Document user-space visible device attributes that
are enabled by ZRAM_MULTI_COMP.
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
---
Documentation/admin-guide/blockdev/zram.rst | 55 +++++++++++++++++++++
drivers/block/zram/zram_drv.c | 7 +--
2 files changed, 57 insertions(+), 5 deletions(-)
diff --git a/Documentation/admin-guide/blockdev/zram.rst b/Documentation/admin-guide/blockdev/zram.rst
index c73b16930449..88957fcb6ad7 100644
--- a/Documentation/admin-guide/blockdev/zram.rst
+++ b/Documentation/admin-guide/blockdev/zram.rst
@@ -401,6 +401,61 @@ budget in next setting is user's job.
If admin wants to measure writeback count in a certain period, they could
know it via /sys/block/zram0/bd_stat's 3rd column.
+recompression
+-------------
+
+With CONFIG_ZRAM_MULTI_COMP, zram can recompress idle/huge pages using
+alternative (secondary) compression algorithm. The basic idea is that
+alternative compression algorithm can provide better compression ratio
+at a price of (potentially) slower compression/decompression speeds.
+Alternative compression algorithm can, for example, be more successful
+compressing huge pages (those that default algorithm failed to compress).
+Another application is idle pages recompression - pages that are cold and
+sit in the memory can be recompressed using more effective algorithm and,
+hence, reduce zsmalloc memory usage.
+
+With CONFIG_ZRAM_MULTI_COMP, zram will setup two compression algorithms
+per-CPU: primary and secondary ones. Primary zram compressor is explained
+in "3) Select compression algorithm", the secondary algorithm is configured
+in a similar way, using recomp_algorithm device attribute:
+
+Examples::
+
+ #show supported recompression algorithms
+ cat /sys/block/zramX/recomp_algorithm
+ zstd [lzo]
+
+ #select zstd recompression algorithm
+ echo zstd > /sys/block/zramX/recomp_algorithm
+
+Another device attribute that CONFIG_ZRAM_MULTI_COMP enables is recompress,
+which controls recompression:
+
+Examples::
+
+ #IDLE pages recompression is activated by `idle` mode
+ echo idle > /sys/block/zramX/recompress
+
+ #HUGE pages recompression is activated by `huge` mode
+ echo huge > /sys/block/zram0/recompress
+
+ #HUGE_IDLE pages recompression is activated by `huge_idle` mode
+ echo huge_idle > /sys/block/zramX/recompress
+
+The number of idle pages can be significant, so user-space can pass a size
+watermark value to the recompress knob, to filter out idle pages for
+recompression: zram will recompress only idle pages of equal or greater
+size:::
+
+ #recompress idle pages larger than 3000 bytes
+ echo 3000 > /sys/block/zramX/recompress
+
+ #recompress idle pages larger than 2000 bytes
+ echo 2000 > /sys/block/zramX/recompress
+
+Recompression is mostly focused on idle pages (except for huge pages
+recompression), so it works better in conjunction with memory tracking.
+
memory tracking
===============
diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index 12368e773f3d..143c33bd6783 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -1623,9 +1623,7 @@ static int zram_bvec_write(struct zram *zram, struct bio_vec *bvec,
*
* Corresponding ZRAM slot should be locked.
*/
-static int zram_recompress(struct zram *zram,
- u32 index,
- struct page *page,
+static int zram_recompress(struct zram *zram, u32 index, struct page *page,
int size_watermark)
{
unsigned long handle_prev;
@@ -1708,8 +1706,7 @@ static int zram_recompress(struct zram *zram,
static ssize_t recompress_store(struct device *dev,
struct device_attribute *attr,
- const char *buf,
- size_t len)
+ const char *buf, size_t len)
{
struct zram *zram = dev_to_zram(dev);
unsigned long nr_pages = zram->disksize >> PAGE_SHIFT;
--
2.38.0.rc1.362.ged0d419d3c-googDocument user-space visible device attributes that are enabled by ZRAM_MULTI_COMP. Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org> --- Documentation/admin-guide/blockdev/zram.rst | 55 +++++++++++++++++++++ 1 file changed, 55 insertions(+) diff --git a/Documentation/admin-guide/blockdev/zram.rst b/Documentation/admin-guide/blockdev/zram.rst index c73b16930449..f09e778bc3dc 100644 --- a/Documentation/admin-guide/blockdev/zram.rst +++ b/Documentation/admin-guide/blockdev/zram.rst @@ -401,6 +401,61 @@ budget in next setting is user's job. If admin wants to measure writeback count in a certain period, they could know it via /sys/block/zram0/bd_stat's 3rd column. +recompression +------------- + +With CONFIG_ZRAM_MULTI_COMP, zram can recompress idle/huge pages using +alternative (secondary) compression algorithm. The basic idea is that +alternative compression algorithm can provide better compression ratio +at a price of (potentially) slower compression/decompression speeds. +Alternative compression algorithm can, for example, be more successful +compressing huge pages (those that default algorithm failed to compress). +Another application is idle pages recompression - pages that are cold and +sit in the memory can be recompressed using more effective algorithm and, +hence, reduce zsmalloc memory usage. + +With CONFIG_ZRAM_MULTI_COMP, zram will setup two compression algorithms +per-CPU: primary and secondary ones. Primary zram compressor is explained +in "3) Select compression algorithm", the secondary algorithm is configured +in a similar way, using recomp_algorithm device attribute: + +Examples:: + + #show supported recompression algorithms + cat /sys/block/zramX/recomp_algorithm + zstd [lzo] + + #select zstd recompression algorithm + echo zstd > /sys/block/zramX/recomp_algorithm + +Another device attribute that CONFIG_ZRAM_MULTI_COMP enables is recompress, +which controls recompression: + +Examples:: + + #IDLE pages recompression is activated by `idle` mode + echo idle > /sys/block/zramX/recompress + + #HUGE pages recompression is activated by `huge` mode + echo huge > /sys/block/zram0/recompress + + #HUGE_IDLE pages recompression is activated by `huge_idle` mode + echo huge_idle > /sys/block/zramX/recompress + +The number of idle pages can be significant, so user-space can pass a size +watermark (in bytes) value to the recompress knob, to filter out idle pages +for recompression: zram will recompress only idle pages of equal or greater +size (bytes)::: + + #recompress idle pages larger than 3000 bytes + echo 3000 > /sys/block/zramX/recompress + + #recompress idle pages larger than 2000 bytes + echo 2000 > /sys/block/zramX/recompress + +Recompression is mostly focused on idle pages (except for huge pages +recompression), so it works better in conjunction with memory tracking. + memory tracking =============== -- 2.38.0.413.g74048e4d9e-goog
On (22/10/09 18:07), Sergey Senozhatsky wrote:
[..]
> diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> index 12368e773f3d..143c33bd6783 100644
> --- a/drivers/block/zram/zram_drv.c
> +++ b/drivers/block/zram/zram_drv.c
> @@ -1623,9 +1623,7 @@ static int zram_bvec_write(struct zram *zram, struct bio_vec *bvec,
> *
> * Corresponding ZRAM slot should be locked.
> */
> -static int zram_recompress(struct zram *zram,
> - u32 index,
> - struct page *page,
> +static int zram_recompress(struct zram *zram, u32 index, struct page *page,
> int size_watermark)
> {
> unsigned long handle_prev;
> @@ -1708,8 +1706,7 @@ static int zram_recompress(struct zram *zram,
>
> static ssize_t recompress_store(struct device *dev,
> struct device_attribute *attr,
> - const char *buf,
> - size_t len)
> + const char *buf, size_t len)
> {
> struct zram *zram = dev_to_zram(dev);
> unsigned long nr_pages = zram->disksize >> PAGE_SHIFT;
D'oh, this patch should not contain any code changes, only
documentation updates. It's the previous patch in the series
that should contain these coding-style tweaks. Rebasing did
not go spectacularly well here...
Andrew, what's the best way to handle this?
On Tue, 18 Oct 2022 11:22:25 +0900 Sergey Senozhatsky <senozhatsky@chromium.org> wrote: > D'oh, this patch should not contain any code changes, only > documentation updates. It's the previous patch in the series > that should contain these coding-style tweaks. Rebasing did > not go spectacularly well here... > > Andrew, what's the best way to handle this? May as well resend the whole patchset.
© 2016 - 2026 Red Hat, Inc.