From: Jan Kara <jack@suse.cz>
To: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Cc: linux-ext4@vger.kernel.org, Theodore Ts'o <tytso@mit.edu>,
Ritesh Harjani <riteshh@linux.ibm.com>,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
Jan Kara <jack@suse.cz>, Kemeng Shi <shikemeng@huaweicloud.com>
Subject: Re: [PATCH v2 12/12] ext4: Give symbolic names to mballoc criterias
Date: Wed, 7 Jun 2023 12:39:36 +0200 [thread overview]
Message-ID: <20230607103936.wqtcrc76tqpbc2ya@quack3> (raw)
In-Reply-To: <a2dc6ec5aea5e5e68cf8e788c2a964ffead9c8b0.1685449706.git.ojaswin@linux.ibm.com>
On Tue 30-05-23 18:03:50, Ojaswin Mujoo wrote:
> mballoc criterias have historically been called by numbers
> like CR0, CR1... however this makes it confusing to understand
> what each criteria is about.
>
> Change these criterias from numbers to symbolic names and add
> relevant comments. While we are at it, also reformat and add some
> comments to ext4_seq_mb_stats_show() for better readability.
>
> Additionally, define CR_FAST which signifies the criteria
> below which we can make quicker decisions like:
> * quitting early if (free block < requested len)
> * avoiding to scan free extents smaller than required len.
> * avoiding to initialize buddy cache and work with existing cache
> * limiting prefetches
>
> Suggested-by: Jan Kara <jack@suse.cz>
> Signed-off-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Thanks for doing this!
> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> index 942e97026a60..c29a4e1fcd5d 100644
> --- a/fs/ext4/ext4.h
> +++ b/fs/ext4/ext4.h
> @@ -135,16 +135,45 @@ enum SHIFT_DIRECTION {
> */
> #define EXT4_MB_NUM_CRS 5
> /*
> - * All possible allocation criterias for mballoc
> + * All possible allocation criterias for mballoc. Lower are faster.
> */
> enum criteria {
> - CR0,
> - CR1,
> - CR1_5,
> - CR2,
> - CR3,
> + /*
> + * Used when number of blocks needed is a power of 2. This doesn't
> + * trigger any disk IO except prefetch and is the fastest criteria.
> + */
> + CR_POWER2_ALIGNED,
> +
> + /*
> + * Tries to lookup in-memory data structures to find the most suitable
> + * group that satisfies goal request. No disk IO except block prefetch.
> + */
> + CR_GOAL_LEN_FAST,
> +
> + /*
> + * Same as CR_GOAL_LEN_FAST but is allowed to reduce the goal length to
> + * the best available length for faster allocation.
Some whitespace damage here...
> + */
> + CR_BEST_AVAIL_LEN,
> +
> + /*
> + * Reads each block group sequentially, performing disk IO if necessary, to
> + * find find_suitable block group. Tries to allocate goal length but might trim
Too long line here.
> + * the request if nothing is found after enough tries.
> + */
> + CR_GOAL_LEN_SLOW,
> +
> + /*
> + * Finds the first free set of blocks and allocates those. This is only
> + * used in rare cases when CR_GOAL_LEN_SLOW also fails to allocate
> + * anything.
> + */
> + CR_ANY_FREE,
> };
>
> +/* criteria below which we use fast block scanning and avoid unnecessary IO */
> +#define CR_FAST CR_GOAL_LEN_SLOW
> +
Maybe instead of defining CR_FAST value we could define
static inline bool mballoc_cr_expensive(enum criteria cr)
{
return cr >= CR_GOAL_LEN_SLOW;
}
And use this. I think it will make the conditions more understandable.
...
> @@ -1064,7 +1068,7 @@ static inline int should_optimize_scan(struct ext4_allocation_context *ac)
> {
> if (unlikely(!test_opt2(ac->ac_sb, MB_OPTIMIZE_SCAN)))
> return 0;
> - if (ac->ac_criteria >= CR2)
> + if (ac->ac_criteria >= CR_GOAL_LEN_SLOW)
Maybe we should use CR_FAST (or the new function) here?
Otherwise the patch looks good!
Honza
--
Jan Kara <jack@suse.com>
SUSE Labs, CR
next prev parent reply other threads:[~2023-06-07 10:39 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-30 12:33 [PATCH v2 00/12] multiblock allocator improvements Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 01/12] Revert "ext4: remove ac->ac_found > sbi->s_mb_min_to_scan dead check in ext4_mb_check_limits" Ojaswin Mujoo
2023-05-30 16:28 ` Sedat Dilek
2023-05-31 8:57 ` Ojaswin Mujoo
2023-06-02 13:45 ` Thorsten Leemhuis
2023-06-02 16:45 ` Theodore Ts'o
2023-05-30 12:33 ` [PATCH v2 02/12] ext4: mballoc: Remove useless setting of ac_criteria Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 03/12] ext4: Remove unused extern variables declaration Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 04/12] ext4: Convert mballoc cr (criteria) to enum Ojaswin Mujoo
2023-06-06 13:13 ` Jan Kara
2023-05-30 12:33 ` [PATCH v2 05/12] ext4: Add per CR extent scanned counter Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 06/12] ext4: Add counter to track successful allocation of goal length Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 07/12] ext4: Avoid scanning smaller extents in BG during CR1 Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 08/12] ext4: Don't skip prefetching BLOCK_UNINIT groups Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 09/12] ext4: Ensure ext4_mb_prefetch_fini() is called for all prefetched BGs Ojaswin Mujoo
2023-06-06 14:00 ` Guoqing Jiang
2023-06-27 6:51 ` Ojaswin Mujoo
2023-06-28 1:33 ` Guoqing Jiang
2023-05-30 12:33 ` [PATCH v2 10/12] ext4: Abstract out logic to search average fragment list Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 11/12] ext4: Add allocation criteria 1.5 (CR1_5) Ojaswin Mujoo
2023-06-07 10:21 ` Jan Kara
2023-06-08 14:45 ` Theodore Ts'o
2023-06-09 10:57 ` Ojaswin Mujoo
2023-05-30 12:33 ` [PATCH v2 12/12] ext4: Give symbolic names to mballoc criterias Ojaswin Mujoo
2023-06-07 10:39 ` Jan Kara [this message]
2023-06-09 3:14 ` [PATCH v2 00/12] multiblock allocator improvements Theodore Ts'o
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=20230607103936.wqtcrc76tqpbc2ya@quack3 \
--to=jack@suse.cz \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ojaswin@linux.ibm.com \
--cc=riteshh@linux.ibm.com \
--cc=shikemeng@huaweicloud.com \
--cc=tytso@mit.edu \
/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).