* [PATCH V2 1/2] zpool: Add malloc_support_movable to zpool_driver @ 2019-06-02 9:46 Hui Zhu 2019-06-02 9:46 ` [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support Hui Zhu 0 siblings, 1 reply; 6+ messages in thread From: Hui Zhu @ 2019-06-02 9:46 UTC (permalink / raw) To: ddstreet, minchan, ngupta, sergey.senozhatsky.work, sjenning, linux-mm, linux-kernel Cc: Hui Zhu As a zpool_driver, zsmalloc can allocate movable memory because it support migate pages. But zbud and z3fold cannot allocate movable memory. This commit adds malloc_support_movable to zpool_driver. If a zpool_driver support allocate movable memory, set it to true. And add zpool_malloc_support_movable check malloc_support_movable to make sure if a zpool support allocate movable memory. Signed-off-by: Hui Zhu <teawaterz@linux.alibaba.com> --- include/linux/zpool.h | 3 +++ mm/zpool.c | 16 ++++++++++++++++ mm/zsmalloc.c | 19 ++++++++++--------- 3 files changed, 29 insertions(+), 9 deletions(-) diff --git a/include/linux/zpool.h b/include/linux/zpool.h index 7238865e75b0..51bf43076165 100644 --- a/include/linux/zpool.h +++ b/include/linux/zpool.h @@ -46,6 +46,8 @@ const char *zpool_get_type(struct zpool *pool); void zpool_destroy_pool(struct zpool *pool); +bool zpool_malloc_support_movable(struct zpool *pool); + int zpool_malloc(struct zpool *pool, size_t size, gfp_t gfp, unsigned long *handle); @@ -90,6 +92,7 @@ struct zpool_driver { struct zpool *zpool); void (*destroy)(void *pool); + bool malloc_support_movable; int (*malloc)(void *pool, size_t size, gfp_t gfp, unsigned long *handle); void (*free)(void *pool, unsigned long handle); diff --git a/mm/zpool.c b/mm/zpool.c index a2dd9107857d..863669212070 100644 --- a/mm/zpool.c +++ b/mm/zpool.c @@ -238,6 +238,22 @@ const char *zpool_get_type(struct zpool *zpool) return zpool->driver->type; } +/** + * zpool_malloc_support_movable() - Check if the zpool support + * allocate movable memory + * @zpool: The zpool to check + * + * This returns if the zpool support allocate movable memory. + * + * Implementations must guarantee this to be thread-safe. + * + * Returns: true if if the zpool support allocate movable memory, false if not + */ +bool zpool_malloc_support_movable(struct zpool *zpool) +{ + return zpool->driver->malloc_support_movable; +} + /** * zpool_malloc() - Allocate memory * @zpool: The zpool to allocate from. diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c index 0787d33b80d8..8f3d9a4d46f4 100644 --- a/mm/zsmalloc.c +++ b/mm/zsmalloc.c @@ -437,15 +437,16 @@ static u64 zs_zpool_total_size(void *pool) } static struct zpool_driver zs_zpool_driver = { - .type = "zsmalloc", - .owner = THIS_MODULE, - .create = zs_zpool_create, - .destroy = zs_zpool_destroy, - .malloc = zs_zpool_malloc, - .free = zs_zpool_free, - .map = zs_zpool_map, - .unmap = zs_zpool_unmap, - .total_size = zs_zpool_total_size, + .type = "zsmalloc", + .owner = THIS_MODULE, + .create = zs_zpool_create, + .destroy = zs_zpool_destroy, + .malloc_support_movable = true, + .malloc = zs_zpool_malloc, + .free = zs_zpool_free, + .map = zs_zpool_map, + .unmap = zs_zpool_unmap, + .total_size = zs_zpool_total_size, }; MODULE_ALIAS("zpool-zsmalloc"); -- 2.20.1 (Apple Git-117) ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support 2019-06-02 9:46 [PATCH V2 1/2] zpool: Add malloc_support_movable to zpool_driver Hui Zhu @ 2019-06-02 9:46 ` Hui Zhu 2019-06-04 17:11 ` Shakeel Butt 0 siblings, 1 reply; 6+ messages in thread From: Hui Zhu @ 2019-06-02 9:46 UTC (permalink / raw) To: ddstreet, minchan, ngupta, sergey.senozhatsky.work, sjenning, linux-mm, linux-kernel Cc: Hui Zhu This is the second version that was updated according to the comments from Sergey Senozhatsky in https://lkml.org/lkml/2019/5/29/73 zswap compresses swap pages into a dynamically allocated RAM-based memory pool. The memory pool should be zbud, z3fold or zsmalloc. All of them will allocate unmovable pages. It will increase the number of unmovable page blocks that will bad for anti-fragment. zsmalloc support page migration if request movable page: handle = zs_malloc(zram->mem_pool, comp_len, GFP_NOIO | __GFP_HIGHMEM | __GFP_MOVABLE); And commit "zpool: Add malloc_support_movable to zpool_driver" add zpool_malloc_support_movable check malloc_support_movable to make sure if a zpool support allocate movable memory. This commit adds module parameter malloc_movable_if_support to enable or disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE if it support allocate movable memory (disabled by default). Following part is test log in a pc that has 8G memory and 2G swap. When it disabled: echo lz4 > /sys/module/zswap/parameters/compressor echo zsmalloc > /sys/module/zswap/parameters/zpool echo 1 > /sys/module/zswap/parameters/enabled swapon /swapfile cd /home/teawater/kernel/vm-scalability/ /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) /home/teawater/kernel/vm-scalability# ./case-anon-w-seq 2717908992 bytes / 3977932 usecs = 667233 KB/s 2717908992 bytes / 4160702 usecs = 637923 KB/s 2717908992 bytes / 4354611 usecs = 609516 KB/s 293359 usecs to free memory 340304 usecs to free memory 205781 usecs to free memory 2717908992 bytes / 5588016 usecs = 474982 KB/s 166124 usecs to free memory /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo Page block order: 9 Pages per block: 512 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type Unmovable 5 10 9 8 8 5 1 2 3 0 0 Node 0, zone DMA32, type Movable 15 16 14 12 14 10 9 6 6 5 776 Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone Normal, type Unmovable 7097 6914 6473 5642 4373 2664 1220 319 78 4 0 Node 0, zone Normal, type Movable 2092 3216 2820 2266 1585 946 559 359 237 258 378 Node 0, zone Normal, type Reclaimable 47 88 122 80 34 9 5 4 2 1 2 Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate Node 0, zone DMA 1 7 0 0 0 0 Node 0, zone DMA32 4 1652 0 0 0 0 Node 0, zone Normal 834 1572 25 0 0 0 When it enabled: echo lz4 > /sys/module/zswap/parameters/compressor echo zsmalloc > /sys/module/zswap/parameters/zpool echo 1 > /sys/module/zswap/parameters/enabled echo 1 > /sys/module/zswap/parameters/malloc_movable_if_support swapon /swapfile cd /home/teawater/kernel/vm-scalability/ /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) /home/teawater/kernel/vm-scalability# ./case-anon-w-seq 2717908992 bytes / 4721401 usecs = 562165 KB/s 2717908992 bytes / 4783167 usecs = 554905 KB/s 2717908992 bytes / 4802125 usecs = 552715 KB/s 2717908992 bytes / 4866579 usecs = 545395 KB/s 323605 usecs to free memory 414817 usecs to free memory 458576 usecs to free memory 355827 usecs to free memory /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo Page block order: 9 Pages per block: 512 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type Unmovable 8 10 8 7 7 6 5 3 2 0 0 Node 0, zone DMA32, type Movable 23 21 18 15 13 14 14 10 11 6 766 Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 1 Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone Normal, type Unmovable 2660 1295 460 102 11 5 3 11 2 4 0 Node 0, zone Normal, type Movable 4178 5760 5045 4137 3324 2306 1482 930 497 254 460 Node 0, zone Normal, type Reclaimable 50 83 114 93 28 12 10 6 3 3 0 Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate Node 0, zone DMA 1 7 0 0 0 0 Node 0, zone DMA32 4 1650 2 0 0 0 Node 0, zone Normal 81 2325 25 0 0 0 You can see that the number of unmovable page blocks is decreased when malloc_movable_if_support is enabled. Signed-off-by: Hui Zhu <teawaterz@linux.alibaba.com> --- mm/zswap.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/mm/zswap.c b/mm/zswap.c index a4e4d36ec085..2fc45de92383 100644 --- a/mm/zswap.c +++ b/mm/zswap.c @@ -123,6 +123,13 @@ static bool zswap_same_filled_pages_enabled = true; module_param_named(same_filled_pages_enabled, zswap_same_filled_pages_enabled, bool, 0644); +/* Enable/disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE + * if it support allocate movable memory (disabled by default). + */ +static bool __read_mostly zswap_malloc_movable_if_support; +module_param_cb(malloc_movable_if_support, ¶m_ops_bool, + &zswap_malloc_movable_if_support, 0644); + /********************************* * data structures **********************************/ @@ -1006,6 +1013,7 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, char *buf; u8 *src, *dst; struct zswap_header zhdr = { .swpentry = swp_entry(type, offset) }; + gfp_t gfp = __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM; /* THP isn't supported */ if (PageTransHuge(page)) { @@ -1079,9 +1087,11 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, /* store */ hlen = zpool_evictable(entry->pool->zpool) ? sizeof(zhdr) : 0; - ret = zpool_malloc(entry->pool->zpool, hlen + dlen, - __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM, - &handle); + if (zswap_malloc_movable_if_support && + zpool_malloc_support_movable(entry->pool->zpool)) { + gfp |= __GFP_HIGHMEM | __GFP_MOVABLE; + } + ret = zpool_malloc(entry->pool->zpool, hlen + dlen, gfp, &handle); if (ret == -ENOSPC) { zswap_reject_compress_poor++; goto put_dstmem; -- 2.20.1 (Apple Git-117) ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support 2019-06-02 9:46 ` [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support Hui Zhu @ 2019-06-04 17:11 ` Shakeel Butt 0 siblings, 0 replies; 6+ messages in thread From: Shakeel Butt @ 2019-06-04 17:11 UTC (permalink / raw) To: Hui Zhu Cc: Dan Streetman, Minchan Kim, ngupta, sergey.senozhatsky.work, Seth Jennings, Linux MM, LKML On Sun, Jun 2, 2019 at 2:47 AM Hui Zhu <teawaterz@linux.alibaba.com> wrote: > > This is the second version that was updated according to the comments > from Sergey Senozhatsky in https://lkml.org/lkml/2019/5/29/73 > > zswap compresses swap pages into a dynamically allocated RAM-based > memory pool. The memory pool should be zbud, z3fold or zsmalloc. > All of them will allocate unmovable pages. It will increase the > number of unmovable page blocks that will bad for anti-fragment. > > zsmalloc support page migration if request movable page: > handle = zs_malloc(zram->mem_pool, comp_len, > GFP_NOIO | __GFP_HIGHMEM | > __GFP_MOVABLE); > > And commit "zpool: Add malloc_support_movable to zpool_driver" add > zpool_malloc_support_movable check malloc_support_movable to make > sure if a zpool support allocate movable memory. > > This commit adds module parameter malloc_movable_if_support to enable > or disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > if it support allocate movable memory (disabled by default). > > Following part is test log in a pc that has 8G memory and 2G swap. > > When it disabled: > echo lz4 > /sys/module/zswap/parameters/compressor > echo zsmalloc > /sys/module/zswap/parameters/zpool > echo 1 > /sys/module/zswap/parameters/enabled > swapon /swapfile > cd /home/teawater/kernel/vm-scalability/ > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > 2717908992 bytes / 3977932 usecs = 667233 KB/s > 2717908992 bytes / 4160702 usecs = 637923 KB/s > 2717908992 bytes / 4354611 usecs = 609516 KB/s > 293359 usecs to free memory > 340304 usecs to free memory > 205781 usecs to free memory > 2717908992 bytes / 5588016 usecs = 474982 KB/s > 166124 usecs to free memory > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > Page block order: 9 > Pages per block: 512 > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Unmovable 5 10 9 8 8 5 1 2 3 0 0 > Node 0, zone DMA32, type Movable 15 16 14 12 14 10 9 6 6 5 776 > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Unmovable 7097 6914 6473 5642 4373 2664 1220 319 78 4 0 > Node 0, zone Normal, type Movable 2092 3216 2820 2266 1585 946 559 359 237 258 378 > Node 0, zone Normal, type Reclaimable 47 88 122 80 34 9 5 4 2 1 2 > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > Node 0, zone DMA 1 7 0 0 0 0 > Node 0, zone DMA32 4 1652 0 0 0 0 > Node 0, zone Normal 834 1572 25 0 0 0 > > When it enabled: > echo lz4 > /sys/module/zswap/parameters/compressor > echo zsmalloc > /sys/module/zswap/parameters/zpool > echo 1 > /sys/module/zswap/parameters/enabled > echo 1 > /sys/module/zswap/parameters/malloc_movable_if_support > swapon /swapfile > cd /home/teawater/kernel/vm-scalability/ > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > 2717908992 bytes / 4721401 usecs = 562165 KB/s > 2717908992 bytes / 4783167 usecs = 554905 KB/s > 2717908992 bytes / 4802125 usecs = 552715 KB/s > 2717908992 bytes / 4866579 usecs = 545395 KB/s > 323605 usecs to free memory > 414817 usecs to free memory > 458576 usecs to free memory > 355827 usecs to free memory > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > Page block order: 9 > Pages per block: 512 > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Unmovable 8 10 8 7 7 6 5 3 2 0 0 > Node 0, zone DMA32, type Movable 23 21 18 15 13 14 14 10 11 6 766 > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 1 > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Unmovable 2660 1295 460 102 11 5 3 11 2 4 0 > Node 0, zone Normal, type Movable 4178 5760 5045 4137 3324 2306 1482 930 497 254 460 > Node 0, zone Normal, type Reclaimable 50 83 114 93 28 12 10 6 3 3 0 > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > Node 0, zone DMA 1 7 0 0 0 0 > Node 0, zone DMA32 4 1650 2 0 0 0 > Node 0, zone Normal 81 2325 25 0 0 0 > > You can see that the number of unmovable page blocks is decreased > when malloc_movable_if_support is enabled. > > Signed-off-by: Hui Zhu <teawaterz@linux.alibaba.com> > --- > mm/zswap.c | 16 +++++++++++++--- > 1 file changed, 13 insertions(+), 3 deletions(-) > > diff --git a/mm/zswap.c b/mm/zswap.c > index a4e4d36ec085..2fc45de92383 100644 > --- a/mm/zswap.c > +++ b/mm/zswap.c > @@ -123,6 +123,13 @@ static bool zswap_same_filled_pages_enabled = true; > module_param_named(same_filled_pages_enabled, zswap_same_filled_pages_enabled, > bool, 0644); > > +/* Enable/disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > + * if it support allocate movable memory (disabled by default). > + */ > +static bool __read_mostly zswap_malloc_movable_if_support; > +module_param_cb(malloc_movable_if_support, ¶m_ops_bool, > + &zswap_malloc_movable_if_support, 0644); > + Any reason for the above tunable? Do we ever want to disable movable for zswap+zsmalloc? > /********************************* > * data structures > **********************************/ > @@ -1006,6 +1013,7 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > char *buf; > u8 *src, *dst; > struct zswap_header zhdr = { .swpentry = swp_entry(type, offset) }; > + gfp_t gfp = __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM; > > /* THP isn't supported */ > if (PageTransHuge(page)) { > @@ -1079,9 +1087,11 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > > /* store */ > hlen = zpool_evictable(entry->pool->zpool) ? sizeof(zhdr) : 0; > - ret = zpool_malloc(entry->pool->zpool, hlen + dlen, > - __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM, > - &handle); > + if (zswap_malloc_movable_if_support && > + zpool_malloc_support_movable(entry->pool->zpool)) { > + gfp |= __GFP_HIGHMEM | __GFP_MOVABLE; > + } > + ret = zpool_malloc(entry->pool->zpool, hlen + dlen, gfp, &handle); > if (ret == -ENOSPC) { > zswap_reject_compress_poor++; > goto put_dstmem; > -- > 2.20.1 (Apple Git-117) > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support @ 2019-06-04 17:11 ` Shakeel Butt 0 siblings, 0 replies; 6+ messages in thread From: Shakeel Butt @ 2019-06-04 17:11 UTC (permalink / raw) To: Hui Zhu Cc: Dan Streetman, Minchan Kim, ngupta, sergey.senozhatsky.work, Seth Jennings, Linux MM, LKML On Sun, Jun 2, 2019 at 2:47 AM Hui Zhu <teawaterz@linux.alibaba.com> wrote: > > This is the second version that was updated according to the comments > from Sergey Senozhatsky in https://lkml.org/lkml/2019/5/29/73 > > zswap compresses swap pages into a dynamically allocated RAM-based > memory pool. The memory pool should be zbud, z3fold or zsmalloc. > All of them will allocate unmovable pages. It will increase the > number of unmovable page blocks that will bad for anti-fragment. > > zsmalloc support page migration if request movable page: > handle = zs_malloc(zram->mem_pool, comp_len, > GFP_NOIO | __GFP_HIGHMEM | > __GFP_MOVABLE); > > And commit "zpool: Add malloc_support_movable to zpool_driver" add > zpool_malloc_support_movable check malloc_support_movable to make > sure if a zpool support allocate movable memory. > > This commit adds module parameter malloc_movable_if_support to enable > or disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > if it support allocate movable memory (disabled by default). > > Following part is test log in a pc that has 8G memory and 2G swap. > > When it disabled: > echo lz4 > /sys/module/zswap/parameters/compressor > echo zsmalloc > /sys/module/zswap/parameters/zpool > echo 1 > /sys/module/zswap/parameters/enabled > swapon /swapfile > cd /home/teawater/kernel/vm-scalability/ > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > 2717908992 bytes / 3977932 usecs = 667233 KB/s > 2717908992 bytes / 4160702 usecs = 637923 KB/s > 2717908992 bytes / 4354611 usecs = 609516 KB/s > 293359 usecs to free memory > 340304 usecs to free memory > 205781 usecs to free memory > 2717908992 bytes / 5588016 usecs = 474982 KB/s > 166124 usecs to free memory > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > Page block order: 9 > Pages per block: 512 > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Unmovable 5 10 9 8 8 5 1 2 3 0 0 > Node 0, zone DMA32, type Movable 15 16 14 12 14 10 9 6 6 5 776 > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Unmovable 7097 6914 6473 5642 4373 2664 1220 319 78 4 0 > Node 0, zone Normal, type Movable 2092 3216 2820 2266 1585 946 559 359 237 258 378 > Node 0, zone Normal, type Reclaimable 47 88 122 80 34 9 5 4 2 1 2 > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > Node 0, zone DMA 1 7 0 0 0 0 > Node 0, zone DMA32 4 1652 0 0 0 0 > Node 0, zone Normal 834 1572 25 0 0 0 > > When it enabled: > echo lz4 > /sys/module/zswap/parameters/compressor > echo zsmalloc > /sys/module/zswap/parameters/zpool > echo 1 > /sys/module/zswap/parameters/enabled > echo 1 > /sys/module/zswap/parameters/malloc_movable_if_support > swapon /swapfile > cd /home/teawater/kernel/vm-scalability/ > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > 2717908992 bytes / 4721401 usecs = 562165 KB/s > 2717908992 bytes / 4783167 usecs = 554905 KB/s > 2717908992 bytes / 4802125 usecs = 552715 KB/s > 2717908992 bytes / 4866579 usecs = 545395 KB/s > 323605 usecs to free memory > 414817 usecs to free memory > 458576 usecs to free memory > 355827 usecs to free memory > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > Page block order: 9 > Pages per block: 512 > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Unmovable 8 10 8 7 7 6 5 3 2 0 0 > Node 0, zone DMA32, type Movable 23 21 18 15 13 14 14 10 11 6 766 > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 1 > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Unmovable 2660 1295 460 102 11 5 3 11 2 4 0 > Node 0, zone Normal, type Movable 4178 5760 5045 4137 3324 2306 1482 930 497 254 460 > Node 0, zone Normal, type Reclaimable 50 83 114 93 28 12 10 6 3 3 0 > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > Node 0, zone DMA 1 7 0 0 0 0 > Node 0, zone DMA32 4 1650 2 0 0 0 > Node 0, zone Normal 81 2325 25 0 0 0 > > You can see that the number of unmovable page blocks is decreased > when malloc_movable_if_support is enabled. > > Signed-off-by: Hui Zhu <teawaterz@linux.alibaba.com> > --- > mm/zswap.c | 16 +++++++++++++--- > 1 file changed, 13 insertions(+), 3 deletions(-) > > diff --git a/mm/zswap.c b/mm/zswap.c > index a4e4d36ec085..2fc45de92383 100644 > --- a/mm/zswap.c > +++ b/mm/zswap.c > @@ -123,6 +123,13 @@ static bool zswap_same_filled_pages_enabled = true; > module_param_named(same_filled_pages_enabled, zswap_same_filled_pages_enabled, > bool, 0644); > > +/* Enable/disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > + * if it support allocate movable memory (disabled by default). > + */ > +static bool __read_mostly zswap_malloc_movable_if_support; > +module_param_cb(malloc_movable_if_support, ¶m_ops_bool, > + &zswap_malloc_movable_if_support, 0644); > + Any reason for the above tunable? Do we ever want to disable movable for zswap+zsmalloc? > /********************************* > * data structures > **********************************/ > @@ -1006,6 +1013,7 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > char *buf; > u8 *src, *dst; > struct zswap_header zhdr = { .swpentry = swp_entry(type, offset) }; > + gfp_t gfp = __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM; > > /* THP isn't supported */ > if (PageTransHuge(page)) { > @@ -1079,9 +1087,11 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > > /* store */ > hlen = zpool_evictable(entry->pool->zpool) ? sizeof(zhdr) : 0; > - ret = zpool_malloc(entry->pool->zpool, hlen + dlen, > - __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM, > - &handle); > + if (zswap_malloc_movable_if_support && > + zpool_malloc_support_movable(entry->pool->zpool)) { > + gfp |= __GFP_HIGHMEM | __GFP_MOVABLE; > + } > + ret = zpool_malloc(entry->pool->zpool, hlen + dlen, gfp, &handle); > if (ret == -ENOSPC) { > zswap_reject_compress_poor++; > goto put_dstmem; > -- > 2.20.1 (Apple Git-117) > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support 2019-06-04 17:11 ` Shakeel Butt @ 2019-06-05 2:33 ` Hui Zhu -1 siblings, 0 replies; 6+ messages in thread From: Hui Zhu @ 2019-06-05 2:33 UTC (permalink / raw) To: Shakeel Butt Cc: Hui Zhu, Dan Streetman, Minchan Kim, ngupta, Sergey Senozhatsky, Seth Jennings, Linux MM, LKML Shakeel Butt <shakeelb@google.com> 于2019年6月5日周三 上午1:12写道： > > On Sun, Jun 2, 2019 at 2:47 AM Hui Zhu <teawaterz@linux.alibaba.com> wrote: > > > > This is the second version that was updated according to the comments > > from Sergey Senozhatsky in https://lkml.org/lkml/2019/5/29/73 > > > > zswap compresses swap pages into a dynamically allocated RAM-based > > memory pool. The memory pool should be zbud, z3fold or zsmalloc. > > All of them will allocate unmovable pages. It will increase the > > number of unmovable page blocks that will bad for anti-fragment. > > > > zsmalloc support page migration if request movable page: > > handle = zs_malloc(zram->mem_pool, comp_len, > > GFP_NOIO | __GFP_HIGHMEM | > > __GFP_MOVABLE); > > > > And commit "zpool: Add malloc_support_movable to zpool_driver" add > > zpool_malloc_support_movable check malloc_support_movable to make > > sure if a zpool support allocate movable memory. > > > > This commit adds module parameter malloc_movable_if_support to enable > > or disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > > if it support allocate movable memory (disabled by default). > > > > Following part is test log in a pc that has 8G memory and 2G swap. > > > > When it disabled: > > echo lz4 > /sys/module/zswap/parameters/compressor > > echo zsmalloc > /sys/module/zswap/parameters/zpool > > echo 1 > /sys/module/zswap/parameters/enabled > > swapon /swapfile > > cd /home/teawater/kernel/vm-scalability/ > > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > > 2717908992 bytes / 3977932 usecs = 667233 KB/s > > 2717908992 bytes / 4160702 usecs = 637923 KB/s > > 2717908992 bytes / 4354611 usecs = 609516 KB/s > > 293359 usecs to free memory > > 340304 usecs to free memory > > 205781 usecs to free memory > > 2717908992 bytes / 5588016 usecs = 474982 KB/s > > 166124 usecs to free memory > > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > > Page block order: 9 > > Pages per block: 512 > > > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Unmovable 5 10 9 8 8 5 1 2 3 0 0 > > Node 0, zone DMA32, type Movable 15 16 14 12 14 10 9 6 6 5 776 > > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Unmovable 7097 6914 6473 5642 4373 2664 1220 319 78 4 0 > > Node 0, zone Normal, type Movable 2092 3216 2820 2266 1585 946 559 359 237 258 378 > > Node 0, zone Normal, type Reclaimable 47 88 122 80 34 9 5 4 2 1 2 > > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > > Node 0, zone DMA 1 7 0 0 0 0 > > Node 0, zone DMA32 4 1652 0 0 0 0 > > Node 0, zone Normal 834 1572 25 0 0 0 > > > > When it enabled: > > echo lz4 > /sys/module/zswap/parameters/compressor > > echo zsmalloc > /sys/module/zswap/parameters/zpool > > echo 1 > /sys/module/zswap/parameters/enabled > > echo 1 > /sys/module/zswap/parameters/malloc_movable_if_support > > swapon /swapfile > > cd /home/teawater/kernel/vm-scalability/ > > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > > 2717908992 bytes / 4721401 usecs = 562165 KB/s > > 2717908992 bytes / 4783167 usecs = 554905 KB/s > > 2717908992 bytes / 4802125 usecs = 552715 KB/s > > 2717908992 bytes / 4866579 usecs = 545395 KB/s > > 323605 usecs to free memory > > 414817 usecs to free memory > > 458576 usecs to free memory > > 355827 usecs to free memory > > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > > Page block order: 9 > > Pages per block: 512 > > > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Unmovable 8 10 8 7 7 6 5 3 2 0 0 > > Node 0, zone DMA32, type Movable 23 21 18 15 13 14 14 10 11 6 766 > > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 1 > > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Unmovable 2660 1295 460 102 11 5 3 11 2 4 0 > > Node 0, zone Normal, type Movable 4178 5760 5045 4137 3324 2306 1482 930 497 254 460 > > Node 0, zone Normal, type Reclaimable 50 83 114 93 28 12 10 6 3 3 0 > > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > > Node 0, zone DMA 1 7 0 0 0 0 > > Node 0, zone DMA32 4 1650 2 0 0 0 > > Node 0, zone Normal 81 2325 25 0 0 0 > > > > You can see that the number of unmovable page blocks is decreased > > when malloc_movable_if_support is enabled. > > > > Signed-off-by: Hui Zhu <teawaterz@linux.alibaba.com> > > --- > > mm/zswap.c | 16 +++++++++++++--- > > 1 file changed, 13 insertions(+), 3 deletions(-) > > > > diff --git a/mm/zswap.c b/mm/zswap.c > > index a4e4d36ec085..2fc45de92383 100644 > > --- a/mm/zswap.c > > +++ b/mm/zswap.c > > @@ -123,6 +123,13 @@ static bool zswap_same_filled_pages_enabled = true; > > module_param_named(same_filled_pages_enabled, zswap_same_filled_pages_enabled, > > bool, 0644); > > > > +/* Enable/disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > > + * if it support allocate movable memory (disabled by default). > > + */ > > +static bool __read_mostly zswap_malloc_movable_if_support; > > +module_param_cb(malloc_movable_if_support, ¶m_ops_bool, > > + &zswap_malloc_movable_if_support, 0644); > > + > > Any reason for the above tunable? Do we ever want to disable movable > for zswap+zsmalloc? Thanks for your remind. I will post a new version that remove this module_param later. Best, Hui > > > /********************************* > > * data structures > > **********************************/ > > @@ -1006,6 +1013,7 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > > char *buf; > > u8 *src, *dst; > > struct zswap_header zhdr = { .swpentry = swp_entry(type, offset) }; > > + gfp_t gfp = __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM; > > > > /* THP isn't supported */ > > if (PageTransHuge(page)) { > > @@ -1079,9 +1087,11 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > > > > /* store */ > > hlen = zpool_evictable(entry->pool->zpool) ? sizeof(zhdr) : 0; > > - ret = zpool_malloc(entry->pool->zpool, hlen + dlen, > > - __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM, > > - &handle); > > + if (zswap_malloc_movable_if_support && > > + zpool_malloc_support_movable(entry->pool->zpool)) { > > + gfp |= __GFP_HIGHMEM | __GFP_MOVABLE; > > + } > > + ret = zpool_malloc(entry->pool->zpool, hlen + dlen, gfp, &handle); > > if (ret == -ENOSPC) { > > zswap_reject_compress_poor++; > > goto put_dstmem; > > -- > > 2.20.1 (Apple Git-117) > > > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support @ 2019-06-05 2:33 ` Hui Zhu 0 siblings, 0 replies; 6+ messages in thread From: Hui Zhu @ 2019-06-05 2:33 UTC (permalink / raw) To: Shakeel Butt Cc: Hui Zhu, Dan Streetman, Minchan Kim, ngupta, Sergey Senozhatsky, Seth Jennings, Linux MM, LKML Shakeel Butt <shakeelb@google.com> 于2019年6月5日周三 上午1:12写道： > > On Sun, Jun 2, 2019 at 2:47 AM Hui Zhu <teawaterz@linux.alibaba.com> wrote: > > > > This is the second version that was updated according to the comments > > from Sergey Senozhatsky in https://lkml.org/lkml/2019/5/29/73 > > > > zswap compresses swap pages into a dynamically allocated RAM-based > > memory pool. The memory pool should be zbud, z3fold or zsmalloc. > > All of them will allocate unmovable pages. It will increase the > > number of unmovable page blocks that will bad for anti-fragment. > > > > zsmalloc support page migration if request movable page: > > handle = zs_malloc(zram->mem_pool, comp_len, > > GFP_NOIO | __GFP_HIGHMEM | > > __GFP_MOVABLE); > > > > And commit "zpool: Add malloc_support_movable to zpool_driver" add > > zpool_malloc_support_movable check malloc_support_movable to make > > sure if a zpool support allocate movable memory. > > > > This commit adds module parameter malloc_movable_if_support to enable > > or disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > > if it support allocate movable memory (disabled by default). > > > > Following part is test log in a pc that has 8G memory and 2G swap. > > > > When it disabled: > > echo lz4 > /sys/module/zswap/parameters/compressor > > echo zsmalloc > /sys/module/zswap/parameters/zpool > > echo 1 > /sys/module/zswap/parameters/enabled > > swapon /swapfile > > cd /home/teawater/kernel/vm-scalability/ > > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > > 2717908992 bytes / 3977932 usecs = 667233 KB/s > > 2717908992 bytes / 4160702 usecs = 637923 KB/s > > 2717908992 bytes / 4354611 usecs = 609516 KB/s > > 293359 usecs to free memory > > 340304 usecs to free memory > > 205781 usecs to free memory > > 2717908992 bytes / 5588016 usecs = 474982 KB/s > > 166124 usecs to free memory > > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > > Page block order: 9 > > Pages per block: 512 > > > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Unmovable 5 10 9 8 8 5 1 2 3 0 0 > > Node 0, zone DMA32, type Movable 15 16 14 12 14 10 9 6 6 5 776 > > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Unmovable 7097 6914 6473 5642 4373 2664 1220 319 78 4 0 > > Node 0, zone Normal, type Movable 2092 3216 2820 2266 1585 946 559 359 237 258 378 > > Node 0, zone Normal, type Reclaimable 47 88 122 80 34 9 5 4 2 1 2 > > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > > Node 0, zone DMA 1 7 0 0 0 0 > > Node 0, zone DMA32 4 1652 0 0 0 0 > > Node 0, zone Normal 834 1572 25 0 0 0 > > > > When it enabled: > > echo lz4 > /sys/module/zswap/parameters/compressor > > echo zsmalloc > /sys/module/zswap/parameters/zpool > > echo 1 > /sys/module/zswap/parameters/enabled > > echo 1 > /sys/module/zswap/parameters/malloc_movable_if_support > > swapon /swapfile > > cd /home/teawater/kernel/vm-scalability/ > > /home/teawater/kernel/vm-scalability# export unit_size=$((9 * 1024 * 1024 * 1024)) > > /home/teawater/kernel/vm-scalability# ./case-anon-w-seq > > 2717908992 bytes / 4721401 usecs = 562165 KB/s > > 2717908992 bytes / 4783167 usecs = 554905 KB/s > > 2717908992 bytes / 4802125 usecs = 552715 KB/s > > 2717908992 bytes / 4866579 usecs = 545395 KB/s > > 323605 usecs to free memory > > 414817 usecs to free memory > > 458576 usecs to free memory > > 355827 usecs to free memory > > /home/teawater/kernel/vm-scalability# cat /proc/pagetypeinfo > > Page block order: 9 > > Pages per block: 512 > > > > Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10 > > Node 0, zone DMA, type Unmovable 1 1 1 0 2 1 1 0 1 0 0 > > Node 0, zone DMA, type Movable 0 0 0 0 0 0 0 0 0 1 3 > > Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Unmovable 8 10 8 7 7 6 5 3 2 0 0 > > Node 0, zone DMA32, type Movable 23 21 18 15 13 14 14 10 11 6 766 > > Node 0, zone DMA32, type Reclaimable 0 0 0 0 0 0 0 0 0 0 1 > > Node 0, zone DMA32, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Unmovable 2660 1295 460 102 11 5 3 11 2 4 0 > > Node 0, zone Normal, type Movable 4178 5760 5045 4137 3324 2306 1482 930 497 254 460 > > Node 0, zone Normal, type Reclaimable 50 83 114 93 28 12 10 6 3 3 0 > > Node 0, zone Normal, type HighAtomic 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type CMA 0 0 0 0 0 0 0 0 0 0 0 > > Node 0, zone Normal, type Isolate 0 0 0 0 0 0 0 0 0 0 0 > > > > Number of blocks type Unmovable Movable Reclaimable HighAtomic CMA Isolate > > Node 0, zone DMA 1 7 0 0 0 0 > > Node 0, zone DMA32 4 1650 2 0 0 0 > > Node 0, zone Normal 81 2325 25 0 0 0 > > > > You can see that the number of unmovable page blocks is decreased > > when malloc_movable_if_support is enabled. > > > > Signed-off-by: Hui Zhu <teawaterz@linux.alibaba.com> > > --- > > mm/zswap.c | 16 +++++++++++++--- > > 1 file changed, 13 insertions(+), 3 deletions(-) > > > > diff --git a/mm/zswap.c b/mm/zswap.c > > index a4e4d36ec085..2fc45de92383 100644 > > --- a/mm/zswap.c > > +++ b/mm/zswap.c > > @@ -123,6 +123,13 @@ static bool zswap_same_filled_pages_enabled = true; > > module_param_named(same_filled_pages_enabled, zswap_same_filled_pages_enabled, > > bool, 0644); > > > > +/* Enable/disable zpool allocate block with gfp __GFP_HIGHMEM | __GFP_MOVABLE > > + * if it support allocate movable memory (disabled by default). > > + */ > > +static bool __read_mostly zswap_malloc_movable_if_support; > > +module_param_cb(malloc_movable_if_support, ¶m_ops_bool, > > + &zswap_malloc_movable_if_support, 0644); > > + > > Any reason for the above tunable? Do we ever want to disable movable > for zswap+zsmalloc? Thanks for your remind. I will post a new version that remove this module_param later. Best, Hui > > > /********************************* > > * data structures > > **********************************/ > > @@ -1006,6 +1013,7 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > > char *buf; > > u8 *src, *dst; > > struct zswap_header zhdr = { .swpentry = swp_entry(type, offset) }; > > + gfp_t gfp = __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM; > > > > /* THP isn't supported */ > > if (PageTransHuge(page)) { > > @@ -1079,9 +1087,11 @@ static int zswap_frontswap_store(unsigned type, pgoff_t offset, > > > > /* store */ > > hlen = zpool_evictable(entry->pool->zpool) ? sizeof(zhdr) : 0; > > - ret = zpool_malloc(entry->pool->zpool, hlen + dlen, > > - __GFP_NORETRY | __GFP_NOWARN | __GFP_KSWAPD_RECLAIM, > > - &handle); > > + if (zswap_malloc_movable_if_support && > > + zpool_malloc_support_movable(entry->pool->zpool)) { > > + gfp |= __GFP_HIGHMEM | __GFP_MOVABLE; > > + } > > + ret = zpool_malloc(entry->pool->zpool, hlen + dlen, gfp, &handle); > > if (ret == -ENOSPC) { > > zswap_reject_compress_poor++; > > goto put_dstmem; > > -- > > 2.20.1 (Apple Git-117) > > > ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2019-06-05 2:33 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-06-02 9:46 [PATCH V2 1/2] zpool: Add malloc_support_movable to zpool_driver Hui Zhu 2019-06-02 9:46 ` [PATCH V2 2/2] zswap: Add module parameter malloc_movable_if_support Hui Zhu 2019-06-04 17:11 ` Shakeel Butt 2019-06-04 17:11 ` Shakeel Butt 2019-06-05 2:33 ` Hui Zhu 2019-06-05 2:33 ` Hui Zhu
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.