qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
To: Hanna Reitz <hreitz@redhat.com>, qemu-block@nongnu.org
Cc: qemu-devel@nongnu.org, kwolf@redhat.com, nsoffer@redhat.com,
	eblake@redhat.com, jsnow@redhat.com, den@openvz.org,
	nikita.lapshin@virtuozzo.com
Subject: Re: [PATCH v2 2/4] qemu-img: make --block-size optional for compare --stat
Date: Thu, 28 Oct 2021 13:04:24 +0300	[thread overview]
Message-ID: <d1f58c9e-f578-7017-282a-26deda1644c4@virtuozzo.com> (raw)
In-Reply-To: <3afe188c-e6de-e956-b068-af6ae1f988da@redhat.com>

26.10.2021 11:07, Hanna Reitz wrote:
> On 21.10.21 12:12, Vladimir Sementsov-Ogievskiy wrote:
>> Let's detect block-size automatically if not specified by user:
>>
>>   If both files define cluster-size, use minimum to be more precise.
>>   If both files don't specify cluster-size, use default of 64K
>>   If only one file specify cluster-size, just use it.
>>
>> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
>> ---
>>   docs/tools/qemu-img.rst |  7 +++-
>>   qemu-img.c              | 71 ++++++++++++++++++++++++++++++++++++-----
>>   qemu-img-cmds.hx        |  4 +--
>>   3 files changed, 71 insertions(+), 11 deletions(-)
> 
> Looks good, just grammar nits and a request for an assertion below.
> 
>> diff --git a/docs/tools/qemu-img.rst b/docs/tools/qemu-img.rst
>> index 21164253d4..4b382ca2b0 100644
>> --- a/docs/tools/qemu-img.rst
>> +++ b/docs/tools/qemu-img.rst
>> @@ -170,6 +170,11 @@ Parameters to compare subcommand:
>>     Block size for comparing with ``--stat``. This doesn't guarantee exact
>>     size of comparing chunks, but that guarantee that data chunks being
>>     compared will never cross aligned block-size boundary.
>> +  When unspecified the following logic is used:
>> +
>> +    - If both files define cluster-size, use minimum to be more precise.
>> +    - If both files don't specify cluster-size, use default of 64K
>> +    - If only one file specify cluster-size, just use it.
> 
> s/specify/specifies/; and perhaps s/it/that/
> 
> [...]
> 
>> diff --git a/qemu-img.c b/qemu-img.c
>> index 79a0589167..61e7f470bb 100644
>> --- a/qemu-img.c
>> +++ b/qemu-img.c
>> @@ -1407,6 +1407,61 @@ static void cmp_stat_print(ImgCmpStat *stat, int64_t total_bytes)
>>       }
>>   }
>> +/* Get default value for qemu-img compare --block-size option. */
>> +static int img_compare_block_size(BlockDriverState *bs1,
>> +                                  BlockDriverState *bs2,
>> +                                  bool quiet)
>> +{
>> +    const int default_block_size = 64 * 1024; /* 64K */
>> +
>> +    int ret;
>> +    BlockDriverInfo bdi;
>> +    int cluster_size1, cluster_size2, block_size;
>> +    const char *note = "Note: to alter it, set --block-size option.";
>> +    const char *fname1 = bs1->filename;
>> +    const char *fname2 = bs2->filename;
>> +
>> +    ret = bdrv_get_info(bs1, &bdi);
>> +    if (ret < 0 && ret != -ENOTSUP) {
>> +        error_report("Failed to get info of %s: %s", fname1, strerror(-ret));
>> +        return ret;
>> +    }
>> +    cluster_size1 = ret < 0 ? 0 : bdi.cluster_size;
>> +
>> +    ret = bdrv_get_info(bs2, &bdi);
>> +    if (ret < 0 && ret != -ENOTSUP) {
>> +        error_report("Failed to get info of %s: %s", fname2, strerror(-ret));
>> +        return ret;
>> +    }
>> +    cluster_size2 = ret < 0 ? 0 : bdi.cluster_size;
>> +
> 
> I’d feel better with an `assert(cluster_size1 >= 0 && cluster_size2 >= 0);` here.

Hmm.. But it seems obvious: bdi.cluster_size should not be <0 on success of bdrv_get_info.

> 
>> +    if (cluster_size1 > 0 && cluster_size2 > 0) {
>> +        if (cluster_size1 == cluster_size2) {
>> +            block_size = cluster_size1;
>> +        } else {
>> +            block_size = MIN(cluster_size1, cluster_size2);
>> +            qprintf(quiet, "%s and %s has different cluster sizes: %d and %d "
> 
> s/has/have/
> 
>> +                    "correspondingly. Use minimum as block-size for "
> 
> s/correspondingly/respectively/; s/Use/Using/ (“Use” sounds like an imperative)
> 
>> +                    "accuracy: %d. %s\n",
>> +                    fname1, fname2, cluster_size1,
>> +                    cluster_size2, block_size, note);
>> +        }
>> +    } else if (cluster_size1 == 0 && cluster_size2 == 0) {
>> +        block_size = default_block_size;
>> +        qprintf(quiet, "Neither of %s and %s has explicit cluster size. Use "
> 
> s/has/have an/; s/Use/Using/
> 
>> +                "default of %d bytes. %s\n", fname1, fname2, block_size, note);
>> +    } else {
>> +        block_size = MAX(cluster_size1, cluster_size2);
>> +        qprintf(quiet, "%s has explicit cluster size of %d and %s "
> 
> s/has/has an/
> 
>> +                "doesn't have one. Use %d as block-size. %s\n",
> 
> s/Use/Using/
> 
> Hanna
> 
>> +                cluster_size1 ? fname1 : fname2, block_size,
>> +                cluster_size1 ? fname2 : fname1,
>> +                block_size, note);
>> +    }
>> +
>> +    return block_size;
>> +}
>> +
>>   /*
>>    * Compares two images. Exit codes:
>>    *
> 


-- 
Best regards,
Vladimir


  reply	other threads:[~2021-10-28 10:06 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-21 10:12 [PATCH v2 0/4] qemu-img compare --stat Vladimir Sementsov-Ogievskiy
2021-10-21 10:12 ` [PATCH v2 1/4] qemu-img: implement " Vladimir Sementsov-Ogievskiy
2021-10-25 16:40   ` Hanna Reitz
2021-10-26  7:53     ` Vladimir Sementsov-Ogievskiy
2021-10-26  8:47       ` Hanna Reitz
2021-10-26  9:18         ` Vladimir Sementsov-Ogievskiy
2021-10-27  8:35           ` Hanna Reitz
2021-10-21 10:12 ` [PATCH v2 2/4] qemu-img: make --block-size optional for " Vladimir Sementsov-Ogievskiy
2021-10-26  8:07   ` Hanna Reitz
2021-10-28 10:04     ` Vladimir Sementsov-Ogievskiy [this message]
2021-10-21 10:12 ` [PATCH v2 3/4] qemu-img: add --shallow option for qemu-img compare Vladimir Sementsov-Ogievskiy
2021-10-26  8:11   ` Hanna Reitz
2021-10-21 10:12 ` [PATCH v2 4/4] iotests: add qemu-img-compare-stat test Vladimir Sementsov-Ogievskiy
2021-10-26  8:26   ` Hanna Reitz

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=d1f58c9e-f578-7017-282a-26deda1644c4@virtuozzo.com \
    --to=vsementsov@virtuozzo.com \
    --cc=den@openvz.org \
    --cc=eblake@redhat.com \
    --cc=hreitz@redhat.com \
    --cc=jsnow@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=nikita.lapshin@virtuozzo.com \
    --cc=nsoffer@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 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).