From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

On A663 (SA8775P) the value matches exactly.

On A610, the value matches on SM6115, but is different on SM6125. That
turns out not to be a problem, as the bits that differ aren't even
interpreted.

Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
 drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
index ae0bb7934e7ed203aa3b81e28767de204f0a4d60..eaf468b67f97ff153e92a73a45581228fcf75e46 100644
--- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c
@@ -598,13 +598,10 @@ static int a6xx_calc_ubwc_config(struct adreno_gpu *gpu)
 	/* Copy the data into the internal struct to drop the const qualifier (temporarily) */
 	*cfg = *common_cfg;
 
-	cfg->ubwc_swizzle = 0x6;
 	cfg->highest_bank_bit = 15;
 
-	if (adreno_is_a610(gpu)) {
+	if (adreno_is_a610(gpu))
 		cfg->highest_bank_bit = 13;
-		cfg->ubwc_swizzle = 0x7;
-	}
 
 	if (adreno_is_a618(gpu))
 		cfg->highest_bank_bit = 14;
@@ -631,10 +628,8 @@ static int a6xx_calc_ubwc_config(struct adreno_gpu *gpu)
 		cfg->highest_bank_bit = 16;
 	}
 
-	if (adreno_is_a663(gpu)) {
+	if (adreno_is_a663(gpu))
 		cfg->highest_bank_bit = 13;
-		cfg->ubwc_swizzle = 0x4;
-	}
 
 	if (adreno_is_7c3(gpu))
 		cfg->highest_bank_bit = 14;

-- 
2.49.0

On Wed, May 14, 2025 at 05:10:32PM +0200, Konrad Dybcio wrote:
> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> On A663 (SA8775P) the value matches exactly.
> 
> On A610, the value matches on SM6115, but is different on SM6125. That
> turns out not to be a problem, as the bits that differ aren't even
> interpreted.

We also don't set swizzle for a lot of UBWC 1.0 targets (as MDSS wasn't
programming those). Should we fix all of them to use 6 by default? Or 7?

> 
> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> ---
>  drivers/gpu/drm/msm/adreno/a6xx_gpu.c | 9 ++-------
>  1 file changed, 2 insertions(+), 7 deletions(-)
> 

-- 
With best wishes
Dmitry

On 5/14/25 10:32 PM, Dmitry Baryshkov wrote:
> On Wed, May 14, 2025 at 05:10:32PM +0200, Konrad Dybcio wrote:
>> From: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
>>
>> On A663 (SA8775P) the value matches exactly.
>>
>> On A610, the value matches on SM6115, but is different on SM6125. That
>> turns out not to be a problem, as the bits that differ aren't even
>> interpreted.
> 
> We also don't set swizzle for a lot of UBWC 1.0 targets (as MDSS wasn't
> programming those). Should we fix all of them to use 6 by default? Or 7?

I don't think any default value is a good idea - this is the sort of
programming error you track down 4 years after you go bald looking
for it

Konrad