From: Anand Jain <anand.jain@oracle.com>
To: David Sterba <dsterba@suse.com>
Cc: linux-btrfs@vger.kernel.org
Subject: Re: [PATCH 6/9] btrfs: reduce compressed_bio members' types
Date: Wed, 26 May 2021 16:41:33 +0800 [thread overview]
Message-ID: <fe145cfe-e8b9-ca83-5bf7-347db9f49ed0@oracle.com> (raw)
In-Reply-To: <1a3c6cddd909d922948a22ac1f287293e0deb665.1621961965.git.dsterba@suse.com>
On 26/05/2021 01:08, David Sterba wrote:
> Several members of compressed_bio are of type that's unnecessarily big
> for the values that they'd hold:
>
> - the size of the uncompressed and compressed data is 128K now, we can
> keep is as int
> - same for number of pages
> - the compress type fits to a byte
> - the errors is 0/1
>
> The size of the unpatched structure is 80 bytes with several holes.
> Reordering nr_pages next to the pages the hole after pending_bios is
> filled and the resulting size is 56 bytes. This keeps the csums array
> aligned to 8 bytes, which is nice. Further size optimizations may be
> possible but right now it looks good to me:
>
> struct compressed_bio {
> refcount_t pending_bios; /* 0 4 */
> unsigned int nr_pages; /* 4 4 */
> struct page * * compressed_pages; /* 8 8 */
> struct inode * inode; /* 16 8 */
> u64 start; /* 24 8 */
> unsigned int len; /* 32 4 */
> unsigned int compressed_len; /* 36 4 */
> u8 compress_type; /* 40 1 */
> u8 errors; /* 41 1 */
>
> /* XXX 2 bytes hole, try to pack */
>
> int mirror_num; /* 44 4 */
> struct bio * orig_bio; /* 48 8 */
> u8 sums[]; /* 56 0 */
>
> /* size: 56, cachelines: 1, members: 12 */
> /* sum members: 54, holes: 1, sum holes: 2 */
> /* last cacheline: 56 bytes */
> };
>
The following member types of struct compressed_bio are changed here:
- unsigned long len;
+ unsigned int len;
- int compress_type;
+ u8 compress_type;
- unsigned long nr_pages;
+ unsigned int nr_pages;
- unsigned long compressed_len;
+ unsigned int compressed_len;
- int errors;
+ u8 errors;
There are no warnings.
But in btrfs_submit_compressed_write()
Essentially struct compressed_bio is updated from struct async_extent.
struct async_extent {
u64 start;
u64 ram_size;
u64 compressed_size;
struct page **pages;
unsigned long nr_pages;
int compress_type;
struct list_head list;
};
which can be looked into later.
For now, with the patch
btrfs: optimize users of members of the struct compressed_bio
which is sent to ML.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
> Signed-off-by: David Sterba <dsterba@suse.com>
> ---
> fs/btrfs/compression.c | 2 +-
> fs/btrfs/compression.h | 20 ++++++++++----------
> 2 files changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
> index 9a0c26e4e389..c006f5d81c2a 100644
> --- a/fs/btrfs/compression.c
> +++ b/fs/btrfs/compression.c
> @@ -507,7 +507,7 @@ blk_status_t btrfs_submit_compressed_write(struct btrfs_inode *inode, u64 start,
> }
> if (bytes_left < PAGE_SIZE) {
> btrfs_info(fs_info,
> - "bytes left %lu compress len %lu nr %lu",
> + "bytes left %lu compress len %u nr %u",
> bytes_left, cb->compressed_len, cb->nr_pages);
> }
> bytes_left -= PAGE_SIZE;
> diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h
> index 8001b700ea3a..00d8439048c9 100644
> --- a/fs/btrfs/compression.h
> +++ b/fs/btrfs/compression.h
> @@ -31,6 +31,9 @@ struct compressed_bio {
> /* number of bios pending for this compressed extent */
> refcount_t pending_bios;
>
> + /* Number of compressed pages in the array */
> + unsigned int nr_pages;
> +
> /* the pages with the compressed data on them */
> struct page **compressed_pages;
>
> @@ -40,20 +43,17 @@ struct compressed_bio {
> /* starting offset in the inode for our pages */
> u64 start;
>
> - /* number of bytes in the inode we're working on */
> - unsigned long len;
> -
> - /* number of bytes on disk */
> - unsigned long compressed_len;
> + /* Number of bytes in the inode we're working on */
> + unsigned int len;
>
> - /* the compression algorithm for this bio */
> - int compress_type;
> + /* Number of bytes on disk */
> + unsigned int compressed_len;
>
> - /* number of compressed pages in the array */
> - unsigned long nr_pages;
> + /* The compression algorithm for this bio */
> + u8 compress_type;
>
> /* IO errors */
> - int errors;
> + u8 errors;
> int mirror_num;
>
> /* for reads, this is the bio we are copying the data into */
>
next prev parent reply other threads:[~2021-05-26 8:41 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-25 17:08 [PATCH 0/9] Misc fixups and cleanups David Sterba
2021-05-25 17:08 ` [PATCH 1/9] btrfs: sysfs: fix format string for some discard stats David Sterba
2021-05-25 23:47 ` Qu Wenruo
2021-05-26 6:01 ` Anand Jain
2021-05-25 17:08 ` [PATCH 2/9] btrfs: clear defrag status of a root if starting transaction fails David Sterba
2021-05-25 23:49 ` Qu Wenruo
2021-05-26 6:05 ` Anand Jain
2021-05-25 17:08 ` [PATCH 3/9] btrfs: clear log tree recovering status " David Sterba
2021-05-25 23:50 ` Qu Wenruo
2021-05-26 6:57 ` Anand Jain
2021-05-25 17:08 ` [PATCH 4/9] btrfs: scrub: factor out common scrub_stripe constraints David Sterba
2021-05-25 23:51 ` Qu Wenruo
2021-05-26 7:16 ` Anand Jain
2021-05-25 17:08 ` [PATCH 5/9] btrfs: document byte swap optimization of root_item::flags accessors David Sterba
2021-05-26 7:28 ` Anand Jain
2021-05-25 17:08 ` [PATCH 6/9] btrfs: reduce compressed_bio members' types David Sterba
2021-05-26 8:40 ` [PATCH] btrfs: optimize users of members of the struct compressed_bio Anand Jain
2021-05-26 16:34 ` David Sterba
2021-05-26 8:41 ` Anand Jain [this message]
2021-05-25 17:08 ` [PATCH 7/9] btrfs: remove extra sb::s_id from message in btrfs_validate_metadata_buffer David Sterba
2021-05-25 23:59 ` Qu Wenruo
2021-05-26 12:54 ` Anand Jain
2021-05-25 17:08 ` [PATCH 8/9] btrfs: simplify eb checksum verification " David Sterba
2021-05-26 0:05 ` Qu Wenruo
2021-05-26 16:31 ` David Sterba
2021-05-26 16:58 ` David Sterba
2021-05-26 23:13 ` Qu Wenruo
2021-05-26 23:13 ` David Sterba
2021-05-25 17:08 ` [PATCH 9/9] btrfs: clean up header members offsets in write helpers David Sterba
2021-05-26 0:07 ` Qu Wenruo
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=fe145cfe-e8b9-ca83-5bf7-347db9f49ed0@oracle.com \
--to=anand.jain@oracle.com \
--cc=dsterba@suse.com \
--cc=linux-btrfs@vger.kernel.org \
/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).