/*
- * Copyright (C) International Business Machines Corp., 2000-2004
+ * Copyright (C) International Business Machines Corp., 2000-2005
* Portions Copyright (C) Christoph Hellwig, 2001-2002
*
* This program is free software; you can redistribute it and/or modify
} TxStat;
#endif
-static int nTxBlock = 512; /* number of transaction blocks */
+static int nTxBlock = -1; /* number of transaction blocks */
module_param(nTxBlock, int, 0);
MODULE_PARM_DESC(nTxBlock,
- "Number of transaction blocks (default:512, max:65536)");
+ "Number of transaction blocks (max:65536)");
-static int nTxLock = 4096; /* number of transaction locks */
+static int nTxLock = -1; /* number of transaction locks */
module_param(nTxLock, int, 0);
MODULE_PARM_DESC(nTxLock,
- "Number of transaction locks (default:4096, max:65536)");
+ "Number of transaction locks (max:65536)");
struct tblock *TxBlock; /* transaction block table */
static int TxLockLWM; /* Low water mark for number of txLocks used */
DECLARE_WAIT_QUEUE_HEAD(jfs_sync_thread_wait);
DECLARE_WAIT_QUEUE_HEAD(jfs_commit_thread_wait);
+static int jfs_commit_thread_waking;
/*
* Retry logic exist outside these macros to protect from spurrious wakeups.
static void txLockFree(lid_t lid)
{
+ TxLock[lid].tid = 0;
TxLock[lid].next = TxAnchor.freelock;
TxAnchor.freelock = lid;
TxAnchor.tlocksInUse--;
int txInit(void)
{
int k, size;
+ struct sysinfo si;
+
+ /* Set defaults for nTxLock and nTxBlock if unset */
+
+ if (nTxLock == -1) {
+ if (nTxBlock == -1) {
+ /* Base default on memory size */
+ si_meminfo(&si);
+ if (si.totalram > (256 * 1024)) /* 1 GB */
+ nTxLock = 64 * 1024;
+ else
+ nTxLock = si.totalram >> 2;
+ } else if (nTxBlock > (8 * 1024))
+ nTxLock = 64 * 1024;
+ else
+ nTxLock = nTxBlock << 3;
+ }
+ if (nTxBlock == -1)
+ nTxBlock = nTxLock >> 3;
/* Verify tunable parameters */
if (nTxBlock < 16)
nTxLock = 256; /* No one should set it this low */
if (nTxLock > 65536)
nTxLock = 65536;
+
+ printk(KERN_INFO "JFS: nTxBlock = %d, nTxLock = %d\n",
+ nTxBlock, nTxLock);
/*
* initialize transaction block (tblock) table
*
* tid = 0 is reserved.
*/
TxLockLWM = (nTxLock * 4) / 10;
- TxLockHWM = (nTxLock * 8) / 10;
- TxLockVHWM = (nTxLock * 9) / 10;
+ TxLockHWM = (nTxLock * 7) / 10;
+ TxLockVHWM = (nTxLock * 8) / 10;
size = sizeof(struct tblock) * nTxBlock;
TxBlock = (struct tblock *) vmalloc(size);
* synchronize with logsync barrier
*/
if (test_bit(log_SYNCBARRIER, &log->flag)) {
- /* forward log syncpt */
- /* lmSync(log); */
-
jfs_info("log barrier off: 0x%x", log->lsn);
/* enable new transactions start */
/* wakeup all waitors for logsync barrier */
TXN_WAKEUP(&log->syncwait);
+
+ TXN_UNLOCK();
+
+ /* forward log syncpt */
+ jfs_syncpt(log);
+
+ goto wakeup;
}
}
+ TXN_UNLOCK();
+wakeup:
/*
* wakeup all waitors for a free tblock
*/
TXN_WAKEUP(&TxAnchor.freewait);
-
- TXN_UNLOCK();
}
/* is page locked by the requester transaction ? */
tlck = lid_to_tlock(lid);
- if ((xtid = tlck->tid) == tid)
+ if ((xtid = tlck->tid) == tid) {
+ TXN_UNLOCK();
goto grantLock;
+ }
/*
* is page locked by anonymous transaction/lock ?
*/
if (xtid == 0) {
tlck->tid = tid;
+ TXN_UNLOCK();
tblk = tid_to_tblock(tid);
/*
* The order of the tlocks in the transaction is important
*/
tlck->tid = tid;
+ TXN_UNLOCK();
+
/* mark tlock for meta-data page */
if (mp->xflag & COMMIT_PAGE) {
tlck->flag = tlckPAGELOCK;
/* mark the page dirty and nohomeok */
- mark_metapage_dirty(mp);
- atomic_inc(&mp->nohomeok);
+ metapage_nohomeok(mp);
jfs_info("locking mp = 0x%p, nohomeok = %d tid = %d tlck = 0x%p",
- mp, atomic_read(&mp->nohomeok), tid, tlck);
+ mp, mp->nohomeok, tid, tlck);
/* if anonymous transaction, and buffer is on the group
* commit synclist, mark inode to show this. This will
if (tlck->next == 0) {
/* This inode's first anonymous transaction */
jfs_ip->atltail = lid;
+ TXN_LOCK();
list_add_tail(&jfs_ip->anon_inode_list,
&TxAnchor.anon_list);
+ TXN_UNLOCK();
}
}
grantLock:
tlck->type |= type;
- TXN_UNLOCK();
-
return tlck;
/*
BUG();
}
INCREMENT(stattx.waitlock); /* statistics */
+ TXN_UNLOCK();
release_metapage(mp);
+ TXN_LOCK();
+ xtid = tlck->tid; /* reaquire after dropping TXN_LOCK */
jfs_info("txLock: in waitLock, tid = %d, xtid = %d, lid = %d",
tid, xtid, lid);
- TXN_SLEEP_DROP_LOCK(&tid_to_tblock(xtid)->waitor);
+
+ /* Recheck everything since dropping TXN_LOCK */
+ if (xtid && (tlck->mp == mp) && (mp->lid == lid))
+ TXN_SLEEP_DROP_LOCK(&tid_to_tblock(xtid)->waitor);
+ else
+ TXN_UNLOCK();
jfs_info("txLock: awakened tid = %d, lid = %d", tid, lid);
return NULL;
struct metapage *mp;
struct jfs_log *log;
int difft, diffp;
+ unsigned long flags;
jfs_info("txUnlock: tblk = 0x%p", tblk);
log = JFS_SBI(tblk->sb)->log;
assert(mp->xflag & COMMIT_PAGE);
/* hold buffer
- *
- * It's possible that someone else has the metapage.
- * The only things were changing are nohomeok, which
- * is handled atomically, and clsn which is protected
- * by the LOGSYNC_LOCK.
*/
- hold_metapage(mp, 1);
+ hold_metapage(mp);
- assert(atomic_read(&mp->nohomeok) > 0);
- atomic_dec(&mp->nohomeok);
+ assert(mp->nohomeok > 0);
+ _metapage_homeok(mp);
/* inherit younger/larger clsn */
- LOGSYNC_LOCK(log);
+ LOGSYNC_LOCK(log, flags);
if (mp->clsn) {
logdiff(difft, tblk->clsn, log);
logdiff(diffp, mp->clsn, log);
mp->clsn = tblk->clsn;
} else
mp->clsn = tblk->clsn;
- LOGSYNC_UNLOCK(log);
+ LOGSYNC_UNLOCK(log, flags);
assert(!(tlck->flag & tlckFREEPAGE));
- if (tlck->flag & tlckWRITEPAGE) {
- write_metapage(mp);
- } else {
- /* release page which has been forced */
- release_metapage(mp);
- }
+ put_metapage(mp);
}
/* insert tlock, and linelock(s) of the tlock if any,
* has been inserted in logsync list at txUpdateMap())
*/
if (tblk->lsn) {
- LOGSYNC_LOCK(log);
+ LOGSYNC_LOCK(log, flags);
log->count--;
list_del(&tblk->synclist);
- LOGSYNC_UNLOCK(log);
+ LOGSYNC_UNLOCK(log, flags);
}
}
/* log after-image for logredo(): */
lrd->type = cpu_to_le16(LOG_REDOPAGE);
- if (JFS_IP(tlck->ip)->next_index < MAX_INLINE_DIRTABLE_ENTRY) {
+ if (jfs_dirtable_inline(tlck->ip)) {
/*
* The table has been truncated, we've must have deleted
* the last entry, so don't bother logging this
*/
mp->lid = 0;
- hold_metapage(mp, 0);
- atomic_dec(&mp->nohomeok);
+ grab_metapage(mp);
+ metapage_homeok(mp);
discard_metapage(mp);
tlck->mp = NULL;
return 0;
struct maplock *maplock;
struct xdlistlock *xadlock;
struct pxd_lock *pxdlock;
- pxd_t *pxd;
+ pxd_t *page_pxd;
int next, lwm, hwm;
ip = tlck->ip;
lrd->log.redopage.type = cpu_to_le16(LOG_XTREE);
lrd->log.redopage.l2linesize = cpu_to_le16(L2XTSLOTSIZE);
- pxd = &lrd->log.redopage.pxd;
+ page_pxd = &lrd->log.redopage.pxd;
if (tlck->type & tlckBTROOT) {
lrd->log.redopage.type |= cpu_to_le16(LOG_BTROOT);
* applying the after-image to the meta-data page.
*/
lrd->type = cpu_to_le16(LOG_REDOPAGE);
-// *pxd = mp->cm_pxd;
- PXDaddress(pxd, mp->index);
- PXDlength(pxd,
+// *page_pxd = mp->cm_pxd;
+ PXDaddress(page_pxd, mp->index);
+ PXDlength(page_pxd,
mp->logical_size >> tblk->sb->s_blocksize_bits);
lrd->backchain = cpu_to_le32(lmLog(log, tblk, lrd, tlck));
tlck->flag |= tlckUPDATEMAP;
xadlock->flag = mlckALLOCXADLIST;
xadlock->count = next - lwm;
- if ((xadlock->count <= 2) && (tblk->xflag & COMMIT_LAZY)) {
+ if ((xadlock->count <= 4) && (tblk->xflag & COMMIT_LAZY)) {
int i;
+ pxd_t *pxd;
/*
* Lazy commit may allow xtree to be modified before
* txUpdateMap runs. Copy xad into linelock to
* preserve correct data.
+ *
+ * We can fit twice as may pxd's as xads in the lock
*/
- xadlock->xdlist = &xtlck->pxdlock;
- memcpy(xadlock->xdlist, &p->xad[lwm],
- sizeof(xad_t) * xadlock->count);
-
- for (i = 0; i < xadlock->count; i++)
+ xadlock->flag = mlckALLOCPXDLIST;
+ pxd = xadlock->xdlist = &xtlck->pxdlock;
+ for (i = 0; i < xadlock->count; i++) {
+ PXDaddress(pxd, addressXAD(&p->xad[lwm + i]));
+ PXDlength(pxd, lengthXAD(&p->xad[lwm + i]));
p->xad[lwm + i].flag &=
~(XAD_NEW | XAD_EXTENDED);
+ pxd++;
+ }
} else {
/*
* xdlist will point to into inode's xtree, ensure
* that transaction is not committed lazily.
*/
+ xadlock->flag = mlckALLOCXADLIST;
xadlock->xdlist = &p->xad[lwm];
tblk->xflag &= ~COMMIT_LAZY;
}
if (tblk->xflag & COMMIT_TRUNCATE) {
/* write NOREDOPAGE for the page */
lrd->type = cpu_to_le16(LOG_NOREDOPAGE);
- PXDaddress(pxd, mp->index);
- PXDlength(pxd,
+ PXDaddress(page_pxd, mp->index);
+ PXDlength(page_pxd,
mp->logical_size >> tblk->sb->
s_blocksize_bits);
lrd->backchain =
* deleted page itself;
*/
tlck->flag |= tlckUPDATEMAP;
- xadlock->flag = mlckFREEXADLIST;
xadlock->count = hwm - XTENTRYSTART + 1;
- if ((xadlock->count <= 2) && (tblk->xflag & COMMIT_LAZY)) {
+ if ((xadlock->count <= 4) && (tblk->xflag & COMMIT_LAZY)) {
+ int i;
+ pxd_t *pxd;
/*
* Lazy commit may allow xtree to be modified before
* txUpdateMap runs. Copy xad into linelock to
* preserve correct data.
+ *
+ * We can fit twice as may pxd's as xads in the lock
*/
- xadlock->xdlist = &xtlck->pxdlock;
- memcpy(xadlock->xdlist, &p->xad[XTENTRYSTART],
- sizeof(xad_t) * xadlock->count);
+ xadlock->flag = mlckFREEPXDLIST;
+ pxd = xadlock->xdlist = &xtlck->pxdlock;
+ for (i = 0; i < xadlock->count; i++) {
+ PXDaddress(pxd,
+ addressXAD(&p->xad[XTENTRYSTART + i]));
+ PXDlength(pxd,
+ lengthXAD(&p->xad[XTENTRYSTART + i]));
+ pxd++;
+ }
} else {
/*
* xdlist will point to into inode's xtree, ensure
* that transaction is not committed lazily.
*/
+ xadlock->flag = mlckFREEXADLIST;
xadlock->xdlist = &p->xad[XTENTRYSTART];
tblk->xflag &= ~COMMIT_LAZY;
}
* header ?
*/
if (tlck->type & tlckTRUNCATE) {
- pxd_t tpxd; /* truncated extent of xad */
+ pxd_t pxd; /* truncated extent of xad */
int twm;
/*
* applying the after-image to the meta-data page.
*/
lrd->type = cpu_to_le16(LOG_REDOPAGE);
- PXDaddress(pxd, mp->index);
- PXDlength(pxd, mp->logical_size >> tblk->sb->s_blocksize_bits);
+ PXDaddress(page_pxd, mp->index);
+ PXDlength(page_pxd,
+ mp->logical_size >> tblk->sb->s_blocksize_bits);
lrd->backchain = cpu_to_le32(lmLog(log, tblk, lrd, tlck));
/*
lrd->log.updatemap.type = cpu_to_le16(LOG_FREEPXD);
lrd->log.updatemap.nxd = cpu_to_le16(1);
lrd->log.updatemap.pxd = pxdlock->pxd;
- tpxd = pxdlock->pxd; /* save to format maplock */
+ pxd = pxdlock->pxd; /* save to format maplock */
lrd->backchain =
cpu_to_le32(lmLog(log, tblk, lrd, NULL));
}
pxdlock = (struct pxd_lock *) xadlock;
pxdlock->flag = mlckFREEPXD;
pxdlock->count = 1;
- pxdlock->pxd = tpxd;
+ pxdlock->pxd = pxd;
jfs_info("xtLog: truncate ip:0x%p mp:0x%p count:%d "
"hwm:%d", ip, mp, pxdlock->count, hwm);
tlck->flag &= ~tlckWRITEPAGE;
/* do not release page to freelist */
-
+ force_metapage(mp);
+#if 0
/*
* The "right" thing to do here is to
* synchronously write the metadata.
* we can get by with synchronously writing
* the pages when they are released.
*/
- assert(atomic_read(&mp->nohomeok));
+ assert(mp->nohomeok);
set_bit(META_dirty, &mp->flag);
set_bit(META_sync, &mp->flag);
+#endif
}
}
}
*/
mp = tlck->mp;
ASSERT(mp->xflag & COMMIT_PAGE);
- hold_metapage(mp, 0);
+ grab_metapage(mp);
}
/*
ASSERT(mp->lid == lid);
tlck->mp->lid = 0;
}
- assert(atomic_read(&mp->nohomeok) == 1);
- atomic_dec(&mp->nohomeok);
+ assert(mp->nohomeok == 1);
+ metapage_homeok(mp);
discard_metapage(mp);
tlck->mp = NULL;
}
do {
LAZY_LOCK(flags);
+ jfs_commit_thread_waking = 0; /* OK to wake another thread */
while (!list_empty(&TxAnchor.unlock_queue)) {
WorkDone = 0;
list_for_each_entry(tblk, &TxAnchor.unlock_queue,
if (!WorkDone)
break;
}
+ /* In case a wakeup came while all threads were active */
+ jfs_commit_thread_waking = 0;
if (current->flags & PF_FREEZE) {
LAZY_UNLOCK(flags);
list_add_tail(&tblk->cqueue, &TxAnchor.unlock_queue);
/*
* Don't wake up a commit thread if there is already one servicing
- * this superblock.
+ * this superblock, or if the last one we woke up hasn't started yet.
*/
- if (!(JFS_SBI(tblk->sb)->commit_state & IN_LAZYCOMMIT))
+ if (!(JFS_SBI(tblk->sb)->commit_state & IN_LAZYCOMMIT) &&
+ !jfs_commit_thread_waking) {
+ jfs_commit_thread_waking = 1;
wake_up(&jfs_commit_thread_wait);
+ }
LAZY_UNLOCK(flags);
}
{
struct jfs_log *log = mp->log;
- assert(atomic_read(&mp->nohomeok));
+ assert(mp->nohomeok);
assert(log);
- atomic_dec(&mp->nohomeok);
-
- if (atomic_read(&mp->nohomeok))
- return;
-
- hold_metapage(mp, 0);
-
- LOGSYNC_LOCK(log);
- mp->log = NULL;
- mp->lsn = 0;
- mp->clsn = 0;
- log->count--;
- list_del_init(&mp->synclist);
- LOGSYNC_UNLOCK(log);
-
- release_metapage(mp);
+ metapage_homeok(mp);
}
/*
git://git.onelab.eu / linux-2.6.git / blobdiff