Re: [PATCH] Btrfs: use a slab for btrfs_dio_private

[Andi Kleen]

Josef Bacik  writes:

> This is in the IO path, let's try to avoid latencies by having our own slab.
> Thanks,

What latencies does this avoid?

-Andi

-- 
a...@linux.intel.com -- Speaking for myself only
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH] Btrfs: use a slab for btrfs_dio_private

[Josef Bacik]

This is in the IO path, let's try to avoid latencies by having our own slab.
Thanks,

Signed-off-by: Josef Bacik 
---
 fs/btrfs/inode.c |   15 ---
 1 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 8a8f0d6..351dd3b 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -75,6 +75,7 @@ struct kmem_cache *btrfs_trans_handle_cachep;
 struct kmem_cache *btrfs_transaction_cachep;
 struct kmem_cache *btrfs_path_cachep;
 struct kmem_cache *btrfs_free_space_cachep;
+struct kmem_cache *btrfs_dip_cachep;
 
 #define S_SHIFT 12
 static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = {
@@ -6076,7 +6077,7 @@ failed:
  dip->logical_offset + dip->bytes - 1);
bio->bi_private = dip->private;
 
-   kfree(dip);
+   kmem_cache_free(btrfs_dip_cachep, dip);
 
/* If we had a csum failure make sure to clear the uptodate flag */
if (err)
@@ -6121,7 +6122,7 @@ out_test:
 out_done:
bio->bi_private = dip->private;
 
-   kfree(dip);
+   kmem_cache_free(btrfs_dip_cachep, dip);
 
/* If we had an error make sure to clear the uptodate flag */
if (err)
@@ -6344,7 +6345,7 @@ static void btrfs_submit_direct(int rw, struct bio *bio, 
struct inode *inode,
 
skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
-   dip = kmalloc(sizeof(*dip), GFP_NOFS);
+   dip = kmem_cache_alloc(btrfs_dip_cachep, GFP_NOFS);
if (!dip) {
ret = -ENOMEM;
goto free_ordered;
@@ -7062,6 +7063,8 @@ void btrfs_destroy_cachep(void)
kmem_cache_destroy(btrfs_path_cachep);
if (btrfs_free_space_cachep)
kmem_cache_destroy(btrfs_free_space_cachep);
+   if (btrfs_dip_cachep)
+   kmem_cache_destroy(btrfs_dip_cachep);
 }
 
 int btrfs_init_cachep(void)
@@ -7096,6 +7099,12 @@ int btrfs_init_cachep(void)
if (!btrfs_free_space_cachep)
goto fail;
 
+   btrfs_dip_cachep = kmem_cache_create("btrfs_dio_private_cache",
+   sizeof(struct btrfs_dio_private), 0,
+   SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
+   if (!btrfs_dip_cachep)
+   goto fail;
+
return 0;
 fail:
btrfs_destroy_cachep();
-- 
1.7.7.6

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html