From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: stefanha@redhat.com
Subject: [Qemu-devel] [PATCH v5 01/21] cow: make reads go at a decent speed
Date: Wed, 4 Sep 2013 19:00:18 +0200 [thread overview]
Message-ID: <1378314038-15525-2-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1378314038-15525-1-git-send-email-pbonzini@redhat.com>
Do not do two reads for each sector; load each sector of the bitmap
and use bitmap operations to process it.
Writes are still dog slow!
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
block/cow.c | 54 ++++++++++++++++++++++++++++++++----------------------
1 file changed, 32 insertions(+), 22 deletions(-)
diff --git a/block/cow.c b/block/cow.c
index 1cc2e89..fa3d41f 100644
--- a/block/cow.c
+++ b/block/cow.c
@@ -126,18 +126,31 @@ static inline int cow_set_bit(BlockDriverState *bs, int64_t bitnum)
return 0;
}
-static inline int is_bit_set(BlockDriverState *bs, int64_t bitnum)
+#define BITS_PER_BITMAP_SECTOR (512 * 8)
+
+/* Cannot use bitmap.c on big-endian machines. */
+static int cow_test_bit(int64_t bitnum, const uint8_t *bitmap)
{
- uint64_t offset = sizeof(struct cow_header_v2) + bitnum / 8;
- uint8_t bitmap;
- int ret;
+ return (bitmap[bitnum / 8] & (1 << (bitnum & 7))) != 0;
+}
- ret = bdrv_pread(bs->file, offset, &bitmap, sizeof(bitmap));
- if (ret < 0) {
- return ret;
+static int cow_find_streak(const uint8_t *bitmap, int value, int start, int nb_sectors)
+{
+ int streak_value = value ? 0xFF : 0;
+ int last = MIN(start + nb_sectors, BITS_PER_BITMAP_SECTOR);
+ int bitnum = start;
+ while (bitnum < last) {
+ if ((bitnum & 7) == 0 && bitmap[bitnum / 8] == streak_value) {
+ bitnum += 8;
+ continue;
+ }
+ if (cow_test_bit(bitnum, bitmap) == value) {
+ bitnum++;
+ continue;
+ }
+ break;
}
-
- return !!(bitmap & (1 << (bitnum % 8)));
+ return MIN(bitnum, last) - start;
}
/* Return true if first block has been changed (ie. current version is
@@ -146,23 +159,20 @@ static inline int is_bit_set(BlockDriverState *bs, int64_t bitnum)
static int coroutine_fn cow_co_is_allocated(BlockDriverState *bs,
int64_t sector_num, int nb_sectors, int *num_same)
{
+ int64_t bitnum = sector_num + sizeof(struct cow_header_v2) * 8;
+ uint64_t offset = (bitnum / 8) & -BDRV_SECTOR_SIZE;
+ uint8_t bitmap[BDRV_SECTOR_SIZE];
+ int ret;
int changed;
- if (nb_sectors == 0) {
- *num_same = nb_sectors;
- return 0;
- }
-
- changed = is_bit_set(bs, sector_num);
- if (changed < 0) {
- return 0; /* XXX: how to return I/O errors? */
- }
-
- for (*num_same = 1; *num_same < nb_sectors; (*num_same)++) {
- if (is_bit_set(bs, sector_num + *num_same) != changed)
- break;
+ ret = bdrv_pread(bs->file, offset, &bitmap, sizeof(bitmap));
+ if (ret < 0) {
+ return ret;
}
+ bitnum &= BITS_PER_BITMAP_SECTOR - 1;
+ changed = cow_test_bit(bitnum, bitmap);
+ *num_same = cow_find_streak(bitmap, changed, bitnum, nb_sectors);
return changed;
}
--
1.8.3.1
next prev parent reply other threads:[~2013-09-04 17:01 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-09-04 17:00 [Qemu-devel] [PATCH v5 00/21] Add qemu-img subcommand to dump file metadata Paolo Bonzini
2013-09-04 17:00 ` Paolo Bonzini [this message]
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 02/21] cow: make writes go at a less indecent speed Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 03/21] cow: do not call bdrv_co_is_allocated Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 04/21] block: keep bs->total_sectors up to date even for growable block devices Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 05/21] block: make bdrv_co_is_allocated static Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 06/21] block: do not use ->total_sectors in bdrv_co_is_allocated Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 07/21] block: remove bdrv_is_allocated_above/bdrv_co_is_allocated_above distinction Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 08/21] block: expect errors from bdrv_co_is_allocated Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 09/21] qemu-img: always probe the input image for allocated sectors Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 10/21] block: make bdrv_has_zero_init return false for copy-on-write-images Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 11/21] block: introduce bdrv_get_block_status API Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 12/21] block: define get_block_status return value Paolo Bonzini
2014-05-05 14:34 ` Kevin Wolf
2014-05-05 14:58 ` Paolo Bonzini
2014-05-06 11:34 ` Kevin Wolf
2014-05-06 12:31 ` Paolo Bonzini
2014-05-06 13:08 ` Kevin Wolf
2014-05-06 13:20 ` Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 13/21] block: return get_block_status data and flags for formats Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 14/21] block: use bdrv_has_zero_init to return BDRV_BLOCK_ZERO Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 15/21] block: return BDRV_BLOCK_ZERO past end of backing file Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 16/21] qemu-img: add a "map" subcommand Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 17/21] docs, qapi: document qemu-img map Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 18/21] raw-posix: return get_block_status data and flags Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 19/21] raw-posix: report unwritten extents as zero Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 20/21] block: add default get_block_status implementation for protocols Paolo Bonzini
2013-09-04 17:00 ` [Qemu-devel] [PATCH v5 21/21] block: look for zero blocks in bs->file Paolo Bonzini
2013-09-05 13:55 ` [Qemu-devel] [PATCH v5 00/21] Add qemu-img subcommand to dump file metadata Stefan Hajnoczi
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=1378314038-15525-2-git-send-email-pbonzini@redhat.com \
--to=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.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).