On Tue, Feb 23, 2021 at 03:31:19PM +0800, Huang Jianan via Linux-erofs wrote:
> lz4 uses LZ4_DISTANCE_MAX to record history preservation. When
> using rolling decompression, a block with a higher compression
> ratio will cause a larger memory allocation (up to 64k). It may
> cause a large resource burden in extreme cases on devices with
> small memory and a large number of concurrent IOs. So appropriately
> reducing this value can improve performance.
>
> Decreasing this value will reduce the compression ratio (except
> when input_size <LZ4_DISTANCE_MAX). But considering that erofs
> currently only supports 4k output, reducing this value will not
> significantly reduce the compression benefits.
>
> The maximum value of LZ4_DISTANCE_MAX defined by lz4 is 64k, and
> we can only reduce this value. For the old kernel, it just can't
> reduce the memory allocation during rolling decompression without
> affecting the decompression result.
>
> Signed-off-by: Huang Jianan <[email protected]>
> Signed-off-by: Guo Weichao <[email protected]>
> ---
>
> change since v2:
> - use z_erofs_load_lz4_config to calculate lz4_distance_pages
> - add description about the compatibility of the old kernel version
> - drop useless comment
>
> fs/erofs/decompressor.c | 22 ++++++++++++++++++----
> fs/erofs/erofs_fs.h | 3 ++-
> fs/erofs/internal.h | 7 +++++++
> fs/erofs/super.c | 2 ++
> 4 files changed, 29 insertions(+), 5 deletions(-)
>
> diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c
> index 1cb1ffd10569..0bb7903e3f9b 100644
> --- a/fs/erofs/decompressor.c
> +++ b/fs/erofs/decompressor.c
> @@ -28,6 +28,18 @@ struct z_erofs_decompressor {
> char *name;
> };
>
> +int z_erofs_load_lz4_config(struct erofs_sb_info *sbi,
> + struct erofs_super_block *dsb)
> +{
> + u16 distance = le16_to_cpu(dsb->lz4_max_distance);
> +
> + sbi->lz4_max_distance_pages = distance ?
> + (DIV_ROUND_UP(distance, PAGE_SIZE) + 1) :
Unneeded parentheses here (I'll update it when applying).
Otherwise it looks good to me.
Reviewed-by: Gao Xiang <[email protected]>
Thanks,
Gao Xiang
On Tue, Feb 23, 2021 at 03:44:18PM +0800, Gao Xiang wrote:
> On Tue, Feb 23, 2021 at 03:31:19PM +0800, Huang Jianan via Linux-erofs wrote:
> > lz4 uses LZ4_DISTANCE_MAX to record history preservation. When
> > using rolling decompression, a block with a higher compression
> > ratio will cause a larger memory allocation (up to 64k). It may
> > cause a large resource burden in extreme cases on devices with
> > small memory and a large number of concurrent IOs. So appropriately
> > reducing this value can improve performance.
> >
> > Decreasing this value will reduce the compression ratio (except
> > when input_size <LZ4_DISTANCE_MAX). But considering that erofs
> > currently only supports 4k output, reducing this value will not
> > significantly reduce the compression benefits.
> >
> > The maximum value of LZ4_DISTANCE_MAX defined by lz4 is 64k, and
> > we can only reduce this value. For the old kernel, it just can't
> > reduce the memory allocation during rolling decompression without
> > affecting the decompression result.
> >
> > Signed-off-by: Huang Jianan <[email protected]>
> > Signed-off-by: Guo Weichao <[email protected]>
> > ---
> >
> > change since v2:
> > - use z_erofs_load_lz4_config to calculate lz4_distance_pages
> > - add description about the compatibility of the old kernel version
> > - drop useless comment
> >
> > fs/erofs/decompressor.c | 22 ++++++++++++++++++----
> > fs/erofs/erofs_fs.h | 3 ++-
> > fs/erofs/internal.h | 7 +++++++
> > fs/erofs/super.c | 2 ++
> > 4 files changed, 29 insertions(+), 5 deletions(-)
> >
> > diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c
> > index 1cb1ffd10569..0bb7903e3f9b 100644
> > --- a/fs/erofs/decompressor.c
> > +++ b/fs/erofs/decompressor.c
> > @@ -28,6 +28,18 @@ struct z_erofs_decompressor {
> > char *name;
> > };
> >
> > +int z_erofs_load_lz4_config(struct erofs_sb_info *sbi,
> > + struct erofs_super_block *dsb)
> > +{
> > + u16 distance = le16_to_cpu(dsb->lz4_max_distance);
> > +
> > + sbi->lz4_max_distance_pages = distance ?
> > + (DIV_ROUND_UP(distance, PAGE_SIZE) + 1) :
>
> Unneeded parentheses here (I'll update it when applying).
> Otherwise it looks good to me.
>
> Reviewed-by: Gao Xiang <[email protected]>
>
> Thanks,
> Gao Xiang
Applied to dev-test temporarily with the following diff
(fix 80 char limitation, etc):
diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c
index 0bb7903e3f9b..49347e681a53 100644
--- a/fs/erofs/decompressor.c
+++ b/fs/erofs/decompressor.c
@@ -34,9 +34,8 @@ int z_erofs_load_lz4_config(struct erofs_sb_info *sbi,
u16 distance = le16_to_cpu(dsb->lz4_max_distance);
sbi->lz4_max_distance_pages = distance ?
- (DIV_ROUND_UP(distance, PAGE_SIZE) + 1) :
+ DIV_ROUND_UP(distance, PAGE_SIZE) + 1 :
LZ4_MAX_DISTANCE_PAGES;
-
return 0;
}
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 4cb2395db45c..77965490dced 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -433,7 +433,7 @@ static inline void erofs_exit_shrinker(void) {}
static inline int z_erofs_init_zip_subsystem(void) { return 0; }
static inline void z_erofs_exit_zip_subsystem(void) {}
static inline int z_erofs_load_lz4_config(struct erofs_sb_info *sbi,
- struct erofs_super_block *dsb) { return 0; }
+ struct erofs_super_block *dsb) { return 0; }
#endif /* !CONFIG_EROFS_FS_ZIP */
#define EFSCORRUPTED EUCLEAN /* Filesystem is corrupted */
diff --git a/fs/erofs/super.c b/fs/erofs/super.c
index 11bc51e488dd..025c25d6e0f3 100644
--- a/fs/erofs/super.c
+++ b/fs/erofs/super.c
@@ -166,8 +166,6 @@ static int erofs_read_superblock(struct super_block *sb)
if (!check_layout_compatibility(sb, dsb))
goto out;
- z_erofs_load_lz4_config(sbi, dsb);
-
sbi->blocks = le32_to_cpu(dsb->blocks);
sbi->meta_blkaddr = le32_to_cpu(dsb->meta_blkaddr);
#ifdef CONFIG_EROFS_FS_XATTR
@@ -189,6 +187,9 @@ static int erofs_read_superblock(struct super_block *sb)
ret = -EFSCORRUPTED;
goto out;
}
+
+ /* parse on-disk compression configurations */
+ z_erofs_load_lz4_config(sbi, dsb);
ret = 0;
out:
kunmap(page);