From da9953b729c12ece6d35fd15d236457eee679228 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Thu, 2 Apr 2020 09:32:35 -0700 Subject: f2fs: introduce sysfs/data_io_flag to attach REQ_META/FUA This patch introduces a way to attach REQ_META/FUA explicitly to all the data writes given temperature. -> attach REQ_FUA to Hot Data writes -> attach REQ_FUA to Hot|Warm Data writes -> attach REQ_FUA to Hot|Warm|Cold Data writes -> attach REQ_FUA to Hot|Warm|Cold Data writes as well as REQ_META to Hot Data writes Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'Documentation') diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index bd8a0d19abe6..c8620ea7022a 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -323,3 +323,12 @@ What: /sys/fs/f2fs//mounted_time_sec Date: February 2020 Contact: "Jaegeuk Kim" Description: Show the mounted time in secs of this partition. + +What: /sys/fs/f2fs//data_io_flag +Date: April 2020 +Contact: "Jaegeuk Kim" +Description: Give a way to attach REQ_META|FUA to data writes + given temperature-based bits. Now the bits indicate: + * REQ_META | REQ_FUA | + * 5 | 4 | 3 | 2 | 1 | 0 | + * Cold | Warm | Hot | Cold | Warm | Hot | -- cgit v1.2.3 From 2bc4bea33848e4c5f441fcec4a94497082c1cc9f Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 30 Mar 2020 03:30:59 +0000 Subject: f2fs: add tracepoint for f2fs iostat Added a tracepoint to see iostat of f2fs. Default period of that is 3 second. This tracepoint can be used to be monitoring I/O statistics periodically. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 6 ++++ fs/f2fs/f2fs.h | 16 +++++++++- fs/f2fs/super.c | 1 + fs/f2fs/sysfs.c | 39 +++++++++++++++++++++++++ include/trace/events/f2fs.h | 52 +++++++++++++++++++++++++++++++++ 5 files changed, 113 insertions(+), 1 deletion(-) (limited to 'Documentation') diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index c8620ea7022a..427f5b45c67f 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -332,3 +332,9 @@ Description: Give a way to attach REQ_META|FUA to data writes * REQ_META | REQ_FUA | * 5 | 4 | 3 | 2 | 1 | 0 | * Cold | Warm | Hot | Cold | Warm | Hot | + +What: /sys/fs/f2fs//iostat_period_ms +Date: April 2020 +Contact: "Daeho Jeong" +Description: Give a way to change iostat_period time. 3secs by default. + The new iostat trace gives stats gap given the period. diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index c2788738aa0d..6cedbfb2067c 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1505,7 +1505,10 @@ struct f2fs_sb_info { /* For app/fs IO statistics */ spinlock_t iostat_lock; unsigned long long write_iostat[NR_IO_TYPE]; + unsigned long long prev_write_iostat[NR_IO_TYPE]; bool iostat_enable; + unsigned long iostat_next_period; + unsigned int iostat_period_ms; /* to attach REQ_META|REQ_FUA flags */ unsigned int data_io_flag; @@ -2999,16 +3002,25 @@ static inline int get_inline_xattr_addrs(struct inode *inode) sizeof((f2fs_inode)->field)) \ <= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize))) \ +#define DEFAULT_IOSTAT_PERIOD_MS 3000 +#define MIN_IOSTAT_PERIOD_MS 100 +/* maximum period of iostat tracing is 1 day */ +#define MAX_IOSTAT_PERIOD_MS 8640000 + static inline void f2fs_reset_iostat(struct f2fs_sb_info *sbi) { int i; spin_lock(&sbi->iostat_lock); - for (i = 0; i < NR_IO_TYPE; i++) + for (i = 0; i < NR_IO_TYPE; i++) { sbi->write_iostat[i] = 0; + sbi->prev_write_iostat[i] = 0; + } spin_unlock(&sbi->iostat_lock); } +extern void f2fs_record_iostat(struct f2fs_sb_info *sbi); + static inline void f2fs_update_iostat(struct f2fs_sb_info *sbi, enum iostat_type type, unsigned long long io_bytes) { @@ -3022,6 +3034,8 @@ static inline void f2fs_update_iostat(struct f2fs_sb_info *sbi, sbi->write_iostat[APP_WRITE_IO] - sbi->write_iostat[APP_DIRECT_IO]; spin_unlock(&sbi->iostat_lock); + + f2fs_record_iostat(sbi); } #define __is_large_section(sbi) ((sbi)->segs_per_sec > 1) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index f2dfc21c6abb..438296e17183 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -3424,6 +3424,7 @@ try_onemore: /* init iostat info */ spin_lock_init(&sbi->iostat_lock); sbi->iostat_enable = false; + sbi->iostat_period_ms = DEFAULT_IOSTAT_PERIOD_MS; for (i = 0; i < NR_PAGE_TYPE; i++) { int n = (i == META) ? 1: NR_TEMP_TYPE; diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index aeebfb5024a2..d05cb68c2637 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -15,6 +15,7 @@ #include "f2fs.h" #include "segment.h" #include "gc.h" +#include static struct proc_dir_entry *f2fs_proc_root; @@ -379,6 +380,15 @@ out: return count; } + if (!strcmp(a->attr.name, "iostat_period_ms")) { + if (t < MIN_IOSTAT_PERIOD_MS || t > MAX_IOSTAT_PERIOD_MS) + return -EINVAL; + spin_lock(&sbi->iostat_lock); + sbi->iostat_period_ms = (unsigned int)t; + spin_unlock(&sbi->iostat_lock); + return count; + } + *ui = (unsigned int)t; return count; @@ -535,6 +545,7 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_idle_interval, interval_time[GC_TIME]); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, umount_discard_timeout, interval_time[UMOUNT_DISCARD_TIMEOUT]); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, iostat_enable, iostat_enable); +F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, iostat_period_ms, iostat_period_ms); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, readdir_ra, readdir_ra); F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_pin_file_thresh, gc_pin_file_threshold); F2FS_RW_ATTR(F2FS_SBI, f2fs_super_block, extension_list, extension_list); @@ -615,6 +626,7 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(gc_idle_interval), ATTR_LIST(umount_discard_timeout), ATTR_LIST(iostat_enable), + ATTR_LIST(iostat_period_ms), ATTR_LIST(readdir_ra), ATTR_LIST(gc_pin_file_thresh), ATTR_LIST(extension_list), @@ -751,6 +763,33 @@ static int __maybe_unused segment_bits_seq_show(struct seq_file *seq, return 0; } +void f2fs_record_iostat(struct f2fs_sb_info *sbi) +{ + unsigned long long iostat_diff[NR_IO_TYPE]; + int i; + + if (time_is_after_jiffies(sbi->iostat_next_period)) + return; + + /* Need double check under the lock */ + spin_lock(&sbi->iostat_lock); + if (time_is_after_jiffies(sbi->iostat_next_period)) { + spin_unlock(&sbi->iostat_lock); + return; + } + sbi->iostat_next_period = jiffies + + msecs_to_jiffies(sbi->iostat_period_ms); + + for (i = 0; i < NR_IO_TYPE; i++) { + iostat_diff[i] = sbi->write_iostat[i] - + sbi->prev_write_iostat[i]; + sbi->prev_write_iostat[i] = sbi->write_iostat[i]; + } + spin_unlock(&sbi->iostat_lock); + + trace_f2fs_iostat(sbi, iostat_diff); +} + static int __maybe_unused iostat_info_seq_show(struct seq_file *seq, void *offset) { diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h index d97adfc327f0..e78c8696e2ad 100644 --- a/include/trace/events/f2fs.h +++ b/include/trace/events/f2fs.h @@ -1812,6 +1812,58 @@ DEFINE_EVENT(f2fs_zip_end, f2fs_decompress_pages_end, TP_ARGS(inode, cluster_idx, compressed_size, ret) ); +TRACE_EVENT(f2fs_iostat, + + TP_PROTO(struct f2fs_sb_info *sbi, unsigned long long *iostat), + + TP_ARGS(sbi, iostat), + + TP_STRUCT__entry( + __field(dev_t, dev) + __field(unsigned long long, app_dio) + __field(unsigned long long, app_bio) + __field(unsigned long long, app_wio) + __field(unsigned long long, app_mio) + __field(unsigned long long, fs_dio) + __field(unsigned long long, fs_nio) + __field(unsigned long long, fs_mio) + __field(unsigned long long, fs_gc_dio) + __field(unsigned long long, fs_gc_nio) + __field(unsigned long long, fs_cp_dio) + __field(unsigned long long, fs_cp_nio) + __field(unsigned long long, fs_cp_mio) + __field(unsigned long long, fs_discard) + ), + + TP_fast_assign( + __entry->dev = sbi->sb->s_dev; + __entry->app_dio = iostat[APP_DIRECT_IO]; + __entry->app_bio = iostat[APP_BUFFERED_IO]; + __entry->app_wio = iostat[APP_WRITE_IO]; + __entry->app_mio = iostat[APP_MAPPED_IO]; + __entry->fs_dio = iostat[FS_DATA_IO]; + __entry->fs_nio = iostat[FS_NODE_IO]; + __entry->fs_mio = iostat[FS_META_IO]; + __entry->fs_gc_dio = iostat[FS_GC_DATA_IO]; + __entry->fs_gc_nio = iostat[FS_GC_NODE_IO]; + __entry->fs_cp_dio = iostat[FS_CP_DATA_IO]; + __entry->fs_cp_nio = iostat[FS_CP_NODE_IO]; + __entry->fs_cp_mio = iostat[FS_CP_META_IO]; + __entry->fs_discard = iostat[FS_DISCARD]; + ), + + TP_printk("dev = (%d,%d), " + "app [write=%llu (direct=%llu, buffered=%llu), mapped=%llu], " + "fs [data=%llu, node=%llu, meta=%llu, discard=%llu], " + "gc [data=%llu, node=%llu], " + "cp [data=%llu, node=%llu, meta=%llu]", + show_dev(__entry->dev), __entry->app_wio, __entry->app_dio, + __entry->app_bio, __entry->app_mio, __entry->fs_dio, + __entry->fs_nio, __entry->fs_mio, __entry->fs_discard, + __entry->fs_gc_dio, __entry->fs_gc_nio, __entry->fs_cp_dio, + __entry->fs_cp_nio, __entry->fs_cp_mio) +); + #endif /* _TRACE_F2FS_H */ /* This part must be outside protection */ -- cgit v1.2.3 From 6d92b201035dfe77426f8814fd5259db385a18b3 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 8 Apr 2020 19:56:32 +0800 Subject: f2fs: compress: support lzo-rle compress algorithm LZO-RLE extension (run length encoding) was introduced to improve performance of LZO algorithm in scenario of data contains many zeros, zram has changed to use this extended algorithm by default, this patch adds to support this algorithm extension, to enable this extension, it needs to enable F2FS_FS_LZO and F2FS_FS_LZORLE config, and specifies "compress_algorithm=lzo-rle" mountoption. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/filesystems/f2fs.rst | 2 +- fs/f2fs/Kconfig | 10 ++++++++++ fs/f2fs/compress.c | 30 ++++++++++++++++++++++++++++++ fs/f2fs/f2fs.h | 1 + fs/f2fs/super.c | 6 ++++++ include/trace/events/f2fs.h | 3 ++- 6 files changed, 50 insertions(+), 2 deletions(-) (limited to 'Documentation') diff --git a/Documentation/filesystems/f2fs.rst b/Documentation/filesystems/f2fs.rst index 87d794bc75a4..96e3f89d7aba 100644 --- a/Documentation/filesystems/f2fs.rst +++ b/Documentation/filesystems/f2fs.rst @@ -244,7 +244,7 @@ checkpoint=%s[:%u[%]] Set to "disable" to turn off checkpointing. Set to "enabl would be unusable can be viewed at /sys/fs/f2fs//unusable This space is reclaimed once checkpoint=enable. compress_algorithm=%s Control compress algorithm, currently f2fs supports "lzo", - "lz4" and "zstd" algorithm. + "lz4", "zstd" and "lzo-rle" algorithm. compress_log_size=%u Support configuring compress cluster size, the size will be 4KB * (1 << %u), 16KB is minimum size, also it's default size. diff --git a/fs/f2fs/Kconfig b/fs/f2fs/Kconfig index bb68d21e1f8c..d13c5c6a9787 100644 --- a/fs/f2fs/Kconfig +++ b/fs/f2fs/Kconfig @@ -127,3 +127,13 @@ config F2FS_FS_ZSTD default y help Support ZSTD compress algorithm, if unsure, say Y. + +config F2FS_FS_LZORLE + bool "LZO-RLE compression support" + depends on F2FS_FS_COMPRESSION + depends on F2FS_FS_LZO + select LZO_COMPRESS + select LZO_DECOMPRESS + default y + help + Support LZO-RLE compress algorithm, if unsure, say Y. diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index 230ea7cd1510..c7c5a8f8a48c 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -442,6 +442,31 @@ static const struct f2fs_compress_ops f2fs_zstd_ops = { }; #endif +#ifdef CONFIG_F2FS_FS_LZO +#ifdef CONFIG_F2FS_FS_LZORLE +static int lzorle_compress_pages(struct compress_ctx *cc) +{ + int ret; + + ret = lzorle1x_1_compress(cc->rbuf, cc->rlen, cc->cbuf->cdata, + &cc->clen, cc->private); + if (ret != LZO_E_OK) { + printk_ratelimited("%sF2FS-fs (%s): lzo-rle compress failed, ret:%d\n", + KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id, ret); + return -EIO; + } + return 0; +} + +static const struct f2fs_compress_ops f2fs_lzorle_ops = { + .init_compress_ctx = lzo_init_compress_ctx, + .destroy_compress_ctx = lzo_destroy_compress_ctx, + .compress_pages = lzorle_compress_pages, + .decompress_pages = lzo_decompress_pages, +}; +#endif +#endif + static const struct f2fs_compress_ops *f2fs_cops[COMPRESS_MAX] = { #ifdef CONFIG_F2FS_FS_LZO &f2fs_lzo_ops, @@ -458,6 +483,11 @@ static const struct f2fs_compress_ops *f2fs_cops[COMPRESS_MAX] = { #else NULL, #endif +#if defined(CONFIG_F2FS_FS_LZO) && defined(CONFIG_F2FS_FS_LZORLE) + &f2fs_lzorle_ops, +#else + NULL, +#endif }; bool f2fs_is_compress_backend_ready(struct inode *inode) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index d098b94ca22d..0dab21e764d9 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1281,6 +1281,7 @@ enum compress_algorithm_type { COMPRESS_LZO, COMPRESS_LZ4, COMPRESS_ZSTD, + COMPRESS_LZORLE, COMPRESS_MAX, }; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 1b170fc48941..582bbf40c559 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -822,6 +822,9 @@ static int parse_options(struct super_block *sb, char *options) } else if (!strcmp(name, "zstd")) { F2FS_OPTION(sbi).compress_algorithm = COMPRESS_ZSTD; + } else if (!strcmp(name, "lzo-rle")) { + F2FS_OPTION(sbi).compress_algorithm = + COMPRESS_LZORLE; } else { kfree(name); return -EINVAL; @@ -1415,6 +1418,9 @@ static inline void f2fs_show_compress_options(struct seq_file *seq, case COMPRESS_ZSTD: algtype = "zstd"; break; + case COMPRESS_LZORLE: + algtype = "lzo-rle"; + break; } seq_printf(seq, ",compress_algorithm=%s", algtype); diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h index 417a486f5c8a..757d3d6031e6 100644 --- a/include/trace/events/f2fs.h +++ b/include/trace/events/f2fs.h @@ -154,7 +154,8 @@ TRACE_DEFINE_ENUM(CP_PAUSE); __print_symbolic(type, \ { COMPRESS_LZO, "LZO" }, \ { COMPRESS_LZ4, "LZ4" }, \ - { COMPRESS_ZSTD, "ZSTD" }) + { COMPRESS_ZSTD, "ZSTD" }, \ + { COMPRESS_LZORLE, "LZO-RLE" }) struct f2fs_sb_info; struct f2fs_io_info; -- cgit v1.2.3 From 32b6aba85c8dbd1d3a155543986574c6969f0a48 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Thu, 4 Jun 2020 11:49:43 -0700 Subject: f2fs: add node_io_flag for bio flags likewise data_io_flag This patch adds another way to attach bio flags to node writes. Description: Give a way to attach REQ_META|FUA to node writes given temperature-based bits. Now the bits indicate: * REQ_META | REQ_FUA | * 5 | 4 | 3 | 2 | 1 | 0 | * Cold | Warm | Hot | Cold | Warm | Hot | Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 9 +++++++++ fs/f2fs/data.c | 24 +++++++++++++++--------- fs/f2fs/f2fs.h | 1 + fs/f2fs/sysfs.c | 2 ++ 4 files changed, 27 insertions(+), 9 deletions(-) (limited to 'Documentation') diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index 427f5b45c67f..4bb93a06d8ab 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -333,6 +333,15 @@ Description: Give a way to attach REQ_META|FUA to data writes * 5 | 4 | 3 | 2 | 1 | 0 | * Cold | Warm | Hot | Cold | Warm | Hot | +What: /sys/fs/f2fs//node_io_flag +Date: June 2020 +Contact: "Jaegeuk Kim" +Description: Give a way to attach REQ_META|FUA to node writes + given temperature-based bits. Now the bits indicate: + * REQ_META | REQ_FUA | + * 5 | 4 | 3 | 2 | 1 | 0 | + * Cold | Warm | Hot | Cold | Warm | Hot | + What: /sys/fs/f2fs//iostat_period_ms Date: April 2020 Contact: "Daeho Jeong" diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index a65bfc07ddb9..798c325a820d 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -514,22 +514,28 @@ void f2fs_submit_bio(struct f2fs_sb_info *sbi, __submit_bio(sbi, bio, type); } -static void __attach_data_io_flag(struct f2fs_io_info *fio) +static void __attach_io_flag(struct f2fs_io_info *fio) { struct f2fs_sb_info *sbi = fio->sbi; unsigned int temp_mask = (1 << NR_TEMP_TYPE) - 1; - unsigned int fua_flag = sbi->data_io_flag & temp_mask; - unsigned int meta_flag = (sbi->data_io_flag >> NR_TEMP_TYPE) & - temp_mask; + unsigned int io_flag, fua_flag, meta_flag; + + if (fio->type == DATA) + io_flag = sbi->data_io_flag; + else if (fio->type == NODE) + io_flag = sbi->node_io_flag; + else + return; + + fua_flag = io_flag & temp_mask; + meta_flag = (io_flag >> NR_TEMP_TYPE) & temp_mask; + /* - * data io flag bits per temp: + * data/node io flag bits per temp: * REQ_META | REQ_FUA | * 5 | 4 | 3 | 2 | 1 | 0 | * Cold | Warm | Hot | Cold | Warm | Hot | */ - if (fio->type != DATA) - return; - if ((1 << fio->temp) & meta_flag) fio->op_flags |= REQ_META; if ((1 << fio->temp) & fua_flag) @@ -543,7 +549,7 @@ static void __submit_merged_bio(struct f2fs_bio_info *io) if (!io->bio) return; - __attach_data_io_flag(fio); + __attach_io_flag(fio); bio_set_op_attrs(io->bio, fio->op, fio->op_flags); if (is_read_io(fio->op)) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 7794e2f0d6e8..c812fb8e2d9c 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1568,6 +1568,7 @@ struct f2fs_sb_info { /* to attach REQ_META|REQ_FUA flags */ unsigned int data_io_flag; + unsigned int node_io_flag; /* For sysfs suppport */ struct kobject s_kobj; diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index a117ae1f9d5f..fc4a46b68904 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -554,6 +554,7 @@ F2FS_RW_ATTR(FAULT_INFO_RATE, f2fs_fault_info, inject_rate, inject_rate); F2FS_RW_ATTR(FAULT_INFO_TYPE, f2fs_fault_info, inject_type, inject_type); #endif F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, data_io_flag, data_io_flag); +F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, node_io_flag, node_io_flag); F2FS_GENERAL_RO_ATTR(dirty_segments); F2FS_GENERAL_RO_ATTR(free_segments); F2FS_GENERAL_RO_ATTR(lifetime_write_kbytes); @@ -635,6 +636,7 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(inject_type), #endif ATTR_LIST(data_io_flag), + ATTR_LIST(node_io_flag), ATTR_LIST(dirty_segments), ATTR_LIST(free_segments), ATTR_LIST(unusable), -- cgit v1.2.3