22  *   1) space_info.  This is the ultimate arbiter of how much space we can use.
25  *   reservations we care about total_bytes - SUM(space_info->bytes_) when
30  *   metadata reservation we have.  You can see the comment in the block_rsv
34  *   3) btrfs_calc*_size.  These are the worst case calculations we used based
35  *   on the number of items we will want to modify.  We have one for changing
36  *   items, and one for inserting new items.  Generally we use these helpers to
42  *   We call into either btrfs_reserve_data_bytes() or
43  *   btrfs_reserve_metadata_bytes(), depending on which we're looking for, with
44  *   num_bytes we want to reserve.
61  *   Assume we are unable to simply make the reservation because we do not have
84  *     Check if ->bytes == 0, if it does we got our reservation and we can carry
85  *     on, if not return the appropriate error (ENOSPC, but can be EINTR if we
90  *   Same as the above, except we add ourselves to the
91  *   space_info->priority_tickets, and we do not use ticket->wait, we simply
97  *   Generally speaking we will have two cases for each state, a "nice" state
98  *   and a "ALL THE THINGS" state.  In btrfs we delay a lot of work in order to
102  *   reclaim space so we can make new reservations.
106  *     for example, we would update the inode item at write time to update the
108  *     isize or bytes.  We keep these delayed items to coalesce these operations
114  *     for delayed allocation.  We can reclaim some of this space simply by
115  *     running delalloc, but usually we need to wait for ordered extents to
119  *     We have a block reserve for the outstanding delayed refs space, and every
121  *     to reclaim space, but we want to hold this until the end because COW can
122  *     churn a lot and we can avoid making some extent tree modifications if we
126  *     We will skip this the first time through space reservation, because of
127  *     overcommit and we don't want to have a lot of useless metadata space when
131  *     If we're freeing inodes we're likely freeing checksums, file extent
136  *     may_commit_transaction() is the ultimate arbiter on whether we commit the
137  *     transaction or not.  In order to avoid constantly churning we do all the
139  *     However we need to take into account things like pinned space that would
140  *     be freed, plus any delayed work we may not have gotten rid of in the case
145  *   Because we hold so many reservations for metadata we will allow you to
150  *   You can see the current logic for when we allow overcommit in
170  * after adding space to the filesystem, we need to clear the full flags
314 	 * If we have dup, raid1 or raid10 then only half of the free  in calc_available_free_space()
316 	 * doesn't include the parity drive, so we don't have to  in calc_available_free_space()
323 	 * If we aren't flushing all things, let us overcommit up to  in calc_available_free_space()
324 	 * 1/2th of the space. If we can flush, don't let us overcommit  in calc_available_free_space()
364  * This is for space we already have accounted in space_info->bytes_may_use, so
365  * basically when we're returning space from block_rsv's.
498 	/* Calc the number of the pages we need flush for space reservation */  in shrink_delalloc()
503 		 * to_reclaim is set to however much metadata we need to  in shrink_delalloc()
506 		 * make sure we're flushing enough delalloc to hopefully reclaim  in shrink_delalloc()
527 	 * If we are doing more ordered than delalloc we need to just wait on  in shrink_delalloc()
528 	 * ordered extents, otherwise we'll waste time trying to flush delalloc  in shrink_delalloc()
529 	 * that likely won't give us the space back we need.  in shrink_delalloc()
565  * @root - the root we're allocating for
566  * @bytes - the number of bytes we want to reserve
620 	 * we have block groups that are going to be freed, allowing us to  in may_commit_transaction()
732 		 * If we have pending delayed iputs then we could free up a  in flush_space()
733 		 * bunch of pinned space, so make sure we run the iputs before  in flush_space()
734 		 * we do our pinned bytes check below.  in flush_space()
768 	 * We may be flushing because suddenly we have less space than we had  in btrfs_calc_reclaim_metadata_size()
769 	 * before, and now we're well over-committed based on our current free  in btrfs_calc_reclaim_metadata_size()
770 	 * space.  If that's the case add in our overage so we make sure to put  in btrfs_calc_reclaim_metadata_size()
807 	/* If we're just plain full then async reclaim just slows us down. */  in need_do_async_reclaim()
847  * maybe_fail_all_tickets - we've exhausted our flushing, start failing tickets
849  * @space_info - the space info we were flushing
851  * We call this when we've exhausted our flushing ability and haven't made
853  * order, so if there is a large ticket first and then smaller ones we could
885 		 * would result in the ticket succeeding.  However if we have a  in maybe_fail_all_tickets()
906 		 * We're just throwing tickets away, so more flushing may not  in maybe_fail_all_tickets()
907 		 * trip over btrfs_try_granting_tickets, so we need to call it  in maybe_fail_all_tickets()
908 		 * here to see if we can make progress with the next ticket in  in maybe_fail_all_tickets()
917  * This is for normal flushers, we can wait all goddamned day if we want to.  We
918  * will loop and continuously try to flush as long as we are making progress.
919  * We count progress as clearing off tickets each time we have to loop.
964 		 * We don't want to force a chunk allocation until we've tried  in btrfs_async_reclaim_metadata_space()
965 		 * pretty hard to reclaim space.  Think of the case where we  in btrfs_async_reclaim_metadata_space()
967 		 * to reclaim.  We would rather use that than possibly create a  in btrfs_async_reclaim_metadata_space()
971 		 * around then we can force a chunk allocation.  in btrfs_async_reclaim_metadata_space()
997  *   1) compression is on and we allocate less space than we reserved
998  *   2) we are overwriting existing space
1005  *   For #2 this is trickier.  Once the ordered extent runs we will drop the
1006  *   extent in the range we are overwriting, which creates a delayed ref for
1011  *   If we are freeing inodes, we want to make sure all delayed iputs have
1020  *   reality if there are outstanding delayed refs.  This is because we adjust
1022  *   any on-disk state which might include more refs.  So for example, if we
1023  *   have an extent with 2 references, but we only drop 1, we'll see that there
1033  *   This is where we reclaim all of the pinned space generated by the previous
1034  *   two stages.  We will not commit the transaction if we don't think we're
1036  *   total_bytes_pinned < reservation we will not commit.  This is why the
1037  *   previous states are actually important, to make sure we know for sure
1041  *   For data we start with alloc chunk force, however we could have been full
1043  *   so if we now have space to allocate do the force chunk allocation.
1196 			 * Delete us from the list. After we unlock the space  in wait_reserve_ticket()
1197 			 * info, we don't want the async reclaim job to reserve  in wait_reserve_ticket()
1222  * @flush - how much we can flush
1262 		 * We were a priority ticket, so we need to delete ourselves  in handle_reserve_ticket()
1263 		 * from the list.  Because we could have other priority tickets  in handle_reserve_ticket()
1279 	 * Check that we can't have an error set if the reservation succeeded,  in handle_reserve_ticket()
1300  * @root - the root we're allocating for
1301  * @space_info - the space info we want to allocate from
1302  * @orig_bytes - the number of bytes we want
1303  * @flush - whether or not we can flush to make our reservation
1335 	 * We don't want NO_FLUSH allocations to jump everybody, they can  in __reserve_bytes()
1346 	 * Carry on if we have enough space (short-circuit) OR call  in __reserve_bytes()
1347 	 * can_overcommit() to ensure we can overcommit to continue.  in __reserve_bytes()
1358 	 * If we couldn't make a reservation then setup our reservation ticket  in __reserve_bytes()
1361 	 * If we are a priority flusher then we just need to add our ticket to  in __reserve_bytes()
1362 	 * the list and we will do our own flushing further down.  in __reserve_bytes()
1389 		 * We will do the space reservation dance during log replay,  in __reserve_bytes()
1390 		 * which means we won't have fs_info->fs_root set, so don't do  in __reserve_bytes()
1391 		 * the async reclaim as we will panic.  in __reserve_bytes()
1411  * @root - the root we're allocating for
1412  * @block_rsv - the block_rsv we're allocating for
1413  * @orig_bytes - the number of bytes we want
1414  * @flush - whether or not we can flush to make our reservation
1454  * @bytes - the number of bytes we need
1455  * @flush - how we are allowed to flush
1458  * space then we will attempt to flush space as specified by flush.