<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Dec 5, 2016 at 2:51 PM, Frediano Ziglio <span dir="ltr"><<a href="mailto:fziglio@redhat.com" target="_blank">fziglio@redhat.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">ping<br>
<br>
Maybe this requires some follow ups for VgaDevice and<br>
possibly avoinind taking into account moves during mapping<br>
computation.<br>
<span class="gmail-HOEnZb"><font color="#888888"><br></font></span></blockquote><div><br></div><div>Indeed, it seems GetMaxSourceMappingHeight(NULL, 0,...) will be also correct.</div><div>From other side, I think mapping of entire source buffer to kernel space is redundant,<br></div><div>as we never touch the source memory in context different than OS call.</div><div>It looks like the copy operation can use original source address and be just protected by exception handler.</div><div>(the prototype - MSFT BDD example executes the copy in different thread, so the mapping is required).</div><div><br></div><div>Regarding Vga device - I do not see why copying from device memory is better than copying from source memory.</div><div>Does VgaDevice have the same problem of misalign?</div><div><br></div><div>Thanks,</div><div>Yuri</div><div> <br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><span class="gmail-HOEnZb"><font color="#888888">
Frediano<br>
</font></span><div class="gmail-HOEnZb"><div class="gmail-h5"><br>
><br>
> Doing some fast check on Windows 8.1 you could note that<br>
> moving windows you got a weird effect were windows were a<br>
> bit misaligned.<br>
> As documented in DXGKARG_PRESENT_DISPLAYONLY page<br>
> NumMoves/pMoves fields point to an array of screen-to-screen<br>
> moves while our code implemented them as an image drawing<br>
> (the same implementation of dirty rects) causing the weird<br>
> effect mentioned.<br>
> This patch implement the moves using QXL_COPY_BITS operation<br>
> instead of a QXL_DRAW_COPY fixing the issue and avoiding sending<br>
> image to the server making the move/scroll operations on the<br>
> guest faster (and taking less bandwidth).<br>
> It seems that Windows 10 doesn't send the move commands but<br>
> instead send only dirty rects so you can't note this problem<br>
> using Windows 10.<br>
><br>
> Signed-off-by: Frediano Ziglio <<a href="mailto:fziglio@redhat.com">fziglio@redhat.com</a>><br>
> ---<br>
> qxldod/QxlDod.cpp | 26 +++++++++++++++++++++-----<br>
> qxldod/QxlDod.h | 1 +<br>
> 2 files changed, 22 insertions(+), 5 deletions(-)<br>
><br>
> diff --git a/qxldod/QxlDod.cpp b/qxldod/QxlDod.cpp<br>
> index 68e3383..7adbec4 100755<br>
> --- a/qxldod/QxlDod.cpp<br>
> +++ b/qxldod/QxlDod.cpp<br>
> @@ -3738,11 +3738,7 @@ QxlDevice::<wbr>ExecutePresentDisplayOnly(<br>
> DbgPrint(TRACE_LEVEL_<wbr>INFORMATION, ("--- %d SourcePoint.x = %ld,<br>
> SourcePoint.y = %ld, DestRect.bottom = %ld, DestRect.left = %ld,<br>
> DestRect.right = %ld, DestRect.top = %ld\n",<br>
> i , pSourcePoint->x, pSourcePoint->y, pDestRect->bottom,<br>
> pDestRect->left, pDestRect->right, pDestRect->top));<br>
><br>
> - BltBits(&DstBltInfo,<br>
> - &SrcBltInfo,<br>
> - 1,<br>
> - pDestRect,<br>
> - pSourcePoint);<br>
> + CopyBits(*pDestRect, *pSourcePoint);<br>
> }<br>
><br>
> // Copy all the dirty rects from source image to video frame buffer.<br>
> @@ -4184,6 +4180,26 @@ VOID QxlDevice::SetImageId(<wbr>InternalImage *internal,<br>
> }<br>
> }<br>
><br>
> +void QxlDevice::CopyBits(const RECT& rect, const POINT& sourcePoint)<br>
> +{<br>
> + PAGED_CODE();<br>
> + QXLDrawable *drawable;<br>
> +<br>
> + DbgPrint(TRACE_LEVEL_VERBOSE, ("---> %s device %d\n",<br>
> __FUNCTION__,m_Id));<br>
> +<br>
> + if (!(drawable = Drawable(QXL_COPY_BITS, &rect, NULL, 0))) {<br>
> + DbgPrint(TRACE_LEVEL_ERROR, ("Cannot get Drawable.\n"));<br>
> + return;<br>
> + }<br>
> +<br>
> + drawable->u.copy_bits.src_pos.<wbr>x = sourcePoint.x;<br>
> + drawable->u.copy_bits.src_pos.<wbr>y = sourcePoint.y;<br>
> +<br>
> + PushDrawable(drawable);<br>
> +<br>
> + DbgPrint(TRACE_LEVEL_VERBOSE, ("<--- %s\n", __FUNCTION__));<br>
> +}<br>
> +<br>
> VOID QxlDevice::BltBits (<br>
> BLT_INFO* pDst,<br>
> CONST BLT_INFO* pSrc,<br>
> diff --git a/qxldod/QxlDod.h b/qxldod/QxlDod.h<br>
> index b151484..324c3d6 100755<br>
> --- a/qxldod/QxlDod.h<br>
> +++ b/qxldod/QxlDod.h<br>
> @@ -487,6 +487,7 @@ protected:<br>
> UINT NumRects,<br>
> _In_reads_(NumRects) CONST RECT *pRects,<br>
> POINT* pSourcePoint);<br>
> + void QxlDevice::CopyBits(const RECT& rect, const POINT& sourcePoint);<br>
> QXLDrawable *Drawable(UINT8 type,<br>
> CONST RECT *area,<br>
> CONST RECT *clip,<br>
<br>
</div></div><div class="gmail-HOEnZb"><div class="gmail-h5">______________________________<wbr>_________________<br>
Spice-devel mailing list<br>
<a href="mailto:Spice-devel@lists.freedesktop.org">Spice-devel@lists.freedesktop.<wbr>org</a><br>
<a href="https://lists.freedesktop.org/mailman/listinfo/spice-devel" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/spice-devel</a><br>
</div></div></blockquote></div><br></div></div>