK_PTR(ip, whichfork); ASSERT(ifp->if_ext_max == XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t)); - if ((wr = (flags & XFS_BMAPI_WRITE)) != 0) - XFS_STATS_INC(xs_blk_mapw); - else - XFS_STATS_INC(xs_blk_mapr); - /* - * IGSTATE flag is used to combine extents which - * differ only due to the state of the extents. - * This technique is used from xfs_getbmap() - * when the caller does not wish to see the - * separation (which is the default). - * - * This technique is also used when writing a - * buffer which has been partially written, - * (usually by being flushed during a chunkread), - * to ensure one write takes place. This also - * prevents a change in the xfs inode extents at - * this time, intentionally. This change occurs - * on completion of the write operation, in - * xfs_strat_comp(), where the xfs_bmapi() call - * is transactioned, and the extents combined. - */ - if ((flags & XFS_BMAPI_IGSTATE) && wr) /* if writing unwritten space */ - wr = 0; /* no allocations are allowed */ + + XFS_STATS_INC(xs_blk_mapw); + logflags = 0; nallocs = 0; cur = NULL; + if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { - ASSERT(wr && tp); - if ((error = xfs_bmap_local_to_extents(tp, ip, - firstblock, total, &logflags, whichfork))) + error = xfs_bmap_local_to_extents(tp, ip, firstblock, total, + &logflags, whichfork); + if (error) goto error0; } - if (wr && *firstblock == NULLFSBLOCK) { + + if (*firstblock == NULLFSBLOCK) { if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE) minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1; else minleft = 1; - } else + } else { minleft = 0; - if (!(ifp->if_flags & XFS_IFEXTENTS) && - (error = xfs_iread_extents(tp, ip, whichfork))) - goto error0; - ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, - &prev); + } + + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + error = xfs_iread_extents(tp, ip, whichfork); + if (error) + goto error0; + } + + xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev); n = 0; end = bno + len; obno = bno; @@ -4945,19 +4932,14 @@ xfs_bmapi( bma.userdata = 0; while (bno < end && n < *nmap) { - /* - * Reading past eof, act as though there's a hole - * up to end. - */ - if (eof && !wr) - got.br_startoff = end; inhole = eof || got.br_startoff > bno; - wasdelay = wr && !inhole && isnullstartblock(got.br_startblock); + wasdelay = !inhole && isnullstartblock(got.br_startblock); + /* * First, deal with the hole before the allocated space * that we found, if any. */ - if (wr && (inhole || wasdelay)) { + if (inhole || wasdelay) { bma.eof = eof; bma.conv = !!(flags & XFS_BMAPI_CONVERT); bma.wasdel = wasdelay; @@ -4975,36 +4957,20 @@ xfs_bmapi( minleft = 0; if (bma.rval == NULLFSBLOCK) break; - } else if (inhole) { - /* - * Reading in a hole. - */ - mval->br_startoff = bno; - mval->br_startblock = HOLESTARTBLOCK; - mval->br_blockcount = - XFS_FILBLKS_MIN(len, got.br_startoff - bno); - mval->br_state = XFS_EXT_NORM; - bno += mval->br_blockcount; - len -= mval->br_blockcount; - mval++; - n++; - continue; } /* Deal with the allocated space we found. */ xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); /* Execute unwritten extent conversion if necessary */ - if (wr) { - error = xfs_bmapi_convert_unwritten(&bma, mval, len, - &lastx, &cur, firstblock, flist, flags, - &tmp_logflags); - logflags |= tmp_logflags; - if (error == EAGAIN) - continue; - if (error) - goto error0; - } + error = xfs_bmapi_convert_unwritten(&bma, mval, len, &lastx, + &cur, firstblock, flist, + flags, &tmp_logflags); + logflags |= tmp_logflags; + if (error == EAGAIN) + continue; + if (error) + goto error0; /* update the extent map to return */ xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags); @@ -5016,24 +4982,22 @@ xfs_bmapi( */ if (bno >= end || n >= *nmap || nallocs >= *nmap) break; - /* - * Else go on to the next record. - */ + + /* Else go on to the next record. */ prev = got; - if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) { - ep = xfs_iext_get_ext(ifp, lastx); - xfs_bmbt_get_all(ep, &got); - } else { + if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got); + else eof = 1; - } } *nmap = n; + /* * Transform from btree to extents, give it cur. */ - if (tp && XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE && + if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE && XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) { - ASSERT(wr && cur); + ASSERT(cur); error = xfs_bmap_btree_to_extents(tp, ip, cur, &tmp_logflags, whichfork); logflags |= tmp_logflags; @@ -5061,10 +5025,9 @@ xfs_bmapi( * detecting a case where the data is changed, there's an error, * and it's not logged so we don't shutdown when we should. */ - if (logflags) { - ASSERT(tp && wr); + if (logflags) xfs_trans_log_inode(tp, ip, logflags); - } + if (cur) { if (!error) { ASSERT(*firstblock == NULLFSBLOCK || diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index c70c19c7f1fa..be235f56d2a6 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -62,25 +62,23 @@ typedef struct xfs_bmap_free #define XFS_BMAP_MAX_NMAP 4 /* - * Flags for xfs_bmapi + * Flags for xfs_bmapi_* */ -#define XFS_BMAPI_WRITE 0x001 /* write operation: allocate space */ -#define XFS_BMAPI_ENTIRE 0x004 /* return entire extent, not trimmed */ -#define XFS_BMAPI_METADATA 0x008 /* mapping metadata not user data */ -#define XFS_BMAPI_ATTRFORK 0x010 /* use attribute fork not data */ -#define XFS_BMAPI_PREALLOC 0x040 /* preallocation op: unwritten space */ -#define XFS_BMAPI_IGSTATE 0x080 /* Ignore state - */ +#define XFS_BMAPI_ENTIRE 0x001 /* return entire extent, not trimmed */ +#define XFS_BMAPI_METADATA 0x002 /* mapping metadata not user data */ +#define XFS_BMAPI_ATTRFORK 0x004 /* use attribute fork not data */ +#define XFS_BMAPI_PREALLOC 0x008 /* preallocation op: unwritten space */ +#define XFS_BMAPI_IGSTATE 0x010 /* Ignore state - */ /* combine contig. space */ -#define XFS_BMAPI_CONTIG 0x100 /* must allocate only one extent */ +#define XFS_BMAPI_CONTIG 0x020 /* must allocate only one extent */ /* * unwritten extent conversion - this needs write cache flushing and no additional * allocation alignments. When specified with XFS_BMAPI_PREALLOC it converts * from written to unwritten, otherwise convert from unwritten to written. */ -#define XFS_BMAPI_CONVERT 0x200 +#define XFS_BMAPI_CONVERT 0x040 #define XFS_BMAPI_FLAGS \ - { XFS_BMAPI_WRITE, "WRITE" }, \ { XFS_BMAPI_ENTIRE, "ENTIRE" }, \ { XFS_BMAPI_METADATA, "METADATA" }, \ { XFS_BMAPI_ATTRFORK, "ATTRFORK" }, \ @@ -265,39 +263,17 @@ xfs_bmap_read_extents( struct xfs_inode *ip, /* incore inode */ int whichfork); /* data or attr fork */ -/* - * Map file blocks to filesystem blocks. - * File range is given by the bno/len pair. - * Adds blocks to file if a write ("flags & XFS_BMAPI_WRITE" set) - * into a hole or past eof. - * Only allocates blocks from a single allocation group, - * to avoid locking problems. - * The returned value in "firstblock" from the first call in a transaction - * must be remembered and presented to subsequent calls in "firstblock". - * An upper bound for the number of blocks to be allocated is supplied to - * the first call in "total"; if no allocation group has that many free - * blocks then the call will fail (return NULLFSBLOCK in "firstblock"). - */ -int /* error */ -xfs_bmapi( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - xfs_fileoff_t bno, /* starting file offs. mapped */ - xfs_filblks_t len, /* length to map in file */ - int flags, /* XFS_BMAPI_... */ - xfs_fsblock_t *firstblock, /* first allocated block - controls a.g. for allocs */ - xfs_extlen_t total, /* total blocks needed */ - struct xfs_bmbt_irec *mval, /* output: map values */ - int *nmap, /* i/o: mval size/count */ - xfs_bmap_free_t *flist); /* i/o: list extents to free */ - int xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, struct xfs_bmbt_irec *mval, int *nmap, int flags); int xfs_bmapi_delay(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, struct xfs_bmbt_irec *mval, int *nmap, int flags); +int xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_fileoff_t bno, xfs_filblks_t len, int flags, + xfs_fsblock_t *firstblock, xfs_extlen_t total, + struct xfs_bmbt_irec *mval, int *nmap, + struct xfs_bmap_free *flist); /* * Unmap (remove) blocks from a file. diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 70a5f580e5ed..46c8aa2740da 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -1578,9 +1578,8 @@ xfs_da_grow_inode_int( */ nmap = 1; ASSERT(args->firstblock != NULL); - error = xfs_bmapi(tp, dp, *bno, count, - xfs_bmapi_aflag(w)|XFS_BMAPI_WRITE|XFS_BMAPI_METADATA| - XFS_BMAPI_CONTIG, + error = xfs_bmapi_write(tp, dp, *bno, count, + xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA|XFS_BMAPI_CONTIG, args->firstblock, args->total, &map, &nmap, args->flist); if (error) @@ -1602,9 +1601,8 @@ xfs_da_grow_inode_int( for (b = *bno, mapi = 0; b < *bno + count; ) { nmap = MIN(XFS_BMAP_MAX_NMAP, count); c = (int)(*bno + count - b); - error = xfs_bmapi(tp, dp, b, c, - xfs_bmapi_aflag(w)|XFS_BMAPI_WRITE| - XFS_BMAPI_METADATA, + error = xfs_bmapi_write(tp, dp, b, c, + xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA, args->firstblock, args->total, &mapp[mapi], &nmap, args->flist); if (error) diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index c377961657ee..179673531f20 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -379,14 +379,12 @@ xfs_qm_dqalloc( xfs_trans_ijoin_ref(tp, quotip, XFS_ILOCK_EXCL); nmaps = 1; - if ((error = xfs_bmapi(tp, quotip, - offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB, - XFS_BMAPI_METADATA | XFS_BMAPI_WRITE, - &firstblock, - XFS_QM_DQALLOC_SPACE_RES(mp), - &map, &nmaps, &flist))) { + error = xfs_bmapi_write(tp, quotip, offset_fsb, + XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA, + &firstblock, XFS_QM_DQALLOC_SPACE_RES(mp), + &map, &nmaps, &flist); + if (error) goto error0; - } ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB); ASSERT(nmaps == 1); ASSERT((map.br_startblock != DELAYSTARTBLOCK) && diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 681ba34c9233..da5bf05c5bb7 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -210,20 +210,18 @@ xfs_iomap_write_direct( xfs_trans_ijoin(tp, ip); - bmapi_flag = XFS_BMAPI_WRITE; + bmapi_flag = 0; if (offset < ip->i_size || extsz) bmapi_flag |= XFS_BMAPI_PREALLOC; /* - * Issue the xfs_bmapi() call to allocate the blocks. - * * From this point onwards we overwrite the imap pointer that the * caller gave to us. */ xfs_bmap_init(&free_list, &firstfsb); nimaps = 1; - error = xfs_bmapi(tp, ip, offset_fsb, count_fsb, bmapi_flag, - &firstfsb, 0, imap, &nimaps, &free_list); + error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, bmapi_flag, + &firstfsb, 0, imap, &nimaps, &free_list); if (error) goto error0; @@ -582,14 +580,12 @@ xfs_iomap_write_allocate( } /* - * Go get the actual blocks. - * * From this point onwards we overwrite the imap * pointer that the caller gave to us. */ - error = xfs_bmapi(tp, ip, map_start_fsb, count_fsb, - XFS_BMAPI_WRITE, &first_block, 1, - imap, &nimaps, &free_list); + error = xfs_bmapi_write(tp, ip, map_start_fsb, + count_fsb, 0, &first_block, 1, + imap, &nimaps, &free_list); if (error) goto trans_cancel; @@ -703,8 +699,8 @@ xfs_iomap_write_unwritten( */ xfs_bmap_init(&free_list, &firstfsb); nimaps = 1; - error = xfs_bmapi(tp, ip, offset_fsb, count_fsb, - XFS_BMAPI_WRITE|XFS_BMAPI_CONVERT, &firstfsb, + error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, + XFS_BMAPI_CONVERT, &firstfsb, 1, &imap, &nimaps, &free_list); if (error) goto error_on_bmapi_transaction; diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index e5f40c6460b2..f29424964521 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -120,9 +120,9 @@ xfs_growfs_rt_alloc( */ nmap = 1; cancelflags |= XFS_TRANS_ABORT; - error = xfs_bmapi(tp, ip, oblocks, nblocks - oblocks, - XFS_BMAPI_WRITE | XFS_BMAPI_METADATA, &firstblock, - resblks, &map, &nmap, &flist); + error = xfs_bmapi_write(tp, ip, oblocks, nblocks - oblocks, + XFS_BMAPI_METADATA, &firstblock, + resblks, &map, &nmap, &flist); if (!error && nmap < 1) error = XFS_ERROR(ENOSPC); if (error) diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 63874a87b378..f47ecee8d437 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1633,10 +1633,9 @@ xfs_symlink( first_fsb = 0; nmaps = SYMLINK_MAPS; - error = xfs_bmapi(tp, ip, first_fsb, fs_blocks, - XFS_BMAPI_WRITE | XFS_BMAPI_METADATA, - &first_block, resblks, mval, &nmaps, - &free_list); + error = xfs_bmapi_write(tp, ip, first_fsb, fs_blocks, + XFS_BMAPI_METADATA, &first_block, resblks, + mval, &nmaps, &free_list); if (error) goto error2; @@ -1782,7 +1781,6 @@ xfs_alloc_file_space( xfs_fileoff_t startoffset_fsb; xfs_fsblock_t firstfsb; int nimaps; - int bmapi_flag; int quota_flag; int rt; xfs_trans_t *tp; @@ -1810,7 +1808,6 @@ xfs_alloc_file_space( count = len; imapp = &imaps[0]; nimaps = 1; - bmapi_flag = XFS_BMAPI_WRITE | alloc_type; startoffset_fsb = XFS_B_TO_FSBT(mp, offset); allocatesize_fsb = XFS_B_TO_FSB(mp, count); @@ -1883,14 +1880,10 @@ xfs_alloc_file_space( xfs_trans_ijoin(tp, ip); - /* - * Issue the xfs_bmapi() call to allocate the blocks - */ xfs_bmap_init(&free_list, &firstfsb); - error = xfs_bmapi(tp, ip, startoffset_fsb, - allocatesize_fsb, bmapi_flag, - &firstfsb, 0, imapp, &nimaps, - &free_list); + error = xfs_bmapi_write(tp, ip, startoffset_fsb, + allocatesize_fsb, alloc_type, &firstfsb, + 0, imapp, &nimaps, &free_list); if (error) { goto error0; } From 27a3f8f2de758205765f277b3428bbf3d15da973 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:53 +0000 Subject: [PATCH 33/69] xfs: introduce xfs_bmap_last_extent Add a common helper for finding the last extent in a file. Largely based on a patch from Dave Chinner. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 226 +++++++++++++++++++++------------------------- 1 file changed, 105 insertions(+), 121 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 595cc6311937..d9de5ae1fd5c 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -203,19 +203,6 @@ xfs_bmap_search_extents( xfs_bmbt_irec_t *gotp, /* out: extent entry found */ xfs_bmbt_irec_t *prevp); /* out: previous extent entry found */ -/* - * Check the last inode extent to determine whether this allocation will result - * in blocks being allocated at the end of the file. When we allocate new data - * blocks at the end of the file which do not start at the previous data block, - * we will try to align the new blocks at stripe unit boundaries. - */ -STATIC int /* error */ -xfs_bmap_isaeof( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_fileoff_t off, /* file offset in fsblocks */ - int whichfork, /* data or attribute fork */ - char *aeof); /* return value */ - /* * Compute the worst-case number of indirect blocks that will be used * for ip's delayed extent of length "len". @@ -3924,42 +3911,122 @@ xfs_bmap_last_before( return 0; } +STATIC int +xfs_bmap_last_extent( + struct xfs_trans *tp, + struct xfs_inode *ip, + int whichfork, + struct xfs_bmbt_irec *rec, + int *is_empty) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); + int error; + int nextents; + + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + error = xfs_iread_extents(tp, ip, whichfork); + if (error) + return error; + } + + nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t); + if (nextents == 0) { + *is_empty = 1; + return 0; + } + + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, nextents - 1), rec); + *is_empty = 0; + return 0; +} + +/* + * Check the last inode extent to determine whether this allocation will result + * in blocks being allocated at the end of the file. When we allocate new data + * blocks at the end of the file which do not start at the previous data block, + * we will try to align the new blocks at stripe unit boundaries. + * + * Returns 0 in *aeof if the file (fork) is empty as any new write will be at, + * or past the EOF. + */ +STATIC int +xfs_bmap_isaeof( + struct xfs_inode *ip, + xfs_fileoff_t off, + int whichfork, + char *aeof) +{ + struct xfs_bmbt_irec rec; + int is_empty; + int error; + + *aeof = 0; + error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty); + if (error || is_empty) + return error; + + /* + * Check if we are allocation or past the last extent, or at least into + * the last delayed allocated extent. + */ + *aeof = off >= rec.br_startoff + rec.br_blockcount || + (off >= rec.br_startoff && isnullstartblock(rec.br_startblock)); + return 0; +} + +/* + * Check if the endoff is outside the last extent. If so the caller will grow + * the allocation to a stripe unit boundary. All offsets are considered outside + * the end of file for an empty fork, so 1 is returned in *eof in that case. + */ +int +xfs_bmap_eof( + struct xfs_inode *ip, + xfs_fileoff_t endoff, + int whichfork, + int *eof) +{ + struct xfs_bmbt_irec rec; + int error; + + error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, eof); + if (error || *eof) + return error; + + *eof = endoff >= rec.br_startoff + rec.br_blockcount; + return 0; +} + /* * Returns the file-relative block number of the first block past eof in * the file. This is not based on i_size, it is based on the extent records. * Returns 0 for local files, as they do not have extent records. */ -int /* error */ +int xfs_bmap_last_offset( - xfs_trans_t *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode */ - xfs_fileoff_t *last_block, /* last block */ - int whichfork) /* data or attr fork */ + struct xfs_trans *tp, + struct xfs_inode *ip, + xfs_fileoff_t *last_block, + int whichfork) { - xfs_bmbt_rec_host_t *ep; /* pointer to last extent */ - int error; /* error return value */ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_extnum_t nextents; /* number of extent entries */ + struct xfs_bmbt_irec rec; + int is_empty; + int error; + + *last_block = 0; + + if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) + return 0; if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE && - XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS && - XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL) + XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS) return XFS_ERROR(EIO); - if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { - *last_block = 0; - return 0; - } - ifp = XFS_IFORK_PTR(ip, whichfork); - if (!(ifp->if_flags & XFS_IFEXTENTS) && - (error = xfs_iread_extents(tp, ip, whichfork))) + + error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty); + if (error || is_empty) return error; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - if (!nextents) { - *last_block = 0; - return 0; - } - ep = xfs_iext_get_ext(ifp, nextents - 1); - *last_block = xfs_bmbt_get_startoff(ep) + xfs_bmbt_get_blockcount(ep); + + *last_block = rec.br_startoff + rec.br_blockcount; return 0; } @@ -5687,89 +5754,6 @@ xfs_getbmap( return error; } -/* - * Check the last inode extent to determine whether this allocation will result - * in blocks being allocated at the end of the file. When we allocate new data - * blocks at the end of the file which do not start at the previous data block, - * we will try to align the new blocks at stripe unit boundaries. - */ -STATIC int /* error */ -xfs_bmap_isaeof( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_fileoff_t off, /* file offset in fsblocks */ - int whichfork, /* data or attribute fork */ - char *aeof) /* return value */ -{ - int error; /* error return value */ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_bmbt_rec_host_t *lastrec; /* extent record pointer */ - xfs_extnum_t nextents; /* number of file extents */ - xfs_bmbt_irec_t s; /* expanded extent record */ - - ASSERT(whichfork == XFS_DATA_FORK); - ifp = XFS_IFORK_PTR(ip, whichfork); - if (!(ifp->if_flags & XFS_IFEXTENTS) && - (error = xfs_iread_extents(NULL, ip, whichfork))) - return error; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - if (nextents == 0) { - *aeof = 1; - return 0; - } - /* - * Go to the last extent - */ - lastrec = xfs_iext_get_ext(ifp, nextents - 1); - xfs_bmbt_get_all(lastrec, &s); - /* - * Check we are allocating in the last extent (for delayed allocations) - * or past the last extent for non-delayed allocations. - */ - *aeof = (off >= s.br_startoff && - off < s.br_startoff + s.br_blockcount && - isnullstartblock(s.br_startblock)) || - off >= s.br_startoff + s.br_blockcount; - return 0; -} - -/* - * Check if the endoff is outside the last extent. If so the caller will grow - * the allocation to a stripe unit boundary. - */ -int /* error */ -xfs_bmap_eof( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_fileoff_t endoff, /* file offset in fsblocks */ - int whichfork, /* data or attribute fork */ - int *eof) /* result value */ -{ - xfs_fsblock_t blockcount; /* extent block count */ - int error; /* error return value */ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_bmbt_rec_host_t *lastrec; /* extent record pointer */ - xfs_extnum_t nextents; /* number of file extents */ - xfs_fileoff_t startoff; /* extent starting file offset */ - - ASSERT(whichfork == XFS_DATA_FORK); - ifp = XFS_IFORK_PTR(ip, whichfork); - if (!(ifp->if_flags & XFS_IFEXTENTS) && - (error = xfs_iread_extents(NULL, ip, whichfork))) - return error; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - if (nextents == 0) { - *eof = 1; - return 0; - } - /* - * Go to the last extent - */ - lastrec = xfs_iext_get_ext(ifp, nextents - 1); - startoff = xfs_bmbt_get_startoff(lastrec); - blockcount = xfs_bmbt_get_blockcount(lastrec); - *eof = endoff >= startoff + blockcount; - return 0; -} - #ifdef DEBUG STATIC struct xfs_buf * xfs_bmap_get_bp( From a5bd606ba65f24e5990edfc0e7b52702720ee6fa Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:40:54 +0000 Subject: [PATCH 34/69] xfs: remove xfs_bmap_add_extent There is no real need to the xfs_bmap_add_extent, as the callers know what kind of extents they need to it. Removing it means duplicating the extents to btree conversion logic in three places, but overall it's still much simpler code and quite a bit less code. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 418 +++++++++++++++++++--------------------------- 1 file changed, 174 insertions(+), 244 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index d9de5ae1fd5c..0c9be0ed606e 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -50,17 +50,22 @@ #include "xfs_trace.h" -#ifdef DEBUG -STATIC void -xfs_bmap_check_leaf_extents(xfs_btree_cur_t *cur, xfs_inode_t *ip, int whichfork); -#endif - kmem_zone_t *xfs_bmap_free_item_zone; /* * Prototypes for internal bmap routines. */ +#ifdef DEBUG +STATIC void +xfs_bmap_check_leaf_extents( + struct xfs_btree_cur *cur, + struct xfs_inode *ip, + int whichfork); +#else +#define xfs_bmap_check_leaf_extents(cur, ip, whichfork) do { } while (0) +#endif + /* * Called from xfs_bmap_add_attrfork to handle extents format files. @@ -84,47 +89,6 @@ xfs_bmap_add_attrfork_local( xfs_bmap_free_t *flist, /* blocks to free at commit */ int *flags); /* inode logging flags */ -/* - * Called by xfs_bmap_add_extent to handle cases converting a delayed - * allocation to a real allocation. - */ -STATIC int /* error */ -xfs_bmap_add_extent_delay_real( - struct xfs_trans *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - xfs_filblks_t *dnew, /* new delayed-alloc indirect blocks */ - xfs_fsblock_t *first, /* pointer to firstblock variable */ - xfs_bmap_free_t *flist, /* list of extents to be freed */ - int *logflagsp); /* inode logging flags */ - -/* - * Called by xfs_bmap_add_extent to handle cases converting a hole - * to a real allocation. - */ -STATIC int /* error */ -xfs_bmap_add_extent_hole_real( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t *cur, /* if null, not a btree */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - int *logflagsp, /* inode logging flags */ - int whichfork); /* data or attr fork */ - -/* - * Called by xfs_bmap_add_extent to handle cases converting an unwritten - * allocation to a real allocation or vice versa. - */ -STATIC int /* error */ -xfs_bmap_add_extent_unwritten_real( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - int *logflagsp); /* inode logging flags */ - /* * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file. * It figures out where to ask the underlying allocator to put the new extent. @@ -407,147 +371,7 @@ xfs_bmap_add_attrfork_local( } /* - * Update file extent records and the btree after allocating space. - */ -STATIC int /* error */ -xfs_bmap_add_extent( - struct xfs_trans *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - xfs_fsblock_t *first, /* pointer to firstblock variable */ - xfs_bmap_free_t *flist, /* list of extents to be freed */ - int *logflagsp, /* inode logging flags */ - int whichfork) /* data or attr fork */ -{ - xfs_btree_cur_t *cur; /* btree cursor or null */ - xfs_filblks_t da_new; /* new count del alloc blocks used */ - xfs_filblks_t da_old; /* old count del alloc blocks used */ - int error; /* error return value */ - xfs_ifork_t *ifp; /* inode fork ptr */ - int logflags; /* returned value */ - xfs_extnum_t nextents; /* number of extents in file now */ - - XFS_STATS_INC(xs_add_exlist); - - cur = *curp; - ifp = XFS_IFORK_PTR(ip, whichfork); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - da_old = da_new = 0; - error = 0; - - ASSERT(*idx >= 0); - ASSERT(*idx <= nextents); - ASSERT(!isnullstartblock(new->br_startblock)); - - /* - * Real allocation off the end of the file. - */ - if (*idx == nextents) { - if (cur) - ASSERT((cur->bc_private.b.flags & - XFS_BTCUR_BPRV_WASDEL) == 0); - error = xfs_bmap_add_extent_hole_real(ip, idx, cur, new, - &logflags, whichfork); - } else { - xfs_bmbt_irec_t prev; /* old extent at offset idx */ - - /* - * Get the record referred to by idx. - */ - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx), &prev); - /* - * If it's a real allocation record, and the new allocation ends - * after the start of the referred to record, then we're filling - * in a delayed or unwritten allocation with a real one, or - * converting real back to unwritten. - */ - if (!isnullstartblock(new->br_startblock) && - new->br_startoff + new->br_blockcount > prev.br_startoff) { - if (prev.br_state != XFS_EXT_UNWRITTEN && - isnullstartblock(prev.br_startblock)) { - da_old = startblockval(prev.br_startblock); - if (cur) - ASSERT(cur->bc_private.b.flags & - XFS_BTCUR_BPRV_WASDEL); - error = xfs_bmap_add_extent_delay_real(tp, ip, - idx, &cur, new, &da_new, - first, flist, &logflags); - } else { - ASSERT(new->br_state == XFS_EXT_NORM || - new->br_state == XFS_EXT_UNWRITTEN); - - error = xfs_bmap_add_extent_unwritten_real(ip, - idx, &cur, new, &logflags); - if (error) - goto done; - } - } - /* - * Otherwise we're filling in a hole with an allocation. - */ - else { - if (cur) - ASSERT((cur->bc_private.b.flags & - XFS_BTCUR_BPRV_WASDEL) == 0); - error = xfs_bmap_add_extent_hole_real(ip, idx, cur, - new, &logflags, whichfork); - } - } - - if (error) - goto done; - ASSERT(*curp == cur || *curp == NULL); - - /* - * Convert to a btree if necessary. - */ - if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS && - XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) { - int tmp_logflags; /* partial log flag return val */ - - ASSERT(cur == NULL); - error = xfs_bmap_extents_to_btree(tp, ip, first, - flist, &cur, da_old > 0, &tmp_logflags, whichfork); - logflags |= tmp_logflags; - if (error) - goto done; - } - /* - * Adjust for changes in reserved delayed indirect blocks. - * Nothing to do for disk quotas here. - */ - if (da_old || da_new) { - xfs_filblks_t nblks; - - nblks = da_new; - if (cur) - nblks += cur->bc_private.b.allocated; - ASSERT(nblks <= da_old); - if (nblks < da_old) - xfs_icsb_modify_counters(ip->i_mount, XFS_SBS_FDBLOCKS, - (int64_t)(da_old - nblks), 0); - } - /* - * Clear out the allocated field, done with it now in any case. - */ - if (cur) { - cur->bc_private.b.allocated = 0; - *curp = cur; - } -done: -#ifdef DEBUG - if (!error) - xfs_bmap_check_leaf_extents(*curp, ip, whichfork); -#endif - *logflagsp = logflags; - return error; -} - -/* - * Called by xfs_bmap_add_extent to handle cases converting a delayed - * allocation to a real allocation. + * Convert a delayed allocation to a real allocation. */ STATIC int /* error */ xfs_bmap_add_extent_delay_real( @@ -556,7 +380,6 @@ xfs_bmap_add_extent_delay_real( xfs_extnum_t *idx, /* extent number to update/insert */ xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ xfs_bmbt_irec_t *new, /* new data to add to file extents */ - xfs_filblks_t *dnew, /* new delayed-alloc indirect blocks */ xfs_fsblock_t *first, /* pointer to firstblock variable */ xfs_bmap_free_t *flist, /* list of extents to be freed */ int *logflagsp) /* inode logging flags */ @@ -572,10 +395,24 @@ xfs_bmap_add_extent_delay_real( /* left is 0, right is 1, prev is 2 */ int rval=0; /* return value (logging flags) */ int state = 0;/* state bits, accessed thru macros */ - xfs_filblks_t temp=0; /* value for dnew calculations */ - xfs_filblks_t temp2=0;/* value for dnew calculations */ + xfs_filblks_t da_new; /* new count del alloc blocks used */ + xfs_filblks_t da_old; /* old count del alloc blocks used */ + xfs_filblks_t temp=0; /* value for da_new calculations */ + xfs_filblks_t temp2=0;/* value for da_new calculations */ int tmp_rval; /* partial logging flags */ + ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + cur = *curp; + + ASSERT(*idx >= 0); + ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(!isnullstartblock(new->br_startblock)); + ASSERT(!cur || (cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); + + XFS_STATS_INC(xs_add_exlist); + + *logflagsp = 0; + #define LEFT r[0] #define RIGHT r[1] #define PREV r[2] @@ -583,14 +420,15 @@ xfs_bmap_add_extent_delay_real( /* * Set up a bunch of variables to make the tests simpler. */ - cur = *curp; - ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); ep = xfs_iext_get_ext(ifp, *idx); xfs_bmbt_get_all(ep, &PREV); new_endoff = new->br_startoff + new->br_blockcount; ASSERT(PREV.br_startoff <= new->br_startoff); ASSERT(PREV.br_startoff + PREV.br_blockcount >= new_endoff); + da_old = startblockval(PREV.br_startblock); + da_new = 0; + /* * Set flags determining what part of the previous delayed allocation * extent is being replaced by a real allocation. @@ -688,7 +526,6 @@ xfs_bmap_add_extent_delay_real( RIGHT.br_blockcount, LEFT.br_state))) goto done; } - *dnew = 0; break; case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG: @@ -719,7 +556,6 @@ xfs_bmap_add_extent_delay_real( PREV.br_blockcount, LEFT.br_state))) goto done; } - *dnew = 0; break; case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG: @@ -749,8 +585,6 @@ xfs_bmap_add_extent_delay_real( RIGHT.br_blockcount, PREV.br_state))) goto done; } - - *dnew = 0; break; case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING: @@ -778,8 +612,6 @@ xfs_bmap_add_extent_delay_real( goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); } - - *dnew = 0; break; case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG: @@ -813,13 +645,12 @@ xfs_bmap_add_extent_delay_real( LEFT.br_state))) goto done; } - temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), startblockval(PREV.br_startblock)); - xfs_bmbt_set_startblock(ep, nullstartblock((int)temp)); + xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); --*idx; - *dnew = temp; break; case BMAP_LEFT_FILLING: @@ -856,14 +687,12 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; } - temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), startblockval(PREV.br_startblock) - (cur ? cur->bc_private.b.allocated : 0)); ep = xfs_iext_get_ext(ifp, *idx + 1); - xfs_bmbt_set_startblock(ep, nullstartblock((int)temp)); + xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); trace_xfs_bmap_post_update(ip, *idx + 1, state, _THIS_IP_); - - *dnew = temp; break; case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG: @@ -896,14 +725,13 @@ xfs_bmap_add_extent_delay_real( goto done; } - temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), startblockval(PREV.br_startblock)); trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); - xfs_bmbt_set_startblock(ep, nullstartblock((int)temp)); + xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); ++*idx; - *dnew = temp; break; case BMAP_RIGHT_FILLING: @@ -939,15 +767,14 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; } - temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), startblockval(PREV.br_startblock) - (cur ? cur->bc_private.b.allocated : 0)); ep = xfs_iext_get_ext(ifp, *idx); - xfs_bmbt_set_startblock(ep, nullstartblock((int)temp)); + xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); ++*idx; - *dnew = temp; break; case 0: @@ -1029,7 +856,7 @@ xfs_bmap_add_extent_delay_real( trace_xfs_bmap_post_update(ip, *idx + 2, state, _THIS_IP_); ++*idx; - *dnew = temp + temp2; + da_new = temp + temp2; break; case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG: @@ -1044,9 +871,39 @@ xfs_bmap_add_extent_delay_real( */ ASSERT(0); } - *curp = cur; + + /* convert to a btree if necessary */ + if (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) == XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) > ifp->if_ext_max) { + int tmp_logflags; /* partial log flag return val */ + + ASSERT(cur == NULL); + error = xfs_bmap_extents_to_btree(tp, ip, first, flist, &cur, + da_old > 0, &tmp_logflags, XFS_DATA_FORK); + *logflagsp |= tmp_logflags; + if (error) + goto done; + } + + /* adjust for changes in reserved delayed indirect blocks */ + if (da_old || da_new) { + temp = da_new; + if (cur) + temp += cur->bc_private.b.allocated; + ASSERT(temp <= da_old); + if (temp < da_old) + xfs_icsb_modify_counters(ip->i_mount, XFS_SBS_FDBLOCKS, + (int64_t)(da_old - temp), 0); + } + + /* clear out the allocated field, done with it now in any case. */ + if (cur) { + cur->bc_private.b.allocated = 0; + *curp = cur; + } + xfs_bmap_check_leaf_extents(cur, ip, XFS_DATA_FORK); done: - *logflagsp = rval; + *logflagsp |= rval; return error; #undef LEFT #undef RIGHT @@ -1054,15 +911,17 @@ xfs_bmap_add_extent_delay_real( } /* - * Called by xfs_bmap_add_extent to handle cases converting an unwritten - * allocation to a real allocation or vice versa. + * Convert an unwritten allocation to a real allocation or vice versa. */ STATIC int /* error */ xfs_bmap_add_extent_unwritten_real( + struct xfs_trans *tp, xfs_inode_t *ip, /* incore inode pointer */ xfs_extnum_t *idx, /* extent number to update/insert */ xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ xfs_bmbt_irec_t *new, /* new data to add to file extents */ + xfs_fsblock_t *first, /* pointer to firstblock variable */ + xfs_bmap_free_t *flist, /* list of extents to be freed */ int *logflagsp) /* inode logging flags */ { xfs_btree_cur_t *cur; /* btree cursor */ @@ -1078,15 +937,25 @@ xfs_bmap_add_extent_unwritten_real( int rval=0; /* return value (logging flags) */ int state = 0;/* state bits, accessed thru macros */ + *logflagsp = 0; + + cur = *curp; + ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + + ASSERT(*idx >= 0); + ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(!isnullstartblock(new->br_startblock)); + + XFS_STATS_INC(xs_add_exlist); + #define LEFT r[0] #define RIGHT r[1] #define PREV r[2] + /* * Set up a bunch of variables to make the tests simpler. */ error = 0; - cur = *curp; - ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); ep = xfs_iext_get_ext(ifp, *idx); xfs_bmbt_get_all(ep, &PREV); newext = new->br_state; @@ -1537,9 +1406,29 @@ xfs_bmap_add_extent_unwritten_real( */ ASSERT(0); } - *curp = cur; + + /* convert to a btree if necessary */ + if (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) == XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) > ifp->if_ext_max) { + int tmp_logflags; /* partial log flag return val */ + + ASSERT(cur == NULL); + error = xfs_bmap_extents_to_btree(tp, ip, first, flist, &cur, + 0, &tmp_logflags, XFS_DATA_FORK); + *logflagsp |= tmp_logflags; + if (error) + goto done; + } + + /* clear out the allocated field, done with it now in any case. */ + if (cur) { + cur->bc_private.b.allocated = 0; + *curp = cur; + } + + xfs_bmap_check_leaf_extents(*curp, ip, XFS_DATA_FORK); done: - *logflagsp = rval; + *logflagsp |= rval; return error; #undef LEFT #undef RIGHT @@ -1691,30 +1580,42 @@ xfs_bmap_add_extent_hole_delay( } /* - * Called by xfs_bmap_add_extent to handle cases converting a hole - * to a real allocation. + * Convert a hole to a real allocation. */ STATIC int /* error */ xfs_bmap_add_extent_hole_real( + struct xfs_trans *tp, xfs_inode_t *ip, /* incore inode pointer */ xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t *cur, /* if null, not a btree */ + xfs_btree_cur_t **curp, /* if null, not a btree */ xfs_bmbt_irec_t *new, /* new data to add to file extents */ + xfs_fsblock_t *first, /* pointer to firstblock variable */ + xfs_bmap_free_t *flist, /* list of extents to be freed */ int *logflagsp, /* inode logging flags */ int whichfork) /* data or attr fork */ { int error; /* error return value */ int i; /* temp state */ + xfs_btree_cur_t *cur; /* if null, not a btree */ xfs_ifork_t *ifp; /* inode fork pointer */ xfs_bmbt_irec_t left; /* left neighbor extent entry */ xfs_bmbt_irec_t right; /* right neighbor extent entry */ int rval=0; /* return value (logging flags) */ int state; /* state bits, accessed thru macros */ - ifp = XFS_IFORK_PTR(ip, whichfork); - ASSERT(*idx <= ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)); - state = 0; + *logflagsp = 0; + ifp = XFS_IFORK_PTR(ip, whichfork); + cur = *curp; + + ASSERT(*idx >= 0); + ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(!isnullstartblock(new->br_startblock)); + ASSERT(!cur || !(cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); + + XFS_STATS_INC(xs_add_exlist); + + state = 0; if (whichfork == XFS_ATTR_FORK) state |= BMAP_ATTRFORK; @@ -1897,8 +1798,28 @@ xfs_bmap_add_extent_hole_real( } break; } + + /* convert to a btree if necessary */ + if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) { + int tmp_logflags; /* partial log flag return val */ + + ASSERT(cur == NULL); + error = xfs_bmap_extents_to_btree(tp, ip, first, + flist, &cur, 0, &tmp_logflags, whichfork); + *logflagsp |= tmp_logflags; + if (error) + goto done; + } + + /* clear out the allocated field, done with it now in any case. */ + if (cur) { + cur->bc_private.b.allocated = 0; + *curp = cur; + } + xfs_bmap_check_leaf_extents(cur, ip, whichfork); done: - *logflagsp = rval; + *logflagsp |= rval; return error; } @@ -4792,14 +4713,22 @@ xfs_bmapi_allocate( xfs_sb_version_hasextflgbit(&mp->m_sb)) bma->gotp->br_state = XFS_EXT_UNWRITTEN; - error = xfs_bmap_add_extent(bma->tp, bma->ip, lastx, cur, bma->gotp, - firstblock, flist, logflags, whichfork); + if (bma->wasdel) { + error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, lastx, + cur, bma->gotp, firstblock, flist, logflags); + } else { + error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, lastx, + cur, bma->gotp, firstblock, flist, logflags, + whichfork); + } + if (error) return error; /* - * Update our extent pointer, given that xfs_bmap_add_extent might - * have merged it into one of the neighbouring ones. + * Update our extent pointer, given that xfs_bmap_add_extent_delay_real + * or xfs_bmap_add_extent_hole_real might have merged it into one of + * the neighbouring ones. */ xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), bma->gotp); @@ -4854,14 +4783,15 @@ xfs_bmapi_convert_unwritten( mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; - error = xfs_bmap_add_extent(bma->tp, bma->ip, lastx, cur, mval, - firstblock, flist, logflags, whichfork); + error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, lastx, + cur, mval, firstblock, flist, logflags); if (error) return error; /* - * Update our extent pointer, given that xfs_bmap_add_extent might - * have merged it into one of the neighbouring ones. + * Update our extent pointer, given that + * xfs_bmap_add_extent_unwritten_real might have merged it into one + * of the neighbouring ones. */ xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), bma->gotp); @@ -5287,9 +5217,9 @@ xfs_bunmapi( del.br_blockcount = mod; } del.br_state = XFS_EXT_UNWRITTEN; - error = xfs_bmap_add_extent(tp, ip, &lastx, &cur, &del, - firstblock, flist, &logflags, - XFS_DATA_FORK); + error = xfs_bmap_add_extent_unwritten_real(tp, ip, + &lastx, &cur, &del, firstblock, flist, + &logflags); if (error) goto error0; goto nodelete; @@ -5345,18 +5275,18 @@ xfs_bunmapi( } prev.br_state = XFS_EXT_UNWRITTEN; lastx--; - error = xfs_bmap_add_extent(tp, ip, &lastx, - &cur, &prev, firstblock, flist, - &logflags, XFS_DATA_FORK); + error = xfs_bmap_add_extent_unwritten_real(tp, + ip, &lastx, &cur, &prev, + firstblock, flist, &logflags); if (error) goto error0; goto nodelete; } else { ASSERT(del.br_state == XFS_EXT_NORM); del.br_state = XFS_EXT_UNWRITTEN; - error = xfs_bmap_add_extent(tp, ip, &lastx, - &cur, &del, firstblock, flist, - &logflags, XFS_DATA_FORK); + error = xfs_bmap_add_extent_unwritten_real(tp, + ip, &lastx, &cur, &del, + firstblock, flist, &logflags); if (error) goto error0; goto nodelete; From 1b16447ba24ae39c7fe7133fcdcb4f174dec1901 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:55 +0000 Subject: [PATCH 35/69] xfs: pass bmalloca structure to xfs_bmap_isaeof All the variables xfs_bmap_isaeof() is passed are contained within the xfs_bmalloca structure. Pass that instead. Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 0c9be0ed606e..e2eb3ba5b420 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -3867,22 +3867,21 @@ xfs_bmap_last_extent( * blocks at the end of the file which do not start at the previous data block, * we will try to align the new blocks at stripe unit boundaries. * - * Returns 0 in *aeof if the file (fork) is empty as any new write will be at, - * or past the EOF. + * Returns 0 in bma->aeof if the file (fork) is empty as any new write will be + * at, or past the EOF. */ STATIC int xfs_bmap_isaeof( - struct xfs_inode *ip, - xfs_fileoff_t off, - int whichfork, - char *aeof) + struct xfs_bmalloca *bma, + int whichfork) { struct xfs_bmbt_irec rec; int is_empty; int error; - *aeof = 0; - error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty); + bma->aeof = 0; + error = xfs_bmap_last_extent(NULL, bma->ip, whichfork, &rec, + &is_empty); if (error || is_empty) return error; @@ -3890,8 +3889,9 @@ xfs_bmap_isaeof( * Check if we are allocation or past the last extent, or at least into * the last delayed allocated extent. */ - *aeof = off >= rec.br_startoff + rec.br_blockcount || - (off >= rec.br_startoff && isnullstartblock(rec.br_startblock)); + bma->aeof = bma->off >= rec.br_startoff + rec.br_blockcount || + (bma->off >= rec.br_startoff && + isnullstartblock(rec.br_startblock)); return 0; } @@ -4658,7 +4658,7 @@ xfs_bmapi_allocate( */ if (mp->m_dalign && alen >= mp->m_dalign && !(flags & XFS_BMAPI_METADATA) && whichfork == XFS_DATA_FORK) { - error = xfs_bmap_isaeof(bma->ip, aoff, whichfork, &bma->aeof); + error = xfs_bmap_isaeof(bma, whichfork); if (error) return error; } From baf41a52b9c62f9a825371806129ed12e2c1e2d8 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:56 +0000 Subject: [PATCH 36/69] xfs: move extent records into bmalloca structure Rather that putting extent records on the stack and then pointing to them in the bmalloca structure which is in the same stack frame, put the extent records directly in the bmalloca structure. This reduces the number of args that need to be passed around. Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 93 +++++++++++++++++++++++------------------------ fs/xfs/xfs_bmap.h | 4 +- 2 files changed, 47 insertions(+), 50 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index e2eb3ba5b420..68edf99bfad6 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2013,18 +2013,18 @@ xfs_bmap_adjacent( * If allocating at eof, and there's a previous real block, * try to use its last block as our starting point. */ - if (ap->eof && ap->prevp->br_startoff != NULLFILEOFF && - !isnullstartblock(ap->prevp->br_startblock) && - ISVALID(ap->prevp->br_startblock + ap->prevp->br_blockcount, - ap->prevp->br_startblock)) { - ap->rval = ap->prevp->br_startblock + ap->prevp->br_blockcount; + if (ap->eof && ap->prev.br_startoff != NULLFILEOFF && + !isnullstartblock(ap->prev.br_startblock) && + ISVALID(ap->prev.br_startblock + ap->prev.br_blockcount, + ap->prev.br_startblock)) { + ap->rval = ap->prev.br_startblock + ap->prev.br_blockcount; /* * Adjust for the gap between prevp and us. */ adjust = ap->off - - (ap->prevp->br_startoff + ap->prevp->br_blockcount); + (ap->prev.br_startoff + ap->prev.br_blockcount); if (adjust && - ISVALID(ap->rval + adjust, ap->prevp->br_startblock)) + ISVALID(ap->rval + adjust, ap->prev.br_startblock)) ap->rval += adjust; } /* @@ -2042,17 +2042,17 @@ xfs_bmap_adjacent( * If there's a previous (left) block, select a requested * start block based on it. */ - if (ap->prevp->br_startoff != NULLFILEOFF && - !isnullstartblock(ap->prevp->br_startblock) && - (prevbno = ap->prevp->br_startblock + - ap->prevp->br_blockcount) && - ISVALID(prevbno, ap->prevp->br_startblock)) { + if (ap->prev.br_startoff != NULLFILEOFF && + !isnullstartblock(ap->prev.br_startblock) && + (prevbno = ap->prev.br_startblock + + ap->prev.br_blockcount) && + ISVALID(prevbno, ap->prev.br_startblock)) { /* * Calculate gap to end of previous block. */ adjust = prevdiff = ap->off - - (ap->prevp->br_startoff + - ap->prevp->br_blockcount); + (ap->prev.br_startoff + + ap->prev.br_blockcount); /* * Figure the startblock based on the previous block's * end and the gap size. @@ -2063,7 +2063,7 @@ xfs_bmap_adjacent( */ if (prevdiff <= XFS_ALLOC_GAP_UNITS * ap->alen && ISVALID(prevbno + prevdiff, - ap->prevp->br_startblock)) + ap->prev.br_startblock)) prevbno += adjust; else prevdiff += adjust; @@ -2084,16 +2084,16 @@ xfs_bmap_adjacent( * If there's a following (right) block, select a requested * start block based on it. */ - if (!isnullstartblock(ap->gotp->br_startblock)) { + if (!isnullstartblock(ap->got.br_startblock)) { /* * Calculate gap to start of next block. */ - adjust = gotdiff = ap->gotp->br_startoff - ap->off; + adjust = gotdiff = ap->got.br_startoff - ap->off; /* * Figure the startblock based on the next block's * start and the gap size. */ - gotbno = ap->gotp->br_startblock; + gotbno = ap->got.br_startblock; /* * Heuristic! * If the gap is large relative to the piece we're @@ -2151,7 +2151,7 @@ xfs_bmap_rtalloc( mp = ap->ip->i_mount; align = xfs_get_extsz_hint(ap->ip); prod = align / mp->m_sb.sb_rextsize; - error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp, + error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, align, 1, ap->eof, 0, ap->conv, &ap->off, &ap->alen); if (error) @@ -2374,7 +2374,7 @@ xfs_bmap_btalloc( mp = ap->ip->i_mount; align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0; if (unlikely(align)) { - error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp, + error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, align, 0, ap->eof, 0, ap->conv, &ap->off, &ap->alen); ASSERT(!error); @@ -4619,17 +4619,17 @@ xfs_bmapi_allocate( * for in this bmap call but that wouldn't be as good. */ if (bma->wasdel) { - alen = (xfs_extlen_t)bma->gotp->br_blockcount; - aoff = bma->gotp->br_startoff; + alen = (xfs_extlen_t)bma->got.br_blockcount; + aoff = bma->got.br_startoff; if (*lastx != NULLEXTNUM && *lastx) { xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx - 1), - bma->prevp); + &bma->prev); } } else { alen = (xfs_extlen_t)XFS_FILBLKS_MIN(bma->alen, MAXEXTLEN); if (!bma->eof) alen = (xfs_extlen_t)XFS_FILBLKS_MIN(alen, - bma->gotp->br_startoff - bma->off); + bma->got.br_startoff - bma->off); aoff = bma->off; } @@ -4700,10 +4700,10 @@ xfs_bmapi_allocate( (*cur)->bc_private.b.flags = bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0; - bma->gotp->br_startoff = aoff; - bma->gotp->br_startblock = abno; - bma->gotp->br_blockcount = alen; - bma->gotp->br_state = XFS_EXT_NORM; + bma->got.br_startoff = aoff; + bma->got.br_startblock = abno; + bma->got.br_blockcount = alen; + bma->got.br_state = XFS_EXT_NORM; /* * A wasdelay extent has been initialized, so shouldn't be flagged @@ -4711,14 +4711,14 @@ xfs_bmapi_allocate( */ if (!bma->wasdel && (flags & XFS_BMAPI_PREALLOC) && xfs_sb_version_hasextflgbit(&mp->m_sb)) - bma->gotp->br_state = XFS_EXT_UNWRITTEN; + bma->got.br_state = XFS_EXT_UNWRITTEN; if (bma->wasdel) { error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, lastx, - cur, bma->gotp, firstblock, flist, logflags); + cur, &bma->got, firstblock, flist, logflags); } else { error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, lastx, - cur, bma->gotp, firstblock, flist, logflags, + cur, &bma->got, firstblock, flist, logflags, whichfork); } @@ -4730,13 +4730,12 @@ xfs_bmapi_allocate( * or xfs_bmap_add_extent_hole_real might have merged it into one of * the neighbouring ones. */ - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), bma->gotp); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), &bma->got); - ASSERT(bma->gotp->br_startoff <= aoff); - ASSERT(bma->gotp->br_startoff + bma->gotp->br_blockcount >= - aoff + alen); - ASSERT(bma->gotp->br_state == XFS_EXT_NORM || - bma->gotp->br_state == XFS_EXT_UNWRITTEN); + ASSERT(bma->got.br_startoff <= aoff); + ASSERT(bma->got.br_startoff + bma->got.br_blockcount >= aoff + alen); + ASSERT(bma->got.br_state == XFS_EXT_NORM || + bma->got.br_state == XFS_EXT_UNWRITTEN); return 0; } @@ -4793,7 +4792,7 @@ xfs_bmapi_convert_unwritten( * xfs_bmap_add_extent_unwritten_real might have merged it into one * of the neighbouring ones. */ - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), bma->gotp); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), &bma->got); /* * We may have combined previously unwritten space with written space, @@ -4837,14 +4836,12 @@ xfs_bmapi_write( xfs_fileoff_t end; /* end of mapped file region */ int eof; /* after the end of extents */ int error; /* error return */ - struct xfs_bmbt_irec got; /* current file extent record */ xfs_extnum_t lastx; /* last useful extent number */ int logflags; /* flags for transaction logging */ xfs_extlen_t minleft; /* min blocks left after allocation */ int n; /* current extent index */ int nallocs; /* number of extents alloc'd */ xfs_fileoff_t obno; /* old block number (offset) */ - struct xfs_bmbt_irec prev; /* previous file extent record */ int tmp_logflags; /* temp flags holder */ int whichfork; /* data or attr fork */ char inhole; /* current location is hole in file */ @@ -4916,21 +4913,20 @@ xfs_bmapi_write( goto error0; } - xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev); + xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &bma.got, + &bma.prev); n = 0; end = bno + len; obno = bno; bma.tp = tp; bma.ip = ip; - bma.prevp = &prev; - bma.gotp = &got; bma.total = total; bma.userdata = 0; while (bno < end && n < *nmap) { - inhole = eof || got.br_startoff > bno; - wasdelay = !inhole && isnullstartblock(got.br_startblock); + inhole = eof || bma.got.br_startoff > bno; + wasdelay = !inhole && isnullstartblock(bma.got.br_startblock); /* * First, deal with the hole before the allocated space @@ -4957,7 +4953,8 @@ xfs_bmapi_write( } /* Deal with the allocated space we found. */ - xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); + xfs_bmapi_trim_map(mval, &bma.got, &bno, len, obno, + end, n, flags); /* Execute unwritten extent conversion if necessary */ error = xfs_bmapi_convert_unwritten(&bma, mval, len, &lastx, @@ -4981,9 +4978,9 @@ xfs_bmapi_write( break; /* Else go on to the next record. */ - prev = got; + bma.prev = bma.got; if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &bma.got); else eof = 1; } diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index be235f56d2a6..6e8f8aee7cdb 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -114,8 +114,8 @@ typedef struct xfs_bmalloca { xfs_fileoff_t off; /* offset in file filling in */ struct xfs_trans *tp; /* transaction pointer */ struct xfs_inode *ip; /* incore inode pointer */ - struct xfs_bmbt_irec *prevp; /* extent before the new one */ - struct xfs_bmbt_irec *gotp; /* extent after, or delayed */ + struct xfs_bmbt_irec prev; /* extent before the new one */ + struct xfs_bmbt_irec got; /* extent after, or delayed */ xfs_extlen_t alen; /* i/o length asked/allocated */ xfs_extlen_t total; /* total blocks needed for xaction */ xfs_extlen_t minlen; /* minimum allocation size (blocks) */ From 0937e0fd8be6f9c26844127d39d677bb752e8741 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:57 +0000 Subject: [PATCH 37/69] xfs: move firstblock and bmap freelist cursor into bmalloca structure Rather than passing the firstblock and freelist structure around, embed it into the bmalloca structure and remove it from the function parameters. This also enables the minleft parameter to be set only once in xfs_bmapi_write(), and the freelist cursor directly queried in xfs_bmapi_allocate to clear it when the lowspace algorithm is activated. Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 92 ++++++++++++++++++++--------------------- fs/xfs/xfs_bmap.h | 4 +- fs/xfs/xfs_filestream.c | 2 +- 3 files changed, 48 insertions(+), 50 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 68edf99bfad6..608a0013791b 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2006,9 +2006,9 @@ xfs_bmap_adjacent( XFS_FSB_TO_AGBNO(mp, x) < mp->m_sb.sb_agblocks) mp = ap->ip->i_mount; - nullfb = ap->firstblock == NULLFSBLOCK; + nullfb = *ap->firstblock == NULLFSBLOCK; rt = XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata; - fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock); + fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock); /* * If allocating at eof, and there's a previous real block, * try to use its last block as our starting point. @@ -2380,8 +2380,8 @@ xfs_bmap_btalloc( ASSERT(!error); ASSERT(ap->alen); } - nullfb = ap->firstblock == NULLFSBLOCK; - fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock); + nullfb = *ap->firstblock == NULLFSBLOCK; + fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock); if (nullfb) { if (ap->userdata && xfs_inode_is_filestream(ap->ip)) { ag = xfs_filestream_lookup_ag(ap->ip); @@ -2391,7 +2391,7 @@ xfs_bmap_btalloc( ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino); } } else - ap->rval = ap->firstblock; + ap->rval = *ap->firstblock; xfs_bmap_adjacent(ap); @@ -2402,7 +2402,7 @@ xfs_bmap_btalloc( if (nullfb || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno) ; else - ap->rval = ap->firstblock; + ap->rval = *ap->firstblock; /* * Normal allocation, done through xfs_alloc_vextent. */ @@ -2413,13 +2413,13 @@ xfs_bmap_btalloc( /* Trim the allocation back to the maximum an AG can fit. */ args.maxlen = MIN(ap->alen, XFS_ALLOC_AG_MAX_USABLE(mp)); - args.firstblock = ap->firstblock; + args.firstblock = *ap->firstblock; blen = 0; if (nullfb) { error = xfs_bmap_btalloc_nullfb(ap, &args, &blen); if (error) return error; - } else if (ap->low) { + } else if (ap->flist->xbf_low) { if (xfs_inode_is_filestream(ap->ip)) args.type = XFS_ALLOCTYPE_FIRST_AG; else @@ -2452,7 +2452,7 @@ xfs_bmap_btalloc( * is >= the stripe unit and the allocation offset is * at the end of file. */ - if (!ap->low && ap->aeof) { + if (!ap->flist->xbf_low && ap->aeof) { if (!ap->off) { args.alignment = mp->m_dalign; atype = args.type; @@ -2540,12 +2540,25 @@ xfs_bmap_btalloc( args.minleft = 0; if ((error = xfs_alloc_vextent(&args))) return error; - ap->low = 1; + ap->flist->xbf_low = 1; } if (args.fsbno != NULLFSBLOCK) { - ap->firstblock = ap->rval = args.fsbno; + /* + * check the allocation happened at the same or higher AG than + * the first block that was allocated. + */ + ASSERT(*ap->firstblock == NULLFSBLOCK || + XFS_FSB_TO_AGNO(mp, *ap->firstblock) == + XFS_FSB_TO_AGNO(mp, args.fsbno) || + (ap->flist->xbf_low && + XFS_FSB_TO_AGNO(mp, *ap->firstblock) < + XFS_FSB_TO_AGNO(mp, args.fsbno))); + + ap->rval = args.fsbno; + if (*ap->firstblock == NULLFSBLOCK) + *ap->firstblock = args.fsbno; ASSERT(nullfb || fb_agno == args.agno || - (ap->low && fb_agno < args.agno)); + (ap->flist->xbf_low && fb_agno < args.agno)); ap->alen = args.len; ap->ip->i_d.di_nblocks += args.len; xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); @@ -4596,8 +4609,6 @@ xfs_bmapi_allocate( struct xfs_bmalloca *bma, xfs_extnum_t *lastx, struct xfs_btree_cur **cur, - xfs_fsblock_t *firstblock, - struct xfs_bmap_free *flist, int flags, int *nallocs, int *logflags) @@ -4647,9 +4658,7 @@ xfs_bmapi_allocate( */ bma->alen = alen; bma->off = aoff; - bma->firstblock = *firstblock; bma->minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; - bma->low = flist->xbf_low; bma->aeof = 0; /* @@ -4671,24 +4680,18 @@ xfs_bmapi_allocate( * Copy out result fields. */ abno = bma->rval; - flist->xbf_low = bma->low; alen = bma->alen; aoff = bma->off; - ASSERT(*firstblock == NULLFSBLOCK || - XFS_FSB_TO_AGNO(mp, *firstblock) == - XFS_FSB_TO_AGNO(mp, bma->firstblock) || - (flist->xbf_low && - XFS_FSB_TO_AGNO(mp, *firstblock) < - XFS_FSB_TO_AGNO(mp, bma->firstblock))); - *firstblock = bma->firstblock; + if (bma->flist->xbf_low) + bma->minleft = 0; if (*cur) - (*cur)->bc_private.b.firstblock = *firstblock; + (*cur)->bc_private.b.firstblock = *bma->firstblock; if (abno == NULLFSBLOCK) return 0; if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { (*cur) = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork); - (*cur)->bc_private.b.firstblock = *firstblock; - (*cur)->bc_private.b.flist = flist; + (*cur)->bc_private.b.firstblock = *bma->firstblock; + (*cur)->bc_private.b.flist = bma->flist; } /* * Bump the number of extents we've allocated @@ -4715,11 +4718,12 @@ xfs_bmapi_allocate( if (bma->wasdel) { error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, lastx, - cur, &bma->got, firstblock, flist, logflags); + cur, &bma->got, bma->firstblock, bma->flist, + logflags); } else { error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, lastx, - cur, &bma->got, firstblock, flist, logflags, - whichfork); + cur, &bma->got, bma->firstblock, bma->flist, + logflags, whichfork); } if (error) @@ -4746,8 +4750,6 @@ xfs_bmapi_convert_unwritten( xfs_filblks_t len, xfs_extnum_t *lastx, struct xfs_btree_cur **cur, - xfs_fsblock_t *firstblock, - struct xfs_bmap_free *flist, int flags, int *logflags) { @@ -4776,14 +4778,14 @@ xfs_bmapi_convert_unwritten( if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { *cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp, bma->ip, whichfork); - (*cur)->bc_private.b.firstblock = *firstblock; - (*cur)->bc_private.b.flist = flist; + (*cur)->bc_private.b.firstblock = *bma->firstblock; + (*cur)->bc_private.b.flist = bma->flist; } mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, lastx, - cur, mval, firstblock, flist, logflags); + cur, mval, bma->firstblock, bma->flist, logflags); if (error) return error; @@ -4838,7 +4840,6 @@ xfs_bmapi_write( int error; /* error return */ xfs_extnum_t lastx; /* last useful extent number */ int logflags; /* flags for transaction logging */ - xfs_extlen_t minleft; /* min blocks left after allocation */ int n; /* current extent index */ int nallocs; /* number of extents alloc'd */ xfs_fileoff_t obno; /* old block number (offset) */ @@ -4900,11 +4901,11 @@ xfs_bmapi_write( if (*firstblock == NULLFSBLOCK) { if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE) - minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1; + bma.minleft = be16_to_cpu(ifp->if_broot->bb_level) + 1; else - minleft = 1; + bma.minleft = 1; } else { - minleft = 0; + bma.minleft = 0; } if (!(ifp->if_flags & XFS_IFEXTENTS)) { @@ -4923,6 +4924,8 @@ xfs_bmapi_write( bma.ip = ip; bma.total = total; bma.userdata = 0; + bma.flist = flist; + bma.firstblock = firstblock; while (bno < end && n < *nmap) { inhole = eof || bma.got.br_startoff > bno; @@ -4938,16 +4941,12 @@ xfs_bmapi_write( bma.wasdel = wasdelay; bma.alen = len; bma.off = bno; - bma.minleft = minleft; - error = xfs_bmapi_allocate(&bma, &lastx, &cur, - firstblock, flist, flags, &nallocs, - &tmp_logflags); + error = xfs_bmapi_allocate(&bma, &lastx, &cur, flags, + &nallocs, &tmp_logflags); logflags |= tmp_logflags; if (error) goto error0; - if (flist && flist->xbf_low) - minleft = 0; if (bma.rval == NULLFSBLOCK) break; } @@ -4958,8 +4957,7 @@ xfs_bmapi_write( /* Execute unwritten extent conversion if necessary */ error = xfs_bmapi_convert_unwritten(&bma, mval, len, &lastx, - &cur, firstblock, flist, - flags, &tmp_logflags); + &cur, flags, &tmp_logflags); logflags |= tmp_logflags; if (error == EAGAIN) continue; diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 6e8f8aee7cdb..6e7c7a50d248 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -109,7 +109,8 @@ static inline void xfs_bmap_init(xfs_bmap_free_t *flp, xfs_fsblock_t *fbp) * Argument structure for xfs_bmap_alloc. */ typedef struct xfs_bmalloca { - xfs_fsblock_t firstblock; /* i/o first block allocated */ + xfs_fsblock_t *firstblock; /* i/o first block allocated */ + struct xfs_bmap_free *flist; /* bmap freelist */ xfs_fsblock_t rval; /* starting block of new extent */ xfs_fileoff_t off; /* offset in file filling in */ struct xfs_trans *tp; /* transaction pointer */ @@ -123,7 +124,6 @@ typedef struct xfs_bmalloca { char eof; /* set if allocating past last extent */ char wasdel; /* replacing a delayed allocation */ char userdata;/* set if is user data */ - char low; /* low on space, using seq'l ags */ char aeof; /* allocated space at eof */ char conv; /* overwriting unwritten extents */ } xfs_bmalloca_t; diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index 3ff3d9e23ded..137f957b8c7e 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -761,7 +761,7 @@ xfs_filestream_new_ag( */ ag = (ag == NULLAGNUMBER) ? 0 : (ag + 1) % mp->m_sb.sb_agcount; flags = (ap->userdata ? XFS_PICK_USERDATA : 0) | - (ap->low ? XFS_PICK_LOWSPACE : 0); + (ap->flist->xbf_low ? XFS_PICK_LOWSPACE : 0); err = _xfs_filestream_pick_ag(mp, ag, agp, flags, minlen); if (err || *agp == NULLAGNUMBER) From 3a75667e902dbdb87718b1ee2b3b745b344a8163 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:58 +0000 Subject: [PATCH 38/69] xfs: rename allocation range fields in struct xfs_bmalloca Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 138 ++++++++++++++++++++-------------------- fs/xfs/xfs_bmap.h | 8 ++- fs/xfs/xfs_filestream.c | 2 +- 3 files changed, 75 insertions(+), 73 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 608a0013791b..b47555cfbd8f 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2017,15 +2017,15 @@ xfs_bmap_adjacent( !isnullstartblock(ap->prev.br_startblock) && ISVALID(ap->prev.br_startblock + ap->prev.br_blockcount, ap->prev.br_startblock)) { - ap->rval = ap->prev.br_startblock + ap->prev.br_blockcount; + ap->blkno = ap->prev.br_startblock + ap->prev.br_blockcount; /* * Adjust for the gap between prevp and us. */ - adjust = ap->off - + adjust = ap->offset - (ap->prev.br_startoff + ap->prev.br_blockcount); if (adjust && - ISVALID(ap->rval + adjust, ap->prev.br_startblock)) - ap->rval += adjust; + ISVALID(ap->blkno + adjust, ap->prev.br_startblock)) + ap->blkno += adjust; } /* * If not at eof, then compare the two neighbor blocks. @@ -2050,7 +2050,7 @@ xfs_bmap_adjacent( /* * Calculate gap to end of previous block. */ - adjust = prevdiff = ap->off - + adjust = prevdiff = ap->offset - (ap->prev.br_startoff + ap->prev.br_blockcount); /* @@ -2061,7 +2061,7 @@ xfs_bmap_adjacent( * allocating, or using it gives us an invalid block * number, then just use the end of the previous block. */ - if (prevdiff <= XFS_ALLOC_GAP_UNITS * ap->alen && + if (prevdiff <= XFS_ALLOC_GAP_UNITS * ap->length && ISVALID(prevbno + prevdiff, ap->prev.br_startblock)) prevbno += adjust; @@ -2088,7 +2088,7 @@ xfs_bmap_adjacent( /* * Calculate gap to start of next block. */ - adjust = gotdiff = ap->got.br_startoff - ap->off; + adjust = gotdiff = ap->got.br_startoff - ap->offset; /* * Figure the startblock based on the next block's * start and the gap size. @@ -2101,12 +2101,12 @@ xfs_bmap_adjacent( * number, then just use the start of the next block * offset by our length. */ - if (gotdiff <= XFS_ALLOC_GAP_UNITS * ap->alen && + if (gotdiff <= XFS_ALLOC_GAP_UNITS * ap->length && ISVALID(gotbno - gotdiff, gotbno)) gotbno -= adjust; - else if (ISVALID(gotbno - ap->alen, gotbno)) { - gotbno -= ap->alen; - gotdiff += adjust - ap->alen; + else if (ISVALID(gotbno - ap->length, gotbno)) { + gotbno -= ap->length; + gotdiff += adjust - ap->length; } else gotdiff += adjust; /* @@ -2124,14 +2124,14 @@ xfs_bmap_adjacent( gotbno = NULLFSBLOCK; /* * If both valid, pick the better one, else the only good - * one, else ap->rval is already set (to 0 or the inode block). + * one, else ap->blkno is already set (to 0 or the inode block). */ if (prevbno != NULLFSBLOCK && gotbno != NULLFSBLOCK) - ap->rval = prevdiff <= gotdiff ? prevbno : gotbno; + ap->blkno = prevdiff <= gotdiff ? prevbno : gotbno; else if (prevbno != NULLFSBLOCK) - ap->rval = prevbno; + ap->blkno = prevbno; else if (gotbno != NULLFSBLOCK) - ap->rval = gotbno; + ap->blkno = gotbno; } #undef ISVALID } @@ -2153,22 +2153,22 @@ xfs_bmap_rtalloc( prod = align / mp->m_sb.sb_rextsize; error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, align, 1, ap->eof, 0, - ap->conv, &ap->off, &ap->alen); + ap->conv, &ap->offset, &ap->length); if (error) return error; - ASSERT(ap->alen); - ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0); + ASSERT(ap->length); + ASSERT(ap->length % mp->m_sb.sb_rextsize == 0); /* * If the offset & length are not perfectly aligned * then kill prod, it will just get us in trouble. */ - if (do_mod(ap->off, align) || ap->alen % align) + if (do_mod(ap->offset, align) || ap->length % align) prod = 1; /* * Set ralen to be the actual requested length in rtextents. */ - ralen = ap->alen / mp->m_sb.sb_rextsize; + ralen = ap->length / mp->m_sb.sb_rextsize; /* * If the old value was close enough to MAXEXTLEN that * we rounded up to it, cut it back so it's valid again. @@ -2189,15 +2189,15 @@ xfs_bmap_rtalloc( * If it's an allocation to an empty file at offset 0, * pick an extent that will space things out in the rt area. */ - if (ap->eof && ap->off == 0) { + if (ap->eof && ap->offset == 0) { xfs_rtblock_t uninitialized_var(rtx); /* realtime extent no */ error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx); if (error) return error; - ap->rval = rtx * mp->m_sb.sb_rextsize; + ap->blkno = rtx * mp->m_sb.sb_rextsize; } else { - ap->rval = 0; + ap->blkno = 0; } xfs_bmap_adjacent(ap); @@ -2205,23 +2205,23 @@ xfs_bmap_rtalloc( /* * Realtime allocation, done through xfs_rtallocate_extent. */ - atype = ap->rval == 0 ? XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO; - do_div(ap->rval, mp->m_sb.sb_rextsize); - rtb = ap->rval; - ap->alen = ralen; - if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen, + atype = ap->blkno == 0 ? XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO; + do_div(ap->blkno, mp->m_sb.sb_rextsize); + rtb = ap->blkno; + ap->length = ralen; + if ((error = xfs_rtallocate_extent(ap->tp, ap->blkno, 1, ap->length, &ralen, atype, ap->wasdel, prod, &rtb))) return error; if (rtb == NULLFSBLOCK && prod > 1 && - (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, - ap->alen, &ralen, atype, + (error = xfs_rtallocate_extent(ap->tp, ap->blkno, 1, + ap->length, &ralen, atype, ap->wasdel, 1, &rtb))) return error; - ap->rval = rtb; - if (ap->rval != NULLFSBLOCK) { - ap->rval *= mp->m_sb.sb_rextsize; + ap->blkno = rtb; + if (ap->blkno != NULLFSBLOCK) { + ap->blkno *= mp->m_sb.sb_rextsize; ralen *= mp->m_sb.sb_rextsize; - ap->alen = ralen; + ap->length = ralen; ap->ip->i_d.di_nblocks += ralen; xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); if (ap->wasdel) @@ -2234,7 +2234,7 @@ xfs_bmap_rtalloc( ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT : XFS_TRANS_DQ_RTBCOUNT, (long) ralen); } else { - ap->alen = 0; + ap->length = 0; } return 0; } @@ -2349,7 +2349,7 @@ xfs_bmap_btalloc_nullfb( * AG as the stream may have moved. */ if (xfs_inode_is_filestream(ap->ip)) - ap->rval = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0); + ap->blkno = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0); return 0; } @@ -2376,9 +2376,9 @@ xfs_bmap_btalloc( if (unlikely(align)) { error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, align, 0, ap->eof, 0, ap->conv, - &ap->off, &ap->alen); + &ap->offset, &ap->length); ASSERT(!error); - ASSERT(ap->alen); + ASSERT(ap->length); } nullfb = *ap->firstblock == NULLFSBLOCK; fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock); @@ -2386,33 +2386,33 @@ xfs_bmap_btalloc( if (ap->userdata && xfs_inode_is_filestream(ap->ip)) { ag = xfs_filestream_lookup_ag(ap->ip); ag = (ag != NULLAGNUMBER) ? ag : 0; - ap->rval = XFS_AGB_TO_FSB(mp, ag, 0); + ap->blkno = XFS_AGB_TO_FSB(mp, ag, 0); } else { - ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino); + ap->blkno = XFS_INO_TO_FSB(mp, ap->ip->i_ino); } } else - ap->rval = *ap->firstblock; + ap->blkno = *ap->firstblock; xfs_bmap_adjacent(ap); /* - * If allowed, use ap->rval; otherwise must use firstblock since + * If allowed, use ap->blkno; otherwise must use firstblock since * it's in the right allocation group. */ - if (nullfb || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno) + if (nullfb || XFS_FSB_TO_AGNO(mp, ap->blkno) == fb_agno) ; else - ap->rval = *ap->firstblock; + ap->blkno = *ap->firstblock; /* * Normal allocation, done through xfs_alloc_vextent. */ tryagain = isaligned = 0; args.tp = ap->tp; args.mp = mp; - args.fsbno = ap->rval; + args.fsbno = ap->blkno; /* Trim the allocation back to the maximum an AG can fit. */ - args.maxlen = MIN(ap->alen, XFS_ALLOC_AG_MAX_USABLE(mp)); + args.maxlen = MIN(ap->length, XFS_ALLOC_AG_MAX_USABLE(mp)); args.firstblock = *ap->firstblock; blen = 0; if (nullfb) { @@ -2433,14 +2433,14 @@ xfs_bmap_btalloc( /* apply extent size hints if obtained earlier */ if (unlikely(align)) { args.prod = align; - if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod))) + if ((args.mod = (xfs_extlen_t)do_mod(ap->offset, args.prod))) args.mod = (xfs_extlen_t)(args.prod - args.mod); } else if (mp->m_sb.sb_blocksize >= PAGE_CACHE_SIZE) { args.prod = 1; args.mod = 0; } else { args.prod = PAGE_CACHE_SIZE >> mp->m_sb.sb_blocklog; - if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod)))) + if ((args.mod = (xfs_extlen_t)(do_mod(ap->offset, args.prod)))) args.mod = (xfs_extlen_t)(args.prod - args.mod); } /* @@ -2453,7 +2453,7 @@ xfs_bmap_btalloc( * at the end of file. */ if (!ap->flist->xbf_low && ap->aeof) { - if (!ap->off) { + if (!ap->offset) { args.alignment = mp->m_dalign; atype = args.type; isaligned = 1; @@ -2506,7 +2506,7 @@ xfs_bmap_btalloc( * turned on. */ args.type = atype; - args.fsbno = ap->rval; + args.fsbno = ap->blkno; args.alignment = mp->m_dalign; args.minlen = nextminlen; args.minalignslop = 0; @@ -2520,7 +2520,7 @@ xfs_bmap_btalloc( * try again. */ args.type = atype; - args.fsbno = ap->rval; + args.fsbno = ap->blkno; args.alignment = 0; if ((error = xfs_alloc_vextent(&args))) return error; @@ -2529,7 +2529,7 @@ xfs_bmap_btalloc( args.minlen > ap->minlen) { args.minlen = ap->minlen; args.type = XFS_ALLOCTYPE_START_BNO; - args.fsbno = ap->rval; + args.fsbno = ap->blkno; if ((error = xfs_alloc_vextent(&args))) return error; } @@ -2554,12 +2554,12 @@ xfs_bmap_btalloc( XFS_FSB_TO_AGNO(mp, *ap->firstblock) < XFS_FSB_TO_AGNO(mp, args.fsbno))); - ap->rval = args.fsbno; + ap->blkno = args.fsbno; if (*ap->firstblock == NULLFSBLOCK) *ap->firstblock = args.fsbno; ASSERT(nullfb || fb_agno == args.agno || (ap->flist->xbf_low && fb_agno < args.agno)); - ap->alen = args.len; + ap->length = args.len; ap->ip->i_d.di_nblocks += args.len; xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); if (ap->wasdel) @@ -2573,8 +2573,8 @@ xfs_bmap_btalloc( XFS_TRANS_DQ_BCOUNT, (long) args.len); } else { - ap->rval = NULLFSBLOCK; - ap->alen = 0; + ap->blkno = NULLFSBLOCK; + ap->length = 0; } return 0; } @@ -3902,8 +3902,8 @@ xfs_bmap_isaeof( * Check if we are allocation or past the last extent, or at least into * the last delayed allocated extent. */ - bma->aeof = bma->off >= rec.br_startoff + rec.br_blockcount || - (bma->off >= rec.br_startoff && + bma->aeof = bma->offset >= rec.br_startoff + rec.br_blockcount || + (bma->offset >= rec.br_startoff && isnullstartblock(rec.br_startblock)); return 0; } @@ -4637,11 +4637,11 @@ xfs_bmapi_allocate( &bma->prev); } } else { - alen = (xfs_extlen_t)XFS_FILBLKS_MIN(bma->alen, MAXEXTLEN); + alen = (xfs_extlen_t)XFS_FILBLKS_MIN(bma->length, MAXEXTLEN); if (!bma->eof) alen = (xfs_extlen_t)XFS_FILBLKS_MIN(alen, - bma->got.br_startoff - bma->off); - aoff = bma->off; + bma->got.br_startoff - bma->offset); + aoff = bma->offset; } /* @@ -4656,8 +4656,8 @@ xfs_bmapi_allocate( /* * Fill in changeable bma fields. */ - bma->alen = alen; - bma->off = aoff; + bma->length = alen; + bma->offset = aoff; bma->minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; bma->aeof = 0; @@ -4679,9 +4679,9 @@ xfs_bmapi_allocate( /* * Copy out result fields. */ - abno = bma->rval; - alen = bma->alen; - aoff = bma->off; + abno = bma->blkno; + alen = bma->length; + aoff = bma->offset; if (bma->flist->xbf_low) bma->minleft = 0; if (*cur) @@ -4939,15 +4939,15 @@ xfs_bmapi_write( bma.eof = eof; bma.conv = !!(flags & XFS_BMAPI_CONVERT); bma.wasdel = wasdelay; - bma.alen = len; - bma.off = bno; + bma.length = len; + bma.offset = bno; error = xfs_bmapi_allocate(&bma, &lastx, &cur, flags, &nallocs, &tmp_logflags); logflags |= tmp_logflags; if (error) goto error0; - if (bma.rval == NULLFSBLOCK) + if (bma.blkno == NULLFSBLOCK) break; } diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 6e7c7a50d248..5f398b1ac70c 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -111,13 +111,15 @@ static inline void xfs_bmap_init(xfs_bmap_free_t *flp, xfs_fsblock_t *fbp) typedef struct xfs_bmalloca { xfs_fsblock_t *firstblock; /* i/o first block allocated */ struct xfs_bmap_free *flist; /* bmap freelist */ - xfs_fsblock_t rval; /* starting block of new extent */ - xfs_fileoff_t off; /* offset in file filling in */ struct xfs_trans *tp; /* transaction pointer */ struct xfs_inode *ip; /* incore inode pointer */ struct xfs_bmbt_irec prev; /* extent before the new one */ struct xfs_bmbt_irec got; /* extent after, or delayed */ - xfs_extlen_t alen; /* i/o length asked/allocated */ + + xfs_fileoff_t offset; /* offset in file filling in */ + xfs_extlen_t length; /* i/o length asked/allocated */ + xfs_fsblock_t blkno; /* starting block of new extent */ + xfs_extlen_t total; /* total blocks needed for xaction */ xfs_extlen_t minlen; /* minimum allocation size (blocks) */ xfs_extlen_t minleft; /* amount must be left after alloc */ diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index 137f957b8c7e..5170306a1009 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -682,7 +682,7 @@ xfs_filestream_new_ag( ip = ap->ip; mp = ip->i_mount; cache = mp->m_filestream; - minlen = ap->alen; + minlen = ap->length; *agp = NULLAGNUMBER; /* From 963c30cf45e8c832ae11438ff9d99c954b9d0114 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:40:59 +0000 Subject: [PATCH 39/69] xfs: do not keep local copies of allocation ranges in xfs_bmapi_allocate Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 43 ++++++++++++++----------------------------- 1 file changed, 14 insertions(+), 29 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index b47555cfbd8f..6a7832d3540e 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4617,9 +4617,6 @@ xfs_bmapi_allocate( int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? XFS_ATTR_FORK : XFS_DATA_FORK; struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); - xfs_fsblock_t abno; - xfs_extlen_t alen; - xfs_fileoff_t aoff; int error; int rt; @@ -4630,18 +4627,17 @@ xfs_bmapi_allocate( * for in this bmap call but that wouldn't be as good. */ if (bma->wasdel) { - alen = (xfs_extlen_t)bma->got.br_blockcount; - aoff = bma->got.br_startoff; + bma->length = (xfs_extlen_t)bma->got.br_blockcount; + bma->offset = bma->got.br_startoff; if (*lastx != NULLEXTNUM && *lastx) { xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx - 1), &bma->prev); } } else { - alen = (xfs_extlen_t)XFS_FILBLKS_MIN(bma->length, MAXEXTLEN); + bma->length = XFS_FILBLKS_MIN(bma->length, MAXEXTLEN); if (!bma->eof) - alen = (xfs_extlen_t)XFS_FILBLKS_MIN(alen, + bma->length = XFS_FILBLKS_MIN(bma->length, bma->got.br_startoff - bma->offset); - aoff = bma->offset; } /* @@ -4649,23 +4645,17 @@ xfs_bmapi_allocate( * user data. */ if (!(flags & XFS_BMAPI_METADATA)) { - bma->userdata = (aoff == 0) ? + bma->userdata = (bma->offset == 0) ? XFS_ALLOC_INITIAL_USER_DATA : XFS_ALLOC_USERDATA; } - /* - * Fill in changeable bma fields. - */ - bma->length = alen; - bma->offset = aoff; - bma->minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1; - bma->aeof = 0; + bma->minlen = (flags & XFS_BMAPI_CONTIG) ? bma->length : 1; /* * Only want to do the alignment at the eof if it is userdata and * allocation length is larger than a stripe unit. */ - if (mp->m_dalign && alen >= mp->m_dalign && + if (mp->m_dalign && bma->length >= mp->m_dalign && !(flags & XFS_BMAPI_METADATA) && whichfork == XFS_DATA_FORK) { error = xfs_bmap_isaeof(bma, whichfork); if (error) @@ -4676,17 +4666,11 @@ xfs_bmapi_allocate( if (error) return error; - /* - * Copy out result fields. - */ - abno = bma->blkno; - alen = bma->length; - aoff = bma->offset; if (bma->flist->xbf_low) bma->minleft = 0; if (*cur) (*cur)->bc_private.b.firstblock = *bma->firstblock; - if (abno == NULLFSBLOCK) + if (bma->blkno == NULLFSBLOCK) return 0; if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { (*cur) = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork); @@ -4703,9 +4687,9 @@ xfs_bmapi_allocate( (*cur)->bc_private.b.flags = bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0; - bma->got.br_startoff = aoff; - bma->got.br_startblock = abno; - bma->got.br_blockcount = alen; + bma->got.br_startoff = bma->offset; + bma->got.br_startblock = bma->blkno; + bma->got.br_blockcount = bma->length; bma->got.br_state = XFS_EXT_NORM; /* @@ -4736,8 +4720,9 @@ xfs_bmapi_allocate( */ xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), &bma->got); - ASSERT(bma->got.br_startoff <= aoff); - ASSERT(bma->got.br_startoff + bma->got.br_blockcount >= aoff + alen); + ASSERT(bma->got.br_startoff <= bma->offset); + ASSERT(bma->got.br_startoff + bma->got.br_blockcount >= + bma->offset + bma->length); ASSERT(bma->got.br_state == XFS_EXT_NORM || bma->got.br_state == XFS_EXT_UNWRITTEN); return 0; From 29c8d17a8938be88e36b93522753f3519aefd05d Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:41:00 +0000 Subject: [PATCH 40/69] xfs: move btree cursor into bmalloca Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 56 ++++++++++++++++++++++------------------------- fs/xfs/xfs_bmap.h | 2 ++ 2 files changed, 28 insertions(+), 30 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 6a7832d3540e..a54326bf9aa0 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4608,7 +4608,6 @@ STATIC int xfs_bmapi_allocate( struct xfs_bmalloca *bma, xfs_extnum_t *lastx, - struct xfs_btree_cur **cur, int flags, int *nallocs, int *logflags) @@ -4668,14 +4667,14 @@ xfs_bmapi_allocate( if (bma->flist->xbf_low) bma->minleft = 0; - if (*cur) - (*cur)->bc_private.b.firstblock = *bma->firstblock; + if (bma->cur) + bma->cur->bc_private.b.firstblock = *bma->firstblock; if (bma->blkno == NULLFSBLOCK) return 0; - if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { - (*cur) = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork); - (*cur)->bc_private.b.firstblock = *bma->firstblock; - (*cur)->bc_private.b.flist = bma->flist; + if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) { + bma->cur = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork); + bma->cur->bc_private.b.firstblock = *bma->firstblock; + bma->cur->bc_private.b.flist = bma->flist; } /* * Bump the number of extents we've allocated @@ -4683,8 +4682,8 @@ xfs_bmapi_allocate( */ (*nallocs)++; - if (*cur) - (*cur)->bc_private.b.flags = + if (bma->cur) + bma->cur->bc_private.b.flags = bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0; bma->got.br_startoff = bma->offset; @@ -4702,12 +4701,12 @@ xfs_bmapi_allocate( if (bma->wasdel) { error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, lastx, - cur, &bma->got, bma->firstblock, bma->flist, - logflags); + &bma->cur, &bma->got, bma->firstblock, + bma->flist, logflags); } else { error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, lastx, - cur, &bma->got, bma->firstblock, bma->flist, - logflags, whichfork); + &bma->cur, &bma->got, bma->firstblock, + bma->flist, logflags, whichfork); } if (error) @@ -4734,7 +4733,6 @@ xfs_bmapi_convert_unwritten( struct xfs_bmbt_irec *mval, xfs_filblks_t len, xfs_extnum_t *lastx, - struct xfs_btree_cur **cur, int flags, int *logflags) { @@ -4760,17 +4758,17 @@ xfs_bmapi_convert_unwritten( * Modify (by adding) the state flag, if writing. */ ASSERT(mval->br_blockcount <= len); - if ((ifp->if_flags & XFS_IFBROOT) && !*cur) { - *cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp, + if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) { + bma->cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp, bma->ip, whichfork); - (*cur)->bc_private.b.firstblock = *bma->firstblock; - (*cur)->bc_private.b.flist = bma->flist; + bma->cur->bc_private.b.firstblock = *bma->firstblock; + bma->cur->bc_private.b.flist = bma->flist; } mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, lastx, - cur, mval, bma->firstblock, bma->flist, logflags); + &bma->cur, mval, bma->firstblock, bma->flist, logflags); if (error) return error; @@ -4819,7 +4817,6 @@ xfs_bmapi_write( struct xfs_mount *mp = ip->i_mount; struct xfs_ifork *ifp; struct xfs_bmalloca bma = { 0 }; /* args for xfs_bmap_alloc */ - struct xfs_btree_cur *cur; /* bmap btree cursor */ xfs_fileoff_t end; /* end of mapped file region */ int eof; /* after the end of extents */ int error; /* error return */ @@ -4875,7 +4872,6 @@ xfs_bmapi_write( logflags = 0; nallocs = 0; - cur = NULL; if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { error = xfs_bmap_local_to_extents(tp, ip, firstblock, total, @@ -4927,7 +4923,7 @@ xfs_bmapi_write( bma.length = len; bma.offset = bno; - error = xfs_bmapi_allocate(&bma, &lastx, &cur, flags, + error = xfs_bmapi_allocate(&bma, &lastx, flags, &nallocs, &tmp_logflags); logflags |= tmp_logflags; if (error) @@ -4942,7 +4938,7 @@ xfs_bmapi_write( /* Execute unwritten extent conversion if necessary */ error = xfs_bmapi_convert_unwritten(&bma, mval, len, &lastx, - &cur, flags, &tmp_logflags); + flags, &tmp_logflags); logflags |= tmp_logflags; if (error == EAGAIN) continue; @@ -4974,8 +4970,8 @@ xfs_bmapi_write( */ if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE && XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) { - ASSERT(cur); - error = xfs_bmap_btree_to_extents(tp, ip, cur, + ASSERT(bma.cur); + error = xfs_bmap_btree_to_extents(tp, ip, bma.cur, &tmp_logflags, whichfork); logflags |= tmp_logflags; if (error) @@ -5005,19 +5001,19 @@ xfs_bmapi_write( if (logflags) xfs_trans_log_inode(tp, ip, logflags); - if (cur) { + if (bma.cur) { if (!error) { ASSERT(*firstblock == NULLFSBLOCK || XFS_FSB_TO_AGNO(mp, *firstblock) == XFS_FSB_TO_AGNO(mp, - cur->bc_private.b.firstblock) || + bma.cur->bc_private.b.firstblock) || (flist->xbf_low && XFS_FSB_TO_AGNO(mp, *firstblock) < XFS_FSB_TO_AGNO(mp, - cur->bc_private.b.firstblock))); - *firstblock = cur->bc_private.b.firstblock; + bma.cur->bc_private.b.firstblock))); + *firstblock = bma.cur->bc_private.b.firstblock; } - xfs_btree_del_cursor(cur, + xfs_btree_del_cursor(bma.cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR); } if (!error) diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 5f398b1ac70c..858d9d509989 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -120,6 +120,8 @@ typedef struct xfs_bmalloca { xfs_extlen_t length; /* i/o length asked/allocated */ xfs_fsblock_t blkno; /* starting block of new extent */ + struct xfs_btree_cur *cur; /* btree cursor */ + xfs_extlen_t total; /* total blocks needed for xaction */ xfs_extlen_t minlen; /* minimum allocation size (blocks) */ xfs_extlen_t minleft; /* amount must be left after alloc */ From e0c3da5d89dc1aeef2275a8b751231e147603f0f Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Sun, 18 Sep 2011 20:41:01 +0000 Subject: [PATCH 41/69] xfs: move lastx and nallocs into bmalloca Signed-off-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 47 +++++++++++++++++++++-------------------------- fs/xfs/xfs_bmap.h | 2 ++ 2 files changed, 23 insertions(+), 26 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index a54326bf9aa0..3ef145307274 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4607,9 +4607,7 @@ xfs_bmapi_delay( STATIC int xfs_bmapi_allocate( struct xfs_bmalloca *bma, - xfs_extnum_t *lastx, int flags, - int *nallocs, int *logflags) { struct xfs_mount *mp = bma->ip->i_mount; @@ -4628,8 +4626,8 @@ xfs_bmapi_allocate( if (bma->wasdel) { bma->length = (xfs_extlen_t)bma->got.br_blockcount; bma->offset = bma->got.br_startoff; - if (*lastx != NULLEXTNUM && *lastx) { - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx - 1), + if (bma->idx != NULLEXTNUM && bma->idx) { + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &bma->prev); } } else { @@ -4680,7 +4678,7 @@ xfs_bmapi_allocate( * Bump the number of extents we've allocated * in this call. */ - (*nallocs)++; + bma->nallocs++; if (bma->cur) bma->cur->bc_private.b.flags = @@ -4700,13 +4698,14 @@ xfs_bmapi_allocate( bma->got.br_state = XFS_EXT_UNWRITTEN; if (bma->wasdel) { - error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, lastx, - &bma->cur, &bma->got, bma->firstblock, - bma->flist, logflags); + error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, + &bma->idx, &bma->cur, &bma->got, + bma->firstblock, bma->flist, logflags); } else { - error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, lastx, - &bma->cur, &bma->got, bma->firstblock, - bma->flist, logflags, whichfork); + error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, + &bma->idx, &bma->cur, &bma->got, + bma->firstblock, bma->flist, logflags, + whichfork); } if (error) @@ -4717,7 +4716,7 @@ xfs_bmapi_allocate( * or xfs_bmap_add_extent_hole_real might have merged it into one of * the neighbouring ones. */ - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), &bma->got); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &bma->got); ASSERT(bma->got.br_startoff <= bma->offset); ASSERT(bma->got.br_startoff + bma->got.br_blockcount >= @@ -4732,7 +4731,6 @@ xfs_bmapi_convert_unwritten( struct xfs_bmalloca *bma, struct xfs_bmbt_irec *mval, xfs_filblks_t len, - xfs_extnum_t *lastx, int flags, int *logflags) { @@ -4767,7 +4765,7 @@ xfs_bmapi_convert_unwritten( mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; - error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, lastx, + error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, &bma->idx, &bma->cur, mval, bma->firstblock, bma->flist, logflags); if (error) return error; @@ -4777,7 +4775,7 @@ xfs_bmapi_convert_unwritten( * xfs_bmap_add_extent_unwritten_real might have merged it into one * of the neighbouring ones. */ - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), &bma->got); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &bma->got); /* * We may have combined previously unwritten space with written space, @@ -4820,10 +4818,8 @@ xfs_bmapi_write( xfs_fileoff_t end; /* end of mapped file region */ int eof; /* after the end of extents */ int error; /* error return */ - xfs_extnum_t lastx; /* last useful extent number */ int logflags; /* flags for transaction logging */ int n; /* current extent index */ - int nallocs; /* number of extents alloc'd */ xfs_fileoff_t obno; /* old block number (offset) */ int tmp_logflags; /* temp flags holder */ int whichfork; /* data or attr fork */ @@ -4871,7 +4867,6 @@ xfs_bmapi_write( XFS_STATS_INC(xs_blk_mapw); logflags = 0; - nallocs = 0; if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { error = xfs_bmap_local_to_extents(tp, ip, firstblock, total, @@ -4895,7 +4890,7 @@ xfs_bmapi_write( goto error0; } - xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &bma.got, + xfs_bmap_search_extents(ip, bno, whichfork, &eof, &bma.idx, &bma.got, &bma.prev); n = 0; end = bno + len; @@ -4923,8 +4918,7 @@ xfs_bmapi_write( bma.length = len; bma.offset = bno; - error = xfs_bmapi_allocate(&bma, &lastx, flags, - &nallocs, &tmp_logflags); + error = xfs_bmapi_allocate(&bma, flags, &tmp_logflags); logflags |= tmp_logflags; if (error) goto error0; @@ -4937,7 +4931,7 @@ xfs_bmapi_write( end, n, flags); /* Execute unwritten extent conversion if necessary */ - error = xfs_bmapi_convert_unwritten(&bma, mval, len, &lastx, + error = xfs_bmapi_convert_unwritten(&bma, mval, len, flags, &tmp_logflags); logflags |= tmp_logflags; if (error == EAGAIN) @@ -4953,14 +4947,15 @@ xfs_bmapi_write( * XFS_BMAP_MAX_NMAP extents no matter what. Otherwise * the transaction may get too big. */ - if (bno >= end || n >= *nmap || nallocs >= *nmap) + if (bno >= end || n >= *nmap || bma.nallocs >= *nmap) break; /* Else go on to the next record. */ bma.prev = bma.got; - if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &bma.got); - else + if (++bma.idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) { + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma.idx), + &bma.got); + } else eof = 1; } *nmap = n; diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 858d9d509989..c2d9e1ef4ba3 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -121,6 +121,8 @@ typedef struct xfs_bmalloca { xfs_fsblock_t blkno; /* starting block of new extent */ struct xfs_btree_cur *cur; /* btree cursor */ + xfs_extnum_t idx; /* current extent index */ + int nallocs;/* number of extents alloc'd */ xfs_extlen_t total; /* total blocks needed for xaction */ xfs_extlen_t minlen; /* minimum allocation size (blocks) */ From c315c90b7d530d1ec3c226052e153b0cffa512c8 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:41:02 +0000 Subject: [PATCH 42/69] xfs: move logflags into bmalloca Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 48 ++++++++++++++++++++++------------------------- fs/xfs/xfs_bmap.h | 1 + 2 files changed, 23 insertions(+), 26 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 3ef145307274..ed094749851b 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4607,13 +4607,13 @@ xfs_bmapi_delay( STATIC int xfs_bmapi_allocate( struct xfs_bmalloca *bma, - int flags, - int *logflags) + int flags) { struct xfs_mount *mp = bma->ip->i_mount; int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? XFS_ATTR_FORK : XFS_DATA_FORK; struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); + int tmp_logflags = 0; int error; int rt; @@ -4700,14 +4700,15 @@ xfs_bmapi_allocate( if (bma->wasdel) { error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, &bma->idx, &bma->cur, &bma->got, - bma->firstblock, bma->flist, logflags); + bma->firstblock, bma->flist, &tmp_logflags); } else { error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, &bma->idx, &bma->cur, &bma->got, - bma->firstblock, bma->flist, logflags, + bma->firstblock, bma->flist, &tmp_logflags, whichfork); } + bma->logflags |= tmp_logflags; if (error) return error; @@ -4731,16 +4732,14 @@ xfs_bmapi_convert_unwritten( struct xfs_bmalloca *bma, struct xfs_bmbt_irec *mval, xfs_filblks_t len, - int flags, - int *logflags) + int flags) { int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? XFS_ATTR_FORK : XFS_DATA_FORK; struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); + int tmp_logflags = 0; int error; - *logflags = 0; - /* check if we need to do unwritten->real conversion */ if (mval->br_state == XFS_EXT_UNWRITTEN && (flags & XFS_BMAPI_PREALLOC)) @@ -4766,7 +4765,9 @@ xfs_bmapi_convert_unwritten( ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, &bma->idx, - &bma->cur, mval, bma->firstblock, bma->flist, logflags); + &bma->cur, mval, bma->firstblock, bma->flist, + &tmp_logflags); + bma->logflags |= tmp_logflags; if (error) return error; @@ -4818,10 +4819,8 @@ xfs_bmapi_write( xfs_fileoff_t end; /* end of mapped file region */ int eof; /* after the end of extents */ int error; /* error return */ - int logflags; /* flags for transaction logging */ int n; /* current extent index */ xfs_fileoff_t obno; /* old block number (offset) */ - int tmp_logflags; /* temp flags holder */ int whichfork; /* data or attr fork */ char inhole; /* current location is hole in file */ char wasdelay; /* old extent was delayed */ @@ -4866,11 +4865,9 @@ xfs_bmapi_write( XFS_STATS_INC(xs_blk_mapw); - logflags = 0; - if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { error = xfs_bmap_local_to_extents(tp, ip, firstblock, total, - &logflags, whichfork); + &bma.logflags, whichfork); if (error) goto error0; } @@ -4918,8 +4915,7 @@ xfs_bmapi_write( bma.length = len; bma.offset = bno; - error = xfs_bmapi_allocate(&bma, flags, &tmp_logflags); - logflags |= tmp_logflags; + error = xfs_bmapi_allocate(&bma, flags); if (error) goto error0; if (bma.blkno == NULLFSBLOCK) @@ -4931,9 +4927,7 @@ xfs_bmapi_write( end, n, flags); /* Execute unwritten extent conversion if necessary */ - error = xfs_bmapi_convert_unwritten(&bma, mval, len, - flags, &tmp_logflags); - logflags |= tmp_logflags; + error = xfs_bmapi_convert_unwritten(&bma, mval, len, flags); if (error == EAGAIN) continue; if (error) @@ -4965,10 +4959,12 @@ xfs_bmapi_write( */ if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE && XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max) { + int tmp_logflags = 0; + ASSERT(bma.cur); error = xfs_bmap_btree_to_extents(tp, ip, bma.cur, &tmp_logflags, whichfork); - logflags |= tmp_logflags; + bma.logflags |= tmp_logflags; if (error) goto error0; } @@ -4982,19 +4978,19 @@ xfs_bmapi_write( * Log everything. Do this after conversion, there's no point in * logging the extent records if we've converted to btree format. */ - if ((logflags & xfs_ilog_fext(whichfork)) && + if ((bma.logflags & xfs_ilog_fext(whichfork)) && XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS) - logflags &= ~xfs_ilog_fext(whichfork); - else if ((logflags & xfs_ilog_fbroot(whichfork)) && + bma.logflags &= ~xfs_ilog_fext(whichfork); + else if ((bma.logflags & xfs_ilog_fbroot(whichfork)) && XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) - logflags &= ~xfs_ilog_fbroot(whichfork); + bma.logflags &= ~xfs_ilog_fbroot(whichfork); /* * Log whatever the flags say, even if error. Otherwise we might miss * detecting a case where the data is changed, there's an error, * and it's not logged so we don't shutdown when we should. */ - if (logflags) - xfs_trans_log_inode(tp, ip, logflags); + if (bma.logflags) + xfs_trans_log_inode(tp, ip, bma.logflags); if (bma.cur) { if (!error) { diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index c2d9e1ef4ba3..243e212b31f6 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -123,6 +123,7 @@ typedef struct xfs_bmalloca { struct xfs_btree_cur *cur; /* btree cursor */ xfs_extnum_t idx; /* current extent index */ int nallocs;/* number of extents alloc'd */ + int logflags;/* flags for transaction logging */ xfs_extlen_t total; /* total blocks needed for xaction */ xfs_extlen_t minlen; /* minimum allocation size (blocks) */ From 572a4cf04ac7f46e9206aabfef03dae602812341 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:41:04 +0000 Subject: [PATCH 43/69] xfs: pass bmalloca to xfs_bmap_add_extent_delay_real All the parameters passed to xfs_bmap_add_extent_delay_real() are in the xfs_bmalloca structure now. Just pass the bmalloca parameter to the function instead of 8 separate parameters. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 332 ++++++++++++++++++++++++---------------------- 1 file changed, 171 insertions(+), 161 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index ed094749851b..c8b9c4ec9f6f 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -375,16 +375,9 @@ xfs_bmap_add_attrfork_local( */ STATIC int /* error */ xfs_bmap_add_extent_delay_real( - struct xfs_trans *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - xfs_fsblock_t *first, /* pointer to firstblock variable */ - xfs_bmap_free_t *flist, /* list of extents to be freed */ - int *logflagsp) /* inode logging flags */ + struct xfs_bmalloca *bma) { - xfs_btree_cur_t *cur; /* btree cursor */ + struct xfs_bmbt_irec *new = &bma->got; int diff; /* temp value */ xfs_bmbt_rec_host_t *ep; /* extent entry for idx */ int error; /* error return value */ @@ -401,18 +394,16 @@ xfs_bmap_add_extent_delay_real( xfs_filblks_t temp2=0;/* value for da_new calculations */ int tmp_rval; /* partial logging flags */ - ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); - cur = *curp; + ifp = XFS_IFORK_PTR(bma->ip, XFS_DATA_FORK); - ASSERT(*idx >= 0); - ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(bma->idx >= 0); + ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); ASSERT(!isnullstartblock(new->br_startblock)); - ASSERT(!cur || (cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); + ASSERT(!bma->cur || + (bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); XFS_STATS_INC(xs_add_exlist); - *logflagsp = 0; - #define LEFT r[0] #define RIGHT r[1] #define PREV r[2] @@ -420,7 +411,7 @@ xfs_bmap_add_extent_delay_real( /* * Set up a bunch of variables to make the tests simpler. */ - ep = xfs_iext_get_ext(ifp, *idx); + ep = xfs_iext_get_ext(ifp, bma->idx); xfs_bmbt_get_all(ep, &PREV); new_endoff = new->br_startoff + new->br_blockcount; ASSERT(PREV.br_startoff <= new->br_startoff); @@ -442,9 +433,9 @@ xfs_bmap_add_extent_delay_real( * Check and set flags if this segment has a left neighbor. * Don't set contiguous if the combined extent would be too large. */ - if (*idx > 0) { + if (bma->idx > 0) { state |= BMAP_LEFT_VALID; - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx - 1), &LEFT); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &LEFT); if (isnullstartblock(LEFT.br_startblock)) state |= BMAP_LEFT_DELAY; @@ -462,9 +453,9 @@ xfs_bmap_add_extent_delay_real( * Don't set contiguous if the combined extent would be too large. * Also check for all-three-contiguous being too large. */ - if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) { + if (bma->idx < bma->ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) { state |= BMAP_RIGHT_VALID; - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx + 1), &RIGHT); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx + 1), &RIGHT); if (isnullstartblock(RIGHT.br_startblock)) state |= BMAP_RIGHT_DELAY; @@ -495,35 +486,39 @@ xfs_bmap_add_extent_delay_real( * Filling in all of a previously delayed allocation extent. * The left and right neighbors are both contiguous with new. */ - --*idx; - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); - xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), + bma->idx--; + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); + xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx), LEFT.br_blockcount + PREV.br_blockcount + RIGHT.br_blockcount); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - xfs_iext_remove(ip, *idx + 1, 2, state); - ip->i_d.di_nextents--; - if (cur == NULL) + xfs_iext_remove(bma->ip, bma->idx + 1, 2, state); + bma->ip->i_d.di_nextents--; + if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff, RIGHT.br_startblock, - RIGHT.br_blockcount, &i))) + RIGHT.br_blockcount, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_btree_delete(cur, &i))) + error = xfs_btree_delete(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_btree_decrement(cur, 0, &i))) + error = xfs_btree_decrement(bma->cur, 0, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, LEFT.br_startoff, + error = xfs_bmbt_update(bma->cur, LEFT.br_startoff, LEFT.br_startblock, LEFT.br_blockcount + PREV.br_blockcount + - RIGHT.br_blockcount, LEFT.br_state))) + RIGHT.br_blockcount, LEFT.br_state); + if (error) goto done; } break; @@ -533,27 +528,29 @@ xfs_bmap_add_extent_delay_real( * Filling in all of a previously delayed allocation extent. * The left neighbor is contiguous, the right is not. */ - --*idx; + bma->idx--; - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); - xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); + xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx), LEFT.br_blockcount + PREV.br_blockcount); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - xfs_iext_remove(ip, *idx + 1, 1, state); - if (cur == NULL) + xfs_iext_remove(bma->ip, bma->idx + 1, 1, state); + if (bma->cur == NULL) rval = XFS_ILOG_DEXT; else { rval = 0; - if ((error = xfs_bmbt_lookup_eq(cur, LEFT.br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, LEFT.br_startoff, LEFT.br_startblock, LEFT.br_blockcount, - &i))) + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, LEFT.br_startoff, + error = xfs_bmbt_update(bma->cur, LEFT.br_startoff, LEFT.br_startblock, LEFT.br_blockcount + - PREV.br_blockcount, LEFT.br_state))) + PREV.br_blockcount, LEFT.br_state); + if (error) goto done; } break; @@ -563,26 +560,28 @@ xfs_bmap_add_extent_delay_real( * Filling in all of a previously delayed allocation extent. * The right neighbor is contiguous, the left is not. */ - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_startblock(ep, new->br_startblock); xfs_bmbt_set_blockcount(ep, PREV.br_blockcount + RIGHT.br_blockcount); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - xfs_iext_remove(ip, *idx + 1, 1, state); - if (cur == NULL) + xfs_iext_remove(bma->ip, bma->idx + 1, 1, state); + if (bma->cur == NULL) rval = XFS_ILOG_DEXT; else { rval = 0; - if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff, RIGHT.br_startblock, - RIGHT.br_blockcount, &i))) + RIGHT.br_blockcount, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, PREV.br_startoff, + error = xfs_bmbt_update(bma->cur, PREV.br_startoff, new->br_startblock, PREV.br_blockcount + - RIGHT.br_blockcount, PREV.br_state))) + RIGHT.br_blockcount, PREV.br_state); + if (error) goto done; } break; @@ -593,22 +592,24 @@ xfs_bmap_add_extent_delay_real( * Neither the left nor right neighbors are contiguous with * the new one. */ - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_startblock(ep, new->br_startblock); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - ip->i_d.di_nextents++; - if (cur == NULL) + bma->ip->i_d.di_nextents++; + if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff, new->br_startblock, new->br_blockcount, - &i))) + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 0, done); - cur->bc_rec.b.br_state = XFS_EXT_NORM; - if ((error = xfs_btree_insert(cur, &i))) + bma->cur->bc_rec.b.br_state = XFS_EXT_NORM; + error = xfs_btree_insert(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); } @@ -619,38 +620,40 @@ xfs_bmap_add_extent_delay_real( * Filling in the first part of a previous delayed allocation. * The left neighbor is contiguous. */ - trace_xfs_bmap_pre_update(ip, *idx - 1, state, _THIS_IP_); - xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx - 1), + trace_xfs_bmap_pre_update(bma->ip, bma->idx - 1, state, _THIS_IP_); + xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx - 1), LEFT.br_blockcount + new->br_blockcount); xfs_bmbt_set_startoff(ep, PREV.br_startoff + new->br_blockcount); - trace_xfs_bmap_post_update(ip, *idx - 1, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx - 1, state, _THIS_IP_); temp = PREV.br_blockcount - new->br_blockcount; - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_blockcount(ep, temp); - if (cur == NULL) + if (bma->cur == NULL) rval = XFS_ILOG_DEXT; else { rval = 0; - if ((error = xfs_bmbt_lookup_eq(cur, LEFT.br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, LEFT.br_startoff, LEFT.br_startblock, LEFT.br_blockcount, - &i))) + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, LEFT.br_startoff, + error = xfs_bmbt_update(bma->cur, LEFT.br_startoff, LEFT.br_startblock, LEFT.br_blockcount + new->br_blockcount, - LEFT.br_state))) + LEFT.br_state); + if (error) goto done; } - da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp), startblockval(PREV.br_startblock)); xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - --*idx; + bma->idx--; break; case BMAP_LEFT_FILLING: @@ -658,41 +661,43 @@ xfs_bmap_add_extent_delay_real( * Filling in the first part of a previous delayed allocation. * The left neighbor is not contiguous. */ - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_startoff(ep, new_endoff); temp = PREV.br_blockcount - new->br_blockcount; xfs_bmbt_set_blockcount(ep, temp); - xfs_iext_insert(ip, *idx, 1, new, state); - ip->i_d.di_nextents++; - if (cur == NULL) + xfs_iext_insert(bma->ip, bma->idx, 1, new, state); + bma->ip->i_d.di_nextents++; + if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff, new->br_startblock, new->br_blockcount, - &i))) + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 0, done); - cur->bc_rec.b.br_state = XFS_EXT_NORM; - if ((error = xfs_btree_insert(cur, &i))) + bma->cur->bc_rec.b.br_state = XFS_EXT_NORM; + error = xfs_btree_insert(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); } - if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS && - ip->i_d.di_nextents > ip->i_df.if_ext_max) { - error = xfs_bmap_extents_to_btree(tp, ip, - first, flist, &cur, 1, &tmp_rval, - XFS_DATA_FORK); + if (bma->ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS && + bma->ip->i_d.di_nextents > bma->ip->i_df.if_ext_max) { + error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, + bma->firstblock, bma->flist, + &bma->cur, 1, &tmp_rval, XFS_DATA_FORK); rval |= tmp_rval; if (error) goto done; } - da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp), startblockval(PREV.br_startblock) - - (cur ? cur->bc_private.b.allocated : 0)); - ep = xfs_iext_get_ext(ifp, *idx + 1); + (bma->cur ? bma->cur->bc_private.b.allocated : 0)); + ep = xfs_iext_get_ext(ifp, bma->idx + 1); xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); - trace_xfs_bmap_post_update(ip, *idx + 1, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx + 1, state, _THIS_IP_); break; case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG: @@ -701,37 +706,39 @@ xfs_bmap_add_extent_delay_real( * The right neighbor is contiguous with the new allocation. */ temp = PREV.br_blockcount - new->br_blockcount; - trace_xfs_bmap_pre_update(ip, *idx + 1, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx + 1, state, _THIS_IP_); xfs_bmbt_set_blockcount(ep, temp); - xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, *idx + 1), + xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, bma->idx + 1), new->br_startoff, new->br_startblock, new->br_blockcount + RIGHT.br_blockcount, RIGHT.br_state); - trace_xfs_bmap_post_update(ip, *idx + 1, state, _THIS_IP_); - if (cur == NULL) + trace_xfs_bmap_post_update(bma->ip, bma->idx + 1, state, _THIS_IP_); + if (bma->cur == NULL) rval = XFS_ILOG_DEXT; else { rval = 0; - if ((error = xfs_bmbt_lookup_eq(cur, RIGHT.br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, RIGHT.br_startoff, RIGHT.br_startblock, - RIGHT.br_blockcount, &i))) + RIGHT.br_blockcount, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, new->br_startoff, + error = xfs_bmbt_update(bma->cur, new->br_startoff, new->br_startblock, new->br_blockcount + RIGHT.br_blockcount, - RIGHT.br_state))) + RIGHT.br_state); + if (error) goto done; } - da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp), startblockval(PREV.br_startblock)); - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - ++*idx; + bma->idx++; break; case BMAP_RIGHT_FILLING: @@ -740,41 +747,43 @@ xfs_bmap_add_extent_delay_real( * The right neighbor is not contiguous. */ temp = PREV.br_blockcount - new->br_blockcount; - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_blockcount(ep, temp); - xfs_iext_insert(ip, *idx + 1, 1, new, state); - ip->i_d.di_nextents++; - if (cur == NULL) + xfs_iext_insert(bma->ip, bma->idx + 1, 1, new, state); + bma->ip->i_d.di_nextents++; + if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff, new->br_startblock, new->br_blockcount, - &i))) + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 0, done); - cur->bc_rec.b.br_state = XFS_EXT_NORM; - if ((error = xfs_btree_insert(cur, &i))) + bma->cur->bc_rec.b.br_state = XFS_EXT_NORM; + error = xfs_btree_insert(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); } - if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS && - ip->i_d.di_nextents > ip->i_df.if_ext_max) { - error = xfs_bmap_extents_to_btree(tp, ip, - first, flist, &cur, 1, &tmp_rval, - XFS_DATA_FORK); + if (bma->ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS && + bma->ip->i_d.di_nextents > bma->ip->i_df.if_ext_max) { + error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, + bma->firstblock, bma->flist, &bma->cur, 1, + &tmp_rval, XFS_DATA_FORK); rval |= tmp_rval; if (error) goto done; } - da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp), + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp), startblockval(PREV.br_startblock) - - (cur ? cur->bc_private.b.allocated : 0)); - ep = xfs_iext_get_ext(ifp, *idx); + (bma->cur ? bma->cur->bc_private.b.allocated : 0)); + ep = xfs_iext_get_ext(ifp, bma->idx); xfs_bmbt_set_startblock(ep, nullstartblock(da_new)); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - ++*idx; + bma->idx++; break; case 0: @@ -800,46 +809,48 @@ xfs_bmap_add_extent_delay_real( */ temp = new->br_startoff - PREV.br_startoff; temp2 = PREV.br_startoff + PREV.br_blockcount - new_endoff; - trace_xfs_bmap_pre_update(ip, *idx, 0, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx, 0, _THIS_IP_); xfs_bmbt_set_blockcount(ep, temp); /* truncate PREV */ LEFT = *new; RIGHT.br_state = PREV.br_state; RIGHT.br_startblock = nullstartblock( - (int)xfs_bmap_worst_indlen(ip, temp2)); + (int)xfs_bmap_worst_indlen(bma->ip, temp2)); RIGHT.br_startoff = new_endoff; RIGHT.br_blockcount = temp2; /* insert LEFT (r[0]) and RIGHT (r[1]) at the same time */ - xfs_iext_insert(ip, *idx + 1, 2, &LEFT, state); - ip->i_d.di_nextents++; - if (cur == NULL) + xfs_iext_insert(bma->ip, bma->idx + 1, 2, &LEFT, state); + bma->ip->i_d.di_nextents++; + if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff, + error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff, new->br_startblock, new->br_blockcount, - &i))) + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 0, done); - cur->bc_rec.b.br_state = XFS_EXT_NORM; - if ((error = xfs_btree_insert(cur, &i))) + bma->cur->bc_rec.b.br_state = XFS_EXT_NORM; + error = xfs_btree_insert(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); } - if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS && - ip->i_d.di_nextents > ip->i_df.if_ext_max) { - error = xfs_bmap_extents_to_btree(tp, ip, - first, flist, &cur, 1, &tmp_rval, - XFS_DATA_FORK); + if (bma->ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS && + bma->ip->i_d.di_nextents > bma->ip->i_df.if_ext_max) { + error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, + bma->firstblock, bma->flist, &bma->cur, + 1, &tmp_rval, XFS_DATA_FORK); rval |= tmp_rval; if (error) goto done; } - temp = xfs_bmap_worst_indlen(ip, temp); - temp2 = xfs_bmap_worst_indlen(ip, temp2); + temp = xfs_bmap_worst_indlen(bma->ip, temp); + temp2 = xfs_bmap_worst_indlen(bma->ip, temp2); diff = (int)(temp + temp2 - startblockval(PREV.br_startblock) - - (cur ? cur->bc_private.b.allocated : 0)); + (bma->cur ? bma->cur->bc_private.b.allocated : 0)); if (diff > 0) { - error = xfs_icsb_modify_counters(ip->i_mount, + error = xfs_icsb_modify_counters(bma->ip->i_mount, XFS_SBS_FDBLOCKS, -((int64_t)diff), 0); ASSERT(!error); @@ -847,15 +858,15 @@ xfs_bmap_add_extent_delay_real( goto done; } - ep = xfs_iext_get_ext(ifp, *idx); + ep = xfs_iext_get_ext(ifp, bma->idx); xfs_bmbt_set_startblock(ep, nullstartblock((int)temp)); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); - trace_xfs_bmap_pre_update(ip, *idx + 2, state, _THIS_IP_); - xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, *idx + 2), + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); + trace_xfs_bmap_pre_update(bma->ip, bma->idx + 2, state, _THIS_IP_); + xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, bma->idx + 2), nullstartblock((int)temp2)); - trace_xfs_bmap_post_update(ip, *idx + 2, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx + 2, state, _THIS_IP_); - ++*idx; + bma->idx++; da_new = temp + temp2; break; @@ -873,14 +884,15 @@ xfs_bmap_add_extent_delay_real( } /* convert to a btree if necessary */ - if (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) == XFS_DINODE_FMT_EXTENTS && - XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) > ifp->if_ext_max) { + if (XFS_IFORK_FORMAT(bma->ip, XFS_DATA_FORK) == XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_NEXTENTS(bma->ip, XFS_DATA_FORK) > ifp->if_ext_max) { int tmp_logflags; /* partial log flag return val */ - ASSERT(cur == NULL); - error = xfs_bmap_extents_to_btree(tp, ip, first, flist, &cur, + ASSERT(bma->cur == NULL); + error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, + bma->firstblock, bma->flist, &bma->cur, da_old > 0, &tmp_logflags, XFS_DATA_FORK); - *logflagsp |= tmp_logflags; + bma->logflags |= tmp_logflags; if (error) goto done; } @@ -888,22 +900,22 @@ xfs_bmap_add_extent_delay_real( /* adjust for changes in reserved delayed indirect blocks */ if (da_old || da_new) { temp = da_new; - if (cur) - temp += cur->bc_private.b.allocated; + if (bma->cur) + temp += bma->cur->bc_private.b.allocated; ASSERT(temp <= da_old); if (temp < da_old) - xfs_icsb_modify_counters(ip->i_mount, XFS_SBS_FDBLOCKS, - (int64_t)(da_old - temp), 0); + xfs_icsb_modify_counters(bma->ip->i_mount, + XFS_SBS_FDBLOCKS, + (int64_t)(da_old - temp), 0); } /* clear out the allocated field, done with it now in any case. */ - if (cur) { - cur->bc_private.b.allocated = 0; - *curp = cur; - } - xfs_bmap_check_leaf_extents(cur, ip, XFS_DATA_FORK); + if (bma->cur) + bma->cur->bc_private.b.allocated = 0; + + xfs_bmap_check_leaf_extents(bma->cur, bma->ip, XFS_DATA_FORK); done: - *logflagsp |= rval; + bma->logflags |= rval; return error; #undef LEFT #undef RIGHT @@ -4698,9 +4710,7 @@ xfs_bmapi_allocate( bma->got.br_state = XFS_EXT_UNWRITTEN; if (bma->wasdel) { - error = xfs_bmap_add_extent_delay_real(bma->tp, bma->ip, - &bma->idx, &bma->cur, &bma->got, - bma->firstblock, bma->flist, &tmp_logflags); + error = xfs_bmap_add_extent_delay_real(bma); } else { error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, &bma->idx, &bma->cur, &bma->got, From c6534249851d062113ab4d8d226be8dba8ecb92e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:41:05 +0000 Subject: [PATCH 44/69] xfs: pass bmalloca to xfs_bmap_add_extent_hole_real All the parameters passed to xfs_bmap_add_extent_hole_real() are in the xfs_bmalloca structure now. Just pass the bmalloca parameter to the function instead of 8 separate parameters. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 161 ++++++++++++++++++++++------------------------ 1 file changed, 78 insertions(+), 83 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index c8b9c4ec9f6f..2b31945ce9fe 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -1596,34 +1596,25 @@ xfs_bmap_add_extent_hole_delay( */ STATIC int /* error */ xfs_bmap_add_extent_hole_real( - struct xfs_trans *tp, - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t *idx, /* extent number to update/insert */ - xfs_btree_cur_t **curp, /* if null, not a btree */ - xfs_bmbt_irec_t *new, /* new data to add to file extents */ - xfs_fsblock_t *first, /* pointer to firstblock variable */ - xfs_bmap_free_t *flist, /* list of extents to be freed */ - int *logflagsp, /* inode logging flags */ - int whichfork) /* data or attr fork */ + struct xfs_bmalloca *bma, + int whichfork) { + struct xfs_bmbt_irec *new = &bma->got; int error; /* error return value */ int i; /* temp state */ - xfs_btree_cur_t *cur; /* if null, not a btree */ xfs_ifork_t *ifp; /* inode fork pointer */ xfs_bmbt_irec_t left; /* left neighbor extent entry */ xfs_bmbt_irec_t right; /* right neighbor extent entry */ int rval=0; /* return value (logging flags) */ int state; /* state bits, accessed thru macros */ - *logflagsp = 0; + ifp = XFS_IFORK_PTR(bma->ip, whichfork); - ifp = XFS_IFORK_PTR(ip, whichfork); - cur = *curp; - - ASSERT(*idx >= 0); - ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(bma->idx >= 0); + ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); ASSERT(!isnullstartblock(new->br_startblock)); - ASSERT(!cur || !(cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); + ASSERT(!bma->cur || + !(bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); XFS_STATS_INC(xs_add_exlist); @@ -1634,9 +1625,9 @@ xfs_bmap_add_extent_hole_real( /* * Check and set flags if this segment has a left neighbor. */ - if (*idx > 0) { + if (bma->idx > 0) { state |= BMAP_LEFT_VALID; - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx - 1), &left); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &left); if (isnullstartblock(left.br_startblock)) state |= BMAP_LEFT_DELAY; } @@ -1645,9 +1636,9 @@ xfs_bmap_add_extent_hole_real( * Check and set flags if this segment has a current value. * Not true if we're inserting into the "hole" at eof. */ - if (*idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) { + if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) { state |= BMAP_RIGHT_VALID; - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx), &right); + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &right); if (isnullstartblock(right.br_startblock)) state |= BMAP_RIGHT_DELAY; } @@ -1684,39 +1675,42 @@ xfs_bmap_add_extent_hole_real( * left and on the right. * Merge all three into a single extent record. */ - --*idx; - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); - xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), + --bma->idx; + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); + xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx), left.br_blockcount + new->br_blockcount + right.br_blockcount); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - xfs_iext_remove(ip, *idx + 1, 1, state); + xfs_iext_remove(bma->ip, bma->idx + 1, 1, state); - XFS_IFORK_NEXT_SET(ip, whichfork, - XFS_IFORK_NEXTENTS(ip, whichfork) - 1); - if (cur == NULL) { + XFS_IFORK_NEXT_SET(bma->ip, whichfork, + XFS_IFORK_NEXTENTS(bma->ip, whichfork) - 1); + if (bma->cur == NULL) { rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork); } else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, - right.br_startoff, - right.br_startblock, - right.br_blockcount, &i))) + error = xfs_bmbt_lookup_eq(bma->cur, right.br_startoff, + right.br_startblock, right.br_blockcount, + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_btree_delete(cur, &i))) + error = xfs_btree_delete(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_btree_decrement(cur, 0, &i))) + error = xfs_btree_decrement(bma->cur, 0, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, left.br_startoff, + error = xfs_bmbt_update(bma->cur, left.br_startoff, left.br_startblock, left.br_blockcount + new->br_blockcount + right.br_blockcount, - left.br_state))) + left.br_state); + if (error) goto done; } break; @@ -1727,27 +1721,28 @@ xfs_bmap_add_extent_hole_real( * on the left. * Merge the new allocation with the left neighbor. */ - --*idx; - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); - xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, *idx), + --bma->idx; + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); + xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, bma->idx), left.br_blockcount + new->br_blockcount); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - if (cur == NULL) { + if (bma->cur == NULL) { rval = xfs_ilog_fext(whichfork); } else { rval = 0; - if ((error = xfs_bmbt_lookup_eq(cur, - left.br_startoff, - left.br_startblock, - left.br_blockcount, &i))) + error = xfs_bmbt_lookup_eq(bma->cur, left.br_startoff, + left.br_startblock, left.br_blockcount, + &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, left.br_startoff, + error = xfs_bmbt_update(bma->cur, left.br_startoff, left.br_startblock, left.br_blockcount + new->br_blockcount, - left.br_state))) + left.br_state); + if (error) goto done; } break; @@ -1758,28 +1753,30 @@ xfs_bmap_add_extent_hole_real( * on the right. * Merge the new allocation with the right neighbor. */ - trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); - xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, *idx), + trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); + xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, bma->idx), new->br_startoff, new->br_startblock, new->br_blockcount + right.br_blockcount, right.br_state); - trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - if (cur == NULL) { + if (bma->cur == NULL) { rval = xfs_ilog_fext(whichfork); } else { rval = 0; - if ((error = xfs_bmbt_lookup_eq(cur, + error = xfs_bmbt_lookup_eq(bma->cur, right.br_startoff, right.br_startblock, - right.br_blockcount, &i))) + right.br_blockcount, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); - if ((error = xfs_bmbt_update(cur, new->br_startoff, + error = xfs_bmbt_update(bma->cur, new->br_startoff, new->br_startblock, new->br_blockcount + right.br_blockcount, - right.br_state))) + right.br_state); + if (error) goto done; } break; @@ -1790,21 +1787,23 @@ xfs_bmap_add_extent_hole_real( * real allocation. * Insert a new entry. */ - xfs_iext_insert(ip, *idx, 1, new, state); - XFS_IFORK_NEXT_SET(ip, whichfork, - XFS_IFORK_NEXTENTS(ip, whichfork) + 1); - if (cur == NULL) { + xfs_iext_insert(bma->ip, bma->idx, 1, new, state); + XFS_IFORK_NEXT_SET(bma->ip, whichfork, + XFS_IFORK_NEXTENTS(bma->ip, whichfork) + 1); + if (bma->cur == NULL) { rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork); } else { rval = XFS_ILOG_CORE; - if ((error = xfs_bmbt_lookup_eq(cur, + error = xfs_bmbt_lookup_eq(bma->cur, new->br_startoff, new->br_startblock, - new->br_blockcount, &i))) + new->br_blockcount, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 0, done); - cur->bc_rec.b.br_state = new->br_state; - if ((error = xfs_btree_insert(cur, &i))) + bma->cur->bc_rec.b.br_state = new->br_state; + error = xfs_btree_insert(bma->cur, &i); + if (error) goto done; XFS_WANT_CORRUPTED_GOTO(i == 1, done); } @@ -1812,26 +1811,26 @@ xfs_bmap_add_extent_hole_real( } /* convert to a btree if necessary */ - if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS && - XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max) { + if (XFS_IFORK_FORMAT(bma->ip, whichfork) == XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_NEXTENTS(bma->ip, whichfork) > ifp->if_ext_max) { int tmp_logflags; /* partial log flag return val */ - ASSERT(cur == NULL); - error = xfs_bmap_extents_to_btree(tp, ip, first, - flist, &cur, 0, &tmp_logflags, whichfork); - *logflagsp |= tmp_logflags; + ASSERT(bma->cur == NULL); + error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, + bma->firstblock, bma->flist, &bma->cur, + 0, &tmp_logflags, whichfork); + bma->logflags |= tmp_logflags; if (error) goto done; } /* clear out the allocated field, done with it now in any case. */ - if (cur) { - cur->bc_private.b.allocated = 0; - *curp = cur; - } - xfs_bmap_check_leaf_extents(cur, ip, whichfork); + if (bma->cur) + bma->cur->bc_private.b.allocated = 0; + + xfs_bmap_check_leaf_extents(bma->cur, bma->ip, whichfork); done: - *logflagsp |= rval; + bma->logflags |= rval; return error; } @@ -4709,14 +4708,10 @@ xfs_bmapi_allocate( xfs_sb_version_hasextflgbit(&mp->m_sb)) bma->got.br_state = XFS_EXT_UNWRITTEN; - if (bma->wasdel) { + if (bma->wasdel) error = xfs_bmap_add_extent_delay_real(bma); - } else { - error = xfs_bmap_add_extent_hole_real(bma->tp, bma->ip, - &bma->idx, &bma->cur, &bma->got, - bma->firstblock, bma->flist, &tmp_logflags, - whichfork); - } + else + error = xfs_bmap_add_extent_hole_real(bma, whichfork); bma->logflags |= tmp_logflags; if (error) From b0eab14e74d2d7b22d065e18a1cdebcf7716debf Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:41:06 +0000 Subject: [PATCH 45/69] xfs: dont ignore error code from xfs_bmbt_update Fix a case in xfs_bmap_add_extent_unwritten_real where we aren't passing the returned error on. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 2b31945ce9fe..bb31d37bf49b 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -1217,10 +1217,11 @@ xfs_bmap_add_extent_unwritten_real( goto done; if ((error = xfs_btree_decrement(cur, 0, &i))) goto done; - if (xfs_bmbt_update(cur, LEFT.br_startoff, + error = xfs_bmbt_update(cur, LEFT.br_startoff, LEFT.br_startblock, LEFT.br_blockcount + new->br_blockcount, - LEFT.br_state)) + LEFT.br_state); + if (error) goto done; } break; From d952e2f81244d6502aff126df5011fab10f92187 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:41:07 +0000 Subject: [PATCH 46/69] xfs: cleanup xfs_bmap.h Convert all function prototypes to the short form used elsewhere, and remove duplicates of comments already placed at the function body. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_bmap.h | 225 +++++++--------------------------------------- 1 file changed, 33 insertions(+), 192 deletions(-) diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 243e212b31f6..89ee672d378a 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -155,121 +155,29 @@ typedef struct xfs_bmalloca { { BMAP_RIGHT_FILLING, "RF" }, \ { BMAP_ATTRFORK, "ATTR" } -/* - * Add bmap trace insert entries for all the contents of the extent list. - * - * Quite excessive tracing. Only do this for debug builds. - */ #if defined(__KERNEL) && defined(DEBUG) -void -xfs_bmap_trace_exlist( - struct xfs_inode *ip, /* incore inode pointer */ - xfs_extnum_t cnt, /* count of entries in list */ - int whichfork, - unsigned long caller_ip); /* data or attr fork */ +void xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt, + int whichfork, unsigned long caller_ip); #define XFS_BMAP_TRACE_EXLIST(ip,c,w) \ xfs_bmap_trace_exlist(ip,c,w, _THIS_IP_) #else #define XFS_BMAP_TRACE_EXLIST(ip,c,w) #endif -/* - * Convert inode from non-attributed to attributed. - * Must not be in a transaction, ip must not be locked. - */ -int /* error code */ -xfs_bmap_add_attrfork( - struct xfs_inode *ip, /* incore inode pointer */ - int size, /* space needed for new attribute */ - int rsvd); /* flag for reserved block allocation */ - -/* - * Add the extent to the list of extents to be free at transaction end. - * The list is maintained sorted (by block number). - */ -void -xfs_bmap_add_free( - xfs_fsblock_t bno, /* fs block number of extent */ - xfs_filblks_t len, /* length of extent */ - xfs_bmap_free_t *flist, /* list of extents */ - struct xfs_mount *mp); /* mount point structure */ - -/* - * Routine to clean up the free list data structure when - * an error occurs during a transaction. - */ -void -xfs_bmap_cancel( - xfs_bmap_free_t *flist); /* free list to clean up */ - -/* - * Compute and fill in the value of the maximum depth of a bmap btree - * in this filesystem. Done once, during mount. - */ -void -xfs_bmap_compute_maxlevels( - struct xfs_mount *mp, /* file system mount structure */ - int whichfork); /* data or attr fork */ - -/* - * Returns the file-relative block number of the first unused block in the file. - * This is the lowest-address hole if the file has holes, else the first block - * past the end of file. - */ -int /* error */ -xfs_bmap_first_unused( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - xfs_extlen_t len, /* size of hole to find */ - xfs_fileoff_t *unused, /* unused block num */ - int whichfork); /* data or attr fork */ - -/* - * Returns the file-relative block number of the last block + 1 before - * last_block (input value) in the file. - * This is not based on i_size, it is based on the extent list. - * Returns 0 for local files, as they do not have an extent list. - */ -int /* error */ -xfs_bmap_last_before( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - xfs_fileoff_t *last_block, /* last block */ - int whichfork); /* data or attr fork */ - -/* - * Returns the file-relative block number of the first block past eof in - * the file. This is not based on i_size, it is based on the extent list. - * Returns 0 for local files, as they do not have an extent list. - */ -int /* error */ -xfs_bmap_last_offset( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - xfs_fileoff_t *unused, /* last block num */ - int whichfork); /* data or attr fork */ - -/* - * Returns whether the selected fork of the inode has exactly one - * block or not. For the data fork we check this matches di_size, - * implying the file's range is 0..bsize-1. - */ -int -xfs_bmap_one_block( - struct xfs_inode *ip, /* incore inode */ - int whichfork); /* data or attr fork */ - -/* - * Read in the extents to iu_extents. - * All inode fields are set up by caller, we just traverse the btree - * and copy the records in. - */ -int /* error */ -xfs_bmap_read_extents( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - int whichfork); /* data or attr fork */ - +int xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd); +void xfs_bmap_add_free(xfs_fsblock_t bno, xfs_filblks_t len, + struct xfs_bmap_free *flist, struct xfs_mount *mp); +void xfs_bmap_cancel(struct xfs_bmap_free *flist); +void xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork); +int xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork); +int xfs_bmap_last_before(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_fileoff_t *last_block, int whichfork); +int xfs_bmap_last_offset(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_fileoff_t *unused, int whichfork); +int xfs_bmap_one_block(struct xfs_inode *ip, int whichfork); +int xfs_bmap_read_extents(struct xfs_trans *tp, struct xfs_inode *ip, + int whichfork); int xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, struct xfs_bmbt_irec *mval, int *nmap, int flags); @@ -281,95 +189,28 @@ int xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fsblock_t *firstblock, xfs_extlen_t total, struct xfs_bmbt_irec *mval, int *nmap, struct xfs_bmap_free *flist); - -/* - * Unmap (remove) blocks from a file. - * If nexts is nonzero then the number of extents to remove is limited to - * that value. If not all extents in the block range can be removed then - * *done is set. - */ -int /* error */ -xfs_bunmapi( - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_inode *ip, /* incore inode */ - xfs_fileoff_t bno, /* starting offset to unmap */ - xfs_filblks_t len, /* length to unmap in file */ - int flags, /* XFS_BMAPI_... */ - xfs_extnum_t nexts, /* number of extents max */ - xfs_fsblock_t *firstblock, /* first allocated block - controls a.g. for allocs */ - xfs_bmap_free_t *flist, /* i/o: list extents to free */ - int *done); /* set if not done yet */ - -/* - * Check an extent list, which has just been read, for - * any bit in the extent flag field. - */ -int -xfs_check_nostate_extents( - struct xfs_ifork *ifp, - xfs_extnum_t idx, - xfs_extnum_t num); - -uint -xfs_default_attroffset( - struct xfs_inode *ip); +int xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_fileoff_t bno, xfs_filblks_t len, int flags, + xfs_extnum_t nexts, xfs_fsblock_t *firstblock, + struct xfs_bmap_free *flist, int *done); +int xfs_check_nostate_extents(struct xfs_ifork *ifp, xfs_extnum_t idx, + xfs_extnum_t num); +uint xfs_default_attroffset(struct xfs_inode *ip); #ifdef __KERNEL__ - -/* - * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi - * caller. Frees all the extents that need freeing, which must be done - * last due to locking considerations. - * - * Return 1 if the given transaction was committed and a new one allocated, - * and 0 otherwise. - */ -int /* error */ -xfs_bmap_finish( - struct xfs_trans **tp, /* transaction pointer addr */ - xfs_bmap_free_t *flist, /* i/o: list extents to free */ - int *committed); /* xact committed or not */ - /* bmap to userspace formatter - copy to user & advance pointer */ typedef int (*xfs_bmap_format_t)(void **, struct getbmapx *, int *); -/* - * Get inode's extents as described in bmv, and format for output. - */ -int /* error code */ -xfs_getbmap( - xfs_inode_t *ip, - struct getbmapx *bmv, /* user bmap structure */ - xfs_bmap_format_t formatter, /* format to user */ - void *arg); /* formatter arg */ - -/* - * Check if the endoff is outside the last extent. If so the caller will grow - * the allocation to a stripe unit boundary - */ -int -xfs_bmap_eof( - struct xfs_inode *ip, - xfs_fileoff_t endoff, - int whichfork, - int *eof); - -/* - * Count fsblocks of the given fork. - */ -int -xfs_bmap_count_blocks( - xfs_trans_t *tp, - struct xfs_inode *ip, - int whichfork, - int *count); - -int -xfs_bmap_punch_delalloc_range( - struct xfs_inode *ip, - xfs_fileoff_t start_fsb, - xfs_fileoff_t length); +int xfs_bmap_finish(struct xfs_trans **tp, struct xfs_bmap_free *flist, + int *committed); +int xfs_getbmap(struct xfs_inode *ip, struct getbmapx *bmv, + xfs_bmap_format_t formatter, void *arg); +int xfs_bmap_eof(struct xfs_inode *ip, xfs_fileoff_t endoff, + int whichfork, int *eof); +int xfs_bmap_count_blocks(struct xfs_trans *tp, struct xfs_inode *ip, + int whichfork, int *count); +int xfs_bmap_punch_delalloc_range(struct xfs_inode *ip, + xfs_fileoff_t start_fsb, xfs_fileoff_t length); #endif /* __KERNEL__ */ #endif /* __XFS_BMAP_H__ */ From c029a50d51b8a9520105ec903639de03389915d0 Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Wed, 21 Sep 2011 09:42:30 +0000 Subject: [PATCH 47/69] xfs: fix possible overflow in xfs_ioc_trim() In xfs_ioc_trim it is possible that computing the last allocation group to discard might overflow for big start & len values, because the result might be bigger then xfs_agnumber_t which is 32 bit long. Fix this by not allowing the start and end block of the range to be beyond the end of the file system. Note that if the start is beyond the end of the file system we have to return -EINVAL, but in the "end" case we have to truncate it to the fs size. Also introduce "end" variable, rather than using start+len which which might be more confusing to get right as this bug shows. Signed-off-by: Lukas Czerner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_discard.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c index 244e797dae32..8a24f0c6c860 100644 --- a/fs/xfs/xfs_discard.c +++ b/fs/xfs/xfs_discard.c @@ -38,7 +38,7 @@ xfs_trim_extents( struct xfs_mount *mp, xfs_agnumber_t agno, xfs_fsblock_t start, - xfs_fsblock_t len, + xfs_fsblock_t end, xfs_fsblock_t minlen, __uint64_t *blocks_trimmed) { @@ -100,7 +100,7 @@ xfs_trim_extents( * down partially overlapping ranges for now. */ if (XFS_AGB_TO_FSB(mp, agno, fbno) + flen < start || - XFS_AGB_TO_FSB(mp, agno, fbno) >= start + len) { + XFS_AGB_TO_FSB(mp, agno, fbno) > end) { trace_xfs_discard_exclude(mp, agno, fbno, flen); goto next_extent; } @@ -145,7 +145,7 @@ xfs_ioc_trim( struct request_queue *q = mp->m_ddev_targp->bt_bdev->bd_disk->queue; unsigned int granularity = q->limits.discard_granularity; struct fstrim_range range; - xfs_fsblock_t start, len, minlen; + xfs_fsblock_t start, end, minlen; xfs_agnumber_t start_agno, end_agno, agno; __uint64_t blocks_trimmed = 0; int error, last_error = 0; @@ -165,19 +165,19 @@ xfs_ioc_trim( * matter as trimming blocks is an advisory interface. */ start = XFS_B_TO_FSBT(mp, range.start); - len = XFS_B_TO_FSBT(mp, range.len); + end = start + XFS_B_TO_FSBT(mp, range.len) - 1; minlen = XFS_B_TO_FSB(mp, max_t(u64, granularity, range.minlen)); - start_agno = XFS_FSB_TO_AGNO(mp, start); - if (start_agno >= mp->m_sb.sb_agcount) + if (start >= mp->m_sb.sb_dblocks) return -XFS_ERROR(EINVAL); + if (end > mp->m_sb.sb_dblocks - 1) + end = mp->m_sb.sb_dblocks - 1; - end_agno = XFS_FSB_TO_AGNO(mp, start + len); - if (end_agno >= mp->m_sb.sb_agcount) - end_agno = mp->m_sb.sb_agcount - 1; + start_agno = XFS_FSB_TO_AGNO(mp, start); + end_agno = XFS_FSB_TO_AGNO(mp, end); for (agno = start_agno; agno <= end_agno; agno++) { - error = -xfs_trim_extents(mp, agno, start, len, minlen, + error = -xfs_trim_extents(mp, agno, start, end, minlen, &blocks_trimmed); if (error) last_error = error; From 815cb21662b914e1e14c256a3d662b1352c8509e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 26 Sep 2011 09:14:34 +0000 Subject: [PATCH 48/69] xfs: XFS_TRANS_SWAPEXT is not a valid flag for xfs_trans_commit XFS_TRANS_SWAPEXT is a transaction type, not a flag for xfs_trans_commit, so don't pass it in xfs_swap_extents. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_dfrag.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c index 9a84a85c03b1..645387b69738 100644 --- a/fs/xfs/xfs_dfrag.c +++ b/fs/xfs/xfs_dfrag.c @@ -438,7 +438,7 @@ xfs_swap_extents( if (mp->m_flags & XFS_MOUNT_WSYNC) xfs_trans_set_sync(tp); - error = xfs_trans_commit(tp, XFS_TRANS_SWAPEXT); + error = xfs_trans_commit(tp, 0); trace_xfs_swap_extent_after(ip, 0); trace_xfs_swap_extent_after(tip, 1); From b10370585349d364ff3c550afa7922e6e21f029d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 19 Sep 2011 14:55:51 +0000 Subject: [PATCH 49/69] xfs: unlock the inode before log force in xfs_fsync Only read the LSN we need to push to with the ilock held, and then release it before we do the log force to improve concurrency. This also removes the only direct caller of _xfs_trans_commit, thus allowing it to be merged into the plain xfs_trans_commit again. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_file.c | 15 ++++++++------- fs/xfs/xfs_trans.c | 11 ++++------- fs/xfs/xfs_trans.h | 5 +---- 3 files changed, 13 insertions(+), 18 deletions(-) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 558543c146b3..d8ef02eb178a 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -137,6 +137,7 @@ xfs_file_fsync( struct xfs_trans *tp; int error = 0; int log_flushed = 0; + xfs_lsn_t lsn = 0; trace_xfs_file_fsync(ip); @@ -214,9 +215,9 @@ xfs_file_fsync( */ xfs_trans_ijoin(tp, ip); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - xfs_trans_set_sync(tp); - error = _xfs_trans_commit(tp, 0, &log_flushed); + error = xfs_trans_commit(tp, 0); + lsn = ip->i_itemp->ili_last_lsn; xfs_iunlock(ip, XFS_ILOCK_EXCL); } else { /* @@ -227,14 +228,14 @@ xfs_file_fsync( * disk yet, the inode will be still be pinned. If it is, * force the log. */ - if (xfs_ipincount(ip)) { - error = _xfs_log_force_lsn(mp, - ip->i_itemp->ili_last_lsn, - XFS_LOG_SYNC, &log_flushed); - } + if (xfs_ipincount(ip)) + lsn = ip->i_itemp->ili_last_lsn; xfs_iunlock(ip, XFS_ILOCK_SHARED); } + if (!error && lsn) + error = _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, &log_flushed); + /* * If we only have a single device, and the log force about was * a no-op we might have to flush the data device cache here. diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index efc147f0e9b6..b64c8c79736a 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -1790,9 +1790,7 @@ xfs_trans_commit_cil( } /* - * xfs_trans_commit - * - * Commit the given transaction to the log a/synchronously. + * Commit the given transaction to the log. * * XFS disk error handling mechanism is not based on a typical * transaction abort mechanism. Logically after the filesystem @@ -1804,10 +1802,9 @@ xfs_trans_commit_cil( * Do not reference the transaction structure after this call. */ int -_xfs_trans_commit( +xfs_trans_commit( struct xfs_trans *tp, - uint flags, - int *log_flushed) + uint flags) { struct xfs_mount *mp = tp->t_mountp; xfs_lsn_t commit_lsn = -1; @@ -1866,7 +1863,7 @@ _xfs_trans_commit( if (sync) { if (!error) { error = _xfs_log_force_lsn(mp, commit_lsn, - XFS_LOG_SYNC, log_flushed); + XFS_LOG_SYNC, NULL); } XFS_STATS_INC(xs_trans_sync); } else { diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 06a9759b6352..54b0b1cc3e3f 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -487,10 +487,7 @@ void xfs_trans_log_efd_extent(xfs_trans_t *, struct xfs_efd_log_item *, xfs_fsblock_t, xfs_extlen_t); -int _xfs_trans_commit(xfs_trans_t *, - uint flags, - int *); -#define xfs_trans_commit(tp, flags) _xfs_trans_commit(tp, flags, NULL) +int xfs_trans_commit(xfs_trans_t *, uint flags); void xfs_trans_cancel(xfs_trans_t *, int); int xfs_trans_ail_init(struct xfs_mount *); void xfs_trans_ail_destroy(struct xfs_mount *); From 8292d88c5c833fc8b837c3a018fd6d72c35a3231 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:47:50 +0000 Subject: [PATCH 50/69] xfs: unlock the inode before log force in xfs_fs_nfs_commit_metadata Only read the LSN we need to push to with the ilock held, and then release it before we do the log force to improve concurrency. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_export.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/xfs/xfs_export.c b/fs/xfs/xfs_export.c index 75e5d322e48f..da108977b21f 100644 --- a/fs/xfs/xfs_export.c +++ b/fs/xfs/xfs_export.c @@ -229,16 +229,16 @@ xfs_fs_nfs_commit_metadata( { struct xfs_inode *ip = XFS_I(inode); struct xfs_mount *mp = ip->i_mount; - int error = 0; + xfs_lsn_t lsn = 0; xfs_ilock(ip, XFS_ILOCK_SHARED); - if (xfs_ipincount(ip)) { - error = _xfs_log_force_lsn(mp, ip->i_itemp->ili_last_lsn, - XFS_LOG_SYNC, NULL); - } + if (xfs_ipincount(ip)) + lsn = ip->i_itemp->ili_last_lsn; xfs_iunlock(ip, XFS_ILOCK_SHARED); - return error; + if (!lsn) + return 0; + return _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, NULL); } const struct export_operations xfs_export_operations = { From 23bb0be1a237c8732ce1a43140e5cb103a676b92 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 18 Sep 2011 20:47:51 +0000 Subject: [PATCH 51/69] xfs: unlock the inode before log force in xfs_change_file_space Let the transaction commit unlock the inode before it potentially causes a synchronous log force. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_vnodeops.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index f47ecee8d437..c9c8e8230b21 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -2349,8 +2349,7 @@ xfs_change_file_space( } xfs_ilock(ip, XFS_ILOCK_EXCL); - - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); if ((attr_flags & XFS_ATTR_DMI) == 0) { ip->i_d.di_mode &= ~S_ISUID; @@ -2375,10 +2374,5 @@ xfs_change_file_space( xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); if (attr_flags & XFS_ATTR_SYNC) xfs_trans_set_sync(tp); - - error = xfs_trans_commit(tp, 0); - - xfs_iunlock(ip, XFS_ILOCK_EXCL); - - return error; + return xfs_trans_commit(tp, 0); } From ddc3415aba1cb2f86d1fcad720cea834ee178f54 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 19 Sep 2011 15:00:54 +0000 Subject: [PATCH 52/69] xfs: simplify xfs_trans_ijoin* again There is no reason to keep a reference to the inode even if we unlock it during transaction commit because we never drop a reference between the ijoin and commit. Also use this fact to merge xfs_trans_ijoin_ref back into xfs_trans_ijoin - the third argument decides if an unlock is needed now. I'm actually starting to wonder if allowing inodes to be unlocked at transaction commit really is worth the effort. The only real benefit is that they can be unlocked earlier when commiting a synchronous transactions, but that could be solved by doing the log force manually after the unlock, too. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 28 ++++++++++++++-------------- fs/xfs/xfs_bmap.c | 4 ++-- fs/xfs/xfs_dfrag.c | 4 ++-- fs/xfs/xfs_dquot.c | 2 +- fs/xfs/xfs_file.c | 2 +- fs/xfs/xfs_inode.c | 6 +++--- fs/xfs/xfs_inode_item.c | 4 +--- fs/xfs/xfs_ioctl.c | 2 +- fs/xfs/xfs_iomap.c | 6 +++--- fs/xfs/xfs_iops.c | 4 ++-- fs/xfs/xfs_qm_syscalls.c | 2 +- fs/xfs/xfs_rename.c | 8 ++++---- fs/xfs/xfs_rtalloc.c | 10 +++++----- fs/xfs/xfs_super.c | 2 +- fs/xfs/xfs_trans.c | 2 +- fs/xfs/xfs_trans.h | 3 +-- fs/xfs/xfs_trans_inode.c | 25 +++++-------------------- fs/xfs/xfs_vnodeops.c | 34 +++++++++++++++++----------------- 18 files changed, 65 insertions(+), 83 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 5484766938f9..981a0624f382 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -319,7 +319,7 @@ xfs_attr_set_int( return (error); } - xfs_trans_ijoin(args.trans, dp); + xfs_trans_ijoin(args.trans, dp, 0); /* * If the attribute list is non-existent or a shortform list, @@ -389,7 +389,7 @@ xfs_attr_set_int( * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args.trans, dp); + xfs_trans_ijoin(args.trans, dp, 0); /* * Commit the leaf transformation. We'll need another (linked) @@ -537,7 +537,7 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags) * No need to make quota reservations here. We expect to release some * blocks not allocate in the common case. */ - xfs_trans_ijoin(args.trans, dp); + xfs_trans_ijoin(args.trans, dp, 0); /* * Decide on what work routines to call based on the inode size. @@ -809,7 +809,7 @@ xfs_attr_inactive(xfs_inode_t *dp) * No need to make quota reservations here. We expect to release some * blocks, not allocate, in the common case. */ - xfs_trans_ijoin(trans, dp); + xfs_trans_ijoin(trans, dp, 0); /* * Decide on what work routines to call based on the inode size. @@ -961,7 +961,7 @@ xfs_attr_leaf_addname(xfs_da_args_t *args) * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); /* * Commit the current trans (including the inode) and start @@ -1063,7 +1063,7 @@ xfs_attr_leaf_addname(xfs_da_args_t *args) * in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); } else xfs_da_buf_done(bp); @@ -1137,7 +1137,7 @@ xfs_attr_leaf_removename(xfs_da_args_t *args) * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); } else xfs_da_buf_done(bp); return(0); @@ -1291,7 +1291,7 @@ xfs_attr_node_addname(xfs_da_args_t *args) * in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); /* * Commit the node conversion and start the next @@ -1328,7 +1328,7 @@ xfs_attr_node_addname(xfs_da_args_t *args) * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); } else { /* * Addition succeeded, update Btree hashvals. @@ -1440,7 +1440,7 @@ xfs_attr_node_addname(xfs_da_args_t *args) * in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); } /* @@ -1572,7 +1572,7 @@ xfs_attr_node_removename(xfs_da_args_t *args) * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); /* * Commit the Btree join operation and start a new trans. @@ -1623,7 +1623,7 @@ xfs_attr_node_removename(xfs_da_args_t *args) * in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); } else xfs_da_brelse(args->trans, bp); } @@ -2060,7 +2060,7 @@ xfs_attr_rmtval_set(xfs_da_args_t *args) * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, dp); + xfs_trans_ijoin(args->trans, dp, 0); ASSERT(nmap == 1); ASSERT((map.br_startblock != DELAYSTARTBLOCK) && @@ -2207,7 +2207,7 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) * a new one. We need the inode to be in all transactions. */ if (committed) - xfs_trans_ijoin(args->trans, args->dp); + xfs_trans_ijoin(args->trans, args->dp, 0); /* * Close out trans and start the next one in the chain. diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index bb31d37bf49b..c68baeb0974a 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2195,7 +2195,7 @@ xfs_bmap_rtalloc( * Lock out other modifications to the RT bitmap inode. */ xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(ap->tp, mp->m_rbmip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(ap->tp, mp->m_rbmip, XFS_ILOCK_EXCL); /* * If it's an allocation to an empty file at offset 0, @@ -3460,7 +3460,7 @@ xfs_bmap_add_attrfork( } ASSERT(ip->i_d.di_anextents == 0); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); switch (ip->i_d.di_format) { diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c index 645387b69738..654dc6f05bac 100644 --- a/fs/xfs/xfs_dfrag.c +++ b/fs/xfs/xfs_dfrag.c @@ -425,8 +425,8 @@ xfs_swap_extents( } - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); - xfs_trans_ijoin_ref(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + xfs_trans_ijoin(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); xfs_trans_log_inode(tp, ip, ilf_fields); xfs_trans_log_inode(tp, tip, tilf_fields); diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 179673531f20..c597bfe4ada0 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -377,7 +377,7 @@ xfs_qm_dqalloc( return (ESRCH); } - xfs_trans_ijoin_ref(tp, quotip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL); nmaps = 1; error = xfs_bmapi_write(tp, quotip, offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA, diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index d8ef02eb178a..0b600b51778c 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -213,7 +213,7 @@ xfs_file_fsync( * transaction. So we play it safe and fire off the * transaction anyway. */ - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); error = xfs_trans_commit(tp, 0); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index f690d3a10b34..b676494a55ca 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1152,7 +1152,7 @@ xfs_ialloc( /* * Log the new values stuffed into the inode. */ - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); xfs_trans_log_inode(tp, ip, flags); /* now that we have an i_mode we can setup inode ops and unlock */ @@ -1297,7 +1297,7 @@ xfs_itruncate_extents( */ error = xfs_bmap_finish(&tp, &free_list, &committed); if (committed) - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); if (error) goto out_bmap_cancel; @@ -1313,7 +1313,7 @@ xfs_itruncate_extents( error = xfs_trans_commit(tp, 0); tp = ntp; - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); if (error) goto out; diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 588406dc6a35..8704a99241d7 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -658,10 +658,8 @@ xfs_inode_item_unlock( lock_flags = iip->ili_lock_flags; iip->ili_lock_flags = 0; - if (lock_flags) { + if (lock_flags) xfs_iunlock(ip, lock_flags); - IRELE(ip); - } } /* diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index f7ce7debe14c..d99a90518909 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -1069,7 +1069,7 @@ xfs_ioctl_setattr( } } - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); /* * Change file ownership. Must be the owner or privileged. diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index da5bf05c5bb7..9afa282aa937 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -208,7 +208,7 @@ xfs_iomap_write_direct( if (error) goto error1; - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); bmapi_flag = 0; if (offset < ip->i_size || extsz) @@ -528,7 +528,7 @@ xfs_iomap_write_allocate( return XFS_ERROR(error); } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); xfs_bmap_init(&free_list, &first_block); @@ -692,7 +692,7 @@ xfs_iomap_write_unwritten( } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); /* * Modify the unwritten extent state of the buffer. diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index e6b3e7620888..556bbe7751b7 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -611,7 +611,7 @@ xfs_setattr_nonsize( } } - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); /* * Change file ownership. Must be the owner or privileged. @@ -863,7 +863,7 @@ xfs_setattr_size( xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); /* * Only change the c/mtime if we are changing the size or we are diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c index 609246f42e6c..5cc3dde1bc90 100644 --- a/fs/xfs/xfs_qm_syscalls.c +++ b/fs/xfs/xfs_qm_syscalls.c @@ -261,7 +261,7 @@ xfs_qm_scall_trunc_qfile( } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); error = xfs_itruncate_data(&tp, ip, 0); if (error) { diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c index df78c297d1a1..866de277079a 100644 --- a/fs/xfs/xfs_rename.c +++ b/fs/xfs/xfs_rename.c @@ -170,12 +170,12 @@ xfs_rename( * we can rely on either trans_commit or trans_cancel to unlock * them. */ - xfs_trans_ijoin_ref(tp, src_dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL); if (new_parent) - xfs_trans_ijoin_ref(tp, target_dp, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, src_ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL); if (target_ip) - xfs_trans_ijoin_ref(tp, target_ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL); /* * If we are using project inheritance, we only allow renames diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index f29424964521..87323f1ded64 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -112,7 +112,7 @@ xfs_growfs_rt_alloc( * Lock the inode. */ xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); xfs_bmap_init(&flist, &firstblock); /* @@ -155,7 +155,7 @@ xfs_growfs_rt_alloc( * Lock the bitmap inode. */ xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); /* * Get a buffer for the block. */ @@ -1960,7 +1960,7 @@ xfs_growfs_rt( * Lock out other callers by grabbing the bitmap inode lock. */ xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, mp->m_rbmip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL); /* * Update the bitmap inode's size. */ @@ -1972,7 +1972,7 @@ xfs_growfs_rt( * Get the summary inode into the transaction. */ xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, mp->m_rsumip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, mp->m_rsumip, XFS_ILOCK_EXCL); /* * Update the summary inode's size. */ @@ -2143,7 +2143,7 @@ xfs_rtfree_extent( * Synchronize by locking the bitmap inode. */ xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, mp->m_rbmip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL); #if defined(__KERNEL__) && defined(DEBUG) /* diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 54d5e102ffe1..6ad05e68abda 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -884,7 +884,7 @@ xfs_log_inode( } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); return xfs_trans_commit(tp, 0); } diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index b64c8c79736a..1f35b2feca97 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -2018,6 +2018,6 @@ xfs_trans_roll( if (error) return error; - xfs_trans_ijoin(trans, dp); + xfs_trans_ijoin(trans, dp, 0); return 0; } diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 54b0b1cc3e3f..f5df16969f82 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -470,8 +470,7 @@ void xfs_trans_stale_inode_buf(xfs_trans_t *, struct xfs_buf *); void xfs_trans_dquot_buf(xfs_trans_t *, struct xfs_buf *, uint); void xfs_trans_inode_alloc_buf(xfs_trans_t *, struct xfs_buf *); void xfs_trans_ichgtime(struct xfs_trans *, struct xfs_inode *, int); -void xfs_trans_ijoin_ref(struct xfs_trans *, struct xfs_inode *, uint); -void xfs_trans_ijoin(struct xfs_trans *, struct xfs_inode *); +void xfs_trans_ijoin(struct xfs_trans *, struct xfs_inode *, uint); void xfs_trans_log_buf(xfs_trans_t *, struct xfs_buf *, uint, uint); void xfs_trans_log_inode(xfs_trans_t *, struct xfs_inode *, uint); struct xfs_efi_log_item *xfs_trans_get_efi(xfs_trans_t *, uint); diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c index c8dea2fd7e68..32f0288ae10f 100644 --- a/fs/xfs/xfs_trans_inode.c +++ b/fs/xfs/xfs_trans_inode.c @@ -47,11 +47,13 @@ xfs_trans_inode_broot_debug( * Add a locked inode to the transaction. * * The inode must be locked, and it cannot be associated with any transaction. + * If lock_flags is non-zero the inode will be unlocked on transaction commit. */ void xfs_trans_ijoin( struct xfs_trans *tp, - struct xfs_inode *ip) + struct xfs_inode *ip, + uint lock_flags) { xfs_inode_log_item_t *iip; @@ -59,7 +61,9 @@ xfs_trans_ijoin( if (ip->i_itemp == NULL) xfs_inode_item_init(ip, ip->i_mount); iip = ip->i_itemp; + ASSERT(iip->ili_lock_flags == 0); + iip->ili_lock_flags = lock_flags; /* * Get a log_item_desc to point at the new item. @@ -69,25 +73,6 @@ xfs_trans_ijoin( xfs_trans_inode_broot_debug(ip); } -/* - * Add a locked inode to the transaction. - * - * - * Grabs a reference to the inode which will be dropped when the transaction - * is committed. The inode will also be unlocked at that point. The inode - * must be locked, and it cannot be associated with any transaction. - */ -void -xfs_trans_ijoin_ref( - struct xfs_trans *tp, - struct xfs_inode *ip, - uint lock_flags) -{ - xfs_trans_ijoin(tp, ip); - IHOLD(ip); - ip->i_itemp->ili_lock_flags = lock_flags; -} - /* * Transactional inode timestamp update. Requires the inode to be locked and * joined to the transaction supplied. Relies on the transaction subsystem to diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index c9c8e8230b21..fc38f15808c6 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -219,7 +219,7 @@ xfs_free_eofblocks( } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); error = xfs_itruncate_data(&tp, ip, ip->i_size); if (error) { @@ -288,7 +288,7 @@ xfs_inactive_symlink_rmt( xfs_ilock(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); size = (int)ip->i_d.di_size; ip->i_d.di_size = 0; - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); /* * Find the block(s) so we can inval and unmap them. @@ -336,7 +336,7 @@ xfs_inactive_symlink_rmt( * Mark it dirty so it will be logged and moved forward in the log as * part of every commit. */ - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); /* * Get a new, empty transaction to return to our caller. @@ -469,7 +469,7 @@ xfs_inactive_attrs( goto error_cancel; xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); xfs_idestroy_fork(ip, XFS_ATTR_FORK); ASSERT(ip->i_d.di_anextents == 0); @@ -663,7 +663,7 @@ xfs_inactive( } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); error = xfs_itruncate_data(&tp, ip, 0); if (error) { @@ -687,7 +687,7 @@ xfs_inactive( return VN_INACTIVE_CACHE; } - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); } else { error = xfs_trans_reserve(tp, 0, XFS_IFREE_LOG_RES(mp), @@ -700,7 +700,7 @@ xfs_inactive( } xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); } /* @@ -940,7 +940,7 @@ xfs_create( * the transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - xfs_trans_ijoin_ref(tp, dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = B_FALSE; error = xfs_dir_createname(tp, dp, name, ip->i_ino, @@ -1261,8 +1261,8 @@ xfs_remove( xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, dp, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); /* * If we're removing a directory perform some additional validation. @@ -1407,8 +1407,8 @@ xfs_link( xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, sip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, tdp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL); /* * If the source has too many links, we can't make any more to it. @@ -1602,7 +1602,7 @@ xfs_symlink( * transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - xfs_trans_ijoin_ref(tp, dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = B_FALSE; /* @@ -1735,7 +1735,7 @@ xfs_set_dmattrs( return error; } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); ip->i_d.di_dmevmask = evmask; ip->i_d.di_dmstate = state; @@ -1878,7 +1878,7 @@ xfs_alloc_file_space( if (error) goto error1; - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); xfs_bmap_init(&free_list, &firstfsb); error = xfs_bmapi_write(tp, ip, startoffset_fsb, @@ -2176,7 +2176,7 @@ xfs_free_file_space( if (error) goto error1; - xfs_trans_ijoin(tp, ip); + xfs_trans_ijoin(tp, ip, 0); /* * issue the bunmapi() call to free the blocks @@ -2349,7 +2349,7 @@ xfs_change_file_space( } xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); if ((attr_flags & XFS_ATTR_DMI) == 0) { ip->i_d.di_mode &= ~S_ISUID; From 3815832a2aa4df9815d15dac05227e0c8551833f Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 30 Sep 2011 04:45:02 +0000 Subject: [PATCH 53/69] xfs: Don't allocate new buffers on every call to _xfs_buf_find Stats show that for an 8-way unlink @ ~80,000 unlinks/s we are doing ~1 million cache hit lookups to ~3000 buffer creates. That's almost 3 orders of magnitude more cahce hits than misses, so optimising for cache hits is quite important. In the cache hit case, we do not need to allocate a new buffer in case of a cache miss, so we are effectively hitting the allocator for no good reason for vast the majority of calls to _xfs_buf_find. 8-way create workloads are showing similar cache hit/miss ratios. The result is profiles that look like this: samples pcnt function DSO _______ _____ _______________________________ _________________ 1036.00 10.0% _xfs_buf_find [kernel.kallsyms] 582.00 5.6% kmem_cache_alloc [kernel.kallsyms] 519.00 5.0% __memcpy [kernel.kallsyms] 468.00 4.5% __ticket_spin_lock [kernel.kallsyms] 388.00 3.7% kmem_cache_free [kernel.kallsyms] 331.00 3.2% xfs_log_commit_cil [kernel.kallsyms] Further, there is a fair bit of work involved in initialising a new buffer once a cache miss has occurred and we currently do that under the rbtree spinlock. That increases spinlock hold time on what are heavily used trees. To fix this, remove the initialisation of the buffer from _xfs_buf_find() and only allocate the new buffer once we've had a cache miss. Initialise the buffer immediately after allocating it in xfs_buf_get, too, so that is it ready for insert if we get another cache miss after allocation. This minimises lock hold time and avoids unnecessary allocator churn. The resulting profiles look like: samples pcnt function DSO _______ _____ ___________________________ _________________ 8111.00 9.1% _xfs_buf_find [kernel.kallsyms] 4380.00 4.9% __memcpy [kernel.kallsyms] 4341.00 4.8% __ticket_spin_lock [kernel.kallsyms] 3401.00 3.8% kmem_cache_alloc [kernel.kallsyms] 2856.00 3.2% xfs_log_commit_cil [kernel.kallsyms] 2625.00 2.9% __kmalloc [kernel.kallsyms] 2380.00 2.7% kfree [kernel.kallsyms] 2016.00 2.3% kmem_cache_free [kernel.kallsyms] Showing a significant reduction in time spent doing allocation and freeing from slabs (kmem_cache_alloc and kmem_cache_free). Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 48 ++++++++++++++++++++++++++++-------------------- 1 file changed, 28 insertions(+), 20 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index e3af85095ddb..6785b7bd952d 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -477,8 +477,6 @@ _xfs_buf_find( /* No match found */ if (new_bp) { - _xfs_buf_initialize(new_bp, btp, range_base, - range_length, flags); rb_link_node(&new_bp->b_rbnode, parent, rbp); rb_insert_color(&new_bp->b_rbnode, &pag->pag_buf_tree); /* the buffer keeps the perag reference until it is freed */ @@ -521,35 +519,53 @@ _xfs_buf_find( } /* - * Assembles a buffer covering the specified range. - * Storage in memory for all portions of the buffer will be allocated, - * although backing storage may not be. + * Assembles a buffer covering the specified range. The code is optimised for + * cache hits, as metadata intensive workloads will see 3 orders of magnitude + * more hits than misses. */ -xfs_buf_t * +struct xfs_buf * xfs_buf_get( xfs_buftarg_t *target,/* target for buffer */ xfs_off_t ioff, /* starting offset of range */ size_t isize, /* length of range */ xfs_buf_flags_t flags) { - xfs_buf_t *bp, *new_bp; + struct xfs_buf *bp; + struct xfs_buf *new_bp; int error = 0; + bp = _xfs_buf_find(target, ioff, isize, flags, NULL); + if (likely(bp)) + goto found; + new_bp = xfs_buf_allocate(flags); if (unlikely(!new_bp)) return NULL; + _xfs_buf_initialize(new_bp, target, + ioff << BBSHIFT, isize << BBSHIFT, flags); + bp = _xfs_buf_find(target, ioff, isize, flags, new_bp); + if (!bp) { + xfs_buf_deallocate(new_bp); + return NULL; + } + if (bp == new_bp) { error = xfs_buf_allocate_memory(bp, flags); if (error) goto no_buffer; - } else { + } else xfs_buf_deallocate(new_bp); - if (unlikely(bp == NULL)) - return NULL; - } + /* + * Now we have a workable buffer, fill in the block number so + * that we can do IO on it. + */ + bp->b_bn = ioff; + bp->b_count_desired = bp->b_buffer_length; + +found: if (!(bp->b_flags & XBF_MAPPED)) { error = _xfs_buf_map_pages(bp, flags); if (unlikely(error)) { @@ -560,18 +576,10 @@ xfs_buf_get( } XFS_STATS_INC(xb_get); - - /* - * Always fill in the block number now, the mapped cases can do - * their own overlay of this later. - */ - bp->b_bn = ioff; - bp->b_count_desired = bp->b_buffer_length; - trace_xfs_buf_get(bp, flags, _RET_IP_); return bp; - no_buffer: +no_buffer: if (flags & (XBF_LOCK | XBF_TRYLOCK)) xfs_buf_unlock(bp); xfs_buf_rele(bp); From 670ce93fef93bba8c8a422a79747385bec8e846a Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Fri, 30 Sep 2011 04:45:03 +0000 Subject: [PATCH 54/69] xfs: reduce the number of log forces from tail pushing The AIL push code will issue a log force on ever single push loop that it exits and has encountered pinned items. It doesn't rescan these pinned items until it revisits the AIL from the start. Hence we only need to force the log once per walk from the start of the AIL to the target LSN. This results in numbers like this: xs_push_ail_flush..... 1456 xs_log_force......... 1485 For an 8-way 50M inode create workload - almost all the log forces are coming from the AIL pushing code. Reduce the number of log forces by only forcing the log if the previous walk found pinned buffers. This reduces the numbers to: xs_push_ail_flush..... 665 xs_log_force......... 682 For the same test. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_trans_ail.c | 33 ++++++++++++++++++++------------- fs/xfs/xfs_trans_priv.h | 1 + 2 files changed, 21 insertions(+), 13 deletions(-) diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index c15aa29fa169..9df7f9f1b5ee 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -372,12 +372,24 @@ xfs_ail_worker( xfs_lsn_t lsn; xfs_lsn_t target; long tout = 10; - int flush_log = 0; int stuck = 0; int count = 0; int push_xfsbufd = 0; + /* + * If last time we ran we encountered pinned items, force the log first + * and wait for it before pushing again. + */ spin_lock(&ailp->xa_lock); + if (ailp->xa_last_pushed_lsn == 0 && ailp->xa_log_flush && + !list_empty(&ailp->xa_ail)) { + ailp->xa_log_flush = 0; + spin_unlock(&ailp->xa_lock); + XFS_STATS_INC(xs_push_ail_flush); + xfs_log_force(mp, XFS_LOG_SYNC); + spin_lock(&ailp->xa_lock); + } + target = ailp->xa_target; lip = xfs_trans_ail_cursor_first(ailp, &cur, ailp->xa_last_pushed_lsn); if (!lip || XFS_FORCED_SHUTDOWN(mp)) { @@ -435,7 +447,7 @@ xfs_ail_worker( case XFS_ITEM_PINNED: XFS_STATS_INC(xs_push_ail_pinned); stuck++; - flush_log = 1; + ailp->xa_log_flush++; break; case XFS_ITEM_LOCKED: @@ -480,16 +492,6 @@ xfs_ail_worker( xfs_trans_ail_cursor_done(ailp, &cur); spin_unlock(&ailp->xa_lock); - if (flush_log) { - /* - * If something we need to push out was pinned, then - * push out the log so it will become unpinned and - * move forward in the AIL. - */ - XFS_STATS_INC(xs_push_ail_flush); - xfs_log_force(mp, 0); - } - if (push_xfsbufd) { /* we've got delayed write buffers to flush */ wake_up_process(mp->m_ddev_targp->bt_task); @@ -500,6 +502,7 @@ xfs_ail_worker( if (!count) { /* We're past our target or empty, so idle */ ailp->xa_last_pushed_lsn = 0; + ailp->xa_log_flush = 0; /* * We clear the XFS_AIL_PUSHING_BIT first before checking @@ -532,9 +535,13 @@ xfs_ail_worker( * were stuck. * * Backoff a bit more to allow some I/O to complete before - * continuing from where we were. + * restarting from the start of the AIL. This prevents us + * from spinning on the same items, and if they are pinned will + * all the restart to issue a log force to unpin the stuck + * items. */ tout = 20; + ailp->xa_last_pushed_lsn = 0; } /* There is more to do, requeue us. */ diff --git a/fs/xfs/xfs_trans_priv.h b/fs/xfs/xfs_trans_priv.h index 212946b97239..0a6eec6d472a 100644 --- a/fs/xfs/xfs_trans_priv.h +++ b/fs/xfs/xfs_trans_priv.h @@ -71,6 +71,7 @@ struct xfs_ail { struct delayed_work xa_work; xfs_lsn_t xa_last_pushed_lsn; unsigned long xa_flags; + int xa_log_flush; }; #define XFS_AIL_PUSHING_BIT 0 From 1da2f2dbf2d2aaa1b0f6ca2f61fcf07e24eb659b Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Sun, 2 Oct 2011 14:25:16 +0000 Subject: [PATCH 55/69] xfs: optimize fsync on directories Directories are only updated transactionally, which means fsync only needs to flush the log the inode is currently dirty, but not bother with checking for dirty data, non-transactional updates, and most importanly doesn't have to flush disk caches except as part of a transaction commit. While the first two optimizations can't easily be measured, the latter actually makes a difference when doing lots of fsync that do not actually have to commit the inode, e.g. because an earlier fsync already pushed the log far enough. The new xfs_dir_fsync is identical to xfs_nfs_commit_metadata except for the prototype, but I'm not sure creating a common helper for the two is worth it given how simple the functions are. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder --- fs/xfs/xfs_file.c | 31 ++++++++++++++++++++++++++++++- fs/xfs/xfs_trace.h | 1 + 2 files changed, 31 insertions(+), 1 deletion(-) diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 0b600b51778c..753ed9b5c70b 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -124,6 +124,35 @@ xfs_iozero( return (-status); } +/* + * Fsync operations on directories are much simpler than on regular files, + * as there is no file data to flush, and thus also no need for explicit + * cache flush operations, and there are no non-transaction metadata updates + * on directories either. + */ +STATIC int +xfs_dir_fsync( + struct file *file, + loff_t start, + loff_t end, + int datasync) +{ + struct xfs_inode *ip = XFS_I(file->f_mapping->host); + struct xfs_mount *mp = ip->i_mount; + xfs_lsn_t lsn = 0; + + trace_xfs_dir_fsync(ip); + + xfs_ilock(ip, XFS_ILOCK_SHARED); + if (xfs_ipincount(ip)) + lsn = ip->i_itemp->ili_last_lsn; + xfs_iunlock(ip, XFS_ILOCK_SHARED); + + if (!lsn) + return 0; + return _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, NULL); +} + STATIC int xfs_file_fsync( struct file *file, @@ -1140,7 +1169,7 @@ const struct file_operations xfs_dir_file_operations = { #ifdef CONFIG_COMPAT .compat_ioctl = xfs_file_compat_ioctl, #endif - .fsync = xfs_file_fsync, + .fsync = xfs_dir_fsync, }; static const struct vm_operations_struct xfs_file_vm_ops = { diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index bb5e660e0fab..b78f2c65438b 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -576,6 +576,7 @@ DEFINE_INODE_EVENT(xfs_vm_bmap); DEFINE_INODE_EVENT(xfs_file_ioctl); DEFINE_INODE_EVENT(xfs_file_compat_ioctl); DEFINE_INODE_EVENT(xfs_ioctl_setattr); +DEFINE_INODE_EVENT(xfs_dir_fsync); DEFINE_INODE_EVENT(xfs_file_fsync); DEFINE_INODE_EVENT(xfs_destroy_inode); DEFINE_INODE_EVENT(xfs_write_inode); From 87c7bec7fc3377b3873eb3a0f4b603981ea16ebb Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 14 Sep 2011 14:08:26 +0000 Subject: [PATCH 56/69] xfs: fix buffer flushing during unmount The code to flush buffers in the umount code is a bit iffy: we first flush all delwri buffers out, but then might be able to queue up a new one when logging the sb counts. On a normal shutdown that one would get flushed out when doing the synchronous superblock write in xfs_unmountfs_writesb, but we skip that one if the filesystem has been shut down. Fix this by moving the delwri list flushing until just before unmounting the log, and while we're at it also remove the superflous delwri list and buffer lru flusing for the rt and log device that can never have cached or delwri buffers. Signed-off-by: Christoph Hellwig Reported-by: Amit Sahrawat Tested-by: Amit Sahrawat Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.h | 1 - fs/xfs/xfs_mount.c | 29 ++++++++++------------------- 2 files changed, 10 insertions(+), 20 deletions(-) diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index c23826685d3d..200f59138b8a 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -318,7 +318,6 @@ extern struct list_head *xfs_get_buftarg_list(void); #define xfs_getsize_buftarg(buftarg) block_size((buftarg)->bt_bdev) #define xfs_readonly_buftarg(buftarg) bdev_read_only((buftarg)->bt_bdev) -#define xfs_binval(buftarg) xfs_flush_buftarg(buftarg, 1) #define XFS_bflush(buftarg) xfs_flush_buftarg(buftarg, 1) #endif /* __XFS_BUF_H__ */ diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index a957e437bee1..e8fe1cb54094 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -44,9 +44,6 @@ #include "xfs_trace.h" -STATIC void xfs_unmountfs_wait(xfs_mount_t *); - - #ifdef HAVE_PERCPU_SB STATIC void xfs_icsb_balance_counter(xfs_mount_t *, xfs_sb_field_t, int); @@ -1496,11 +1493,6 @@ xfs_unmountfs( */ xfs_log_force(mp, XFS_LOG_SYNC); - xfs_binval(mp->m_ddev_targp); - if (mp->m_rtdev_targp) { - xfs_binval(mp->m_rtdev_targp); - } - /* * Unreserve any blocks we have so that when we unmount we don't account * the reserved free space as used. This is really only necessary for @@ -1526,7 +1518,16 @@ xfs_unmountfs( xfs_warn(mp, "Unable to update superblock counters. " "Freespace may not be correct on next mount."); xfs_unmountfs_writesb(mp); - xfs_unmountfs_wait(mp); /* wait for async bufs */ + + /* + * Make sure all buffers have been flushed and completed before + * unmounting the log. + */ + error = xfs_flush_buftarg(mp->m_ddev_targp, 1); + if (error) + xfs_warn(mp, "%d busy buffers during unmount.", error); + xfs_wait_buftarg(mp->m_ddev_targp); + xfs_log_unmount_write(mp); xfs_log_unmount(mp); xfs_uuid_unmount(mp); @@ -1537,16 +1538,6 @@ xfs_unmountfs( xfs_free_perag(mp); } -STATIC void -xfs_unmountfs_wait(xfs_mount_t *mp) -{ - if (mp->m_logdev_targp != mp->m_ddev_targp) - xfs_wait_buftarg(mp->m_logdev_targp); - if (mp->m_rtdev_targp) - xfs_wait_buftarg(mp->m_rtdev_targp); - xfs_wait_buftarg(mp->m_ddev_targp); -} - int xfs_fs_writable(xfs_mount_t *mp) { From b17b833443a3b65907f5ecb36f8af33996f6ec78 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:43 +0000 Subject: [PATCH 57/69] xfs: remove xfs_get_buftarg_list The code is unused and under a config option that doesn't exist, remove it. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 8 -------- fs/xfs/xfs_buf.h | 4 ---- 2 files changed, 12 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 6785b7bd952d..e0339a4a0bc8 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1835,11 +1835,3 @@ xfs_buf_terminate(void) destroy_workqueue(xfslogd_workqueue); kmem_zone_destroy(xfs_buf_zone); } - -#ifdef CONFIG_KDB_MODULES -struct list_head * -xfs_get_buftarg_list(void) -{ - return &xfs_buftarg_list; -} -#endif diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 200f59138b8a..cb54cc234ed1 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -311,10 +311,6 @@ extern void xfs_wait_buftarg(xfs_buftarg_t *); extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int); extern int xfs_flush_buftarg(xfs_buftarg_t *, int); -#ifdef CONFIG_KDB_MODULES -extern struct list_head *xfs_get_buftarg_list(void); -#endif - #define xfs_getsize_buftarg(buftarg) block_size((buftarg)->bt_bdev) #define xfs_readonly_buftarg(buftarg) bdev_read_only((buftarg)->bt_bdev) From 5fde0326ddb1472ef31034c8ed952a19d4679191 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:44 +0000 Subject: [PATCH 58/69] xfs: remove XFS_BUF_FINISH_IOWAIT Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 2 +- fs/xfs/xfs_buf.h | 2 -- 2 files changed, 1 insertion(+), 3 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index e0339a4a0bc8..36fed03da26f 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1100,7 +1100,7 @@ xfs_bioerror_relse( * ASYNC buffers. */ xfs_buf_ioerror(bp, EIO); - XFS_BUF_FINISH_IOWAIT(bp); + complete(&bp->b_iowait); } else { xfs_buf_relse(bp); } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index cb54cc234ed1..65181220c9a2 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -293,8 +293,6 @@ static inline int xfs_buf_ispinned(struct xfs_buf *bp) return atomic_read(&bp->b_pin_count); } -#define XFS_BUF_FINISH_IOWAIT(bp) complete(&bp->b_iowait); - static inline void xfs_buf_relse(xfs_buf_t *bp) { xfs_buf_unlock(bp); From 38f23232449c9d2c0bc8e9541cb8ab08b7c2b9ce Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:45 +0000 Subject: [PATCH 59/69] xfs: remove XFS_BUF_SET_VTYPE and XFS_BUF_SET_VTYPE_REF Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_alloc.c | 4 ++-- fs/xfs/xfs_btree.c | 8 ++++---- fs/xfs/xfs_buf.h | 7 +------ fs/xfs/xfs_da_btree.c | 11 ++++------- fs/xfs/xfs_dquot.c | 2 +- fs/xfs/xfs_ialloc.c | 2 +- fs/xfs/xfs_inode.c | 6 ------ 7 files changed, 13 insertions(+), 27 deletions(-) diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index bdd9cb54d63b..ce84ffd0264c 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -452,7 +452,7 @@ xfs_alloc_read_agfl( if (error) return error; ASSERT(!xfs_buf_geterror(bp)); - XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGFL, XFS_AGFL_REF); + xfs_buf_set_ref(bp, XFS_AGFL_REF); *bpp = bp; return 0; } @@ -2139,7 +2139,7 @@ xfs_read_agf( xfs_trans_brelse(tp, *bpp); return XFS_ERROR(EFSCORRUPTED); } - XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_AGF, XFS_AGF_REF); + xfs_buf_set_ref(*bpp, XFS_AGF_REF); return 0; } diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c index 28cc0199dc1f..1f19f03af9d3 100644 --- a/fs/xfs/xfs_btree.c +++ b/fs/xfs/xfs_btree.c @@ -631,7 +631,7 @@ xfs_btree_read_bufl( } ASSERT(!xfs_buf_geterror(bp)); if (bp) - XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, refval); + xfs_buf_set_ref(bp, refval); *bpp = bp; return 0; } @@ -939,13 +939,13 @@ xfs_btree_set_refs( switch (cur->bc_btnum) { case XFS_BTNUM_BNO: case XFS_BTNUM_CNT: - XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, XFS_ALLOC_BTREE_REF); + xfs_buf_set_ref(bp, XFS_ALLOC_BTREE_REF); break; case XFS_BTNUM_INO: - XFS_BUF_SET_VTYPE_REF(bp, B_FS_INOMAP, XFS_INO_BTREE_REF); + xfs_buf_set_ref(bp, XFS_INO_BTREE_REF); break; case XFS_BTNUM_BMAP: - XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, XFS_BMAP_BTREE_REF); + xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF); break; default: ASSERT(0); diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 65181220c9a2..ca2934717343 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -278,15 +278,10 @@ void xfs_buf_stale(struct xfs_buf *bp); #define XFS_BUF_SIZE(bp) ((bp)->b_buffer_length) #define XFS_BUF_SET_SIZE(bp, cnt) ((bp)->b_buffer_length = (cnt)) -static inline void -xfs_buf_set_ref( - struct xfs_buf *bp, - int lru_ref) +static inline void xfs_buf_set_ref(struct xfs_buf *bp, int lru_ref) { atomic_set(&bp->b_lru_ref, lru_ref); } -#define XFS_BUF_SET_VTYPE_REF(bp, type, ref) xfs_buf_set_ref(bp, ref) -#define XFS_BUF_SET_VTYPE(bp, type) do { } while (0) static inline int xfs_buf_ispinned(struct xfs_buf *bp) { diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c index 46c8aa2740da..77c74257c2a3 100644 --- a/fs/xfs/xfs_da_btree.c +++ b/fs/xfs/xfs_da_btree.c @@ -2053,13 +2053,10 @@ xfs_da_do_buf( if (!bp) continue; if (caller == 1) { - if (whichfork == XFS_ATTR_FORK) { - XFS_BUF_SET_VTYPE_REF(bp, B_FS_ATTR_BTREE, - XFS_ATTR_BTREE_REF); - } else { - XFS_BUF_SET_VTYPE_REF(bp, B_FS_DIR_BTREE, - XFS_DIR_BTREE_REF); - } + if (whichfork == XFS_ATTR_FORK) + xfs_buf_set_ref(bp, XFS_ATTR_BTREE_REF); + else + xfs_buf_set_ref(bp, XFS_DIR_BTREE_REF); } if (bplist) { bplist[nbplist++] = bp; diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index c597bfe4ada0..25d7280e9f6b 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -605,7 +605,7 @@ xfs_qm_dqread( dqp->q_res_rtbcount = be64_to_cpu(ddqp->d_rtbcount); /* Mark the buf so that this will stay incore a little longer */ - XFS_BUF_SET_VTYPE_REF(bp, B_FS_DQUOT, XFS_DQUOT_REF); + xfs_buf_set_ref(bp, XFS_DQUOT_REF); /* * We got the buffer with a xfs_trans_read_buf() (in dqtobp()) diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 207e0b0c0730..169380e66057 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -1505,7 +1505,7 @@ xfs_read_agi( return XFS_ERROR(EFSCORRUPTED); } - XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_AGI, XFS_AGI_REF); + xfs_buf_set_ref(*bpp, XFS_AGI_REF); xfs_check_agi_unlinked(agi); return 0; diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index b676494a55ca..21cec6cdf453 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -190,12 +190,6 @@ xfs_imap_to_bp( } xfs_inobp_check(mp, bp); - - /* - * Mark the buffer as an inode buffer now that it looks good - */ - XFS_BUF_SET_VTYPE(bp, B_FS_INO); - *bpp = bp; return 0; } From c867cb61641751fd3d86350232d64ae2a10137d4 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:46 +0000 Subject: [PATCH 60/69] xfs: remove XFS_BUF_STALE and XFS_BUF_SUPER_STALE Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 2 +- fs/xfs/xfs_buf.c | 4 ++-- fs/xfs/xfs_buf.h | 6 ------ fs/xfs/xfs_buf_item.c | 6 ++++-- fs/xfs/xfs_inode.c | 4 ++-- fs/xfs/xfs_log.c | 4 ++-- fs/xfs/xfs_log_recover.c | 2 +- fs/xfs/xfs_rw.c | 2 +- fs/xfs/xfs_trans_buf.c | 13 +++++++++---- 9 files changed, 22 insertions(+), 21 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index 981a0624f382..ae8f917490d4 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2168,7 +2168,7 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) */ bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt, XBF_TRYLOCK); if (bp) { - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_buf_delwri_dequeue(bp); xfs_buf_relse(bp); bp = NULL; diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 36fed03da26f..f88eab9e8144 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1061,7 +1061,7 @@ xfs_bioerror( XFS_BUF_UNREAD(bp); xfs_buf_delwri_dequeue(bp); XFS_BUF_UNDONE(bp); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_buf_ioend(bp, 0); @@ -1090,7 +1090,7 @@ xfs_bioerror_relse( XFS_BUF_UNREAD(bp); xfs_buf_delwri_dequeue(bp); XFS_BUF_DONE(bp); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); bp->b_iodone = NULL; if (!(fl & XBF_ASYNC)) { /* diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index ca2934717343..fa38401449d9 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -242,14 +242,8 @@ xfs_buf_target_name(struct xfs_buftarg *target) XBF_SYNCIO|XBF_FUA|XBF_FLUSH)) void xfs_buf_stale(struct xfs_buf *bp); -#define XFS_BUF_STALE(bp) xfs_buf_stale(bp); #define XFS_BUF_UNSTALE(bp) ((bp)->b_flags &= ~XBF_STALE) #define XFS_BUF_ISSTALE(bp) ((bp)->b_flags & XBF_STALE) -#define XFS_BUF_SUPER_STALE(bp) do { \ - XFS_BUF_STALE(bp); \ - xfs_buf_delwri_dequeue(bp); \ - XFS_BUF_DONE(bp); \ - } while (0) #define XFS_BUF_ISDELAYWRITE(bp) ((bp)->b_flags & XBF_DELWRI) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 3243083d8693..8213f4a753dc 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -966,7 +966,9 @@ xfs_buf_iodone_callbacks( * I/O errors, there's no point in giving this a retry. */ if (XFS_FORCED_SHUTDOWN(mp)) { - XFS_BUF_SUPER_STALE(bp); + xfs_buf_stale(bp); + xfs_buf_delwri_dequeue(bp); + XFS_BUF_DONE(bp); trace_xfs_buf_item_iodone(bp, _RET_IP_); goto do_callbacks; } @@ -1005,7 +1007,7 @@ xfs_buf_iodone_callbacks( * If the write of the buffer was synchronous, we want to make * sure to return the error to the caller of xfs_bwrite(). */ - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); XFS_BUF_DONE(bp); xfs_buf_delwri_dequeue(bp); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 21cec6cdf453..c0237c602f11 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2469,11 +2469,11 @@ xfs_iflush_cluster( */ if (bp->b_iodone) { XFS_BUF_UNDONE(bp); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_buf_ioerror(bp, EIO); xfs_buf_ioend(bp, 0); } else { - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_buf_relse(bp); } } diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 3a8d4f66d702..493447dc4f55 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -881,7 +881,7 @@ xlog_iodone(xfs_buf_t *bp) if (XFS_TEST_ERROR((xfs_buf_geterror(bp)), l->l_mp, XFS_ERRTAG_IODONE_IOERR, XFS_RANDOM_IODONE_IOERR)) { xfs_ioerror_alert("xlog_iodone", l->l_mp, bp, XFS_BUF_ADDR(bp)); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_force_shutdown(l->l_mp, SHUTDOWN_LOG_IO_ERROR); /* * This flag will be propagated to the trans-committed @@ -1247,7 +1247,7 @@ xlog_bdstrat( if (iclog->ic_state & XLOG_STATE_IOERROR) { xfs_buf_ioerror(bp, EIO); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_buf_ioend(bp, 0); /* * It would seem logical to return EIO here, but we rely on diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index be173852b2ca..8f70f3469997 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2174,7 +2174,7 @@ xlog_recover_buffer_pass2( be16_to_cpu(*((__be16 *)xfs_buf_offset(bp, 0))) && (XFS_BUF_COUNT(bp) != MAX(log->l_mp->m_sb.sb_blocksize, (__uint32_t)XFS_INODE_CLUSTER_SIZE(log->l_mp)))) { - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); error = xfs_bwrite(bp); } else { ASSERT(bp->b_target->bt_mount == mp); diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index 99823c3b9aca..ff33645fe62d 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -150,7 +150,7 @@ xfs_read_buf( if (bp) { XFS_BUF_UNDONE(bp); xfs_buf_delwri_dequeue(bp); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); /* * brelse clears B_ERROR and b_error */ diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 5e5196a269dd..d03a8ee19172 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -160,8 +160,11 @@ xfs_trans_get_buf(xfs_trans_t *tp, bp = xfs_trans_buf_item_match(tp, target_dev, blkno, len); if (bp != NULL) { ASSERT(xfs_buf_islocked(bp)); - if (XFS_FORCED_SHUTDOWN(tp->t_mountp)) - XFS_BUF_SUPER_STALE(bp); + if (XFS_FORCED_SHUTDOWN(tp->t_mountp)) { + xfs_buf_stale(bp); + xfs_buf_delwri_dequeue(bp); + XFS_BUF_DONE(bp); + } /* * If the buffer is stale then it was binval'ed @@ -387,7 +390,9 @@ xfs_trans_read_buf( } if (bp->b_error) { error = bp->b_error; - XFS_BUF_SUPER_STALE(bp); + xfs_buf_stale(bp); + xfs_buf_delwri_dequeue(bp); + XFS_BUF_DONE(bp); xfs_ioerror_alert("xfs_trans_read_buf", mp, bp, blkno); if (tp->t_flags & XFS_TRANS_DIRTY) @@ -740,7 +745,7 @@ xfs_trans_binval( * rid of it. */ xfs_buf_delwri_dequeue(bp); - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); bip->bli_flags |= XFS_BLI_STALE; bip->bli_flags &= ~(XFS_BLI_INODE_BUF | XFS_BLI_LOGGED | XFS_BLI_DIRTY); bip->bli_format.blf_flags &= ~XFS_BLF_INODE_BUF; From af5c4bee499eb68bc36ca046030394d82d0e3669 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:47 +0000 Subject: [PATCH 61/69] xfs: remove buffers from the delwri list in xfs_buf_stale For each call to xfs_buf_stale we call xfs_buf_delwri_dequeue either directly before or after it, or are guaranteed by the surrounding conditionals that we are never called on delwri buffers. Simply this situation by moving the call to xfs_buf_delwri_dequeue into xfs_buf_stale. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_attr.c | 1 - fs/xfs/xfs_buf.c | 3 +-- fs/xfs/xfs_buf_item.c | 2 -- fs/xfs/xfs_rw.c | 1 - fs/xfs/xfs_trans_buf.c | 3 --- 5 files changed, 1 insertion(+), 9 deletions(-) diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index ae8f917490d4..1e5d97f86ea8 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2169,7 +2169,6 @@ xfs_attr_rmtval_remove(xfs_da_args_t *args) bp = xfs_incore(mp->m_ddev_targp, dblkno, blkcnt, XBF_TRYLOCK); if (bp) { xfs_buf_stale(bp); - xfs_buf_delwri_dequeue(bp); xfs_buf_relse(bp); bp = NULL; } diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index f88eab9e8144..3df7d0a2b245 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -151,6 +151,7 @@ xfs_buf_stale( struct xfs_buf *bp) { bp->b_flags |= XBF_STALE; + xfs_buf_delwri_dequeue(bp); atomic_set(&(bp)->b_lru_ref, 0); if (!list_empty(&bp->b_lru)) { struct xfs_buftarg *btp = bp->b_target; @@ -1059,7 +1060,6 @@ xfs_bioerror( * We're calling xfs_buf_ioend, so delete XBF_DONE flag. */ XFS_BUF_UNREAD(bp); - xfs_buf_delwri_dequeue(bp); XFS_BUF_UNDONE(bp); xfs_buf_stale(bp); @@ -1088,7 +1088,6 @@ xfs_bioerror_relse( * change that interface. */ XFS_BUF_UNREAD(bp); - xfs_buf_delwri_dequeue(bp); XFS_BUF_DONE(bp); xfs_buf_stale(bp); bp->b_iodone = NULL; diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 8213f4a753dc..06a76ca99659 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -967,7 +967,6 @@ xfs_buf_iodone_callbacks( */ if (XFS_FORCED_SHUTDOWN(mp)) { xfs_buf_stale(bp); - xfs_buf_delwri_dequeue(bp); XFS_BUF_DONE(bp); trace_xfs_buf_item_iodone(bp, _RET_IP_); goto do_callbacks; @@ -1009,7 +1008,6 @@ xfs_buf_iodone_callbacks( */ xfs_buf_stale(bp); XFS_BUF_DONE(bp); - xfs_buf_delwri_dequeue(bp); trace_xfs_buf_error_relse(bp, _RET_IP_); diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index ff33645fe62d..86f1928b4cfa 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -149,7 +149,6 @@ xfs_read_buf( } if (bp) { XFS_BUF_UNDONE(bp); - xfs_buf_delwri_dequeue(bp); xfs_buf_stale(bp); /* * brelse clears B_ERROR and b_error diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index d03a8ee19172..5bab5980a6f9 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -162,7 +162,6 @@ xfs_trans_get_buf(xfs_trans_t *tp, ASSERT(xfs_buf_islocked(bp)); if (XFS_FORCED_SHUTDOWN(tp->t_mountp)) { xfs_buf_stale(bp); - xfs_buf_delwri_dequeue(bp); XFS_BUF_DONE(bp); } @@ -391,7 +390,6 @@ xfs_trans_read_buf( if (bp->b_error) { error = bp->b_error; xfs_buf_stale(bp); - xfs_buf_delwri_dequeue(bp); XFS_BUF_DONE(bp); xfs_ioerror_alert("xfs_trans_read_buf", mp, bp, blkno); @@ -744,7 +742,6 @@ xfs_trans_binval( * We set the stale bit in the buffer as well since we're getting * rid of it. */ - xfs_buf_delwri_dequeue(bp); xfs_buf_stale(bp); bip->bli_flags |= XFS_BLI_STALE; bip->bli_flags &= ~(XFS_BLI_INODE_BUF | XFS_BLI_LOGGED | XFS_BLI_DIRTY); From 4347b9d7ad4223474d315c3ab6bc1ce7cce7fa2d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:48 +0000 Subject: [PATCH 62/69] xfs: clean up buffer allocation Change _xfs_buf_initialize to allocate the buffer directly and rename it to xfs_buf_alloc now that is the only buffer allocation routine. Also remove the xfs_buf_deallocate wrapper around the kmem_zone_free calls for buffers. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 50 +++++++++++++++++------------------------------- fs/xfs/xfs_buf.h | 3 ++- fs/xfs/xfs_log.c | 2 +- 3 files changed, 21 insertions(+), 34 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 3df7d0a2b245..1f24ee5f0d7a 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -65,10 +65,6 @@ struct workqueue_struct *xfsconvertd_workqueue; #define xb_to_km(flags) \ (((flags) & XBF_DONT_BLOCK) ? KM_NOFS : KM_SLEEP) -#define xfs_buf_allocate(flags) \ - kmem_zone_alloc(xfs_buf_zone, xb_to_km(flags)) -#define xfs_buf_deallocate(bp) \ - kmem_zone_free(xfs_buf_zone, (bp)); static inline int xfs_buf_is_vmapped( @@ -167,14 +163,19 @@ xfs_buf_stale( ASSERT(atomic_read(&bp->b_hold) >= 1); } -STATIC void -_xfs_buf_initialize( - xfs_buf_t *bp, - xfs_buftarg_t *target, +struct xfs_buf * +xfs_buf_alloc( + struct xfs_buftarg *target, xfs_off_t range_base, size_t range_length, xfs_buf_flags_t flags) { + struct xfs_buf *bp; + + bp = kmem_zone_alloc(xfs_buf_zone, xb_to_km(flags)); + if (unlikely(!bp)) + return NULL; + /* * We don't want certain flags to appear in b_flags. */ @@ -203,8 +204,9 @@ _xfs_buf_initialize( init_waitqueue_head(&bp->b_waiters); XFS_STATS_INC(xb_create); - trace_xfs_buf_init(bp, _RET_IP_); + + return bp; } /* @@ -277,7 +279,7 @@ xfs_buf_free( } else if (bp->b_flags & _XBF_KMEM) kmem_free(bp->b_addr); _xfs_buf_free_pages(bp); - xfs_buf_deallocate(bp); + kmem_zone_free(xfs_buf_zone, bp); } /* @@ -539,16 +541,14 @@ xfs_buf_get( if (likely(bp)) goto found; - new_bp = xfs_buf_allocate(flags); + new_bp = xfs_buf_alloc(target, ioff << BBSHIFT, isize << BBSHIFT, + flags); if (unlikely(!new_bp)) return NULL; - _xfs_buf_initialize(new_bp, target, - ioff << BBSHIFT, isize << BBSHIFT, flags); - bp = _xfs_buf_find(target, ioff, isize, flags, new_bp); if (!bp) { - xfs_buf_deallocate(new_bp); + kmem_zone_free(xfs_buf_zone, new_bp); return NULL; } @@ -557,7 +557,7 @@ xfs_buf_get( if (error) goto no_buffer; } else - xfs_buf_deallocate(new_bp); + kmem_zone_free(xfs_buf_zone, new_bp); /* * Now we have a workable buffer, fill in the block number so @@ -694,19 +694,6 @@ xfs_buf_read_uncached( return bp; } -xfs_buf_t * -xfs_buf_get_empty( - size_t len, - xfs_buftarg_t *target) -{ - xfs_buf_t *bp; - - bp = xfs_buf_allocate(0); - if (bp) - _xfs_buf_initialize(bp, target, 0, len, 0); - return bp; -} - /* * Return a buffer allocated as an empty buffer and associated to external * memory via xfs_buf_associate_memory() back to it's empty state. @@ -792,10 +779,9 @@ xfs_buf_get_uncached( int error, i; xfs_buf_t *bp; - bp = xfs_buf_allocate(0); + bp = xfs_buf_alloc(target, 0, len, 0); if (unlikely(bp == NULL)) goto fail; - _xfs_buf_initialize(bp, target, 0, len, 0); error = _xfs_buf_get_pages(bp, page_count, 0); if (error) @@ -823,7 +809,7 @@ xfs_buf_get_uncached( __free_page(bp->b_pages[i]); _xfs_buf_free_pages(bp); fail_free_buf: - xfs_buf_deallocate(bp); + kmem_zone_free(xfs_buf_zone, bp); fail: return NULL; } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index fa38401449d9..26b909417deb 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -175,7 +175,8 @@ extern xfs_buf_t *xfs_buf_get(xfs_buftarg_t *, xfs_off_t, size_t, extern xfs_buf_t *xfs_buf_read(xfs_buftarg_t *, xfs_off_t, size_t, xfs_buf_flags_t); -extern xfs_buf_t *xfs_buf_get_empty(size_t, xfs_buftarg_t *); +struct xfs_buf *xfs_buf_alloc(struct xfs_buftarg *, xfs_off_t, size_t, + xfs_buf_flags_t); extern void xfs_buf_set_empty(struct xfs_buf *bp, size_t len); extern xfs_buf_t *xfs_buf_get_uncached(struct xfs_buftarg *, size_t, int); extern int xfs_buf_associate_memory(xfs_buf_t *, void *, size_t); diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 493447dc4f55..8c9db4e5ddd2 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -1047,7 +1047,7 @@ xlog_alloc_log(xfs_mount_t *mp, xlog_get_iclog_buffer_size(mp, log); error = ENOMEM; - bp = xfs_buf_get_empty(log->l_iclog_size, mp->m_logdev_targp); + bp = xfs_buf_alloc(mp->m_logdev_targp, 0, log->l_iclog_size, 0); if (!bp) goto out_free_log; bp->b_iodone = xlog_iodone; From 901796afca0d31d97bf6d1bf2ab251a93a4b8c83 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:49 +0000 Subject: [PATCH 63/69] xfs: clean up xfs_ioerror_alert Instead of passing the block number and mount structure explicitly get them off the bp and fix make the argument order more natural. Also move it to xfs_buf.c and stop printing the device name given that we already get the fs name as part of xfs_alert, and we know what device is operates on because of the caller that gets printed, finally rename it to xfs_buf_ioerror_alert and pass __func__ as argument where it makes sense. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 11 +++++++++++ fs/xfs/xfs_buf.h | 1 + fs/xfs/xfs_log.c | 14 +++++++------- fs/xfs/xfs_log_recover.c | 25 ++++++++----------------- fs/xfs/xfs_mount.c | 3 +-- fs/xfs/xfs_rw.c | 20 +------------------- fs/xfs/xfs_rw.h | 2 -- fs/xfs/xfs_trans_buf.c | 9 +++------ fs/xfs/xfs_vnodeops.c | 11 +++++------ 9 files changed, 37 insertions(+), 59 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 1f24ee5f0d7a..0a767fca0305 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1004,6 +1004,17 @@ xfs_buf_ioerror( trace_xfs_buf_ioerror(bp, error, _RET_IP_); } +void +xfs_buf_ioerror_alert( + struct xfs_buf *bp, + const char *func) +{ + xfs_alert(bp->b_target->bt_mount, +"metadata I/O error: block 0x%llx (\"%s\") error %d buf count %zd", + (__uint64_t)XFS_BUF_ADDR(bp), func, + bp->b_error, XFS_BUF_COUNT(bp)); +} + int xfs_bwrite( struct xfs_buf *bp) diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 26b909417deb..357a3371cae7 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -205,6 +205,7 @@ extern int xfs_bdstrat_cb(struct xfs_buf *); extern void xfs_buf_ioend(xfs_buf_t *, int); extern void xfs_buf_ioerror(xfs_buf_t *, int); +extern void xfs_buf_ioerror_alert(struct xfs_buf *, const char *func); extern int xfs_buf_iorequest(xfs_buf_t *); extern int xfs_buf_iowait(xfs_buf_t *); extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *, diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 8c9db4e5ddd2..2758a6277c52 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -880,7 +880,7 @@ xlog_iodone(xfs_buf_t *bp) */ if (XFS_TEST_ERROR((xfs_buf_geterror(bp)), l->l_mp, XFS_ERRTAG_IODONE_IOERR, XFS_RANDOM_IODONE_IOERR)) { - xfs_ioerror_alert("xlog_iodone", l->l_mp, bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); xfs_buf_stale(bp); xfs_force_shutdown(l->l_mp, SHUTDOWN_LOG_IO_ERROR); /* @@ -1387,9 +1387,9 @@ xlog_sync(xlog_t *log, */ XFS_BUF_WRITE(bp); - if ((error = xlog_bdstrat(bp))) { - xfs_ioerror_alert("xlog_sync", log->l_mp, bp, - XFS_BUF_ADDR(bp)); + error = xlog_bdstrat(bp); + if (error) { + xfs_buf_ioerror_alert(bp, "xlog_sync"); return error; } if (split) { @@ -1423,9 +1423,9 @@ xlog_sync(xlog_t *log, /* account for internal log which doesn't start at block #0 */ XFS_BUF_SET_ADDR(bp, XFS_BUF_ADDR(bp) + log->l_logBBstart); XFS_BUF_WRITE(bp); - if ((error = xlog_bdstrat(bp))) { - xfs_ioerror_alert("xlog_sync (split)", log->l_mp, - bp, XFS_BUF_ADDR(bp)); + error = xlog_bdstrat(bp); + if (error) { + xfs_buf_ioerror_alert(bp, "xlog_sync (split)"); return error; } } diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 8f70f3469997..82ee9db628ed 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -183,8 +183,7 @@ xlog_bread_noalign( xfsbdstrat(log->l_mp, bp); error = xfs_buf_iowait(bp); if (error) - xfs_ioerror_alert("xlog_bread", log->l_mp, - bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); return error; } @@ -269,10 +268,8 @@ xlog_bwrite( XFS_BUF_SET_COUNT(bp, BBTOB(nbblks)); error = xfs_bwrite(bp); - if (error) { - xfs_ioerror_alert("xlog_bwrite", log->l_mp, - bp, XFS_BUF_ADDR(bp)); - } + if (error) + xfs_buf_ioerror_alert(bp, __func__); xfs_buf_relse(bp); return error; } @@ -364,9 +361,7 @@ xlog_recover_iodone( * We're not going to bother about retrying * this during recovery. One strike! */ - xfs_ioerror_alert("xlog_recover_iodone", - bp->b_target->bt_mount, bp, - XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); xfs_force_shutdown(bp->b_target->bt_mount, SHUTDOWN_META_IO_ERROR); } @@ -2138,8 +2133,7 @@ xlog_recover_buffer_pass2( return XFS_ERROR(ENOMEM); error = bp->b_error; if (error) { - xfs_ioerror_alert("xlog_recover_do..(read#1)", mp, - bp, buf_f->blf_blkno); + xfs_buf_ioerror_alert(bp, "xlog_recover_do..(read#1)"); xfs_buf_relse(bp); return error; } @@ -2234,8 +2228,7 @@ xlog_recover_inode_pass2( } error = bp->b_error; if (error) { - xfs_ioerror_alert("xlog_recover_do..(read#2)", mp, - bp, in_f->ilf_blkno); + xfs_buf_ioerror_alert(bp, "xlog_recover_do..(read#2)"); xfs_buf_relse(bp); goto error; } @@ -2542,8 +2535,7 @@ xlog_recover_dquot_pass2( XFS_FSB_TO_BB(mp, dq_f->qlf_len), 0, &bp); if (error) { - xfs_ioerror_alert("xlog_recover_do..(read#3)", mp, - bp, dq_f->qlf_blkno); + xfs_buf_ioerror_alert(bp, "xlog_recover_do..(read#3)"); return error; } ASSERT(bp); @@ -3695,8 +3687,7 @@ xlog_do_recover( xfsbdstrat(log->l_mp, bp); error = xfs_buf_iowait(bp); if (error) { - xfs_ioerror_alert("xlog_do_recover", - log->l_mp, bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); ASSERT(0); xfs_buf_relse(bp); return error; diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index e8fe1cb54094..f3b1cec38b81 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1610,8 +1610,7 @@ xfs_unmountfs_writesb(xfs_mount_t *mp) xfsbdstrat(mp, sbp); error = xfs_buf_iowait(sbp); if (error) - xfs_ioerror_alert("xfs_unmountfs_writesb", - mp, sbp, XFS_BUF_ADDR(sbp)); + xfs_buf_ioerror_alert(sbp, __func__); xfs_buf_relse(sbp); } return error; diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index 86f1928b4cfa..597d044a09a1 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -91,24 +91,6 @@ xfs_do_force_shutdown( } } -/* - * Prints out an ALERT message about I/O error. - */ -void -xfs_ioerror_alert( - char *func, - struct xfs_mount *mp, - xfs_buf_t *bp, - xfs_daddr_t blkno) -{ - xfs_alert(mp, - "I/O error occurred: meta-data dev %s block 0x%llx" - " (\"%s\") error %d buf count %zd", - xfs_buf_target_name(bp->b_target), - (__uint64_t)blkno, func, - bp->b_error, XFS_BUF_COUNT(bp)); -} - /* * This isn't an absolute requirement, but it is * just a good idea to call xfs_read_buf instead of @@ -143,7 +125,7 @@ xfs_read_buf( } else { *bpp = NULL; if (error) { - xfs_ioerror_alert("xfs_read_buf", mp, bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); } else { error = XFS_ERROR(EIO); } diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h index 11c41ec6ed75..bbdb9ad6a4ba 100644 --- a/fs/xfs/xfs_rw.h +++ b/fs/xfs/xfs_rw.h @@ -42,8 +42,6 @@ xfs_fsb_to_db(struct xfs_inode *ip, xfs_fsblock_t fsb) extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp, xfs_daddr_t blkno, int len, uint flags, struct xfs_buf **bpp); -extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp, - xfs_buf_t *bp, xfs_daddr_t blkno); extern xfs_extlen_t xfs_get_extsz_hint(struct xfs_inode *ip); #endif /* __XFS_RW_H__ */ diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 5bab5980a6f9..475a4ded4f41 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -296,8 +296,7 @@ xfs_trans_read_buf( if (bp->b_error) { error = bp->b_error; - xfs_ioerror_alert("xfs_trans_read_buf", mp, - bp, blkno); + xfs_buf_ioerror_alert(bp, __func__); xfs_buf_relse(bp); return error; } @@ -339,8 +338,7 @@ xfs_trans_read_buf( xfsbdstrat(tp->t_mountp, bp); error = xfs_buf_iowait(bp); if (error) { - xfs_ioerror_alert("xfs_trans_read_buf", mp, - bp, blkno); + xfs_buf_ioerror_alert(bp, __func__); xfs_buf_relse(bp); /* * We can gracefully recover from most read @@ -391,8 +389,7 @@ xfs_trans_read_buf( error = bp->b_error; xfs_buf_stale(bp); XFS_BUF_DONE(bp); - xfs_ioerror_alert("xfs_trans_read_buf", mp, - bp, blkno); + xfs_buf_ioerror_alert(bp, __func__); if (tp->t_flags & XFS_TRANS_DIRTY) xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR); xfs_buf_relse(bp); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index fc38f15808c6..4ecf2a549060 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -87,8 +87,7 @@ xfs_readlink_bmap( return XFS_ERROR(ENOMEM); error = bp->b_error; if (error) { - xfs_ioerror_alert("xfs_readlink", - ip->i_mount, bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); xfs_buf_relse(bp); goto out; } @@ -1993,8 +1992,8 @@ xfs_zero_remaining_bytes( xfsbdstrat(mp, bp); error = xfs_buf_iowait(bp); if (error) { - xfs_ioerror_alert("xfs_zero_remaining_bytes(read)", - mp, bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, + "xfs_zero_remaining_bytes(read)"); break; } memset(bp->b_addr + @@ -2006,8 +2005,8 @@ xfs_zero_remaining_bytes( xfsbdstrat(mp, bp); error = xfs_buf_iowait(bp); if (error) { - xfs_ioerror_alert("xfs_zero_remaining_bytes(write)", - mp, bp, XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, + "xfs_zero_remaining_bytes(write)"); break; } } From b38505b09b7854d446b2f60b4414e3231277aa1a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:50 +0000 Subject: [PATCH 64/69] xfs: use xfs_ioerror_alert in xfs_buf_iodone_callbacks Use xfs_ioerror_alert instead of opencoding a very similar error message. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf_item.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 06a76ca99659..65d6f4432d28 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -975,9 +975,7 @@ xfs_buf_iodone_callbacks( if (bp->b_target != lasttarg || time_after(jiffies, (lasttime + 5*HZ))) { lasttime = jiffies; - xfs_alert(mp, "Device %s: metadata write error block 0x%llx", - xfs_buf_target_name(bp->b_target), - (__uint64_t)XFS_BUF_ADDR(bp)); + xfs_buf_ioerror_alert(bp, __func__); } lasttarg = bp->b_target; From 02b102df1502a7ea4167d115510e1e8fe6467f12 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:51 +0000 Subject: [PATCH 65/69] xfs: remove xfs_buf_target_name The calling convention that returns a pointer to a static buffer is fairly nasty, so just opencode it in the only caller that is left. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 6 +++++- fs/xfs/xfs_buf.h | 9 --------- 2 files changed, 5 insertions(+), 10 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 0a767fca0305..6f615c259411 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1457,9 +1457,13 @@ xfs_setsize_buftarg_flags( btp->bt_smask = sectorsize - 1; if (set_blocksize(btp->bt_bdev, sectorsize)) { + char name[BDEVNAME_SIZE]; + + bdevname(btp->bt_bdev, name); + xfs_warn(btp->bt_mount, "Cannot set_blocksize to %u on device %s\n", - sectorsize, xfs_buf_target_name(btp)); + sectorsize, name); return EINVAL; } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 357a3371cae7..be19dd2b0212 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -230,15 +230,6 @@ extern void xfs_buf_delwri_promote(struct xfs_buf *); extern int xfs_buf_init(void); extern void xfs_buf_terminate(void); -static inline const char * -xfs_buf_target_name(struct xfs_buftarg *target) -{ - static char __b[BDEVNAME_SIZE]; - - return bdevname(target->bt_bdev, __b); -} - - #define XFS_BUF_ZEROFLAGS(bp) \ ((bp)->b_flags &= ~(XBF_READ|XBF_WRITE|XBF_ASYNC|XBF_DELWRI| \ XBF_SYNCIO|XBF_FUA|XBF_FLUSH)) From a9add83e5abd29bf2b7b3658311199eeabbdefc6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:52 +0000 Subject: [PATCH 66/69] xfs: remove XFS_bflush Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.h | 2 -- fs/xfs/xfs_log_recover.c | 2 +- fs/xfs/xfs_mount.c | 2 +- fs/xfs/xfs_qm.c | 2 +- fs/xfs/xfs_super.c | 4 ++-- fs/xfs/xfs_sync.c | 2 +- 6 files changed, 6 insertions(+), 8 deletions(-) diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index be19dd2b0212..5bab046e859f 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -294,6 +294,4 @@ extern int xfs_flush_buftarg(xfs_buftarg_t *, int); #define xfs_getsize_buftarg(buftarg) block_size((buftarg)->bt_bdev) #define xfs_readonly_buftarg(buftarg) bdev_read_only((buftarg)->bt_bdev) -#define XFS_bflush(buftarg) xfs_flush_buftarg(buftarg, 1) - #endif /* __XFS_BUF_H__ */ diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 82ee9db628ed..541a508adea1 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3654,7 +3654,7 @@ xlog_do_recover( return error; } - XFS_bflush(log->l_mp->m_ddev_targp); + xfs_flush_buftarg(log->l_mp->m_ddev_targp, 1); /* * If IO errors happened during recovery, bail out. diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index f3b1cec38b81..d06afbc3540d 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1481,7 +1481,7 @@ xfs_unmountfs( * state as much as possible. */ xfs_reclaim_inodes(mp, 0); - XFS_bflush(mp->m_ddev_targp); + xfs_flush_buftarg(mp->m_ddev_targp, 1); xfs_reclaim_inodes(mp, SYNC_WAIT); xfs_qm_unmount(mp); diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index ddaf97a57ec6..5cff443f6cdb 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1681,7 +1681,7 @@ xfs_qm_quotacheck( * quotacheck'd stamp on the superblock. So, here we do a synchronous * flush. */ - XFS_bflush(mp->m_ddev_targp); + xfs_flush_buftarg(mp->m_ddev_targp, 1); /* * If one type of quotas is off, then it will lose its diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 6ad05e68abda..ba16248bcf24 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1015,7 +1015,7 @@ xfs_fs_put_super( */ xfs_filestream_unmount(mp); - XFS_bflush(mp->m_ddev_targp); + xfs_flush_buftarg(mp->m_ddev_targp, 1); xfs_unmountfs(mp); xfs_freesb(mp); @@ -1439,7 +1439,7 @@ xfs_fs_fill_super( */ xfs_filestream_unmount(mp); - XFS_bflush(mp->m_ddev_targp); + xfs_flush_buftarg(mp->m_ddev_targp, 1); xfs_unmountfs(mp); goto out_free_sb; diff --git a/fs/xfs/xfs_sync.c b/fs/xfs/xfs_sync.c index bf2b38c21caa..aa3dc1a4d53d 100644 --- a/fs/xfs/xfs_sync.c +++ b/fs/xfs/xfs_sync.c @@ -377,7 +377,7 @@ xfs_quiesce_data( /* flush data-only devices */ if (mp->m_rtdev_targp) - XFS_bflush(mp->m_rtdev_targp); + xfs_flush_buftarg(mp->m_rtdev_targp, 1); return error ? error : error2; } From 5a93a064d27b42e4af1772b0599b53e3241191ac Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 10 Oct 2011 16:52:53 +0000 Subject: [PATCH 67/69] xfs: do not flush data workqueues in xfs_flush_buftarg When we call xfs_flush_buftarg (generally from sync or umount) it already is too late to flush the data workqueues, as I/O completion is signalled for them and we are thus already done with the data we would flush here. There are places where flushing them might be useful, but the current sync interface doesn't give us that opportunity. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_buf.c | 11 +---------- 1 file changed, 1 insertion(+), 10 deletions(-) diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 6f615c259411..cf0ac056815f 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1623,13 +1623,6 @@ xfs_buf_delwri_promote( spin_unlock(&btp->bt_delwri_lock); } -STATIC void -xfs_buf_runall_queues( - struct workqueue_struct *queue) -{ - flush_workqueue(queue); -} - /* * Move as many buffers as specified to the supplied list * idicating if we skipped any buffers to prevent deadlocks. @@ -1752,9 +1745,7 @@ xfs_flush_buftarg( LIST_HEAD(wait_list); struct blk_plug plug; - xfs_buf_runall_queues(xfsconvertd_workqueue); - xfs_buf_runall_queues(xfsdatad_workqueue); - xfs_buf_runall_queues(xfslogd_workqueue); + flush_workqueue(xfslogd_workqueue); set_bit(XBT_FORCE_FLUSH, &target->bt_flags); pincount = xfs_buf_delwri_split(target, &tmp_list, 0); From 2900b33999e2fc8a8edf0dddaafffec4da25ee10 Mon Sep 17 00:00:00 2001 From: Alex Elder Date: Tue, 18 Oct 2011 20:00:14 +0000 Subject: [PATCH 68/69] xfs: put in missed fix for merge problem I intended to do this as part of fixing part of the conflict with the merge with Linus' tree, but evidently it didn't get included in the commit. Signed-off-by: Alex Elder Reviewed-by: Christoph Hellwig --- fs/xfs/xfs_trans_ail.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 512ff646d01c..4e3f9bbe0141 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -434,7 +434,7 @@ xfsaild_push( if (!IOP_PUSHBUF(lip)) { stuck++; - flush_log = 1; + ailp->xa_log_flush++; } else { ailp->xa_last_pushed_lsn = lsn; } From 9e4c109ac822395e0aae650e4e3c9e4903f6602f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 11 Oct 2011 15:14:11 +0000 Subject: [PATCH 69/69] xfs: add AIL pushing tracepoints Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder --- fs/xfs/xfs_trace.h | 37 +++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_trans_ail.c | 8 ++++++++ 2 files changed, 45 insertions(+) diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index b78f2c65438b..f1d2802b2f07 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -30,6 +30,7 @@ struct xfs_buf_log_item; struct xfs_da_args; struct xfs_da_node_entry; struct xfs_dquot; +struct xfs_log_item; struct xlog_ticket; struct log; struct xlog_recover; @@ -853,6 +854,42 @@ DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_enter); DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_exit); DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_sub); +DECLARE_EVENT_CLASS(xfs_log_item_class, + TP_PROTO(struct xfs_log_item *lip), + TP_ARGS(lip), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(void *, lip) + __field(uint, type) + __field(uint, flags) + __field(xfs_lsn_t, lsn) + ), + TP_fast_assign( + __entry->dev = lip->li_mountp->m_super->s_dev; + __entry->lip = lip; + __entry->type = lip->li_type; + __entry->flags = lip->li_flags; + __entry->lsn = lip->li_lsn; + ), + TP_printk("dev %d:%d lip 0x%p lsn %d/%d type %s flags %s", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->lip, + CYCLE_LSN(__entry->lsn), BLOCK_LSN(__entry->lsn), + __print_symbolic(__entry->type, XFS_LI_TYPE_DESC), + __print_flags(__entry->flags, "|", XFS_LI_FLAGS)) +) + +#define DEFINE_LOG_ITEM_EVENT(name) \ +DEFINE_EVENT(xfs_log_item_class, name, \ + TP_PROTO(struct xfs_log_item *lip), \ + TP_ARGS(lip)) +DEFINE_LOG_ITEM_EVENT(xfs_ail_push); +DEFINE_LOG_ITEM_EVENT(xfs_ail_pushbuf); +DEFINE_LOG_ITEM_EVENT(xfs_ail_pushbuf_pinned); +DEFINE_LOG_ITEM_EVENT(xfs_ail_pinned); +DEFINE_LOG_ITEM_EVENT(xfs_ail_locked); + + DECLARE_EVENT_CLASS(xfs_file_class, TP_PROTO(struct xfs_inode *ip, size_t count, loff_t offset, int flags), TP_ARGS(ip, count, offset, flags), diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 4e3f9bbe0141..ed9252bcdac9 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -26,6 +26,7 @@ #include "xfs_ag.h" #include "xfs_mount.h" #include "xfs_trans_priv.h" +#include "xfs_trace.h" #include "xfs_error.h" #ifdef DEBUG @@ -425,14 +426,18 @@ xfsaild_push( switch (lock_result) { case XFS_ITEM_SUCCESS: XFS_STATS_INC(xs_push_ail_success); + trace_xfs_ail_push(lip); + IOP_PUSH(lip); ailp->xa_last_pushed_lsn = lsn; break; case XFS_ITEM_PUSHBUF: XFS_STATS_INC(xs_push_ail_pushbuf); + trace_xfs_ail_pushbuf(lip); if (!IOP_PUSHBUF(lip)) { + trace_xfs_ail_pushbuf_pinned(lip); stuck++; ailp->xa_log_flush++; } else { @@ -443,12 +448,15 @@ xfsaild_push( case XFS_ITEM_PINNED: XFS_STATS_INC(xs_push_ail_pinned); + trace_xfs_ail_pinned(lip); + stuck++; ailp->xa_log_flush++; break; case XFS_ITEM_LOCKED: XFS_STATS_INC(xs_push_ail_locked); + trace_xfs_ail_locked(lip); stuck++; break;