From: Kevin Wolf <kwolf@redhat.com>
To: Max Reitz <mreitz@redhat.com>
Cc: qemu-devel@nongnu.org, Stefan Hajnoczi <stefanha@redhat.com>
Subject: Re: [Qemu-devel] [PATCH v2 1/4] qcow2: Constant cache size in bytes
Date: Tue, 19 Aug 2014 15:27:41 +0200 [thread overview]
Message-ID: <20140819132741.GF4638@noname.redhat.com> (raw)
In-Reply-To: <1408392454-22044-2-git-send-email-mreitz@redhat.com>
Am 18.08.2014 um 22:07 hat Max Reitz geschrieben:
> Specifying the metadata cache sizes in clusters results in less clusters
> (and much less bytes) covered for small cluster sizes and vice versa.
> Using a constant byte size reduces this difference, and makes it
> possible to manually specify the cache size in an easily comprehensible
> unit.
>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/qcow2.c | 16 ++++++++++++++--
> block/qcow2.h | 10 ++++++++--
> 2 files changed, 22 insertions(+), 4 deletions(-)
>
> diff --git a/block/qcow2.c b/block/qcow2.c
> index 435e0e1..910d9cf 100644
> --- a/block/qcow2.c
> +++ b/block/qcow2.c
> @@ -470,6 +470,7 @@ static int qcow2_open(BlockDriverState *bs, QDict *options, int flags,
> uint64_t l1_vm_state_index;
> const char *opt_overlap_check;
> int overlap_check_template = 0;
> + uint64_t l2_cache_size, refcount_cache_size;
>
> ret = bdrv_pread(bs->file, 0, &header, sizeof(header));
> if (ret < 0) {
> @@ -707,8 +708,19 @@ static int qcow2_open(BlockDriverState *bs, QDict *options, int flags,
> }
>
> /* alloc L2 table/refcount block cache */
> - s->l2_table_cache = qcow2_cache_create(bs, L2_CACHE_SIZE);
> - s->refcount_block_cache = qcow2_cache_create(bs, REFCOUNT_CACHE_SIZE);
> + l2_cache_size = DEFAULT_L2_CACHE_BYTE_SIZE / s->cluster_size;
> + if (l2_cache_size < MIN_L2_CACHE_SIZE) {
> + l2_cache_size = MIN_L2_CACHE_SIZE;
> + }
> +
> + refcount_cache_size = l2_cache_size
> + / (DEFAULT_L2_REFCOUNT_SIZE_RATIO * s->cluster_size);
The factor s->cluster_size is too much, l2_cache_size is already in
clusters. I think one of the other patches in the series fixes this
again, at least I can't see the bug in the final applied version.
If you agree, I'll change this patch while applying to make the series
bisectable.
Kevin
> + if (refcount_cache_size < MIN_REFCOUNT_CACHE_SIZE) {
> + refcount_cache_size = MIN_REFCOUNT_CACHE_SIZE;
> + }
> +
> + s->l2_table_cache = qcow2_cache_create(bs, l2_cache_size);
> + s->refcount_block_cache = qcow2_cache_create(bs, refcount_cache_size);
> if (s->l2_table_cache == NULL || s->refcount_block_cache == NULL) {
> error_setg(errp, "Could not allocate metadata caches");
> ret = -ENOMEM;
> diff --git a/block/qcow2.h b/block/qcow2.h
> index b49424b..671783d 100644
> --- a/block/qcow2.h
> +++ b/block/qcow2.h
> @@ -64,10 +64,16 @@
> #define MIN_CLUSTER_BITS 9
> #define MAX_CLUSTER_BITS 21
>
> -#define L2_CACHE_SIZE 16
> +#define MIN_L2_CACHE_SIZE 1 /* cluster */
>
> /* Must be at least 4 to cover all cases of refcount table growth */
> -#define REFCOUNT_CACHE_SIZE 4
> +#define MIN_REFCOUNT_CACHE_SIZE 4 /* clusters */
> +
> +#define DEFAULT_L2_CACHE_BYTE_SIZE 1048576 /* bytes */
> +
> +/* The refblock cache needs only a fourth of the L2 cache size to cover as many
> + * clusters */
> +#define DEFAULT_L2_REFCOUNT_SIZE_RATIO 4
>
> #define DEFAULT_CLUSTER_SIZE 65536
>
> --
> 2.0.4
>
next prev parent reply other threads:[~2014-08-19 13:27 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-08-18 20:07 [Qemu-devel] [PATCH v2 0/4] qcow2: Allow runtime specification of cache sizes Max Reitz
2014-08-18 20:07 ` [Qemu-devel] [PATCH v2 1/4] qcow2: Constant cache size in bytes Max Reitz
2014-08-19 13:27 ` Kevin Wolf [this message]
2014-08-18 20:07 ` [Qemu-devel] [PATCH v2 2/4] qcow2: Use g_try_new0() for cache array Max Reitz
2014-08-18 20:07 ` [Qemu-devel] [PATCH v2 3/4] qcow2: Add runtime options for cache sizes Max Reitz
2014-08-18 20:07 ` [Qemu-devel] [PATCH v2 4/4] iotests: Add test for qcow2's cache options Max Reitz
2014-08-19 14:00 ` [Qemu-devel] [PATCH v2 0/4] qcow2: Allow runtime specification of cache sizes Kevin Wolf
2014-08-19 14:18 ` Eric Blake
2014-08-19 14:39 ` Kevin Wolf
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=20140819132741.GF4638@noname.redhat.com \
--to=kwolf@redhat.com \
--cc=mreitz@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).