Commit 32fb9b57 authored by Tim Shimmin's avatar Tim Shimmin Committed by Nathan Scott

[XFS] Fix up the calculation of the reservation overhead to hopefully

include all the components which make up the transaction in the ondisk
log. Having this incomplete has shown up as problems on IRIX when some v2
log changes went in. The symptom was the msg of "xfs_log_write:
reservation ran out. Need to up reservation" and was seen on synchronous
writes on files with lots of holes (and therefore lots of extents).

SGI-PV: 931457
SGI-Modid: xfs-linux:xfs-kern:23095a
Signed-off-by: default avatarTim Shimmin <tes@sgi.com>
Signed-off-by: default avatarNathan Scott <nathans@sgi.com>
parent d52b44d0
......@@ -3180,28 +3180,56 @@ xlog_ticket_get(xlog_t *log,
*
* The following lines of code account for non-transaction data
* which occupy space in the on-disk log.
*/
*
* Normal form of a transaction is:
* <oph><trans-hdr><start-oph><reg1-oph><reg1><reg2-oph>...<commit-oph>
* and then there are LR hdrs, split-recs and roundoff at end of syncs.
*
* We need to account for all the leadup data and trailer data
* around the transaction data.
* And then we need to account for the worst case in terms of using
* more space.
* The worst case will happen if:
* - the placement of the transaction happens to be such that the
* roundoff is at its maximum
* - the transaction data is synced before the commit record is synced
* i.e. <transaction-data><roundoff> | <commit-rec><roundoff>
* Therefore the commit record is in its own Log Record.
* This can happen as the commit record is called with its
* own region to xlog_write().
* This then means that in the worst case, roundoff can happen for
* the commit-rec as well.
* The commit-rec is smaller than padding in this scenario and so it is
* not added separately.
*/
/* for trans header */
unit_bytes += sizeof(xlog_op_header_t);
unit_bytes += sizeof(xfs_trans_header_t);
/* for start-rec */
unit_bytes += sizeof(xlog_op_header_t);
/* for padding */
/* for LR headers */
num_headers = ((unit_bytes + log->l_iclog_size-1) >> log->l_iclog_size_log);
unit_bytes += log->l_iclog_hsize * num_headers;
/* for commit-rec LR header - note: padding will subsume the ophdr */
unit_bytes += log->l_iclog_hsize;
/* for split-recs - ophdrs added when data split over LRs */
unit_bytes += sizeof(xlog_op_header_t) * num_headers;
/* for roundoff padding for transaction data and one for commit record */
if (XFS_SB_VERSION_HASLOGV2(&log->l_mp->m_sb) &&
log->l_mp->m_sb.sb_logsunit > 1) {
/* log su roundoff */
unit_bytes += log->l_mp->m_sb.sb_logsunit;
unit_bytes += 2*log->l_mp->m_sb.sb_logsunit;
} else {
/* BB roundoff */
unit_bytes += BBSIZE;
unit_bytes += 2*BBSIZE;
}
/* for commit-rec */
unit_bytes += sizeof(xlog_op_header_t);
/* for LR headers */
num_headers = ((unit_bytes + log->l_iclog_size-1) >> log->l_iclog_size_log);
unit_bytes += log->l_iclog_hsize * num_headers;
tic->t_unit_res = unit_bytes;
tic->t_curr_res = unit_bytes;
tic->t_cnt = cnt;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment