From: Peter Xu <peterx@redhat.com>
To: Felipe Franciosi <felipe@nutanix.com>
Cc: Juan Quintela <quintela@redhat.com>,
"Jason J. Herne" <jjherne@linux.vnet.ibm.com>,
"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>
Subject: Re: [Qemu-devel] [PATCH] migration: keep bytes_xfer_prev init'd to zero
Date: Tue, 23 May 2017 12:27:02 +0800 [thread overview]
Message-ID: <20170523042702.GC3873@pxdev.xzpeter.org> (raw)
In-Reply-To: <1495231142-19055-1-git-send-email-felipe@nutanix.com>
On Fri, May 19, 2017 at 10:59:02PM +0100, Felipe Franciosi wrote:
> The first time migration_bitmap_sync() is called, bytes_xfer_prev is set
> to ram_state.bytes_transferred which is, at this point, zero. The next
> time migration_bitmap_sync() is called, an iteration has happened and
> bytes_xfer_prev is set to 'x' bytes. Most likely, more than one second
> has passed, so the auto converge logic will be triggered and
> bytes_xfer_now will also be set to 'x' bytes.
>
> This condition is currently masked by dirty_rate_high_cnt, which will
> wait for a few iterations before throttling. It would otherwise always
> assume zero bytes have been copied and therefore throttle the guest
> (possibly) prematurely.
>
> Given bytes_xfer_prev is only used by the auto convergence logic, it
> makes sense to only set its value after a check has been made against
> bytes_xfer_now.
>
> Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
> ~
> ---
> migration/ram.c | 4 ----
> 1 file changed, 4 deletions(-)
>
> diff --git a/migration/ram.c b/migration/ram.c
> index f59fdd4..793af39 100644
> --- a/migration/ram.c
> +++ b/migration/ram.c
> @@ -670,10 +670,6 @@ static void migration_bitmap_sync(RAMState *rs)
>
> rs->bitmap_sync_count++;
>
> - if (!rs->bytes_xfer_prev) {
> - rs->bytes_xfer_prev = ram_bytes_transferred();
> - }
> -
> if (!rs->time_last_bitmap_sync) {
> rs->time_last_bitmap_sync = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
> }
> --
> 1.9.5
>
>
I feel like this patch wants to correctly initialize bytes_xfer_prev,
however I still see problem. E.g., when user specify auto-convergence
during migration, and in the first iteration we'll always have a very
small bytes_xfer_prev (with this patch, it'll be zero) with a very big
bytes_xfer_now (which is the ram_bytes_transferred() value). If so, do
you think squash below change together with current one would be more
accurate?
diff --git a/migration/ram.c b/migration/ram.c
index f59fdd4..e01a218 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -703,7 +703,8 @@ static void migration_bitmap_sync(RAMState *rs)
throttling */
bytes_xfer_now = ram_bytes_transferred();
- if (rs->dirty_pages_rate &&
+ /* Skip first iteration when bytes_xfer_prev not inited */
+ if (rs->bytes_xfer_prev && rs->dirty_pages_rate &&
(rs->num_dirty_pages_period * TARGET_PAGE_SIZE >
(bytes_xfer_now - rs->bytes_xfer_prev) / 2) &&
(rs->dirty_rate_high_cnt++ >= 2)) {
Even I am not sure whether we can move bytes_xfer_prev out of RAMState
since it's only used by migration_bitmap_sync().
Thanks,
--
Peter Xu
next prev parent reply other threads:[~2017-05-23 4:27 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-19 21:59 [Qemu-devel] [PATCH] migration: keep bytes_xfer_prev init'd to zero Felipe Franciosi
2017-05-23 4:27 ` Peter Xu [this message]
2017-05-24 13:02 ` Felipe Franciosi
2017-05-25 0:36 ` Peter Xu
2017-05-25 10:50 ` Felipe Franciosi
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=20170523042702.GC3873@pxdev.xzpeter.org \
--to=peterx@redhat.com \
--cc=felipe@nutanix.com \
--cc=jjherne@linux.vnet.ibm.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@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).