diff options
author | Marek Olšák <[email protected]> | 2011-01-25 05:37:52 +0100 |
---|---|---|
committer | Marek Olšák <[email protected]> | 2011-01-27 18:12:01 +0100 |
commit | db299a9f8244d53d9041fcdbd396a77ebe1f9e3e (patch) | |
tree | 60adc94f4a132101be84579bf57988283e20ef90 /src/gallium/drivers/r300/r300_chipset.h | |
parent | 7a4345fd83605695dc641af503f6e87b808b48d7 (diff) |
r300g: fix some bugs with zbuffer compression (v4)
This drops the memblock manager for ZMASK. Instead, only one zbuffer can be
compressed at a time. Note that this does not necessarily have to be slower.
When there is a large number of zbuffers, compression might be used more often
than it was before. It's also easier to debug.
How it works:
1) 'clear' turns the compression on.
2) If some other zbuffer is set or the currently-bound zbuffer is used
for texturing, the driver decompresses it and then turns the compression off.
Notes:
- The ZMASK clear has been refactored, so that only one packet3 is used to clear
ZMASK.
- The 8x8 compression mode is disabled. I couldn't make it work without issues.
- Also removed driver-specific stuff from u_blitter.
Driver status:
- RV530 and R580 appear to just work (finally).
- RV570 should work, but there may be an issue that we don't correctly
calculate the number of dwords to clear, resulting in a partially
uninitialized zbuffer.
- RS690 misrenders as if no ZMASK clear happened. No idea what's going on.
- RV350 may even hardlock. This issue was already present and this patch doesn't
fix it.
I think we are still missing some hardware info we need to make the zbuffer
compression work fully.
Note that there is also an issue with HiZ, resulting in a sort of blocky
zigzagged corruption around some objects.
Diffstat (limited to 'src/gallium/drivers/r300/r300_chipset.h')
-rw-r--r-- | src/gallium/drivers/r300/r300_chipset.h | 13 |
1 files changed, 11 insertions, 2 deletions
diff --git a/src/gallium/drivers/r300/r300_chipset.h b/src/gallium/drivers/r300/r300_chipset.h index f2035d20092..0be161fa07a 100644 --- a/src/gallium/drivers/r300/r300_chipset.h +++ b/src/gallium/drivers/r300/r300_chipset.h @@ -33,6 +33,13 @@ #define PIPE_ZMASK_SIZE 4096 #define RV3xx_ZMASK_SIZE 5120 +/* The size of a compressed tile. Each compressed tile takes 2 bits + * in the ZMASK RAM, so there is always 16 tiles per one dword. */ +enum r300_zmask_compression { + R300_ZCOMP_4X4 = 4, + R300_ZCOMP_8X8 = 8 +}; + /* Structure containing all the possible information about a specific Radeon * in the R3xx, R4xx, and R5xx families. */ struct r300_capabilities { @@ -50,10 +57,12 @@ struct r300_capabilities { unsigned num_tex_units; /* Whether or not TCL is physically present */ boolean has_tcl; - /* Some chipsets do not have HiZ RAM - other have varying amounts . */ + /* Some chipsets do not have HiZ RAM - other have varying amounts. */ int hiz_ram; - /* some chipsets have zmask ram per pipe some don't */ + /* Some chipsets have zmask ram per pipe some don't. */ int zmask_ram; + /* Compression mode for ZMASK. */ + enum r300_zmask_compression z_compress; /* Whether or not this is RV350 or newer, including all r400 and r500 * chipsets. The differences compared to the oldest r300 chips are: * - Blend LTE/GTE thresholds |