summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--fs/xfs/xfs_attr_remote.c2
-rw-r--r--fs/xfs/xfs_bmap.c23
-rw-r--r--fs/xfs/xfs_bmap_util.c15
-rw-r--r--fs/xfs/xfs_buf.c37
-rw-r--r--fs/xfs/xfs_buf.h11
-rw-r--r--fs/xfs/xfs_buf_item.c124
-rw-r--r--fs/xfs/xfs_dquot_item.c67
-rw-r--r--fs/xfs/xfs_dquot_item.h3
-rw-r--r--fs/xfs/xfs_extfree_item.c21
-rw-r--r--fs/xfs/xfs_ialloc.c53
-rw-r--r--fs/xfs/xfs_ialloc.h21
-rw-r--r--fs/xfs/xfs_icreate_item.c10
-rw-r--r--fs/xfs/xfs_inode.c23
-rw-r--r--fs/xfs/xfs_inode_fork.c15
-rw-r--r--fs/xfs/xfs_inode_item.c400
-rw-r--r--fs/xfs/xfs_inode_item.h5
-rw-r--r--fs/xfs/xfs_itable.c22
-rw-r--r--fs/xfs/xfs_log.h46
-rw-r--r--fs/xfs/xfs_log_cil.c74
-rw-r--r--fs/xfs/xfs_log_recover.c29
-rw-r--r--fs/xfs/xfs_trans.h2
-rw-r--r--fs/xfs/xfs_trans_buf.c13
-rw-r--r--fs/xfs/xfs_trans_resv.c10
-rw-r--r--fs/xfs/xfs_trans_space.h2
24 files changed, 493 insertions, 535 deletions
diff --git a/fs/xfs/xfs_attr_remote.c b/fs/xfs/xfs_attr_remote.c
index 739e0a52deda..5549d69ddb45 100644
--- a/fs/xfs/xfs_attr_remote.c
+++ b/fs/xfs/xfs_attr_remote.c
@@ -110,7 +110,7 @@ xfs_attr3_rmt_verify(
if (be32_to_cpu(rmt->rm_bytes) > fsbsize - sizeof(*rmt))
return false;
if (be32_to_cpu(rmt->rm_offset) +
- be32_to_cpu(rmt->rm_bytes) >= XATTR_SIZE_MAX)
+ be32_to_cpu(rmt->rm_bytes) > XATTR_SIZE_MAX)
return false;
if (rmt->rm_owner == 0)
return false;
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index dbf27384b744..152543c4ca70 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -3648,10 +3648,19 @@ xfs_bmap_btalloc(
int isaligned;
int tryagain;
int error;
+ int stripe_align;
ASSERT(ap->length);
mp = ap->ip->i_mount;
+
+ /* stripe alignment for allocation is determined by mount parameters */
+ stripe_align = 0;
+ if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC))
+ stripe_align = mp->m_swidth;
+ else if (mp->m_dalign)
+ stripe_align = mp->m_dalign;
+
align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0;
if (unlikely(align)) {
error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
@@ -3660,6 +3669,8 @@ xfs_bmap_btalloc(
ASSERT(!error);
ASSERT(ap->length);
}
+
+
nullfb = *ap->firstblock == NULLFSBLOCK;
fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock);
if (nullfb) {
@@ -3735,7 +3746,7 @@ xfs_bmap_btalloc(
*/
if (!ap->flist->xbf_low && ap->aeof) {
if (!ap->offset) {
- args.alignment = mp->m_dalign;
+ args.alignment = stripe_align;
atype = args.type;
isaligned = 1;
/*
@@ -3760,13 +3771,13 @@ xfs_bmap_btalloc(
* of minlen+alignment+slop doesn't go up
* between the calls.
*/
- if (blen > mp->m_dalign && blen <= args.maxlen)
- nextminlen = blen - mp->m_dalign;
+ if (blen > stripe_align && blen <= args.maxlen)
+ nextminlen = blen - stripe_align;
else
nextminlen = args.minlen;
- if (nextminlen + mp->m_dalign > args.minlen + 1)
+ if (nextminlen + stripe_align > args.minlen + 1)
args.minalignslop =
- nextminlen + mp->m_dalign -
+ nextminlen + stripe_align -
args.minlen - 1;
else
args.minalignslop = 0;
@@ -3788,7 +3799,7 @@ xfs_bmap_btalloc(
*/
args.type = atype;
args.fsbno = ap->blkno;
- args.alignment = mp->m_dalign;
+ args.alignment = stripe_align;
args.minlen = nextminlen;
args.minalignslop = 0;
isaligned = 1;
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 202a51f7c450..f264616080ca 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -287,6 +287,7 @@ xfs_bmapi_allocate(
INIT_WORK_ONSTACK(&args->work, xfs_bmapi_allocate_worker);
queue_work(xfs_alloc_wq, &args->work);
wait_for_completion(&done);
+ destroy_work_on_stack(&args->work);
return args->result;
}
@@ -1198,7 +1199,12 @@ xfs_zero_remaining_bytes(
XFS_BUF_UNWRITE(bp);
XFS_BUF_READ(bp);
XFS_BUF_SET_ADDR(bp, xfs_fsb_to_db(ip, imap.br_startblock));
- xfsbdstrat(mp, bp);
+
+ if (XFS_FORCED_SHUTDOWN(mp)) {
+ error = XFS_ERROR(EIO);
+ break;
+ }
+ xfs_buf_iorequest(bp);
error = xfs_buf_iowait(bp);
if (error) {
xfs_buf_ioerror_alert(bp,
@@ -1211,7 +1217,12 @@ xfs_zero_remaining_bytes(
XFS_BUF_UNDONE(bp);
XFS_BUF_UNREAD(bp);
XFS_BUF_WRITE(bp);
- xfsbdstrat(mp, bp);
+
+ if (XFS_FORCED_SHUTDOWN(mp)) {
+ error = XFS_ERROR(EIO);
+ break;
+ }
+ xfs_buf_iorequest(bp);
error = xfs_buf_iowait(bp);
if (error) {
xfs_buf_ioerror_alert(bp,
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index ce01c1a17cc1..9fccfb594291 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -698,7 +698,11 @@ xfs_buf_read_uncached(
bp->b_flags |= XBF_READ;
bp->b_ops = ops;
- xfsbdstrat(target->bt_mount, bp);
+ if (XFS_FORCED_SHUTDOWN(target->bt_mount)) {
+ xfs_buf_relse(bp);
+ return NULL;
+ }
+ xfs_buf_iorequest(bp);
xfs_buf_iowait(bp);
return bp;
}
@@ -1089,7 +1093,7 @@ xfs_bioerror(
* This is meant for userdata errors; metadata bufs come with
* iodone functions attached, so that we can track down errors.
*/
-STATIC int
+int
xfs_bioerror_relse(
struct xfs_buf *bp)
{
@@ -1152,7 +1156,7 @@ xfs_bwrite(
ASSERT(xfs_buf_islocked(bp));
bp->b_flags |= XBF_WRITE;
- bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q);
+ bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q | XBF_WRITE_FAIL);
xfs_bdstrat_cb(bp);
@@ -1164,25 +1168,6 @@ xfs_bwrite(
return error;
}
-/*
- * Wrapper around bdstrat so that we can stop data from going to disk in case
- * we are shutting down the filesystem. Typically user data goes thru this
- * path; one of the exceptions is the superblock.
- */
-void
-xfsbdstrat(
- struct xfs_mount *mp,
- struct xfs_buf *bp)
-{
- if (XFS_FORCED_SHUTDOWN(mp)) {
- trace_xfs_bdstrat_shut(bp, _RET_IP_);
- xfs_bioerror_relse(bp);
- return;
- }
-
- xfs_buf_iorequest(bp);
-}
-
STATIC void
_xfs_buf_ioend(
xfs_buf_t *bp,
@@ -1516,6 +1501,12 @@ xfs_wait_buftarg(
struct xfs_buf *bp;
bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
list_del_init(&bp->b_lru);
+ if (bp->b_flags & XBF_WRITE_FAIL) {
+ xfs_alert(btp->bt_mount,
+"Corruption Alert: Buffer at block 0x%llx had permanent write failures!\n"
+"Please run xfs_repair to determine the extent of the problem.",
+ (long long)bp->b_bn);
+ }
xfs_buf_rele(bp);
}
if (loop++ != 0)
@@ -1789,7 +1780,7 @@ __xfs_buf_delwri_submit(
blk_start_plug(&plug);
list_for_each_entry_safe(bp, n, io_list, b_list) {
- bp->b_flags &= ~(_XBF_DELWRI_Q | XBF_ASYNC);
+ bp->b_flags &= ~(_XBF_DELWRI_Q | XBF_ASYNC | XBF_WRITE_FAIL);
bp->b_flags |= XBF_WRITE;
if (!wait) {
diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h
index e65683361017..1cf21a4a9f22 100644
--- a/fs/xfs/xfs_buf.h
+++ b/fs/xfs/xfs_buf.h
@@ -45,6 +45,7 @@ typedef enum {
#define XBF_ASYNC (1 << 4) /* initiator will not wait for completion */
#define XBF_DONE (1 << 5) /* all pages in the buffer uptodate */
#define XBF_STALE (1 << 6) /* buffer has been staled, do not find it */
+#define XBF_WRITE_FAIL (1 << 24)/* async writes have failed on this buffer */
/* I/O hints for the BIO layer */
#define XBF_SYNCIO (1 << 10)/* treat this buffer as synchronous I/O */
@@ -70,6 +71,7 @@ typedef unsigned int xfs_buf_flags_t;
{ XBF_ASYNC, "ASYNC" }, \
{ XBF_DONE, "DONE" }, \
{ XBF_STALE, "STALE" }, \
+ { XBF_WRITE_FAIL, "WRITE_FAIL" }, \
{ XBF_SYNCIO, "SYNCIO" }, \
{ XBF_FUA, "FUA" }, \
{ XBF_FLUSH, "FLUSH" }, \
@@ -80,6 +82,7 @@ typedef unsigned int xfs_buf_flags_t;
{ _XBF_DELWRI_Q, "DELWRI_Q" }, \
{ _XBF_COMPOUND, "COMPOUND" }
+
/*
* Internal state flags.
*/
@@ -269,9 +272,6 @@ extern void xfs_buf_unlock(xfs_buf_t *);
/* Buffer Read and Write Routines */
extern int xfs_bwrite(struct xfs_buf *bp);
-
-extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *);
-
extern void xfs_buf_ioend(xfs_buf_t *, int);
extern void xfs_buf_ioerror(xfs_buf_t *, int);
extern void xfs_buf_ioerror_alert(struct xfs_buf *, const char *func);
@@ -282,6 +282,8 @@ extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *,
#define xfs_buf_zero(bp, off, len) \
xfs_buf_iomove((bp), (off), (len), NULL, XBRW_ZERO)
+extern int xfs_bioerror_relse(struct xfs_buf *);
+
static inline int xfs_buf_geterror(xfs_buf_t *bp)
{
return bp ? bp->b_error : ENOMEM;
@@ -301,7 +303,8 @@ extern void xfs_buf_terminate(void);
#define XFS_BUF_ZEROFLAGS(bp) \
((bp)->b_flags &= ~(XBF_READ|XBF_WRITE|XBF_ASYNC| \
- XBF_SYNCIO|XBF_FUA|XBF_FLUSH))
+ XBF_SYNCIO|XBF_FUA|XBF_FLUSH| \
+ XBF_WRITE_FAIL))
void xfs_buf_stale(struct xfs_buf *bp);
#define XFS_BUF_UNSTALE(bp) ((bp)->b_flags &= ~XBF_STALE)
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index a64f67ba25d3..33149113e333 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -182,21 +182,47 @@ xfs_buf_item_size(
trace_xfs_buf_item_size(bip);
}
-static struct xfs_log_iovec *
+static inline void
+xfs_buf_item_copy_iovec(
+ struct xfs_log_vec *lv,
+ struct xfs_log_iovec **vecp,
+ struct xfs_buf *bp,
+ uint offset,
+ int first_bit,
+ uint nbits)
+{
+ offset += first_bit * XFS_BLF_CHUNK;
+ xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_BCHUNK,
+ xfs_buf_offset(bp, offset),
+ nbits * XFS_BLF_CHUNK);
+}
+
+static inline bool
+xfs_buf_item_straddle(
+ struct xfs_buf *bp,
+ uint offset,
+ int next_bit,
+ int last_bit)
+{
+ return xfs_buf_offset(bp, offset + (next_bit << XFS_BLF_SHIFT)) !=
+ (xfs_buf_offset(bp, offset + (last_bit << XFS_BLF_SHIFT)) +
+ XFS_BLF_CHUNK);
+}
+
+static void
xfs_buf_item_format_segment(
struct xfs_buf_log_item *bip,
- struct xfs_log_iovec *vecp,
+ struct xfs_log_vec *lv,
+ struct xfs_log_iovec **vecp,
uint offset,
struct xfs_buf_log_format *blfp)
{
struct xfs_buf *bp = bip->bli_buf;
uint base_size;
- uint nvecs;
int first_bit;
int last_bit;
int next_bit;
uint nbits;
- uint buffer_offset;
/* copy the flags across from the base format item */
blfp->blf_flags = bip->__bli_format.blf_flags;
@@ -208,21 +234,17 @@ xfs_buf_item_format_segment(
*/
base_size = xfs_buf_log_format_size(blfp);
- nvecs = 0;
first_bit = xfs_next_bit(blfp->blf_data_map, blfp->blf_map_size, 0);
if (!(bip->bli_flags & XFS_BLI_STALE) && first_bit == -1) {
/*
* If the map is not be dirty in the transaction, mark
* the size as zero and do not advance the vector pointer.
*/
- goto out;
+ return;
}
- vecp->i_addr = blfp;
- vecp->i_len = base_size;
- vecp->i_type = XLOG_REG_TYPE_BFORMAT;
- vecp++;
- nvecs = 1;
+ blfp = xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_BFORMAT, blfp, base_size);
+ blfp->blf_size = 1;
if (bip->bli_flags & XFS_BLI_STALE) {
/*
@@ -232,14 +254,13 @@ xfs_buf_item_format_segment(
*/
trace_xfs_buf_item_format_stale(bip);
ASSERT(blfp->blf_flags & XFS_BLF_CANCEL);
- goto out;
+ return;
}
/*
* Fill in an iovec for each set of contiguous chunks.
*/
-
last_bit = first_bit;
nbits = 1;
for (;;) {
@@ -252,42 +273,22 @@ xfs_buf_item_format_segment(
next_bit = xfs_next_bit(blfp->blf_data_map, blfp->blf_map_size,
(uint)last_bit + 1);
/*
- * If we run out of bits fill in the last iovec and get
- * out of the loop.
- * Else if we start a new set of bits then fill in the
- * iovec for the series we were looking at and start
- * counting the bits in the new one.
- * Else we're still in the same set of bits so just
- * keep counting and scanning.
+ * If we run out of bits fill in the last iovec and get out of
+ * the loop. Else if we start a new set of bits then fill in
+ * the iovec for the series we were looking at and start
+ * counting the bits in the new one. Else we're still in the
+ * same set of bits so just keep counting and scanning.
*/
if (next_bit == -1) {
- buffer_offset = offset + first_bit * XFS_BLF_CHUNK;
- vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
- vecp->i_len = nbits * XFS_BLF_CHUNK;
- vecp->i_type = XLOG_REG_TYPE_BCHUNK;
- nvecs++;
+ xfs_buf_item_copy_iovec(lv, vecp, bp, offset,
+ first_bit, nbits);
+ blfp->blf_size++;
break;
- } else if (next_bit != last_bit + 1) {
- buffer_offset = offset + first_bit * XFS_BLF_CHUNK;
- vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
- vecp->i_len = nbits * XFS_BLF_CHUNK;
- vecp->i_type = XLOG_REG_TYPE_BCHUNK;
- nvecs++;
- vecp++;
- first_bit = next_bit;
- last_bit = next_bit;
- nbits = 1;
- } else if (xfs_buf_offset(bp, offset +
- (next_bit << XFS_BLF_SHIFT)) !=
- (xfs_buf_offset(bp, offset +
- (last_bit << XFS_BLF_SHIFT)) +
- XFS_BLF_CHUNK)) {
- buffer_offset = offset + first_bit * XFS_BLF_CHUNK;
- vecp->i_addr = xfs_buf_offset(bp, buffer_offset);
- vecp->i_len = nbits * XFS_BLF_CHUNK;
- vecp->i_type = XLOG_REG_TYPE_BCHUNK;
- nvecs++;
- vecp++;
+ } else if (next_bit != last_bit + 1 ||
+ xfs_buf_item_straddle(bp, offset, next_bit, last_bit)) {
+ xfs_buf_item_copy_iovec(lv, vecp, bp, offset,
+ first_bit, nbits);
+ blfp->blf_size++;
first_bit = next_bit;
last_bit = next_bit;
nbits = 1;
@@ -296,9 +297,6 @@ xfs_buf_item_format_segment(
nbits++;
}
}
-out:
- blfp->blf_size = nvecs;
- return vecp;
}
/*
@@ -310,10 +308,11 @@ out:
STATIC void
xfs_buf_item_format(
struct xfs_log_item *lip,
- struct xfs_log_iovec *vecp)
+ struct xfs_log_vec *lv)
{
struct xfs_buf_log_item *bip = BUF_ITEM(lip);
struct xfs_buf *bp = bip->bli_buf;
+ struct xfs_log_iovec *vecp = NULL;
uint offset = 0;
int i;
@@ -354,8 +353,8 @@ xfs_buf_item_format(
}
for (i = 0; i < bip->bli_format_count; i++) {
- vecp = xfs_buf_item_format_segment(bip, vecp, offset,
- &bip->bli_formats[i]);
+ xfs_buf_item_format_segment(bip, lv, &vecp, offset,
+ &bip->bli_formats[i]);
offset += bp->b_maps[i].bm_len;
}
@@ -496,6 +495,14 @@ xfs_buf_item_unpin(
}
}
+/*
+ * Buffer IO error rate limiting. Limit it to no more than 10 messages per 30
+ * seconds so as to not spam logs too much on repeated detection of the same
+ * buffer being bad..
+ */
+
+DEFINE_RATELIMIT_STATE(xfs_buf_write_fail_rl_state, 30 * HZ, 10);
+
STATIC uint
xfs_buf_item_push(
struct xfs_log_item *lip,
@@ -524,6 +531,14 @@ xfs_buf_item_push(
trace_xfs_buf_item_push(bip);
+ /* has a previous flush failed due to IO errors? */
+ if ((bp->b_flags & XBF_WRITE_FAIL) &&
+ ___ratelimit(&xfs_buf_write_fail_rl_state, "XFS:")) {
+ xfs_warn(bp->b_target->bt_mount,
+"Detected failing async write on buffer block 0x%llx. Retrying async write.\n",
+ (long long)bp->b_bn);
+ }
+
if (!xfs_buf_delwri_queue(bp, buffer_list))
rval = XFS_ITEM_FLUSHING;
xfs_buf_unlock(bp);
@@ -1096,8 +1111,9 @@ xfs_buf_iodone_callbacks(
xfs_buf_ioerror(bp, 0); /* errno of 0 unsets the flag */
- if (!XFS_BUF_ISSTALE(bp)) {
- bp->b_flags |= XBF_WRITE | XBF_ASYNC | XBF_DONE;
+ if (!(bp->b_flags & (XBF_STALE|XBF_WRITE_FAIL))) {
+ bp->b_flags |= XBF_WRITE | XBF_ASYNC |
+ XBF_DONE | XBF_WRITE_FAIL;
xfs_buf_iorequest(bp);
} else {
xfs_buf_relse(bp);
diff --git a/fs/xfs/xfs_dquot_item.c b/fs/xfs/xfs_dquot_item.c
index 92e5f62eefc6..f33fbaaa4d8a 100644
--- a/fs/xfs/xfs_dquot_item.c
+++ b/fs/xfs/xfs_dquot_item.c
@@ -57,20 +57,24 @@ xfs_qm_dquot_logitem_size(
STATIC void
xfs_qm_dquot_logitem_format(
struct xfs_log_item *lip,
- struct xfs_log_iovec *logvec)
+ struct xfs_log_vec *lv)
{
struct xfs_dq_logitem *qlip = DQUOT_ITEM(lip);
-
- logvec->i_addr = &qlip->qli_format;
- logvec->i_len = sizeof(xfs_dq_logformat_t);
- logvec->i_type = XLOG_REG_TYPE_QFORMAT;
- logvec++;
- logvec->i_addr = &qlip->qli_dquot->q_core;
- logvec->i_len = sizeof(xfs_disk_dquot_t);
- logvec->i_type = XLOG_REG_TYPE_DQUOT;
-
- qlip->qli_format.qlf_size = 2;
-
+ struct xfs_log_iovec *vecp = NULL;
+ struct xfs_dq_logformat *qlf;
+
+ qlf = xlog_prepare_iovec(lv, &vecp, XLOG_REG_TYPE_QFORMAT);
+ qlf->qlf_type = XFS_LI_DQUOT;
+ qlf->qlf_size = 2;
+ qlf->qlf_id = be32_to_cpu(qlip->qli_dquot->q_core.d_id);
+ qlf->qlf_blkno = qlip->qli_dquot->q_blkno;
+ qlf->qlf_len = 1;
+ qlf->qlf_boffset = qlip->qli_dquot->q_bufoffset;
+ xlog_finish_iovec(lv, vecp, sizeof(struct xfs_dq_logformat));
+
+ xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_DQUOT,
+ &qlip->qli_dquot->q_core,
+ sizeof(struct xfs_disk_dquot));
}
/*
@@ -257,18 +261,6 @@ xfs_qm_dquot_logitem_init(
xfs_log_item_init(dqp->q_mount, &lp->qli_item, XFS_LI_DQUOT,
&xfs_dquot_item_ops);
lp->qli_dquot = dqp;
- lp->qli_format.qlf_type = XFS_LI_DQUOT;
- lp->qli_format.qlf_id = be32_to_cpu(dqp->q_core.d_id);
- lp->qli_format.qlf_blkno = dqp->q_blkno;
- lp->qli_format.qlf_len = 1;
- /*
- * This is just the offset of this dquot within its buffer
- * (which is currently 1 FSB and probably won't change).
- * Hence 32 bits for this offset should be just fine.
- * Alternatively, we can store (bufoffset / sizeof(xfs_dqblk_t))
- * here, and recompute it at recovery time.
- */
- lp->qli_format.qlf_boffset = (__uint32_t)dqp->q_bufoffset;
}
/*------------------ QUOTAOFF LOG ITEMS -------------------*/
@@ -294,26 +286,20 @@ xfs_qm_qoff_logitem_size(
*nbytes += sizeof(struct xfs_qoff_logitem);
}
-/*
- * This is called to fill in the vector of log iovecs for the
- * given quotaoff log item. We use only 1 iovec, and we point that
- * at the quotaoff_log_format structure embedded in the quotaoff item.
- * It is at this point that we assert that all of the extent
- * slots in the quotaoff item have been filled.
- */
STATIC void
xfs_qm_qoff_logitem_format(
struct xfs_log_item *lip,
- struct xfs_log_iovec *log_vector)
+ struct xfs_log_vec *lv)
{
struct xfs_qoff_logitem *qflip = QOFF_ITEM(lip);
-
- ASSERT(qflip->qql_format.qf_type == XFS_LI_QUOTAOFF);
-
- log_vector->i_addr = &qflip->qql_format;
- log_vector->i_len = sizeof(xfs_qoff_logitem_t);
- log_vector->i_type = XLOG_REG_TYPE_QUOTAOFF;
- qflip->qql_format.qf_size = 1;
+ struct xfs_log_iovec *vecp = NULL;
+ struct xfs_qoff_logformat *qlf;
+
+ qlf = xlog_prepare_iovec(lv, &vecp, XLOG_REG_TYPE_QUOTAOFF);
+ qlf->qf_type = XFS_LI_QUOTAOFF;
+ qlf->qf_size = 1;
+ qlf->qf_flags = qflip->qql_flags;
+ xlog_finish_iovec(lv, vecp, sizeof(struct xfs_qoff_logitem));
}
/*
@@ -453,8 +439,7 @@ xfs_qm_qoff_logitem_init(
xfs_log_item_init(mp, &qf->qql_item, XFS_LI_QUOTAOFF, start ?
&xfs_qm_qoffend_logitem_ops : &xfs_qm_qoff_logitem_ops);
qf->qql_item.li_mountp = mp;
- qf->qql_format.qf_type = XFS_LI_QUOTAOFF;
- qf->qql_format.qf_flags = flags;
qf->qql_start_lip = start;
+ qf->qql_flags = flags;
return qf;
}
diff --git a/fs/xfs/xfs_dquot_item.h b/fs/xfs/xfs_dquot_item.h
index 5acae2ada70b..502e9464634a 100644
--- a/fs/xfs/xfs_dquot_item.h
+++ b/fs/xfs/xfs_dquot_item.h
@@ -27,13 +27,12 @@ typedef struct xfs_dq_logitem {
xfs_log_item_t qli_item; /* common portion */
struct xfs_dquot *qli_dquot; /* dquot ptr */
xfs_lsn_t qli_flush_lsn; /* lsn at last flush */
- xfs_dq_logformat_t qli_format; /* logged structure */
} xfs_dq_logitem_t;
typedef struct xfs_qoff_logitem {
xfs_log_item_t qql_item; /* common portion */
struct xfs_qoff_logitem *qql_start_lip; /* qoff-start logitem, if any */
- xfs_qoff_logformat_t qql_format; /* logged structure */
+ unsigned int qql_flags;
} xfs_qoff_logitem_t;
diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c
index 3680d04f973f..fb7a4c1ce1c5 100644
--- a/fs/xfs/xfs_extfree_item.c
+++ b/fs/xfs/xfs_extfree_item.c
@@ -26,6 +26,7 @@
#include "xfs_trans_priv.h"
#include "xfs_buf_item.h"
#include "xfs_extfree_item.h"
+#include "xfs_log.h"
kmem_zone_t *xfs_efi_zone;
@@ -101,9 +102,10 @@ xfs_efi_item_size(
STATIC void
xfs_efi_item_format(
struct xfs_log_item *lip,
- struct xfs_log_iovec *log_vector)
+ struct xfs_log_vec *lv)
{
struct xfs_efi_log_item *efip = EFI_ITEM(lip);
+ struct xfs_log_iovec *vecp = NULL;
ASSERT(atomic_read(&efip->efi_next_extent) ==
efip->efi_format.efi_nextents);
@@ -111,10 +113,9 @@ xfs_efi_item_format(
efip->efi_format.efi_type = XFS_LI_EFI;
efip->efi_format.efi_size = 1;
- log_vector->i_addr = &efip->efi_format;
- log_vector->i_len = xfs_efi_item_sizeof(efip);
- log_vector->i_type = XLOG_REG_TYPE_EFI_FORMAT;
- ASSERT(log_vector->i_len >= sizeof(xfs_efi_log_format_t));
+ xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_EFI_FORMAT,
+ &efip->efi_format,
+ xfs_efi_item_sizeof(efip));
}
@@ -368,19 +369,19 @@ xfs_efd_item_size(
STATIC void
xfs_efd_item_format(
struct xfs_log_item *lip,
- struct xfs_log_iovec *log_vector)
+ struct xfs_log_vec *lv)
{
struct xfs_efd_log_item *efdp = EFD_ITEM(lip);
+ struct xfs_log_iovec *vecp = NULL;
ASSERT(efdp->efd_next_extent == efdp->efd_format.efd_nextents);
efdp->efd_format.efd_type = XFS_LI_EFD;
efdp->efd_format.efd_size = 1;
- log_vector->i_addr = &efdp->efd_format;
- log_vector->i_len = xfs_efd_item_sizeof(efdp);
- log_vector->i_type = XLOG_REG_TYPE_EFD_FORMAT;
- ASSERT(log_vector->i_len >= sizeof(xfs_efd_log_format_t));
+ xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_EFD_FORMAT,
+ &efdp->efd_format,
+ xfs_efd_item_sizeof(efdp));
}
/*
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 7a728f9fc0be..5d7f105a1c82 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -52,7 +52,7 @@ xfs_ialloc_cluster_alignment(
{
if (xfs_sb_version_hasalign(&args->mp->m_sb) &&
args->mp->m_sb.sb_inoalignmt >=
- XFS_B_TO_FSBT(args->mp, XFS_INODE_CLUSTER_SIZE(args->mp)))
+ XFS_B_TO_FSBT(args->mp, args->mp->m_inode_cluster_size))
return args->mp->m_sb.sb_inoalignmt;
return 1;
}
@@ -170,27 +170,20 @@ xfs_ialloc_inode_init(
{
struct xfs_buf *fbuf;
struct xfs_dinode *free;
- int blks_per_cluster, nbufs, ninodes;
+ int nbufs, blks_per_cluster, inodes_per_cluster;
int version;
int i, j;
xfs_daddr_t d;
xfs_ino_t ino = 0;
/*
- * Loop over the new block(s), filling in the inodes.
- * For small block sizes, manipulate the inodes in buffers
- * which are multiples of the blocks size.
+ * Loop over the new block(s), filling in the inodes. For small block
+ * sizes, manipulate the inodes in buffers which are multiples of the
+ * blocks size.
*/
- if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
- blks_per_cluster = 1;
- nbufs = length;
- ninodes = mp->m_sb.sb_inopblock;
- } else {
- blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) /
- mp->m_sb.sb_blocksize;
- nbufs = length / blks_per_cluster;
- ninodes = blks_per_cluster * mp->m_sb.sb_inopblock;
- }
+ blks_per_cluster = xfs_icluster_size_fsb(mp);
+ inodes_per_cluster = blks_per_cluster << mp->m_sb.sb_inopblog;
+ nbufs = length / blks_per_cluster;
/*
* Figure out what version number to use in the inodes we create. If
@@ -225,7 +218,7 @@ xfs_ialloc_inode_init(
* they track in the AIL as if they were physically logged.
*/
if (tp)
- xfs_icreate_log(tp, agno, agbno, XFS_IALLOC_INODES(mp),
+ xfs_icreate_log(tp, agno, agbno, mp->m_ialloc_inos,
mp->m_sb.sb_inodesize, length, gen);
} else if (xfs_sb_version_hasnlink(&mp->m_sb))
version = 2;
@@ -246,7 +239,7 @@ xfs_ialloc_inode_init(
/* Initialize the inode buffers and log them appropriately. */
fbuf->b_ops = &xfs_inode_buf_ops;
xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
- for (i = 0; i < ninodes; i++) {
+ for (i = 0; i < inodes_per_cluster; i++) {
int ioffset = i << mp->m_sb.sb_inodelog;
uint isize = xfs_dinode_size(version);
@@ -329,11 +322,11 @@ xfs_ialloc_ag_alloc(
* Locking will ensure that we don't have two callers in here
* at one time.
*/
- newlen = XFS_IALLOC_INODES(args.mp);
+ newlen = args.mp->m_ialloc_inos;
if (args.mp->m_maxicount &&
args.mp->m_sb.sb_icount + newlen > args.mp->m_maxicount)
return XFS_ERROR(ENOSPC);
- args.minlen = args.maxlen = XFS_IALLOC_BLOCKS(args.mp);
+ args.minlen = args.maxlen = args.mp->m_ialloc_blks;
/*
* First try to allocate inodes contiguous with the last-allocated
* chunk of inodes. If the filesystem is striped, this will fill
@@ -343,7 +336,7 @@ xfs_ialloc_ag_alloc(
newino = be32_to_cpu(agi->agi_newino);
agno = be32_to_cpu(agi->agi_seqno);
args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
- XFS_IALLOC_BLOCKS(args.mp);
+ args.mp->m_ialloc_blks;
if (likely(newino != NULLAGINO &&
(args.agbno < be32_to_cpu(agi->agi_length)))) {
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
@@ -585,7 +578,7 @@ xfs_ialloc_ag_select(
* Is there enough free space for the file plus a block of
* inodes? (if we need to allocate some)?
*/
- ineed = XFS_IALLOC_BLOCKS(mp);
+ ineed = mp->m_ialloc_blks;
longest = pag->pagf_longest;
if (!longest)
longest = pag->pagf_flcount > 0;
@@ -999,7 +992,7 @@ xfs_dialloc(
* inode.
*/
if (mp->m_maxicount &&
- mp->m_sb.sb_icount + XFS_IALLOC_INODES(mp) > mp->m_maxicount) {
+ mp->m_sb.sb_icount + mp->m_ialloc_inos > mp->m_maxicount) {
noroom = 1;
okalloc = 0;
}
@@ -1202,7 +1195,7 @@ xfs_difree(
* When an inode cluster is free, it becomes eligible for removal
*/
if (!(mp->m_flags & XFS_MOUNT_IKEEP) &&
- (rec.ir_freecount == XFS_IALLOC_INODES(mp))) {
+ (rec.ir_freecount == mp->m_ialloc_inos)) {
*delete = 1;
*first_ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino);
@@ -1212,7 +1205,7 @@ xfs_difree(
* AGI and Superblock inode counts, and mark the disk space
* to be freed when the transaction is committed.
*/
- ilen = XFS_IALLOC_INODES(mp);
+ ilen = mp->m_ialloc_inos;
be32_add_cpu(&agi->agi_count, -ilen);
be32_add_cpu(&agi->agi_freecount, -(ilen - 1));
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_COUNT | XFS_AGI_FREECOUNT);
@@ -1228,9 +1221,9 @@ xfs_difree(
goto error0;
}
- xfs_bmap_add_free(XFS_AGB_TO_FSB(mp,
- agno, XFS_AGINO_TO_AGBNO(mp, rec.ir_startino)),
- XFS_IALLOC_BLOCKS(mp), flist, mp);
+ xfs_bmap_add_free(XFS_AGB_TO_FSB(mp, agno,
+ XFS_AGINO_TO_AGBNO(mp, rec.ir_startino)),
+ mp->m_ialloc_blks, flist, mp);
} else {
*delete = 0;
@@ -1311,7 +1304,7 @@ xfs_imap_lookup(
/* check that the returned record contains the required inode */
if (rec.ir_startino > agino ||
- rec.ir_startino + XFS_IALLOC_INODES(mp) <= agino)
+ rec.ir_startino + mp->m_ialloc_inos <= agino)
return EINVAL;
/* for untrusted inodes check it is allocated first */
@@ -1384,7 +1377,7 @@ xfs_imap(
return XFS_ERROR(EINVAL);
}
- blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_blocklog;
+ blks_per_cluster = xfs_icluster_size_fsb(mp);
/*
* For bulkstat and handle lookups, we have an untrusted inode number
@@ -1405,7 +1398,7 @@ xfs_imap(
* If the inode cluster size is the same as the blocksize or
* smaller we get to the buffer by simple arithmetics.
*/
- if (XFS_INODE_CLUSTER_SIZE(mp) <= mp->m_sb.sb_blocksize) {
+ if (blks_per_cluster == 1) {
offset = XFS_INO_TO_OFFSET(mp, ino);
ASSERT(offset < mp->m_sb.sb_inopblock);
diff --git a/fs/xfs/xfs_ialloc.h b/fs/xfs/xfs_ialloc.h
index a8f76a5ff418..812365d17e67 100644
--- a/fs/xfs/xfs_ialloc.h
+++ b/fs/xfs/xfs_ialloc.h
@@ -25,17 +25,18 @@ struct xfs_mount;
struct xfs_trans;
struct xfs_btree_cur;
-/*
- * Allocation parameters for inode allocation.
- */
-#define XFS_IALLOC_INODES(mp) (mp)->m_ialloc_inos
-#define XFS_IALLOC_BLOCKS(mp) (mp)->m_ialloc_blks
-
-/*
- * Move inodes in clusters of this size.
- */
+/* Move inodes in clusters of this size */
#define XFS_INODE_BIG_CLUSTER_SIZE 8192
-#define XFS_INODE_CLUSTER_SIZE(mp) (mp)->m_inode_cluster_size
+
+/* Calculate and return the number of filesystem blocks per inode cluster */
+static inline int
+xfs_icluster_size_fsb(
+ struct xfs_mount *mp)
+{
+ if (mp->m_sb.sb_blocksize >= mp->m_inode_cluster_size)
+ return 1;
+ return mp->m_inode_cluster_size >> mp->m_sb.sb_blocklog;
+}
/*
* Make an inode pointer out of the buffer/offset.
diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c
index d2eaccfa73f4..7e4549233251 100644
--- a/fs/xfs/xfs_icreate_item.c
+++ b/fs/xfs/xfs_icreate_item.c
@@ -28,6 +28,7 @@
#include "xfs_trans_priv.h"
#include "xfs_error.h"
#include "xfs_icreate_item.h"
+#include "xfs_log.h"
kmem_zone_t *xfs_icreate_zone; /* inode create item zone */
@@ -58,13 +59,14 @@ xfs_icreate_item_size(
STATIC void
xfs_icreate_item_format(
struct xfs_log_item *lip,
- struct xfs_log_iovec *log_vector)
+ struct xfs_log_vec *lv)
{
struct xfs_icreate_item *icp = ICR_ITEM(lip);
+ struct xfs_log_iovec *vecp = NULL;
- log_vector->i_addr = (xfs_caddr_t)&icp->ic_format;
- log_vector->i_len = sizeof(struct xfs_icreate_log);
- log_vector->i_type = XLOG_REG_TYPE_ICREATE;
+ xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_ICREATE,
+ &icp->ic_format,
+ sizeof(struct xfs_icreate_log));
}
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index e655bb07e8bb..3a137e9f9a7d 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -2137,8 +2137,8 @@ xfs_ifree_cluster(
{
xfs_mount_t *mp = free_ip->i_mount;
int blks_per_cluster;
+ int inodes_per_cluster;
int nbufs;
- int ninodes;
int i, j;
xfs_daddr_t blkno;
xfs_buf_t *bp;
@@ -2148,18 +2148,11 @@ xfs_ifree_cluster(
struct xfs_perag *pag;
pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum));
- if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
- blks_per_cluster = 1;
- ninodes = mp->m_sb.sb_inopblock;
- nbufs = XFS_IALLOC_BLOCKS(mp);
- } else {
- blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) /
- mp->m_sb.sb_blocksize;
- ninodes = blks_per_cluster * mp->m_sb.sb_inopblock;
- nbufs = XFS_IALLOC_BLOCKS(mp) / blks_per_cluster;
- }
+ blks_per_cluster = xfs_icluster_size_fsb(mp);
+ inodes_per_cluster = blks_per_cluster << mp->m_sb.sb_inopblog;
+ nbufs = mp->m_ialloc_blks / blks_per_cluster;
- for (j = 0; j < nbufs; j++, inum += ninodes) {
+ for (j = 0; j < nbufs; j++, inum += inodes_per_cluster) {
blkno = XFS_AGB_TO_DADDR(mp, XFS_INO_TO_AGNO(mp, inum),
XFS_INO_TO_AGBNO(mp, inum));
@@ -2221,7 +2214,7 @@ xfs_ifree_cluster(
* transaction stale above, which means there is no point in
* even trying to lock them.
*/
- for (i = 0; i < ninodes; i++) {
+ for (i = 0; i < inodes_per_cluster; i++) {
retry:
rcu_read_lock();
ip = radix_tree_lookup(&pag->pag_ici_root,
@@ -2902,13 +2895,13 @@ xfs_iflush_cluster(
pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
- inodes_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog;
+ inodes_per_cluster = mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog;
ilist_size = inodes_per_cluster * sizeof(xfs_inode_t *);
ilist = kmem_alloc(ilist_size, KM_MAYFAIL|KM_NOFS);
if (!ilist)
goto out_put;
- mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1);
+ mask = ~(((mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog)) - 1);
first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask;
rcu_read_lock();
/* really need a gang lookup range call here */
diff --git a/fs/xfs/xfs_inode_fork.c b/fs/xfs/xfs_inode_fork.c
index e16985e1c2fd..73514c0486b7 100644
--- a/fs/xfs/xfs_inode_fork.c
+++ b/fs/xfs/xfs_inode_fork.c
@@ -723,15 +723,16 @@ xfs_idestroy_fork(
}
/*
- * xfs_iextents_copy()
+ * Convert in-core extents to on-disk form
*
- * This is called to copy the REAL extents (as opposed to the delayed
- * allocation extents) from the inode into the given buffer. It
- * returns the number of bytes copied into the buffer.
+ * For either the data or attr fork in extent format, we need to endian convert
+ * the in-core extent as we place them into the on-disk inode.
*
- * If there are no delayed allocation extents, then we can just
- * memcpy() the extents into the buffer. Otherwise, we need to
- * examine each extent in turn and skip those which are delayed.
+ * In the case of the data fork, the in-core and on-disk fork sizes can be
+ * different due to delayed allocation extents. We only copy on-disk extents
+ * here, so callers must always use the physical fork size to determine the
+ * size of the buffer passed to this routine. We will return the size actually
+ * used.
*/
int
xfs_iextents_copy(
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 7c0d391f9a6e..686889b4a1e5 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -30,6 +30,7 @@
#include "xfs_trace.h"
#include "xfs_trans_priv.h"
#include "xfs_dinode.h"
+#include "xfs_log.h"
kmem_zone_t *xfs_ili_zone; /* inode log item zone */
@@ -39,27 +40,14 @@ static inline struct xfs_inode_log_item *INODE_ITEM(struct xfs_log_item *lip)
return container_of(lip, struct xfs_inode_log_item, ili_item);
}
-
-/*
- * This returns the number of iovecs needed to log the given inode item.
- *
- * We need one iovec for the inode log format structure, one for the
- * inode core, and possibly one for the inode data/extents/b-tree root
- * and one for the inode attribute data/extents/b-tree root.
- */
STATIC void
-xfs_inode_item_size(
- struct xfs_log_item *lip,
+xfs_inode_item_data_fork_size(
+ struct xfs_inode_log_item *iip,
int *nvecs,
int *nbytes)
{
- struct xfs_inode_log_item *iip = INODE_ITEM(lip);
struct xfs_inode *ip = iip->ili_inode;
- *nvecs += 2;
- *nbytes += sizeof(struct xfs_inode_log_format) +
- xfs_icdinode_size(ip->i_d.di_version);
-
switch (ip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
if ((iip->ili_fields & XFS_ILOG_DEXT) &&
@@ -70,7 +58,6 @@ xfs_inode_item_size(
*nvecs += 1;
}
break;
-
case XFS_DINODE_FMT_BTREE:
if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
ip->i_df.if_broot_bytes > 0) {
@@ -78,7 +65,6 @@ xfs_inode_item_size(
*nvecs += 1;
}
break;
-
case XFS_DINODE_FMT_LOCAL:
if ((iip->ili_fields & XFS_ILOG_DDATA) &&
ip->i_df.if_bytes > 0) {
@@ -90,19 +76,20 @@ xfs_inode_item_size(
case XFS_DINODE_FMT_DEV:
case XFS_DINODE_FMT_UUID:
break;
-
default:
ASSERT(0);
break;
}
+}
- if (!XFS_IFORK_Q(ip))
- return;
-
+STATIC void
+xfs_inode_item_attr_fork_size(
+ struct xfs_inode_log_item *iip,
+ int *nvecs,
+ int *nbytes)
+{
+ struct xfs_inode *ip = iip->ili_inode;
- /*
- * Log any necessary attribute data.
- */
switch (ip->i_d.di_aformat) {
case XFS_DINODE_FMT_EXTENTS:
if ((iip->ili_fields & XFS_ILOG_AEXT) &&
@@ -113,7 +100,6 @@ xfs_inode_item_size(
*nvecs += 1;
}
break;
-
case XFS_DINODE_FMT_BTREE:
if ((iip->ili_fields & XFS_ILOG_ABROOT) &&
ip->i_afp->if_broot_bytes > 0) {
@@ -121,7 +107,6 @@ xfs_inode_item_size(
*nvecs += 1;
}
break;
-
case XFS_DINODE_FMT_LOCAL:
if ((iip->ili_fields & XFS_ILOG_ADATA) &&
ip->i_afp->if_bytes > 0) {
@@ -129,7 +114,6 @@ xfs_inode_item_size(
*nvecs += 1;
}
break;
-
default:
ASSERT(0);
break;
@@ -137,98 +121,67 @@ xfs_inode_item_size(
}
/*
- * xfs_inode_item_format_extents - convert in-core extents to on-disk form
- *
- * For either the data or attr fork in extent format, we need to endian convert
- * the in-core extent as we place them into the on-disk inode. In this case, we
- * need to do this conversion before we write the extents into the log. Because
- * we don't have the disk inode to write into here, we allocate a buffer and
- * format the extents into it via xfs_iextents_copy(). We free the buffer in
- * the unlock routine after the copy for the log has been made.
+ * This returns the number of iovecs needed to log the given inode item.
*
- * In the case of the data fork, the in-core and on-disk fork sizes can be
- * different due to delayed allocation extents. We only log on-disk extents
- * here, so always use the physical fork size to determine the size of the
- * buffer we need to allocate.
+ * We need one iovec for the inode log format structure, one for the
+ * inode core, and possibly one for the inode data/extents/b-tree root
+ * and one for the inode attribute data/extents/b-tree root.
*/
STATIC void
-xfs_inode_item_format_extents(
- struct xfs_inode *ip,
- struct xfs_log_iovec *vecp,
- int whichfork,
- int type)
+xfs_inode_item_size(
+ struct xfs_log_item *lip,
+ int *nvecs,
+ int *nbytes)
{
- xfs_bmbt_rec_t *ext_buffer;
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
- ext_buffer = kmem_alloc(XFS_IFORK_SIZE(ip, whichfork), KM_SLEEP);
- if (whichfork == XFS_DATA_FORK)
- ip->i_itemp->ili_extents_buf = ext_buffer;
- else
- ip->i_itemp->ili_aextents_buf = ext_buffer;
+ *nvecs += 2;
+ *nbytes += sizeof(struct xfs_inode_log_format) +
+ xfs_icdinode_size(ip->i_d.di_version);
- vecp->i_addr = ext_buffer;
- vecp->i_len = xfs_iextents_copy(ip, ext_buffer, whichfork);
- vecp->i_type = type;
+ xfs_inode_item_data_fork_size(iip, nvecs, nbytes);
+ if (XFS_IFORK_Q(ip))
+ xfs_inode_item_attr_fork_size(iip, nvecs, nbytes);
}
/*
- * This is called to fill in the vector of log iovecs for the
- * given inode log item. It fills the first item with an inode
- * log format structure, the second with the on-disk inode structure,
- * and a possible third and/or fourth with the inode data/extents/b-tree
- * root and inode attributes data/extents/b-tree root.
+ * If this is a v1 format inode, then we need to log it as such. This means
+ * that we have to copy the link count from the new field to the old. We
+ * don't have to worry about the new fields, because nothing trusts them as
+ * long as the old inode version number is there.
*/
STATIC void
-xfs_inode_item_format(
- struct xfs_log_item *lip,
- struct xfs_log_iovec *vecp)
+xfs_inode_item_format_v1_inode(
+ struct xfs_inode *ip)
+{
+ if (!xfs_sb_version_hasnlink(&ip->i_mount->m_sb)) {
+ /*
+ * Convert it back.
+ */
+ ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
+ ip->i_d.di_onlink = ip->i_d.di_nlink;
+ } else {
+ /*
+ * The superblock version has already been bumped,
+ * so just make the conversion to the new inode
+ * format permanent.
+ */
+ ip->i_d.di_version = 2;
+ ip->i_d.di_onlink = 0;
+ memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
+ }
+}
+
+STATIC void
+xfs_inode_item_format_data_fork(
+ struct xfs_inode_log_item *iip,
+ struct xfs_inode_log_format *ilf,
+ struct xfs_log_vec *lv,
+ struct xfs_log_iovec **vecp)
{
- struct xfs_inode_log_item *iip = INODE_ITEM(lip);
struct xfs_inode *ip = iip->ili_inode;
- uint nvecs;
size_t data_bytes;
- xfs_mount_t *mp;
-
- vecp->i_addr = &iip->ili_format;
- vecp->i_len = sizeof(xfs_inode_log_format_t);
- vecp->i_type = XLOG_REG_TYPE_IFORMAT;
- vecp++;
- nvecs = 1;
-
- vecp->i_addr = &ip->i_d;
- vecp->i_len = xfs_icdinode_size(ip->i_d.di_version);
- vecp->i_type = XLOG_REG_TYPE_ICORE;
- vecp++;
- nvecs++;
-
- /*
- * If this is really an old format inode, then we need to
- * log it as such. This means that we have to copy the link
- * count from the new field to the old. We don't have to worry
- * about the new fields, because nothing trusts them as long as
- * the old inode version number is there. If the superblock already
- * has a new version number, then we don't bother converting back.
- */
- mp = ip->i_mount;
- ASSERT(ip->i_d.di_version == 1 || xfs_sb_version_hasnlink(&mp->m_sb));
- if (ip->i_d.di_version == 1) {
- if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
- /*
- * Convert it back.
- */
- ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
- ip->i_d.di_onlink = ip->i_d.di_nlink;
- } else {
- /*
- * The superblock version has already been bumped,
- * so just make the conversion to the new inode
- * format permanent.
- */
- ip->i_d.di_version = 2;
- ip->i_d.di_onlink = 0;
- memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
- }
- }
switch (ip->i_d.di_format) {
case XFS_DINODE_FMT_EXTENTS:
@@ -239,36 +192,23 @@ xfs_inode_item_format(
if ((iip->ili_fields & XFS_ILOG_DEXT) &&
ip->i_d.di_nextents > 0 &&
ip->i_df.if_bytes > 0) {
+ struct xfs_bmbt_rec *p;
+
ASSERT(ip->i_df.if_u1.if_extents != NULL);
ASSERT(ip->i_df.if_bytes / sizeof(xfs_bmbt_rec_t) > 0);
- ASSERT(iip->ili_extents_buf == NULL);
-
-#ifdef XFS_NATIVE_HOST
- if (ip->i_d.di_nextents == ip->i_df.if_bytes /
- (uint)sizeof(xfs_bmbt_rec_t)) {
- /*
- * There are no delayed allocation
- * extents, so just point to the
- * real extents array.
- */
- vecp->i_addr = ip->i_df.if_u1.if_extents;
- vecp->i_len = ip->i_df.if_bytes;
- vecp->i_type = XLOG_REG_TYPE_IEXT;
- } else
-#endif
- {
- xfs_inode_item_format_extents(ip, vecp,
- XFS_DATA_FORK, XLOG_REG_TYPE_IEXT);
- }
- ASSERT(vecp->i_len <= ip->i_df.if_bytes);
- iip->ili_format.ilf_dsize = vecp->i_len;
- vecp++;
- nvecs++;
+
+ p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IEXT);
+ data_bytes = xfs_iextents_copy(ip, p, XFS_DATA_FORK);
+ xlog_finish_iovec(lv, *vecp, data_bytes);
+
+ ASSERT(data_bytes <= ip->i_df.if_bytes);
+
+ ilf->ilf_dsize = data_bytes;
+ ilf->ilf_size++;
} else {
iip->ili_fields &= ~XFS_ILOG_DEXT;
}
break;
-
case XFS_DINODE_FMT_BTREE:
iip->ili_fields &=
~(XFS_ILOG_DDATA | XFS_ILOG_DEXT |
@@ -277,80 +217,70 @@ xfs_inode_item_format(
if ((iip->ili_fields & XFS_ILOG_DBROOT) &&
ip->i_df.if_broot_bytes > 0) {
ASSERT(ip->i_df.if_broot != NULL);
- vecp->i_addr = ip->i_df.if_broot;
- vecp->i_len = ip->i_df.if_broot_bytes;
- vecp->i_type = XLOG_REG_TYPE_IBROOT;
- vecp++;
- nvecs++;
- iip->ili_format.ilf_dsize = ip->i_df.if_broot_bytes;
+ xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IBROOT,
+ ip->i_df.if_broot,
+ ip->i_df.if_broot_bytes);
+ ilf->ilf_dsize = ip->i_df.if_broot_bytes;
+ ilf->ilf_size++;
} else {
ASSERT(!(iip->ili_fields &
XFS_ILOG_DBROOT));
iip->ili_fields &= ~XFS_ILOG_DBROOT;
}
break;
-
case XFS_DINODE_FMT_LOCAL:
iip->ili_fields &=
~(XFS_ILOG_DEXT | XFS_ILOG_DBROOT |
XFS_ILOG_DEV | XFS_ILOG_UUID);
if ((iip->ili_fields & XFS_ILOG_DDATA) &&
ip->i_df.if_bytes > 0) {
- ASSERT(ip->i_df.if_u1.if_data != NULL);
- ASSERT(ip->i_d.di_size > 0);
-
- vecp->i_addr = ip->i_df.if_u1.if_data;
/*
* Round i_bytes up to a word boundary.
* The underlying memory is guaranteed to
* to be there by xfs_idata_realloc().
*/
data_bytes = roundup(ip->i_df.if_bytes, 4);
- ASSERT((ip->i_df.if_real_bytes == 0) ||
- (ip->i_df.if_real_bytes == data_bytes));
- vecp->i_len = (int)data_bytes;
- vecp->i_type = XLOG_REG_TYPE_ILOCAL;
- vecp++;
- nvecs++;
- iip->ili_format.ilf_dsize = (unsigned)data_bytes;
+ ASSERT(ip->i_df.if_real_bytes == 0 ||
+ ip->i_df.if_real_bytes == data_bytes);
+ ASSERT(ip->i_df.if_u1.if_data != NULL);
+ ASSERT(ip->i_d.di_size > 0);
+ xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_ILOCAL,
+ ip->i_df.if_u1.if_data, data_bytes);
+ ilf->ilf_dsize = (unsigned)data_bytes;
+ ilf->ilf_size++;
} else {
iip->ili_fields &= ~XFS_ILOG_DDATA;
}
break;
-
case XFS_DINODE_FMT_DEV:
iip->ili_fields &=
~(XFS_ILOG_DDATA | XFS_ILOG_DBROOT |
XFS_ILOG_DEXT | XFS_ILOG_UUID);
- if (iip->ili_fields & XFS_ILOG_DEV) {
- iip->ili_format.ilf_u.ilfu_rdev =
- ip->i_df.if_u2.if_rdev;
- }
+ if (iip->ili_fields & XFS_ILOG_DEV)
+ ilf->ilf_u.ilfu_rdev = ip->i_df.if_u2.if_rdev;
break;
-
case XFS_DINODE_FMT_UUID:
iip->ili_fields &=
~(XFS_ILOG_DDATA | XFS_ILOG_DBROOT |
XFS_ILOG_DEXT | XFS_ILOG_DEV);
- if (iip->ili_fields & XFS_ILOG_UUID) {
- iip->ili_format.ilf_u.ilfu_uuid =
- ip->i_df.if_u2.if_uuid;
- }
+ if (iip->ili_fields & XFS_ILOG_UUID)
+ ilf->ilf_u.ilfu_uuid = ip->i_df.if_u2.if_uuid;
break;
-
default:
ASSERT(0);
break;
}
+}
- /*
- * If there are no attributes associated with the file, then we're done.
- */
- if (!XFS_IFORK_Q(ip)) {
- iip->ili_fields &=
- ~(XFS_ILOG_ADATA | XFS_ILOG_ABROOT | XFS_ILOG_AEXT);
- goto out;
- }
+STATIC void
+xfs_inode_item_format_attr_fork(
+ struct xfs_inode_log_item *iip,
+ struct xfs_inode_log_format *ilf,
+ struct xfs_log_vec *lv,
+ struct xfs_log_iovec **vecp)
+{
+ struct xfs_inode *ip = iip->ili_inode;
+ size_t data_bytes;
switch (ip->i_d.di_aformat) {
case XFS_DINODE_FMT_EXTENTS:
@@ -360,30 +290,22 @@ xfs_inode_item_format(
if ((iip->ili_fields & XFS_ILOG_AEXT) &&
ip->i_d.di_anextents > 0 &&
ip->i_afp->if_bytes > 0) {
+ struct xfs_bmbt_rec *p;
+
ASSERT(ip->i_afp->if_bytes / sizeof(xfs_bmbt_rec_t) ==
ip->i_d.di_anextents);
ASSERT(ip->i_afp->if_u1.if_extents != NULL);
-#ifdef XFS_NATIVE_HOST
- /*
- * There are not delayed allocation extents
- * for attributes, so just point at the array.
- */
- vecp->i_addr = ip->i_afp->if_u1.if_extents;
- vecp->i_len = ip->i_afp->if_bytes;
- vecp->i_type = XLOG_REG_TYPE_IATTR_EXT;
-#else
- ASSERT(iip->ili_aextents_buf == NULL);
- xfs_inode_item_format_extents(ip, vecp,
- XFS_ATTR_FORK, XLOG_REG_TYPE_IATTR_EXT);
-#endif
- iip->ili_format.ilf_asize = vecp->i_len;
- vecp++;
- nvecs++;
+
+ p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_EXT);
+ data_bytes = xfs_iextents_copy(ip, p, XFS_ATTR_FORK);
+ xlog_finish_iovec(lv, *vecp, data_bytes);
+
+ ilf->ilf_asize = data_bytes;
+ ilf->ilf_size++;
} else {
iip->ili_fields &= ~XFS_ILOG_AEXT;
}
break;
-
case XFS_DINODE_FMT_BTREE:
iip->ili_fields &=
~(XFS_ILOG_ADATA | XFS_ILOG_AEXT);
@@ -392,61 +314,89 @@ xfs_inode_item_format(
ip->i_afp->if_broot_bytes > 0) {
ASSERT(ip->i_afp->if_broot != NULL);
- vecp->i_addr = ip->i_afp->if_broot;
- vecp->i_len = ip->i_afp->if_broot_bytes;
- vecp->i_type = XLOG_REG_TYPE_IATTR_BROOT;
- vecp++;
- nvecs++;
- iip->ili_format.ilf_asize = ip->i_afp->if_broot_bytes;
+ xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_BROOT,
+ ip->i_afp->if_broot,
+ ip->i_afp->if_broot_bytes);
+ ilf->ilf_asize = ip->i_afp->if_broot_bytes;
+ ilf->ilf_size++;
} else {
iip->ili_fields &= ~XFS_ILOG_ABROOT;
}
break;
-
case XFS_DINODE_FMT_LOCAL:
iip->ili_fields &=
~(XFS_ILOG_AEXT | XFS_ILOG_ABROOT);
if ((iip->ili_fields & XFS_ILOG_ADATA) &&
ip->i_afp->if_bytes > 0) {
- ASSERT(ip->i_afp->if_u1.if_data != NULL);
-
- vecp->i_addr = ip->i_afp->if_u1.if_data;
/*
* Round i_bytes up to a word boundary.
* The underlying memory is guaranteed to
* to be there by xfs_idata_realloc().
*/
data_bytes = roundup(ip->i_afp->if_bytes, 4);
- ASSERT((ip->i_afp->if_real_bytes == 0) ||
- (ip->i_afp->if_real_bytes == data_bytes));
- vecp->i_len = (int)data_bytes;
- vecp->i_type = XLOG_REG_TYPE_IATTR_LOCAL;
- vecp++;
- nvecs++;
- iip->ili_format.ilf_asize = (unsigned)data_bytes;
+ ASSERT(ip->i_afp->if_real_bytes == 0 ||
+ ip->i_afp->if_real_bytes == data_bytes);
+ ASSERT(ip->i_afp->if_u1.if_data != NULL);
+ xlog_copy_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_LOCAL,
+ ip->i_afp->if_u1.if_data,
+ data_bytes);
+ ilf->ilf_asize = (unsigned)data_bytes;
+ ilf->ilf_size++;
} else {
iip->ili_fields &= ~XFS_ILOG_ADATA;
}
break;
-
default:
ASSERT(0);
break;
}
-
-out:
- /*
- * Now update the log format that goes out to disk from the in-core
- * values. We always write the inode core to make the arithmetic
- * games in recovery easier, which isn't a big deal as just about any
- * transaction would dirty it anyway.
- */
- iip->ili_format.ilf_fields = XFS_ILOG_CORE |
- (iip->ili_fields & ~XFS_ILOG_TIMESTAMP);
- iip->ili_format.ilf_size = nvecs;
}
+/*
+ * This is called to fill in the vector of log iovecs for the given inode
+ * log item. It fills the first item with an inode log format structure,
+ * the second with the on-disk inode structure, and a possible third and/or
+ * fourth with the inode data/extents/b-tree root and inode attributes
+ * data/extents/b-tree root.
+ */
+STATIC void
+xfs_inode_item_format(
+ struct xfs_log_item *lip,
+ struct xfs_log_vec *lv)
+{
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
+ struct xfs_inode_log_format *ilf;
+ struct xfs_log_iovec *vecp = NULL;
+
+ ilf = xlog_prepare_iovec(lv, &vecp, XLOG_REG_TYPE_IFORMAT);
+ ilf->ilf_type = XFS_LI_INODE;
+ ilf->ilf_ino = ip->i_ino;
+ ilf->ilf_blkno = ip->i_imap.im_blkno;
+ ilf->ilf_len = ip->i_imap.im_len;
+ ilf->ilf_boffset = ip->i_imap.im_boffset;
+ ilf->ilf_fields = XFS_ILOG_CORE;
+ ilf->ilf_size = 2; /* format + core */
+ xlog_finish_iovec(lv, vecp, sizeof(struct xfs_inode_log_format));
+
+ if (ip->i_d.di_version == 1)
+ xfs_inode_item_format_v1_inode(ip);
+ xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_ICORE,
+ &ip->i_d,
+ xfs_icdinode_size(ip->i_d.di_version));
+
+ xfs_inode_item_format_data_fork(iip, ilf, lv, &vecp);
+ if (XFS_IFORK_Q(ip)) {
+ xfs_inode_item_format_attr_fork(iip, ilf, lv, &vecp);
+ } else {
+ iip->ili_fields &=
+ ~(XFS_ILOG_ADATA | XFS_ILOG_ABROOT | XFS_ILOG_AEXT);
+ }
+
+ /* update the format with the exact fields we actually logged */
+ ilf->ilf_fields |= (iip->ili_fields & ~XFS_ILOG_TIMESTAMP);
+}
/*
* This is called to pin the inode associated with the inode log
@@ -563,27 +513,6 @@ xfs_inode_item_unlock(
ASSERT(ip->i_itemp != NULL);
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
- /*
- * If the inode needed a separate buffer with which to log
- * its extents, then free it now.
- */
- if (iip->ili_extents_buf != NULL) {
- ASSERT(ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS);
- ASSERT(ip->i_d.di_nextents > 0);
- ASSERT(iip->ili_fields & XFS_ILOG_DEXT);
- ASSERT(ip->i_df.if_bytes > 0);
- kmem_free(iip->ili_extents_buf);
- iip->ili_extents_buf = NULL;
- }
- if (iip->ili_aextents_buf != NULL) {
- ASSERT(ip->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS);
- ASSERT(ip->i_d.di_anextents > 0);
- ASSERT(iip->ili_fields & XFS_ILOG_AEXT);
- ASSERT(ip->i_afp->if_bytes > 0);
- kmem_free(iip->ili_aextents_buf);
- iip->ili_aextents_buf = NULL;
- }
-
lock_flags = iip->ili_lock_flags;
iip->ili_lock_flags = 0;
if (lock_flags)
@@ -670,11 +599,6 @@ xfs_inode_item_init(
iip->ili_inode = ip;
xfs_log_item_init(mp, &iip->ili_item, XFS_LI_INODE,
&xfs_inode_item_ops);
- iip->ili_format.ilf_type = XFS_LI_INODE;
- iip->ili_format.ilf_ino = ip->i_ino;
- iip->ili_format.ilf_blkno = ip->i_imap.im_blkno;
- iip->ili_format.ilf_len = ip->i_imap.im_len;
- iip->ili_format.ilf_boffset = ip->i_imap.im_boffset;
}
/*
diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h
index dce4d656768c..488d81254e28 100644
--- a/fs/xfs/xfs_inode_item.h
+++ b/fs/xfs/xfs_inode_item.h
@@ -34,11 +34,6 @@ typedef struct xfs_inode_log_item {
unsigned short ili_logged; /* flushed logged data */
unsigned int ili_last_fields; /* fields when flushed */
unsigned int ili_fields; /* fields to be logged */
- struct xfs_bmbt_rec *ili_extents_buf; /* array of logged
- data exts */
- struct xfs_bmbt_rec *ili_aextents_buf; /* array of logged
- attr exts */
- xfs_inode_log_format_t ili_format; /* logged structure */
} xfs_inode_log_item_t;
static inline int xfs_inode_clean(xfs_inode_t *ip)
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index c237ad15d500..f46338285152 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -209,9 +209,8 @@ xfs_bulkstat(
xfs_inobt_rec_incore_t *irbuf; /* start of irec buffer */
xfs_inobt_rec_incore_t *irbufend; /* end of good irec buffer entries */
xfs_ino_t lastino; /* last inode number returned */
- int nbcluster; /* # of blocks in a cluster */
- int nicluster; /* # of inodes in a cluster */
- int nimask; /* mask for inode clusters */
+ int blks_per_cluster; /* # of blocks per cluster */
+ int inodes_per_cluster;/* # of inodes per cluster */
int nirbuf; /* size of irbuf */
int rval; /* return value error code */
int tmp; /* result value from btree calls */
@@ -243,11 +242,8 @@ xfs_bulkstat(
*done = 0;
fmterror = 0;
ubufp = ubuffer;
- nicluster = mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp) ?
- mp->m_sb.sb_inopblock :
- (XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog);
- nimask = ~(nicluster - 1);
- nbcluster = nicluster >> mp->m_sb.sb_inopblog;
+ blks_per_cluster = xfs_icluster_size_fsb(mp);
+ inodes_per_cluster = blks_per_cluster << mp->m_sb.sb_inopblog;
irbuf = kmem_zalloc_greedy(&irbsize, PAGE_SIZE, PAGE_SIZE * 4);
if (!irbuf)
return ENOMEM;
@@ -390,12 +386,12 @@ xfs_bulkstat(
agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino);
for (chunkidx = 0;
chunkidx < XFS_INODES_PER_CHUNK;
- chunkidx += nicluster,
- agbno += nbcluster) {
- if (xfs_inobt_maskn(chunkidx, nicluster)
- & ~r.ir_free)
+ chunkidx += inodes_per_cluster,
+ agbno += blks_per_cluster) {
+ if (xfs_inobt_maskn(chunkidx,
+ inodes_per_cluster) & ~r.ir_free)
xfs_btree_reada_bufs(mp, agno,
- agbno, nbcluster,
+ agbno, blks_per_cluster,
&xfs_inode_buf_ops);
}
blk_finish_plug(&plug);
diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h
index e148719e0a5d..b0f4ef77fa70 100644
--- a/fs/xfs/xfs_log.h
+++ b/fs/xfs/xfs_log.h
@@ -30,6 +30,52 @@ struct xfs_log_vec {
#define XFS_LOG_VEC_ORDERED (-1)
+static inline void *
+xlog_prepare_iovec(struct xfs_log_vec *lv, struct xfs_log_iovec **vecp,
+ uint type)
+{
+ struct xfs_log_iovec *vec = *vecp;
+
+ if (vec) {
+ ASSERT(vec - lv->lv_iovecp < lv->lv_niovecs);
+ vec++;
+ } else {
+ vec = &lv->lv_iovecp[0];
+ }
+
+ vec->i_type = type;
+ vec->i_addr = lv->lv_buf + lv->lv_buf_len;
+
+ ASSERT(IS_ALIGNED((unsigned long)vec->i_addr, sizeof(uint64_t)));
+
+ *vecp = vec;
+ return vec->i_addr;
+}
+
+static inline void
+xlog_finish_iovec(struct xfs_log_vec *lv, struct xfs_log_iovec *vec, int len)
+{
+ /*
+ * We need to make sure the next buffer is naturally aligned for the
+ * biggest basic data type we put into it. We already accounted for
+ * this when sizing the buffer.
+ */
+ lv->lv_buf_len += round_up(len, sizeof(uint64_t));
+ vec->i_len = len;
+}
+
+static inline void *
+xlog_copy_iovec(struct xfs_log_vec *lv, struct xfs_log_iovec **vecp,
+ uint type, void *data, int len)
+{
+ void *buf;
+
+ buf = xlog_prepare_iovec(lv, vecp, type);
+ memcpy(buf, data, len);
+ xlog_finish_iovec(lv, *vecp, len);
+ return buf;
+}
+
/*
* Structure used to pass callback function and the function's argument
* to the log manager.
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
index 5eb51fc5eb84..cdebd832c3db 100644
--- a/fs/xfs/xfs_log_cil.c
+++ b/fs/xfs/xfs_log_cil.c
@@ -82,36 +82,6 @@ xlog_cil_init_post_recovery(
log->l_curr_block);
}
-STATIC int
-xlog_cil_lv_item_format(
- struct xfs_log_item *lip,
- struct xfs_log_vec *lv)
-{
- int index;
- char *ptr;
-
- /* format new vectors into array */
- lip->li_ops->iop_format(lip, lv->lv_iovecp);
-
- /* copy data into existing array */
- ptr = lv->lv_buf;
- for (index = 0; index < lv->lv_niovecs; index++) {
- struct xfs_log_iovec *vec = &lv->lv_iovecp[index];
-
- memcpy(ptr, vec->i_addr, vec->i_len);
- vec->i_addr = ptr;
- ptr += vec->i_len;
- }
-
- /*
- * some size calculations for log vectors over-estimate, so the caller
- * doesn't know the amount of space actually used by the item. Return
- * the byte count to the caller so they can check and store it
- * appropriately.
- */
- return ptr - lv->lv_buf;
-}
-
/*
* Prepare the log item for insertion into the CIL. Calculate the difference in
* log space and vectors it will consume, and if it is a new item pin it as
@@ -232,6 +202,13 @@ xlog_cil_insert_format_items(
nbytes = 0;
}
+ /*
+ * We 64-bit align the length of each iovec so that the start
+ * of the next one is naturally aligned. We'll need to
+ * account for that slack space here.
+ */
+ nbytes += niovecs * sizeof(uint64_t);
+
/* grab the old item if it exists for reservation accounting */
old_lv = lip->li_lv;
@@ -254,34 +231,27 @@ xlog_cil_insert_format_items(
*/
*diff_iovecs -= lv->lv_niovecs;
*diff_len -= lv->lv_buf_len;
-
- /* Ensure the lv is set up according to ->iop_size */
- lv->lv_niovecs = niovecs;
- lv->lv_buf = (char *)lv + buf_size - nbytes;
-
- lv->lv_buf_len = xlog_cil_lv_item_format(lip, lv);
- goto insert;
+ } else {
+ /* allocate new data chunk */
+ lv = kmem_zalloc(buf_size, KM_SLEEP|KM_NOFS);
+ lv->lv_item = lip;
+ lv->lv_size = buf_size;
+ if (ordered) {
+ /* track as an ordered logvec */
+ ASSERT(lip->li_lv == NULL);
+ lv->lv_buf_len = XFS_LOG_VEC_ORDERED;
+ goto insert;
+ }
+ lv->lv_iovecp = (struct xfs_log_iovec *)&lv[1];
}
- /* allocate new data chunk */
- lv = kmem_zalloc(buf_size, KM_SLEEP|KM_NOFS);
- lv->lv_item = lip;
- lv->lv_size = buf_size;
+ /* Ensure the lv is set up according to ->iop_size */
lv->lv_niovecs = niovecs;
- if (ordered) {
- /* track as an ordered logvec */
- ASSERT(lip->li_lv == NULL);
- lv->lv_buf_len = XFS_LOG_VEC_ORDERED;
- goto insert;
- }
-
- /* The allocated iovec region lies beyond the log vector. */
- lv->lv_iovecp = (struct xfs_log_iovec *)&lv[1];
/* The allocated data region lies beyond the iovec region */
+ lv->lv_buf_len = 0;
lv->lv_buf = (char *)lv + buf_size - nbytes;
-
- lv->lv_buf_len = xlog_cil_lv_item_format(lip, lv);
+ lip->li_ops->iop_format(lip, lv);
insert:
ASSERT(lv->lv_buf_len <= nbytes);
xfs_cil_prepare_item(log, lv, old_lv, diff_len, diff_iovecs);
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 07ab52ca8aba..bce53ac81096 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -193,7 +193,10 @@ xlog_bread_noalign(
bp->b_io_length = nbblks;
bp->b_error = 0;
- xfsbdstrat(log->l_mp, bp);
+ if (XFS_FORCED_SHUTDOWN(log->l_mp))
+ return XFS_ERROR(EIO);
+
+ xfs_buf_iorequest(bp);
error = xfs_buf_iowait(bp);
if (error)
xfs_buf_ioerror_alert(bp, __func__);
@@ -2523,19 +2526,19 @@ xlog_recover_buffer_pass2(
*
* Also make sure that only inode buffers with good sizes stay in
* the buffer cache. The kernel moves inodes in buffers of 1 block
- * or XFS_INODE_CLUSTER_SIZE bytes, whichever is bigger. The inode
+ * or mp->m_inode_cluster_size bytes, whichever is bigger. The inode
* buffers in the log can be a different size if the log was generated
* by an older kernel using unclustered inode buffers or a newer kernel
* running with a different inode cluster size. Regardless, if the
- * the inode buffer size isn't MAX(blocksize, XFS_INODE_CLUSTER_SIZE)
- * for *our* value of XFS_INODE_CLUSTER_SIZE, then we need to keep
+ * the inode buffer size isn't MAX(blocksize, mp->m_inode_cluster_size)
+ * for *our* value of mp->m_inode_cluster_size, then we need to keep
* the buffer out of the buffer cache so that the buffer won't
* overlap with future reads of those inodes.
*/
if (XFS_DINODE_MAGIC ==
be16_to_cpu(*((__be16 *)xfs_buf_offset(bp, 0))) &&
(BBTOB(bp->b_io_length) != MAX(log->l_mp->m_sb.sb_blocksize,
- (__uint32_t)XFS_INODE_CLUSTER_SIZE(log->l_mp)))) {
+ (__uint32_t)log->l_mp->m_inode_cluster_size))) {
xfs_buf_stale(bp);
error = xfs_bwrite(bp);
} else {
@@ -3208,10 +3211,10 @@ xlog_recover_do_icreate_pass2(
}
/* existing allocation is fixed value */
- ASSERT(count == XFS_IALLOC_INODES(mp));
- ASSERT(length == XFS_IALLOC_BLOCKS(mp));
- if (count != XFS_IALLOC_INODES(mp) ||
- length != XFS_IALLOC_BLOCKS(mp)) {
+ ASSERT(count == mp->m_ialloc_inos);
+ ASSERT(length == mp->m_ialloc_blks);
+ if (count != mp->m_ialloc_inos ||
+ length != mp->m_ialloc_blks) {
xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad count 2");
return EINVAL;
}
@@ -4408,7 +4411,13 @@ xlog_do_recover(
XFS_BUF_READ(bp);
XFS_BUF_UNASYNC(bp);
bp->b_ops = &xfs_sb_buf_ops;
- xfsbdstrat(log->l_mp, bp);
+
+ if (XFS_FORCED_SHUTDOWN(log->l_mp)) {
+ xfs_buf_relse(bp);
+ return XFS_ERROR(EIO);
+ }
+
+ xfs_buf_iorequest(bp);
error = xfs_buf_iowait(bp);
if (error) {
xfs_buf_ioerror_alert(bp, __func__);
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 9b96d35e483d..b5bc1ab3c4da 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -64,7 +64,7 @@ typedef struct xfs_log_item {
struct xfs_item_ops {
void (*iop_size)(xfs_log_item_t *, int *, int *);
- void (*iop_format)(xfs_log_item_t *, struct xfs_log_iovec *);
+ void (*iop_format)(xfs_log_item_t *, struct xfs_log_vec *);
void (*iop_pin)(xfs_log_item_t *);
void (*iop_unpin)(xfs_log_item_t *, int remove);
uint (*iop_push)(struct xfs_log_item *, struct list_head *);
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index c035d11b7734..647b6f1d8923 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -314,7 +314,18 @@ xfs_trans_read_buf_map(
ASSERT(bp->b_iodone == NULL);
XFS_BUF_READ(bp);
bp->b_ops = ops;
- xfsbdstrat(tp->t_mountp, bp);
+
+ /*
+ * XXX(hch): clean up the error handling here to be less
+ * of a mess..
+ */
+ if (XFS_FORCED_SHUTDOWN(mp)) {
+ trace_xfs_bdstrat_shut(bp, _RET_IP_);
+ xfs_bioerror_relse(bp);
+ } else {
+ xfs_buf_iorequest(bp);
+ }
+
error = xfs_buf_iowait(bp);
if (error) {
xfs_buf_ioerror_alert(bp, __func__);
diff --git a/fs/xfs/xfs_trans_resv.c b/fs/xfs/xfs_trans_resv.c
index 2fd59c0dae66..2ffd3e331b49 100644
--- a/fs/xfs/xfs_trans_resv.c
+++ b/fs/xfs/xfs_trans_resv.c
@@ -174,7 +174,7 @@ xfs_calc_itruncate_reservation(
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) +
+ xfs_calc_buf_res(2 + mp->m_ialloc_blks +
mp->m_in_maxlevels, 0)));
}
@@ -282,7 +282,7 @@ xfs_calc_create_resv_modify(
* 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 blocks allocated: mp->m_ialloc_blks * blocksize
* the inode btree: max depth * blocksize
* the allocation btrees: 2 trees * (max depth - 1) * block size
*/
@@ -292,7 +292,7 @@ xfs_calc_create_resv_alloc(
{
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_ialloc_blks, 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));
@@ -385,9 +385,9 @@ xfs_calc_ifree_reservation(
xfs_calc_inode_res(mp, 1) +
xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
xfs_calc_buf_res(1, XFS_FSB_TO_B(mp, 1)) +
- max_t(uint, XFS_FSB_TO_B(mp, 1), XFS_INODE_CLUSTER_SIZE(mp)) +
+ max_t(uint, XFS_FSB_TO_B(mp, 1), mp->m_inode_cluster_size) +
xfs_calc_buf_res(1, 0) +
- xfs_calc_buf_res(2 + XFS_IALLOC_BLOCKS(mp) +
+ xfs_calc_buf_res(2 + mp->m_ialloc_blks +
mp->m_in_maxlevels, 0) +
xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
XFS_FSB_TO_B(mp, 1));
diff --git a/fs/xfs/xfs_trans_space.h b/fs/xfs/xfs_trans_space.h
index 7d2c920dfb9c..af5dbe06cb65 100644
--- a/fs/xfs/xfs_trans_space.h
+++ b/fs/xfs/xfs_trans_space.h
@@ -47,7 +47,7 @@
#define XFS_DIRREMOVE_SPACE_RES(mp) \
XFS_DAREMOVE_SPACE_RES(mp, XFS_DATA_FORK)
#define XFS_IALLOC_SPACE_RES(mp) \
- (XFS_IALLOC_BLOCKS(mp) + (mp)->m_in_maxlevels - 1)
+ ((mp)->m_ialloc_blks + (mp)->m_in_maxlevels - 1)
/*
* Space reservation values for various transactions.
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-30 15:23:21 +0000