Skip to contentSkip to conversation

xfs/558: scale blk IO size based on the filesystem blksz

PATCH In review Pankaj Raghav (Samsung) 2 participants 4 messages Mar 13, 2024 – Mar 27, 2024

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xfs/558: scale blk IO size based on the filesystem blksz

From: Pankaj Raghav <[email protected]>

This test fails for 64k filesystem block size on a 4k PAGE_SIZE system. Scale the blksz based on the filesystem block size instead of fixing it as 64k so that we do get some iomap invalidations while doing concurrent writes.

Cap the blksz to be at least 64k to retain the same behaviour as before for smaller filesystem blocksizes.

This fixes the "Expected to hear about writeback iomap invalidations?" message for 64k filesystems.

---
 tests/xfs/558 | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
tests/xfs/558 +6 -1
index 9e9b3be8..270f458c 100755
@@ -127,7 +127,12 @@ _scratch_mount >> $seqres.full
 $XFS_IO_PROG -c 'chattr -x' $SCRATCH_MNT &> $seqres.full
 _require_pagecache_access $SCRATCH_MNT
 
-blksz=65536
+min_blksz=65536
+file_blksz=$(_get_file_block_size "$SCRATCH_MNT")
+blksz=$(( 8 * $file_blksz ))
+
+blksz=$(( blksz > min_blksz ? blksz : min_blksz ))
+
 _require_congruent_file_oplen $SCRATCH_MNT $blksz
 
 # Make sure we have sufficient extent size to create speculative CoW
Signature · 1 line
2.43.0
#1 lore raw
Quoted 10 lines · Pankaj Raghav (Samsung)

From: Pankaj Raghav <[email protected]>

This test fails for 64k filesystem block size on a 4k PAGE_SIZE system. Scale the blksz based on the filesystem block size instead of fixing it as 64k so that we do get some iomap invalidations while doing concurrent writes.

Cap the blksz to be at least 64k to retain the same behaviour as before for smaller filesystem blocksizes.

Quoted 7 lines

This fixes the "Expected to hear about writeback iomap invalidations?" message for 64k filesystems.

---

Hi Zorro,
I checked xfstests/for-next and didn't see this patch queued up. Do I need
to do something or will it get queued up for the next release?

Quoted 21 lines
 tests/xfs/558 | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
tests/xfs/558 +6 -1
index 9e9b3be8..270f458c 100755
@@ -127,7 +127,12 @@ _scratch_mount >> $seqres.full
 $XFS_IO_PROG -c 'chattr -x' $SCRATCH_MNT &> $seqres.full
 _require_pagecache_access $SCRATCH_MNT
 
-blksz=65536
+min_blksz=65536
+file_blksz=$(_get_file_block_size "$SCRATCH_MNT")
+blksz=$(( 8 * $file_blksz ))
+
+blksz=$(( blksz > min_blksz ? blksz : min_blksz ))
+
 _require_congruent_file_oplen $SCRATCH_MNT $blksz
 
 # Make sure we have sufficient extent size to create speculative CoW
#2 lore raw
Quoted 14 lines · Pankaj Raghav
Quoted 13 lines · Pankaj Raghav (Samsung)

From: Pankaj Raghav <[email protected]>

This test fails for 64k filesystem block size on a 4k PAGE_SIZE system. Scale the blksz based on the filesystem block size instead of fixing it as 64k so that we do get some iomap invalidations while doing concurrent writes.

Cap the blksz to be at least 64k to retain the same behaviour as before for smaller filesystem blocksizes.

This fixes the "Expected to hear about writeback iomap invalidations?" message for 64k filesystems.

Quoted 8 lines
Quoted 4 lines

---

Hi Zorro,
I checked xfstests/for-next and didn't see this patch queued up. Do I need
to do something or will it get queued up for the next release?

Oh sorry, due to this patch was in a patchset with:

[PATCH 2/2] xfs/161: adapt the test case for LBS filesystem

So I don't know if you'd like to have them both together, or merge this one (with RVB) at first:)

If you hope to have this one at first, I'll merge it in patches-in-queue branch, then have it in next branch.

Thanks,
Zorro

Quoted 23 lines
Quoted 22 lines
 tests/xfs/558 | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
tests/xfs/558 +6 -1
index 9e9b3be8..270f458c 100755
@@ -127,7 +127,12 @@ _scratch_mount >> $seqres.full
 $XFS_IO_PROG -c 'chattr -x' $SCRATCH_MNT &> $seqres.full
 _require_pagecache_access $SCRATCH_MNT
 
-blksz=65536
+min_blksz=65536
+file_blksz=$(_get_file_block_size "$SCRATCH_MNT")
+blksz=$(( 8 * $file_blksz ))
+
+blksz=$(( blksz > min_blksz ? blksz : min_blksz ))
+
 _require_congruent_file_oplen $SCRATCH_MNT $blksz
 
 # Make sure we have sufficient extent size to create speculative CoW
#3 lore raw
Quoted 9 lines
Quoted 8 lines
Quoted 4 lines

---

Hi Zorro,
I checked xfstests/for-next and didn't see this patch queued up. Do I need
to do something or will it get queued up for the next release?

Quoted 7 lines

Oh sorry, due to this patch was in a patchset with:

[PATCH 2/2] xfs/161: adapt the test case for LBS filesystem

So I don't know if you'd like to have them both together, or merge this one (with RVB) at first:)

The other patch needs some more discussion I think. Please feel free to merge this one in as it already has RVB :)

Quoted 30 lines

If you hope to have this one at first, I'll merge it in patches-in-queue branch, then have it in next branch.

Thanks,
Zorro

Quoted 23 lines
Quoted 22 lines
 tests/xfs/558 | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
tests/xfs/558 +6 -1
index 9e9b3be8..270f458c 100755
@@ -127,7 +127,12 @@ _scratch_mount >> $seqres.full
 $XFS_IO_PROG -c 'chattr -x' $SCRATCH_MNT &> $seqres.full
 _require_pagecache_access $SCRATCH_MNT
 
-blksz=65536
+min_blksz=65536
+file_blksz=$(_get_file_block_size "$SCRATCH_MNT")
+blksz=$(( 8 * $file_blksz ))
+
+blksz=$(( blksz > min_blksz ? blksz : min_blksz ))
+
 _require_congruent_file_oplen $SCRATCH_MNT $blksz
 
 # Make sure we have sufficient extent size to create speculative CoW