All of lore.kernel.org
 help / color / mirror / Atom feed
From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
To: Alberto Garcia <berto@igalia.com>, qemu-devel@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>,
	Anton Nefedov <anton.nefedov@virtuozzo.com>,
	qemu-block@nongnu.org, Max Reitz <mreitz@redhat.com>,
	"Denis V . Lunev" <den@openvz.org>
Subject: Re: [PATCH v4 11/30] qcow2: Add l2_entry_size()
Date: Tue, 14 Apr 2020 12:44:57 +0300	[thread overview]
Message-ID: <58d1fa17-91ea-9f8d-c39a-4141783d1234@virtuozzo.com> (raw)
In-Reply-To: <fd0f93353a218ff4518f34ebdbca05c2fc0f1085.1584468723.git.berto@igalia.com>

17.03.2020 21:16, Alberto Garcia wrote:
> qcow2 images with subclusters have 128-bit L2 entries. The first 64
> bits contain the same information as traditional images and the last
> 64 bits form a bitmap with the status of each individual subcluster.
> 
> Because of that we cannot assume that L2 entries are sizeof(uint64_t)
> anymore. This function returns the proper value for the image.
> 
> Signed-off-by: Alberto Garcia <berto@igalia.com>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
> ---
>   block/qcow2.h          |  9 +++++++++
>   block/qcow2-cluster.c  | 12 ++++++------
>   block/qcow2-refcount.c | 14 ++++++++------
>   block/qcow2.c          |  8 ++++----
>   4 files changed, 27 insertions(+), 16 deletions(-)
> 
> diff --git a/block/qcow2.h b/block/qcow2.h
> index 06929072d2..1eb4b46807 100644
> --- a/block/qcow2.h
> +++ b/block/qcow2.h
> @@ -80,6 +80,10 @@
>   
>   #define QCOW_EXTL2_SUBCLUSTERS_PER_CLUSTER 32
>   
> +/* Size of normal and extended L2 entries */
> +#define L2E_SIZE_NORMAL   (sizeof(uint64_t))
> +#define L2E_SIZE_EXTENDED (sizeof(uint64_t) * 2)
> +
>   #define MIN_CLUSTER_BITS 9
>   #define MAX_CLUSTER_BITS 21
>   
> @@ -506,6 +510,11 @@ static inline bool has_subclusters(BDRVQcow2State *s)
>       return false;
>   }
>   
> +static inline size_t l2_entry_size(BDRVQcow2State *s)
> +{
> +    return has_subclusters(s) ? L2E_SIZE_EXTENDED : L2E_SIZE_NORMAL;
> +}
> +
>   static inline uint64_t get_l2_entry(BDRVQcow2State *s, uint64_t *l2_slice,
>                                       int idx)
>   {
> diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
> index cd48ab0223..41a23c5305 100644
> --- a/block/qcow2-cluster.c
> +++ b/block/qcow2-cluster.c
> @@ -208,7 +208,7 @@ static int l2_load(BlockDriverState *bs, uint64_t offset,
>                      uint64_t l2_offset, uint64_t **l2_slice)
>   {
>       BDRVQcow2State *s = bs->opaque;
> -    int start_of_slice = sizeof(uint64_t) *
> +    int start_of_slice = l2_entry_size(s) *
>           (offset_to_l2_index(s, offset) - offset_to_l2_slice_index(s, offset));
>   
>       return qcow2_cache_get(bs, s->l2_table_cache, l2_offset + start_of_slice,
> @@ -281,7 +281,7 @@ static int l2_allocate(BlockDriverState *bs, int l1_index)
>   
>       /* allocate a new l2 entry */
>   
> -    l2_offset = qcow2_alloc_clusters(bs, s->l2_size * sizeof(uint64_t));
> +    l2_offset = qcow2_alloc_clusters(bs, s->l2_size * l2_entry_size(s));

hmm. s->l2_size * l2_entry_size, isn't it just s->cluster_size always? Maybe, just refactor these things?


>       if (l2_offset < 0) {
>           ret = l2_offset;
>           goto fail;
> @@ -305,7 +305,7 @@ static int l2_allocate(BlockDriverState *bs, int l1_index)

[...]

> @@ -1425,7 +1425,7 @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options,
>           bs->encrypted = true;
>       }
>   
> -    s->l2_bits = s->cluster_bits - 3; /* L2 is always one cluster */
> +    s->l2_bits = s->cluster_bits - ctz32(l2_entry_size(s));
>       s->l2_size = 1 << s->l2_bits;
>       /* 2^(s->refcount_order - 3) is the refcount width in bytes */
>       s->refcount_block_bits = s->cluster_bits - (s->refcount_order - 3);
> @@ -4104,7 +4104,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,
>            *  preallocation. All that matters is that we will not have to allocate
>            *  new refcount structures for them.) */
>           nb_new_l2_tables = DIV_ROUND_UP(nb_new_data_clusters,
> -                                        s->cluster_size / sizeof(uint64_t));
> +                                        s->cluster_size / l2_entry_size(s));

Isn't it just s->l2_size ?

>           /* The cluster range may not be aligned to L2 boundaries, so add one L2
>            * table for a potential head/tail */
>           nb_new_l2_tables++;
> 


Conversions looks correct, but how to check that we have converted everything?

Trying at least

    cd block; git grep 'sizeof(uint64_t)' qcow2* | grep -v 'l1_size \*' | grep -v 'l1_sz \*' | grep -v refcount | grep -v reftable

I found this not converted chunk:

     /* total size of L2 tables */
     nl2e = aligned_total_size / cluster_size;
     nl2e = ROUND_UP(nl2e, cluster_size / sizeof(uint64_t));
     meta_size += nl2e * sizeof(uint64_t);


Hmm. How to avoid it? Maybe, at least, refactor the code, to drop all sizeof(uint64_t), converting them to L2_ENTRY_SIZE, L1_ENTRY_SIZE, REFTABLE_ENTRY_SIZE etc?
And all occurrences of pure '8' (not many of them exist)

-- 
Best regards,
Vladimir


  reply	other threads:[~2020-04-14  9:46 UTC|newest]

Thread overview: 128+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-17 18:15 [PATCH v4 00/30] Add subcluster allocation to qcow2 Alberto Garcia
2020-03-17 18:15 ` [PATCH v4 01/30] qcow2: Make Qcow2AioTask store the full host offset Alberto Garcia
2020-03-18 11:23   ` Eric Blake
2020-04-08 10:23   ` Max Reitz
2020-04-09  6:49   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:15 ` [PATCH v4 02/30] qcow2: Convert qcow2_get_cluster_offset() into qcow2_get_host_offset() Alberto Garcia
2020-03-18 12:08   ` Eric Blake
2020-04-08 10:51   ` Max Reitz
2020-04-08 17:29     ` Alberto Garcia
2020-04-09  7:57     ` Vladimir Sementsov-Ogievskiy
2020-04-09 14:35       ` Alberto Garcia
2020-04-09  7:50   ` Vladimir Sementsov-Ogievskiy
2020-04-09 14:45     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 03/30] qcow2: Add calculate_l2_meta() Alberto Garcia
2020-04-09  8:30   ` Vladimir Sementsov-Ogievskiy
2020-04-09 15:12     ` Alberto Garcia
2020-04-09 18:47       ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 04/30] qcow2: Split cluster_needs_cow() out of count_cow_clusters() Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 05/30] qcow2: Process QCOW2_CLUSTER_ZERO_ALLOC clusters in handle_copied() Alberto Garcia
2020-04-09 10:59   ` Vladimir Sementsov-Ogievskiy
2020-04-09 16:08     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 06/30] qcow2: Add get_l2_entry() and set_l2_entry() Alberto Garcia
2020-04-10  8:48   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 07/30] qcow2: Document the Extended L2 Entries feature Alberto Garcia
2020-04-08 11:09   ` Max Reitz
2020-04-09 15:12   ` Eric Blake
2020-04-10  9:29     ` Vladimir Sementsov-Ogievskiy
2020-04-14 14:50       ` Alberto Garcia
2020-04-14 16:19         ` Vladimir Sementsov-Ogievskiy
2020-04-14 16:30           ` Alberto Garcia
2020-04-14 18:06             ` Vladimir Sementsov-Ogievskiy
2020-04-14 18:13               ` Alberto Garcia
2020-04-15 19:11       ` Alberto Garcia
2020-04-15 21:13         ` Eric Blake
2020-04-10 12:01     ` Alberto Garcia
2020-04-14 18:16     ` Alberto Garcia
2020-04-14 18:23       ` Eric Blake
2020-04-14 18:25         ` Eric Blake
2020-03-17 18:16 ` [PATCH v4 08/30] qcow2: Add dummy has_subclusters() function Alberto Garcia
2020-04-10  9:11   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 09/30] qcow2: Add subcluster-related fields to BDRVQcow2State Alberto Garcia
2020-04-08 11:12   ` Max Reitz
2020-04-10  9:45   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 10/30] qcow2: Add offset_to_sc_index() Alberto Garcia
2020-04-13 11:02   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 11/30] qcow2: Add l2_entry_size() Alberto Garcia
2020-04-14  9:44   ` Vladimir Sementsov-Ogievskiy [this message]
2020-04-14 12:20     ` Alberto Garcia
2020-04-14 12:29       ` Vladimir Sementsov-Ogievskiy
2020-04-14 12:33         ` Alberto Garcia
2020-04-14 12:39           ` Vladimir Sementsov-Ogievskiy
2020-04-14 16:01       ` Eric Blake
2020-04-14 16:16         ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 12/30] qcow2: Update get/set_l2_entry() and add get/set_l2_bitmap() Alberto Garcia
2020-04-14  9:49   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 13/30] qcow2: Add QCow2SubclusterType and qcow2_get_subcluster_type() Alberto Garcia
2020-04-08 11:23   ` Max Reitz
2020-04-08 17:46     ` Alberto Garcia
2020-04-09  8:22       ` Max Reitz
2020-04-14 11:10   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 14/30] qcow2: Add cluster type parameter to qcow2_get_host_offset() Alberto Garcia
2020-04-08 12:15   ` Max Reitz
2020-04-14 12:30   ` Vladimir Sementsov-Ogievskiy
2020-04-14 12:38     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 15/30] qcow2: Replace QCOW2_CLUSTER_* with QCOW2_SUBCLUSTER_* Alberto Garcia
2020-04-08 12:42   ` Max Reitz
2020-04-15  7:10   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 16/30] qcow2: Handle QCOW2_SUBCLUSTER_UNALLOCATED_ALLOC Alberto Garcia
2020-04-15  7:28   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 17/30] qcow2: Add subcluster support to calculate_l2_meta() Alberto Garcia
2020-04-15  8:39   ` Vladimir Sementsov-Ogievskiy
2020-04-16 20:01     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 18/30] qcow2: Add subcluster support to qcow2_get_host_offset() Alberto Garcia
2020-04-08 12:49   ` Max Reitz
2020-04-08 17:35     ` Alberto Garcia
2020-04-22  8:07   ` Vladimir Sementsov-Ogievskiy
2020-04-22 11:54     ` Alberto Garcia
2020-04-22 12:18       ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 19/30] qcow2: Add subcluster support to zero_in_l2_slice() Alberto Garcia
2020-04-22 11:06   ` Vladimir Sementsov-Ogievskiy
2020-04-22 12:53     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 20/30] qcow2: Add subcluster support to discard_in_l2_slice() Alberto Garcia
2020-04-09 10:05   ` Max Reitz
2020-04-10 12:47     ` Alberto Garcia
2020-04-14 10:13       ` Max Reitz
2020-04-22 11:35   ` Vladimir Sementsov-Ogievskiy
2020-04-22 17:42     ` Alberto Garcia
2020-04-22 18:09       ` Vladimir Sementsov-Ogievskiy
2020-04-23 14:18         ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 21/30] qcow2: Add subcluster support to check_refcounts_l2() Alberto Garcia
2020-04-22 12:06   ` Vladimir Sementsov-Ogievskiy
2020-04-23 15:45     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 22/30] qcow2: Fix offset calculation in handle_dependencies() Alberto Garcia
2020-04-22 12:38   ` Vladimir Sementsov-Ogievskiy
2020-04-23 15:50     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 23/30] qcow2: Update L2 bitmap in qcow2_alloc_cluster_link_l2() Alberto Garcia
2020-04-24 19:39   ` Eric Blake
2020-04-27 13:17     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 24/30] qcow2: Clear the L2 bitmap when allocating a compressed cluster Alberto Garcia
2020-04-24 17:02   ` Alberto Garcia
2020-04-24 17:11     ` Eric Blake
2020-04-24 17:21       ` Alberto Garcia
2020-04-24 17:44         ` Eric Blake
2020-04-24 17:56           ` Alberto Garcia
2020-04-24 18:25             ` Vladimir Sementsov-Ogievskiy
2020-04-24 18:37               ` Alberto Garcia
2020-04-24 18:47                 ` Eric Blake
2020-04-27  7:49                   ` Max Reitz
2020-04-27 18:12                     ` Alberto Garcia
2020-04-24 18:15           ` Vladimir Sementsov-Ogievskiy
2020-04-24 18:41             ` Alberto Garcia
2020-04-25  6:38               ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 25/30] qcow2: Add subcluster support to handle_alloc_space() Alberto Garcia
2020-04-27 11:54   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 26/30] qcow2: Restrict qcow2_co_pwrite_zeroes() to full clusters only Alberto Garcia
2020-04-27 11:59   ` Vladimir Sementsov-Ogievskiy
2020-03-17 18:16 ` [PATCH v4 27/30] qcow2: Assert that expand_zero_clusters_in_l1() does not support subclusters Alberto Garcia
2020-04-09 10:27   ` Max Reitz
2020-04-10 16:42     ` Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 28/30] qcow2: Add the 'extended_l2' option and the QCOW2_INCOMPAT_EXTL2 bit Alberto Garcia
2020-04-09 14:49   ` Eric Blake
2020-03-17 18:16 ` [PATCH v4 29/30] qcow2: Add subcluster support to qcow2_measure() Alberto Garcia
2020-03-17 18:16 ` [PATCH v4 30/30] iotests: Add tests for qcow2 images with extended L2 entries Alberto Garcia
2020-04-09 12:22   ` Max Reitz
2020-04-13 17:16     ` Alberto Garcia
2020-04-14 10:14       ` Max Reitz
2020-04-21  5:06 ` [PATCH v4 00/30] Add subcluster allocation to qcow2 Derek Su
2020-04-21 10:35   ` Alberto Garcia

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=58d1fa17-91ea-9f8d-c39a-4141783d1234@virtuozzo.com \
    --to=vsementsov@virtuozzo.com \
    --cc=anton.nefedov@virtuozzo.com \
    --cc=berto@igalia.com \
    --cc=den@openvz.org \
    --cc=kwolf@redhat.com \
    --cc=mreitz@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.