linux-f2fs-devel.lists.sourceforge.net archive mirror
 help / color / mirror / Atom feed
From: Jaegeuk Kim <jaegeuk@kernel.org>
To: Yonggil Song <yonggil.song@samsung.com>
Cc: Seokhwan Kim <sukka.kim@samsung.com>,
	Jorn Lee <lunar.lee@samsung.com>,
	"linux-f2fs-devel@lists.sourceforge.net"
	<linux-f2fs-devel@lists.sourceforge.net>
Subject: Re: [f2fs-dev] [PATCH v1] mkfs.f2fs: Introduce configurable reserved sections
Date: Mon, 27 Mar 2023 14:26:59 -0700	[thread overview]
Message-ID: <ZCIKIzPmJyjA44JK@google.com> (raw)
In-Reply-To: <20230317044151epcms2p5872591b638dab3261a9dd563253f1b94@epcms2p5>

On 03/17, Yonggil Song wrote:
> Overview
> ========
> 
> This option allows zoned block device users to configure GC reserved and
> overprovision area manually according to their demands on performance of
> sustained write latency and WAF.
> 
> Problem
> =======
> 
> The overprovision segments that mkfs generates are mostly occupied by GC
> reserved. This degrades WAF performance.
> 
> Experiment
> ==========
> 
> The following experiment evaluated the application of configurable reserved.
> The experimental environment is as follows.
> 
>   System info
>     - 4.2Ghz, 8 core CPU
>     - 64GiB Memory
>   Device info
>     - a conventional null_blk with 448MiB capacity(meta area) and
>     - a sequential null_blk with 953 zones of 64MiB
>   Format
>     - as-is (find out ovp ratio): mkfs.f2fs <conv null_blk> -c <seq null_blk> -m
>         Info: Overprovision ratio = 3.700%
>         Info: Overprovision segments = 1152 (GC reserved = 1088)
>     - config rsvd: mkfs.f2fs <conv null_blk> -c <seq null_blk> -m 8 -o 2.965
>         Info: Overprovision ratio = 2.965%
>         Info: Overprovision segments = 1152 (GC reserved = 256)
>   Mount
>     - mount <conv null_blk> <mount point>
>   Fio script
>     - fio --rw=randwrite --bs=4k --ba=4k --filesize=58630m --norandommap --overwrite=1 --name=job1 --filename=<mount point>/sustain --time_based --runtime=2h
>   WAF calculation
>     - (IOs on conv. null_blk + IOs on seq. null_blk) / random write IOs
> 
> Conclusion
> ==========
> 
> In the experiment, it can be shown that reducing the reserved segments
> decreases WAF to 10% (from 222 to 23) although it triggers checkpoint more
> frequently during gc. With direct IO, the WAF of as-is gets much higher.
> In other words, a user can configure more reserved segments for lower GC
> latency or allocate less reserved segments for lower WAF on the same number
> of OP segments.
> 
> Signed-off-by: Yonggil Song <yonggil.song@samsung.com>
> ---
>  include/f2fs_fs.h       | 22 ++++++++++++++++++++--
>  lib/libf2fs.c           | 22 ++++++++++++++++++++++
>  man/mkfs.f2fs.8         |  9 +++++++--
>  mkfs/f2fs_format.c      | 29 +++++++++++++++++++++++------
>  mkfs/f2fs_format_main.c |  5 +++--
>  5 files changed, 75 insertions(+), 12 deletions(-)
> 
> diff --git a/include/f2fs_fs.h b/include/f2fs_fs.h
> index 333ae07a5ebd..1d41e9f8397e 100644
> --- a/include/f2fs_fs.h
> +++ b/include/f2fs_fs.h
> @@ -375,6 +375,10 @@ static inline uint64_t bswap_64(uint64_t val)
>  
>  #define LPF "lost+found"
>  
> +/* one for gc buffer, the other for node */
> +#define MIN_RSVD_SECS	(uint32_t)(NR_CURSEG_TYPE + 2)
> +#define CONFIG_RSVD_DEFAULT_OP_RATIO	3.0
> +
>  enum f2fs_config_func {
>  	MKFS,
>  	FSCK,
> @@ -460,6 +464,7 @@ typedef struct {
>  #define ALIGN_UP(addrs, size)	ALIGN_DOWN(((addrs) + (size) - 1), (size))
>  
>  struct f2fs_configuration {
> +	uint32_t conf_reserved_sections;
>  	uint32_t reserved_segments;
>  	uint32_t new_reserved_segments;
>  	int sparse_mode;
> @@ -1614,6 +1619,20 @@ extern uint32_t f2fs_get_usable_segments(struct f2fs_super_block *sb);
>  #define ZONE_ALIGN(blks)	SIZE_ALIGN(blks, c.blks_per_seg * \
>  					c.segs_per_zone)
>  
> +static inline double get_reserved(struct f2fs_super_block *sb, double ovp)
> +{
> +	double reserved;
> +	uint32_t usable_main_segs = f2fs_get_usable_segments(sb);
> +	uint32_t segs_per_sec = round_up(usable_main_segs, get_sb(section_count));
> +
> +	if (c.conf_reserved_sections)
> +		reserved = c.conf_reserved_sections * segs_per_sec;
> +	else
> +		reserved = (100 / ovp + 1 + NR_CURSEG_TYPE) * segs_per_sec;
> +
> +	return reserved;
> +}
> +
>  static inline double get_best_overprovision(struct f2fs_super_block *sb)
>  {
>  	double reserved, ovp, candidate, end, diff, space;
> @@ -1631,8 +1650,7 @@ static inline double get_best_overprovision(struct f2fs_super_block *sb)
>  	}
>  
>  	for (; candidate <= end; candidate += diff) {
> -		reserved = (100 / candidate + 1 + NR_CURSEG_TYPE) *
> -				round_up(usable_main_segs, get_sb(section_count));
> +		reserved = get_reserved(sb, candidate);
>  		ovp = (usable_main_segs - reserved) * candidate / 100;
>  		if (ovp < 0)
>  			continue;
> diff --git a/lib/libf2fs.c b/lib/libf2fs.c
> index f63307a42a08..b5644ff6ebdd 100644
> --- a/lib/libf2fs.c
> +++ b/lib/libf2fs.c
> @@ -1069,6 +1069,28 @@ int get_device_info(int i)
>  				dev->nr_rnd_zones);
>  		MSG(0, "      %zu blocks per zone\n",
>  				dev->zone_blocks);
> +		if (c.conf_reserved_sections) {
> +			if (c.conf_reserved_sections < MIN_RSVD_SECS) {
> +				MSG(0, "      Too small sections are reserved(%u secs)\n",
> +				    c.conf_reserved_sections);
> +				c.conf_reserved_sections =
> +					max(c.conf_reserved_sections, MIN_RSVD_SECS);
> +				MSG(0, "      It is operated as a minimum reserved sections(%u secs)\n",
> +				    c.conf_reserved_sections);
> +			} else {
> +				MSG(0, "      %u sections are reserved\n",
> +				c.conf_reserved_sections);
> +			}
> +			if (!c.overprovision) {
> +				c.overprovision = CONFIG_RSVD_DEFAULT_OP_RATIO;
> +				MSG(0, "      Overprovision ratio is set to default(%.1lf%%)\n",
> +				    c.overprovision);
> +			}
> +		} else {
> +			MSG(0,
> +				"      0 reserved sections are received.\n"
> +				"      Reserved sections are calculating\n");
> +		}
>  	}
>  #endif
>  	/* adjust wanted_total_sectors */
> diff --git a/man/mkfs.f2fs.8 b/man/mkfs.f2fs.8
> index a6249f6ef6ed..800c4c79bf37 100644
> --- a/man/mkfs.f2fs.8
> +++ b/man/mkfs.f2fs.8
> @@ -43,6 +43,7 @@ mkfs.f2fs \- create an F2FS file system
>  ]
>  [
>  .B \-m
> +.I #-of-reserved-sections
>  ]
>  [
>  .B \-o
> @@ -152,8 +153,12 @@ Enable extended node bitmap.
>  .BI \-l " volume-label"
>  Specify the volume label to the partition mounted as F2FS.
>  .TP
> -.BI \-m
> -Specify f2fs filesystem to supports the block zoned feature.
> +.BI \-m " #-of-reserved-sections"
> +Specify f2fs filesystem to supports the block zoned feature and reserved sections.
> +If specified to non-zero, reserved segments count is set to the larger size
> +between 8 sections and the input value.  If specified to zero, the best number
> +will be assigned automatically according to the partition size. If overprovision-ratio-percentage
> +is not specifed, it will set to default 3.0%.
>  Without it, the filesystem doesn't support the feature.
>  .TP
>  .BI \-o " overprovision-ratio-percentage"
> diff --git a/mkfs/f2fs_format.c b/mkfs/f2fs_format.c
> index f4a49acc498c..22e53be5003d 100644
> --- a/mkfs/f2fs_format.c
> +++ b/mkfs/f2fs_format.c
> @@ -483,9 +483,7 @@ static int f2fs_prepare_super_block(void)
>  	if (c.overprovision == 0)
>  		c.overprovision = get_best_overprovision(sb);
>  
> -	c.reserved_segments =
> -			(100 / c.overprovision + 1 + NR_CURSEG_TYPE) *
> -			round_up(f2fs_get_usable_segments(sb), get_sb(section_count));
> +	c.reserved_segments = get_reserved(sb, c.overprovision);
>  
>  	if (c.feature & cpu_to_le32(F2FS_FEATURE_RO)) {
>  		c.overprovision = 0;
> @@ -765,11 +763,30 @@ static int f2fs_write_check_point_pack(void)
>  			get_cp(rsvd_segment_count)) *
>  			c.overprovision / 100);
>  
> -	if (get_cp(overprov_segment_count) < get_cp(rsvd_segment_count))
> +	if (!(c.conf_reserved_sections) &&
> +	    get_cp(overprov_segment_count) < get_cp(rsvd_segment_count))
>  		set_cp(overprov_segment_count, get_cp(rsvd_segment_count));
>  
> -	set_cp(overprov_segment_count, get_cp(overprov_segment_count) +
> -			2 * get_sb(segs_per_sec));
> +	/*
> +	 * If conf_reserved_sections has a non zero value, overprov_segment_count
> +	 * is set to overprov_segment_count + rsvd_segment_count.
> +	 */
> +	if (c.conf_reserved_sections) {
> +		/*
> +		 * Overprovision segments must be bigger than two sections.
> +		 * In non configurable reserved section case, overprovision
> +		 * segments are always bigger than two sections.
> +		 */
> +		if (get_cp(overprov_segment_count) < 2 * get_sb(segs_per_sec)) {
> +			MSG(0, "\tError: Not enough overprovision segments (%u)\n",
> +			    get_cp(overprov_segment_count));
> +			goto free_cp_payload;
> +		}
> +		set_cp(overprov_segment_count, get_cp(overprov_segment_count) +
> +				get_cp(rsvd_segment_count));
> +	 } else
> +		set_cp(overprov_segment_count, get_cp(overprov_segment_count) +
> +				2 * get_sb(segs_per_sec));
>  
>  	if (f2fs_get_usable_segments(sb) <= get_cp(overprov_segment_count)) {
>  		MSG(0, "\tError: Not enough segments to create F2FS Volume\n");
> diff --git a/mkfs/f2fs_format_main.c b/mkfs/f2fs_format_main.c
> index f50971c4591c..a4accda91780 100644
> --- a/mkfs/f2fs_format_main.c
> +++ b/mkfs/f2fs_format_main.c
> @@ -61,7 +61,7 @@ static void mkfs_usage()
>  	MSG(0, "  -i extended node bitmap, node ratio is 20%% by default\n");
>  	MSG(0, "  -l label\n");
>  	MSG(0, "  -U uuid\n");
> -	MSG(0, "  -m support zoned block device [default:0]\n");
> +	MSG(0, "  -m support zoned block device, # of reserved sections [default:0]\n");
>  	MSG(0, "  -o overprovision percentage [default:auto]\n");
>  	MSG(0, "  -O feature1[,feature2,...] e.g. \"encrypt\"\n");
>  	MSG(0, "  -C [encoding[:flag1,...]] Support casefolding with optional flags\n");
> @@ -176,7 +176,7 @@ static void add_default_options(void)
>  
>  static void f2fs_parse_options(int argc, char *argv[])
>  {
> -	static const char *option_string = "qa:c:C:d:e:E:g:hil:mo:O:rR:s:S:z:t:T:U:Vfw:";
> +	static const char *option_string = "qa:c:C:d:e:E:g:hil:m:o:O:rR:s:S:z:t:T:U:Vfw:";
>  	static const struct option long_opts[] = {
>  		{ .name = "help", .has_arg = 0, .flag = NULL, .val = 'h' },
>  		{ .name = NULL, .has_arg = 0, .flag = NULL, .val = 0 }
> @@ -235,6 +235,7 @@ static void f2fs_parse_options(int argc, char *argv[])
>  			break;
>  		case 'm':
>  			c.zoned_mode = 1;
> +			c.conf_reserved_sections = atoi(optarg);

This breaks the existing "-m" option. Can you address to support that?

>  			break;
>  		case 'o':
>  			c.overprovision = atof(optarg);
> -- 
> 2.34.1


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

  reply	other threads:[~2023-03-27 21:27 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CGME20230317044151epcms2p5872591b638dab3261a9dd563253f1b94@epcms2p5>
2023-03-17  4:41 ` [f2fs-dev] [PATCH v1] mkfs.f2fs: Introduce configurable reserved sections Yonggil Song
2023-03-27 21:26   ` Jaegeuk Kim [this message]
     [not found]   ` <CGME20230317044151epcms2p5872591b638dab3261a9dd563253f1b94@epcms2p3>
2023-03-28  8:06     ` [f2fs-dev] (2) " Yonggil Song
     [not found]   ` <CGME20230317044151epcms2p5872591b638dab3261a9dd563253f1b94@epcms2p2>
2023-04-04  6:21     ` [f2fs-dev] [PATCH v2] " Yonggil Song

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ZCIKIzPmJyjA44JK@google.com \
    --to=jaegeuk@kernel.org \
    --cc=linux-f2fs-devel@lists.sourceforge.net \
    --cc=lunar.lee@samsung.com \
    --cc=sukka.kim@samsung.com \
    --cc=yonggil.song@samsung.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).