// SPDX-License-Identifier: GPL-2.0 /* * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc. * All Rights Reserved. */ #include "xfs.h" #include "xfs_fs.h" #include "xfs_shared.h" #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" #include "xfs_inode.h" #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_trans_priv.h" #include "xfs_error.h" #include "xfs_trace.h" /* * Check to see if a buffer matching the given parameters is already * a part of the given transaction. */ STATIC struct xfs_buf * xfs_trans_buf_item_match( struct xfs_trans *tp, struct xfs_buftarg *target, struct xfs_buf_map *map, int nmaps) { struct xfs_log_item *lip; struct xfs_buf_log_item *blip; int len = 0; int i; for (i = 0; i < nmaps; i++) len += map[i].bm_len; list_for_each_entry(lip, &tp->t_items, li_trans) { blip = (struct xfs_buf_log_item *)lip; if (blip->bli_item.li_type == XFS_LI_BUF && blip->bli_buf->b_target == target && XFS_BUF_ADDR(blip->bli_buf) == map[0].bm_bn && blip->bli_buf->b_length == len) { ASSERT(blip->bli_buf->b_map_count == nmaps); return blip->bli_buf; } } return NULL; } /* * Add the locked buffer to the transaction. * * The buffer must be locked, and it cannot be associated with any * transaction. * * If the buffer does not yet have a buf log item associated with it, * then allocate one for it. Then add the buf item to the transaction. */ STATIC void _xfs_trans_bjoin( struct xfs_trans *tp, struct xfs_buf *bp, int reset_recur) { struct xfs_buf_log_item *bip; ASSERT(bp->b_transp == NULL); /* * The xfs_buf_log_item pointer is stored in b_log_item. If * it doesn't have one yet, then allocate one and initialize it. * The checks to see if one is there are in xfs_buf_item_init(). */ xfs_buf_item_init(bp, tp->t_mountp); bip = bp->b_log_item; ASSERT(!(bip->bli_flags & XFS_BLI_STALE)); ASSERT(!(bip->__bli_format.blf_flags & XFS_BLF_CANCEL)); ASSERT(!(bip->bli_flags & XFS_BLI_LOGGED)); if (reset_recur) bip->bli_recur = 0; /* * Take a reference for this transaction on the buf item. */ atomic_inc(&bip->bli_refcount); /* * Attach the item to the transaction so we can find it in * xfs_trans_get_buf() and friends. */ xfs_trans_add_item(tp, &bip->bli_item); bp->b_transp = tp; } void xfs_trans_bjoin( struct xfs_trans *tp, struct xfs_buf *bp) { _xfs_trans_bjoin(tp, bp, 0); trace_xfs_trans_bjoin(bp->b_log_item); } /* * Get and lock the buffer for the caller if it is not already * locked within the given transaction. If it is already locked * within the transaction, just increment its lock recursion count * and return a pointer to it. * * If the transaction pointer is NULL, make this just a normal * get_buf() call. */ struct xfs_buf * xfs_trans_get_buf_map( struct xfs_trans *tp, struct xfs_buftarg *target, struct xfs_buf_map *map, int nmaps, xfs_buf_flags_t flags) { xfs_buf_t *bp; struct xfs_buf_log_item *bip; if (!tp) return xfs_buf_get_map(target, map, nmaps, flags); /* * If we find the buffer in the cache with this transaction * pointer in its b_fsprivate2 field, then we know we already * have it locked. In this case we just increment the lock * recursion count and return the buffer to the caller. */ bp = xfs_trans_buf_item_match(tp, target, map, nmaps); if (bp != NULL) { ASSERT(xfs_buf_islocked(bp)); if (XFS_FORCED_SHUTDOWN(tp->t_mountp)) { xfs_buf_stale(bp); bp->b_flags |= XBF_DONE; } ASSERT(bp->b_transp == tp); bip = bp->b_log_item; ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); bip->bli_recur++; trace_xfs_trans_get_buf_recur(bip); return bp; } bp = xfs_buf_get_map(target, map, nmaps, flags); if (bp == NULL) { return NULL; } ASSERT(!bp->b_error); _xfs_trans_bjoin(tp, bp, 1); trace_xfs_trans_get_buf(bp->b_log_item); return bp; } /* * Get and lock the superblock buffer of this file system for the * given transaction. * * We don't need to use incore_match() here, because the superblock * buffer is a private buffer which we keep a pointer to in the * mount structure. */ xfs_buf_t * xfs_trans_getsb( xfs_trans_t *tp, struct xfs_mount *mp, int flags) { xfs_buf_t *bp; struct xfs_buf_log_item *bip; /* * Default to just trying to lock the superblock buffer * if tp is NULL. */ if (tp == NULL) return xfs_getsb(mp, flags); /* * If the superblock buffer already has this transaction * pointer in its b_fsprivate2 field, then we know we already * have it locked. In this case we just increment the lock * recursion count and return the buffer to the caller. */ bp = mp->m_sb_bp; if (bp->b_transp == tp) { bip = bp->b_log_item; ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); bip->bli_recur++; trace_xfs_trans_getsb_recur(bip); return bp; } bp = xfs_getsb(mp, flags); if (bp == NULL) return NULL; _xfs_trans_bjoin(tp, bp, 1); trace_xfs_trans_getsb(bp->b_log_item); return bp; } /* * Get and lock the buffer for the caller if it is not already * locked within the given transaction. If it has not yet been * read in, read it from disk. If it is already locked * within the transaction and already read in, just increment its * lock recursion count and return a pointer to it. * * If the transaction pointer is NULL, make this just a normal * read_buf() call. */ int xfs_trans_read_buf_map( struct xfs_mount *mp, struct xfs_trans *tp, struct xfs_buftarg *target, struct xfs_buf_map *map, int nmaps, xfs_buf_flags_t flags, struct xfs_buf **bpp, const struct xfs_buf_ops *ops) { struct xfs_buf *bp = NULL; struct xfs_buf_log_item *bip; int error; *bpp = NULL; /* * If we find the buffer in the cache with this transaction * pointer in its b_fsprivate2 field, then we know we already * have it locked. If it is already read in we just increment * the lock recursion count and return the buffer to the caller. * If the buffer is not yet read in, then we read it in, increment * the lock recursion count, and return it to the caller. */ if (tp) bp = xfs_trans_buf_item_match(tp, target, map, nmaps); if (bp) { ASSERT(xfs_buf_islocked(bp)); ASSERT(bp->b_transp == tp); ASSERT(bp->b_log_item != NULL); ASSERT(!bp->b_error); ASSERT(bp->b_flags & XBF_DONE); /* * We never locked this buf ourselves, so we shouldn't * brelse it either. Just get out. */ if (XFS_FORCED_SHUTDOWN(mp)) { trace_xfs_trans_read_buf_shut(bp, _RET_IP_); return -EIO; } /* * Check if the caller is trying to read a buffer that is * already attached to the transaction yet has no buffer ops * assigned. Ops are usually attached when the buffer is * attached to the transaction, or by the read caller if * special circumstances. That didn't happen, which is not * how this is supposed to go. * * If the buffer passes verification we'll let this go, but if * not we have to shut down. Let the transaction cleanup code * release this buffer when it kills the tranaction. */ ASSERT(bp->b_ops != NULL); error = xfs_buf_ensure_ops(bp, ops); if (error) { xfs_buf_ioerror_alert(bp, __func__); if (tp->t_flags & XFS_TRANS_DIRTY) xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR); /* bad CRC means corrupted metadata */ if (error == -EFSBADCRC) error = -EFSCORRUPTED; return error; } bip = bp->b_log_item; bip->bli_recur++; ASSERT(atomic_read(&bip->bli_refcount) > 0); trace_xfs_trans_read_buf_recur(bip); ASSERT(bp->b_ops != NULL || ops == NULL); *bpp = bp; return 0; } bp = xfs_buf_read_map(target, map, nmaps, flags, ops); if (!bp) { if (!(flags & XBF_TRYLOCK)) return -ENOMEM; return tp ? 0 : -EAGAIN; } /* * If we've had a read error, then the contents of the buffer are * invalid and should not be used. To ensure that a followup read tries * to pull the buffer from disk again, we clear the XBF_DONE flag and * mark the buffer stale. This ensures that anyone who has a current * reference to the buffer will interpret it's contents correctly and * future cache lookups will also treat it as an empty, uninitialised * buffer. */ if (bp->b_error) { error = bp->b_error; if (!XFS_FORCED_SHUTDOWN(mp)) xfs_buf_ioerror_alert(bp, __func__); bp->b_flags &= ~XBF_DONE; xfs_buf_stale(bp); if (tp && (tp->t_flags & XFS_TRANS_DIRTY)) xfs_force_shutdown(tp->t_mountp, SHUTDOWN_META_IO_ERROR); xfs_buf_relse(bp); /* bad CRC means corrupted metadata */ if (error == -EFSBADCRC) error = -EFSCORRUPTED; return error; } if (XFS_FORCED_SHUTDOWN(mp)) { xfs_buf_relse(bp); trace_xfs_trans_read_buf_shut(bp, _RET_IP_); return -EIO; } if (tp) { _xfs_trans_bjoin(tp, bp, 1); trace_xfs_trans_read_buf(bp->b_log_item); } ASSERT(bp->b_ops != NULL || ops == NULL); *bpp = bp; return 0; } /* * Release a buffer previously joined to the transaction. If the buffer is * modified within this transaction, decrement the recursion count but do not * release the buffer even if the count goes to 0. If the buffer is not modified * within the transaction, decrement the recursion count and release the buffer * if the recursion count goes to 0. * * If the buffer is to be released and it was not already dirty before this * transaction began, then also free the buf_log_item associated with it. * * If the transaction pointer is NULL, this is a normal xfs_buf_relse() call. */ void xfs_trans_brelse( struct xfs_trans *tp, struct xfs_buf *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); if (!tp) { xfs_buf_relse(bp); return; } trace_xfs_trans_brelse(bip); ASSERT(bip->bli_item.li_type == XFS_LI_BUF); ASSERT(atomic_read(&bip->bli_refcount) > 0); /* * If the release is for a recursive lookup, then decrement the count * and return. */ if (bip->bli_recur > 0) { bip->bli_recur--; return; } /* * If the buffer is invalidated or dirty in this transaction, we can't * release it until we commit. */ if (test_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags)) return; if (bip->bli_flags & XFS_BLI_STALE) return; /* * Unlink the log item from the transaction and clear the hold flag, if * set. We wouldn't want the next user of the buffer to get confused. */ ASSERT(!(bip->bli_flags & XFS_BLI_LOGGED)); xfs_trans_del_item(&bip->bli_item); bip->bli_flags &= ~XFS_BLI_HOLD; /* drop the reference to the bli */ xfs_buf_item_put(bip); bp->b_transp = NULL; xfs_buf_relse(bp); } /* * Mark the buffer as not needing to be unlocked when the buf item's * iop_unlock() routine is called. The buffer must already be locked * and associated with the given transaction. */ /* ARGSUSED */ void xfs_trans_bhold( xfs_trans_t *tp, xfs_buf_t *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(!(bip->bli_flags & XFS_BLI_STALE)); ASSERT(!(bip->__bli_format.blf_flags & XFS_BLF_CANCEL)); ASSERT(atomic_read(&bip->bli_refcount) > 0); bip->bli_flags |= XFS_BLI_HOLD; trace_xfs_trans_bhold(bip); } /* * Cancel the previous buffer hold request made on this buffer * for this transaction. */ void xfs_trans_bhold_release( xfs_trans_t *tp, xfs_buf_t *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(!(bip->bli_flags & XFS_BLI_STALE)); ASSERT(!(bip->__bli_format.blf_flags & XFS_BLF_CANCEL)); ASSERT(atomic_read(&bip->bli_refcount) > 0); ASSERT(bip->bli_flags & XFS_BLI_HOLD); bip->bli_flags &= ~XFS_BLI_HOLD; trace_xfs_trans_bhold_release(bip); } /* * Mark a buffer dirty in the transaction. */ void xfs_trans_dirty_buf( struct xfs_trans *tp, struct xfs_buf *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(bp->b_iodone == NULL || bp->b_iodone == xfs_buf_iodone_callbacks); /* * Mark the buffer as needing to be written out eventually, * and set its iodone function to remove the buffer's buf log * item from the AIL and free it when the buffer is flushed * to disk. See xfs_buf_attach_iodone() for more details * on li_cb and xfs_buf_iodone_callbacks(). * If we end up aborting this transaction, we trap this buffer * inside the b_bdstrat callback so that this won't get written to * disk. */ bp->b_flags |= XBF_DONE; ASSERT(atomic_read(&bip->bli_refcount) > 0); bp->b_iodone = xfs_buf_iodone_callbacks; bip->bli_item.li_cb = xfs_buf_iodone; /* * If we invalidated the buffer within this transaction, then * cancel the invalidation now that we're dirtying the buffer * again. There are no races with the code in xfs_buf_item_unpin(), * because we have a reference to the buffer this entire time. */ if (bip->bli_flags & XFS_BLI_STALE) { bip->bli_flags &= ~XFS_BLI_STALE; ASSERT(bp->b_flags & XBF_STALE); bp->b_flags &= ~XBF_STALE; bip->__bli_format.blf_flags &= ~XFS_BLF_CANCEL; } bip->bli_flags |= XFS_BLI_DIRTY | XFS_BLI_LOGGED; tp->t_flags |= XFS_TRANS_DIRTY; set_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags); } /* * This is called to mark bytes first through last inclusive of the given * buffer as needing to be logged when the transaction is committed. * The buffer must already be associated with the given transaction. * * First and last are numbers relative to the beginning of this buffer, * so the first byte in the buffer is numbered 0 regardless of the * value of b_blkno. */ void xfs_trans_log_buf( struct xfs_trans *tp, struct xfs_buf *bp, uint first, uint last) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(first <= last && last < BBTOB(bp->b_length)); ASSERT(!(bip->bli_flags & XFS_BLI_ORDERED)); xfs_trans_dirty_buf(tp, bp); trace_xfs_trans_log_buf(bip); xfs_buf_item_log(bip, first, last); } /* * Invalidate a buffer that is being used within a transaction. * * Typically this is because the blocks in the buffer are being freed, so we * need to prevent it from being written out when we're done. Allowing it * to be written again might overwrite data in the free blocks if they are * reallocated to a file. * * We prevent the buffer from being written out by marking it stale. We can't * get rid of the buf log item at this point because the buffer may still be * pinned by another transaction. If that is the case, then we'll wait until * the buffer is committed to disk for the last time (we can tell by the ref * count) and free it in xfs_buf_item_unpin(). Until that happens we will * keep the buffer locked so that the buffer and buf log item are not reused. * * We also set the XFS_BLF_CANCEL flag in the buf log format structure and log * the buf item. This will be used at recovery time to determine that copies * of the buffer in the log before this should not be replayed. * * We mark the item descriptor and the transaction dirty so that we'll hold * the buffer until after the commit. * * Since we're invalidating the buffer, we also clear the state about which * parts of the buffer have been logged. We also clear the flag indicating * that this is an inode buffer since the data in the buffer will no longer * be valid. * * We set the stale bit in the buffer as well since we're getting rid of it. */ void xfs_trans_binval( xfs_trans_t *tp, xfs_buf_t *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; int i; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); trace_xfs_trans_binval(bip); if (bip->bli_flags & XFS_BLI_STALE) { /* * If the buffer is already invalidated, then * just return. */ ASSERT(bp->b_flags & XBF_STALE); ASSERT(!(bip->bli_flags & (XFS_BLI_LOGGED | XFS_BLI_DIRTY))); ASSERT(!(bip->__bli_format.blf_flags & XFS_BLF_INODE_BUF)); ASSERT(!(bip->__bli_format.blf_flags & XFS_BLFT_MASK)); ASSERT(bip->__bli_format.blf_flags & XFS_BLF_CANCEL); ASSERT(test_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags)); ASSERT(tp->t_flags & XFS_TRANS_DIRTY); return; } 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; bip->__bli_format.blf_flags |= XFS_BLF_CANCEL; bip->__bli_format.blf_flags &= ~XFS_BLFT_MASK; for (i = 0; i < bip->bli_format_count; i++) { memset(bip->bli_formats[i].blf_data_map, 0, (bip->bli_formats[i].blf_map_size * sizeof(uint))); } set_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags); tp->t_flags |= XFS_TRANS_DIRTY; } /* * This call is used to indicate that the buffer contains on-disk inodes which * must be handled specially during recovery. They require special handling * because only the di_next_unlinked from the inodes in the buffer should be * recovered. The rest of the data in the buffer is logged via the inodes * themselves. * * All we do is set the XFS_BLI_INODE_BUF flag in the items flags so it can be * transferred to the buffer's log format structure so that we'll know what to * do at recovery time. */ void xfs_trans_inode_buf( xfs_trans_t *tp, xfs_buf_t *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); bip->bli_flags |= XFS_BLI_INODE_BUF; xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF); } /* * This call is used to indicate that the buffer is going to * be staled and was an inode buffer. This means it gets * special processing during unpin - where any inodes * associated with the buffer should be removed from ail. * There is also special processing during recovery, * any replay of the inodes in the buffer needs to be * prevented as the buffer may have been reused. */ void xfs_trans_stale_inode_buf( xfs_trans_t *tp, xfs_buf_t *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); bip->bli_flags |= XFS_BLI_STALE_INODE; bip->bli_item.li_cb = xfs_buf_iodone; xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF); } /* * Mark the buffer as being one which contains newly allocated * inodes. We need to make sure that even if this buffer is * relogged as an 'inode buf' we still recover all of the inode * images in the face of a crash. This works in coordination with * xfs_buf_item_committed() to ensure that the buffer remains in the * AIL at its original location even after it has been relogged. */ /* ARGSUSED */ void xfs_trans_inode_alloc_buf( xfs_trans_t *tp, xfs_buf_t *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); bip->bli_flags |= XFS_BLI_INODE_ALLOC_BUF; xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF); } /* * Mark the buffer as ordered for this transaction. This means that the contents * of the buffer are not recorded in the transaction but it is tracked in the * AIL as though it was. This allows us to record logical changes in * transactions rather than the physical changes we make to the buffer without * changing writeback ordering constraints of metadata buffers. */ bool xfs_trans_ordered_buf( struct xfs_trans *tp, struct xfs_buf *bp) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); if (xfs_buf_item_dirty_format(bip)) return false; bip->bli_flags |= XFS_BLI_ORDERED; trace_xfs_buf_item_ordered(bip); /* * We don't log a dirty range of an ordered buffer but it still needs * to be marked dirty and that it has been logged. */ xfs_trans_dirty_buf(tp, bp); return true; } /* * Set the type of the buffer for log recovery so that it can correctly identify * and hence attach the correct buffer ops to the buffer after replay. */ void xfs_trans_buf_set_type( struct xfs_trans *tp, struct xfs_buf *bp, enum xfs_blft type) { struct xfs_buf_log_item *bip = bp->b_log_item; if (!tp) return; ASSERT(bp->b_transp == tp); ASSERT(bip != NULL); ASSERT(atomic_read(&bip->bli_refcount) > 0); xfs_blft_to_flags(&bip->__bli_format, type); } void xfs_trans_buf_copy_type( struct xfs_buf *dst_bp, struct xfs_buf *src_bp) { struct xfs_buf_log_item *sbip = src_bp->b_log_item; struct xfs_buf_log_item *dbip = dst_bp->b_log_item; enum xfs_blft type; type = xfs_blft_from_flags(&sbip->__bli_format); xfs_blft_to_flags(&dbip->__bli_format, type); } /* * Similar to xfs_trans_inode_buf(), this marks the buffer as a cluster of * dquots. However, unlike in inode buffer recovery, dquot buffers get * recovered in their entirety. (Hence, no XFS_BLI_DQUOT_ALLOC_BUF flag). * The only thing that makes dquot buffers different from regular * buffers is that we must not replay dquot bufs when recovering * if a _corresponding_ quotaoff has happened. We also have to distinguish * between usr dquot bufs and grp dquot bufs, because usr and grp quotas * can be turned off independently. */ /* ARGSUSED */ void xfs_trans_dquot_buf( xfs_trans_t *tp, xfs_buf_t *bp, uint type) { struct xfs_buf_log_item *bip = bp->b_log_item; ASSERT(type == XFS_BLF_UDQUOT_BUF || type == XFS_BLF_PDQUOT_BUF || type == XFS_BLF_GDQUOT_BUF); bip->__bli_format.blf_flags |= type; switch (type) { case XFS_BLF_UDQUOT_BUF: type = XFS_BLFT_UDQUOT_BUF; break; case XFS_BLF_PDQUOT_BUF: type = XFS_BLFT_PDQUOT_BUF; break; case XFS_BLF_GDQUOT_BUF: type = XFS_BLFT_GDQUOT_BUF; break; default: type = XFS_BLFT_UNKNOWN_BUF; break; } xfs_trans_buf_set_type(tp, bp, type); }