qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Alon Levy <alevy@redhat.com>
To: Yonit Halperin <yhalperi@redhat.com>
Cc: qemu-devel@nongnu.org, spice-devel@freedesktop.org
Subject: Re: [Qemu-devel] [Spice-devel] [PATCH v2 1/2] qxl: set only off-screen surfaces dirty instead of the whole vram
Date: Wed, 15 Feb 2012 11:37:34 +0200	[thread overview]
Message-ID: <20120215093734.GS15812@garlic.tlv.redhat.com> (raw)
In-Reply-To: <1329297736-3049-1-git-send-email-yhalperi@redhat.com>

On Wed, Feb 15, 2012 at 11:22:15AM +0200, Yonit Halperin wrote:
> We used to assure the guest surfaces were saved before migration by
> setting the whole vram dirty. This patch sets dirty only the areas
> that are actually used in the vram.
> 

Series is
Reviewed-by: Alon Levy <alevy@redhat.com>

> Signed-off-by: Yonit Halperin <yhalperi@redhat.com>
> ---
>  hw/qxl.c |   53 ++++++++++++++++++++++++++++++++++++++++++++---------
>  1 files changed, 44 insertions(+), 9 deletions(-)
> 
> diff --git a/hw/qxl.c b/hw/qxl.c
> index bc03c1d..df55de1 100644
> --- a/hw/qxl.c
> +++ b/hw/qxl.c
> @@ -1006,7 +1006,7 @@ static void qxl_reset_surfaces(PCIQXLDevice *d)
>      qxl_spice_destroy_surfaces(d, QXL_SYNC);
>  }
>  
> -/* called from spice server thread context only */
> +/* can be also called from spice server thread context */
>  void *qxl_phys2virt(PCIQXLDevice *qxl, QXLPHYSICAL pqxl, int group_id)
>  {
>      uint64_t phys   = le64_to_cpu(pqxl);
> @@ -1465,6 +1465,46 @@ static void qxl_hw_text_update(void *opaque, console_ch_t *chardata)
>      }
>  }
>  
> +static void qxl_dirty_surfaces(PCIQXLDevice *qxl)
> +{
> +    intptr_t vram_start;
> +    int i;
> +
> +    if (qxl->mode != QXL_MODE_NATIVE) {
> +        return;
> +    }
> +
> +    /* dirty the primary surface */
> +    qxl_set_dirty(&qxl->vga.vram, qxl->shadow_rom.draw_area_offset,
> +                  qxl->shadow_rom.surface0_area_size);
> +
> +    vram_start =  (intptr_t)memory_region_get_ram_ptr(&qxl->vram_bar);
> +
> +    /* dirty the off-screen surfaces */
> +    for (i = 0; i < NUM_SURFACES; i++) {
> +        QXLSurfaceCmd *cmd;
> +        intptr_t surface_offset;
> +        int surface_size;
> +
> +        if (qxl->guest_surfaces.cmds[i] == 0) {
> +            continue;
> +        }
> +
> +        cmd = qxl_phys2virt(qxl, qxl->guest_surfaces.cmds[i],
> +                            MEMSLOT_GROUP_GUEST);
> +        assert(cmd->type == QXL_SURFACE_CMD_CREATE);
> +        surface_offset = (intptr_t)qxl_phys2virt(qxl,
> +                                                 cmd->u.surface_create.data,
> +                                                 MEMSLOT_GROUP_GUEST);
> +        surface_offset -= vram_start;
> +        surface_size = cmd->u.surface_create.height *
> +                       abs(cmd->u.surface_create.stride);
> +        dprint(qxl, 3, "%s: dirty surface %d, offset %d, size %d\n", __func__,
> +               i, (int)surface_offset, surface_size);
> +        qxl_set_dirty(&qxl->vram_bar, surface_offset, surface_size);
> +    }
> +}
> +
>  static void qxl_vm_change_state_handler(void *opaque, int running,
>                                          RunState state)
>  {
> @@ -1478,14 +1518,9 @@ static void qxl_vm_change_state_handler(void *opaque, int running,
>           * called
>           */
>           qxl_update_irq(qxl);
> -    } else if (qxl->mode == QXL_MODE_NATIVE) {
> -        /* dirty all vram (which holds surfaces) and devram (primary surface)
> -         * to make sure they are saved */
> -        /* FIXME #1: should go out during "live" stage */
> -        /* FIXME #2: we only need to save the areas which are actually used */
> -        qxl_set_dirty(&qxl->vram_bar, 0, qxl->vram_size);
> -        qxl_set_dirty(&qxl->vga.vram, qxl->shadow_rom.draw_area_offset,
> -                      qxl->shadow_rom.surface0_area_size);
> +    } else {
> +        /* make sure surfaces are saved before migration */
> +        qxl_dirty_surfaces(qxl);
>      }
>  }
>  
> -- 
> 1.7.7.6
> 
> _______________________________________________
> Spice-devel mailing list
> Spice-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/spice-devel

  parent reply	other threads:[~2012-02-15  9:37 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-15  9:22 [Qemu-devel] [PATCH v2 1/2] qxl: set only off-screen surfaces dirty instead of the whole vram Yonit Halperin
2012-02-15  9:22 ` [Qemu-devel] [PATCH v2 2/2] qxl: make sure primary surface is saved on migration also in compat mode Yonit Halperin
2012-02-15  9:37 ` Alon Levy [this message]
2012-02-15 13:15 ` [Qemu-devel] [PATCH v2 1/2] qxl: set only off-screen surfaces dirty instead of the whole vram Gerd Hoffmann

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=20120215093734.GS15812@garlic.tlv.redhat.com \
    --to=alevy@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=spice-devel@freedesktop.org \
    --cc=yhalperi@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).