P: Artem Bityutskiy
M: dedekind@infradead.org
P: Adrian Hunter
-M: ext-adrian.hunter@nokia.com
+M: adrian.hunter@nokia.com
L: linux-mtd@lists.infradead.org
T: git git://git.infradead.org/ubifs-2.6.git
W: http://www.linux-mtd.infradead.org/doc/ubifs.html
{
struct ubifs_wbuf *wbuf = container_of(timer, struct ubifs_wbuf, timer);
+ dbg_io("jhead %d", wbuf->jhead);
wbuf->need_sync = 1;
wbuf->c->need_wbuf_sync = 1;
ubifs_wake_up_bgt(wbuf->c);
{
ubifs_assert(!hrtimer_active(&wbuf->timer));
- if (!ktime_to_ns(wbuf->softlimit))
+ if (wbuf->no_timer)
return;
+ dbg_io("set timer for jhead %d, %llu-%llu millisecs", wbuf->jhead,
+ div_u64(ktime_to_ns(wbuf->softlimit), USEC_PER_SEC),
+ div_u64(ktime_to_ns(wbuf->softlimit) + wbuf->delta,
+ USEC_PER_SEC));
hrtimer_start_range_ns(&wbuf->timer, wbuf->softlimit, wbuf->delta,
HRTIMER_MODE_REL);
}
*/
static void cancel_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
{
- /*
- * If the syncer is waiting for the lock (from the background thread's
- * context) and another task is changing write-buffer then the syncing
- * should be canceled.
- */
+ if (wbuf->no_timer)
+ return;
wbuf->need_sync = 0;
hrtimer_cancel(&wbuf->timer);
}
/* Write-buffer is empty or not seeked */
return 0;
- dbg_io("LEB %d:%d, %d bytes",
- wbuf->lnum, wbuf->offs, wbuf->used);
+ dbg_io("LEB %d:%d, %d bytes, jhead %d",
+ wbuf->lnum, wbuf->offs, wbuf->used, wbuf->jhead);
ubifs_assert(!(c->vfs_sb->s_flags & MS_RDONLY));
ubifs_assert(!(wbuf->avail & 7));
ubifs_assert(wbuf->offs + c->min_io_size <= c->leb_size);
* @offs: logical eraseblock offset to seek to
* @dtype: data type
*
- * This function targets the write buffer to logical eraseblock @lnum:@offs.
+ * This function targets the write-buffer to logical eraseblock @lnum:@offs.
* The write-buffer is synchronized if it is not empty. Returns zero in case of
* success and a negative error code in case of failure.
*/
{
const struct ubifs_info *c = wbuf->c;
- dbg_io("LEB %d:%d", lnum, offs);
+ dbg_io("LEB %d:%d, jhead %d", lnum, offs, wbuf->jhead);
ubifs_assert(lnum >= 0 && lnum < c->leb_cnt);
ubifs_assert(offs >= 0 && offs <= c->leb_size);
ubifs_assert(offs % c->min_io_size == 0 && !(offs & 7));
struct ubifs_info *c = wbuf->c;
int err, written, n, aligned_len = ALIGN(len, 8), offs;
- dbg_io("%d bytes (%s) to wbuf at LEB %d:%d", len,
- dbg_ntype(((struct ubifs_ch *)buf)->node_type), wbuf->lnum,
- wbuf->offs + wbuf->used);
+ dbg_io("%d bytes (%s) to jhead %d wbuf at LEB %d:%d", len,
+ dbg_ntype(((struct ubifs_ch *)buf)->node_type), wbuf->jhead,
+ wbuf->lnum, wbuf->offs + wbuf->used);
ubifs_assert(len > 0 && wbuf->lnum >= 0 && wbuf->lnum < c->leb_cnt);
ubifs_assert(wbuf->offs >= 0 && wbuf->offs % c->min_io_size == 0);
ubifs_assert(!(wbuf->offs & 7) && wbuf->offs <= c->leb_size);
memcpy(wbuf->buf + wbuf->used, buf, len);
if (aligned_len == wbuf->avail) {
- dbg_io("flush wbuf to LEB %d:%d", wbuf->lnum,
- wbuf->offs);
+ dbg_io("flush jhead %d wbuf to LEB %d:%d",
+ wbuf->jhead, wbuf->lnum, wbuf->offs);
err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf,
wbuf->offs, c->min_io_size,
wbuf->dtype);
* minimal I/O unit. We have to fill and flush write-buffer and switch
* to the next min. I/O unit.
*/
- dbg_io("flush wbuf to LEB %d:%d", wbuf->lnum, wbuf->offs);
+ dbg_io("flush jhead %d wbuf to LEB %d:%d",
+ wbuf->jhead, wbuf->lnum, wbuf->offs);
memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs,
c->min_io_size, wbuf->dtype);
int err, rlen, overlap;
struct ubifs_ch *ch = buf;
- dbg_io("LEB %d:%d, %s, length %d", lnum, offs, dbg_ntype(type), len);
+ dbg_io("LEB %d:%d, %s, length %d, jhead %d", lnum, offs,
+ dbg_ntype(type), len, wbuf->jhead);
ubifs_assert(wbuf && lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
ubifs_assert(!(offs & 7) && offs < c->leb_size);
ubifs_assert(type >= 0 && type < UBIFS_NODE_TYPES_CNT);
* @c: UBIFS file-system description object
* @wbuf: write-buffer to initialize
*
- * This function initializes write buffer. Returns zero in case of success
+ * This function initializes write-buffer. Returns zero in case of success
* %-ENOMEM in case of failure.
*/
int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
{
size_t size;
- ktime_t hardlimit;
wbuf->buf = kmalloc(c->min_io_size, GFP_KERNEL);
if (!wbuf->buf)
hrtimer_init(&wbuf->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
wbuf->timer.function = wbuf_timer_callback_nolock;
- /*
- * Make write-buffer soft limit to be 20% of the hard limit. The
- * write-buffer timer is allowed to expire any time between the soft
- * and hard limits.
- */
- hardlimit = ktime_set(DEFAULT_WBUF_TIMEOUT_SECS, 0);
- wbuf->delta = (DEFAULT_WBUF_TIMEOUT_SECS * NSEC_PER_SEC) * 2 / 10;
- wbuf->softlimit = ktime_sub_ns(hardlimit, wbuf->delta);
- hrtimer_set_expires_range_ns(&wbuf->timer, wbuf->softlimit,
- wbuf->delta);
+ wbuf->softlimit = ktime_set(WBUF_TIMEOUT_SOFTLIMIT, 0);
+ wbuf->delta = WBUF_TIMEOUT_HARDLIMIT - WBUF_TIMEOUT_SOFTLIMIT;
+ wbuf->delta *= 1000000000ULL;
+ ubifs_assert(wbuf->delta <= ULONG_MAX);
return 0;
}
/**
* ubifs_wbuf_add_ino_nolock - add an inode number into the wbuf inode array.
- * @wbuf: the write-buffer whereto add
+ * @wbuf: the write-buffer where to add
* @inum: the inode number
*
* This function adds an inode number to the inode array of the write-buffer.
return 1;
}
+/**
+ * first_non_ff - find offset of the first non-0xff byte.
+ * @buf: buffer to search in
+ * @len: length of buffer
+ *
+ * This function returns offset of the first non-0xff byte in @buf or %-1 if
+ * the buffer contains only 0xff bytes.
+ */
+static int first_non_ff(void *buf, int len)
+{
+ uint8_t *p = buf;
+ int i;
+
+ for (i = 0; i < len; i++)
+ if (*p++ != 0xff)
+ return i;
+ return -1;
+}
+
/**
* get_master_node - get the last valid master node allowing for corruption.
* @c: UBIFS file-system description object
empty_offs = ALIGN(offs + 1, c->min_io_size);
check_len = c->leb_size - empty_offs;
p = buf + empty_offs - offs;
-
- for (; check_len > 0; check_len--)
- if (*p++ != 0xff)
- return 0;
- return 1;
+ return is_empty(p, check_len);
}
/**
*
* This function does a scan of a LEB, but caters for errors that might have
* been caused by the unclean unmount from which we are attempting to recover.
- *
- * This function returns %0 on success and a negative error code on failure.
+ * Returns %0 in case of success, %-EUCLEAN if an unrecoverable corruption is
+ * found, and a negative error code in case of failure.
*/
struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
int offs, void *sbuf, int grouped)
goto corrupted;
default:
dbg_err("unknown");
- goto corrupted;
+ err = -EINVAL;
+ goto error;
}
}
clean_buf(c, &buf, lnum, &offs, &len);
need_clean = 1;
} else {
- ubifs_err("corrupt empty space at LEB %d:%d",
- lnum, offs);
+ int corruption = first_non_ff(buf, len);
+
+ ubifs_err("corrupt empty space LEB %d:%d, corruption "
+ "starts at %d", lnum, offs, corruption);
+ /* Make sure we dump interesting non-0xFF data */
+ offs = corruption;
+ buf += corruption;
goto corrupted;
}
}
static int recover_head(const struct ubifs_info *c, int lnum, int offs,
void *sbuf)
{
- int len, err, need_clean = 0;
+ int len, err;
if (c->min_io_size > 1)
len = c->min_io_size;
/* Read at the head location and check it is empty flash */
err = ubi_read(c->ubi, lnum, sbuf, offs, len);
- if (err)
- need_clean = 1;
- else {
- uint8_t *p = sbuf;
-
- while (len--)
- if (*p++ != 0xff) {
- need_clean = 1;
- break;
- }
- }
-
- if (need_clean) {
+ if (err || !is_empty(sbuf, len)) {
dbg_rcvry("cleaning head at %d:%d", lnum, offs);
if (offs == 0)
return ubifs_leb_unmap(c, lnum);
dbg_mnt("replay log LEB %d:%d", lnum, offs);
sleb = ubifs_scan(c, lnum, offs, sbuf);
- if (IS_ERR(sleb)) {
- if (c->need_recovery)
- sleb = ubifs_recover_log_leb(c, lnum, offs, sbuf);
+ if (IS_ERR(sleb) ) {
+ if (PTR_ERR(sleb) != -EUCLEAN || !c->need_recovery)
+ return PTR_ERR(sleb);
+ sleb = ubifs_recover_log_leb(c, lnum, offs, sbuf);
if (IS_ERR(sleb))
return PTR_ERR(sleb);
}
return err;
out_dump:
- ubifs_err("log error detected while replying the log at LEB %d:%d",
+ ubifs_err("log error detected while replaying the log at LEB %d:%d",
lnum, offs + snod->offs);
dbg_dump_node(c, snod->node);
ubifs_scan_destroy(sleb);
{
int len;
- ubifs_err("corrupted data at LEB %d:%d", lnum, offs);
+ ubifs_err("corruption at LEB %d:%d", lnum, offs);
if (dbg_failure_mode)
return;
len = c->leb_size - offs;
- if (len > 4096)
- len = 4096;
+ if (len > 8192)
+ len = 8192;
dbg_err("first %d bytes from LEB %d:%d", len, lnum, offs);
print_hex_dump(KERN_DEBUG, "", DUMP_PREFIX_OFFSET, 32, 4, buf, len, 1);
}
* @sbuf: scan buffer (must be c->leb_size)
*
* This function scans LEB number @lnum and returns complete information about
- * its contents. Returns an error code in case of failure.
+ * its contents. Returns the scaned information in case of success and,
+ * %-EUCLEAN if the LEB neads recovery, and other negative error codes in case
+ * of failure.
*/
struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
int offs, void *sbuf)
cond_resched();
ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 0);
-
if (ret > 0) {
/* Padding bytes or a valid padding node */
offs += ret;
goto corrupted;
default:
dbg_err("unknown");
- goto corrupted;
+ err = -EINVAL;
+ goto error;
}
err = ubifs_add_snod(c, sleb, buf, offs);
len -= node_len;
}
- if (offs % c->min_io_size)
- goto corrupted;
+ if (offs % c->min_io_size) {
+ ubifs_err("empty space starts at non-aligned offset %d", offs);
+ goto corrupted;;
+ }
ubifs_end_scan(c, sleb, lnum, offs);
* does not need to be synchronized by timer.
*/
c->jheads[GCHD].wbuf.dtype = UBI_LONGTERM;
- c->jheads[GCHD].wbuf.softlimit = ktime_set(0, 0);
+ c->jheads[GCHD].wbuf.no_timer = 1;
return 0;
}
switch (token) {
/*
* %Opt_fast_unmount and %Opt_norm_unmount options are ignored.
- * We accepte them in order to be backware-compatible. But this
+ * We accept them in order to be backward-compatible. But this
* should be removed at some point.
*/
case Opt_fast_unmount:
if (err)
goto out_journal;
+ /* Calculate 'min_idx_lebs' after journal replay */
+ c->min_idx_lebs = ubifs_calc_min_idx_lebs(c);
+
err = ubifs_mount_orphans(c, c->need_recovery, mounted_read_only);
if (err)
goto out_orphans;
/* Synchronize write-buffers */
if (c->jheads)
- for (i = 0; i < c->jhead_cnt; i++) {
+ for (i = 0; i < c->jhead_cnt; i++)
ubifs_wbuf_sync(&c->jheads[i].wbuf);
- hrtimer_cancel(&c->jheads[i].wbuf.timer);
- }
/*
* On fatal errors c->ro_media is set to 1, in which case we do
err = bdi_init(&c->bdi);
if (err)
goto out_close;
- err = bdi_register(&c->bdi, NULL, "ubifs");
+ err = bdi_register(&c->bdi, NULL, "ubifs_%d_%d",
+ c->vi.ubi_num, c->vi.vol_id);
if (err)
goto out_bdi;
*/
#define BGT_NAME_PATTERN "ubifs_bgt%d_%d"
-/* Default write-buffer synchronization timeout in seconds */
-#define DEFAULT_WBUF_TIMEOUT_SECS 5
+/* Write-buffer synchronization timeout interval in seconds */
+#define WBUF_TIMEOUT_SOFTLIMIT 3
+#define WBUF_TIMEOUT_HARDLIMIT 5
/* Maximum possible inode number (only 32-bit inodes are supported now) */
#define MAX_INUM 0xFFFFFFFF
* @delta: hard and soft timeouts delta (the timer expire inteval is @softlimit
* and @softlimit + @delta)
* @timer: write-buffer timer
- * @need_sync: it is set if its timer expired and needs sync
+ * @no_timer: non-zero if this write-buffer does not have a timer
+ * @need_sync: non-zero if the timer expired and the wbuf needs sync'ing
* @next_ino: points to the next position of the following inode number
* @inodes: stores the inode numbers of the nodes which are in wbuf
*
ktime_t softlimit;
unsigned long long delta;
struct hrtimer timer;
- int need_sync;
+ unsigned int no_timer:1;
+ unsigned int need_sync:1;
int next_ino;
ino_t *inodes;
};