[Intel-gfx] [PATCH] Xv: free tearing on textured video
Xiang, Haihao
haihao.xiang at intel.com
Wed Feb 25 09:18:02 CET 2009
On Wed, 2009-02-25 at 15:56 +0800, Zou, Nanhai wrote:
> >-----Original Message-----
> >From: intel-gfx-bounces at lists.freedesktop.org
> >[mailto:intel-gfx-bounces at lists.freedesktop.org] On Behalf Of Xiang, Haihao
> >Sent: 2009年2月25日 15:35
> >To: intel-gfx at lists.freedesktop.org
> >Subject: [Intel-gfx] [PATCH] Xv: free tearing on textured video
> >
> >Synchronize the screen update to vblank and add an option
> >XvTexturedVsync to control it.
> >---
> > man/intel.man | 5 +++++
> > src/i810_reg.h | 4 ++++
> > src/i830.h | 1 +
> > src/i830_driver.c | 3 +++
> > src/i830_video.c | 53
> >++++++++++++++++++++++++++++++++++++++---------------
> > 5 files changed, 51 insertions(+), 15 deletions(-)
> >
> >diff --git a/man/intel.man b/man/intel.man
> >index c7a3c61..2738ef6 100644
> >--- a/man/intel.man
> >+++ b/man/intel.man
> >@@ -202,6 +202,11 @@ information.
> > Enable XvMC driver. Current support MPEG2 MC on 915/945 and G33 series.
> > User should provide absolute path to libIntelXvMC.so in XvMCConfig file.
> > Default: Disabled.
> >+.TP
> >+.BI "Option \*qXvTexturedVsync\*q \*q" boolean \*q
> >+This option controls whether textured adapter synchronizes the screen update
> >to
> >+the vblank.
> >+Default: disabled.
> >
> > .SH OUTPUT CONFIGURATION
> > On 830M and better chipsets, the driver supports runtime configuration of
> >diff --git a/src/i810_reg.h b/src/i810_reg.h
> >index e2ffba1..b6aa769 100644
> >--- a/src/i810_reg.h
> >+++ b/src/i810_reg.h
> >@@ -2432,7 +2432,11 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
> >
> > /* Wait for Events */
> > #define MI_WAIT_FOR_EVENT (0x03<<23)
> >+#define MI_WAIT_FOR_PIPEB_SVBLANK (1<<18)
> >+#define MI_WAIT_FOR_PIPEA_SVBLANK (1<<17)
> > #define MI_WAIT_FOR_OVERLAY_FLIP (1<<16)
> >+#define MI_WAIT_FOR_PIPEB_VBLANK (1<<7)
> >+#define MI_WAIT_FOR_PIPEA_VBLANK (1<<3)
> >
> > /* Flush */
> > #define MI_FLUSH (0x04<<23)
> >diff --git a/src/i830.h b/src/i830.h
> >index 7904b9f..735b0c1 100644
> >--- a/src/i830.h
> >+++ b/src/i830.h
> >@@ -555,6 +555,7 @@ typedef struct _I830Rec {
> > Bool XvDisabled; /* Xv disabled in PreInit. */
> > Bool XvEnabled; /* Xv enabled for this generation. */
> > Bool XvPreferOverlay;
> >+ Bool XvTexturedVsync;
> >
> > #ifdef I830_XV
> > int colorKey;
> >diff --git a/src/i830_driver.c b/src/i830_driver.c
> >index 0a8a9c6..5d7285c 100644
> >--- a/src/i830_driver.c
> >+++ b/src/i830_driver.c
> >@@ -318,6 +318,7 @@ typedef enum {
> > OPTION_XVMC,
> > #endif
> > OPTION_PREFER_OVERLAY,
> >+ OPTION_TEXTURED_VSYNC,
> > } I830Opts;
> >
> > static OptionInfoRec I830Options[] = {
> >@@ -343,6 +344,7 @@ static OptionInfoRec I830Options[] = {
> > {OPTION_XVMC, "XvMC", OPTV_BOOLEAN, {0}, TRUE},
> > #endif
> > {OPTION_PREFER_OVERLAY, "XvPreferOverlay", OPTV_BOOLEAN, {0}, FALSE},
> >+ {OPTION_TEXTURED_VSYNC, "XvTexturedVsync", OPTV_BOOLEAN, {0}, FALSE},
> > {-1, NULL, OPTV_NONE, {0}, FALSE}
> > };
> > /* *INDENT-ON* */
> >@@ -1777,6 +1779,7 @@ I830XvInit(ScrnInfoPtr pScrn)
> > !xf86ReturnOptValBool(pI830->Options, OPTION_XVIDEO, TRUE);
> >
> > pI830->XvPreferOverlay = xf86ReturnOptValBool(pI830->Options,
> >OPTION_PREFER_OVERLAY, FALSE);
> >+ pI830->XvTexturedVsync = xf86ReturnOptValBool(pI830->Options,
> >OPTION_TEXTURED_VSYNC, FALSE);
> >
> > #ifdef I830_XV
> > if (xf86GetOptValInteger(pI830->Options, OPTION_VIDEO_KEY,
> >diff --git a/src/i830_video.c b/src/i830_video.c
> >index c9a0181..eea6a5b 100644
> >--- a/src/i830_video.c
> >+++ b/src/i830_video.c
> >@@ -2303,7 +2303,7 @@ I830PutImage(ScrnInfoPtr pScrn,
> > dstBox.y2 = drw_y + drw_h;
> >
> > if (!i830_clip_video_helper(pScrn,
> >- pPriv->textured ? NULL : &crtc,
> >+ &crtc,
> > &dstBox, &x1, &x2, &y1, &y2, clipBoxes,
> > width, height))
> > return Success;
> >@@ -2528,22 +2528,45 @@ I830PutImage(ScrnInfoPtr pScrn,
> > REGION_COPY(pScrn->pScreen, &pPriv->clip, clipBoxes);
> > i830_fill_colorkey (pScreen, pPriv->colorKey, clipBoxes);
> > }
> >- } else if (IS_I965G(pI830)) {
> >-
> >+ } else {
> >+ if (pI830->XvTexturedVsync) {
> >+ I830CrtcPrivatePtr intel_crtc = crtc->driver_private;
> >+ int event;
> >+
> >+ if (IS_I965G(pI830)) {
> >+ if (intel_crtc->pipe == 0)
> >+ event = MI_WAIT_FOR_PIPEA_SVBLANK;
> >+ else
> >+ event = MI_WAIT_FOR_PIPEB_SVBLANK;
> >+ } else {
> >+ if (intel_crtc->pipe == 0)
> >+ event = MI_WAIT_FOR_PIPEA_VBLANK;
> >+ else
> >+ event = MI_WAIT_FOR_PIPEB_VBLANK;
> >+ }
> >+
> >+ BEGIN_BATCH(2);
> >+ OUT_BATCH(MI_WAIT_FOR_EVENT | event);
> >+ OUT_BATCH(MI_NOOP);
> >+ ADVANCE_BATCH();
> >+ }
> >+
> >+ if (IS_I965G(pI830)) {
> > #ifdef INTEL_XVMC
> >- if (id == FOURCC_XVMC && pPriv->rotation == RR_Rotate_0) {
> >- pPriv->YBuf0offset = buf - pI830->FbBase;
> >- pPriv->UBuf0offset = pPriv->YBuf0offset + height*width;
> >- pPriv->VBuf0offset = pPriv->UBuf0offset + height*width/4;
> >- }
> >+ if (id == FOURCC_XVMC && pPriv->rotation == RR_Rotate_0) {
> >+ pPriv->YBuf0offset = buf - pI830->FbBase;
> >+ pPriv->UBuf0offset = pPriv->YBuf0offset + height*width;
> >+ pPriv->VBuf0offset = pPriv->UBuf0offset + height*width/4;
> >+ }
> > #endif
> >- I965DisplayVideoTextured(pScrn, pPriv, destId, clipBoxes, width, height,
> >- dstPitch, x1, y1, x2, y2,
> >- src_w, src_h, drw_w, drw_h, pPixmap);
> >- } else {
> >- I915DisplayVideoTextured(pScrn, pPriv, destId, clipBoxes, width, height,
> >- dstPitch, dstPitch2, x1, y1, x2, y2,
> >- src_w, src_h, drw_w, drw_h, pPixmap);
> >+ I965DisplayVideoTextured(pScrn, pPriv, destId, clipBoxes, width,
> >height,
> >+ dstPitch, x1, y1, x2, y2,
> >+ src_w, src_h, drw_w, drw_h, pPixmap);
> >+ } else {
> >+ I915DisplayVideoTextured(pScrn, pPriv, destId, clipBoxes, width,
> >height,
> >+ dstPitch, dstPitch2, x1, y1, x2, y2,
> >+ src_w, src_h, drw_w, drw_h, pPixmap);
> >+ }
> > }
> > if (pPriv->textured) {
> > DamageDamageRegion(pDraw, clipBoxes);
> >--
> >1.5.6.3
> >
>
> I suggest turn this option default to be true or not use an option at all.
> We already have too many options to handle.
Users maybe care about X performance and they need to control it and
there is not any dependency of this option.
Haihao
>
> Thanks
> Zou Nanhai
> >_______________________________________________
> >Intel-gfx mailing list
> >Intel-gfx at lists.freedesktop.org
> >http://lists.freedesktop.org/mailman/listinfo/intel-gfx
More information about the Intel-gfx
mailing list