diff options
author | Kenneth Graunke <[email protected]> | 2017-04-21 01:28:13 -0700 |
---|---|---|
committer | Andres Gomez <[email protected]> | 2017-04-26 12:34:24 +0300 |
commit | de9483a6cb97a34d47a04660061afd18a9fc6f41 (patch) | |
tree | c20b66c8faf7d7b0d01942000c389320629fd33a /src | |
parent | 4c2356f13cb199dad474d38fecd424d200e4cffd (diff) |
i965/vec4: Avoid reswizzling MACH instructions in opt_register_coalesce().
opt_register_coalesce() was optimizing sequences such as:
mul(8) acc0:D, attr18.xyyy:D, attr19.xyyy:D
mach(8) vgrf5.xy:D, attr18.xyyy:D, attr19.xyyy:D
mov(8) m4.zw:F, vgrf5.xxxy:F
into:
mul(8) acc0:D, attr18.xyyy:D, attr19.xyyy:D
mach(8) m4.zw:D, attr18.xxxy:D, attr19.xxxy:D
This doesn't work - if we're going to reswizzle MACH, we'd need to
reswizzle the MUL as well. Here, the MUL fills the accumulator's .zw
components with attr18.yy * attr19.yy. But the MACH instruction expects
.z to contain attr18.x * attr19.x. Bogus results ensue.
No change in shader-db on Haswell. Prevents regressions in Timothy's
patches to use enhanced layouts for varying packing (which rearrange
code just enough to trigger this pre-existing bug, but were fine
themselves).
Acked-by: Timothy Arceri <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
(cherry picked from commit 2faf227ec2e22c7a37e0a54783a3f0a0062ac852)
Squashed with commit:
i965/vec4: Use reads_accumulator_implicitly(), not MACH checks.
Curro pointed out that I should not just check for MACH, but use
the reads_accumulator_implicitly() helper, which would also prevent
the same bug with MAC and SADA2 (if we ever decide to use them).
Cc: [email protected]
Reviewed-by: Francisco Jerez <[email protected]>
(cherry picked from commit 6b10c37b9c3a73add73f444fe1aee73c9ec82c94)
Diffstat (limited to 'src')
-rw-r--r-- | src/mesa/drivers/dri/i965/brw_vec4.cpp | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/src/mesa/drivers/dri/i965/brw_vec4.cpp b/src/mesa/drivers/dri/i965/brw_vec4.cpp index 5e60eb657a7..a7694089e93 100644 --- a/src/mesa/drivers/dri/i965/brw_vec4.cpp +++ b/src/mesa/drivers/dri/i965/brw_vec4.cpp @@ -1066,6 +1066,13 @@ vec4_instruction::can_reswizzle(const struct gen_device_info *devinfo, if (devinfo->gen == 6 && is_math() && swizzle != BRW_SWIZZLE_XYZW) return false; + /* We can't swizzle implicit accumulator access. We'd have to + * reswizzle the producer of the accumulator value in addition + * to the consumer (i.e. both MUL and MACH). Just skip this. + */ + if (reads_accumulator_implicitly()) + return false; + if (!can_do_writemask(devinfo) && dst_writemask != WRITEMASK_XYZW) return false; |