KonstantinSeurer
/
mesa
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

intel/isl: Don't align linear images to 64K on Gen12+ 

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4048>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4048>
This commit is contained in:
Jason Ekstrand 2020-03-04 11:09:50 -06:00
parent 94e37859a9
commit 0905d5a14a
1 changed files with 12 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
src/intel/isl/isl.c
View File

@ -1651,10 +1651,19 @@ isl_surf_init_s(const struct isl_device *dev,
*/
if (tiling == ISL_TILING_GEN12_CCS)
base_alignment_B = MAX(base_alignment_B, 4096);
}
if (ISL_DEV_GEN(dev) >= 12) {
base_alignment_B = MAX(base_alignment_B, 64 * 1024);
/* Gen12+ requires that images be 64K-aligned if they're going to used
* with CCS. This is because the Aux translation table maps main
* surface addresses to aux addresses at a 64K (in the main surface)
* granularity. Because we don't know for sure in ISL if a surface will
* use CCS, we have to guess based on the DISABLE_AUX usage bit. The
* one thing we do know is that we haven't enable CCS on linear images
* yet so we can avoid the extra alignment there.
*/
if (ISL_DEV_GEN(dev) >= 12 &&
!(info->usage & ISL_SURF_USAGE_DISABLE_AUX_BIT)) {
base_alignment_B = MAX(base_alignment_B, 64 * 1024);
}
}
if (ISL_DEV_GEN(dev) < 9) {