kernel.git/fs/xfs/xfs_alloc.h, branch linux-3.4.y

xfs: fix fstrim offset calculations

2012-03-27T21:07:03Z

xfs_ioc_fstrim() doesn't treat the incoming offset and length correctly. It treats them as a filesystem block address, rather than a disk address. This is wrong because the range passed in is a linear representation, while the filesystem block address notation is a sparse representation. Hence we cannot convert the range direct to filesystem block units and then use that for calculating the range to trim. While this sounds dangerous, the problem is limited to calculating what AGs need to be trimmed. The code that calcuates the actual ranges to trim gets the right result (i.e. only ever discards free space), even though it uses the wrong ranges to limit what is trimmed. Hence this is not a bug that endangers user data. Fix this by treating the range as a disk address range and use the appropriate functions to convert the range into the desired formats for calculations. Further, fix the first free extent lookup (the longest) to actually find the largest free extent. Currently this lookup uses a <= lookup, which results in finding the extent to the left of the largest because we can never get an exact match on the largest extent. This is due to the fact that while we know it's size, we don't know it's location and so the exact match fails and we move one record to the left to get the next largest extent. Instead, use a >= search so that the lookup returns the largest extent regardless of the fact we don't get an exact match on it. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers

xfs: introduce an allocation workqueue

2012-03-22T21:12:24Z

We currently have significant issues with the amount of stack that allocation in XFS uses, especially in the writeback path. We can easily consume 4k of stack between mapping the page, manipulating the bmap btree and allocating blocks from the free list. Not to mention btree block readahead and other functionality that issues IO in the allocation path. As a result, we can no longer fit allocation in the writeback path in the stack space provided on x86_64. To alleviate this problem, introduce an allocation workqueue and move all allocations to a seperate context. This can be easily added as an interposing layer into xfs_alloc_vextent(), which takes a single argument structure and does not return until the allocation is complete or has failed. To do this, add a work structure and a completion to the allocation args structure. This allows xfs_alloc_vextent to queue the args onto the workqueue and wait for it to be completed by the worker. This can be done completely transparently to the caller. The worker function needs to ensure that it sets and clears the PF_TRANS flag appropriately as it is being run in an active transaction context. Work can also be queued in a memory reclaim context, so a rescuer is needed for the workqueue. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers

xfs: do not discard alloc btree blocks

2011-05-24T16:17:22Z

Blocks for the allocation btree are allocated from and released to the AGFL, and thus frequently reused. Even worse we do not have an easy way to avoid using an AGFL block when it is discarded due to the simple FILO list of free blocks, and thus can frequently stall on blocks that are currently undergoing a discard. Add a flag to the busy extent tracking structure to skip the discard for allocation btree blocks. In normal operation these blocks are reused frequently enough that there is no need to discard them anyway, but if they spill over to the allocation btree as part of a balance we "leak" blocks that we would otherwise discard. We could fix this by adding another flag and keeping these block in the rbtree even after they aren't busy any more so that we could discard them when they migrate out of the AGFL. Given that this would cause significant overhead I don't think it's worthwile for now. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder

xfs: add online discard support

2011-05-24T16:17:13Z

Now that we have reliably tracking of deleted extents in a transaction we can easily implement "online" discard support which calls blkdev_issue_discard once a transaction commits. The actual discard is a two stage operation as we first have to mark the busy extent as not available for reuse before we can start the actual discard. Note that we don't bother supporting discard for the non-delaylog mode. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder

xfs: reduce the number of pagb_lock roundtrips in xfs_alloc_clear_busy

2011-04-28T18:18:09Z

Instead of finding the per-ag and then taking and releasing the pagb_lock for every single busy extent completed sort the list of busy extents and only switch betweens AGs where nessecary. This becomes especially important with the online discard support which will hit this lock more often. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder

xfs: exact busy extent tracking

2011-04-28T18:18:04Z

Update the extent tree in case we have to reuse a busy extent, so that it always is kept uptodate. This is done by replacing the busy list searches with a new xfs_alloc_busy_reuse helper, which updates the busy extent tree in case of a reuse. This allows us to allow reusing metadata extents unconditionally, and thus avoid log forces especially for allocation btree blocks. Signed-off-by: Christoph Hellwig Signed-off-by: Alex Elder

xfs: limit extent length for allocation to AG size

2011-01-28T15:05:35Z

Delayed allocation extents can be larger than AGs, so when trying to convert a large range we may scan every AG inside xfs_bmap_alloc_nullfb() trying to find an AG with a size larger than an AG. We should stop when we find the first AG with a maximum possible allocation size. This causes excessive CPU usage when there are lots of AGs. The same problem occurs when doing preallocation of a range larger than an AG. Fix the problem by limiting real allocation lengths to the maximum that an AG can support. This means if we have empty AGs, we'll stop the search at the first of them. If there are no empty AGs, we'll still scan them all, but that is a different problem.... Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Reviewed-by: Alex Elder

xfs: add FITRIM support

2011-01-12T02:28:29Z

Allow manual discards from userspace using the FITRIM ioctl. This is not intended to be run during normal workloads, as the freepsace btree walks can cause large performance degradation. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Alex Elder

xfs: do not use emums for flags used in tracing

2010-07-26T18:16:43Z

The tracing code can't print flags defined as enums. Most flags that we want to print are defines as macros already, but move the few remaining ones over to make the trace output more useful. Signed-off-by: Christoph Hellwig Reviewed-by: Dave Chinner

xfs: Improve scalability of busy extent tracking

2010-05-24T15:34:00Z

When we free a metadata extent, we record it in the per-AG busy extent array so that it is not re-used before the freeing transaction hits the disk. This array is fixed size, so when it overflows we make further allocation transactions synchronous because we cannot track more freed extents until those transactions hit the disk and are completed. Under heavy mixed allocation and freeing workloads with large log buffers, we can overflow this array quite easily. Further, the array is sparsely populated, which means that inserts need to search for a free slot, and array searches often have to search many more slots that are actually used to check all the busy extents. Quite inefficient, really. To enable this aspect of extent freeing to scale better, we need a structure that can grow dynamically. While in other areas of XFS we have used radix trees, the extents being freed are at random locations on disk so are better suited to being indexed by an rbtree. So, use a per-AG rbtree indexed by block number to track busy extents. This incures a memory allocation when marking an extent busy, but should not occur too often in low memory situations. This should scale to an arbitrary number of extents so should not be a limitation for features such as in-memory aggregation of transactions. However, there are still situations where we can't avoid allocating busy extents (such as allocation from the AGFL). To minimise the overhead of such occurences, we need to avoid doing a synchronous log force while holding the AGF locked to ensure that the previous transactions are safely on disk before we use the extent. We can do this by marking the transaction doing the allocation as synchronous rather issuing a log force. Because of the locking involved and the ordering of transactions, the synchronous transaction provides the same guarantees as a synchronous log force because it ensures that all the prior transactions are already on disk when the synchronous transaction hits the disk. i.e. it preserves the free->allocate order of the extent correctly in recovery. By doing this, we avoid holding the AGF locked while log writes are in progress, hence reducing the length of time the lock is held and therefore we increase the rate at which we can allocate and free from the allocation group, thereby increasing overall throughput. The only problem with this approach is that when a metadata buffer is marked stale (e.g. a directory block is removed), then buffer remains pinned and locked until the log goes to disk. The issue here is that if that stale buffer is reallocated in a subsequent transaction, the attempt to lock that buffer in the transaction will hang waiting the log to go to disk to unlock and unpin the buffer. Hence if someone tries to lock a pinned, stale, locked buffer we need to push on the log to get it unlocked ASAP. Effectively we are trading off a guaranteed log force for a much less common trigger for log force to occur. Ideally we should not reallocate busy extents. That is a much more complex fix to the problem as it involves direct intervention in the allocation btree searches in many places. This is left to a future set of modifications. Finally, now that we track busy extents in allocated memory, we don't need the descriptors in the transaction structure to point to them. We can replace the complex busy chunk infrastructure with a simple linked list of busy extents. This allows us to remove a large chunk of code, making the overall change a net reduction in code size. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Alex Elder