From 56e5a103a721d0ef139bba7ff3d3ada6c8217d5b Mon Sep 17 00:00:00 2001 From: Nhat Pham Date: Wed, 2 Apr 2025 13:44:16 -0700 Subject: zsmalloc: prefer the the original page's node for compressed data Currently, zsmalloc, zswap's and zram's backend memory allocator, does not enforce any policy for the allocation of memory for the compressed data, instead just adopting the memory policy of the task entering reclaim, or the default policy (prefer local node) if no such policy is specified. This can lead to several pathological behaviors in multi-node NUMA systems: 1. Systems with CXL-based memory tiering can encounter the following inversion with zswap/zram: the coldest pages demoted to the CXL tier can return to the high tier when they are reclaimed to compressed swap, creating memory pressure on the high tier. 2. Consider a direct reclaimer scanning nodes in order of allocation preference. If it ventures into remote nodes, the memory it compresses there should stay there. Trying to shift those contents over to the reclaiming thread's preferred node further *increases* its local pressure, and provoking more spills. The remote node is also the most likely to refault this data again. This undesirable behavior was pointed out by Johannes Weiner in [1]. 3. For zswap writeback, the zswap entries are organized in node-specific LRUs, based on the node placement of the original pages, allowing for targeted zswap writeback for specific nodes. However, the compressed data of a zswap entry can be placed on a different node from the LRU it is placed on. This means that reclaim targeted at one node might not free up memory used for zswap entries in that node, but instead reclaiming memory in a different node. All of these issues will be resolved if the compressed data go to the same node as the original page. This patch encourages this behavior by having zswap and zram pass the node of the original page to zsmalloc, and have zsmalloc prefer the specified node if we need to allocate new (zs)pages for the compressed data. Note that we are not strictly binding the allocation to the preferred node. We still allow the allocation to fall back to other nodes when the preferred node is full, or if we have zspages with slots available on a different node. This is OK, and still a strict improvement over the status quo: 1. On a system with demotion enabled, we will generally prefer demotions over compressed swapping, and only swap when pages have already gone to the lowest tier. This patch should achieve the desired effect for the most part. 2. If the preferred node is out of memory, letting the compressed data going to other nodes can be better than the alternative (OOMs, keeping cold memory unreclaimed, disk swapping, etc.). 3. If the allocation go to a separate node because we have a zspage with slots available, at least we're not creating extra immediate memory pressure (since the space is already allocated). 3. While there can be mixings, we generally reclaim pages in same-node batches, which encourage zspage grouping that is more likely to go to the right node. 4. A strict binding would require partitioning zsmalloc by node, which is more complicated, and more prone to regression, since it reduces the storage density of zsmalloc. We need to evaluate the tradeoff and benchmark carefully before adopting such an involved solution. [1]: https://lore.kernel.org/linux-mm/20250331165306.GC2110528@cmpxchg.org/ [senozhatsky@chromium.org: coding-style fixes] Link: https://lkml.kernel.org/r/mnvexa7kseswglcqbhlot4zg3b3la2ypv2rimdl5mh5glbmhvz@wi6bgqn47hge Link: https://lkml.kernel.org/r/20250402204416.3435994-1-nphamcs@gmail.com Signed-off-by: Nhat Pham Suggested-by: Gregory Price Acked-by: Dan Williams Reviewed-by: Chengming Zhou Acked-by: Sergey Senozhatsky [zram, zsmalloc] Acked-by: Johannes Weiner Acked-by: Yosry Ahmed [zswap/zsmalloc] Cc: "Huang, Ying" Cc: Joanthan Cameron Cc: Minchan Kim Cc: SeongJae Park Signed-off-by: Andrew Morton --- drivers/block/zram/zram_drv.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'drivers/block') diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c index fda7d8624889..0ba18277ed7b 100644 --- a/drivers/block/zram/zram_drv.c +++ b/drivers/block/zram/zram_drv.c @@ -1694,7 +1694,7 @@ static int write_incompressible_page(struct zram *zram, struct page *page, */ handle = zs_malloc(zram->mem_pool, PAGE_SIZE, GFP_NOIO | __GFP_NOWARN | - __GFP_HIGHMEM | __GFP_MOVABLE); + __GFP_HIGHMEM | __GFP_MOVABLE, page_to_nid(page)); if (IS_ERR_VALUE(handle)) return PTR_ERR((void *)handle); @@ -1761,7 +1761,7 @@ static int zram_write_page(struct zram *zram, struct page *page, u32 index) handle = zs_malloc(zram->mem_pool, comp_len, GFP_NOIO | __GFP_NOWARN | - __GFP_HIGHMEM | __GFP_MOVABLE); + __GFP_HIGHMEM | __GFP_MOVABLE, page_to_nid(page)); if (IS_ERR_VALUE(handle)) { zcomp_stream_put(zstrm); return PTR_ERR((void *)handle); @@ -1981,10 +1981,15 @@ static int recompress_slot(struct zram *zram, u32 index, struct page *page, * We are holding per-CPU stream mutex and entry lock so better * avoid direct reclaim. Allocation error is not fatal since * we still have the old object in the mem_pool. + * + * XXX: technically, the node we really want here is the node that holds + * the original compressed data. But that would require us to modify + * zsmalloc API to return this information. For now, we will make do with + * the node of the page allocated for recompression. */ handle_new = zs_malloc(zram->mem_pool, comp_len_new, GFP_NOIO | __GFP_NOWARN | - __GFP_HIGHMEM | __GFP_MOVABLE); + __GFP_HIGHMEM | __GFP_MOVABLE, page_to_nid(page)); if (IS_ERR_VALUE(handle_new)) { zcomp_stream_put(zstrm); return PTR_ERR((void *)handle_new); -- cgit v1.2.3 From cf42d4cccf0d01e375175393776a16dc47b9996f Mon Sep 17 00:00:00 2001 From: Sergey Senozhatsky Date: Thu, 27 Mar 2025 10:58:09 +0900 Subject: zram: modernize writeback interface The writeback interface supports a page_index=N parameter which performs writeback of the given page. Since we rarely need to writeback just one single page, the typical use case involves a number of writeback calls, each performing writeback of one page: echo page_index=100 > zram0/writeback ... echo page_index=200 > zram0/writeback echo page_index=500 > zram0/writeback ... echo page_index=700 > zram0/writeback One obvious downside of this is that it increases the number of syscalls. Less obvious, but a significantly more important downside, is that when given only one page to post-process zram cannot perform an optimal target selection. This becomes a critical limitation when writeback_limit is enabled, because under writeback_limit we want to guarantee the highest memory savings hence we first need to writeback pages that release the highest amount of zsmalloc pool memory. This patch adds page_indexes=LOW-HIGH parameter to the writeback interface: echo page_indexes=100-200 page_indexes=500-700 > zram0/writeback This gives zram a chance to apply an optimal target selection strategy on each iteration of the writeback loop. We also now permit multiple page_index parameters per call (previously zram would recognize only one page_index) and a mix or single pages and page ranges: echo page_index=42 page_index=99 page_indexes=100-200 \ page_indexes=500-700 > zram0/writeback Apart from that the patch also unifies parameters passing and resembles other "modern" zram device attributes (e.g. recompression), while the old interface used a mixed scheme: values-less parameters for mode and a key=value format for page_index. We still support the "old" value-less format for compatibility reasons. [senozhatsky@chromium.org: simplify parse_page_index() range checks, per Brian] nk: https://lkml.kernel.org/r/20250404015327.2427684-1-senozhatsky@chromium.org [sozhatsky@chromium.org: fix uninitialized variable in zram_writeback_slots(), per Dan] nk: https://lkml.kernel.org/r/20250409112611.1154282-1-senozhatsky@chromium.org Link: https://lkml.kernel.org/r/20250327015818.4148660-1-senozhatsky@chromium.org Signed-off-by: Sergey Senozhatsky Reviewed-by: Brian Geffon Cc: Minchan Kim Cc: Richard Chang Cc: Sergey Senozhatsky Cc: Dan Carpenter Signed-off-by: Andrew Morton --- Documentation/admin-guide/blockdev/zram.rst | 17 ++ drivers/block/zram/zram_drv.c | 320 +++++++++++++++++++--------- 2 files changed, 232 insertions(+), 105 deletions(-) (limited to 'drivers/block') diff --git a/Documentation/admin-guide/blockdev/zram.rst b/Documentation/admin-guide/blockdev/zram.rst index 9bdb30901a93..b8d36134a151 100644 --- a/Documentation/admin-guide/blockdev/zram.rst +++ b/Documentation/admin-guide/blockdev/zram.rst @@ -369,6 +369,23 @@ they could write a page index into the interface:: echo "page_index=1251" > /sys/block/zramX/writeback +In Linux 6.16 this interface underwent some rework. First, the interface +now supports `key=value` format for all of its parameters (`type=huge_idle`, +etc.) Second, the support for `page_indexes` was introduced, which specify +`LOW-HIGH` range (or ranges) of pages to be written-back. This reduces the +number of syscalls, but more importantly this enables optimal post-processing +target selection strategy. Usage example:: + + echo "type=idle" > /sys/block/zramX/writeback + echo "page_indexes=1-100 page_indexes=200-300" > \ + /sys/block/zramX/writeback + +We also now permit multiple page_index params per call and a mix of +single pages and page ranges:: + + echo page_index=42 page_index=99 page_indexes=100-200 \ + page_indexes=500-700 > /sys/block/zramX/writeback + If there are lots of write IO with flash device, potentially, it has flash wearout problem so that admin needs to design write limitation to guarantee storage health for entire product life. diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c index 0ba18277ed7b..94e6e9b80bf0 100644 --- a/drivers/block/zram/zram_drv.c +++ b/drivers/block/zram/zram_drv.c @@ -734,114 +734,19 @@ static void read_from_bdev_async(struct zram *zram, struct page *page, submit_bio(bio); } -#define PAGE_WB_SIG "page_index=" - -#define PAGE_WRITEBACK 0 -#define HUGE_WRITEBACK (1<<0) -#define IDLE_WRITEBACK (1<<1) -#define INCOMPRESSIBLE_WRITEBACK (1<<2) - -static int scan_slots_for_writeback(struct zram *zram, u32 mode, - unsigned long nr_pages, - unsigned long index, - struct zram_pp_ctl *ctl) +static int zram_writeback_slots(struct zram *zram, struct zram_pp_ctl *ctl) { - for (; nr_pages != 0; index++, nr_pages--) { - bool ok = true; - - zram_slot_lock(zram, index); - if (!zram_allocated(zram, index)) - goto next; - - if (zram_test_flag(zram, index, ZRAM_WB) || - zram_test_flag(zram, index, ZRAM_SAME)) - goto next; - - if (mode & IDLE_WRITEBACK && - !zram_test_flag(zram, index, ZRAM_IDLE)) - goto next; - if (mode & HUGE_WRITEBACK && - !zram_test_flag(zram, index, ZRAM_HUGE)) - goto next; - if (mode & INCOMPRESSIBLE_WRITEBACK && - !zram_test_flag(zram, index, ZRAM_INCOMPRESSIBLE)) - goto next; - - ok = place_pp_slot(zram, ctl, index); -next: - zram_slot_unlock(zram, index); - if (!ok) - break; - } - - return 0; -} - -static ssize_t writeback_store(struct device *dev, - struct device_attribute *attr, const char *buf, size_t len) -{ - struct zram *zram = dev_to_zram(dev); - unsigned long nr_pages = zram->disksize >> PAGE_SHIFT; - struct zram_pp_ctl *ctl = NULL; + unsigned long blk_idx = 0; + struct page *page = NULL; struct zram_pp_slot *pps; - unsigned long index = 0; - struct bio bio; struct bio_vec bio_vec; - struct page *page = NULL; - ssize_t ret = len; - int mode, err; - unsigned long blk_idx = 0; - - if (sysfs_streq(buf, "idle")) - mode = IDLE_WRITEBACK; - else if (sysfs_streq(buf, "huge")) - mode = HUGE_WRITEBACK; - else if (sysfs_streq(buf, "huge_idle")) - mode = IDLE_WRITEBACK | HUGE_WRITEBACK; - else if (sysfs_streq(buf, "incompressible")) - mode = INCOMPRESSIBLE_WRITEBACK; - else { - if (strncmp(buf, PAGE_WB_SIG, sizeof(PAGE_WB_SIG) - 1)) - return -EINVAL; - - if (kstrtol(buf + sizeof(PAGE_WB_SIG) - 1, 10, &index) || - index >= nr_pages) - return -EINVAL; - - nr_pages = 1; - mode = PAGE_WRITEBACK; - } - - down_read(&zram->init_lock); - if (!init_done(zram)) { - ret = -EINVAL; - goto release_init_lock; - } - - /* Do not permit concurrent post-processing actions. */ - if (atomic_xchg(&zram->pp_in_progress, 1)) { - up_read(&zram->init_lock); - return -EAGAIN; - } - - if (!zram->backing_dev) { - ret = -ENODEV; - goto release_init_lock; - } + struct bio bio; + int ret = 0, err; + u32 index; page = alloc_page(GFP_KERNEL); - if (!page) { - ret = -ENOMEM; - goto release_init_lock; - } - - ctl = init_pp_ctl(); - if (!ctl) { - ret = -ENOMEM; - goto release_init_lock; - } - - scan_slots_for_writeback(zram, mode, nr_pages, index, ctl); + if (!page) + return -ENOMEM; while ((pps = select_pp_slot(ctl))) { spin_lock(&zram->wb_limit_lock); @@ -929,10 +834,215 @@ next: if (blk_idx) free_block_bdev(zram, blk_idx); - -release_init_lock: if (page) __free_page(page); + + return ret; +} + +#define PAGE_WRITEBACK 0 +#define HUGE_WRITEBACK (1 << 0) +#define IDLE_WRITEBACK (1 << 1) +#define INCOMPRESSIBLE_WRITEBACK (1 << 2) + +static int parse_page_index(char *val, unsigned long nr_pages, + unsigned long *lo, unsigned long *hi) +{ + int ret; + + ret = kstrtoul(val, 10, lo); + if (ret) + return ret; + if (*lo >= nr_pages) + return -ERANGE; + *hi = *lo + 1; + return 0; +} + +static int parse_page_indexes(char *val, unsigned long nr_pages, + unsigned long *lo, unsigned long *hi) +{ + char *delim; + int ret; + + delim = strchr(val, '-'); + if (!delim) + return -EINVAL; + + *delim = 0x00; + ret = kstrtoul(val, 10, lo); + if (ret) + return ret; + if (*lo >= nr_pages) + return -ERANGE; + + ret = kstrtoul(delim + 1, 10, hi); + if (ret) + return ret; + if (*hi >= nr_pages || *lo > *hi) + return -ERANGE; + *hi += 1; + return 0; +} + +static int parse_mode(char *val, u32 *mode) +{ + *mode = 0; + + if (!strcmp(val, "idle")) + *mode = IDLE_WRITEBACK; + if (!strcmp(val, "huge")) + *mode = HUGE_WRITEBACK; + if (!strcmp(val, "huge_idle")) + *mode = IDLE_WRITEBACK | HUGE_WRITEBACK; + if (!strcmp(val, "incompressible")) + *mode = INCOMPRESSIBLE_WRITEBACK; + + if (*mode == 0) + return -EINVAL; + return 0; +} + +static int scan_slots_for_writeback(struct zram *zram, u32 mode, + unsigned long lo, unsigned long hi, + struct zram_pp_ctl *ctl) +{ + u32 index = lo; + + while (index < hi) { + bool ok = true; + + zram_slot_lock(zram, index); + if (!zram_allocated(zram, index)) + goto next; + + if (zram_test_flag(zram, index, ZRAM_WB) || + zram_test_flag(zram, index, ZRAM_SAME)) + goto next; + + if (mode & IDLE_WRITEBACK && + !zram_test_flag(zram, index, ZRAM_IDLE)) + goto next; + if (mode & HUGE_WRITEBACK && + !zram_test_flag(zram, index, ZRAM_HUGE)) + goto next; + if (mode & INCOMPRESSIBLE_WRITEBACK && + !zram_test_flag(zram, index, ZRAM_INCOMPRESSIBLE)) + goto next; + + ok = place_pp_slot(zram, ctl, index); +next: + zram_slot_unlock(zram, index); + if (!ok) + break; + index++; + } + + return 0; +} + +static ssize_t writeback_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t len) +{ + struct zram *zram = dev_to_zram(dev); + u64 nr_pages = zram->disksize >> PAGE_SHIFT; + unsigned long lo = 0, hi = nr_pages; + struct zram_pp_ctl *ctl = NULL; + char *args, *param, *val; + ssize_t ret = len; + int err, mode = 0; + + down_read(&zram->init_lock); + if (!init_done(zram)) { + up_read(&zram->init_lock); + return -EINVAL; + } + + /* Do not permit concurrent post-processing actions. */ + if (atomic_xchg(&zram->pp_in_progress, 1)) { + up_read(&zram->init_lock); + return -EAGAIN; + } + + if (!zram->backing_dev) { + ret = -ENODEV; + goto release_init_lock; + } + + ctl = init_pp_ctl(); + if (!ctl) { + ret = -ENOMEM; + goto release_init_lock; + } + + args = skip_spaces(buf); + while (*args) { + args = next_arg(args, ¶m, &val); + + /* + * Workaround to support the old writeback interface. + * + * The old writeback interface has a minor inconsistency and + * requires key=value only for page_index parameter, while the + * writeback mode is a valueless parameter. + * + * This is not the case anymore and now all parameters are + * required to have values, however, we need to support the + * legacy writeback interface format so we check if we can + * recognize a valueless parameter as the (legacy) writeback + * mode. + */ + if (!val || !*val) { + err = parse_mode(param, &mode); + if (err) { + ret = err; + goto release_init_lock; + } + + scan_slots_for_writeback(zram, mode, lo, hi, ctl); + break; + } + + if (!strcmp(param, "type")) { + err = parse_mode(val, &mode); + if (err) { + ret = err; + goto release_init_lock; + } + + scan_slots_for_writeback(zram, mode, lo, hi, ctl); + break; + } + + if (!strcmp(param, "page_index")) { + err = parse_page_index(val, nr_pages, &lo, &hi); + if (err) { + ret = err; + goto release_init_lock; + } + + scan_slots_for_writeback(zram, mode, lo, hi, ctl); + continue; + } + + if (!strcmp(param, "page_indexes")) { + err = parse_page_indexes(val, nr_pages, &lo, &hi); + if (err) { + ret = err; + goto release_init_lock; + } + + scan_slots_for_writeback(zram, mode, lo, hi, ctl); + continue; + } + } + + err = zram_writeback_slots(zram, ctl); + if (err) + ret = err; + +release_init_lock: release_pp_ctl(zram, ctl); atomic_set(&zram->pp_in_progress, 0); up_read(&zram->init_lock); -- cgit v1.2.3