xfs: limit speculative prealloc near ENOSPC thresholds

There is a window on small filesytsems where specualtive preallocation can be larger than that ENOSPC throttling thresholds, resulting in specualtive preallocation trying to reserve more space than there is space available. This causes immediate ENOSPC to be triggered, prealloc to be turned off and flushing to occur. One the next write (i.e. next 4k page), we do exactly the same thing, and so effective drive into synchronous 4k writes by triggering ENOSPC flushing on every page while in the window between the prealloc size and the ENOSPC prealloc throttle threshold. Fix this by checking to see if the prealloc size would consume all free space, and throttle it appropriately to avoid premature ENOSPC... Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Ben Myers <bpm@sgi.com>

xfs: limit speculative prealloc near ENOSPC thresholds
There is a window on small filesytsems where specualtive preallocation can be larger than that ENOSPC throttling thresholds, resulting in specualtive preallocation trying to reserve more space than there is space available. This causes immediate ENOSPC to be triggered, prealloc to be turned off and flushing to occur. One the next write (i.e. next 4k page), we do exactly the same thing, and so effective drive into synchronous 4k writes by triggering ENOSPC flushing on every page while in the window between the prealloc size and the ENOSPC prealloc throttle threshold. Fix this by checking to see if the prealloc size would consume all free space, and throttle it appropriately to avoid premature ENOSPC... Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Ben Myers <bpm@sgi.com>
Dave Chinner · Ben Myers
1 parent eb178619f9
Showing 1 changed file with 9 additions and 0 deletions Side-by-side Diff
fs/xfs/xfs_iomap.c
@@ -351,6 +351,15 @@
 		}
 		if (shift)
 			alloc_blocks >>= shift;
+
+		/*
+		 * If we are still trying to allocate more space than is
+		 * available, squash the prealloc hard. This can happen if we
+		 * have a large file on a small filesystem and the above
+		 * lowspace thresholds are smaller than MAXEXTLEN.
+		 */
+		while (alloc_blocks >= freesp)
+			alloc_blocks >>= 4;
 	}
  
 	if (alloc_blocks < mp->m_writeio_blocks)
...	...	@@ -351,6 +351,15 @@
351	351	}
352	352	if (shift)
353	353	alloc_blocks >>= shift;
	354	+
	355	+ /*
	356	+ * If we are still trying to allocate more space than is
	357	+ * available, squash the prealloc hard. This can happen if we
	358	+ * have a large file on a small filesystem and the above
	359	+ * lowspace thresholds are smaller than MAXEXTLEN.
	360	+ */
	361	+ while (alloc_blocks >= freesp)
	362	+ alloc_blocks >>= 4;
354	363	}
355	364
356	365	if (alloc_blocks < mp->m_writeio_blocks)