qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Hanna Czenczek <hreitz@redhat.com>
To: Alexander Ivanov <alexander.ivanov@virtuozzo.com>, qemu-devel@nongnu.org
Cc: qemu-block@nongnu.org, den@virtuozzo.com, stefanha@redhat.com,
	vsementsov@yandex-team.ru, kwolf@redhat.com
Subject: Re: [PATCH v5 2/5] parallels: Split image leak handling to separate check and fix helpers
Date: Fri, 2 Jun 2023 16:08:37 +0200	[thread overview]
Message-ID: <db062a52-3ee2-2f9b-7ef6-8c7e28e4e217@redhat.com> (raw)
In-Reply-To: <20230529151503.34006-3-alexander.ivanov@virtuozzo.com>

On 29.05.23 17:15, Alexander Ivanov wrote:
> We need to fix leak after deduplication in the next patch. Move leak
> fixing to a separate helper parallels_fix_leak() and add
> parallels_get_leak_size() helper wich used in parallels_fix_leak() and
> parallels_check_leak().
>
> Signed-off-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
> ---
>   block/parallels.c | 86 +++++++++++++++++++++++++++++++++--------------
>   1 file changed, 61 insertions(+), 25 deletions(-)
>
> diff --git a/block/parallels.c b/block/parallels.c
> index 1ec98c722b..64850b9655 100644
> --- a/block/parallels.c
> +++ b/block/parallels.c
> @@ -482,43 +482,79 @@ parallels_check_outside_image(BlockDriverState *bs, BdrvCheckResult *res,
>       return 0;
>   }
>   
> +static int64_t parallels_get_leak_size(BlockDriverState *bs,
> +                                       BdrvCheckResult *res)
> +{
> +    int64_t size;
> +
> +    size = bdrv_getlength(bs->file->bs);
> +    if (size < 0) {
> +        return size;
> +    }
> +
> +    /*
> +     * Before any usage of this function, image_end_offset has to be set to the
> +     * the highest offset in the BAT, excluding out-of-image offsets.
> +     */
> +    assert(size >= res->image_end_offset);

If `high_off == 0` in parallels_check_outside_image(), it will use 
s->data_end to determine image_end_offset, which is originally read from 
the image header.  I don’t see any place where we ensure that 
`s->data_end <= bdrv_getlength(bs->file->bs)`, so can we be certain the 
assertion holds even in that case?

> +
> +    return size - res->image_end_offset;
> +}
> +
> +static int parallels_fix_leak(BlockDriverState *bs,
> +                              BdrvCheckResult *res)
> +{
> +    Error *local_err = NULL;
> +    int64_t size;
> +    int ret;
> +
> +    size = parallels_get_leak_size(bs, res);
> +    if (size <= 0) {
> +        return size;
> +    }
> +
> +    /*
> +     * In order to really repair the image, we must shrink it.
> +     * That means we have to pass exact=true.
> +     */
> +    ret = bdrv_co_truncate(bs->file, res->image_end_offset, true,
> +                           PREALLOC_MODE_OFF, 0, &local_err);
> +    if (ret < 0) {
> +        error_report_err(local_err);
> +        return ret;
> +    }
> +
> +    return 0;
> +}
> +
>   static int coroutine_fn GRAPH_RDLOCK
>   parallels_check_leak(BlockDriverState *bs, BdrvCheckResult *res,
>                        BdrvCheckMode fix)
>   {
>       BDRVParallelsState *s = bs->opaque;
> -    int64_t size;
> +    int64_t count, leak_size;
>       int ret;
>   
> -    size = bdrv_getlength(bs->file->bs);
> -    if (size < 0) {
> +    leak_size = parallels_get_leak_size(bs, res);
> +    if (leak_size < 0) {
>           res->check_errors++;
> -        return size;
> +        return leak_size;
> +    }
> +    if (leak_size == 0) {
> +        return 0;
>       }
>   
> -    if (size > res->image_end_offset) {
> -        int64_t count;
> -        count = DIV_ROUND_UP(size - res->image_end_offset, s->cluster_size);
> -        fprintf(stderr, "%s space leaked at the end of the image %" PRId64 "\n",
> -                fix & BDRV_FIX_LEAKS ? "Repairing" : "ERROR",
> -                size - res->image_end_offset);
> -        res->leaks += count;
> -        if (fix & BDRV_FIX_LEAKS) {
> -            Error *local_err = NULL;
> +    count = DIV_ROUND_UP(leak_size, s->cluster_size);
> +    res->leaks += count;
> +    fprintf(stderr, "%s space leaked at the end of the image %" PRId64 "\n",
> +            fix & BDRV_FIX_LEAKS ? "Repairing" : "ERROR", leak_size);
>   
> -            /*
> -             * In order to really repair the image, we must shrink it.
> -             * That means we have to pass exact=true.
> -             */
> -            ret = bdrv_co_truncate(bs->file, res->image_end_offset, true,
> -                                   PREALLOC_MODE_OFF, 0, &local_err);
> -            if (ret < 0) {
> -                error_report_err(local_err);
> -                res->check_errors++;
> -                return ret;
> -            }
> -            res->leaks_fixed += count;
> +    if (fix & BDRV_FIX_LEAKS) {
> +        ret = parallels_fix_leak(bs, res);
> +        if (ret < 0) {
> +            return ret;

We used to increment res->check_errors here – should we still do that?

Hanna

>           }
> +        res->leaks_fixed += count;
>       }
>   
>       return 0;



  reply	other threads:[~2023-06-02 14:09 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-05-29 15:14 [PATCH v5 0/5] parallels: Add duplication check, repair at open, fix bugs Alexander Ivanov
2023-05-29 15:14 ` [PATCH v5 1/5] parallels: Incorrect data end calculation in parallels_open() Alexander Ivanov
2023-06-02 14:44   ` Hanna Czenczek
2023-05-29 15:15 ` [PATCH v5 2/5] parallels: Split image leak handling to separate check and fix helpers Alexander Ivanov
2023-06-02 14:08   ` Hanna Czenczek [this message]
2023-06-05 13:13     ` Alexander Ivanov
2023-05-29 15:15 ` [PATCH v5 3/5] parallels: Add checking and repairing duplicate offsets in BAT Alexander Ivanov
2023-06-02 14:43   ` Hanna Czenczek
2023-06-05 16:55     ` Alexander Ivanov
2023-05-29 15:15 ` [PATCH v5 4/5] parallels: Replace fprintf by qemu_log in check Alexander Ivanov
2023-06-02 14:48   ` Hanna Czenczek
2023-06-09 10:36     ` Alexander Ivanov
2023-06-09 10:59   ` Peter Maydell
2023-05-29 15:15 ` [PATCH v5 5/5] parallels: Image repairing in parallels_open() Alexander Ivanov
2023-06-02 14:59   ` Hanna Czenczek
2023-06-09 13:21     ` Alexander Ivanov
2023-06-09 13:41       ` Hanna Czenczek
2023-06-11 14:45         ` Alexander Ivanov

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=db062a52-3ee2-2f9b-7ef6-8c7e28e4e217@redhat.com \
    --to=hreitz@redhat.com \
    --cc=alexander.ivanov@virtuozzo.com \
    --cc=den@virtuozzo.com \
    --cc=kwolf@redhat.com \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    --cc=vsementsov@yandex-team.ru \
    /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).