On Wed, 2008-02-06 at 11:05 -0600, Eric Sandeen wrote:
> struct ext4_allocation_context is rather large, and this bloats
> the stack of many functions which use it. Allocating it from
> a named slab cache will alleviate this.
>
> For example, with this change (on top of the noinline patch sent earlier):
>
> -ext4_mb_new_blocks 200
> +ext4_mb_new_blocks 40
>
> -ext4_mb_free_blocks 344
> +ext4_mb_free_blocks 168
>
> -ext4_mb_release_inode_pa 216
> +ext4_mb_release_inode_pa 40
>
> -ext4_mb_release_group_pa 192
> +ext4_mb_release_group_pa 24
>
> Most of these stack-allocated structs are actually used only for
> mballoc history; and in those cases often a smaller struct would do.
> So changing that may be another way around it, at least for those
> functions, if preferred. For now, in those cases where the ac
> is only for history, an allocation failure simply skips the history
> recording, and does not cause any other failures.
>
> Comments?
>
> Signed-off-by: Eric Sandeen <sandeen@xxxxxxxxxx>
> ---
>
> Index: linux-2.6.24-rc6-mm1/fs/ext4/mballoc.c
> ===================================================================
> --- linux-2.6.24-rc6-mm1.orig/fs/ext4/mballoc.c
> +++ linux-2.6.24-rc6-mm1/fs/ext4/mballoc.c
> @@ -419,6 +419,7 @@
> #define MB_DEFAULT_GROUP_PREALLOC 512
>
> static struct kmem_cache *ext4_pspace_cachep;
> +static struct kmem_cache *ext4_ac_cachep;
>
> #ifdef EXT4_BB_MAX_BLOCKS
> #undef EXT4_BB_MAX_BLOCKS
> @@ -2958,11 +2959,18 @@ int __init init_ext4_mballoc(void)
> if (ext4_pspace_cachep == NULL)
> return -ENOMEM;
>
> -#ifdef CONFIG_PROC_FS
> + ext4_ac_cachep =
> + kmem_cache_create("ext4_alloc_context",
> + sizeof(struct ext4_allocation_context),
> + 0, SLAB_RECLAIM_ACCOUNT, NULL);
> + if (ext4_ac_cachep == NULL) {
> + kmem_cache_destroy(ext4_pspace_cachep);
> + return -ENOMEM;
> + }
> +
> proc_root_ext4 = proc_mkdir(EXT4_ROOT, proc_root_fs);
> if (proc_root_ext4 == NULL)
> printk(KERN_ERR "EXT4-fs: Unable to create %s\n", EXT4_ROOT);
> -#endif
>
> return 0;
> }
> @@ -2971,9 +2979,8 @@ void exit_ext4_mballoc(void)
> {
> /* XXX: synchronize_rcu(); */
> kmem_cache_destroy(ext4_pspace_cachep);
> -#ifdef CONFIG_PROC_FS
> + kmem_cache_destroy(ext4_ac_cachep);
> remove_proc_entry(EXT4_ROOT, proc_root_fs);
> -#endif
> }
>
>
Do you intend to remove the #ifdef CONFIG_PROC_FS, or it's a accident? I
think we need keep that to allow ext4 build without procfs configured.
Other than this, the patch looks fine to me.:)
Mingming
-
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
