| Message ID | 7cb1f22ef8d575238c84f091f03dee8e7fedb78d.1579232679.git.dirty@apple.com |
|---|---|
| State | New |
| Headers | show |
| Series | [v1] vnc: fix VNC artifacts | expand |
On Thu, Jan 16, 2020 at 07:50:58PM -0800, Cameron Esfahani wrote: > 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. So just revert de3f7de7f4e257 then ... > In my investigation, ZRLE always compresses better than ZLIB so > prioritize ZRLE over ZLIB, even if the client hints that ZLIB is > preferred. ... and make this a separate patch? cheers, Gerd
Yes. Personally, I'd also take the change to vnc-enc-zrle.c: because vs->zrle->zlib is reset at the top of the function, using vs->zrle->zlib.offset in determining zstream->next_out and zstream->avail_out is useless. Cameron Esfahani dirty@apple.com "All that is necessary for the triumph of evil is that good men do nothing." Edmund Burke > On Jan 16, 2020, at 11:45 PM, Gerd Hoffmann <kraxel@redhat.com> wrote: > > On Thu, Jan 16, 2020 at 07:50:58PM -0800, Cameron Esfahani wrote: >> 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. > > So just revert de3f7de7f4e257 then ... > >> In my investigation, ZRLE always compresses better than ZLIB so >> prioritize ZRLE over ZLIB, even if the client hints that ZLIB is >> preferred. > > ... and make this a separate patch? > > cheers, > Gerd >
I’m new to this process, what are the next steps? Cameron Esfahani dirty@apple.com > On Jan 16, 2020, at 11:47 PM, Gerd Hoffmann <kraxel@redhat.com> wrote: > > On Thu, Jan 16, 2020 at 07:50:58PM -0800, Cameron Esfahani wrote: >> 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. > > So just revert de3f7de7f4e257 then ... > >> In my investigation, ZRLE always compresses better than ZLIB so >> prioritize ZRLE over ZLIB, even if the client hints that ZLIB is >> preferred. > > ... and make this a separate patch? > > cheers, > Gerd > >
On 1/17/20 10:43 PM, Cameron Esfahani via wrote: > I’m new to this process, what are the next steps? Assuming your patch is in a branch named vncfix_v1: 1/ Start new branch based on the commit previous to your patch: - git checkout -b vncfix_v2 vncfix_v1~ 2/ Revert the offending patch, explain in commit description why: - git revert de3f7de7f4e257 3/ Apply the rest of your patch on top, git-cherry-pick is smart to directly use the diff context. Verify the patch is correct and rewrite the commit description: - git cherry-pick vncfix_v1 4/ Send the 2 patches as a series to the mailing list > > Cameron Esfahani > dirty@apple.com > >> On Jan 16, 2020, at 11:47 PM, Gerd Hoffmann <kraxel@redhat.com> wrote: >> >> On Thu, Jan 16, 2020 at 07:50:58PM -0800, Cameron Esfahani wrote: >>> 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. >> >> So just revert de3f7de7f4e257 then ... >> >>> In my investigation, ZRLE always compresses better than ZLIB so >>> prioritize ZRLE over ZLIB, even if the client hints that ZLIB is >>> preferred. >> >> ... and make this a separate patch? >> >> cheers, >> Gerd >> >> >
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;
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(-)