Lines Matching full:we

69  * By covering, we mean changing the h_tail_lsn in the last on-disk
78  * might include space beyond the EOF. So if we just push the EOF a
86  * system is idle. We need two dummy transaction because the h_tail_lsn
98  *		we are done covering previous transactions.
99  *  NEED -- logging has occurred and we need a dummy transaction
101  *  DONE -- we were in the NEED state and have committed a dummy
103  *  NEED2 -- we detected that a dummy transaction has gone to the
105  *  DONE2 -- we committed a dummy transaction when in the NEED2 state.
107  * There are two places where we switch states:
109  * 1.) In xfs_sync, when we detect an idle log and are in NEED or NEED2.
110  *	We commit the dummy transaction and switch to DONE or DONE2,
111  *	respectively. In all other states, we don't do anything.
113  * 2.) When we finish writing the on-disk log (xlog_state_clean_log).
115  *	No matter what state we are in, if this isn't the dummy
117  *	So, if we aren't in the DONE or DONE2 states, the next state
118  *	is NEED. We can't be finishing a write of the dummy record
121  *	If we are in the DONE state and this was a write of the
122  *		dummy transaction, we move to NEED2.
124  *	If we are in the DONE2 state and this was a write of the
125  *		dummy transaction, we move to IDLE.
131  * This is why we have the NEED2 and DONE2 states before going idle.
148  * we don't care about.
191  * Because of cacheline contention on large machines, we need to separate
200  * so we need to ensure that these fields are located in separate cachelines.
201  * We'll put all the read-only and l_icloglock fields in the first cacheline,
259  * we can avoid the problem of having to hold out new transactions during a
260  * flush until we have a the commit record LSN of the checkpoint. We can
263  * checkpoint is still in the process of committing, we can block waiting for
284  * The amount of log space we allow the CIL to aggregate is difficult to size.
285  * Whatever we choose, we have to make sure we can get a reservation for the
288  * the log or induces too much latency when writing out through the iclogs. We
290  * context, so we need to decide which to use for limiting.
292  * Every log buffer we write out during a push needs a header reserved, which
293  * is at least one sector and more for v2 logs. Hence we need a reservation of
295  * 16KB of reservation per megabyte of delayed logging space we will consume,
305  * can block on log pushes. Hence if we have to regrant log space during a log
306  * push, we can deadlock.
308  * However, we can avoid this by use of a dynamic "reservation stealing"
311  * space needed by the checkpoint transaction. This means that we never need to
312  * specifically reserve space for the CIL checkpoint transaction, nor do we
317  * With dynamic reservations, we can effectively make up arbitrary limits for
319  * Recovery imposes a rule that no transaction exceed half the log, so we are
321  * tries to keep 25% of the log free, so we need to keep below that limit or we
324  * In order to keep background CIL push efficient, we will set a lower
327  * enforced to ensure we stay within our maximum checkpoint size bounds.
458  * When we crack an atomic LSN, we sample it first so that the value will not
459  * change while we are cracking it into the component values. This means we
483  * When we crack the grant head, we sample it first so that the value will not
484  * change while we are cracking it into the component values. This means we
580 	 * a transiently forward state. Instead, we can see the LSN in a  in xlog_valid_lsn()
581 	 * transiently behind state if we happen to race with a cycle wrap.  in xlog_valid_lsn()