xfs: split out transaction reservation code

The transaction reservation size calculations is used by both kernel
and userspace, but most of the transaction code in xfs_trans.c is
kernel specific. Split all the transaction reservation code out into
it's own files to make sharing with userspace simpler. This just
leaves kernel-only definitions in xfs_trans.h, so it doesn't need to
be shared with userspace anymore, either.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>
This commit is contained in:
Dave Chinner 2013-08-12 20:49:32 +10:00 committed by Ben Myers
parent d386b32b55
commit 7fd36c4418
10 changed files with 851 additions and 772 deletions

View File

@ -47,6 +47,7 @@ xfs-y += xfs_aops.o \
xfs_mru_cache.o \ xfs_mru_cache.o \
xfs_rename.o \ xfs_rename.o \
xfs_super.o \ xfs_super.o \
xfs_trans.o \
xfs_utils.o \ xfs_utils.o \
xfs_vnodeops.o \ xfs_vnodeops.o \
xfs_xattr.o \ xfs_xattr.o \
@ -76,7 +77,7 @@ xfs-y += xfs_alloc.o \
xfs_log_recover.o \ xfs_log_recover.o \
xfs_mount.o \ xfs_mount.o \
xfs_symlink.o \ xfs_symlink.o \
xfs_trans.o xfs_trans_resv.o
# low-level transaction/log code # low-level transaction/log code
xfs-y += xfs_log.o \ xfs-y += xfs_log.o \

View File

@ -17,6 +17,7 @@
*/ */
#include "xfs.h" #include "xfs.h"
#include "xfs_log_format.h" #include "xfs_log_format.h"
#include "xfs_trans_resv.h"
#include "xfs_acl.h" #include "xfs_acl.h"
#include "xfs_attr.h" #include "xfs_attr.h"
#include "xfs_bmap_btree.h" #include "xfs_bmap_btree.h"

View File

@ -35,6 +35,7 @@
#include <linux/freezer.h> #include <linux/freezer.h>
#include "xfs_sb.h" #include "xfs_sb.h"
#include "xfs_trans_resv.h"
#include "xfs_log.h" #include "xfs_log.h"
#include "xfs_ag.h" #include "xfs_ag.h"
#include "xfs_mount.h" #include "xfs_mount.h"

View File

@ -17,12 +17,12 @@
*/ */
#include "xfs.h" #include "xfs.h"
#include "xfs_format.h" #include "xfs_format.h"
#include "xfs_sb.h"
#include "xfs_log.h" #include "xfs_log.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h" #include "xfs_ag.h"
#include "xfs_mount.h" #include "xfs_mount.h"
#include "xfs_quota.h" #include "xfs_quota.h"
#include "xfs_trans.h"
#include "xfs_alloc_btree.h" #include "xfs_alloc_btree.h"
#include "xfs_bmap_btree.h" #include "xfs_bmap_btree.h"
#include "xfs_ialloc_btree.h" #include "xfs_ialloc_btree.h"

View File

@ -18,45 +18,7 @@
#ifndef __XFS_MOUNT_H__ #ifndef __XFS_MOUNT_H__
#define __XFS_MOUNT_H__ #define __XFS_MOUNT_H__
typedef struct xfs_trans_reservations { #ifdef __KERNEL__
uint tr_write; /* extent alloc trans */
uint tr_itruncate; /* truncate trans */
uint tr_rename; /* rename trans */
uint tr_link; /* link trans */
uint tr_remove; /* unlink trans */
uint tr_symlink; /* symlink trans */
uint tr_create; /* create trans */
uint tr_mkdir; /* mkdir trans */
uint tr_ifree; /* inode free trans */
uint tr_ichange; /* inode update trans */
uint tr_growdata; /* fs data section grow trans */
uint tr_swrite; /* sync write inode trans */
uint tr_addafork; /* cvt inode to attributed trans */
uint tr_writeid; /* write setuid/setgid file */
uint tr_attrinval; /* attr fork buffer invalidation */
uint tr_attrsetm; /* set/create an attribute at mount time */
uint tr_attrsetrt; /* set/create an attribute at runtime */
uint tr_attrrm; /* remove an attribute */
uint tr_clearagi; /* clear bad agi unlinked ino bucket */
uint tr_growrtalloc; /* grow realtime allocations */
uint tr_growrtzero; /* grow realtime zeroing */
uint tr_growrtfree; /* grow realtime freeing */
uint tr_qm_sbchange; /* change quota flags */
uint tr_qm_setqlim; /* adjust quota limits */
uint tr_qm_dqalloc; /* allocate quota on disk */
uint tr_qm_quotaoff; /* turn quota off */
uint tr_qm_equotaoff;/* end of turn quota off */
uint tr_sb; /* modify superblock */
} xfs_trans_reservations_t;
#ifndef __KERNEL__
#define xfs_daddr_to_agno(mp,d) \
((xfs_agnumber_t)(XFS_BB_TO_FSBT(mp, d) / (mp)->m_sb.sb_agblocks))
#define xfs_daddr_to_agbno(mp,d) \
((xfs_agblock_t)(XFS_BB_TO_FSBT(mp, d) % (mp)->m_sb.sb_agblocks))
#else /* __KERNEL__ */
struct xlog; struct xlog;
struct xfs_inode; struct xfs_inode;
@ -174,7 +136,7 @@ typedef struct xfs_mount {
int m_ialloc_blks; /* blocks in inode allocation */ int m_ialloc_blks; /* blocks in inode allocation */
int m_inoalign_mask;/* mask sb_inoalignmt if used */ int m_inoalign_mask;/* mask sb_inoalignmt if used */
uint m_qflags; /* quota status flags */ uint m_qflags; /* quota status flags */
xfs_trans_reservations_t m_reservations;/* precomputed res values */ struct xfs_trans_resv m_reservations; /* precomputed res values */
__uint64_t m_maxicount; /* maximum inode count */ __uint64_t m_maxicount; /* maximum inode count */
__uint64_t m_resblks; /* total reserved blocks */ __uint64_t m_resblks; /* total reserved blocks */
__uint64_t m_resblks_avail;/* available reserved blocks */ __uint64_t m_resblks_avail;/* available reserved blocks */

View File

@ -17,8 +17,9 @@
*/ */
#include "xfs.h" #include "xfs.h"
#include "xfs_format.h" #include "xfs_format.h"
#include "xfs_sb.h" #include "xfs_trans_resv.h"
#include "xfs_log.h" #include "xfs_log.h"
#include "xfs_sb.h"
#include "xfs_ag.h" #include "xfs_ag.h"
#include "xfs_mount.h" #include "xfs_mount.h"
#include "xfs_quota.h" #include "xfs_quota.h"

View File

@ -48,629 +48,6 @@
kmem_zone_t *xfs_trans_zone; kmem_zone_t *xfs_trans_zone;
kmem_zone_t *xfs_log_item_desc_zone; kmem_zone_t *xfs_log_item_desc_zone;
/*
* A buffer has a format structure overhead in the log in addition
* to the data, so we need to take this into account when reserving
* space in a transaction for a buffer. Round the space required up
* to a multiple of 128 bytes so that we don't change the historical
* reservation that has been used for this overhead.
*/
STATIC uint
xfs_buf_log_overhead(void)
{
return round_up(sizeof(struct xlog_op_header) +
sizeof(struct xfs_buf_log_format), 128);
}
/*
* Calculate out transaction log reservation per item in bytes.
*
* The nbufs argument is used to indicate the number of items that
* will be changed in a transaction. size is used to tell how many
* bytes should be reserved per item.
*/
STATIC uint
xfs_calc_buf_res(
uint nbufs,
uint size)
{
return nbufs * (size + xfs_buf_log_overhead());
}
/*
* Various log reservation values.
*
* These are based on the size of the file system block because that is what
* most transactions manipulate. Each adds in an additional 128 bytes per
* item logged to try to account for the overhead of the transaction mechanism.
*
* Note: Most of the reservations underestimate the number of allocation
* groups into which they could free extents in the xfs_bmap_finish() call.
* This is because the number in the worst case is quite high and quite
* unusual. In order to fix this we need to change xfs_bmap_finish() to free
* extents in only a single AG at a time. This will require changes to the
* EFI code as well, however, so that the EFI for the extents not freed is
* logged again in each transaction. See SGI PV #261917.
*
* Reservation functions here avoid a huge stack in xfs_trans_init due to
* register overflow from temporaries in the calculations.
*/
/*
* In a write transaction we can allocate a maximum of 2
* extents. This gives:
* the inode getting the new extents: inode size
* the inode's bmap btree: max depth * block size
* the agfs of the ags from which the extents are allocated: 2 * sector
* the superblock free block counter: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
* And the bmap_finish transaction can free bmap blocks in a join:
* the agfs of the ags containing the blocks: 2 * sector size
* the agfls of the ags containing the blocks: 2 * sector size
* the super block free block counter: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_write_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))));
}
/*
* In truncating a file we free up to two extents at once. We can modify:
* the inode being truncated: inode size
* the inode's bmap btree: (max depth + 1) * block size
* And the bmap_finish transaction can free the blocks and bmap blocks:
* the agf for each of the ags: 4 * sector size
* the agfl for each of the ags: 4 * sector size
* the super block to reflect the freed blocks: sector size
* worst case split in allocation btrees per extent assuming 4 extents:
* 4 exts * 2 trees * (2 * max depth - 1) * block size
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_itruncate_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) + 1,
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(9, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 4),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(5, 0) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(2 + XFS_IALLOC_BLOCKS(mp) +
mp->m_in_maxlevels, 0)));
}
/*
* In renaming a files we can modify:
* the four inodes involved: 4 * inode size
* the two directory btrees: 2 * (max depth + v2) * dir block size
* the two directory bmap btrees: 2 * max depth * block size
* And the bmap_finish transaction can free dir and bmap blocks (two sets
* of bmap blocks) giving:
* the agf for the ags in which the blocks live: 3 * sector size
* the agfl for the ags in which the blocks live: 3 * sector size
* the superblock for the free block count: sector size
* the allocation btrees: 3 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_rename_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(4, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(2 * XFS_DIROP_LOG_COUNT(mp),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(7, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 3),
XFS_FSB_TO_B(mp, 1))));
}
/*
* For creating a link to an inode:
* the parent directory inode: inode size
* the linked inode: inode size
* the directory btree could split: (max depth + v2) * dir block size
* the directory bmap btree could join or split: (max depth + v2) * blocksize
* And the bmap_finish transaction can free some bmap blocks giving:
* the agf for the ag in which the blocks live: sector size
* the agfl for the ag in which the blocks live: sector size
* the superblock for the free block count: sector size
* the allocation btrees: 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_link_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1))));
}
/*
* For removing a directory entry we can modify:
* the parent directory inode: inode size
* the removed inode: inode size
* the directory btree could join: (max depth + v2) * dir block size
* the directory bmap btree could join or split: (max depth + v2) * blocksize
* And the bmap_finish transaction can free the dir and bmap blocks giving:
* the agf for the ag in which the blocks live: 2 * sector size
* the agfl for the ag in which the blocks live: 2 * sector size
* the superblock for the free block count: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_remove_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))));
}
/*
* For create, break it in to the two cases that the transaction
* covers. We start with the modify case - allocation done by modification
* of the state of existing inodes - and the allocation case.
*/
/*
* For create we can modify:
* the parent directory inode: inode size
* the new inode: inode size
* the inode btree entry: block size
* the superblock for the nlink flag: sector size
* the directory btree: (max depth + v2) * dir block size
* the directory inode's bmap btree: (max depth + v2) * block size
*/
STATIC uint
xfs_calc_create_resv_modify(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
(uint)XFS_FSB_TO_B(mp, 1) +
xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1));
}
/*
* For create we can allocate some inodes giving:
* the agi and agf of the ag getting the new inodes: 2 * sectorsize
* the superblock for the nlink flag: sector size
* the inode blocks allocated: XFS_IALLOC_BLOCKS * blocksize
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_create_resv_alloc(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
mp->m_sb.sb_sectsize +
xfs_calc_buf_res(XFS_IALLOC_BLOCKS(mp), XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
STATIC uint
__xfs_calc_create_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX(xfs_calc_create_resv_alloc(mp),
xfs_calc_create_resv_modify(mp));
}
/*
* For icreate we can allocate some inodes giving:
* the agi and agf of the ag getting the new inodes: 2 * sectorsize
* the superblock for the nlink flag: sector size
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_icreate_resv_alloc(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
mp->m_sb.sb_sectsize +
xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
STATIC uint
xfs_calc_icreate_reservation(xfs_mount_t *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX(xfs_calc_icreate_resv_alloc(mp),
xfs_calc_create_resv_modify(mp));
}
STATIC uint
xfs_calc_create_reservation(
struct xfs_mount *mp)
{
if (xfs_sb_version_hascrc(&mp->m_sb))
return xfs_calc_icreate_reservation(mp);
return __xfs_calc_create_reservation(mp);
}
/*
* Making a new directory is the same as creating a new file.
*/
STATIC uint
xfs_calc_mkdir_reservation(
struct xfs_mount *mp)
{
return xfs_calc_create_reservation(mp);
}
/*
* Making a new symplink is the same as creating a new file, but
* with the added blocks for remote symlink data which can be up to 1kB in
* length (MAXPATHLEN).
*/
STATIC uint
xfs_calc_symlink_reservation(
struct xfs_mount *mp)
{
return xfs_calc_create_reservation(mp) +
xfs_calc_buf_res(1, MAXPATHLEN);
}
/*
* In freeing an inode we can modify:
* the inode being freed: inode size
* the super block free inode counter: sector size
* the agi hash list and counters: sector size
* the inode btree entry: block size
* the on disk inode before ours in the agi hash list: inode cluster size
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_ifree_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(1, XFS_FSB_TO_B(mp, 1)) +
MAX((__uint16_t)XFS_FSB_TO_B(mp, 1),
XFS_INODE_CLUSTER_SIZE(mp)) +
xfs_calc_buf_res(1, 0) +
xfs_calc_buf_res(2 + XFS_IALLOC_BLOCKS(mp) +
mp->m_in_maxlevels, 0) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* When only changing the inode we log the inode and possibly the superblock
* We also add a bit of slop for the transaction stuff.
*/
STATIC uint
xfs_calc_ichange_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
mp->m_sb.sb_inodesize +
mp->m_sb.sb_sectsize +
512;
}
/*
* Growing the data section of the filesystem.
* superblock
* agi and agf
* allocation btrees
*/
STATIC uint
xfs_calc_growdata_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* Growing the rt section of the filesystem.
* In the first set of transactions (ALLOC) we allocate space to the
* bitmap or summary files.
* superblock: sector size
* agf of the ag from which the extent is allocated: sector size
* bmap btree for bitmap/summary inode: max depth * blocksize
* bitmap/summary inode: inode size
* allocation btrees for 1 block alloc: 2 * (2 * maxdepth - 1) * blocksize
*/
STATIC uint
xfs_calc_growrtalloc_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* Growing the rt section of the filesystem.
* In the second set of transactions (ZERO) we zero the new metadata blocks.
* one bitmap/summary block: blocksize
*/
STATIC uint
xfs_calc_growrtzero_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_blocksize);
}
/*
* Growing the rt section of the filesystem.
* In the third set of transactions (FREE) we update metadata without
* allocating any new blocks.
* superblock: sector size
* bitmap inode: inode size
* summary inode: inode size
* one bitmap block: blocksize
* summary blocks: new summary size
*/
STATIC uint
xfs_calc_growrtfree_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(1, mp->m_sb.sb_blocksize) +
xfs_calc_buf_res(1, mp->m_rsumsize);
}
/*
* Logging the inode modification timestamp on a synchronous write.
* inode
*/
STATIC uint
xfs_calc_swrite_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_inodesize);
}
/*
* Logging the inode mode bits when writing a setuid/setgid file
* inode
*/
STATIC uint
xfs_calc_writeid_reservation(xfs_mount_t *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_inodesize);
}
/*
* Converting the inode from non-attributed to attributed.
* the inode being converted: inode size
* agf block and superblock (for block allocation)
* the new block (directory sized)
* bmap blocks for the new directory block
* allocation btrees
*/
STATIC uint
xfs_calc_addafork_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(1, mp->m_dirblksize) +
xfs_calc_buf_res(XFS_DAENTER_BMAP1B(mp, XFS_DATA_FORK) + 1,
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* Removing the attribute fork of a file
* the inode being truncated: inode size
* the inode's bmap btree: max depth * block size
* And the bmap_finish transaction can free the blocks and bmap blocks:
* the agf for each of the ags: 4 * sector size
* the agfl for each of the ags: 4 * sector size
* the super block to reflect the freed blocks: sector size
* worst case split in allocation btrees per extent assuming 4 extents:
* 4 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_attrinval_reservation(
struct xfs_mount *mp)
{
return MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(9, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 4),
XFS_FSB_TO_B(mp, 1))));
}
/*
* Setting an attribute at mount time.
* the inode getting the attribute
* the superblock for allocations
* the agfs extents are allocated from
* the attribute btree * max depth
* the inode allocation btree
* Since attribute transaction space is dependent on the size of the attribute,
* the calculation is done partially at mount time and partially at runtime(see
* below).
*/
STATIC uint
xfs_calc_attrsetm_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_DA_NODE_MAXDEPTH, XFS_FSB_TO_B(mp, 1));
}
/*
* Setting an attribute at runtime, transaction space unit per block.
* the superblock for allocations: sector size
* the inode bmap btree could join or split: max depth * block size
* Since the runtime attribute transaction space is dependent on the total
* blocks needed for the 1st bmap, here we calculate out the space unit for
* one block so that the caller could figure out the total space according
* to the attibute extent length in blocks by: ext * XFS_ATTRSETRT_LOG_RES(mp).
*/
STATIC uint
xfs_calc_attrsetrt_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK),
XFS_FSB_TO_B(mp, 1));
}
/*
* Removing an attribute.
* the inode: inode size
* the attribute btree could join: max depth * block size
* the inode bmap btree could join or split: max depth * block size
* And the bmap_finish transaction can free the attr blocks freed giving:
* the agf for the ag in which the blocks live: 2 * sector size
* the agfl for the ag in which the blocks live: 2 * sector size
* the superblock for the free block count: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_attrrm_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_DA_NODE_MAXDEPTH,
XFS_FSB_TO_B(mp, 1)) +
(uint)XFS_FSB_TO_B(mp,
XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK)) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK), 0)),
(xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))));
}
/*
* Clearing a bad agino number in an agi hash bucket.
*/
STATIC uint
xfs_calc_clear_agi_bucket_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/*
* Clearing the quotaflags in the superblock.
* the super block for changing quota flags: sector size
*/
STATIC uint
xfs_calc_qm_sbchange_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/*
* Adjusting quota limits.
* the xfs_disk_dquot_t: sizeof(struct xfs_disk_dquot)
*/
STATIC uint
xfs_calc_qm_setqlim_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, sizeof(struct xfs_disk_dquot));
}
/*
* Allocating quota on disk if needed.
* the write transaction log space: XFS_WRITE_LOG_RES(mp)
* the unit of quota allocation: one system block size
*/
STATIC uint
xfs_calc_qm_dqalloc_reservation(
struct xfs_mount *mp)
{
return XFS_WRITE_LOG_RES(mp) +
xfs_calc_buf_res(1,
XFS_FSB_TO_B(mp, XFS_DQUOT_CLUSTER_SIZE_FSB) - 1);
}
/*
* Turning off quotas.
* the xfs_qoff_logitem_t: sizeof(struct xfs_qoff_logitem) * 2
* the superblock for the quota flags: sector size
*/
STATIC uint
xfs_calc_qm_quotaoff_reservation(
struct xfs_mount *mp)
{
return sizeof(struct xfs_qoff_logitem) * 2 +
xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/*
* End of turning off quotas.
* the xfs_qoff_logitem_t: sizeof(struct xfs_qoff_logitem) * 2
*/
STATIC uint
xfs_calc_qm_quotaoff_end_reservation(
struct xfs_mount *mp)
{
return sizeof(struct xfs_qoff_logitem) * 2;
}
/*
* Syncing the incore super block changes to disk.
* the super block to reflect the changes: sector size
*/
STATIC uint
xfs_calc_sb_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/* /*
* Initialize the precomputed transaction reservation values * Initialize the precomputed transaction reservation values
* in the mount structure. * in the mount structure.
@ -679,36 +56,7 @@ void
xfs_trans_init( xfs_trans_init(
struct xfs_mount *mp) struct xfs_mount *mp)
{ {
struct xfs_trans_reservations *resp = &mp->m_reservations; xfs_trans_resv_calc(mp, &mp->m_reservations);
resp->tr_write = xfs_calc_write_reservation(mp);
resp->tr_itruncate = xfs_calc_itruncate_reservation(mp);
resp->tr_rename = xfs_calc_rename_reservation(mp);
resp->tr_link = xfs_calc_link_reservation(mp);
resp->tr_remove = xfs_calc_remove_reservation(mp);
resp->tr_symlink = xfs_calc_symlink_reservation(mp);
resp->tr_create = xfs_calc_create_reservation(mp);
resp->tr_mkdir = xfs_calc_mkdir_reservation(mp);
resp->tr_ifree = xfs_calc_ifree_reservation(mp);
resp->tr_ichange = xfs_calc_ichange_reservation(mp);
resp->tr_growdata = xfs_calc_growdata_reservation(mp);
resp->tr_swrite = xfs_calc_swrite_reservation(mp);
resp->tr_writeid = xfs_calc_writeid_reservation(mp);
resp->tr_addafork = xfs_calc_addafork_reservation(mp);
resp->tr_attrinval = xfs_calc_attrinval_reservation(mp);
resp->tr_attrsetm = xfs_calc_attrsetm_reservation(mp);
resp->tr_attrsetrt = xfs_calc_attrsetrt_reservation(mp);
resp->tr_attrrm = xfs_calc_attrrm_reservation(mp);
resp->tr_clearagi = xfs_calc_clear_agi_bucket_reservation(mp);
resp->tr_growrtalloc = xfs_calc_growrtalloc_reservation(mp);
resp->tr_growrtzero = xfs_calc_growrtzero_reservation(mp);
resp->tr_growrtfree = xfs_calc_growrtfree_reservation(mp);
resp->tr_qm_sbchange = xfs_calc_qm_sbchange_reservation(mp);
resp->tr_qm_setqlim = xfs_calc_qm_setqlim_reservation(mp);
resp->tr_qm_dqalloc = xfs_calc_qm_dqalloc_reservation(mp);
resp->tr_qm_quotaoff = xfs_calc_qm_quotaoff_reservation(mp);
resp->tr_qm_equotaoff = xfs_calc_qm_quotaoff_end_reservation(mp);
resp->tr_sb = xfs_calc_sb_reservation(mp);
} }
/* /*

View File

@ -20,80 +20,9 @@
struct xfs_log_item; struct xfs_log_item;
/* #include "xfs_trans_resv.h"
* Per-extent log reservation for the allocation btree changes
* involved in freeing or allocating an extent.
* 2 trees * (2 blocks/level * max depth - 1)
*/
#define XFS_ALLOCFREE_LOG_COUNT(mp,nx) \
((nx) * (2 * (2 * XFS_AG_MAXLEVELS(mp) - 1)))
/* /* kernel only transaction subsystem defines */
* Per-directory log reservation for any directory change.
* dir blocks: (1 btree block per level + data block + free block)
* bmap btree: (levels + 2) * max depth
* v2 directory blocks can be fragmented below the dirblksize down to the fsb
* size, so account for that in the DAENTER macros.
*/
#define XFS_DIROP_LOG_COUNT(mp) \
(XFS_DAENTER_BLOCKS(mp, XFS_DATA_FORK) + \
XFS_DAENTER_BMAPS(mp, XFS_DATA_FORK) + 1)
#define XFS_WRITE_LOG_RES(mp) ((mp)->m_reservations.tr_write)
#define XFS_ITRUNCATE_LOG_RES(mp) ((mp)->m_reservations.tr_itruncate)
#define XFS_RENAME_LOG_RES(mp) ((mp)->m_reservations.tr_rename)
#define XFS_LINK_LOG_RES(mp) ((mp)->m_reservations.tr_link)
#define XFS_REMOVE_LOG_RES(mp) ((mp)->m_reservations.tr_remove)
#define XFS_SYMLINK_LOG_RES(mp) ((mp)->m_reservations.tr_symlink)
#define XFS_CREATE_LOG_RES(mp) ((mp)->m_reservations.tr_create)
#define XFS_MKDIR_LOG_RES(mp) ((mp)->m_reservations.tr_mkdir)
#define XFS_IFREE_LOG_RES(mp) ((mp)->m_reservations.tr_ifree)
#define XFS_ICHANGE_LOG_RES(mp) ((mp)->m_reservations.tr_ichange)
#define XFS_GROWDATA_LOG_RES(mp) ((mp)->m_reservations.tr_growdata)
#define XFS_GROWRTALLOC_LOG_RES(mp) ((mp)->m_reservations.tr_growrtalloc)
#define XFS_GROWRTZERO_LOG_RES(mp) ((mp)->m_reservations.tr_growrtzero)
#define XFS_GROWRTFREE_LOG_RES(mp) ((mp)->m_reservations.tr_growrtfree)
#define XFS_SWRITE_LOG_RES(mp) ((mp)->m_reservations.tr_swrite)
/*
* Logging the inode timestamps on an fsync -- same as SWRITE
* as long as SWRITE logs the entire inode core
*/
#define XFS_FSYNC_TS_LOG_RES(mp) ((mp)->m_reservations.tr_swrite)
#define XFS_WRITEID_LOG_RES(mp) ((mp)->m_reservations.tr_swrite)
#define XFS_ADDAFORK_LOG_RES(mp) ((mp)->m_reservations.tr_addafork)
#define XFS_ATTRINVAL_LOG_RES(mp) ((mp)->m_reservations.tr_attrinval)
#define XFS_ATTRSETM_LOG_RES(mp) ((mp)->m_reservations.tr_attrsetm)
#define XFS_ATTRSETRT_LOG_RES(mp) ((mp)->m_reservations.tr_attrsetrt)
#define XFS_ATTRRM_LOG_RES(mp) ((mp)->m_reservations.tr_attrrm)
#define XFS_CLEAR_AGI_BUCKET_LOG_RES(mp) ((mp)->m_reservations.tr_clearagi)
#define XFS_QM_SBCHANGE_LOG_RES(mp) ((mp)->m_reservations.tr_qm_sbchange)
#define XFS_QM_SETQLIM_LOG_RES(mp) ((mp)->m_reservations.tr_qm_setqlim)
#define XFS_QM_DQALLOC_LOG_RES(mp) ((mp)->m_reservations.tr_qm_dqalloc)
#define XFS_QM_QUOTAOFF_LOG_RES(mp) ((mp)->m_reservations.tr_qm_quotaoff)
#define XFS_QM_QUOTAOFF_END_LOG_RES(mp) ((mp)->m_reservations.tr_qm_equotaoff)
#define XFS_SB_LOG_RES(mp) ((mp)->m_reservations.tr_sb)
/*
* Various log count values.
*/
#define XFS_DEFAULT_LOG_COUNT 1
#define XFS_DEFAULT_PERM_LOG_COUNT 2
#define XFS_ITRUNCATE_LOG_COUNT 2
#define XFS_INACTIVE_LOG_COUNT 2
#define XFS_CREATE_LOG_COUNT 2
#define XFS_MKDIR_LOG_COUNT 3
#define XFS_SYMLINK_LOG_COUNT 3
#define XFS_REMOVE_LOG_COUNT 2
#define XFS_LINK_LOG_COUNT 2
#define XFS_RENAME_LOG_COUNT 2
#define XFS_WRITE_LOG_COUNT 2
#define XFS_ADDAFORK_LOG_COUNT 2
#define XFS_ATTRINVAL_LOG_COUNT 1
#define XFS_ATTRSET_LOG_COUNT 3
#define XFS_ATTRRM_LOG_COUNT 3
#ifdef __KERNEL__
struct xfs_buf; struct xfs_buf;
struct xfs_buftarg; struct xfs_buftarg;
@ -323,6 +252,4 @@ void xfs_trans_ail_destroy(struct xfs_mount *);
extern kmem_zone_t *xfs_trans_zone; extern kmem_zone_t *xfs_trans_zone;
extern kmem_zone_t *xfs_log_item_desc_zone; extern kmem_zone_t *xfs_log_item_desc_zone;
#endif /* __KERNEL__ */
#endif /* __XFS_TRANS_H__ */ #endif /* __XFS_TRANS_H__ */

701
fs/xfs/xfs_trans_resv.c Normal file
View File

@ -0,0 +1,701 @@
/*
* Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
* Copyright (C) 2010 Red Hat, Inc.
* All Rights Reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it would be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_format.h"
#include "xfs_log.h"
#include "xfs_trans_resv.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_extent_busy.h"
#include "xfs_bmap.h"
#include "xfs_quota.h"
#include "xfs_qm.h"
#include "xfs_trans_space.h"
#include "xfs_trace.h"
/*
* A buffer has a format structure overhead in the log in addition
* to the data, so we need to take this into account when reserving
* space in a transaction for a buffer. Round the space required up
* to a multiple of 128 bytes so that we don't change the historical
* reservation that has been used for this overhead.
*/
STATIC uint
xfs_buf_log_overhead(void)
{
return round_up(sizeof(struct xlog_op_header) +
sizeof(struct xfs_buf_log_format), 128);
}
/*
* Calculate out transaction log reservation per item in bytes.
*
* The nbufs argument is used to indicate the number of items that
* will be changed in a transaction. size is used to tell how many
* bytes should be reserved per item.
*/
STATIC uint
xfs_calc_buf_res(
uint nbufs,
uint size)
{
return nbufs * (size + xfs_buf_log_overhead());
}
/*
* Various log reservation values.
*
* These are based on the size of the file system block because that is what
* most transactions manipulate. Each adds in an additional 128 bytes per
* item logged to try to account for the overhead of the transaction mechanism.
*
* Note: Most of the reservations underestimate the number of allocation
* groups into which they could free extents in the xfs_bmap_finish() call.
* This is because the number in the worst case is quite high and quite
* unusual. In order to fix this we need to change xfs_bmap_finish() to free
* extents in only a single AG at a time. This will require changes to the
* EFI code as well, however, so that the EFI for the extents not freed is
* logged again in each transaction. See SGI PV #261917.
*
* Reservation functions here avoid a huge stack in xfs_trans_init due to
* register overflow from temporaries in the calculations.
*/
/*
* In a write transaction we can allocate a maximum of 2
* extents. This gives:
* the inode getting the new extents: inode size
* the inode's bmap btree: max depth * block size
* the agfs of the ags from which the extents are allocated: 2 * sector
* the superblock free block counter: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
* And the bmap_finish transaction can free bmap blocks in a join:
* the agfs of the ags containing the blocks: 2 * sector size
* the agfls of the ags containing the blocks: 2 * sector size
* the super block free block counter: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_write_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))));
}
/*
* In truncating a file we free up to two extents at once. We can modify:
* the inode being truncated: inode size
* the inode's bmap btree: (max depth + 1) * block size
* And the bmap_finish transaction can free the blocks and bmap blocks:
* the agf for each of the ags: 4 * sector size
* the agfl for each of the ags: 4 * sector size
* the super block to reflect the freed blocks: sector size
* worst case split in allocation btrees per extent assuming 4 extents:
* 4 exts * 2 trees * (2 * max depth - 1) * block size
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_itruncate_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) + 1,
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(9, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 4),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(5, 0) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(2 + XFS_IALLOC_BLOCKS(mp) +
mp->m_in_maxlevels, 0)));
}
/*
* In renaming a files we can modify:
* the four inodes involved: 4 * inode size
* the two directory btrees: 2 * (max depth + v2) * dir block size
* the two directory bmap btrees: 2 * max depth * block size
* And the bmap_finish transaction can free dir and bmap blocks (two sets
* of bmap blocks) giving:
* the agf for the ags in which the blocks live: 3 * sector size
* the agfl for the ags in which the blocks live: 3 * sector size
* the superblock for the free block count: sector size
* the allocation btrees: 3 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_rename_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(4, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(2 * XFS_DIROP_LOG_COUNT(mp),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(7, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 3),
XFS_FSB_TO_B(mp, 1))));
}
/*
* For creating a link to an inode:
* the parent directory inode: inode size
* the linked inode: inode size
* the directory btree could split: (max depth + v2) * dir block size
* the directory bmap btree could join or split: (max depth + v2) * blocksize
* And the bmap_finish transaction can free some bmap blocks giving:
* the agf for the ag in which the blocks live: sector size
* the agfl for the ag in which the blocks live: sector size
* the superblock for the free block count: sector size
* the allocation btrees: 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_link_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1))));
}
/*
* For removing a directory entry we can modify:
* the parent directory inode: inode size
* the removed inode: inode size
* the directory btree could join: (max depth + v2) * dir block size
* the directory bmap btree could join or split: (max depth + v2) * blocksize
* And the bmap_finish transaction can free the dir and bmap blocks giving:
* the agf for the ag in which the blocks live: 2 * sector size
* the agfl for the ag in which the blocks live: 2 * sector size
* the superblock for the free block count: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_remove_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))));
}
/*
* For create, break it in to the two cases that the transaction
* covers. We start with the modify case - allocation done by modification
* of the state of existing inodes - and the allocation case.
*/
/*
* For create we can modify:
* the parent directory inode: inode size
* the new inode: inode size
* the inode btree entry: block size
* the superblock for the nlink flag: sector size
* the directory btree: (max depth + v2) * dir block size
* the directory inode's bmap btree: (max depth + v2) * block size
*/
STATIC uint
xfs_calc_create_resv_modify(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
(uint)XFS_FSB_TO_B(mp, 1) +
xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1));
}
/*
* For create we can allocate some inodes giving:
* the agi and agf of the ag getting the new inodes: 2 * sectorsize
* the superblock for the nlink flag: sector size
* the inode blocks allocated: XFS_IALLOC_BLOCKS * blocksize
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_create_resv_alloc(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
mp->m_sb.sb_sectsize +
xfs_calc_buf_res(XFS_IALLOC_BLOCKS(mp), XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
STATIC uint
__xfs_calc_create_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX(xfs_calc_create_resv_alloc(mp),
xfs_calc_create_resv_modify(mp));
}
/*
* For icreate we can allocate some inodes giving:
* the agi and agf of the ag getting the new inodes: 2 * sectorsize
* the superblock for the nlink flag: sector size
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_icreate_resv_alloc(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
mp->m_sb.sb_sectsize +
xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
STATIC uint
xfs_calc_icreate_reservation(xfs_mount_t *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX(xfs_calc_icreate_resv_alloc(mp),
xfs_calc_create_resv_modify(mp));
}
STATIC uint
xfs_calc_create_reservation(
struct xfs_mount *mp)
{
if (xfs_sb_version_hascrc(&mp->m_sb))
return xfs_calc_icreate_reservation(mp);
return __xfs_calc_create_reservation(mp);
}
/*
* Making a new directory is the same as creating a new file.
*/
STATIC uint
xfs_calc_mkdir_reservation(
struct xfs_mount *mp)
{
return xfs_calc_create_reservation(mp);
}
/*
* Making a new symplink is the same as creating a new file, but
* with the added blocks for remote symlink data which can be up to 1kB in
* length (MAXPATHLEN).
*/
STATIC uint
xfs_calc_symlink_reservation(
struct xfs_mount *mp)
{
return xfs_calc_create_reservation(mp) +
xfs_calc_buf_res(1, MAXPATHLEN);
}
/*
* In freeing an inode we can modify:
* the inode being freed: inode size
* the super block free inode counter: sector size
* the agi hash list and counters: sector size
* the inode btree entry: block size
* the on disk inode before ours in the agi hash list: inode cluster size
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
STATIC uint
xfs_calc_ifree_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(1, XFS_FSB_TO_B(mp, 1)) +
MAX((__uint16_t)XFS_FSB_TO_B(mp, 1),
XFS_INODE_CLUSTER_SIZE(mp)) +
xfs_calc_buf_res(1, 0) +
xfs_calc_buf_res(2 + XFS_IALLOC_BLOCKS(mp) +
mp->m_in_maxlevels, 0) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* When only changing the inode we log the inode and possibly the superblock
* We also add a bit of slop for the transaction stuff.
*/
STATIC uint
xfs_calc_ichange_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
mp->m_sb.sb_inodesize +
mp->m_sb.sb_sectsize +
512;
}
/*
* Growing the data section of the filesystem.
* superblock
* agi and agf
* allocation btrees
*/
STATIC uint
xfs_calc_growdata_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* Growing the rt section of the filesystem.
* In the first set of transactions (ALLOC) we allocate space to the
* bitmap or summary files.
* superblock: sector size
* agf of the ag from which the extent is allocated: sector size
* bmap btree for bitmap/summary inode: max depth * blocksize
* bitmap/summary inode: inode size
* allocation btrees for 1 block alloc: 2 * (2 * maxdepth - 1) * blocksize
*/
STATIC uint
xfs_calc_growrtalloc_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK),
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* Growing the rt section of the filesystem.
* In the second set of transactions (ZERO) we zero the new metadata blocks.
* one bitmap/summary block: blocksize
*/
STATIC uint
xfs_calc_growrtzero_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_blocksize);
}
/*
* Growing the rt section of the filesystem.
* In the third set of transactions (FREE) we update metadata without
* allocating any new blocks.
* superblock: sector size
* bitmap inode: inode size
* summary inode: inode size
* one bitmap block: blocksize
* summary blocks: new summary size
*/
STATIC uint
xfs_calc_growrtfree_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(2, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(1, mp->m_sb.sb_blocksize) +
xfs_calc_buf_res(1, mp->m_rsumsize);
}
/*
* Logging the inode modification timestamp on a synchronous write.
* inode
*/
STATIC uint
xfs_calc_swrite_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_inodesize);
}
/*
* Logging the inode mode bits when writing a setuid/setgid file
* inode
*/
STATIC uint
xfs_calc_writeid_reservation(xfs_mount_t *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_inodesize);
}
/*
* Converting the inode from non-attributed to attributed.
* the inode being converted: inode size
* agf block and superblock (for block allocation)
* the new block (directory sized)
* bmap blocks for the new directory block
* allocation btrees
*/
STATIC uint
xfs_calc_addafork_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(1, mp->m_dirblksize) +
xfs_calc_buf_res(XFS_DAENTER_BMAP1B(mp, XFS_DATA_FORK) + 1,
XFS_FSB_TO_B(mp, 1)) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
}
/*
* Removing the attribute fork of a file
* the inode being truncated: inode size
* the inode's bmap btree: max depth * block size
* And the bmap_finish transaction can free the blocks and bmap blocks:
* the agf for each of the ags: 4 * sector size
* the agfl for each of the ags: 4 * sector size
* the super block to reflect the freed blocks: sector size
* worst case split in allocation btrees per extent assuming 4 extents:
* 4 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_attrinval_reservation(
struct xfs_mount *mp)
{
return MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK),
XFS_FSB_TO_B(mp, 1))),
(xfs_calc_buf_res(9, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 4),
XFS_FSB_TO_B(mp, 1))));
}
/*
* Setting an attribute at mount time.
* the inode getting the attribute
* the superblock for allocations
* the agfs extents are allocated from
* the attribute btree * max depth
* the inode allocation btree
* Since attribute transaction space is dependent on the size of the attribute,
* the calculation is done partially at mount time and partially at runtime(see
* below).
*/
STATIC uint
xfs_calc_attrsetm_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_DA_NODE_MAXDEPTH, XFS_FSB_TO_B(mp, 1));
}
/*
* Setting an attribute at runtime, transaction space unit per block.
* the superblock for allocations: sector size
* the inode bmap btree could join or split: max depth * block size
* Since the runtime attribute transaction space is dependent on the total
* blocks needed for the 1st bmap, here we calculate out the space unit for
* one block so that the caller could figure out the total space according
* to the attibute extent length in blocks by: ext * XFS_ATTRSETRT_LOG_RES(mp).
*/
STATIC uint
xfs_calc_attrsetrt_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK),
XFS_FSB_TO_B(mp, 1));
}
/*
* Removing an attribute.
* the inode: inode size
* the attribute btree could join: max depth * block size
* the inode bmap btree could join or split: max depth * block size
* And the bmap_finish transaction can free the attr blocks freed giving:
* the agf for the ag in which the blocks live: 2 * sector size
* the agfl for the ag in which the blocks live: 2 * sector size
* the superblock for the free block count: sector size
* the allocation btrees: 2 exts * 2 trees * (2 * max depth - 1) * block size
*/
STATIC uint
xfs_calc_attrrm_reservation(
struct xfs_mount *mp)
{
return XFS_DQUOT_LOGRES(mp) +
MAX((xfs_calc_buf_res(1, mp->m_sb.sb_inodesize) +
xfs_calc_buf_res(XFS_DA_NODE_MAXDEPTH,
XFS_FSB_TO_B(mp, 1)) +
(uint)XFS_FSB_TO_B(mp,
XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK)) +
xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK), 0)),
(xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2),
XFS_FSB_TO_B(mp, 1))));
}
/*
* Clearing a bad agino number in an agi hash bucket.
*/
STATIC uint
xfs_calc_clear_agi_bucket_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/*
* Clearing the quotaflags in the superblock.
* the super block for changing quota flags: sector size
*/
STATIC uint
xfs_calc_qm_sbchange_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/*
* Adjusting quota limits.
* the xfs_disk_dquot_t: sizeof(struct xfs_disk_dquot)
*/
STATIC uint
xfs_calc_qm_setqlim_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, sizeof(struct xfs_disk_dquot));
}
/*
* Allocating quota on disk if needed.
* the write transaction log space: XFS_WRITE_LOG_RES(mp)
* the unit of quota allocation: one system block size
*/
STATIC uint
xfs_calc_qm_dqalloc_reservation(
struct xfs_mount *mp)
{
return XFS_WRITE_LOG_RES(mp) +
xfs_calc_buf_res(1,
XFS_FSB_TO_B(mp, XFS_DQUOT_CLUSTER_SIZE_FSB) - 1);
}
/*
* Turning off quotas.
* the xfs_qoff_logitem_t: sizeof(struct xfs_qoff_logitem) * 2
* the superblock for the quota flags: sector size
*/
STATIC uint
xfs_calc_qm_quotaoff_reservation(
struct xfs_mount *mp)
{
return sizeof(struct xfs_qoff_logitem) * 2 +
xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
/*
* End of turning off quotas.
* the xfs_qoff_logitem_t: sizeof(struct xfs_qoff_logitem) * 2
*/
STATIC uint
xfs_calc_qm_quotaoff_end_reservation(
struct xfs_mount *mp)
{
return sizeof(struct xfs_qoff_logitem) * 2;
}
/*
* Syncing the incore super block changes to disk.
* the super block to reflect the changes: sector size
*/
STATIC uint
xfs_calc_sb_reservation(
struct xfs_mount *mp)
{
return xfs_calc_buf_res(1, mp->m_sb.sb_sectsize);
}
void
xfs_trans_resv_calc(
struct xfs_mount *mp,
struct xfs_trans_resv *resp)
{
resp->tr_write = xfs_calc_write_reservation(mp);
resp->tr_itruncate = xfs_calc_itruncate_reservation(mp);
resp->tr_rename = xfs_calc_rename_reservation(mp);
resp->tr_link = xfs_calc_link_reservation(mp);
resp->tr_remove = xfs_calc_remove_reservation(mp);
resp->tr_symlink = xfs_calc_symlink_reservation(mp);
resp->tr_create = xfs_calc_create_reservation(mp);
resp->tr_mkdir = xfs_calc_mkdir_reservation(mp);
resp->tr_ifree = xfs_calc_ifree_reservation(mp);
resp->tr_ichange = xfs_calc_ichange_reservation(mp);
resp->tr_growdata = xfs_calc_growdata_reservation(mp);
resp->tr_swrite = xfs_calc_swrite_reservation(mp);
resp->tr_writeid = xfs_calc_writeid_reservation(mp);
resp->tr_addafork = xfs_calc_addafork_reservation(mp);
resp->tr_attrinval = xfs_calc_attrinval_reservation(mp);
resp->tr_attrsetm = xfs_calc_attrsetm_reservation(mp);
resp->tr_attrsetrt = xfs_calc_attrsetrt_reservation(mp);
resp->tr_attrrm = xfs_calc_attrrm_reservation(mp);
resp->tr_clearagi = xfs_calc_clear_agi_bucket_reservation(mp);
resp->tr_growrtalloc = xfs_calc_growrtalloc_reservation(mp);
resp->tr_growrtzero = xfs_calc_growrtzero_reservation(mp);
resp->tr_growrtfree = xfs_calc_growrtfree_reservation(mp);
resp->tr_qm_sbchange = xfs_calc_qm_sbchange_reservation(mp);
resp->tr_qm_setqlim = xfs_calc_qm_setqlim_reservation(mp);
resp->tr_qm_dqalloc = xfs_calc_qm_dqalloc_reservation(mp);
resp->tr_qm_quotaoff = xfs_calc_qm_quotaoff_reservation(mp);
resp->tr_qm_equotaoff = xfs_calc_qm_quotaoff_end_reservation(mp);
resp->tr_sb = xfs_calc_sb_reservation(mp);
}

137
fs/xfs/xfs_trans_resv.h Normal file
View File

@ -0,0 +1,137 @@
/*
* Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
* All Rights Reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it would be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __XFS_TRANS_RESV_H__
#define __XFS_TRANS_RESV_H__
struct xfs_mount;
/*
* structure for maintaining pre-calculated transaction reservations.
*/
struct xfs_trans_resv {
uint tr_write; /* extent alloc trans */
uint tr_itruncate; /* truncate trans */
uint tr_rename; /* rename trans */
uint tr_link; /* link trans */
uint tr_remove; /* unlink trans */
uint tr_symlink; /* symlink trans */
uint tr_create; /* create trans */
uint tr_mkdir; /* mkdir trans */
uint tr_ifree; /* inode free trans */
uint tr_ichange; /* inode update trans */
uint tr_growdata; /* fs data section grow trans */
uint tr_swrite; /* sync write inode trans */
uint tr_addafork; /* cvt inode to attributed trans */
uint tr_writeid; /* write setuid/setgid file */
uint tr_attrinval; /* attr fork buffer invalidation */
uint tr_attrsetm; /* set/create an attribute at mount time */
uint tr_attrsetrt; /* set/create an attribute at runtime */
uint tr_attrrm; /* remove an attribute */
uint tr_clearagi; /* clear bad agi unlinked ino bucket */
uint tr_growrtalloc; /* grow realtime allocations */
uint tr_growrtzero; /* grow realtime zeroing */
uint tr_growrtfree; /* grow realtime freeing */
uint tr_qm_sbchange; /* change quota flags */
uint tr_qm_setqlim; /* adjust quota limits */
uint tr_qm_dqalloc; /* allocate quota on disk */
uint tr_qm_quotaoff; /* turn quota off */
uint tr_qm_equotaoff;/* end of turn quota off */
uint tr_sb; /* modify superblock */
};
/*
* Per-extent log reservation for the allocation btree changes
* involved in freeing or allocating an extent.
* 2 trees * (2 blocks/level * max depth - 1) * block size
*/
#define XFS_ALLOCFREE_LOG_RES(mp,nx) \
((nx) * (2 * XFS_FSB_TO_B((mp), 2 * XFS_AG_MAXLEVELS(mp) - 1)))
#define XFS_ALLOCFREE_LOG_COUNT(mp,nx) \
((nx) * (2 * (2 * XFS_AG_MAXLEVELS(mp) - 1)))
/*
* Per-directory log reservation for any directory change.
* dir blocks: (1 btree block per level + data block + free block) * dblock size
* bmap btree: (levels + 2) * max depth * block size
* v2 directory blocks can be fragmented below the dirblksize down to the fsb
* size, so account for that in the DAENTER macros.
*/
#define XFS_DIROP_LOG_RES(mp) \
(XFS_FSB_TO_B(mp, XFS_DAENTER_BLOCKS(mp, XFS_DATA_FORK)) + \
(XFS_FSB_TO_B(mp, XFS_DAENTER_BMAPS(mp, XFS_DATA_FORK) + 1)))
#define XFS_DIROP_LOG_COUNT(mp) \
(XFS_DAENTER_BLOCKS(mp, XFS_DATA_FORK) + \
XFS_DAENTER_BMAPS(mp, XFS_DATA_FORK) + 1)
#define XFS_WRITE_LOG_RES(mp) ((mp)->m_reservations.tr_write)
#define XFS_ITRUNCATE_LOG_RES(mp) ((mp)->m_reservations.tr_itruncate)
#define XFS_RENAME_LOG_RES(mp) ((mp)->m_reservations.tr_rename)
#define XFS_LINK_LOG_RES(mp) ((mp)->m_reservations.tr_link)
#define XFS_REMOVE_LOG_RES(mp) ((mp)->m_reservations.tr_remove)
#define XFS_SYMLINK_LOG_RES(mp) ((mp)->m_reservations.tr_symlink)
#define XFS_CREATE_LOG_RES(mp) ((mp)->m_reservations.tr_create)
#define XFS_MKDIR_LOG_RES(mp) ((mp)->m_reservations.tr_mkdir)
#define XFS_IFREE_LOG_RES(mp) ((mp)->m_reservations.tr_ifree)
#define XFS_ICHANGE_LOG_RES(mp) ((mp)->m_reservations.tr_ichange)
#define XFS_GROWDATA_LOG_RES(mp) ((mp)->m_reservations.tr_growdata)
#define XFS_GROWRTALLOC_LOG_RES(mp) ((mp)->m_reservations.tr_growrtalloc)
#define XFS_GROWRTZERO_LOG_RES(mp) ((mp)->m_reservations.tr_growrtzero)
#define XFS_GROWRTFREE_LOG_RES(mp) ((mp)->m_reservations.tr_growrtfree)
#define XFS_SWRITE_LOG_RES(mp) ((mp)->m_reservations.tr_swrite)
/*
* Logging the inode timestamps on an fsync -- same as SWRITE
* as long as SWRITE logs the entire inode core
*/
#define XFS_FSYNC_TS_LOG_RES(mp) ((mp)->m_reservations.tr_swrite)
#define XFS_WRITEID_LOG_RES(mp) ((mp)->m_reservations.tr_swrite)
#define XFS_ADDAFORK_LOG_RES(mp) ((mp)->m_reservations.tr_addafork)
#define XFS_ATTRINVAL_LOG_RES(mp) ((mp)->m_reservations.tr_attrinval)
#define XFS_ATTRSETM_LOG_RES(mp) ((mp)->m_reservations.tr_attrsetm)
#define XFS_ATTRSETRT_LOG_RES(mp) ((mp)->m_reservations.tr_attrsetrt)
#define XFS_ATTRRM_LOG_RES(mp) ((mp)->m_reservations.tr_attrrm)
#define XFS_CLEAR_AGI_BUCKET_LOG_RES(mp) ((mp)->m_reservations.tr_clearagi)
#define XFS_QM_SBCHANGE_LOG_RES(mp) ((mp)->m_reservations.tr_qm_sbchange)
#define XFS_QM_SETQLIM_LOG_RES(mp) ((mp)->m_reservations.tr_qm_setqlim)
#define XFS_QM_DQALLOC_LOG_RES(mp) ((mp)->m_reservations.tr_qm_dqalloc)
#define XFS_QM_QUOTAOFF_LOG_RES(mp) ((mp)->m_reservations.tr_qm_quotaoff)
#define XFS_QM_QUOTAOFF_END_LOG_RES(mp) ((mp)->m_reservations.tr_qm_equotaoff)
#define XFS_SB_LOG_RES(mp) ((mp)->m_reservations.tr_sb)
/*
* Various log count values.
*/
#define XFS_DEFAULT_LOG_COUNT 1
#define XFS_DEFAULT_PERM_LOG_COUNT 2
#define XFS_ITRUNCATE_LOG_COUNT 2
#define XFS_INACTIVE_LOG_COUNT 2
#define XFS_CREATE_LOG_COUNT 2
#define XFS_MKDIR_LOG_COUNT 3
#define XFS_SYMLINK_LOG_COUNT 3
#define XFS_REMOVE_LOG_COUNT 2
#define XFS_LINK_LOG_COUNT 2
#define XFS_RENAME_LOG_COUNT 2
#define XFS_WRITE_LOG_COUNT 2
#define XFS_ADDAFORK_LOG_COUNT 2
#define XFS_ATTRINVAL_LOG_COUNT 1
#define XFS_ATTRSET_LOG_COUNT 3
#define XFS_ATTRRM_LOG_COUNT 3
void xfs_trans_resv_calc(struct xfs_mount *mp, struct xfs_trans_resv *resp);
#endif /* __XFS_TRANS_RESV_H__ */