* [RFC PATCH 01/16] DM: dm-inplace-compress: an inplace compressed DM target
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 02/16] DM: Ability to choose the compressor Ram Pai
` (14 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc
Cc: snitzer, corbet, Shaohua Li, Ram Pai, shli, agk
From: Shaohua Li <shli@kernel.org>
This is a simple DM target supporting inplace compression. Its best suited for
SSD. The underlying disk must support 512B sector size, the target only
supports 4k sector size.
Disk layout:
|super|...meta...|..data...|
Store unit is 4k (a block). Super is 1 block, which stores meta and data
size and compression algorithm. Meta is a bitmap. For each data block,
there are 5 bits meta.
Data:
Data of a block is compressed. Compressed data is round up to 512B, which
is the payload. In disk, payload is stored at the beginning of logical
sector of the block. Let's look at an example. Say we store data to block
A, which is in sector B(A*8), its orginal size is 4k, compressed size is
1500. Compressed data (CD) will use 3 sectors (512B). The 3 sectors are the
payload. Payload will be stored at sector B.
---------------------------------------------------
... | CD1 | CD2 | CD3 | | | | | | ...
---------------------------------------------------
^B ^B+1 ^B+2 ^B+7 ^B+8
For this block, we will not use sector B+3 to B+7 (a hole). We use 4 meta
bits to present payload size. The compressed size (1500) isn't stored in
meta directly. Instead, we store it at the last 32bits of payload. In this
example, we store it at the end of sector B+2. If compressed size +
sizeof(32bits) crosses a sector, payload size will increase one sector. If
payload uses 8 sectors, we store uncompressed data directly.
If IO size is bigger than one block, we can store the data as an extent.
Data of the whole extent will compressed and stored in the similar way like
above. The first block of the extent is the head, all others are the tail.
If extent is 1 block, the block is head. We have 1 bit of meta to present
if a block is head or tail. If 4 meta bits of head block can't store extent
payload size, we will borrow tail block meta bits to store payload size.
Max allowd extent size is 128k, so we don't compress/decompress too big
size data.
Meta:
Modifying data will modify meta too. Meta will be written(flush) to disk
depending on meta write policy. We support writeback and writethrough mode.
In writeback mode, meta will be written to disk in an interval or a FLUSH
request. In writethrough mode, data and meta data will be written to disk
together.
Advantages:
1. Simple. Since we store compressed data in-place, we don't need complicated
disk data management.
2. Efficient. For each 4k, we only need 5 bits meta. 1T data will use less than
200M meta, so we can load all meta into memory. And actual compression size is
in payload. So if IO doesn't need RMW and we use writeback meta flush, we don't
need extra IO for meta.
Disadvantages:
1. hole. Since we store compressed data in-place, there are a lot of holes
(in above example, B+3 - B+7) Hole can impact IO, because we can't do IO
merge.
2. 1:1 size. Compression doesn't change disk size. If disk is 1T, we can
only store 1T data even we do compression.
But this is for SSD only. Generally SSD firmware has a FTL layer to map
disk sectors to flash nand. High end SSD firmware has filesystem-like FTL.
1. hole. Disk has a lot of holes, but SSD FTL can still store data continuous
in nand. Even if we can't do IO merge in OS layer, SSD firmware can do it.
2. 1:1 size. On one side, we write compressed data to SSD, which means less
data is written to SSD. This will be very helpful to improve SSD garbage
collection, and so write speed and life cycle. So even this is a problem, the
target is still helpful. On the other side, advanced SSD FTL can easily do thin
provision. For example, if nand is 1T and we let SSD report it as 2T, and use
the SSD as compressed target. In such SSD, we don't have the 1:1 size issue.
So even if SSD FTL cannot map non-continuous disk sectors to continuous nand,
the compression target can still function well.
Signed-off-by: Shaohua Li <shli@fusionio.com>
Signed-off-by: Ram Pai <ram.n.pai@gmail.com>
---
drivers/md/Kconfig | 6 +
drivers/md/Makefile | 1 +
drivers/md/dm-inplace-compress.c | 1487 ++++++++++++++++++++++++++++++++++++++
drivers/md/dm-inplace-compress.h | 140 ++++
4 files changed, 1634 insertions(+), 0 deletions(-)
create mode 100644 drivers/md/dm-inplace-compress.c
create mode 100644 drivers/md/dm-inplace-compress.h
diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig
index 02a5345..cdb1984 100644
--- a/drivers/md/Kconfig
+++ b/drivers/md/Kconfig
@@ -343,6 +343,12 @@ config DM_MIRROR
Allow volume managers to mirror logical volumes, also
needed for live data migration tools such as 'pvmove'.
+config DM_INPLACE_COMPRESS
+ tristate "Inplace Compression target"
+ depends on BLK_DEV_DM
+ ---help---
+ Allow volume managers to compress data for SSD.
+
config DM_LOG_USERSPACE
tristate "Mirror userspace logging"
depends on DM_MIRROR && NET
diff --git a/drivers/md/Makefile b/drivers/md/Makefile
index 52ba8dd..966eb2c 100644
--- a/drivers/md/Makefile
+++ b/drivers/md/Makefile
@@ -58,6 +58,7 @@ obj-$(CONFIG_DM_CACHE_SMQ) += dm-cache-smq.o
obj-$(CONFIG_DM_CACHE_CLEANER) += dm-cache-cleaner.o
obj-$(CONFIG_DM_ERA) += dm-era.o
obj-$(CONFIG_DM_LOG_WRITES) += dm-log-writes.o
+obj-$(CONFIG_DM_INPLACE_COMPRESS) += dm-inplace-compress.o
ifeq ($(CONFIG_DM_UEVENT),y)
dm-mod-objs += dm-uevent.o
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
new file mode 100644
index 0000000..c3c3750
--- /dev/null
+++ b/drivers/md/dm-inplace-compress.c
@@ -0,0 +1,1487 @@
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/blkdev.h>
+#include <linux/bio.h>
+#include <linux/slab.h>
+#include <linux/device-mapper.h>
+#include <linux/dm-io.h>
+#include <linux/crypto.h>
+#include <linux/lzo.h>
+#include <linux/kthread.h>
+#include <linux/page-flags.h>
+#include <linux/completion.h>
+#include <linux/vmalloc.h>
+#include "dm-inplace-compress.h"
+
+#define DM_MSG_PREFIX "dm-inplace-compress"
+
+static struct dm_icomp_compressor_data compressors[] = {
+ [DMCP_COMP_ALG_LZO] = {
+ .name = "lzo",
+ .comp_len = lzo_comp_len,
+ },
+};
+static int default_compressor;
+
+static struct kmem_cache *dm_icomp_req_cachep;
+static struct kmem_cache *dm_icomp_io_range_cachep;
+static struct kmem_cache *dm_icomp_meta_io_cachep;
+
+static struct dm_icomp_io_worker dm_icomp_io_workers[NR_CPUS];
+static struct workqueue_struct *dm_icomp_wq;
+
+static u8 dm_icomp_get_meta(struct dm_icomp_info *info, u64 block_index)
+{
+ u64 first_bit = block_index * DMCP_META_BITS;
+ int bits, offset;
+ u8 data, ret = 0;
+
+ offset = first_bit & 7;
+ bits = min_t(u8, DMCP_META_BITS, 8 - offset);
+
+ data = info->meta_bitmap[first_bit >> 3];
+ ret = (data >> offset) & ((1 << bits) - 1);
+
+ if (bits < DMCP_META_BITS) {
+ data = info->meta_bitmap[(first_bit >> 3) + 1];
+ bits = DMCP_META_BITS - bits;
+ ret |= (data & ((1 << bits) - 1)) << (DMCP_META_BITS - bits);
+ }
+ return ret;
+}
+
+static void dm_icomp_set_meta(struct dm_icomp_info *info, u64 block_index,
+ u8 meta, bool dirty_meta)
+{
+ u64 first_bit = block_index * DMCP_META_BITS;
+ int bits, offset;
+ u8 data;
+ struct page *page;
+
+ offset = first_bit & 7;
+ bits = min_t(u8, DMCP_META_BITS, 8 - offset);
+
+ data = info->meta_bitmap[first_bit >> 3];
+ data &= ~(((1 << bits) - 1) << offset);
+ data |= (meta & ((1 << bits) - 1)) << offset;
+ info->meta_bitmap[first_bit >> 3] = data;
+
+ if (info->write_mode == DMCP_WRITE_BACK) {
+ page = vmalloc_to_page(&info->meta_bitmap[first_bit >> 3]);
+ if (dirty_meta)
+ SetPageDirty(page);
+ else
+ ClearPageDirty(page);
+ }
+
+ if (bits < DMCP_META_BITS) {
+ meta >>= bits;
+ data = info->meta_bitmap[(first_bit >> 3) + 1];
+ bits = DMCP_META_BITS - bits;
+ data = (data >> bits) << bits;
+ data |= meta & ((1 << bits) - 1);
+ info->meta_bitmap[(first_bit >> 3) + 1] = data;
+
+ if (info->write_mode == DMCP_WRITE_BACK) {
+ page = vmalloc_to_page(&info->meta_bitmap[
+ (first_bit >> 3) + 1]);
+ if (dirty_meta)
+ SetPageDirty(page);
+ else
+ ClearPageDirty(page);
+ }
+ }
+}
+
+static void dm_icomp_set_extent(struct dm_icomp_req *req, u64 block,
+ u16 logical_blocks, sector_t data_sectors)
+{
+ int i;
+ u8 data;
+
+ for (i = 0; i < logical_blocks; i++) {
+ data = min_t(sector_t, data_sectors, 8);
+ data_sectors -= data;
+ if (i != 0)
+ data |= DMCP_TAIL_MASK;
+ /* For FUA, we write out meta data directly */
+ dm_icomp_set_meta(req->info, block + i, data,
+ !(req->bio->bi_rw & REQ_FUA));
+ }
+}
+
+static void dm_icomp_get_extent(struct dm_icomp_info *info, u64 block_index,
+ u64 *first_block_index, u16 *logical_sectors, u16 *data_sectors)
+{
+ u8 data;
+
+ data = dm_icomp_get_meta(info, block_index);
+ while (data & DMCP_TAIL_MASK) {
+ block_index--;
+ data = dm_icomp_get_meta(info, block_index);
+ }
+ *first_block_index = block_index;
+ *logical_sectors = DMCP_BLOCK_SIZE >> 9;
+ *data_sectors = data & DMCP_LENGTH_MASK;
+ block_index++;
+ while (block_index < info->data_blocks) {
+ data = dm_icomp_get_meta(info, block_index);
+ if (!(data & DMCP_TAIL_MASK))
+ break;
+ *logical_sectors += DMCP_BLOCK_SIZE >> 9;
+ *data_sectors += data & DMCP_LENGTH_MASK;
+ block_index++;
+ }
+}
+
+static int dm_icomp_access_super(struct dm_icomp_info *info, void *addr, int rw)
+{
+ struct dm_io_region region;
+ struct dm_io_request req;
+ unsigned long io_error = 0;
+ int ret;
+
+ region.bdev = info->dev->bdev;
+ region.sector = 0;
+ region.count = DMCP_BLOCK_SIZE >> 9;
+
+ req.bi_rw = rw;
+ req.mem.type = DM_IO_KMEM;
+ req.mem.offset = 0;
+ req.mem.ptr.addr = addr;
+ req.notify.fn = NULL;
+ req.client = info->io_client;
+
+ ret = dm_io(&req, 1, ®ion, &io_error);
+ if (ret || io_error)
+ return -EIO;
+ return 0;
+}
+
+static void dm_icomp_meta_io_done(unsigned long error, void *context)
+{
+ struct dm_icomp_meta_io *meta_io = context;
+
+ meta_io->fn(meta_io->data, error);
+ kmem_cache_free(dm_icomp_meta_io_cachep, meta_io);
+}
+
+static int dm_icomp_write_meta(struct dm_icomp_info *info, u64 start_page,
+ u64 end_page, void *data,
+ void (*fn)(void *data, unsigned long error), int rw)
+{
+ struct dm_icomp_meta_io *meta_io;
+
+ WARN_ON(end_page > info->meta_bitmap_pages);
+
+ meta_io = kmem_cache_alloc(dm_icomp_meta_io_cachep, GFP_NOIO);
+ if (!meta_io) {
+ fn(data, -ENOMEM);
+ return -ENOMEM;
+ }
+ meta_io->data = data;
+ meta_io->fn = fn;
+
+ meta_io->io_region.bdev = info->dev->bdev;
+ meta_io->io_region.sector = DMCP_META_START_SECTOR +
+ (start_page << (PAGE_SHIFT - 9));
+ meta_io->io_region.count = (end_page - start_page) << (PAGE_SHIFT - 9);
+
+ atomic64_add(meta_io->io_region.count << 9, &info->meta_write_size);
+
+ meta_io->io_req.bi_rw = rw;
+ meta_io->io_req.mem.type = DM_IO_VMA;
+ meta_io->io_req.mem.offset = 0;
+ meta_io->io_req.mem.ptr.addr = info->meta_bitmap +
+ (start_page << PAGE_SHIFT);
+ meta_io->io_req.notify.fn = dm_icomp_meta_io_done;
+ meta_io->io_req.notify.context = meta_io;
+ meta_io->io_req.client = info->io_client;
+
+ dm_io(&meta_io->io_req, 1, &meta_io->io_region, NULL);
+ return 0;
+}
+
+struct writeback_flush_data {
+ struct completion complete;
+ atomic_t cnt;
+};
+
+static void writeback_flush_io_done(void *data, unsigned long error)
+{
+ struct writeback_flush_data *wb = data;
+
+ if (atomic_dec_return(&wb->cnt))
+ return;
+ complete(&wb->complete);
+}
+
+static void dm_icomp_flush_dirty_meta(struct dm_icomp_info *info,
+ struct writeback_flush_data *data)
+{
+ struct page *page;
+ u64 start = 0, index;
+ u32 pending = 0, cnt = 0;
+ bool dirty;
+ struct blk_plug plug;
+
+ blk_start_plug(&plug);
+ for (index = 0; index < info->meta_bitmap_pages; index++, cnt++) {
+ if (cnt == 256) {
+ cnt = 0;
+ cond_resched();
+ }
+
+ page = vmalloc_to_page(info->meta_bitmap +
+ (index << PAGE_SHIFT));
+ dirty = TestClearPageDirty(page);
+
+ if (pending == 0 && dirty) {
+ start = index;
+ pending++;
+ continue;
+ } else if (pending == 0)
+ continue;
+ else if (pending > 0 && dirty) {
+ pending++;
+ continue;
+ }
+
+ /* pending > 0 && !dirty */
+ atomic_inc(&data->cnt);
+ dm_icomp_write_meta(info, start, start + pending, data,
+ writeback_flush_io_done, WRITE);
+ pending = 0;
+ }
+
+ if (pending > 0) {
+ atomic_inc(&data->cnt);
+ dm_icomp_write_meta(info, start, start + pending, data,
+ writeback_flush_io_done, WRITE);
+ }
+ blkdev_issue_flush(info->dev->bdev, GFP_NOIO, NULL);
+ blk_finish_plug(&plug);
+}
+
+static int dm_icomp_meta_writeback_thread(void *data)
+{
+ struct dm_icomp_info *info = data;
+ struct writeback_flush_data wb;
+
+ atomic_set(&wb.cnt, 1);
+ init_completion(&wb.complete);
+
+ while (!kthread_should_stop()) {
+ schedule_timeout_interruptible(
+ msecs_to_jiffies(info->writeback_delay * 1000));
+ dm_icomp_flush_dirty_meta(info, &wb);
+ }
+
+ dm_icomp_flush_dirty_meta(info, &wb);
+
+ writeback_flush_io_done(&wb, 0);
+ wait_for_completion(&wb.complete);
+ return 0;
+}
+
+static int dm_icomp_init_meta(struct dm_icomp_info *info, bool new)
+{
+ struct dm_io_region region;
+ struct dm_io_request req;
+ unsigned long io_error = 0;
+ struct blk_plug plug;
+ int ret;
+ ssize_t len = DIV_ROUND_UP_ULL(info->meta_bitmap_bits, BITS_PER_LONG);
+
+ len *= sizeof(unsigned long);
+
+ region.bdev = info->dev->bdev;
+ region.sector = DMCP_META_START_SECTOR;
+ region.count = (len + 511) >> 9;
+
+ req.mem.type = DM_IO_VMA;
+ req.mem.offset = 0;
+ req.mem.ptr.addr = info->meta_bitmap;
+ req.notify.fn = NULL;
+ req.client = info->io_client;
+
+ blk_start_plug(&plug);
+ if (new) {
+ memset(info->meta_bitmap, 0, len);
+ req.bi_rw = WRITE_FLUSH;
+ ret = dm_io(&req, 1, ®ion, &io_error);
+ } else {
+ req.bi_rw = READ;
+ ret = dm_io(&req, 1, ®ion, &io_error);
+ }
+ blk_finish_plug(&plug);
+
+ if (ret || io_error) {
+ info->ti->error = "Access metadata error";
+ return -EIO;
+ }
+
+ if (info->write_mode == DMCP_WRITE_BACK) {
+ info->writeback_tsk = kthread_run(
+ dm_icomp_meta_writeback_thread,
+ info, "dm_icomp_writeback");
+ if (!info->writeback_tsk) {
+ info->ti->error = "Create writeback thread error";
+ return -EINVAL;
+ }
+ }
+
+ return 0;
+}
+
+static int dm_icomp_alloc_compressor(struct dm_icomp_info *info)
+{
+ int i;
+
+ for_each_possible_cpu(i) {
+ info->tfm[i] = crypto_alloc_comp(
+ compressors[info->comp_alg].name, 0, 0);
+ if (IS_ERR(info->tfm[i])) {
+ info->tfm[i] = NULL;
+ goto err;
+ }
+ }
+ return 0;
+err:
+ for_each_possible_cpu(i) {
+ if (info->tfm[i]) {
+ crypto_free_comp(info->tfm[i]);
+ info->tfm[i] = NULL;
+ }
+ }
+ return -ENOMEM;
+}
+
+static void dm_icomp_free_compressor(struct dm_icomp_info *info)
+{
+ int i;
+
+ for_each_possible_cpu(i) {
+ if (info->tfm[i]) {
+ crypto_free_comp(info->tfm[i]);
+ info->tfm[i] = NULL;
+ }
+ }
+}
+
+static int dm_icomp_read_or_create_super(struct dm_icomp_info *info)
+{
+ void *addr;
+ struct dm_icomp_super_block *super;
+ u64 total_blocks;
+ u64 data_blocks, meta_blocks;
+ u32 rem, cnt;
+ bool new_super = false;
+ int ret;
+ ssize_t len;
+
+ total_blocks = i_size_read(info->dev->bdev->bd_inode) >>
+ DMCP_BLOCK_SHIFT;
+ data_blocks = total_blocks - 1;
+ rem = do_div(data_blocks, DMCP_BLOCK_SIZE * 8 + DMCP_META_BITS);
+ meta_blocks = data_blocks * DMCP_META_BITS;
+ data_blocks *= DMCP_BLOCK_SIZE * 8;
+
+ cnt = rem;
+ rem /= (DMCP_BLOCK_SIZE * 8 / DMCP_META_BITS + 1);
+ data_blocks += rem * (DMCP_BLOCK_SIZE * 8 / DMCP_META_BITS);
+ meta_blocks += rem;
+
+ cnt %= (DMCP_BLOCK_SIZE * 8 / DMCP_META_BITS + 1);
+ meta_blocks += 1;
+ data_blocks += cnt - 1;
+
+ info->data_blocks = data_blocks;
+ info->data_start = (1 + meta_blocks) << DMCP_BLOCK_SECTOR_SHIFT;
+
+ addr = kzalloc(DMCP_BLOCK_SIZE, GFP_KERNEL);
+ if (!addr) {
+ info->ti->error = "Cannot allocate super";
+ return -ENOMEM;
+ }
+
+ super = addr;
+ ret = dm_icomp_access_super(info, addr, READ);
+ if (ret)
+ goto out;
+
+ if (le64_to_cpu(super->magic) == DMCP_SUPER_MAGIC) {
+ if (le64_to_cpu(super->meta_blocks) != meta_blocks ||
+ le64_to_cpu(super->data_blocks) != data_blocks) {
+ info->ti->error = "Super is invalid";
+ ret = -EINVAL;
+ goto out;
+ }
+ if (!crypto_has_comp(compressors[super->comp_alg].name, 0, 0)) {
+ info->ti->error =
+ "Compressor algorithm doesn't support";
+ ret = -EINVAL;
+ goto out;
+ }
+ } else {
+ super->magic = cpu_to_le64(DMCP_SUPER_MAGIC);
+ super->meta_blocks = cpu_to_le64(meta_blocks);
+ super->data_blocks = cpu_to_le64(data_blocks);
+ super->comp_alg = default_compressor;
+ ret = dm_icomp_access_super(info, addr, WRITE_FUA);
+ if (ret) {
+ info->ti->error = "Access super fails";
+ goto out;
+ }
+ new_super = true;
+ }
+
+ info->comp_alg = super->comp_alg;
+ if (dm_icomp_alloc_compressor(info)) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ info->meta_bitmap_bits = data_blocks * DMCP_META_BITS;
+ len = DIV_ROUND_UP_ULL(info->meta_bitmap_bits, BITS_PER_LONG);
+ len *= sizeof(unsigned long);
+ info->meta_bitmap_pages = (len + PAGE_SIZE - 1) >> PAGE_SHIFT;
+ info->meta_bitmap = vmalloc(info->meta_bitmap_pages * PAGE_SIZE);
+ if (!info->meta_bitmap) {
+ ret = -ENOMEM;
+ goto bitmap_err;
+ }
+
+ ret = dm_icomp_init_meta(info, new_super);
+ if (ret)
+ goto meta_err;
+
+ return 0;
+meta_err:
+ vfree(info->meta_bitmap);
+bitmap_err:
+ dm_icomp_free_compressor(info);
+out:
+ kfree(addr);
+ return ret;
+}
+
+/*
+ * <dev> <writethough>/<writeback> <meta_commit_delay>
+ */
+static int dm_icomp_ctr(struct dm_target *ti, unsigned int argc, char **argv)
+{
+ struct dm_icomp_info *info;
+ char write_mode[15];
+ int ret, i;
+
+ if (argc < 2) {
+ ti->error = "Invalid argument count";
+ return -EINVAL;
+ }
+
+ info = kzalloc(sizeof(*info), GFP_KERNEL);
+ if (!info) {
+ ti->error = "dm-inplace-compress: Cannot allocate context";
+ return -ENOMEM;
+ }
+ info->ti = ti;
+
+ if (sscanf(argv[1], "%s", write_mode) != 1) {
+ ti->error = "Invalid argument";
+ ret = -EINVAL;
+ goto err_para;
+ }
+
+ if (strcmp(write_mode, "writeback") == 0) {
+ if (argc != 3) {
+ ti->error = "Invalid argument";
+ ret = -EINVAL;
+ goto err_para;
+ }
+ info->write_mode = DMCP_WRITE_BACK;
+ if (sscanf(argv[2], "%u", &info->writeback_delay) != 1) {
+ ti->error = "Invalid argument";
+ ret = -EINVAL;
+ goto err_para;
+ }
+ } else if (strcmp(write_mode, "writethrough") == 0) {
+ info->write_mode = DMCP_WRITE_THROUGH;
+ } else {
+ ti->error = "Invalid argument";
+ ret = -EINVAL;
+ goto err_para;
+ }
+
+ if (dm_get_device(ti, argv[0], dm_table_get_mode(ti->table),
+ &info->dev)) {
+ ti->error = "Can't get device";
+ ret = -EINVAL;
+ goto err_para;
+ }
+
+ info->io_client = dm_io_client_create();
+ if (!info->io_client) {
+ ti->error = "Can't create io client";
+ ret = -EINVAL;
+ goto err_ioclient;
+ }
+
+ if (bdev_logical_block_size(info->dev->bdev) != 512) {
+ ti->error = "Can't logical block size too big";
+ ret = -EINVAL;
+ goto err_blocksize;
+ }
+
+ ret = dm_icomp_read_or_create_super(info);
+ if (ret)
+ goto err_blocksize;
+
+ for (i = 0; i < BITMAP_HASH_LEN; i++) {
+ info->bitmap_locks[i].io_running = 0;
+ spin_lock_init(&info->bitmap_locks[i].wait_lock);
+ INIT_LIST_HEAD(&info->bitmap_locks[i].wait_list);
+ }
+
+ atomic64_set(&info->compressed_write_size, 0);
+ atomic64_set(&info->uncompressed_write_size, 0);
+ atomic64_set(&info->meta_write_size, 0);
+ ti->num_flush_bios = 1;
+ /* ti->num_discard_bios = 1; */
+ ti->private = info;
+ return 0;
+err_blocksize:
+ dm_io_client_destroy(info->io_client);
+err_ioclient:
+ dm_put_device(ti, info->dev);
+err_para:
+ kfree(info);
+ return ret;
+}
+
+static void dm_icomp_dtr(struct dm_target *ti)
+{
+ struct dm_icomp_info *info = ti->private;
+
+ if (info->write_mode == DMCP_WRITE_BACK)
+ kthread_stop(info->writeback_tsk);
+ dm_icomp_free_compressor(info);
+ vfree(info->meta_bitmap);
+ dm_io_client_destroy(info->io_client);
+ dm_put_device(ti, info->dev);
+ kfree(info);
+}
+
+static u64 dm_icomp_sector_to_block(sector_t sect)
+{
+ return sect >> DMCP_BLOCK_SECTOR_SHIFT;
+}
+
+static struct dm_icomp_hash_lock *dm_icomp_block_hash_lock(
+ struct dm_icomp_info *info, u64 block_index)
+{
+ return &info->bitmap_locks[(block_index >> BITMAP_HASH_SHIFT) &
+ BITMAP_HASH_MASK];
+}
+
+static struct dm_icomp_hash_lock *dm_icomp_trylock_block(
+ struct dm_icomp_info *info,
+ struct dm_icomp_req *req, u64 block_index)
+{
+ struct dm_icomp_hash_lock *hash_lock;
+
+ hash_lock = dm_icomp_block_hash_lock(req->info, block_index);
+
+ spin_lock_irq(&hash_lock->wait_lock);
+ if (!hash_lock->io_running) {
+ hash_lock->io_running = 1;
+ spin_unlock_irq(&hash_lock->wait_lock);
+ return hash_lock;
+ }
+ list_add_tail(&req->sibling, &hash_lock->wait_list);
+ spin_unlock_irq(&hash_lock->wait_lock);
+ return NULL;
+}
+
+static void dm_icomp_queue_req_list(struct dm_icomp_info *info,
+ struct list_head *list);
+
+static void dm_icomp_unlock_block(struct dm_icomp_info *info,
+ struct dm_icomp_req *req, struct dm_icomp_hash_lock *hash_lock)
+{
+ LIST_HEAD(pending_list);
+ unsigned long flags;
+
+ spin_lock_irqsave(&hash_lock->wait_lock, flags);
+ /* wakeup all pending reqs to avoid live lock */
+ list_splice_init(&hash_lock->wait_list, &pending_list);
+ hash_lock->io_running = 0;
+ spin_unlock_irqrestore(&hash_lock->wait_lock, flags);
+
+ dm_icomp_queue_req_list(info, &pending_list);
+}
+
+static int dm_icomp_lock_req_range(struct dm_icomp_req *req)
+{
+ u64 block_index, first_block_index;
+ u64 first_lock_block, second_lock_block;
+ u16 logical_sectors, data_sectors;
+
+ block_index = dm_icomp_sector_to_block(req->bio->bi_iter.bi_sector);
+ req->locks[0] = dm_icomp_trylock_block(req->info, req, block_index);
+ if (!req->locks[0])
+ return 0;
+ dm_icomp_get_extent(req->info, block_index, &first_block_index,
+ &logical_sectors, &data_sectors);
+ if (dm_icomp_block_hash_lock(req->info, first_block_index) !=
+ req->locks[0]) {
+ dm_icomp_unlock_block(req->info, req, req->locks[0]);
+
+ first_lock_block = first_block_index;
+ second_lock_block = block_index;
+ goto two_locks;
+ }
+
+ block_index = dm_icomp_sector_to_block(bio_end_sector(req->bio) - 1);
+ dm_icomp_get_extent(req->info, block_index, &first_block_index,
+ &logical_sectors, &data_sectors);
+ first_block_index += dm_icomp_sector_to_block(logical_sectors) - 1;
+ if (dm_icomp_block_hash_lock(req->info, first_block_index) !=
+ req->locks[0]) {
+ second_lock_block = first_block_index;
+ goto second_lock;
+ }
+ req->locked_locks = 1;
+ return 1;
+
+two_locks:
+ req->locks[0] = dm_icomp_trylock_block(req->info, req,
+ first_lock_block);
+ if (!req->locks[0])
+ return 0;
+second_lock:
+ req->locks[1] = dm_icomp_trylock_block(req->info, req,
+ second_lock_block);
+ if (!req->locks[1]) {
+ dm_icomp_unlock_block(req->info, req, req->locks[0]);
+ return 0;
+ }
+ /* Don't need check if meta is changed */
+ req->locked_locks = 2;
+ return 1;
+}
+
+static void dm_icomp_unlock_req_range(struct dm_icomp_req *req)
+{
+ int i;
+
+ for (i = req->locked_locks - 1; i >= 0; i--)
+ dm_icomp_unlock_block(req->info, req, req->locks[i]);
+}
+
+static void dm_icomp_queue_req(struct dm_icomp_info *info,
+ struct dm_icomp_req *req)
+{
+ unsigned long flags;
+ struct dm_icomp_io_worker *worker = &dm_icomp_io_workers[req->cpu];
+
+ spin_lock_irqsave(&worker->lock, flags);
+ list_add_tail(&req->sibling, &worker->pending);
+ spin_unlock_irqrestore(&worker->lock, flags);
+
+ queue_work_on(req->cpu, dm_icomp_wq, &worker->work);
+}
+
+static void dm_icomp_queue_req_list(struct dm_icomp_info *info,
+ struct list_head *list)
+{
+ struct dm_icomp_req *req;
+
+ while (!list_empty(list)) {
+ req = list_first_entry(list, struct dm_icomp_req, sibling);
+ list_del_init(&req->sibling);
+ dm_icomp_queue_req(info, req);
+ }
+}
+
+static void dm_icomp_get_req(struct dm_icomp_req *req)
+{
+ atomic_inc(&req->io_pending);
+}
+
+static void dm_icomp_free_io_range(struct dm_icomp_io_range *io)
+{
+ kfree(io->decomp_data);
+ kfree(io->comp_data);
+ kmem_cache_free(dm_icomp_io_range_cachep, io);
+}
+
+static void dm_icomp_put_req(struct dm_icomp_req *req)
+{
+ struct dm_icomp_io_range *io;
+
+ if (atomic_dec_return(&req->io_pending))
+ return;
+
+ if (req->stage == STAGE_INIT) /* waiting for locking */
+ return;
+
+ if (req->stage == STAGE_READ_DECOMP ||
+ req->stage == STAGE_WRITE_COMP ||
+ req->result)
+ req->stage = STAGE_DONE;
+
+ if (req->stage != STAGE_DONE) {
+ dm_icomp_queue_req(req->info, req);
+ return;
+ }
+
+ while (!list_empty(&req->all_io)) {
+ io = list_entry(req->all_io.next,
+ struct dm_icomp_io_range, next);
+ list_del(&io->next);
+ dm_icomp_free_io_range(io);
+ }
+
+ dm_icomp_unlock_req_range(req);
+
+ req->bio->bi_error = req->result;
+ bio_endio(req->bio);
+ kmem_cache_free(dm_icomp_req_cachep, req);
+}
+
+static void dm_icomp_io_range_done(unsigned long error, void *context)
+{
+ struct dm_icomp_io_range *io = context;
+
+ if (error)
+ io->req->result = error;
+ dm_icomp_put_req(io->req);
+}
+
+static inline int dm_icomp_compressor_len(struct dm_icomp_info *info, int len)
+{
+ if (compressors[info->comp_alg].comp_len)
+ return compressors[info->comp_alg].comp_len(len);
+ return len;
+}
+
+/*
+ * caller should set region.sector, region.count. bi_rw. IO always to/from
+ * comp_data
+ */
+static struct dm_icomp_io_range *dm_icomp_create_io_range(
+ struct dm_icomp_req *req, int comp_len, int decomp_len)
+{
+ struct dm_icomp_io_range *io;
+
+ io = kmem_cache_alloc(dm_icomp_io_range_cachep, GFP_NOIO);
+ if (!io)
+ return NULL;
+
+ io->comp_data = kmalloc(dm_icomp_compressor_len(req->info, comp_len),
+ GFP_NOIO);
+ io->decomp_data = kmalloc(decomp_len, GFP_NOIO);
+ if (!io->decomp_data || !io->comp_data) {
+ kfree(io->decomp_data);
+ kfree(io->comp_data);
+ kmem_cache_free(dm_icomp_io_range_cachep, io);
+ return NULL;
+ }
+
+ io->io_req.notify.fn = dm_icomp_io_range_done;
+ io->io_req.notify.context = io;
+ io->io_req.client = req->info->io_client;
+ io->io_req.mem.type = DM_IO_KMEM;
+ io->io_req.mem.ptr.addr = io->comp_data;
+ io->io_req.mem.offset = 0;
+
+ io->io_region.bdev = req->info->dev->bdev;
+
+ io->decomp_len = decomp_len;
+ io->comp_len = comp_len;
+ io->req = req;
+ return io;
+}
+
+static void dm_icomp_bio_copy(struct bio *bio, off_t bio_off, void *buf,
+ ssize_t len, bool to_buf)
+{
+ struct bio_vec bv;
+ struct bvec_iter iter;
+ off_t buf_off = 0;
+ ssize_t size;
+ void *addr;
+
+ WARN_ON(bio_off + len > (bio_sectors(bio) << 9));
+
+ bio_for_each_segment(bv, bio, iter) {
+ int length = bv.bv_len;
+
+ if (bio_off >= length) {
+ bio_off -= length;
+ continue;
+ }
+ addr = kmap_atomic(bv.bv_page);
+ size = min_t(ssize_t, len, length - bio_off);
+ if (to_buf)
+ memcpy(buf + buf_off, addr + bio_off + bv.bv_offset,
+ size);
+ else
+ memcpy(addr + bio_off + bv.bv_offset, buf + buf_off,
+ size);
+ kunmap_atomic(addr);
+ bio_off = 0;
+ buf_off += size;
+ len -= size;
+ }
+}
+
+/*
+ * return value:
+ * < 0 : error
+ * == 0 : ok
+ * == 1 : ok, but comp/decomp is skipped
+ * Compressed data size is roundup of 512, which makes the payload.
+ * We store the actual compressed len in the last u32 of the payload.
+ * If there is no free space, we add 512 to the payload size.
+ */
+static int dm_icomp_io_range_comp(struct dm_icomp_info *info, void *comp_data,
+ unsigned int *comp_len, void *decomp_data, unsigned int decomp_len,
+ bool do_comp)
+{
+ struct crypto_comp *tfm;
+ u32 *addr;
+ unsigned int actual_comp_len;
+ int ret;
+
+ if (do_comp) {
+ actual_comp_len = *comp_len;
+
+ tfm = info->tfm[get_cpu()];
+ ret = crypto_comp_compress(tfm, decomp_data, decomp_len,
+ comp_data, &actual_comp_len);
+ put_cpu();
+
+ atomic64_add(decomp_len, &info->uncompressed_write_size);
+ if (ret || decomp_len < actual_comp_len + sizeof(u32) + 512) {
+ *comp_len = decomp_len;
+ atomic64_add(*comp_len, &info->compressed_write_size);
+ return 1;
+ }
+
+ *comp_len = round_up(actual_comp_len, 512);
+ if (*comp_len - actual_comp_len < sizeof(u32))
+ *comp_len += 512;
+ atomic64_add(*comp_len, &info->compressed_write_size);
+ addr = comp_data + *comp_len;
+ addr--;
+ *addr = cpu_to_le32(actual_comp_len);
+ } else {
+ if (*comp_len == decomp_len)
+ return 1;
+ addr = comp_data + *comp_len;
+ addr--;
+ actual_comp_len = le32_to_cpu(*addr);
+
+ tfm = info->tfm[get_cpu()];
+ ret = crypto_comp_decompress(tfm, comp_data, actual_comp_len,
+ decomp_data, &decomp_len);
+ put_cpu();
+ if (ret)
+ return -EINVAL;
+ }
+ return 0;
+}
+
+static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
+{
+ struct dm_icomp_io_range *io;
+ off_t bio_off = 0;
+ int ret;
+
+ req->stage = STAGE_READ_DECOMP;
+
+ if (req->result)
+ return;
+
+ list_for_each_entry(io, &req->all_io, next) {
+ ssize_t dst_off = 0, src_off = 0, len;
+
+ io->io_region.sector -= req->info->data_start;
+
+ /* Do decomp here */
+ ret = dm_icomp_io_range_comp(req->info, io->comp_data,
+ &io->comp_len, io->decomp_data, io->decomp_len, false);
+ if (ret < 0) {
+ req->result = -EIO;
+ return;
+ }
+
+ if (io->io_region.sector >= req->bio->bi_iter.bi_sector)
+ dst_off = (io->io_region.sector -
+ req->bio->bi_iter.bi_sector) << 9;
+ else
+ src_off = (req->bio->bi_iter.bi_sector -
+ io->io_region.sector) << 9;
+
+ len = min_t(ssize_t, io->decomp_len - src_off,
+ (bio_sectors(req->bio) << 9) - dst_off);
+
+ /* io range in all_io list is ordered for read IO */
+ while (bio_off != dst_off) {
+ ssize_t size = min_t(ssize_t, PAGE_SIZE,
+ dst_off - bio_off);
+ dm_icomp_bio_copy(req->bio, bio_off, empty_zero_page,
+ size, false);
+ bio_off += size;
+ }
+
+ if (ret == 1)
+ dm_icomp_bio_copy(req->bio, dst_off,
+ io->comp_data + src_off, len, false);
+ else
+ dm_icomp_bio_copy(req->bio, dst_off,
+ io->decomp_data + src_off, len, false);
+ bio_off = dst_off + len;
+ }
+
+ while (bio_off != (bio_sectors(req->bio) << 9)) {
+ ssize_t size = min_t(ssize_t, PAGE_SIZE,
+ (bio_sectors(req->bio) << 9) - bio_off);
+ dm_icomp_bio_copy(req->bio, bio_off, empty_zero_page,
+ size, false);
+ bio_off += size;
+ }
+}
+
+static void dm_icomp_read_one_extent(struct dm_icomp_req *req, u64 block,
+ u16 logical_sectors, u16 data_sectors)
+{
+ struct dm_icomp_io_range *io;
+
+ io = dm_icomp_create_io_range(req, data_sectors << 9,
+ logical_sectors << 9);
+ if (!io) {
+ req->result = -EIO;
+ return;
+ }
+
+ dm_icomp_get_req(req);
+ list_add_tail(&io->next, &req->all_io);
+
+ io->io_region.sector = (block << DMCP_BLOCK_SECTOR_SHIFT) +
+ req->info->data_start;
+ io->io_region.count = data_sectors;
+
+ io->io_req.bi_rw = READ;
+ dm_io(&io->io_req, 1, &io->io_region, NULL);
+}
+
+static void dm_icomp_handle_read_read_existing(struct dm_icomp_req *req)
+{
+ u64 block_index, first_block_index;
+ u16 logical_sectors, data_sectors;
+
+ req->stage = STAGE_READ_EXISTING;
+
+ block_index = dm_icomp_sector_to_block(req->bio->bi_iter.bi_sector);
+again:
+ dm_icomp_get_extent(req->info, block_index, &first_block_index,
+ &logical_sectors, &data_sectors);
+ if (data_sectors > 0)
+ dm_icomp_read_one_extent(req, first_block_index,
+ logical_sectors, data_sectors);
+
+ if (req->result)
+ return;
+
+ block_index = first_block_index + (logical_sectors >>
+ DMCP_BLOCK_SECTOR_SHIFT);
+ if ((block_index << DMCP_BLOCK_SECTOR_SHIFT) < bio_end_sector(req->bio))
+ goto again;
+
+ /* A shortcut if all data is in already */
+ if (list_empty(&req->all_io))
+ dm_icomp_handle_read_decomp(req);
+}
+
+static void dm_icomp_handle_read_request(struct dm_icomp_req *req)
+{
+ dm_icomp_get_req(req);
+
+ if (req->stage == STAGE_INIT) {
+ if (!dm_icomp_lock_req_range(req)) {
+ dm_icomp_put_req(req);
+ return;
+ }
+
+ dm_icomp_handle_read_read_existing(req);
+ } else if (req->stage == STAGE_READ_EXISTING)
+ dm_icomp_handle_read_decomp(req);
+
+ dm_icomp_put_req(req);
+}
+
+static void dm_icomp_write_meta_done(void *context, unsigned long error)
+{
+ struct dm_icomp_req *req = context;
+
+ dm_icomp_put_req(req);
+}
+
+static u64 dm_icomp_block_meta_page_index(u64 block, bool end)
+{
+ u64 bits = block * DMCP_META_BITS - !!end;
+ /* (1 << 3) bits per byte */
+ return bits >> (3 + PAGE_SHIFT);
+}
+
+static int dm_icomp_handle_write_modify(struct dm_icomp_io_range *io,
+ u64 *meta_start, u64 *meta_end, bool *handle_bio)
+{
+ struct dm_icomp_req *req = io->req;
+ sector_t start, count;
+ unsigned int comp_len;
+ off_t offset;
+ u64 page_index;
+ int ret;
+
+ io->io_region.sector -= req->info->data_start;
+
+ /* decompress original data */
+ ret = dm_icomp_io_range_comp(req->info, io->comp_data, &io->comp_len,
+ io->decomp_data, io->decomp_len, false);
+ if (ret < 0) {
+ req->result = -EINVAL;
+ return -EIO;
+ }
+
+ start = io->io_region.sector;
+ count = io->decomp_len >> 9;
+ if (start < req->bio->bi_iter.bi_sector && start + count >
+ bio_end_sector(req->bio)) {
+ /* we don't split an extent */
+ if (ret == 1) {
+ memcpy(io->decomp_data, io->comp_data, io->decomp_len);
+ dm_icomp_bio_copy(req->bio, 0,
+ io->decomp_data +
+ ((req->bio->bi_iter.bi_sector - start) << 9),
+ bio_sectors(req->bio) << 9, true);
+ } else {
+ dm_icomp_bio_copy(req->bio, 0,
+ io->decomp_data +
+ ((req->bio->bi_iter.bi_sector - start) << 9),
+ bio_sectors(req->bio) << 9, true);
+
+ kfree(io->comp_data);
+ /* New compressed len might be bigger */
+ io->comp_data = kmalloc(
+ dm_icomp_compressor_len(req->info,
+ io->decomp_len), GFP_NOIO);
+ io->comp_len = io->decomp_len;
+ if (!io->comp_data) {
+ req->result = -ENOMEM;
+ return -EIO;
+ }
+ io->io_req.mem.ptr.addr = io->comp_data;
+ }
+ /* need compress data */
+ ret = 0;
+ offset = 0;
+ *handle_bio = false;
+ } else if (start < req->bio->bi_iter.bi_sector) {
+ count = req->bio->bi_iter.bi_sector - start;
+ offset = 0;
+ } else {
+ offset = bio_end_sector(req->bio) - start;
+ start = bio_end_sector(req->bio);
+ count = count - offset;
+ }
+
+ /* Original data is uncompressed, we don't need writeback */
+ if (ret == 1) {
+ comp_len = count << 9;
+ goto handle_meta;
+ }
+
+ /* assume compress less data uses less space (at least 4k lsess data) */
+ comp_len = io->comp_len;
+ ret = dm_icomp_io_range_comp(req->info, io->comp_data, &comp_len,
+ io->decomp_data + (offset << 9), count << 9, true);
+ if (ret < 0) {
+ req->result = -EIO;
+ return -EIO;
+ }
+
+ dm_icomp_get_req(req);
+ if (ret == 1)
+ io->io_req.mem.ptr.addr = io->decomp_data + (offset << 9);
+ io->io_region.count = comp_len >> 9;
+ io->io_region.sector = start + req->info->data_start;
+
+ io->io_req.bi_rw = req->bio->bi_rw;
+ dm_io(&io->io_req, 1, &io->io_region, NULL);
+handle_meta:
+ dm_icomp_set_extent(req, start >> DMCP_BLOCK_SECTOR_SHIFT,
+ count >> DMCP_BLOCK_SECTOR_SHIFT, comp_len >> 9);
+
+ page_index = dm_icomp_block_meta_page_index(start >>
+ DMCP_BLOCK_SECTOR_SHIFT, false);
+ if (*meta_start > page_index)
+ *meta_start = page_index;
+ page_index = dm_icomp_block_meta_page_index(
+ (start + count) >> DMCP_BLOCK_SECTOR_SHIFT, true);
+ if (*meta_end < page_index)
+ *meta_end = page_index;
+ return 0;
+}
+
+static void dm_icomp_handle_write_comp(struct dm_icomp_req *req)
+{
+ struct dm_icomp_io_range *io;
+ sector_t count;
+ unsigned int comp_len;
+ u64 meta_start = -1L, meta_end = 0, page_index;
+ int ret;
+ bool handle_bio = true;
+
+ req->stage = STAGE_WRITE_COMP;
+
+ if (req->result)
+ return;
+
+ list_for_each_entry(io, &req->all_io, next) {
+ if (dm_icomp_handle_write_modify(io, &meta_start, &meta_end,
+ &handle_bio))
+ return;
+ }
+
+ if (!handle_bio)
+ goto update_meta;
+
+ count = bio_sectors(req->bio);
+ io = dm_icomp_create_io_range(req, count << 9, count << 9);
+ if (!io) {
+ req->result = -EIO;
+ return;
+ }
+ dm_icomp_bio_copy(req->bio, 0, io->decomp_data, count << 9, true);
+
+ /* compress data */
+ comp_len = io->comp_len;
+ ret = dm_icomp_io_range_comp(req->info, io->comp_data, &comp_len,
+ io->decomp_data, count << 9, true);
+ if (ret < 0) {
+ dm_icomp_free_io_range(io);
+ req->result = -EIO;
+ return;
+ }
+
+ dm_icomp_get_req(req);
+ list_add_tail(&io->next, &req->all_io);
+ io->io_region.sector = req->bio->bi_iter.bi_sector +
+ req->info->data_start;
+
+ if (ret == 1)
+ io->io_req.mem.ptr.addr = io->decomp_data;
+
+ io->io_region.count = comp_len >> 9;
+ io->io_req.bi_rw = req->bio->bi_rw;
+ dm_io(&io->io_req, 1, &io->io_region, NULL);
+ dm_icomp_set_extent(req,
+ req->bio->bi_iter.bi_sector >> DMCP_BLOCK_SECTOR_SHIFT,
+ count >> DMCP_BLOCK_SECTOR_SHIFT, comp_len >> 9);
+
+ page_index = dm_icomp_block_meta_page_index(
+ req->bio->bi_iter.bi_sector >> DMCP_BLOCK_SECTOR_SHIFT, false);
+ if (meta_start > page_index)
+ meta_start = page_index;
+
+ page_index = dm_icomp_block_meta_page_index(
+ (req->bio->bi_iter.bi_sector + count) >> DMCP_BLOCK_SECTOR_SHIFT,
+ true);
+
+ if (meta_end < page_index)
+ meta_end = page_index;
+update_meta:
+ if (req->info->write_mode == DMCP_WRITE_THROUGH ||
+ (req->bio->bi_rw & REQ_FUA)) {
+ dm_icomp_get_req(req);
+ dm_icomp_write_meta(req->info, meta_start, meta_end + 1, req,
+ dm_icomp_write_meta_done, req->bio->bi_rw);
+ }
+}
+
+static void dm_icomp_handle_write_read_existing(struct dm_icomp_req *req)
+{
+ u64 block_index, first_block_index;
+ u16 logical_sectors, data_sectors;
+
+ req->stage = STAGE_READ_EXISTING;
+
+ block_index = dm_icomp_sector_to_block(req->bio->bi_iter.bi_sector);
+ dm_icomp_get_extent(req->info, block_index, &first_block_index,
+ &logical_sectors, &data_sectors);
+ if (data_sectors > 0 && (first_block_index < block_index ||
+ first_block_index + dm_icomp_sector_to_block(logical_sectors) >
+ dm_icomp_sector_to_block(bio_end_sector(req->bio))))
+ dm_icomp_read_one_extent(req, first_block_index,
+ logical_sectors, data_sectors);
+
+ if (req->result)
+ return;
+
+ if (first_block_index + dm_icomp_sector_to_block(logical_sectors) >=
+ dm_icomp_sector_to_block(bio_end_sector(req->bio)))
+ goto out;
+
+ block_index = dm_icomp_sector_to_block(bio_end_sector(req->bio)) - 1;
+ dm_icomp_get_extent(req->info, block_index, &first_block_index,
+ &logical_sectors, &data_sectors);
+ if (data_sectors > 0 &&
+ first_block_index + dm_icomp_sector_to_block(logical_sectors) >
+ block_index + 1)
+ dm_icomp_read_one_extent(req, first_block_index,
+ logical_sectors, data_sectors);
+
+ if (req->result)
+ return;
+out:
+ if (list_empty(&req->all_io))
+ dm_icomp_handle_write_comp(req);
+}
+
+static void dm_icomp_handle_write_request(struct dm_icomp_req *req)
+{
+ dm_icomp_get_req(req);
+
+ if (req->stage == STAGE_INIT) {
+ if (!dm_icomp_lock_req_range(req)) {
+ dm_icomp_put_req(req);
+ return;
+ }
+
+ dm_icomp_handle_write_read_existing(req);
+ } else if (req->stage == STAGE_READ_EXISTING)
+ dm_icomp_handle_write_comp(req);
+
+ dm_icomp_put_req(req);
+}
+
+/* For writeback mode */
+static void dm_icomp_handle_flush_request(struct dm_icomp_req *req)
+{
+ struct writeback_flush_data wb;
+
+ atomic_set(&wb.cnt, 1);
+ init_completion(&wb.complete);
+
+ dm_icomp_flush_dirty_meta(req->info, &wb);
+
+ writeback_flush_io_done(&wb, 0);
+ wait_for_completion(&wb.complete);
+
+ req->bio->bi_error = 0;
+ bio_endio(req->bio);
+ kmem_cache_free(dm_icomp_req_cachep, req);
+}
+
+static void dm_icomp_handle_request(struct dm_icomp_req *req)
+{
+ if (req->bio->bi_rw & REQ_FLUSH)
+ dm_icomp_handle_flush_request(req);
+ else if (req->bio->bi_rw & REQ_WRITE)
+ dm_icomp_handle_write_request(req);
+ else
+ dm_icomp_handle_read_request(req);
+}
+
+static void dm_icomp_do_request_work(struct work_struct *work)
+{
+ struct dm_icomp_io_worker *worker = container_of(work,
+ struct dm_icomp_io_worker, work);
+ LIST_HEAD(list);
+ struct dm_icomp_req *req;
+ struct blk_plug plug;
+ bool repeat;
+
+ blk_start_plug(&plug);
+again:
+ spin_lock_irq(&worker->lock);
+ list_splice_init(&worker->pending, &list);
+ spin_unlock_irq(&worker->lock);
+
+ repeat = !list_empty(&list);
+ while (!list_empty(&list)) {
+ req = list_first_entry(&list, struct dm_icomp_req, sibling);
+ list_del(&req->sibling);
+
+ dm_icomp_handle_request(req);
+ }
+ if (repeat)
+ goto again;
+ blk_finish_plug(&plug);
+}
+
+static int dm_icomp_map(struct dm_target *ti, struct bio *bio)
+{
+ struct dm_icomp_info *info = ti->private;
+ struct dm_icomp_req *req;
+
+ if ((bio->bi_rw & REQ_FLUSH) &&
+ info->write_mode == DMCP_WRITE_THROUGH) {
+ bio->bi_bdev = info->dev->bdev;
+ return DM_MAPIO_REMAPPED;
+ }
+ req = kmem_cache_alloc(dm_icomp_req_cachep, GFP_NOIO);
+ if (!req)
+ return -EIO;
+
+ req->bio = bio;
+ req->info = info;
+ atomic_set(&req->io_pending, 0);
+ INIT_LIST_HEAD(&req->all_io);
+ req->result = 0;
+ req->stage = STAGE_INIT;
+ req->locked_locks = 0;
+
+ req->cpu = raw_smp_processor_id();
+ dm_icomp_queue_req(info, req);
+
+ return DM_MAPIO_SUBMITTED;
+}
+
+static void dm_icomp_status(struct dm_target *ti, status_type_t type,
+ unsigned int status_flags, char *result, unsigned int maxlen)
+{
+ struct dm_icomp_info *info = ti->private;
+ unsigned int sz = 0;
+
+ switch (type) {
+ case STATUSTYPE_INFO:
+ DMEMIT("%lu %lu %lu",
+ atomic64_read(&info->uncompressed_write_size),
+ atomic64_read(&info->compressed_write_size),
+ atomic64_read(&info->meta_write_size));
+ break;
+ case STATUSTYPE_TABLE:
+ if (info->write_mode == DMCP_WRITE_BACK)
+ DMEMIT("%s %s %d", info->dev->name, "writeback",
+ info->writeback_delay);
+ else
+ DMEMIT("%s %s", info->dev->name, "writethrough");
+ break;
+ }
+}
+
+static int dm_icomp_iterate_devices(struct dm_target *ti,
+ iterate_devices_callout_fn fn, void *data)
+{
+ struct dm_icomp_info *info = ti->private;
+
+ return fn(ti, info->dev, info->data_start,
+ info->data_blocks << DMCP_BLOCK_SECTOR_SHIFT, data);
+}
+
+static void dm_icomp_io_hints(struct dm_target *ti,
+ struct queue_limits *limits)
+{
+ /* No blk_limits_logical_block_size */
+ limits->logical_block_size = limits->physical_block_size =
+ limits->io_min = DMCP_BLOCK_SIZE;
+}
+
+static struct target_type dm_icomp_target = {
+ .name = "inplacecompress",
+ .version = {1, 0, 0},
+ .module = THIS_MODULE,
+ .ctr = dm_icomp_ctr,
+ .dtr = dm_icomp_dtr,
+ .map = dm_icomp_map,
+ .status = dm_icomp_status,
+ .iterate_devices = dm_icomp_iterate_devices,
+ .io_hints = dm_icomp_io_hints,
+};
+
+static int __init dm_icomp_init(void)
+{
+ int r;
+
+ for (r = 0; r < ARRAY_SIZE(compressors); r++)
+ if (crypto_has_comp(compressors[r].name, 0, 0))
+ break;
+ if (r >= ARRAY_SIZE(compressors)) {
+ DMWARN("No crypto compressors are supported");
+ return -EINVAL;
+ }
+
+ default_compressor = r;
+
+ r = -ENOMEM;
+ dm_icomp_req_cachep = kmem_cache_create("dm_icomp_requests",
+ sizeof(struct dm_icomp_req), 0, 0, NULL);
+ if (!dm_icomp_req_cachep) {
+ DMWARN("Can't create request cache");
+ goto err;
+ }
+
+ dm_icomp_io_range_cachep = kmem_cache_create("dm_icomp_io_range",
+ sizeof(struct dm_icomp_io_range), 0, 0, NULL);
+ if (!dm_icomp_io_range_cachep) {
+ DMWARN("Can't create io_range cache");
+ goto err;
+ }
+
+ dm_icomp_meta_io_cachep = kmem_cache_create("dm_icomp_meta_io",
+ sizeof(struct dm_icomp_meta_io), 0, 0, NULL);
+ if (!dm_icomp_meta_io_cachep) {
+ DMWARN("Can't create meta_io cache");
+ goto err;
+ }
+
+ dm_icomp_wq = alloc_workqueue("dm_icomp_io",
+ WQ_UNBOUND|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 0);
+ if (!dm_icomp_wq) {
+ DMWARN("Can't create io workqueue");
+ goto err;
+ }
+
+ r = dm_register_target(&dm_icomp_target);
+ if (r < 0) {
+ DMWARN("target registration failed");
+ goto err;
+ }
+
+ for_each_possible_cpu(r) {
+ INIT_LIST_HEAD(&dm_icomp_io_workers[r].pending);
+ spin_lock_init(&dm_icomp_io_workers[r].lock);
+ INIT_WORK(&dm_icomp_io_workers[r].work,
+ dm_icomp_do_request_work);
+ }
+ return 0;
+err:
+ kmem_cache_destroy(dm_icomp_req_cachep);
+ kmem_cache_destroy(dm_icomp_io_range_cachep);
+ kmem_cache_destroy(dm_icomp_meta_io_cachep);
+ if (dm_icomp_wq)
+ destroy_workqueue(dm_icomp_wq);
+
+ return r;
+}
+
+static void __exit dm_icomp_exit(void)
+{
+ dm_unregister_target(&dm_icomp_target);
+ kmem_cache_destroy(dm_icomp_req_cachep);
+ kmem_cache_destroy(dm_icomp_io_range_cachep);
+ kmem_cache_destroy(dm_icomp_meta_io_cachep);
+ destroy_workqueue(dm_icomp_wq);
+}
+
+module_init(dm_icomp_init);
+module_exit(dm_icomp_exit);
+
+MODULE_AUTHOR("Shaohua Li <shli@kernel.org>");
+MODULE_DESCRIPTION(DM_NAME " target with data inplace-compression");
+MODULE_LICENSE("GPL");
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
new file mode 100644
index 0000000..e07b9b7
--- /dev/null
+++ b/drivers/md/dm-inplace-compress.h
@@ -0,0 +1,140 @@
+#ifndef __DM_INPLACE_COMPRESS_H__
+#define __DM_INPLACE_COMPRESS_H__
+#include <linux/types.h>
+
+#define DMCP_SUPER_MAGIC 0x106526c206506c09
+struct dm_icomp_super_block {
+ __le64 magic;
+ __le64 meta_blocks;
+ __le64 data_blocks;
+ u8 comp_alg;
+} __packed;
+
+#define DMCP_COMP_ALG_LZO 0
+
+#ifdef __KERNEL__
+struct dm_icomp_compressor_data {
+ char *name;
+ int (*comp_len)(int comp_len);
+};
+
+static inline int lzo_comp_len(int comp_len)
+{
+ return lzo1x_worst_compress(comp_len);
+}
+
+/*
+ * Minium logical sector size of this target is 4096 byte, which is a block.
+ * Data of a block is compressed. Compressed data is round up to 512B, which is
+ * the payload. For each block, we have 5 bits meta data. bit 0 - 3 stands
+ * payload length. 0 - 8 sectors. If compressed payload length is 8 sectors, we
+ * just store uncompressed data. Actual compressed data length is stored at the
+ * last 32 bits of payload if data is compressed. In disk, payload is stored at
+ * the beginning of logical sector of the block. If IO size is bigger than one
+ * block, we store the whole data as an extent. Bit 4 stands tail for an
+ * extent. Max allowed extent size is 128k.
+ */
+#define DMCP_BLOCK_SIZE 4096
+#define DMCP_BLOCK_SHIFT 12
+#define DMCP_BLOCK_SECTOR_SHIFT (DMCP_BLOCK_SHIFT - 9)
+
+#define DMCP_MIN_SIZE 4096
+#define DMCP_MAX_SIZE (128 * 1024)
+
+#define DMCP_LENGTH_MASK ((1 << 4) - 1)
+#define DMCP_TAIL_MASK (1 << 4)
+#define DMCP_META_BITS 5
+
+#define DMCP_META_START_SECTOR (DMCP_BLOCK_SIZE >> 9)
+
+enum DMCP_WRITE_MODE {
+ DMCP_WRITE_BACK,
+ DMCP_WRITE_THROUGH,
+};
+
+/* 128*4 = 512k, since max IO size is 128k, an IO crosses at most 2 hash */
+#define BITMAP_HASH_SHIFT 7
+#define BITMAP_HASH_MASK ((1 << 6) - 1)
+#define BITMAP_HASH_LEN 64
+struct dm_icomp_hash_lock {
+ int io_running;
+ spinlock_t wait_lock;
+ struct list_head wait_list;
+};
+
+struct dm_icomp_info {
+ struct dm_target *ti;
+ struct dm_dev *dev;
+
+ int comp_alg;
+ struct crypto_comp *tfm[NR_CPUS];
+
+ sector_t data_start;
+ u64 data_blocks;
+
+ char *meta_bitmap;
+ u64 meta_bitmap_bits;
+ u64 meta_bitmap_pages;
+ struct dm_icomp_hash_lock bitmap_locks[BITMAP_HASH_LEN];
+
+ enum DMCP_WRITE_MODE write_mode;
+ unsigned int writeback_delay; /* second */
+ struct task_struct *writeback_tsk;
+ struct dm_io_client *io_client;
+
+ atomic64_t compressed_write_size;
+ atomic64_t uncompressed_write_size;
+ atomic64_t meta_write_size;
+};
+
+struct dm_icomp_meta_io {
+ struct dm_io_request io_req;
+ struct dm_io_region io_region;
+ void *data;
+ void (*fn)(void *data, unsigned long error);
+};
+
+struct dm_icomp_io_range {
+ struct dm_io_request io_req;
+ struct dm_io_region io_region;
+ void *decomp_data;
+ unsigned int decomp_len;
+ void *comp_data;
+ unsigned int comp_len; /* For write, this is estimated */
+ struct list_head next;
+ struct dm_icomp_req *req;
+};
+
+enum DMCP_REQ_STAGE {
+ STAGE_INIT,
+ STAGE_READ_EXISTING,
+ STAGE_READ_DECOMP,
+ STAGE_WRITE_COMP,
+ STAGE_DONE,
+};
+
+struct dm_icomp_req {
+ struct bio *bio;
+ struct dm_icomp_info *info;
+ struct list_head sibling;
+
+ struct list_head all_io;
+ atomic_t io_pending;
+ enum DMCP_REQ_STAGE stage;
+
+ struct dm_icomp_hash_lock *locks[2];
+ int locked_locks;
+ int result;
+
+ int cpu;
+ struct work_struct work;
+};
+
+struct dm_icomp_io_worker {
+ struct list_head pending;
+ spinlock_t lock;
+ struct work_struct work;
+};
+#endif
+
+#endif
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 02/16] DM: Ability to choose the compressor.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
2016-08-15 17:36 ` [RFC PATCH 01/16] DM: dm-inplace-compress: an inplace compressed DM target Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 03/16] DM: Error if enough space is not available Ram Pai
` (13 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
Ability to create a block device with a compression algorithm of the users
choice. Currently lzo and nx842 compressors are supported.
If the compressor algorithm is not specified the default setting in
/sys/modules/dm-inplace-compress/parameters/compress is used.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 129 ++++++++++++++++++++++++++++----------
drivers/md/dm-inplace-compress.h | 8 ++-
2 files changed, 103 insertions(+), 34 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index c3c3750..70d6c0e 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -20,8 +20,62 @@ static struct dm_icomp_compressor_data compressors[] = {
.name = "lzo",
.comp_len = lzo_comp_len,
},
+ [DMCP_COMP_ALG_842] = {
+ .name = "842",
+ .comp_len = nx842_comp_len,
+ },
+};
+static int default_compressor = -1;
+
+#define DMCP_ALGO_LENGTH 9
+static char dm_icomp_algorithm[DMCP_ALGO_LENGTH] = "lzo";
+static struct kparam_string dm_icomp_compressor_kparam = {
+ .string = dm_icomp_algorithm,
+ .maxlen = sizeof(dm_icomp_algorithm),
};
-static int default_compressor;
+static int dm_icomp_compressor_param_set(const char *,
+ const struct kernel_param *);
+static struct kernel_param_ops dm_icomp_compressor_param_ops = {
+ .set = dm_icomp_compressor_param_set,
+ .get = param_get_string,
+};
+module_param_cb(compress_algorithm, &dm_icomp_compressor_param_ops,
+ &dm_icomp_compressor_kparam, 0644);
+
+static int dm_icomp_get_compressor(const char *s)
+{
+ int r, val_len;
+
+ if (crypto_has_comp(s, 0, 0)) {
+ for (r = 0; r < ARRAY_SIZE(compressors); r++) {
+ val_len = strlen(compressors[r].name);
+ if (strncmp(s, compressors[r].name, val_len) == 0)
+ return r;
+ }
+ }
+ return -1;
+}
+
+static int dm_icomp_compressor_param_set(const char *val,
+ const struct kernel_param *kp)
+{
+ int ret;
+ char str[kp->str->maxlen], *s;
+ int val_len = strlen(val)+1;
+
+ strlcpy(str, val, val_len);
+ s = strim(str);
+ ret = dm_icomp_get_compressor(s);
+ if (ret < 0) {
+ DMWARN("Compressor %s not supported", s);
+ return -1;
+ }
+ DMWARN("compressor is %s", s);
+ default_compressor = ret;
+ strlcpy(dm_icomp_algorithm, compressors[ret].name,
+ sizeof(dm_icomp_algorithm));
+ return 0;
+}
static struct kmem_cache *dm_icomp_req_cachep;
static struct kmem_cache *dm_icomp_io_range_cachep;
@@ -417,7 +471,7 @@ static int dm_icomp_read_or_create_super(struct dm_icomp_info *info)
ret = -EINVAL;
goto out;
}
- if (!crypto_has_comp(compressors[super->comp_alg].name, 0, 0)) {
+ if (!crypto_has_comp(compressors[info->comp_alg].name, 0, 0)) {
info->ti->error =
"Compressor algorithm doesn't support";
ret = -EINVAL;
@@ -436,7 +490,6 @@ static int dm_icomp_read_or_create_super(struct dm_icomp_info *info)
new_super = true;
}
- info->comp_alg = super->comp_alg;
if (dm_icomp_alloc_compressor(info)) {
ret = -ENOMEM;
goto out;
@@ -467,50 +520,56 @@ out:
}
/*
- * <dev> <writethough>/<writeback> <meta_commit_delay>
+ * <dev> [ <writethough>/<writeback> <meta_commit_delay> ]
+ * [ <compressor> <type> ]
*/
static int dm_icomp_ctr(struct dm_target *ti, unsigned int argc, char **argv)
{
struct dm_icomp_info *info;
- char write_mode[15];
+ char mode[15];
+ int par = 0;
int ret, i;
- if (argc < 2) {
- ti->error = "Invalid argument count";
- return -EINVAL;
- }
-
info = kzalloc(sizeof(*info), GFP_KERNEL);
if (!info) {
ti->error = "dm-inplace-compress: Cannot allocate context";
return -ENOMEM;
}
info->ti = ti;
-
- if (sscanf(argv[1], "%s", write_mode) != 1) {
- ti->error = "Invalid argument";
- ret = -EINVAL;
- goto err_para;
- }
-
- if (strcmp(write_mode, "writeback") == 0) {
- if (argc != 3) {
+ info->comp_alg = default_compressor;
+ while (++par < argc) {
+ if (sscanf(argv[par], "%s", mode) != 1) {
ti->error = "Invalid argument";
ret = -EINVAL;
goto err_para;
}
- info->write_mode = DMCP_WRITE_BACK;
- if (sscanf(argv[2], "%u", &info->writeback_delay) != 1) {
- ti->error = "Invalid argument";
- ret = -EINVAL;
- goto err_para;
+
+ if (strcmp(mode, "writeback") == 0) {
+ info->write_mode = DMCP_WRITE_BACK;
+ if (sscanf(argv[++par], "%u",
+ &info->writeback_delay) != 1) {
+ ti->error = "Invalid argument";
+ ret = -EINVAL;
+ goto err_para;
+ }
+ } else if (strcmp(mode, "writethrough") == 0) {
+ info->write_mode = DMCP_WRITE_THROUGH;
+ } else if (strcmp(mode, "compressor") == 0) {
+ if (sscanf(argv[++par], "%s", mode) != 1) {
+ ti->error = "Invalid argument";
+ ret = -EINVAL;
+ goto err_para;
+ }
+ ret = dm_icomp_get_compressor(mode);
+ if (ret >= 0) {
+ DMWARN("compressor is %s", mode);
+ info->comp_alg = ret;
+ } else {
+ ti->error = "Unsupported compressor";
+ ret = -EINVAL;
+ goto err_para;
+ }
}
- } else if (strcmp(write_mode, "writethrough") == 0) {
- info->write_mode = DMCP_WRITE_THROUGH;
- } else {
- ti->error = "Invalid argument";
- ret = -EINVAL;
- goto err_para;
}
if (dm_get_device(ti, argv[0], dm_table_get_mode(ti->table),
@@ -1407,16 +1466,20 @@ static struct target_type dm_icomp_target = {
static int __init dm_icomp_init(void)
{
int r;
+ int arr_size = ARRAY_SIZE(compressors);
- for (r = 0; r < ARRAY_SIZE(compressors); r++)
+ for (r = 0; r < arr_size; r++)
if (crypto_has_comp(compressors[r].name, 0, 0))
break;
- if (r >= ARRAY_SIZE(compressors)) {
+ if (r >= arr_size) {
DMWARN("No crypto compressors are supported");
return -EINVAL;
}
-
default_compressor = r;
+ strlcpy(dm_icomp_algorithm, compressors[r].name,
+ sizeof(dm_icomp_algorithm));
+ DMWARN(" %s crypto compressor used ",
+ compressors[default_compressor].name);
r = -ENOMEM;
dm_icomp_req_cachep = kmem_cache_create("dm_icomp_requests",
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
index e07b9b7..b61ff0d 100644
--- a/drivers/md/dm-inplace-compress.h
+++ b/drivers/md/dm-inplace-compress.h
@@ -10,7 +10,8 @@ struct dm_icomp_super_block {
u8 comp_alg;
} __packed;
-#define DMCP_COMP_ALG_LZO 0
+#define DMCP_COMP_ALG_LZO 1
+#define DMCP_COMP_ALG_842 0
#ifdef __KERNEL__
struct dm_icomp_compressor_data {
@@ -23,6 +24,11 @@ static inline int lzo_comp_len(int comp_len)
return lzo1x_worst_compress(comp_len);
}
+static inline int nx842_comp_len(int comp_len)
+{
+ return comp_len;
+}
+
/*
* Minium logical sector size of this target is 4096 byte, which is a block.
* Data of a block is compressed. Compressed data is round up to 512B, which is
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 03/16] DM: Error if enough space is not available.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
2016-08-15 17:36 ` [RFC PATCH 01/16] DM: dm-inplace-compress: an inplace compressed DM target Ram Pai
2016-08-15 17:36 ` [RFC PATCH 02/16] DM: Ability to choose the compressor Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 04/16] DM: Ensure that the read request is within the device range Ram Pai
` (12 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
if enough space is not available to create a block of the specified size error
out.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 6 ++++++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index 70d6c0e..17221a1 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -453,6 +453,12 @@ static int dm_icomp_read_or_create_super(struct dm_icomp_info *info)
info->data_blocks = data_blocks;
info->data_start = (1 + meta_blocks) << DMCP_BLOCK_SECTOR_SHIFT;
+ if ((data_blocks << DMCP_BLOCK_SECTOR_SHIFT) < info->ti->len) {
+ info->ti->error =
+ "Insufficient sectors to satisfy requested size";
+ return -ENOMEM;
+ }
+
addr = kzalloc(DMCP_BLOCK_SIZE, GFP_KERNEL);
if (!addr) {
info->ti->error = "Cannot allocate super";
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 04/16] DM: Ensure that the read request is within the device range.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (2 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 03/16] DM: Error if enough space is not available Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 05/16] DM: allocation/free helper routines Ram Pai
` (11 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, snitzer, agk, corbet
If a read request is not within the device range return error.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 10 +++++++++-
1 files changed, 9 insertions(+), 1 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index 17221a1..bf18028 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -1025,6 +1025,12 @@ static void dm_icomp_read_one_extent(struct dm_icomp_req *req, u64 block,
{
struct dm_icomp_io_range *io;
+ if (block+(data_sectors>>DMCP_BLOCK_SECTOR_SHIFT) >=
+ req->info->data_blocks) {
+ req->result = -EIO;
+ return;
+ }
+
io = dm_icomp_create_io_range(req, data_sectors << 9,
logical_sectors << 9);
if (!io) {
@@ -1063,7 +1069,9 @@ again:
block_index = first_block_index + (logical_sectors >>
DMCP_BLOCK_SECTOR_SHIFT);
- if ((block_index << DMCP_BLOCK_SECTOR_SHIFT) < bio_end_sector(req->bio))
+ if (((block_index << DMCP_BLOCK_SECTOR_SHIFT) <
+ bio_end_sector(req->bio)) &&
+ ((block_index) < req->info->data_blocks))
goto again;
/* A shortcut if all data is in already */
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 05/16] DM: allocation/free helper routines.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (3 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 04/16] DM: Ensure that the read request is within the device range Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 06/16] DM: separate out compression and decompression routines Ram Pai
` (10 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, snitzer, agk, corbet
Helper functions to allocate/reallocate and free memory.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 17 +++++++++++++++++
1 files changed, 17 insertions(+), 0 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index bf18028..c11567c 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -774,6 +774,23 @@ static void dm_icomp_get_req(struct dm_icomp_req *req)
atomic_inc(&req->io_pending);
}
+static void *dm_icomp_kmalloc(size_t size, gfp_t flags)
+{
+ return kmalloc(size, flags);
+}
+
+static void *dm_icomp_krealloc(void *addr, size_t size,
+ size_t orig_size, gfp_t flags)
+{
+ return krealloc(addr, size, flags);
+}
+
+static void dm_icomp_kfree(void *addr, unsigned int size)
+{
+ kfree(addr);
+}
+
+
static void dm_icomp_free_io_range(struct dm_icomp_io_range *io)
{
kfree(io->decomp_data);
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 06/16] DM: separate out compression and decompression routines.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (4 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 05/16] DM: allocation/free helper routines Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 07/16] DM: Optimize memory allocated to hold compressed buffer Ram Pai
` (9 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
Simplify the code by separating out the compression and decompression routines.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 127 +++++++++++++++++++++-----------------
1 files changed, 71 insertions(+), 56 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index c11567c..5c39169 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -793,8 +793,8 @@ static void dm_icomp_kfree(void *addr, unsigned int size)
static void dm_icomp_free_io_range(struct dm_icomp_io_range *io)
{
- kfree(io->decomp_data);
- kfree(io->comp_data);
+ dm_icomp_kfree(io->decomp_data, io->decomp_len);
+ dm_icomp_kfree(io->comp_data, io->comp_len);
kmem_cache_free(dm_icomp_io_range_cachep, io);
}
@@ -861,12 +861,12 @@ static struct dm_icomp_io_range *dm_icomp_create_io_range(
if (!io)
return NULL;
- io->comp_data = kmalloc(dm_icomp_compressor_len(req->info, comp_len),
- GFP_NOIO);
- io->decomp_data = kmalloc(decomp_len, GFP_NOIO);
+ io->comp_data = dm_icomp_kmalloc(
+ dm_icomp_compressor_len(req->info, comp_len), GFP_NOIO);
+ io->decomp_data = dm_icomp_kmalloc(decomp_len, GFP_NOIO);
if (!io->decomp_data || !io->comp_data) {
- kfree(io->decomp_data);
- kfree(io->comp_data);
+ dm_icomp_kfree(io->decomp_data, io->decomp_len);
+ dm_icomp_kfree(io->comp_data, io->comp_len);
kmem_cache_free(dm_icomp_io_range_cachep, io);
return NULL;
}
@@ -928,51 +928,66 @@ static void dm_icomp_bio_copy(struct bio *bio, off_t bio_off, void *buf,
* We store the actual compressed len in the last u32 of the payload.
* If there is no free space, we add 512 to the payload size.
*/
-static int dm_icomp_io_range_comp(struct dm_icomp_info *info, void *comp_data,
- unsigned int *comp_len, void *decomp_data, unsigned int decomp_len,
- bool do_comp)
+static int dm_icomp_io_range_compress(struct dm_icomp_info *info,
+ struct dm_icomp_io_range *io, unsigned int *comp_len,
+ void *decomp_data, unsigned int decomp_len)
{
- struct crypto_comp *tfm;
+ unsigned int actual_comp_len = io->comp_len;
u32 *addr;
- unsigned int actual_comp_len;
+ struct crypto_comp *tfm = info->tfm[get_cpu()];
int ret;
- if (do_comp) {
- actual_comp_len = *comp_len;
-
- tfm = info->tfm[get_cpu()];
- ret = crypto_comp_compress(tfm, decomp_data, decomp_len,
- comp_data, &actual_comp_len);
- put_cpu();
+ ret = crypto_comp_compress(tfm, decomp_data, decomp_len,
+ io->comp_data, &actual_comp_len);
- atomic64_add(decomp_len, &info->uncompressed_write_size);
- if (ret || decomp_len < actual_comp_len + sizeof(u32) + 512) {
- *comp_len = decomp_len;
- atomic64_add(*comp_len, &info->compressed_write_size);
- return 1;
- }
+ put_cpu();
+ if (ret < 0)
+ DMWARN("CO Error %d ", ret);
- *comp_len = round_up(actual_comp_len, 512);
- if (*comp_len - actual_comp_len < sizeof(u32))
- *comp_len += 512;
+ atomic64_add(decomp_len, &info->uncompressed_write_size);
+ if (ret || decomp_len < actual_comp_len + sizeof(u32) + 512) {
+ *comp_len = decomp_len;
atomic64_add(*comp_len, &info->compressed_write_size);
- addr = comp_data + *comp_len;
- addr--;
- *addr = cpu_to_le32(actual_comp_len);
- } else {
- if (*comp_len == decomp_len)
- return 1;
- addr = comp_data + *comp_len;
- addr--;
- actual_comp_len = le32_to_cpu(*addr);
-
- tfm = info->tfm[get_cpu()];
- ret = crypto_comp_decompress(tfm, comp_data, actual_comp_len,
- decomp_data, &decomp_len);
- put_cpu();
- if (ret)
- return -EINVAL;
+ return 1;
}
+
+ *comp_len = round_up(actual_comp_len, 512);
+ if (*comp_len - actual_comp_len < sizeof(u32))
+ *comp_len += 512;
+ atomic64_add(*comp_len, &info->compressed_write_size);
+ addr = io->comp_data + *comp_len;
+ addr--;
+ *addr = cpu_to_le32(actual_comp_len);
+ return 0;
+}
+
+/*
+ * return value:
+ * < 0 : error
+ * == 0 : ok
+ * == 1 : ok, but comp/decomp is skipped
+ */
+static int dm_icomp_io_range_decompress(struct dm_icomp_info *info,
+ void *comp_data, unsigned int comp_len, void *decomp_data,
+ unsigned int decomp_len)
+{
+ struct crypto_comp *tfm;
+ u32 *addr;
+ int ret;
+
+ if (comp_len == decomp_len)
+ return 1;
+
+ addr = comp_data + comp_len;
+ addr--;
+ comp_len = le32_to_cpu(*addr);
+
+ tfm = info->tfm[get_cpu()];
+ ret = crypto_comp_decompress(tfm, comp_data, comp_len,
+ decomp_data, &decomp_len);
+ put_cpu();
+ if (ret)
+ return -EINVAL;
return 0;
}
@@ -993,8 +1008,8 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
io->io_region.sector -= req->info->data_start;
/* Do decomp here */
- ret = dm_icomp_io_range_comp(req->info, io->comp_data,
- &io->comp_len, io->decomp_data, io->decomp_len, false);
+ ret = dm_icomp_io_range_decompress(req->info, io->comp_data,
+ io->comp_len, io->decomp_data, io->decomp_len);
if (ret < 0) {
req->result = -EIO;
return;
@@ -1140,8 +1155,8 @@ static int dm_icomp_handle_write_modify(struct dm_icomp_io_range *io,
io->io_region.sector -= req->info->data_start;
/* decompress original data */
- ret = dm_icomp_io_range_comp(req->info, io->comp_data, &io->comp_len,
- io->decomp_data, io->decomp_len, false);
+ ret = dm_icomp_io_range_decompress(req->info, io->comp_data,
+ io->comp_len, io->decomp_data, io->decomp_len);
if (ret < 0) {
req->result = -EINVAL;
return -EIO;
@@ -1164,11 +1179,11 @@ static int dm_icomp_handle_write_modify(struct dm_icomp_io_range *io,
((req->bio->bi_iter.bi_sector - start) << 9),
bio_sectors(req->bio) << 9, true);
- kfree(io->comp_data);
+ dm_icomp_kfree(io->comp_data, io->comp_len);
/* New compressed len might be bigger */
- io->comp_data = kmalloc(
- dm_icomp_compressor_len(req->info,
- io->decomp_len), GFP_NOIO);
+ io->comp_data = dm_icomp_kmalloc(
+ dm_icomp_compressor_len(
+ req->info, io->decomp_len), GFP_NOIO);
io->comp_len = io->decomp_len;
if (!io->comp_data) {
req->result = -ENOMEM;
@@ -1197,8 +1212,8 @@ static int dm_icomp_handle_write_modify(struct dm_icomp_io_range *io,
/* assume compress less data uses less space (at least 4k lsess data) */
comp_len = io->comp_len;
- ret = dm_icomp_io_range_comp(req->info, io->comp_data, &comp_len,
- io->decomp_data + (offset << 9), count << 9, true);
+ ret = dm_icomp_io_range_compress(req->info, io, &comp_len,
+ io->decomp_data + (offset << 9), count << 9);
if (ret < 0) {
req->result = -EIO;
return -EIO;
@@ -1260,8 +1275,8 @@ static void dm_icomp_handle_write_comp(struct dm_icomp_req *req)
/* compress data */
comp_len = io->comp_len;
- ret = dm_icomp_io_range_comp(req->info, io->comp_data, &comp_len,
- io->decomp_data, count << 9, true);
+ ret = dm_icomp_io_range_compress(req->info, io, &comp_len,
+ io->decomp_data, count << 9);
if (ret < 0) {
dm_icomp_free_io_range(io);
req->result = -EIO;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 07/16] DM: Optimize memory allocated to hold compressed buffer.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (5 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 06/16] DM: separate out compression and decompression routines Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 08/16] DM: Tag a magicmarker at the end of each compressed segment Ram Pai
` (8 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, snitzer, agk, corbet
On an average the compressed size is less than 50% of the original buffer. Use
this knowledge to optimize the amount of space allocated to hold the compressed
buffer. If the allocated size is determined to be insufficient than reallocate
the required size.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 39 ++++++++++++++++++++++++++++++++++++++
drivers/md/dm-inplace-compress.h | 11 ++++++++++
2 files changed, 50 insertions(+), 0 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index 5c39169..fe4a4c1 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -19,10 +19,12 @@ static struct dm_icomp_compressor_data compressors[] = {
[DMCP_COMP_ALG_LZO] = {
.name = "lzo",
.comp_len = lzo_comp_len,
+ .max_comp_len = lzo_max_comp_len,
},
[DMCP_COMP_ALG_842] = {
.name = "842",
.comp_len = nx842_comp_len,
+ .max_comp_len = nx842_max_comp_len,
},
};
static int default_compressor = -1;
@@ -848,6 +850,14 @@ static inline int dm_icomp_compressor_len(struct dm_icomp_info *info, int len)
return len;
}
+static inline int dm_icomp_compressor_maxlen(struct dm_icomp_info *info,
+ int len)
+{
+ if (compressors[info->comp_alg].max_comp_len)
+ return compressors[info->comp_alg].max_comp_len(len);
+ return len;
+}
+
/*
* caller should set region.sector, region.count. bi_rw. IO always to/from
* comp_data
@@ -919,6 +929,25 @@ static void dm_icomp_bio_copy(struct bio *bio, off_t bio_off, void *buf,
}
}
+static int dm_icomp_mod_to_max_io_range(struct dm_icomp_info *info,
+ struct dm_icomp_io_range *io)
+{
+ unsigned int maxlen = dm_icomp_compressor_maxlen(info, io->decomp_len);
+
+ if (maxlen <= io->comp_len)
+ return -ENOSPC;
+ io->io_req.mem.ptr.addr = io->comp_data =
+ dm_icomp_krealloc(io->comp_data, maxlen,
+ io->comp_len, GFP_NOIO);
+ if (!io->comp_data) {
+ DMWARN("UNFORTUNE allocation failure ");
+ io->comp_len = 0;
+ return -ENOSPC;
+ }
+ io->comp_len = maxlen;
+ return 0;
+}
+
/*
* return value:
* < 0 : error
@@ -940,7 +969,17 @@ static int dm_icomp_io_range_compress(struct dm_icomp_info *info,
ret = crypto_comp_compress(tfm, decomp_data, decomp_len,
io->comp_data, &actual_comp_len);
+ if (ret || actual_comp_len > io->comp_len) {
+ ret = dm_icomp_mod_to_max_io_range(info, io);
+ if (!ret) {
+ actual_comp_len = io->comp_len;
+ ret = crypto_comp_compress(tfm, decomp_data, decomp_len,
+ io->comp_data, &actual_comp_len);
+ }
+ }
+
put_cpu();
+
if (ret < 0)
DMWARN("CO Error %d ", ret);
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
index b61ff0d..86c0ce6 100644
--- a/drivers/md/dm-inplace-compress.h
+++ b/drivers/md/dm-inplace-compress.h
@@ -17,15 +17,26 @@ struct dm_icomp_super_block {
struct dm_icomp_compressor_data {
char *name;
int (*comp_len)(int comp_len);
+ int (*max_comp_len)(int comp_len);
};
static inline int lzo_comp_len(int comp_len)
{
+ return lzo1x_worst_compress(comp_len) >> 1;
+}
+
+static inline int lzo_max_comp_len(int comp_len)
+{
return lzo1x_worst_compress(comp_len);
}
static inline int nx842_comp_len(int comp_len)
{
+ return (comp_len>>4)*7; /* less than half: 7/16 */
+}
+
+static inline int nx842_max_comp_len(int comp_len)
+{
return comp_len;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 08/16] DM: Tag a magicmarker at the end of each compressed segment.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (6 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 07/16] DM: Optimize memory allocated to hold compressed buffer Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 09/16] DM: Delay allocation of decompression buffer during read Ram Pai
` (7 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, snitzer, agk, corbet
We store the size of the compressed segment, on the sector boundary. And later
use that location to determine the size of the compressed segment. However if
that location is corrupted for any reason we would'nt know. Hence add a
magicmarker to catch for such corruptions.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 24 ++++++++++++++++--------
drivers/md/dm-inplace-compress.h | 1 +
2 files changed, 17 insertions(+), 8 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index fe4a4c1..04decdd 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -984,19 +984,21 @@ static int dm_icomp_io_range_compress(struct dm_icomp_info *info,
DMWARN("CO Error %d ", ret);
atomic64_add(decomp_len, &info->uncompressed_write_size);
- if (ret || decomp_len < actual_comp_len + sizeof(u32) + 512) {
+ if (ret || decomp_len < actual_comp_len + 2*sizeof(u32) + 512) {
*comp_len = decomp_len;
atomic64_add(*comp_len, &info->compressed_write_size);
return 1;
}
*comp_len = round_up(actual_comp_len, 512);
- if (*comp_len - actual_comp_len < sizeof(u32))
+ if (*comp_len - actual_comp_len < 2*sizeof(u32))
*comp_len += 512;
atomic64_add(*comp_len, &info->compressed_write_size);
addr = io->comp_data + *comp_len;
addr--;
*addr = cpu_to_le32(actual_comp_len);
+ addr--;
+ *addr = cpu_to_le32(DMCP_COMPRESS_MAGIC);
return 0;
}
@@ -1020,13 +1022,19 @@ static int dm_icomp_io_range_decompress(struct dm_icomp_info *info,
addr = comp_data + comp_len;
addr--;
comp_len = le32_to_cpu(*addr);
+ addr--;
- tfm = info->tfm[get_cpu()];
- ret = crypto_comp_decompress(tfm, comp_data, comp_len,
- decomp_data, &decomp_len);
- put_cpu();
- if (ret)
- return -EINVAL;
+ if (comp_len == decomp_len)
+ return 1;
+ if (le32_to_cpu(*addr) == DMCP_COMPRESS_MAGIC) {
+ tfm = info->tfm[get_cpu()];
+ ret = crypto_comp_decompress(tfm, comp_data, comp_len,
+ decomp_data, &decomp_len);
+ put_cpu();
+ if (ret)
+ return -EINVAL;
+ } else
+ memset(decomp_data, 0, decomp_len);
return 0;
}
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
index 86c0ce6..1ce7a6e 100644
--- a/drivers/md/dm-inplace-compress.h
+++ b/drivers/md/dm-inplace-compress.h
@@ -3,6 +3,7 @@
#include <linux/types.h>
#define DMCP_SUPER_MAGIC 0x106526c206506c09
+#define DMCP_COMPRESS_MAGIC 0xfaceecaf
struct dm_icomp_super_block {
__le64 magic;
__le64 meta_blocks;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 09/16] DM: Delay allocation of decompression buffer during read.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (7 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 08/16] DM: Tag a magicmarker at the end of each compressed segment Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 10/16] DM: Try to use the bio buffer for decompression instead of allocating one Ram Pai
` (6 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
The read path allocates a temporary buffer each to hold compressed data and
decompressed data. The buffer to hold the decompressed data is not needed till
the compressed data is read from the device and is uncompressed.
Hence delay the allocation of decompress buffer till it is really needed.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 76 +++++++++++++++++++++++++++++++++-----
drivers/md/dm-inplace-compress.h | 3 +-
2 files changed, 68 insertions(+), 11 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index 04decdd..c56d9b7 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -863,7 +863,7 @@ static inline int dm_icomp_compressor_maxlen(struct dm_icomp_info *info,
* comp_data
*/
static struct dm_icomp_io_range *dm_icomp_create_io_range(
- struct dm_icomp_req *req, int comp_len, int decomp_len)
+ struct dm_icomp_req *req, int comp_len)
{
struct dm_icomp_io_range *io;
@@ -871,12 +871,8 @@ static struct dm_icomp_io_range *dm_icomp_create_io_range(
if (!io)
return NULL;
- io->comp_data = dm_icomp_kmalloc(
- dm_icomp_compressor_len(req->info, comp_len), GFP_NOIO);
- io->decomp_data = dm_icomp_kmalloc(decomp_len, GFP_NOIO);
- if (!io->decomp_data || !io->comp_data) {
- dm_icomp_kfree(io->decomp_data, io->decomp_len);
- dm_icomp_kfree(io->comp_data, io->comp_len);
+ io->comp_data = dm_icomp_kmalloc(comp_len, GFP_NOIO);
+ if (!io->comp_data) {
kmem_cache_free(dm_icomp_io_range_cachep, io);
return NULL;
}
@@ -890,12 +886,42 @@ static struct dm_icomp_io_range *dm_icomp_create_io_range(
io->io_region.bdev = req->info->dev->bdev;
- io->decomp_len = decomp_len;
io->comp_len = comp_len;
io->req = req;
+
+ io->decomp_data = NULL;
+ io->decomp_len = 0;
+ io->decomp_req_len = 0;
return io;
}
+static struct dm_icomp_io_range *dm_icomp_create_io_read_range(
+ struct dm_icomp_req *req, int comp_len, int decomp_len)
+{
+ struct dm_icomp_io_range *io = dm_icomp_create_io_range(req, comp_len);
+
+ if (io) {
+ /* note down the requested length for decompress buffer.
+ * but dont allocate it yet.
+ */
+ io->decomp_req_len = decomp_len;
+ }
+ return io;
+}
+
+static int dm_icomp_update_io_read_range(struct dm_icomp_io_range *io)
+{
+ if (io->decomp_len)
+ return 0;
+
+ io->decomp_data = dm_icomp_kmalloc(io->decomp_req_len, GFP_NOIO);
+ if (!io->decomp_data)
+ return 1;
+ io->decomp_len = io->decomp_req_len;
+
+ return 0;
+}
+
static void dm_icomp_bio_copy(struct bio *bio, off_t bio_off, void *buf,
ssize_t len, bool to_buf)
{
@@ -948,6 +974,31 @@ static int dm_icomp_mod_to_max_io_range(struct dm_icomp_info *info,
return 0;
}
+static struct dm_icomp_io_range *dm_icomp_create_io_write_range(
+ struct dm_icomp_req *req)
+{
+ struct dm_icomp_io_range *io;
+ sector_t size = bio_sectors(req->bio)<<9;
+ int comp_len = dm_icomp_compressor_len(req->info, size);
+ void *addr;
+
+ addr = dm_icomp_kmalloc(size, GFP_NOIO);
+ if (!addr)
+ return NULL;
+
+ io = dm_icomp_create_io_range(req, comp_len);
+ if (!io) {
+ dm_icomp_kfree(addr, size);
+ return NULL;
+ }
+
+ io->decomp_data = addr;
+ io->decomp_len = size;
+
+ dm_icomp_bio_copy(req->bio, 0, io->decomp_data, size, true);
+ return io;
+}
+
/*
* return value:
* < 0 : error
@@ -1054,6 +1105,11 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
io->io_region.sector -= req->info->data_start;
+ if (dm_icomp_update_io_read_range(io)) {
+ req->result = -EIO;
+ return;
+ }
+
/* Do decomp here */
ret = dm_icomp_io_range_decompress(req->info, io->comp_data,
io->comp_len, io->decomp_data, io->decomp_len);
@@ -1110,7 +1166,7 @@ static void dm_icomp_read_one_extent(struct dm_icomp_req *req, u64 block,
return;
}
- io = dm_icomp_create_io_range(req, data_sectors << 9,
+ io = dm_icomp_create_io_read_range(req, data_sectors << 9,
logical_sectors << 9);
if (!io) {
req->result = -EIO;
@@ -1313,7 +1369,7 @@ static void dm_icomp_handle_write_comp(struct dm_icomp_req *req)
goto update_meta;
count = bio_sectors(req->bio);
- io = dm_icomp_create_io_range(req, count << 9, count << 9);
+ io = dm_icomp_create_io_write_range(req);
if (!io) {
req->result = -EIO;
return;
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
index 1ce7a6e..d9cf05a 100644
--- a/drivers/md/dm-inplace-compress.h
+++ b/drivers/md/dm-inplace-compress.h
@@ -116,7 +116,8 @@ struct dm_icomp_io_range {
struct dm_io_request io_req;
struct dm_io_region io_region;
void *decomp_data;
- unsigned int decomp_len;
+ unsigned int decomp_req_len;/* originally requested length */
+ unsigned int decomp_len; /* actual allocated/mapped length */
void *comp_data;
unsigned int comp_len; /* For write, this is estimated */
struct list_head next;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 10/16] DM: Try to use the bio buffer for decompression instead of allocating one.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (8 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 09/16] DM: Delay allocation of decompression buffer during read Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 11/16] DM: Try to avoid temporary buffer allocation to hold compressed data Ram Pai
` (5 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, snitzer, agk, corbet
The read path allocates a temporary buffer to hold decompressed data, which is
than copied into the caller's bio buffer.
Instead of allocating a temporary buffer to hold the decompressed data,
decompress the data in the caller's bio buffer. This can be done only if the
destination in the bio-buffer is contigious within the same bio-segment.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 143 ++++++++++++++++++++++++++-----------
drivers/md/dm-inplace-compress.h | 2 +
2 files changed, 102 insertions(+), 43 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index c56d9b7..f6b95e3 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -792,11 +792,34 @@ static void dm_icomp_kfree(void *addr, unsigned int size)
kfree(addr);
}
+static void dm_icomp_release_decomp_buffer(struct dm_icomp_io_range *io)
+{
+ if (!io->decomp_data)
+ return;
-static void dm_icomp_free_io_range(struct dm_icomp_io_range *io)
+ if (io->decomp_kmap)
+ kunmap(io->decomp_real_data);
+ else
+ dm_icomp_kfree(io->decomp_real_data, io->decomp_req_len);
+ io->decomp_data = io->decomp_real_data = NULL;
+ io->decomp_len = 0;
+ io->decomp_kmap = false;
+}
+
+static void dm_icomp_release_comp_buffer(struct dm_icomp_io_range *io)
{
- dm_icomp_kfree(io->decomp_data, io->decomp_len);
+ if (!io->comp_data)
+ return;
dm_icomp_kfree(io->comp_data, io->comp_len);
+
+ io->comp_data = NULL;
+ io->comp_len = 0;
+}
+
+static void dm_icomp_free_io_range(struct dm_icomp_io_range *io)
+{
+ dm_icomp_release_decomp_buffer(io);
+ dm_icomp_release_comp_buffer(io);
kmem_cache_free(dm_icomp_io_range_cachep, io);
}
@@ -890,7 +913,9 @@ static struct dm_icomp_io_range *dm_icomp_create_io_range(
io->req = req;
io->decomp_data = NULL;
+ io->decomp_real_data = NULL;
io->decomp_len = 0;
+ io->decomp_kmap = false;
io->decomp_req_len = 0;
return io;
}
@@ -909,15 +934,43 @@ static struct dm_icomp_io_range *dm_icomp_create_io_read_range(
return io;
}
-static int dm_icomp_update_io_read_range(struct dm_icomp_io_range *io)
+static int dm_icomp_update_io_read_range(struct dm_icomp_io_range *io,
+ struct bio *bio, ssize_t bio_off)
{
+ struct bvec_iter iter;
+ struct bio_vec bv;
+ bool just_use = false;
+
if (io->decomp_len)
return 0;
+ /* use a bio buffer long enough to hold the uncompressed data */
+ bio_for_each_segment(bv, bio, iter) {
+ int avail_len;
+ int length = bv.bv_len;
+
+ if (!just_use && bio_off >= length) {
+ bio_off -= length;
+ continue;
+ }
+ avail_len = just_use ? length : length-bio_off;
+ if (avail_len >= io->decomp_req_len) {
+ io->decomp_real_data = kmap(bv.bv_page);
+ io->decomp_data = io->decomp_real_data + bio_off;
+ io->decomp_len = io->decomp_req_len = avail_len;
+ io->decomp_kmap = true;
+ return 0;
+ }
+ just_use = true;
+ }
+
+ /* none available. :( Allocate one */
io->decomp_data = dm_icomp_kmalloc(io->decomp_req_len, GFP_NOIO);
if (!io->decomp_data)
return 1;
+ io->decomp_real_data = io->decomp_data;
io->decomp_len = io->decomp_req_len;
+ io->decomp_kmap = false;
return 0;
}
@@ -978,24 +1031,34 @@ static struct dm_icomp_io_range *dm_icomp_create_io_write_range(
struct dm_icomp_req *req)
{
struct dm_icomp_io_range *io;
+ struct bio *bio = req->bio;
sector_t size = bio_sectors(req->bio)<<9;
+ int segments = bio_segments(bio);
int comp_len = dm_icomp_compressor_len(req->info, size);
void *addr;
- addr = dm_icomp_kmalloc(size, GFP_NOIO);
+ if (segments == 1) {
+ struct bio_vec bv = bio_iovec(bio);
+
+ addr = kmap(bv.bv_page);
+ } else
+ addr = dm_icomp_kmalloc(size, GFP_NOIO);
+
if (!addr)
return NULL;
io = dm_icomp_create_io_range(req, comp_len);
if (!io) {
- dm_icomp_kfree(addr, size);
+ (segments == 1) ? kunmap(addr) : dm_icomp_kfree(addr, size);
return NULL;
}
- io->decomp_data = addr;
+ io->decomp_data = io->decomp_real_data = addr;
io->decomp_len = size;
- dm_icomp_bio_copy(req->bio, 0, io->decomp_data, size, true);
+ io->decomp_kmap = (segments == 1);
+ if (!io->decomp_kmap)
+ dm_icomp_bio_copy(req->bio, 0, io->decomp_data, size, true);
return io;
}
@@ -1105,7 +1168,15 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
io->io_region.sector -= req->info->data_start;
- if (dm_icomp_update_io_read_range(io)) {
+ if (io->io_region.sector >=
+ req->bio->bi_iter.bi_sector)
+ dst_off = (io->io_region.sector -
+ req->bio->bi_iter.bi_sector) << 9;
+ else
+ src_off = (req->bio->bi_iter.bi_sector -
+ io->io_region.sector) << 9;
+
+ if (dm_icomp_update_io_read_range(io, req->bio, dst_off)) {
req->result = -EIO;
return;
}
@@ -1114,20 +1185,19 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
ret = dm_icomp_io_range_decompress(req->info, io->comp_data,
io->comp_len, io->decomp_data, io->decomp_len);
if (ret < 0) {
+ dm_icomp_release_decomp_buffer(io);
+ dm_icomp_release_comp_buffer(io);
req->result = -EIO;
return;
}
- if (io->io_region.sector >= req->bio->bi_iter.bi_sector)
- dst_off = (io->io_region.sector -
- req->bio->bi_iter.bi_sector) << 9;
- else
- src_off = (req->bio->bi_iter.bi_sector -
- io->io_region.sector) << 9;
-
len = min_t(ssize_t, io->decomp_len - src_off,
(bio_sectors(req->bio) << 9) - dst_off);
+ dm_icomp_bio_copy(req->bio, dst_off,
+ ((ret == 1) ? io->comp_data : io->decomp_data) + src_off,
+ len, false);
+
/* io range in all_io list is ordered for read IO */
while (bio_off != dst_off) {
ssize_t size = min_t(ssize_t, PAGE_SIZE,
@@ -1137,13 +1207,9 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
bio_off += size;
}
- if (ret == 1)
- dm_icomp_bio_copy(req->bio, dst_off,
- io->comp_data + src_off, len, false);
- else
- dm_icomp_bio_copy(req->bio, dst_off,
- io->decomp_data + src_off, len, false);
bio_off = dst_off + len;
+ dm_icomp_release_decomp_buffer(io);
+ dm_icomp_release_comp_buffer(io);
}
while (bio_off != (bio_sectors(req->bio) << 9)) {
@@ -1270,29 +1336,20 @@ static int dm_icomp_handle_write_modify(struct dm_icomp_io_range *io,
if (start < req->bio->bi_iter.bi_sector && start + count >
bio_end_sector(req->bio)) {
/* we don't split an extent */
- if (ret == 1) {
- memcpy(io->decomp_data, io->comp_data, io->decomp_len);
- dm_icomp_bio_copy(req->bio, 0,
- io->decomp_data +
- ((req->bio->bi_iter.bi_sector - start) << 9),
- bio_sectors(req->bio) << 9, true);
- } else {
- dm_icomp_bio_copy(req->bio, 0,
- io->decomp_data +
- ((req->bio->bi_iter.bi_sector - start) << 9),
- bio_sectors(req->bio) << 9, true);
-
- dm_icomp_kfree(io->comp_data, io->comp_len);
- /* New compressed len might be bigger */
- io->comp_data = dm_icomp_kmalloc(
- dm_icomp_compressor_len(
- req->info, io->decomp_len), GFP_NOIO);
- io->comp_len = io->decomp_len;
- if (!io->comp_data) {
- req->result = -ENOMEM;
- return -EIO;
+ if (!io->decomp_kmap) {
+ if (ret == 1) {
+ memcpy(io->decomp_data, io->comp_data,
+ io->decomp_len);
+ dm_icomp_bio_copy(req->bio, 0,
+ io->decomp_data +
+ ((req->bio->bi_iter.bi_sector - start) << 9),
+ bio_sectors(req->bio) << 9, true);
+ } else {
+ dm_icomp_bio_copy(req->bio, 0,
+ io->decomp_data +
+ ((req->bio->bi_iter.bi_sector - start) << 9),
+ bio_sectors(req->bio) << 9, true);
}
- io->io_req.mem.ptr.addr = io->comp_data;
}
/* need compress data */
ret = 0;
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
index d9cf05a..e144e96 100644
--- a/drivers/md/dm-inplace-compress.h
+++ b/drivers/md/dm-inplace-compress.h
@@ -115,7 +115,9 @@ struct dm_icomp_meta_io {
struct dm_icomp_io_range {
struct dm_io_request io_req;
struct dm_io_region io_region;
+ int decomp_kmap; /* Is the decomp_data kmapped'? */
void *decomp_data;
+ void *decomp_real_data; /* holds the actual start of the buffer */
unsigned int decomp_req_len;/* originally requested length */
unsigned int decomp_len; /* actual allocated/mapped length */
void *comp_data;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 11/16] DM: Try to avoid temporary buffer allocation to hold compressed data.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (9 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 10/16] DM: Try to use the bio buffer for decompression instead of allocating one Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 12/16] DM: release unneeded buffer as soon as possible Ram Pai
` (4 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
The read path
a) allocates a temporary buffer to hold the compressed data.
b) reads the compressed data into the temporary buffer.
c) decompresses the compressed data into the caller's bio-buffer.
We know that the caller's bio-buffer will be atleast as large as the compressed
data. So we could save a buffer allocation in step (a) if we use caller's
bio-buffer to read in the compressed data from the disk. NOTE: this is not
possible all the time, but is possible if the destination within the bio-buffer
falls within the same segment.
This saves us from holding the temporary buffer across read boundaries. Which
is a huge advantage especially if we are operating under acute memory
starvation; mostly when this block device is used as a swap device.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 185 ++++++++++++++++++-------------------
drivers/md/dm-inplace-compress.h | 1 +
2 files changed, 91 insertions(+), 95 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index f6b95e3..bc1cf70 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -810,10 +810,15 @@ static void dm_icomp_release_comp_buffer(struct dm_icomp_io_range *io)
{
if (!io->comp_data)
return;
- dm_icomp_kfree(io->comp_data, io->comp_len);
+
+ if (io->comp_kmap)
+ kunmap(io->comp_data);
+ else
+ dm_icomp_kfree(io->comp_data, io->comp_len);
io->comp_data = NULL;
io->comp_len = 0;
+ io->comp_kmap = false;
}
static void dm_icomp_free_io_range(struct dm_icomp_io_range *io)
@@ -857,6 +862,39 @@ static void dm_icomp_put_req(struct dm_icomp_req *req)
kmem_cache_free(dm_icomp_req_cachep, req);
}
+static void dm_icomp_bio_copy(struct bio *bio, off_t bio_off, void *buf,
+ ssize_t len, bool to_buf)
+{
+ struct bio_vec bv;
+ struct bvec_iter iter;
+ off_t buf_off = 0;
+ ssize_t size;
+ void *addr;
+
+ WARN_ON(bio_off + len > (bio_sectors(bio) << 9));
+
+ bio_for_each_segment(bv, bio, iter) {
+ int length = bv.bv_len;
+
+ if (bio_off >= length) {
+ bio_off -= length;
+ continue;
+ }
+ addr = kmap_atomic(bv.bv_page);
+ size = min_t(ssize_t, len, length - bio_off);
+ if (to_buf)
+ memcpy(buf + buf_off, addr + bio_off + bv.bv_offset,
+ size);
+ else
+ memcpy(addr + bio_off + bv.bv_offset, buf + buf_off,
+ size);
+ kunmap_atomic(addr);
+ bio_off = 0;
+ buf_off += size;
+ len -= size;
+ }
+}
+
static void dm_icomp_io_range_done(unsigned long error, void *context)
{
struct dm_icomp_io_range *io = context;
@@ -886,7 +924,7 @@ static inline int dm_icomp_compressor_maxlen(struct dm_icomp_info *info,
* comp_data
*/
static struct dm_icomp_io_range *dm_icomp_create_io_range(
- struct dm_icomp_req *req, int comp_len)
+ struct dm_icomp_req *req)
{
struct dm_icomp_io_range *io;
@@ -894,120 +932,67 @@ static struct dm_icomp_io_range *dm_icomp_create_io_range(
if (!io)
return NULL;
- io->comp_data = dm_icomp_kmalloc(comp_len, GFP_NOIO);
- if (!io->comp_data) {
- kmem_cache_free(dm_icomp_io_range_cachep, io);
- return NULL;
- }
-
io->io_req.notify.fn = dm_icomp_io_range_done;
io->io_req.notify.context = io;
io->io_req.client = req->info->io_client;
io->io_req.mem.type = DM_IO_KMEM;
- io->io_req.mem.ptr.addr = io->comp_data;
io->io_req.mem.offset = 0;
io->io_region.bdev = req->info->dev->bdev;
-
- io->comp_len = comp_len;
io->req = req;
- io->decomp_data = NULL;
- io->decomp_real_data = NULL;
- io->decomp_len = 0;
- io->decomp_kmap = false;
- io->decomp_req_len = 0;
+ io->comp_data = io->decomp_data = io->decomp_real_data = NULL;
+ io->comp_len = io->decomp_len = io->decomp_req_len = 0;
+ io->comp_kmap = io->decomp_kmap = false;
return io;
}
static struct dm_icomp_io_range *dm_icomp_create_io_read_range(
struct dm_icomp_req *req, int comp_len, int decomp_len)
{
- struct dm_icomp_io_range *io = dm_icomp_create_io_range(req, comp_len);
-
- if (io) {
- /* note down the requested length for decompress buffer.
- * but dont allocate it yet.
- */
- io->decomp_req_len = decomp_len;
- }
- return io;
-}
+ struct bio *bio = req->bio;
+ sector_t size = bio_sectors(bio)<<9;
+ int segments = bio_segments(bio);
+ void *addr;
+ struct dm_icomp_io_range *io = dm_icomp_create_io_range(req);
-static int dm_icomp_update_io_read_range(struct dm_icomp_io_range *io,
- struct bio *bio, ssize_t bio_off)
-{
- struct bvec_iter iter;
- struct bio_vec bv;
- bool just_use = false;
+ if (!io)
+ return NULL;
- if (io->decomp_len)
- return 0;
+ /* try reusing the bio buffer for compress data. */
+ if (segments == 1) {
+ struct bio_vec bv = bio_iovec(bio);
- /* use a bio buffer long enough to hold the uncompressed data */
- bio_for_each_segment(bv, bio, iter) {
- int avail_len;
- int length = bv.bv_len;
+ WARN_ON(size < comp_len);
+ addr = kmap(bv.bv_page);
+ } else
+ addr = dm_icomp_kmalloc(comp_len, GFP_NOIO);
- if (!just_use && bio_off >= length) {
- bio_off -= length;
- continue;
- }
- avail_len = just_use ? length : length-bio_off;
- if (avail_len >= io->decomp_req_len) {
- io->decomp_real_data = kmap(bv.bv_page);
- io->decomp_data = io->decomp_real_data + bio_off;
- io->decomp_len = io->decomp_req_len = avail_len;
- io->decomp_kmap = true;
- return 0;
- }
- just_use = true;
+ if (!addr) {
+ kmem_cache_free(dm_icomp_io_range_cachep, io);
+ return NULL;
}
+ io->comp_data = io->io_req.mem.ptr.addr = addr;
+ io->comp_len = comp_len;
+ io->comp_kmap = (segments == 1);
+ /* note down the requested length for decompress buffer.
+ * but dont allocate it yet.
+ /
+ io->decomp_req_len = decomp_len;
+ return io;
+}
- /* none available. :( Allocate one */
+static int dm_icomp_update_io_read_range(struct dm_icomp_io_range *io)
+{
io->decomp_data = dm_icomp_kmalloc(io->decomp_req_len, GFP_NOIO);
if (!io->decomp_data)
return 1;
io->decomp_real_data = io->decomp_data;
io->decomp_len = io->decomp_req_len;
io->decomp_kmap = false;
-
return 0;
}
-static void dm_icomp_bio_copy(struct bio *bio, off_t bio_off, void *buf,
- ssize_t len, bool to_buf)
-{
- struct bio_vec bv;
- struct bvec_iter iter;
- off_t buf_off = 0;
- ssize_t size;
- void *addr;
-
- WARN_ON(bio_off + len > (bio_sectors(bio) << 9));
-
- bio_for_each_segment(bv, bio, iter) {
- int length = bv.bv_len;
-
- if (bio_off >= length) {
- bio_off -= length;
- continue;
- }
- addr = kmap_atomic(bv.bv_page);
- size = min_t(ssize_t, len, length - bio_off);
- if (to_buf)
- memcpy(buf + buf_off, addr + bio_off + bv.bv_offset,
- size);
- else
- memcpy(addr + bio_off + bv.bv_offset, buf + buf_off,
- size);
- kunmap_atomic(addr);
- bio_off = 0;
- buf_off += size;
- len -= size;
- }
-}
-
static int dm_icomp_mod_to_max_io_range(struct dm_icomp_info *info,
struct dm_icomp_io_range *io)
{
@@ -1030,13 +1015,26 @@ static int dm_icomp_mod_to_max_io_range(struct dm_icomp_info *info,
static struct dm_icomp_io_range *dm_icomp_create_io_write_range(
struct dm_icomp_req *req)
{
- struct dm_icomp_io_range *io;
struct bio *bio = req->bio;
sector_t size = bio_sectors(req->bio)<<9;
int segments = bio_segments(bio);
int comp_len = dm_icomp_compressor_len(req->info, size);
void *addr;
+ struct dm_icomp_io_range *io = dm_icomp_create_io_range(req);
+
+ if (!io)
+ return NULL;
+
+ addr = dm_icomp_kmalloc(comp_len, GFP_NOIO);
+ if (!addr) {
+ kmem_cache_free(dm_icomp_io_range_cachep, io);
+ return NULL;
+ }
+ io->comp_data = io->io_req.mem.ptr.addr = addr;
+ io->comp_len = comp_len;
+ io->comp_kmap = false;
+ /* try reusing the bio buffer for decomp data. */
if (segments == 1) {
struct bio_vec bv = bio_iovec(bio);
@@ -1044,21 +1042,18 @@ static struct dm_icomp_io_range *dm_icomp_create_io_write_range(
} else
addr = dm_icomp_kmalloc(size, GFP_NOIO);
- if (!addr)
- return NULL;
-
- io = dm_icomp_create_io_range(req, comp_len);
- if (!io) {
- (segments == 1) ? kunmap(addr) : dm_icomp_kfree(addr, size);
+ if (!addr) {
+ dm_icomp_kfree(io->comp_data, comp_len);
+ kmem_cache_free(dm_icomp_io_range_cachep, io);
return NULL;
}
-
io->decomp_data = io->decomp_real_data = addr;
io->decomp_len = size;
io->decomp_kmap = (segments == 1);
if (!io->decomp_kmap)
dm_icomp_bio_copy(req->bio, 0, io->decomp_data, size, true);
+
return io;
}
@@ -1176,7 +1171,7 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
src_off = (req->bio->bi_iter.bi_sector -
io->io_region.sector) << 9;
- if (dm_icomp_update_io_read_range(io, req->bio, dst_off)) {
+ if (dm_icomp_update_io_read_range(io)) {
req->result = -EIO;
return;
}
diff --git a/drivers/md/dm-inplace-compress.h b/drivers/md/dm-inplace-compress.h
index e144e96..775ccbf 100644
--- a/drivers/md/dm-inplace-compress.h
+++ b/drivers/md/dm-inplace-compress.h
@@ -120,6 +120,7 @@ struct dm_icomp_io_range {
void *decomp_real_data; /* holds the actual start of the buffer */
unsigned int decomp_req_len;/* originally requested length */
unsigned int decomp_len; /* actual allocated/mapped length */
+ int comp_kmap; /* Is the comp_data kmapped'? */
void *comp_data;
unsigned int comp_len; /* For write, this is estimated */
struct list_head next;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 12/16] DM: release unneeded buffer as soon as possible.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (10 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 11/16] DM: Try to avoid temporary buffer allocation to hold compressed data Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 13/16] DM: macros to set and get the state of the request Ram Pai
` (3 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
Done to conserve as much free space as possible. Waiting to release till the
bio is done will unneccesarily hog up memory.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 14 +++++++++++---
1 files changed, 11 insertions(+), 3 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index bc1cf70..b1c3e5f 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -977,7 +977,7 @@ static struct dm_icomp_io_range *dm_icomp_create_io_read_range(
io->comp_kmap = (segments == 1);
/* note down the requested length for decompress buffer.
* but dont allocate it yet.
- /
+ */
io->decomp_req_len = decomp_len;
return io;
}
@@ -1375,8 +1375,13 @@ static int dm_icomp_handle_write_modify(struct dm_icomp_io_range *io,
}
dm_icomp_get_req(req);
- if (ret == 1)
+
+ if (ret == 1) {
io->io_req.mem.ptr.addr = io->decomp_data + (offset << 9);
+ dm_icomp_release_comp_buffer(io);
+ } else
+ dm_icomp_release_decomp_buffer(io);
+
io->io_region.count = comp_len >> 9;
io->io_region.sector = start + req->info->data_start;
@@ -1443,8 +1448,11 @@ static void dm_icomp_handle_write_comp(struct dm_icomp_req *req)
io->io_region.sector = req->bio->bi_iter.bi_sector +
req->info->data_start;
- if (ret == 1)
+ if (ret == 1) {
io->io_req.mem.ptr.addr = io->decomp_data;
+ dm_icomp_release_comp_buffer(io);
+ } else
+ dm_icomp_release_decomp_buffer(io);
io->io_region.count = comp_len >> 9;
io->io_req.bi_rw = req->bio->bi_rw;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 13/16] DM: macros to set and get the state of the request.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (11 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 12/16] DM: release unneeded buffer as soon as possible Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 14/16] DM: Wasted bio copy Ram Pai
` (2 subsequent siblings)
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
Consolidate code to set and get the state of a request.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 31 +++++++++++++++++--------------
1 files changed, 17 insertions(+), 14 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index b1c3e5f..55a515b 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -44,6 +44,9 @@ static struct kernel_param_ops dm_icomp_compressor_param_ops = {
module_param_cb(compress_algorithm, &dm_icomp_compressor_param_ops,
&dm_icomp_compressor_kparam, 0644);
+#define SET_REQ_STAGE(req, value) (req->stage = value)
+#define GET_REQ_STAGE(req) req->stage
+
static int dm_icomp_get_compressor(const char *s)
{
int r, val_len;
@@ -835,15 +838,15 @@ static void dm_icomp_put_req(struct dm_icomp_req *req)
if (atomic_dec_return(&req->io_pending))
return;
- if (req->stage == STAGE_INIT) /* waiting for locking */
+ if (GET_REQ_STAGE(req) == STAGE_INIT) /* waiting for locking */
return;
- if (req->stage == STAGE_READ_DECOMP ||
- req->stage == STAGE_WRITE_COMP ||
+ if (GET_REQ_STAGE(req) == STAGE_READ_DECOMP ||
+ GET_REQ_STAGE(req) == STAGE_WRITE_COMP ||
req->result)
- req->stage = STAGE_DONE;
+ SET_REQ_STAGE(req, STAGE_DONE);
- if (req->stage != STAGE_DONE) {
+ if (GET_REQ_STAGE(req) != STAGE_DONE) {
dm_icomp_queue_req(req->info, req);
return;
}
@@ -1153,7 +1156,7 @@ static void dm_icomp_handle_read_decomp(struct dm_icomp_req *req)
off_t bio_off = 0;
int ret;
- req->stage = STAGE_READ_DECOMP;
+ SET_REQ_STAGE(req, STAGE_READ_DECOMP);
if (req->result)
return;
@@ -1250,7 +1253,7 @@ static void dm_icomp_handle_read_read_existing(struct dm_icomp_req *req)
u64 block_index, first_block_index;
u16 logical_sectors, data_sectors;
- req->stage = STAGE_READ_EXISTING;
+ SET_REQ_STAGE(req, STAGE_READ_EXISTING);
block_index = dm_icomp_sector_to_block(req->bio->bi_iter.bi_sector);
again:
@@ -1279,14 +1282,14 @@ static void dm_icomp_handle_read_request(struct dm_icomp_req *req)
{
dm_icomp_get_req(req);
- if (req->stage == STAGE_INIT) {
+ if (GET_REQ_STAGE(req) == STAGE_INIT) {
if (!dm_icomp_lock_req_range(req)) {
dm_icomp_put_req(req);
return;
}
dm_icomp_handle_read_read_existing(req);
- } else if (req->stage == STAGE_READ_EXISTING)
+ } else if (GET_REQ_STAGE(req) == STAGE_READ_EXISTING)
dm_icomp_handle_read_decomp(req);
dm_icomp_put_req(req);
@@ -1411,7 +1414,7 @@ static void dm_icomp_handle_write_comp(struct dm_icomp_req *req)
int ret;
bool handle_bio = true;
- req->stage = STAGE_WRITE_COMP;
+ SET_REQ_STAGE(req, STAGE_WRITE_COMP);
if (req->result)
return;
@@ -1486,7 +1489,7 @@ static void dm_icomp_handle_write_read_existing(struct dm_icomp_req *req)
u64 block_index, first_block_index;
u16 logical_sectors, data_sectors;
- req->stage = STAGE_READ_EXISTING;
+ SET_REQ_STAGE(req, STAGE_READ_EXISTING);
block_index = dm_icomp_sector_to_block(req->bio->bi_iter.bi_sector);
dm_icomp_get_extent(req->info, block_index, &first_block_index,
@@ -1524,14 +1527,14 @@ static void dm_icomp_handle_write_request(struct dm_icomp_req *req)
{
dm_icomp_get_req(req);
- if (req->stage == STAGE_INIT) {
+ if (GET_REQ_STAGE(req) == STAGE_INIT) {
if (!dm_icomp_lock_req_range(req)) {
dm_icomp_put_req(req);
return;
}
dm_icomp_handle_write_read_existing(req);
- } else if (req->stage == STAGE_READ_EXISTING)
+ } else if (GET_REQ_STAGE(req) == STAGE_READ_EXISTING)
dm_icomp_handle_write_comp(req);
dm_icomp_put_req(req);
@@ -1611,7 +1614,7 @@ static int dm_icomp_map(struct dm_target *ti, struct bio *bio)
atomic_set(&req->io_pending, 0);
INIT_LIST_HEAD(&req->all_io);
req->result = 0;
- req->stage = STAGE_INIT;
+ SET_REQ_STAGE(req, STAGE_INIT);
req->locked_locks = 0;
req->cpu = raw_smp_processor_id();
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 14/16] DM: Wasted bio copy.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (12 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 13/16] DM: macros to set and get the state of the request Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 15/16] DM: Add sysfs parameters to track total memory saved and allocated Ram Pai
2016-08-15 17:36 ` [RFC PATCH 16/16] DM: add documentation for dm-inplace-compress Ram Pai
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
No point doing it.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index 55a515b..31b144b 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -1434,7 +1434,6 @@ static void dm_icomp_handle_write_comp(struct dm_icomp_req *req)
req->result = -EIO;
return;
}
- dm_icomp_bio_copy(req->bio, 0, io->decomp_data, count << 9, true);
/* compress data */
comp_len = io->comp_len;
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 15/16] DM: Add sysfs parameters to track total memory saved and allocated.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (13 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 14/16] DM: Wasted bio copy Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
2016-08-15 17:36 ` [RFC PATCH 16/16] DM: add documentation for dm-inplace-compress Ram Pai
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
Add parameters to monitor the memory efficiency of the module.
dm_icomp_total_alloc_size: total memory currently in use.
dm_icomp_total_bio_save: total memory allocation saved by the optimizations.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
drivers/md/dm-inplace-compress.c | 36 ++++++++++++++++++++++++++++++++++--
1 files changed, 34 insertions(+), 2 deletions(-)
diff --git a/drivers/md/dm-inplace-compress.c b/drivers/md/dm-inplace-compress.c
index 31b144b..0a7790a 100644
--- a/drivers/md/dm-inplace-compress.c
+++ b/drivers/md/dm-inplace-compress.c
@@ -82,6 +82,23 @@ static int dm_icomp_compressor_param_set(const char *val,
return 0;
}
+static const struct kernel_param_ops dm_icomp_alloc_param_ops = {
+ .set = param_set_ulong,
+ .get = param_get_ulong,
+};
+
+static atomic64_t dm_icomp_total_alloc_size;
+#define DMCP_ALLOC(s) {atomic64_add(s, &dm_icomp_total_alloc_size); }
+#define DMCP_FREE_ALLOC(s) {atomic64_sub(s, &dm_icomp_total_alloc_size); }
+module_param_cb(dm_icomp_total_alloc_size, &dm_icomp_alloc_param_ops,
+ &dm_icomp_total_alloc_size, 0644);
+
+static atomic64_t dm_icomp_total_bio_save;
+#define DMCP_ALLOC_SAVE(s) {atomic64_add(s, &dm_icomp_total_bio_save); }
+module_param_cb(dm_icomp_total_bio_save, &dm_icomp_alloc_param_ops,
+ &dm_icomp_total_bio_save, 0644);
+
+
static struct kmem_cache *dm_icomp_req_cachep;
static struct kmem_cache *dm_icomp_io_range_cachep;
static struct kmem_cache *dm_icomp_meta_io_cachep;
@@ -616,6 +633,9 @@ static int dm_icomp_ctr(struct dm_target *ti, unsigned int argc, char **argv)
atomic64_set(&info->compressed_write_size, 0);
atomic64_set(&info->uncompressed_write_size, 0);
atomic64_set(&info->meta_write_size, 0);
+ atomic64_set(&dm_icomp_total_alloc_size, 0);
+ atomic64_set(&dm_icomp_total_bio_save, 0);
+
ti->num_flush_bios = 1;
/* ti->num_discard_bios = 1; */
ti->private = info;
@@ -781,18 +801,28 @@ static void dm_icomp_get_req(struct dm_icomp_req *req)
static void *dm_icomp_kmalloc(size_t size, gfp_t flags)
{
- return kmalloc(size, flags);
+ void *addr = kmalloc(size, flags);
+
+ if (addr)
+ DMCP_ALLOC(size);
+ return addr;
}
static void *dm_icomp_krealloc(void *addr, size_t size,
size_t orig_size, gfp_t flags)
{
- return krealloc(addr, size, flags);
+ void *taddr = krealloc(addr, size, flags);
+
+ DMCP_FREE_ALLOC(orig_size);
+ if (taddr)
+ DMCP_ALLOC(size);
+ return taddr;
}
static void dm_icomp_kfree(void *addr, unsigned int size)
{
kfree(addr);
+ DMCP_FREE_ALLOC(size);
}
static void dm_icomp_release_decomp_buffer(struct dm_icomp_io_range *io)
@@ -1056,6 +1086,8 @@ static struct dm_icomp_io_range *dm_icomp_create_io_write_range(
io->decomp_kmap = (segments == 1);
if (!io->decomp_kmap)
dm_icomp_bio_copy(req->bio, 0, io->decomp_data, size, true);
+ else
+ DMCP_ALLOC_SAVE(size);
return io;
}
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread* [RFC PATCH 16/16] DM: add documentation for dm-inplace-compress.
2016-08-15 17:36 [RFC PATCH 00/16] dm-inplace-compression block device Ram Pai
` (14 preceding siblings ...)
2016-08-15 17:36 ` [RFC PATCH 15/16] DM: Add sysfs parameters to track total memory saved and allocated Ram Pai
@ 2016-08-15 17:36 ` Ram Pai
15 siblings, 0 replies; 17+ messages in thread
From: Ram Pai @ 2016-08-15 17:36 UTC (permalink / raw)
To: LKML, linux-raid, dm-devel, linux-doc; +Cc: shli, agk, snitzer, corbet, Ram Pai
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
---
.../device-mapper/dm-inplace-compress.text | 138 ++++++++++++++++++++
1 files changed, 138 insertions(+), 0 deletions(-)
create mode 100644 Documentation/device-mapper/dm-inplace-compress.text
diff --git a/Documentation/device-mapper/dm-inplace-compress.text b/Documentation/device-mapper/dm-inplace-compress.text
new file mode 100644
index 0000000..c31e69e
--- /dev/null
+++ b/Documentation/device-mapper/dm-inplace-compress.text
@@ -0,0 +1,138 @@
+dm-inplace-compress
+====================
+
+Device-Mapper's "inplace-compress" target provides inplace compression of block
+devices using the kernel compression API.
+
+Parameters: <device path> \
+ [ <#opt_params writethough> ]
+ [ <#opt_params <writeback> <meta_commit_delay> ]
+ [ <#opt_params compressor> <type> ]
+
+
+<writethrough>
+ Write data and metadata together.
+
+<writeback> <meta_commit_delay>
+ Write metadata every 'meta_commit_delay' interval.
+
+<device path>
+ This is the device that is going to be used as backend and contains the
+ compressed data. You can specify it as a path like /dev/xxx or a device
+ number <major>:<minor>.
+
+<compressor> <type>
+ Choose the compressor algorithm. 'lzo' and '842'
+ compressors are supported.
+
+Example scripts
+===============
+
+create a inplace-compress block device using lzo compression. Write metadata
+and data together.
+[[
+#!/bin/sh
+# Create a inplace-compress device using dmsetup
+dmsetup create comp1 --table "0 `blockdev --getsize $1` inplacecompress $1
+ writethrough compressor lzo"
+]]
+
+
+create a inplace-compress block device using nx-842 hardware compression. Write
+metadata periodially every 5sec.
+
+[[
+#!/bin/sh
+# Create a inplace-compress device using dmsetup
+dmsetup create comp1 --table "0 `blockdev --getsize $1` inplacecompress $1
+ writeback 5 compressor 842"
+]]
+
+Description
+===========
+ This is a simple DM target supporting inplace compression. Its best suited for
+ SSD. The underlying disk must support 512B sector size, the target only
+ supports 4k sector size.
+
+ Disk layout:
+ |super|...meta...|..data...|
+
+ Store unit is 4k (a block). Super is 1 block, which stores meta and data
+ size and compression algorithm. Meta is a bitmap. For each data block,
+ there are 5 bits meta.
+
+ Data:
+
+ Data of a block is compressed. Compressed data is round up to 512B, which
+ is the payload. In disk, payload is stored at the beginning of logical
+ sector of the block. Let's look at an example. Say we store data to block
+ A, which is in sector B(A*8), its orginal size is 4k, compressed size is
+ 1500. Compressed data (CD) will use 3 sectors (512B). The 3 sectors are the
+ payload. Payload will be stored at sector B.
+
+ ---------------------------------------------------
+ ... | CD1 | CD2 | CD3 | | | | | | ...
+ ---------------------------------------------------
+ ^B ^B+1 ^B+2 ^B+7 ^B+8
+
+ For this block, we will not use sector B+3 to B+7 (a hole). We use 4 meta
+ bits to present payload size. The compressed size (1500) isn't stored in
+ meta directly. Instead, we store it at the last 32bits of payload. In this
+ example, we store it at the end of sector B+2. If compressed size +
+ sizeof(32bits) crosses a sector, payload size will increase one sector. If
+ payload uses 8 sectors, we store uncompressed data directly.
+
+ If IO size is bigger than one block, we can store the data as an extent.
+ Data of the whole extent will compressed and stored in the similar way like
+ above. The first block of the extent is the head, all others are the tail.
+ If extent is 1 block, the block is head. We have 1 bit of meta to present
+ if a block is head or tail. If 4 meta bits of head block can't store extent
+ payload size, we will borrow tail block meta bits to store payload size.
+ Max allowd extent size is 128k, so we don't compress/decompress too big
+ size data.
+
+ Meta:
+ Modifying data will modify meta too. Meta will be written(flush) to disk
+ depending on meta write policy. We support writeback and writethrough mode.
+ In writeback mode, meta will be written to disk in an interval or a FLUSH
+ request. In writethrough mode, data and meta data will be written to disk
+ together.
+
+ Advantages:
+
+ 1. Simple. Since we store compressed data in-place, we don't need complicated
+ disk data management.
+ 2. Efficient. For each 4k, we only need 5 bits meta. 1T data will use less than
+ 200M meta, so we can load all meta into memory. And actual compression size is
+ in payload. So if IO doesn't need RMW and we use writeback meta flush, we don't
+ need extra IO for meta.
+
+ Disadvantages:
+
+ 1. hole. Since we store compressed data in-place, there are a lot of holes
+ (in above example, B+3 - B+7) Hole can impact IO, because we can't do IO
+ merge.
+
+ 2. 1:1 size. Compression doesn't change disk size. If disk is 1T, we can
+ only store 1T data even we do compression.
+
+ But this is for SSD only. Generally SSD firmware has a FTL layer to map
+ disk sectors to flash nand. High end SSD firmware has filesystem-like FTL.
+
+ 1. hole. Disk has a lot of holes, but SSD FTL can still store data continuous
+ in nand. Even if we can't do IO merge in OS layer, SSD firmware can do it.
+
+ 2. 1:1 size. On one side, we write compressed data to SSD, which means less
+ data is written to SSD. This will be very helpful to improve SSD garbage
+ collection, and so write speed and life cycle. So even this is a problem, the
+ target is still helpful. On the other side, advanced SSD FTL can easily do thin
+ provision. For example, if nand is 1T and we let SSD report it as 2T, and use
+ the SSD as compressed target. In such SSD, we don't have the 1:1 size issue.
+
+ So even if SSD FTL cannot map non-continuous disk sectors to continuous nand,
+ the compression target can still function well.
+
+
+Author:
+ Shaohua Li <shli@fusionio.com>
+ Ram Pai <ram.n.pai@gmail.com>
--
1.7.1
^ permalink raw reply related [flat|nested] 17+ messages in thread