| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2009-05-08 | R3xx/R4xx tex vid: increase the guardband limit for rendering with a tri | Alex Deucher | |
| limit goes from 2880 to 4021 now that we've switched to 1/12 subpixel mode. This allows us to render with a clipped tri at higher resolutions preventing diagonal tearing. | |||
| 2009-05-08 | r3xx-r5xx: switch to 1/12 subpixel precision | Alex Deucher | |
| - based on similar patch in Jerome's cs ddx tree - also fix clipping offsets - should eventually allow for 4k render targets - mesa driver uses 1/12 mode, this avoids changing the subpixel mode when switching between ddx and mesa | |||
| 2009-05-08 | R3xx-R5xx: don't set TX_OFFSET_RS in RS_INST_COUNT | Alex Deucher | |
| Isn't necessary and seems to cause problems for RS690 users | |||
| 2009-04-07 | radeon: fix Xv vsync for multi-head | Alex Deucher | |
| 2009-03-24 | don't convert planar yuv to packed for r200 | Roland Scheidegger | |
| uses 3 textures for planar yuv and does yuv->rgb conversion in the shader. Similar to r300 code, but might have precision issues - hardware alu should have enough precision but hardware consts are only 8bit and we'd want at least 11. This also enables textured video on rv250 (and also supports packed yuv on that chip by using basically the same shader with packed data). | |||
| 2009-03-24 | don't convert planar yuv to packed for r300 | Roland Scheidegger | |
| uses 3 textures for planar yuv and does yuv->rgb conversion in the shader. small performance advantage, but manual texture cache setting is necessary otherwise it may be measurably slower (but probably not relevant) in some cases. Unlike some other drivers, using MADs instead of DP3s, since this requires less instructions due to no MOVs are required, the end result is the same though the constants need to be different. Use of this is user settable for now (XV_HWPLANAR attrib). | |||
| 2009-02-24 | Only call RADEONWaitForVLine if it might actually do anything useful. | Michel Dänzer | |
| Otherwise it and RADEONBiggerCrtcArea show up pretty high in x11perf -aa*text profiles. | |||
| 2009-01-12 | R3xx/R4xx: Maximize the use of clipped triangles for Xv rendering | Christiaan van Dijk | |
| While trying to get tear free video on my RS690 I found some problems in the code for rendering the video as a clipped triangle. The size limitations are quite strict and scaling the video output window creates all kinds of weird effects. This patch maximizes the use of this function and resolves the scaling issues. There could be a problem with zero target height/width, not sure if this is handled elsewhere. | |||
| 2008-12-22 | R3xx-R5xx: better fix for xv primitive tearing issues | Alex Deucher | |
| R5xx: always use single clipped triangle R3xx/R4xx: use single clipped triangle up to guardband limit, then use quad. | |||
| 2008-12-17 | R2xx: switch EXA composite to rects rather quads | Alex Deucher | |
| For conistency with Xv path and to reduce potential diagonal tearing | |||
| 2008-12-15 | R3xx-R5xx: switch back to quad rendering for Xv | Alex Deucher | |
| The large clipped triangle trick hits the guardband limits on large displays. This gets textured video working again until we sort out how to deal with the situation better. See bug 19046 | |||
| 2008-12-05 | Fix scissor setup for Xv | Pierre Ossman | |
| We need to adjust the scissoring for each clip box, so move the setup to inside the while loop. | |||
| 2008-12-05 | Make VSync for EXA and Xv configurable | Alex Deucher | |
| 2008-12-05 | Optimise RADEONWaitForVLine | Pierre Ossman | |
| Only avoid the vlines we are rendering to, instead of the entire screen. This way we don't stall the card for longer than we absolutely have to. EXA calls fixed by Alex Deucher. | |||
| 2008-12-04 | Improve tearing avoidance for Xvideo in two steps | Pierre Ossman | |
| - Fix up VLINE handling to trigger whenever scanout is outside the visible area. - Render the video as a scissored triangle as R300+ cannot render a quad in a single pass. | |||
| 2008-12-04 | First pass at tear-free accel | Alex Deucher | |
| if the dest pixmap is the front buffer, stall the pipe until the vline is outside the active area. For EXA, pick crtc based on the larger mode area; ideally we'd have one pixmap per crtc. For Xv, use dst window area to determine crtc. | |||
| 2008-12-04 | Switch r200 Xv to use rect lists rather than quads to avoid tearing | Pierre Ossman | |
| Patch from Pierre with fixes from me. | |||
| 2008-12-04 | Fix up US setup after r3xx bicubic merge | Alex Deucher | |
| 2008-12-04 | Merge branch 'bicubic' of git://git.infradead.org/users/drzeus/xf86-video-ati | Alex Deucher | |
| 2008-12-03 | Make sure 3D state is emited when switching to 3D IB | Alex Deucher | |
| - fixes bug 18864 | |||
| 2008-12-03 | Make the R300 bicubic shader program a bit easier to understand by | Pierre Ossman | |
| having comments that describe the operation, not how it has to be implemented in the hardware. | |||
| 2008-12-03 | Now that we have a bicubic code path that messes up the pixel shader, | Pierre Ossman | |
| we need to restore it in the non-bicubic code. | |||
| 2008-12-03 | Fix node setup on R300 bicubic pixel shader. It was backwards and used | Pierre Ossman | |
| the wrong base for sizes. | |||
| 2008-12-02 | move 3d state init after RADEON_SWITCH_TO_3D() | Alex Deucher | |
| 2008-12-02 | Stability fixes from radeon-gem-cs | Dave Airlie | |
| 2008-12-01 | Fix comments for R500 fragment shader to reflect the code. | Pierre Ossman | |
| 2008-12-01 | Fix indentation on IS_R300_3D Xv code. | Corbin Simpson | |
| 2008-12-01 | r3xx: Various Xv fixes. | Corbin Simpson | |
| 2008-12-01 | Switch to Mesa-style 24-bit float packing. | Corbin Simpson | |
| Seems like Mesa's got it right, so we should follow suit. | |||
| 2008-12-01 | Add r3xx-specific bicubic filtering code. FP, RS, mostly, everything else is ↵ | Maciej Cencora | |
| shared with r5xx. | |||
| 2008-08-25 | Move accel state to a separate struct | Alex Deucher | |
| 2008-08-25 | match textured video macro names with other accel code | Alex Deucher | |
| OUT_VIDEO_REG() -> OUT_ACCEL_REG() etc. | |||
| 2008-08-25 | Additional cleanups and re-arragement following bicubic merge | Alex Deucher | |
| 2008-08-25 | Whitespace cleanup from bucubic merge | Alex Deucher | |
| 2008-08-25 | Fix typos. | Dennis Kasprzyk | |
| 2008-08-25 | Fix bicubic fp calculation. | Dennis Kasprzyk | |
| 2008-08-25 | Fixed bicubic fragment program comments. | Dennis Kasprzyk | |
| 2008-08-25 | Switch from 32-bit floats to 16-bit half-floats. | Corbin Simpson | |
| Massive bandwidth savings, or so I'm told. Yay? | |||
| 2008-08-25 | Move some ALU instructions after the TEX instructions, so we can do ↵ | Dennis Kasprzyk | |
| something usefull while we are waiting for the texture values. | |||
| 2008-08-25 | Another uneeded SEM_AQUIRE. | Dennis Kasprzyk | |
| 2008-08-25 | Smarter usage of the texture semaphore. | Dennis Kasprzyk | |
| 2008-08-25 | Remove one constant. | Dennis Kasprzyk | |
| 2008-08-25 | Heavy optimizations. | Dennis Kasprzyk | |
| 2008-08-25 | Remove uneeded negations. | Dennis Kasprzyk | |
| 2008-08-25 | Implement LPR in one instruction. | Dennis Kasprzyk | |
| 2008-08-25 | Set helper texture filter correctly. | Dennis Kasprzyk | |
| 2008-08-25 | Oops, made a mistake with vertices. | Corbin Simpson | |
| 2008-08-25 | Merge upstream changes to vertices, and also add Xv attributes for textured ↵ | Corbin Simpson | |
| video, including bicubic filtering. | |||
| 2008-08-25 | Fix constants. | Corbin Simpson | |
| 2008-08-25 | Finally got the fragment program fully working for bicubic filtering on r5xx. | Corbin Simpson | |
 |
index : driver/xf86-video-ati | |
| ATI video cards driver |
| summaryrefslogtreecommitdiff |