X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=fs%2Fxfs%2Fxfs_iomap.c;h=7545dcdaa8aaeb8b7e15b4cc1071cb8dd38e8d12;hb=207d041602cead1c1a16288f6225aea9da1f5bc4;hp=5ecf3e3e86aaa4feab769541b0a7a5c1802a73cd;hpb=dd9f438e32900d67def49fa1b8961b3e19b6fefc;p=safe%2Fjmp%2Flinux-2.6 diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 5ecf3e3..7545dcd 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2000-2005 Silicon Graphics, Inc. + * Copyright (c) 2000-2006 Silicon Graphics, Inc. * All Rights Reserved. * * This program is free software; you can redistribute it and/or @@ -23,7 +23,6 @@ #include "xfs_trans.h" #include "xfs_sb.h" #include "xfs_ag.h" -#include "xfs_dir.h" #include "xfs_dir2.h" #include "xfs_alloc.h" #include "xfs_dmapi.h" @@ -32,7 +31,6 @@ #include "xfs_bmap_btree.h" #include "xfs_alloc_btree.h" #include "xfs_ialloc_btree.h" -#include "xfs_dir_sf.h" #include "xfs_dir2_sf.h" #include "xfs_attr_sf.h" #include "xfs_dinode.h" @@ -40,206 +38,71 @@ #include "xfs_ialloc.h" #include "xfs_btree.h" #include "xfs_bmap.h" -#include "xfs_bit.h" #include "xfs_rtalloc.h" #include "xfs_error.h" #include "xfs_itable.h" #include "xfs_rw.h" -#include "xfs_acl.h" -#include "xfs_cap.h" -#include "xfs_mac.h" #include "xfs_attr.h" #include "xfs_buf_item.h" #include "xfs_trans_space.h" #include "xfs_utils.h" #include "xfs_iomap.h" +#include "xfs_trace.h" -#if defined(XFS_RW_TRACE) -void -xfs_iomap_enter_trace( - int tag, - xfs_iocore_t *io, - xfs_off_t offset, - ssize_t count) -{ - xfs_inode_t *ip = XFS_IO_INODE(io); - - if (!ip->i_rwtrace) - return; - - ktrace_enter(ip->i_rwtrace, - (void *)((unsigned long)tag), - (void *)ip, - (void *)((unsigned long)((ip->i_d.di_size >> 32) & 0xffffffff)), - (void *)((unsigned long)(ip->i_d.di_size & 0xffffffff)), - (void *)((unsigned long)((offset >> 32) & 0xffffffff)), - (void *)((unsigned long)(offset & 0xffffffff)), - (void *)((unsigned long)count), - (void *)((unsigned long)((io->io_new_size >> 32) & 0xffffffff)), - (void *)((unsigned long)(io->io_new_size & 0xffffffff)), - (void *)NULL, - (void *)NULL, - (void *)NULL, - (void *)NULL, - (void *)NULL, - (void *)NULL, - (void *)NULL); -} - -void -xfs_iomap_map_trace( - int tag, - xfs_iocore_t *io, - xfs_off_t offset, - ssize_t count, - xfs_iomap_t *iomapp, - xfs_bmbt_irec_t *imapp, - int flags) -{ - xfs_inode_t *ip = XFS_IO_INODE(io); - - if (!ip->i_rwtrace) - return; - - ktrace_enter(ip->i_rwtrace, - (void *)((unsigned long)tag), - (void *)ip, - (void *)((unsigned long)((ip->i_d.di_size >> 32) & 0xffffffff)), - (void *)((unsigned long)(ip->i_d.di_size & 0xffffffff)), - (void *)((unsigned long)((offset >> 32) & 0xffffffff)), - (void *)((unsigned long)(offset & 0xffffffff)), - (void *)((unsigned long)count), - (void *)((unsigned long)flags), - (void *)((unsigned long)((iomapp->iomap_offset >> 32) & 0xffffffff)), - (void *)((unsigned long)(iomapp->iomap_offset & 0xffffffff)), - (void *)((unsigned long)(iomapp->iomap_delta)), - (void *)((unsigned long)(iomapp->iomap_bsize)), - (void *)((unsigned long)(iomapp->iomap_bn)), - (void *)(__psint_t)(imapp->br_startoff), - (void *)((unsigned long)(imapp->br_blockcount)), - (void *)(__psint_t)(imapp->br_startblock)); -} -#else -#define xfs_iomap_enter_trace(tag, io, offset, count) -#define xfs_iomap_map_trace(tag, io, offset, count, iomapp, imapp, flags) -#endif #define XFS_WRITEIO_ALIGN(mp,off) (((off) >> mp->m_writeio_log) \ << mp->m_writeio_log) #define XFS_STRAT_WRITE_IMAPS 2 #define XFS_WRITE_IMAPS XFS_BMAP_MAX_NMAP -STATIC int -xfs_imap_to_bmap( - xfs_iocore_t *io, - xfs_off_t offset, - xfs_bmbt_irec_t *imap, - xfs_iomap_t *iomapp, - int imaps, /* Number of imap entries */ - int iomaps, /* Number of iomap entries */ - int flags) -{ - xfs_mount_t *mp; - xfs_fsize_t nisize; - int pbm; - xfs_fsblock_t start_block; - - mp = io->io_mount; - nisize = XFS_SIZE(mp, io); - if (io->io_new_size > nisize) - nisize = io->io_new_size; - - for (pbm = 0; imaps && pbm < iomaps; imaps--, iomapp++, imap++, pbm++) { - iomapp->iomap_offset = XFS_FSB_TO_B(mp, imap->br_startoff); - iomapp->iomap_delta = offset - iomapp->iomap_offset; - iomapp->iomap_bsize = XFS_FSB_TO_B(mp, imap->br_blockcount); - iomapp->iomap_flags = flags; - - if (io->io_flags & XFS_IOCORE_RT) { - iomapp->iomap_flags |= IOMAP_REALTIME; - iomapp->iomap_target = mp->m_rtdev_targp; - } else { - iomapp->iomap_target = mp->m_ddev_targp; - } - start_block = imap->br_startblock; - if (start_block == HOLESTARTBLOCK) { - iomapp->iomap_bn = IOMAP_DADDR_NULL; - iomapp->iomap_flags |= IOMAP_HOLE; - } else if (start_block == DELAYSTARTBLOCK) { - iomapp->iomap_bn = IOMAP_DADDR_NULL; - iomapp->iomap_flags |= IOMAP_DELAY; - } else { - iomapp->iomap_bn = XFS_FSB_TO_DB_IO(io, start_block); - if (ISUNWRITTEN(imap)) - iomapp->iomap_flags |= IOMAP_UNWRITTEN; - } - - if ((iomapp->iomap_offset + iomapp->iomap_bsize) >= nisize) { - iomapp->iomap_flags |= IOMAP_EOF; - } - - offset += iomapp->iomap_bsize - iomapp->iomap_delta; - } - return pbm; /* Return the number filled */ -} - int xfs_iomap( - xfs_iocore_t *io, - xfs_off_t offset, - ssize_t count, - int flags, - xfs_iomap_t *iomapp, - int *niomaps) + struct xfs_inode *ip, + xfs_off_t offset, + ssize_t count, + int flags, + struct xfs_bmbt_irec *imap, + int *nimaps, + int *new) { - xfs_mount_t *mp = io->io_mount; - xfs_fileoff_t offset_fsb, end_fsb; - int error = 0; - int lockmode = 0; - xfs_bmbt_irec_t imap; - int nimaps = 1; - int bmapi_flags = 0; - int iomap_flags = 0; + struct xfs_mount *mp = ip->i_mount; + xfs_fileoff_t offset_fsb, end_fsb; + int error = 0; + int lockmode = 0; + int bmapi_flags = 0; + + ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFREG); + + *new = 0; if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); - switch (flags & - (BMAPI_READ | BMAPI_WRITE | BMAPI_ALLOCATE | - BMAPI_UNWRITTEN | BMAPI_DEVICE)) { + trace_xfs_iomap_enter(ip, offset, count, flags, NULL); + + switch (flags & (BMAPI_READ | BMAPI_WRITE | BMAPI_ALLOCATE)) { case BMAPI_READ: - xfs_iomap_enter_trace(XFS_IOMAP_READ_ENTER, io, offset, count); - lockmode = XFS_LCK_MAP_SHARED(mp, io); + lockmode = xfs_ilock_map_shared(ip); bmapi_flags = XFS_BMAPI_ENTIRE; break; case BMAPI_WRITE: - xfs_iomap_enter_trace(XFS_IOMAP_WRITE_ENTER, io, offset, count); - lockmode = XFS_ILOCK_EXCL|XFS_EXTSIZE_WR; + lockmode = XFS_ILOCK_EXCL; if (flags & BMAPI_IGNSTATE) bmapi_flags |= XFS_BMAPI_IGSTATE|XFS_BMAPI_ENTIRE; - XFS_ILOCK(mp, io, lockmode); + xfs_ilock(ip, lockmode); break; case BMAPI_ALLOCATE: - xfs_iomap_enter_trace(XFS_IOMAP_ALLOC_ENTER, io, offset, count); - lockmode = XFS_ILOCK_SHARED|XFS_EXTSIZE_RD; + lockmode = XFS_ILOCK_SHARED; bmapi_flags = XFS_BMAPI_ENTIRE; + /* Attempt non-blocking lock */ if (flags & BMAPI_TRYLOCK) { - if (!XFS_ILOCK_NOWAIT(mp, io, lockmode)) + if (!xfs_ilock_nowait(ip, lockmode)) return XFS_ERROR(EAGAIN); } else { - XFS_ILOCK(mp, io, lockmode); + xfs_ilock(ip, lockmode); } break; - case BMAPI_UNWRITTEN: - goto phase2; - case BMAPI_DEVICE: - lockmode = XFS_LCK_MAP_SHARED(mp, io); - iomapp->iomap_target = io->io_flags & XFS_IOCORE_RT ? - mp->m_rtdev_targp : mp->m_ddev_targp; - error = 0; - *niomaps = 1; - goto out; default: BUG(); } @@ -250,78 +113,63 @@ xfs_iomap( end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count); offset_fsb = XFS_B_TO_FSBT(mp, offset); - error = XFS_BMAPI(mp, NULL, io, offset_fsb, + error = xfs_bmapi(NULL, ip, offset_fsb, (xfs_filblks_t)(end_fsb - offset_fsb), - bmapi_flags, NULL, 0, &imap, - &nimaps, NULL); + bmapi_flags, NULL, 0, imap, + nimaps, NULL, NULL); if (error) goto out; -phase2: - switch (flags & (BMAPI_WRITE|BMAPI_ALLOCATE|BMAPI_UNWRITTEN)) { + switch (flags & (BMAPI_WRITE|BMAPI_ALLOCATE)) { case BMAPI_WRITE: /* If we found an extent, return it */ - if (nimaps && - (imap.br_startblock != HOLESTARTBLOCK) && - (imap.br_startblock != DELAYSTARTBLOCK)) { - xfs_iomap_map_trace(XFS_IOMAP_WRITE_MAP, io, - offset, count, iomapp, &imap, flags); + if (*nimaps && + (imap->br_startblock != HOLESTARTBLOCK) && + (imap->br_startblock != DELAYSTARTBLOCK)) { + trace_xfs_iomap_found(ip, offset, count, flags, imap); break; } if (flags & (BMAPI_DIRECT|BMAPI_MMAP)) { - error = XFS_IOMAP_WRITE_DIRECT(mp, io, offset, - count, flags, &imap, &nimaps, nimaps); + error = xfs_iomap_write_direct(ip, offset, count, flags, + imap, nimaps, *nimaps); } else { - error = XFS_IOMAP_WRITE_DELAY(mp, io, offset, count, - flags, &imap, &nimaps); + error = xfs_iomap_write_delay(ip, offset, count, flags, + imap, nimaps); } if (!error) { - xfs_iomap_map_trace(XFS_IOMAP_ALLOC_MAP, io, - offset, count, iomapp, &imap, flags); + trace_xfs_iomap_alloc(ip, offset, count, flags, imap); } - iomap_flags = IOMAP_NEW; + *new = 1; break; case BMAPI_ALLOCATE: /* If we found an extent, return it */ - XFS_IUNLOCK(mp, io, lockmode); + xfs_iunlock(ip, lockmode); lockmode = 0; - if (nimaps && !ISNULLSTARTBLOCK(imap.br_startblock)) { - xfs_iomap_map_trace(XFS_IOMAP_WRITE_MAP, io, - offset, count, iomapp, &imap, flags); + if (*nimaps && !isnullstartblock(imap->br_startblock)) { + trace_xfs_iomap_found(ip, offset, count, flags, imap); break; } - error = XFS_IOMAP_WRITE_ALLOCATE(mp, io, offset, count, - &imap, &nimaps); - break; - case BMAPI_UNWRITTEN: - lockmode = 0; - error = XFS_IOMAP_WRITE_UNWRITTEN(mp, io, offset, count); - nimaps = 0; + error = xfs_iomap_write_allocate(ip, offset, count, + imap, nimaps); break; } - if (nimaps) { - *niomaps = xfs_imap_to_bmap(io, offset, &imap, - iomapp, nimaps, *niomaps, iomap_flags); - } else if (niomaps) { - *niomaps = 0; - } + ASSERT(*nimaps <= 1); out: if (lockmode) - XFS_IUNLOCK(mp, io, lockmode); + xfs_iunlock(ip, lockmode); return XFS_ERROR(error); } STATIC int xfs_iomap_eof_align_last_fsb( xfs_mount_t *mp, - xfs_iocore_t *io, - xfs_fsize_t isize, + xfs_inode_t *ip, xfs_extlen_t extsize, xfs_fileoff_t *last_fsb) { @@ -329,7 +177,7 @@ xfs_iomap_eof_align_last_fsb( xfs_extlen_t align; int eof, error; - if (io->io_flags & XFS_IOCORE_RT) + if (XFS_IS_REALTIME_INODE(ip)) ; /* * If mounted with the "-o swalloc" option, roundup the allocation @@ -337,14 +185,14 @@ xfs_iomap_eof_align_last_fsb( * stripe width and we are allocating past the allocation eof. */ else if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC) && - (isize >= XFS_FSB_TO_B(mp, mp->m_swidth))) + (ip->i_size >= XFS_FSB_TO_B(mp, mp->m_swidth))) new_last_fsb = roundup_64(*last_fsb, mp->m_swidth); /* * Roundup the allocation request to a stripe unit (m_dalign) boundary * if the file size is >= stripe unit size, and we are allocating past * the allocation eof. */ - else if (mp->m_dalign && (isize >= XFS_FSB_TO_B(mp, mp->m_dalign))) + else if (mp->m_dalign && (ip->i_size >= XFS_FSB_TO_B(mp, mp->m_dalign))) new_last_fsb = roundup_64(*last_fsb, mp->m_dalign); /* @@ -360,7 +208,7 @@ xfs_iomap_eof_align_last_fsb( } if (new_last_fsb) { - error = XFS_BMAP_EOF(mp, io, new_last_fsb, XFS_DATA_FORK, &eof); + error = xfs_bmap_eof(ip, new_last_fsb, XFS_DATA_FORK, &eof); if (error) return error; if (eof) @@ -370,35 +218,20 @@ xfs_iomap_eof_align_last_fsb( } STATIC int -xfs_flush_space( +xfs_cmn_err_fsblock_zero( xfs_inode_t *ip, - int *fsynced, - int *ioflags) + xfs_bmbt_irec_t *imap) { - switch (*fsynced) { - case 0: - if (ip->i_delayed_blks) { - xfs_iunlock(ip, XFS_ILOCK_EXCL); - xfs_flush_inode(ip); - xfs_ilock(ip, XFS_ILOCK_EXCL); - *fsynced = 1; - } else { - *ioflags |= BMAPI_SYNC; - *fsynced = 2; - } - return 0; - case 1: - *fsynced = 2; - *ioflags |= BMAPI_SYNC; - return 0; - case 2: - xfs_iunlock(ip, XFS_ILOCK_EXCL); - xfs_flush_device(ip); - xfs_ilock(ip, XFS_ILOCK_EXCL); - *fsynced = 3; - return 0; - } - return 1; + xfs_cmn_err(XFS_PTAG_FSBLOCK_ZERO, CE_ALERT, ip->i_mount, + "Access to block zero in inode %llu " + "start_block: %llx start_off: %llx " + "blkcnt: %llx extent-state: %x\n", + (unsigned long long)ip->i_ino, + (unsigned long long)imap->br_startblock, + (unsigned long long)imap->br_startoff, + (unsigned long long)imap->br_blockcount, + imap->br_state); + return EFSCORRUPTED; } int @@ -412,13 +245,11 @@ xfs_iomap_write_direct( int found) { xfs_mount_t *mp = ip->i_mount; - xfs_iocore_t *io = &ip->i_iocore; xfs_fileoff_t offset_fsb; xfs_fileoff_t last_fsb; xfs_filblks_t count_fsb, resaligned; xfs_fsblock_t firstfsb; xfs_extlen_t extsz, temp; - xfs_fsize_t isize; int nimaps; int bmapi_flag; int quota_flag; @@ -434,27 +265,17 @@ xfs_iomap_write_direct( * Make sure that the dquots are there. This doesn't hold * the ilock across a disk read. */ - error = XFS_QM_DQATTACH(ip->i_mount, ip, XFS_QMOPT_ILOCKED); + error = xfs_qm_dqattach_locked(ip, 0); if (error) return XFS_ERROR(error); rt = XFS_IS_REALTIME_INODE(ip); - if (unlikely(rt)) { - if (!(extsz = ip->i_d.di_extsize)) - extsz = mp->m_sb.sb_rextsize; - } else { - extsz = ip->i_d.di_extsize; - } - - isize = ip->i_d.di_size; - if (io->io_new_size > isize) - isize = io->io_new_size; + extsz = xfs_get_extsz_hint(ip); - offset_fsb = XFS_B_TO_FSBT(mp, offset); - last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count))); - if ((offset + count) > isize) { - error = xfs_iomap_eof_align_last_fsb(mp, io, isize, extsz, - &last_fsb); + offset_fsb = XFS_B_TO_FSBT(mp, offset); + last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count))); + if ((offset + count) > ip->i_size) { + error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb); if (error) goto error_out; } else { @@ -477,13 +298,13 @@ xfs_iomap_write_direct( if (unlikely(rt)) { resrtextents = qblocks = resaligned; resrtextents /= mp->m_sb.sb_rextsize; - resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0); - quota_flag = XFS_QMOPT_RES_RTBLKS; - } else { - resrtextents = 0; + resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0); + quota_flag = XFS_QMOPT_RES_RTBLKS; + } else { + resrtextents = 0; resblks = qblocks = XFS_DIOSTRAT_SPACE_RES(mp, resaligned); - quota_flag = XFS_QMOPT_RES_REGBLKS; - } + quota_flag = XFS_QMOPT_RES_REGBLKS; + } /* * Allocate and setup the transaction @@ -503,8 +324,7 @@ xfs_iomap_write_direct( if (error) goto error_out; - error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, tp, ip, - qblocks, 0, quota_flag); + error = xfs_trans_reserve_quota_nblks(tp, ip, qblocks, 0, quota_flag); if (error) goto error1; @@ -512,26 +332,26 @@ xfs_iomap_write_direct( xfs_trans_ihold(tp, ip); bmapi_flag = XFS_BMAPI_WRITE; - if ((flags & BMAPI_DIRECT) && (offset < ip->i_d.di_size || extsz)) + if ((flags & BMAPI_DIRECT) && (offset < ip->i_size || extsz)) bmapi_flag |= XFS_BMAPI_PREALLOC; /* * Issue the xfs_bmapi() call to allocate the blocks */ - XFS_BMAP_INIT(&free_list, &firstfsb); + 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(tp, ip, offset_fsb, count_fsb, bmapi_flag, + &firstfsb, 0, &imap, &nimaps, &free_list, NULL); if (error) goto error0; /* * Complete the transaction */ - error = xfs_bmap_finish(&tp, &free_list, firstfsb, &committed); + error = xfs_bmap_finish(&tp, &free_list, &committed); if (error) goto error0; - error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES, NULL); + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); if (error) goto error_out; @@ -539,26 +359,22 @@ xfs_iomap_write_direct( * Copy any maps to caller's array and return any error. */ if (nimaps == 0) { - error = (ENOSPC); + error = ENOSPC; + goto error_out; + } + + if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip))) { + error = xfs_cmn_err_fsblock_zero(ip, &imap); goto error_out; } *ret_imap = imap; *nmaps = 1; - if ( !(io->io_flags & XFS_IOCORE_RT) && !ret_imap->br_startblock) { - cmn_err(CE_PANIC,"Access to block zero: fs <%s> inode: %lld " - "start_block : %llx start_off : %llx blkcnt : %llx " - "extent-state : %x \n", - (ip->i_mount)->m_fsname, - (long long)ip->i_ino, - ret_imap->br_startblock, ret_imap->br_startoff, - ret_imap->br_blockcount,ret_imap->br_state); - } return 0; error0: /* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */ xfs_bmap_cancel(&free_list); - XFS_TRANS_UNRESERVE_QUOTA_NBLKS(mp, tp, ip, qblocks, 0, quota_flag); + xfs_trans_unreserve_quota_nblks(tp, ip, qblocks, 0, quota_flag); error1: /* Just cancel transaction */ xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT); @@ -569,21 +385,14 @@ error_out: } /* - * If the caller is doing a write at the end of the file, - * then extend the allocation out to the file system's write - * iosize. We clean up any extra space left over when the - * file is closed in xfs_inactive(). - * - * For sync writes, we are flushing delayed allocate space to - * try to make additional space available for allocation near - * the filesystem full boundary - preallocation hurts in that - * situation, of course. + * If the caller is doing a write at the end of the file, then extend the + * allocation out to the file system's write iosize. We clean up any extra + * space left over when the file is closed in xfs_inactive(). */ STATIC int xfs_iomap_eof_want_preallocate( xfs_mount_t *mp, - xfs_iocore_t *io, - xfs_fsize_t isize, + xfs_inode_t *ip, xfs_off_t offset, size_t count, int ioflag, @@ -597,7 +406,7 @@ xfs_iomap_eof_want_preallocate( int n, error, imaps; *prealloc = 0; - if ((ioflag & BMAPI_SYNC) || (offset + count) <= isize) + if ((offset + count) <= ip->i_size) return 0; /* @@ -608,8 +417,9 @@ xfs_iomap_eof_want_preallocate( count_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp)); while (count_fsb > 0) { imaps = nimaps; - error = XFS_BMAPI(mp, NULL, io, start_fsb, count_fsb, - 0, &firstblock, 0, imap, &imaps, NULL); + firstblock = NULLFSBLOCK; + error = xfs_bmapi(NULL, ip, start_fsb, count_fsb, 0, + &firstblock, 0, imap, &imaps, NULL, NULL); if (error) return error; for (n = 0; n < imaps; n++) { @@ -634,48 +444,36 @@ xfs_iomap_write_delay( int *nmaps) { xfs_mount_t *mp = ip->i_mount; - xfs_iocore_t *io = &ip->i_iocore; xfs_fileoff_t offset_fsb; xfs_fileoff_t last_fsb; xfs_off_t aligned_offset; xfs_fileoff_t ioalign; xfs_fsblock_t firstblock; xfs_extlen_t extsz; - xfs_fsize_t isize; int nimaps; xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS]; - int prealloc, fsynced = 0; + int prealloc, flushed = 0; int error; - ASSERT(ismrlocked(&ip->i_lock, MR_UPDATE) != 0); + ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); /* * Make sure that the dquots are there. This doesn't hold * the ilock across a disk read. */ - error = XFS_QM_DQATTACH(mp, ip, XFS_QMOPT_ILOCKED); + error = xfs_qm_dqattach_locked(ip, 0); if (error) return XFS_ERROR(error); - if (XFS_IS_REALTIME_INODE(ip)) { - if (!(extsz = ip->i_d.di_extsize)) - extsz = mp->m_sb.sb_rextsize; - } else { - extsz = ip->i_d.di_extsize; - } - + extsz = xfs_get_extsz_hint(ip); offset_fsb = XFS_B_TO_FSBT(mp, offset); -retry: - isize = ip->i_d.di_size; - if (io->io_new_size > isize) - isize = io->io_new_size; - - error = xfs_iomap_eof_want_preallocate(mp, io, isize, offset, count, + error = xfs_iomap_eof_want_preallocate(mp, ip, offset, count, ioflag, imap, XFS_WRITE_IMAPS, &prealloc); if (error) return error; +retry: if (prealloc) { aligned_offset = XFS_WRITEIO_ALIGN(mp, (offset + count - 1)); ioalign = XFS_B_TO_FSBT(mp, aligned_offset); @@ -685,8 +483,7 @@ retry: } if (prealloc || extsz) { - error = xfs_iomap_eof_align_last_fsb(mp, io, isize, extsz, - &last_fsb); + error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb); if (error) return error; } @@ -697,33 +494,32 @@ retry: (xfs_filblks_t)(last_fsb - offset_fsb), XFS_BMAPI_DELAY | XFS_BMAPI_WRITE | XFS_BMAPI_ENTIRE, &firstblock, 1, imap, - &nimaps, NULL); + &nimaps, NULL, NULL); if (error && (error != ENOSPC)) return XFS_ERROR(error); /* * If bmapi returned us nothing, and if we didn't get back EDQUOT, - * then we must have run out of space - flush delalloc, and retry.. + * then we must have run out of space - flush all other inodes with + * delalloc blocks and retry without EOF preallocation. */ if (nimaps == 0) { - xfs_iomap_enter_trace(XFS_IOMAP_WRITE_NOSPACE, - io, offset, count); - if (xfs_flush_space(ip, &fsynced, &ioflag)) + trace_xfs_delalloc_enospc(ip, offset, count); + if (flushed) return XFS_ERROR(ENOSPC); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + xfs_flush_inodes(ip); + xfs_ilock(ip, XFS_ILOCK_EXCL); + + flushed = 1; error = 0; + prealloc = 0; goto retry; } - if (!(io->io_flags & XFS_IOCORE_RT) && !ret_imap->br_startblock) { - cmn_err(CE_PANIC,"Access to block zero: fs <%s> inode: %lld " - "start_block : %llx start_off : %llx blkcnt : %llx " - "extent-state : %x \n", - (ip->i_mount)->m_fsname, - (long long)ip->i_ino, - ret_imap->br_startblock, ret_imap->br_startoff, - ret_imap->br_blockcount,ret_imap->br_state); - } + if (!(imap[0].br_startblock || XFS_IS_REALTIME_INODE(ip))) + return xfs_cmn_err_fsblock_zero(ip, &imap[0]); *ret_imap = imap[0]; *nmaps = 1; @@ -737,6 +533,9 @@ retry: * the originating callers request. * * Called without a lock on the inode. + * + * We no longer bother to look at the incoming map - all we have to + * guarantee is that whatever we allocate fills the required range. */ int xfs_iomap_write_allocate( @@ -747,15 +546,14 @@ xfs_iomap_write_allocate( int *retmap) { xfs_mount_t *mp = ip->i_mount; - xfs_iocore_t *io = &ip->i_iocore; xfs_fileoff_t offset_fsb, last_block; xfs_fileoff_t end_fsb, map_start_fsb; xfs_fsblock_t first_block; xfs_bmap_free_t free_list; xfs_filblks_t count_fsb; - xfs_bmbt_irec_t imap[XFS_STRAT_WRITE_IMAPS]; + xfs_bmbt_irec_t imap; xfs_trans_t *tp; - int i, nimaps, committed; + int nimaps, committed; int error = 0; int nres; @@ -764,7 +562,8 @@ xfs_iomap_write_allocate( /* * Make sure that the dquots are there. */ - if ((error = XFS_QM_DQATTACH(mp, ip, 0))) + error = xfs_qm_dqattach(ip, 0); + if (error) return XFS_ERROR(error); offset_fsb = XFS_B_TO_FSBT(mp, offset); @@ -786,18 +585,12 @@ xfs_iomap_write_allocate( nimaps = 0; while (nimaps == 0) { tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE); + tp->t_flags |= XFS_TRANS_RESERVE; nres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK); error = xfs_trans_reserve(tp, nres, XFS_WRITE_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_WRITE_LOG_COUNT); - if (error == ENOSPC) { - error = xfs_trans_reserve(tp, 0, - XFS_WRITE_LOG_RES(mp), - 0, - XFS_TRANS_PERM_LOG_RES, - XFS_WRITE_LOG_COUNT); - } if (error) { xfs_trans_cancel(tp, 0); return XFS_ERROR(error); @@ -806,18 +599,46 @@ xfs_iomap_write_allocate( xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); xfs_trans_ihold(tp, ip); - XFS_BMAP_INIT(&free_list, &first_block); + xfs_bmap_init(&free_list, &first_block); - nimaps = XFS_STRAT_WRITE_IMAPS; /* - * Ensure we don't go beyond eof - it is possible - * the extents changed since we did the read call, - * we dropped the ilock in the interim. + * it is possible that the extents have changed since + * we did the read call as we dropped the ilock for a + * while. We have to be careful about truncates or hole + * punchs here - we are not allowed to allocate + * non-delalloc blocks here. + * + * The only protection against truncation is the pages + * for the range we are being asked to convert are + * locked and hence a truncate will block on them + * first. + * + * As a result, if we go beyond the range we really + * need and hit an delalloc extent boundary followed by + * a hole while we have excess blocks in the map, we + * will fill the hole incorrectly and overrun the + * transaction reservation. + * + * Using a single map prevents this as we are forced to + * check each map we look for overlap with the desired + * range and abort as soon as we find it. Also, given + * that we only return a single map, having one beyond + * what we can return is probably a bit silly. + * + * We also need to check that we don't go beyond EOF; + * this is a truncate optimisation as a truncate sets + * the new file size before block on the pages we + * currently have locked under writeback. Because they + * are about to be tossed, we don't need to write them + * back.... */ + nimaps = 1; + end_fsb = XFS_B_TO_FSB(mp, ip->i_size); + error = xfs_bmap_last_offset(NULL, ip, &last_block, + XFS_DATA_FORK); + if (error) + goto trans_cancel; - end_fsb = XFS_B_TO_FSB(mp, ip->i_d.di_size); - xfs_bmap_last_offset(NULL, ip, &last_block, - XFS_DATA_FORK); last_block = XFS_FILEOFF_MAX(last_block, end_fsb); if ((map_start_fsb + count_fsb) > last_block) { count_fsb = last_block - map_start_fsb; @@ -830,17 +651,15 @@ xfs_iomap_write_allocate( /* Go get the actual blocks */ error = xfs_bmapi(tp, ip, map_start_fsb, count_fsb, XFS_BMAPI_WRITE, &first_block, 1, - imap, &nimaps, &free_list); + &imap, &nimaps, &free_list, NULL); if (error) goto trans_cancel; - error = xfs_bmap_finish(&tp, &free_list, - first_block, &committed); + error = xfs_bmap_finish(&tp, &free_list, &committed); if (error) goto trans_cancel; - error = xfs_trans_commit(tp, - XFS_TRANS_RELEASE_LOG_RES, NULL); + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); if (error) goto error0; @@ -851,37 +670,24 @@ xfs_iomap_write_allocate( * See if we were able to allocate an extent that * covers at least part of the callers request */ - - for (i = 0; i < nimaps; i++) { - if ( !(io->io_flags & XFS_IOCORE_RT) && - !imap[i].br_startblock) { - cmn_err(CE_PANIC,"Access to block zero: " - "fs <%s> inode: %lld " - "start_block : %llx start_off : %llx " - "blkcnt : %llx extent-state : %x \n", - (ip->i_mount)->m_fsname, - (long long)ip->i_ino, - imap[i].br_startblock, - imap[i].br_startoff, - imap[i].br_blockcount,imap[i].br_state); - } - if ((offset_fsb >= imap[i].br_startoff) && - (offset_fsb < (imap[i].br_startoff + - imap[i].br_blockcount))) { - *map = imap[i]; - *retmap = 1; - XFS_STATS_INC(xs_xstrat_quick); - return 0; - } - count_fsb -= imap[i].br_blockcount; + if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip))) + return xfs_cmn_err_fsblock_zero(ip, &imap); + + if ((offset_fsb >= imap.br_startoff) && + (offset_fsb < (imap.br_startoff + + imap.br_blockcount))) { + *map = imap; + *retmap = 1; + XFS_STATS_INC(xs_xstrat_quick); + return 0; } - /* So far we have not mapped the requested part of the + /* + * So far we have not mapped the requested part of the * file, just surrounding data, try again. */ - nimaps--; - map_start_fsb = imap[nimaps].br_startoff + - imap[nimaps].br_blockcount; + count_fsb -= imap.br_blockcount; + map_start_fsb = imap.br_startoff + imap.br_blockcount; } trans_cancel: @@ -899,7 +705,6 @@ xfs_iomap_write_unwritten( size_t count) { xfs_mount_t *mp = ip->i_mount; - xfs_iocore_t *io = &ip->i_iocore; xfs_fileoff_t offset_fsb; xfs_filblks_t count_fsb; xfs_filblks_t numblks_fsb; @@ -912,13 +717,22 @@ xfs_iomap_write_unwritten( int committed; int error; - xfs_iomap_enter_trace(XFS_IOMAP_UNWRITTEN, - &ip->i_iocore, offset, count); + trace_xfs_unwritten_convert(ip, offset, count); offset_fsb = XFS_B_TO_FSBT(mp, offset); count_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count); count_fsb = (xfs_filblks_t)(count_fsb - offset_fsb); + /* + * Reserve enough blocks in this transaction for two complete extent + * btree splits. We may be converting the middle part of an unwritten + * extent and in this case we will insert two new extents in the btree + * each of which could cause a full split. + * + * This reservation amount will be used in the first call to + * xfs_bmbt_split() to select an AG with enough space to satisfy the + * rest of the operation. + */ resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0) << 1; do { @@ -926,16 +740,23 @@ xfs_iomap_write_unwritten( * set up a transaction to convert the range of extents * from unwritten to real. Do allocations in a loop until * we have covered the range passed in. + * + * Note that we open code the transaction allocation here + * to pass KM_NOFS--we can't risk to recursing back into + * the filesystem here as we might be asked to write out + * the same inode that we complete here and might deadlock + * on the iolock. */ - - tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE); + xfs_wait_for_freeze(mp, SB_FREEZE_TRANS); + tp = _xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE, KM_NOFS); + tp->t_flags |= XFS_TRANS_RESERVE; error = xfs_trans_reserve(tp, resblks, XFS_WRITE_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_WRITE_LOG_COUNT); if (error) { xfs_trans_cancel(tp, 0); - goto error0; + return XFS_ERROR(error); } xfs_ilock(ip, XFS_ILOCK_EXCL); @@ -945,33 +766,25 @@ xfs_iomap_write_unwritten( /* * Modify the unwritten extent state of the buffer. */ - XFS_BMAP_INIT(&free_list, &firstfsb); + xfs_bmap_init(&free_list, &firstfsb); nimaps = 1; error = xfs_bmapi(tp, ip, offset_fsb, count_fsb, XFS_BMAPI_WRITE|XFS_BMAPI_CONVERT, &firstfsb, - 1, &imap, &nimaps, &free_list); + 1, &imap, &nimaps, &free_list, NULL); if (error) goto error_on_bmapi_transaction; - error = xfs_bmap_finish(&(tp), &(free_list), - firstfsb, &committed); + error = xfs_bmap_finish(&(tp), &(free_list), &committed); if (error) goto error_on_bmapi_transaction; - error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES, NULL); + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); xfs_iunlock(ip, XFS_ILOCK_EXCL); if (error) - goto error0; - - if ( !(io->io_flags & XFS_IOCORE_RT) && !imap.br_startblock) { - cmn_err(CE_PANIC,"Access to block zero: fs <%s> " - "inode: %lld start_block : %llx start_off : " - "%llx blkcnt : %llx extent-state : %x \n", - (ip->i_mount)->m_fsname, - (long long)ip->i_ino, - imap.br_startblock,imap.br_startoff, - imap.br_blockcount,imap.br_state); - } + return XFS_ERROR(error); + + if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip))) + return xfs_cmn_err_fsblock_zero(ip, &imap); if ((numblks_fsb = imap.br_blockcount) == 0) { /* @@ -991,6 +804,5 @@ error_on_bmapi_transaction: xfs_bmap_cancel(&free_list); xfs_trans_cancel(tp, (XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT)); xfs_iunlock(ip, XFS_ILOCK_EXCL); -error0: return XFS_ERROR(error); }