All of lore.kernel.org
 help / color / mirror / Atom feed
From: Cameron Esfahani via <qemu-devel@nongnu.org>
To: qemu-devel@nongnu.org
Cc: kraxel@redhat.com
Subject: [PATCH v1] vnc: fix VNC artifacts
Date: Thu, 16 Jan 2020 19:50:58 -0800	[thread overview]
Message-ID: <7cb1f22ef8d575238c84f091f03dee8e7fedb78d.1579232679.git.dirty@apple.com> (raw)

Remove VNC optimization to reencode framebuffer update as raw if it's
smaller than the default encoding.  QEMU's implementation was naive and
didn't account for the ZLIB z_stream mutating with each compression.  Just
saving and restoring the output buffer offset wasn't sufficient to "rewind"
the previous encoding.  Considering that ZRLE is never larger than raw and
even though ZLIB can occasionally be fractionally larger than raw, the
overhead of implementing this optimization correctly isn't worth it.

In my investigation, ZRLE always compresses better than ZLIB so
prioritize ZRLE over ZLIB, even if the client hints that ZLIB is
preferred.

Fixes: <de3f7de7f4e257> ("vnc: allow fall back to RAW encoding")
Signed-off-by: Cameron Esfahani <dirty@apple.com>
---
 ui/vnc-enc-zrle.c |  4 ++--
 ui/vnc.c          | 30 +++++++++++-------------------
 2 files changed, 13 insertions(+), 21 deletions(-)

diff --git a/ui/vnc-enc-zrle.c b/ui/vnc-enc-zrle.c
index 17fd28a2e2..b4f71e32cf 100644
--- a/ui/vnc-enc-zrle.c
+++ b/ui/vnc-enc-zrle.c
@@ -98,8 +98,8 @@ static int zrle_compress_data(VncState *vs, int level)
     /* set pointers */
     zstream->next_in = vs->zrle->zrle.buffer;
     zstream->avail_in = vs->zrle->zrle.offset;
-    zstream->next_out = vs->zrle->zlib.buffer + vs->zrle->zlib.offset;
-    zstream->avail_out = vs->zrle->zlib.capacity - vs->zrle->zlib.offset;
+    zstream->next_out = vs->zrle->zlib.buffer;
+    zstream->avail_out = vs->zrle->zlib.capacity;
     zstream->data_type = Z_BINARY;
 
     /* start encoding */
diff --git a/ui/vnc.c b/ui/vnc.c
index 4100d6e404..f085e1b747 100644
--- a/ui/vnc.c
+++ b/ui/vnc.c
@@ -898,8 +898,6 @@ int vnc_raw_send_framebuffer_update(VncState *vs, int x, int y, int w, int h)
 int vnc_send_framebuffer_update(VncState *vs, int x, int y, int w, int h)
 {
     int n = 0;
-    bool encode_raw = false;
-    size_t saved_offs = vs->output.offset;
 
     switch(vs->vnc_encoding) {
         case VNC_ENCODING_ZLIB:
@@ -922,24 +920,11 @@ int vnc_send_framebuffer_update(VncState *vs, int x, int y, int w, int h)
             n = vnc_zywrle_send_framebuffer_update(vs, x, y, w, h);
             break;
         default:
-            encode_raw = true;
+            vnc_framebuffer_update(vs, x, y, w, h, VNC_ENCODING_RAW);
+            n = vnc_raw_send_framebuffer_update(vs, x, y, w, h);
             break;
     }
 
-    /* If the client has the same pixel format as our internal buffer and
-     * a RAW encoding would need less space fall back to RAW encoding to
-     * save bandwidth and processing power in the client. */
-    if (!encode_raw && vs->write_pixels == vnc_write_pixels_copy &&
-        12 + h * w * VNC_SERVER_FB_BYTES <= (vs->output.offset - saved_offs)) {
-        vs->output.offset = saved_offs;
-        encode_raw = true;
-    }
-
-    if (encode_raw) {
-        vnc_framebuffer_update(vs, x, y, w, h, VNC_ENCODING_RAW);
-        n = vnc_raw_send_framebuffer_update(vs, x, y, w, h);
-    }
-
     return n;
 }
 
@@ -2087,8 +2072,15 @@ static void set_encodings(VncState *vs, int32_t *encodings, size_t n_encodings)
             break;
 #endif
         case VNC_ENCODING_ZLIB:
-            vs->features |= VNC_FEATURE_ZLIB_MASK;
-            vs->vnc_encoding = enc;
+            /*
+             * VNC_ENCODING_ZRLE compresses better than VNC_ENCODING_ZLIB.
+             * So prioritize ZRLE, even if the client hints that it prefers
+             * ZLIB.
+             */
+            if ((vs->features & VNC_FEATURE_ZRLE_MASK) == 0) {
+                vs->features |= VNC_FEATURE_ZLIB_MASK;
+                vs->vnc_encoding = enc;
+            }
             break;
         case VNC_ENCODING_ZRLE:
             vs->features |= VNC_FEATURE_ZRLE_MASK;
-- 
2.24.0



             reply	other threads:[~2020-01-17  3:51 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-17  3:50 Cameron Esfahani via [this message]
2020-01-17  7:45 ` [PATCH v1] vnc: fix VNC artifacts Gerd Hoffmann
2020-01-17  7:48   ` Cameron Esfahani via
2020-01-17 21:43   ` Cameron Esfahani via
2020-01-18  6:24     ` Philippe Mathieu-Daudé

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=7cb1f22ef8d575238c84f091f03dee8e7fedb78d.1579232679.git.dirty@apple.com \
    --to=qemu-devel@nongnu.org \
    --cc=dirty@apple.com \
    --cc=kraxel@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 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.