All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
To: Chuan Zheng <zhengchuan@huawei.com>
Cc: zhang.zhanghailiang@huawei.com, quintela@redhat.com,
	linyilu@huawei.com, qemu-devel@nongnu.org, alex.chen@huawei.com,
	ann.zhuangyanying@huawei.com, fangying1@huawei.com
Subject: Re: [RFC PATCH 5/8] migration/dirtyrate: Compare hash results for recorded ramblock
Date: Tue, 4 Aug 2020 18:29:01 +0100	[thread overview]
Message-ID: <20200804172901.GH2659@work-vm> (raw)
In-Reply-To: <1595646669-109310-6-git-send-email-zhengchuan@huawei.com>

* Chuan Zheng (zhengchuan@huawei.com) wrote:
> From: Zheng Chuan <zhengchuan@huawei.com>
> 
> Compare hash results for recorded ramblock.
> 
> Signed-off-by: Zheng Chuan <zhengchuan@huawei.com>
> Signed-off-by: YanYing Zhang <ann.zhuangyanying@huawei.com>
> ---
>  migration/dirtyrate.c | 77 +++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 77 insertions(+)
> 
> diff --git a/migration/dirtyrate.c b/migration/dirtyrate.c
> index 45cfc91..7badc53 100644
> --- a/migration/dirtyrate.c
> +++ b/migration/dirtyrate.c
> @@ -202,6 +202,83 @@ static int record_block_hash_info(struct dirtyrate_config config,
>      return 0;
>  }
>  
> +static int cal_block_dirty_rate(struct block_dirty_info *info)
> +{
> +    uint8_t *md = NULL;
> +    size_t hash_len;
> +    int i;
> +    int ret = 0;
> +
> +    hash_len = qcrypto_hash_digest_len(QCRYPTO_HASH_ALG_MD5);
> +    md = g_new0(uint8_t, hash_len);

Is 'hash_len' actually constant for a given algorithm, like MD5 ?
i.e. can we just have a nice fixed size array?

> +    for (i = 0; i < info->sample_pages_count; i++) {
> +        ret = get_block_vfn_hash(info, info->sample_page_vfn[i], &md, &hash_len);
> +        if (ret < 0) {
> +            goto out;
> +        }
> +
> +        if (memcmp(md, info->hash_result + i * hash_len, hash_len) != 0) {
> +            info->sample_dirty_count++;

When the page doesn't match, do we have to update info->hash_result with
the new hash?   If the page is only modified once, and we catch it on
this cycle, we wouldn't want to catch it next time around.

> +        }
> +    }
> +
> +out:
> +    g_free(md);
> +    return ret;
> +}
> +
> +static bool find_block_matched(RAMBlock *block, struct block_dirty_info *infos,
> +                               int count, struct block_dirty_info **matched)
> +{
> +    int i;
> +
> +    for (i = 0; i < count; i++) {
> +        if (!strcmp(infos[i].idstr, qemu_ram_get_idstr(block))) {
> +            break;
> +        }
> +    }
> +
> +    if (i == count) {
> +        return false;
> +    }
> +
> +    if (infos[i].block_addr != qemu_ram_get_host_addr(block) ||
> +        infos[i].block_pages !=
> +            (qemu_ram_get_used_length(block) >> DIRTYRATE_PAGE_SIZE_SHIFT)) {

How does this happen?

> +        return false;
> +    }
> +
> +    *matched = &infos[i];
> +    return true;
> +}
> +
> +static int compare_block_hash_info(struct block_dirty_info *info, int block_index)
> +{
> +    struct block_dirty_info *block_dinfo = NULL;
> +    RAMBlock *block = NULL;
> +
> +    RAMBLOCK_FOREACH_MIGRATABLE(block) {
> +        if (ram_block_skip(block) < 0) {
> +            continue;
> +        }
> +        block_dinfo = NULL;
> +        if (!find_block_matched(block, info, block_index + 1, &block_dinfo)) {
> +            continue;
> +        }
> +        if (cal_block_dirty_rate(block_dinfo) < 0) {
> +            return -1;
> +        }
> +        update_dirtyrate_stat(block_dinfo);
> +    }
> +    if (!dirty_stat.total_sample_count) {
> +        return -1;
> +    }
> +
> +    return 0;
> +}
> +
> +
>  static void calculate_dirtyrate(struct dirtyrate_config config, int64_t time)
>  {
>      /* todo */
> -- 
> 1.8.3.1
> 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK



  reply	other threads:[~2020-08-04 17:33 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-25  3:11 [RFC PATCH 0/8] *** A Method for evaluating dirty page rate *** Chuan Zheng
2020-07-25  3:11 ` [RFC PATCH 1/8] migration/dirtyrate: Add get_dirtyrate_thread() function Chuan Zheng
2020-08-04 16:23   ` Dr. David Alan Gilbert
2020-08-06  7:36     ` Zheng Chuan
2020-07-25  3:11 ` [RFC PATCH 2/8] migration/dirtyrate: Add block_dirty_info to store dirtypage info Chuan Zheng
2020-08-04 16:28   ` Dr. David Alan Gilbert
2020-08-06  7:37     ` Zheng Chuan
2020-08-06 16:59       ` Dr. David Alan Gilbert
2020-08-07  6:19         ` Zheng Chuan
2020-07-25  3:11 ` [RFC PATCH 3/8] migration/dirtyrate: Add dirtyrate statistics series functions Chuan Zheng
2020-08-04 16:44   ` Dr. David Alan Gilbert
2020-07-25  3:11 ` [RFC PATCH 4/8] migration/dirtyrate: Record hash results for each ramblock Chuan Zheng
2020-08-04 17:00   ` Dr. David Alan Gilbert
2020-08-06  7:37     ` Zheng Chuan
2020-07-25  3:11 ` [RFC PATCH 5/8] migration/dirtyrate: Compare hash results for recorded ramblock Chuan Zheng
2020-08-04 17:29   ` Dr. David Alan Gilbert [this message]
2020-08-11  8:42     ` Zheng Chuan
2020-07-25  3:11 ` [RFC PATCH 6/8] migration/dirtyrate: Implement get_sample_gap_period() and block_sample_gap_period() Chuan Zheng
2020-08-04 17:52   ` Dr. David Alan Gilbert
2020-07-25  3:11 ` [RFC PATCH 7/8] migration/dirtyrate: Implement calculate_dirtyrate() function Chuan Zheng
2020-08-04 17:57   ` Dr. David Alan Gilbert
2020-07-25  3:11 ` [RFC PATCH 8/8] migration/dirtyrate: Implement qmp_cal_dirty_rate()/qmp_get_dirty_rate() function Chuan Zheng
2020-08-04 16:28   ` Eric Blake
2020-08-06  7:37     ` Zheng Chuan
2020-08-04 16:34   ` Eric Blake
2020-08-04 18:02   ` Dr. David Alan Gilbert
2020-08-04 16:19 ` [RFC PATCH 0/8] *** A Method for evaluating dirty page rate *** Dr. David Alan Gilbert
2020-08-06  7:36   ` Zheng Chuan
2020-08-06 16:58     ` Dr. David Alan Gilbert
2020-08-07  6:13       ` Zheng Chuan

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=20200804172901.GH2659@work-vm \
    --to=dgilbert@redhat.com \
    --cc=alex.chen@huawei.com \
    --cc=ann.zhuangyanying@huawei.com \
    --cc=fangying1@huawei.com \
    --cc=linyilu@huawei.com \
    --cc=qemu-devel@nongnu.org \
    --cc=quintela@redhat.com \
    --cc=zhang.zhanghailiang@huawei.com \
    --cc=zhengchuan@huawei.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 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.