geemili
/
drm
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
drm/radeon/radeon_surface.c

2568 lines
82 KiB
C
Raw Normal View History

radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/*
* Copyright © 2011 Red Hat All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sub license, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
* OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NON-INFRINGEMENT. IN NO EVENT SHALL THE COPYRIGHT HOLDERS, AUTHORS
* AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
* USE OR OTHER DEALINGS IN THE SOFTWARE.
*
* The above copyright notice and this permission notice (including the
* next paragraph) shall be included in all copies or substantial portions
* of the Software.
*/
/*
* Authors:
* Jérôme Glisse <jglisse@redhat.com>
*/
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
#include <stdbool.h>
#include <assert.h>
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/ioctl.h>
#include "drm.h"
drm: rename libdrm{,_macros}.h Provide a more meaningful name, considering what it does. Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
2015-04-05 08:51:59 -06:00
#include "libdrm_macros.h"
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
#include "xf86drm.h"
#include "radeon_drm.h"
#include "radeon_surface.h"
radeon: sync radeon_drm.h with the kernel the CIK tile mode definitions are moved out, userspace doesn't use them Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
2016-08-19 08:07:50 -06:00
#define CIK_TILE_MODE_COLOR_2D 14
#define CIK_TILE_MODE_COLOR_2D_SCANOUT 10
#define CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_64 0
#define CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_128 1
#define CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_256 2
#define CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_512 3
#define CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_ROW_SIZE 4
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
#define ALIGN(value, alignment) (((value) + alignment - 1) & ~(alignment - 1))
#define MAX2(A, B) ((A) > (B) ? (A) : (B))
#define MIN2(A, B) ((A) < (B) ? (A) : (B))
/* keep this private */
enum radeon_family {
CHIP_UNKNOWN,
CHIP_R600,
CHIP_RV610,
CHIP_RV630,
CHIP_RV670,
CHIP_RV620,
CHIP_RV635,
CHIP_RS780,
CHIP_RS880,
CHIP_RV770,
CHIP_RV730,
CHIP_RV710,
CHIP_RV740,
CHIP_CEDAR,
CHIP_REDWOOD,
CHIP_JUNIPER,
CHIP_CYPRESS,
CHIP_HEMLOCK,
CHIP_PALM,
CHIP_SUMO,
CHIP_SUMO2,
CHIP_BARTS,
CHIP_TURKS,
CHIP_CAICOS,
CHIP_CAYMAN,
radeon: add TN surface support Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-02-14 09:32:17 -07:00
CHIP_ARUBA,
radeon: Add Southern Islands PCI IDs. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2012-05-16 10:49:18 -06:00
CHIP_TAHITI,
CHIP_PITCAIRN,
CHIP_VERDE,
radeon: add OLAND family Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2013-01-24 15:37:10 -07:00
CHIP_OLAND,
radeon: add HAINAN family Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2013-05-13 14:15:17 -06:00
CHIP_HAINAN,
radeon: add CIK chip families Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2013-06-07 11:45:30 -06:00
CHIP_BONAIRE,
CHIP_KAVERI,
CHIP_KABINI,
radeon: add hawaii chip family Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2013-09-24 09:27:11 -06:00
CHIP_HAWAII,
radeon: add Mullins chip family Signed-off-by: Samuel Li <samuel.li@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2014-04-17 12:30:08 -06:00
CHIP_MULLINS,
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
CHIP_LAST,
};
typedef int (*hw_init_surface_t)(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf);
typedef int (*hw_best_surface_t)(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf);
struct radeon_hw_info {
/* apply to r6, eg */
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
uint32_t group_bytes;
uint32_t num_banks;
uint32_t num_pipes;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* apply to eg */
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
uint32_t row_size;
unsigned allow_2d;
/* apply to si */
uint32_t tile_mode_array[32];
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
/* apply to cik */
uint32_t macrotile_mode_array[16];
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
};
struct radeon_surface_manager {
int fd;
uint32_t device_id;
struct radeon_hw_info hw_info;
unsigned family;
hw_init_surface_t surface_init;
hw_best_surface_t surface_best;
};
/* helper */
static int radeon_get_value(int fd, unsigned req, uint32_t *value)
{
struct drm_radeon_info info = {};
int r;
*value = 0;
info.request = req;
info.value = (uintptr_t)value;
r = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info,
sizeof(struct drm_radeon_info));
return r;
}
static int radeon_get_family(struct radeon_surface_manager *surf_man)
{
switch (surf_man->device_id) {
#define CHIPSET(pci_id, name, fam) case pci_id: surf_man->family = CHIP_##fam; break;
#include "r600_pci_ids.h"
#undef CHIPSET
default:
return -EINVAL;
}
return 0;
}
static unsigned next_power_of_two(unsigned x)
{
if (x <= 1)
return 1;
return (1 << ((sizeof(unsigned) * 8) - __builtin_clz(x - 1)));
}
static unsigned mip_minify(unsigned size, unsigned level)
{
unsigned val;
val = MAX2(1, size >> level);
if (level > 0)
val = next_power_of_two(val);
return val;
}
static void surf_minify(struct radeon_surface *surf,
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
struct radeon_surface_level *surflevel,
unsigned bpe, unsigned level,
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
uint32_t xalign, uint32_t yalign, uint32_t zalign,
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
uint64_t offset)
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
{
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surflevel->npix_x = mip_minify(surf->npix_x, level);
surflevel->npix_y = mip_minify(surf->npix_y, level);
surflevel->npix_z = mip_minify(surf->npix_z, level);
surflevel->nblk_x = (surflevel->npix_x + surf->blk_w - 1) / surf->blk_w;
surflevel->nblk_y = (surflevel->npix_y + surf->blk_h - 1) / surf->blk_h;
surflevel->nblk_z = (surflevel->npix_z + surf->blk_d - 1) / surf->blk_d;
radeon: add RADEON_SURF_FMASK flag which disables 2D->1D tiling transition Signed-off-by: Marek Olšák <maraeo@gmail.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-04-24 12:39:45 -06:00
if (surf->nsamples == 1 && surflevel->mode == RADEON_SURF_MODE_2D &&
!(surf->flags & RADEON_SURF_FMASK)) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
if (surflevel->nblk_x < xalign || surflevel->nblk_y < yalign) {
surflevel->mode = RADEON_SURF_MODE_1D;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
return;
}
}
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surflevel->nblk_x = ALIGN(surflevel->nblk_x, xalign);
surflevel->nblk_y = ALIGN(surflevel->nblk_y, yalign);
surflevel->nblk_z = ALIGN(surflevel->nblk_z, zalign);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surflevel->offset = offset;
surflevel->pitch_bytes = surflevel->nblk_x * bpe * surf->nsamples;
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
surflevel->slice_size = (uint64_t)surflevel->pitch_bytes * surflevel->nblk_y;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf->bo_size = offset + surflevel->slice_size * surflevel->nblk_z * surf->array_size;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
/* ===========================================================================
* r600/r700 family
*/
static int r6_init_hw_info(struct radeon_surface_manager *surf_man)
{
uint32_t tiling_config;
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
drmVersionPtr version;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
int r;
r = radeon_get_value(surf_man->fd, RADEON_INFO_TILING_CONFIG,
&tiling_config);
if (r) {
return r;
}
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
surf_man->hw_info.allow_2d = 0;
version = drmGetVersion(surf_man->fd);
if (version && version->version_minor >= 14) {
surf_man->hw_info.allow_2d = 1;
}
radeon/surface: free version after using it. fixes leak in valgrind. Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-06-17 02:18:03 -06:00
drmFreeVersion(version);
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
switch ((tiling_config & 0xe) >> 1) {
case 0:
surf_man->hw_info.num_pipes = 1;
break;
case 1:
surf_man->hw_info.num_pipes = 2;
break;
case 2:
surf_man->hw_info.num_pipes = 4;
break;
case 3:
surf_man->hw_info.num_pipes = 8;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.num_pipes = 8;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
switch ((tiling_config & 0x30) >> 4) {
case 0:
surf_man->hw_info.num_banks = 4;
break;
case 1:
surf_man->hw_info.num_banks = 8;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.num_banks = 8;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
switch ((tiling_config & 0xc0) >> 6) {
case 0:
surf_man->hw_info.group_bytes = 256;
break;
case 1:
surf_man->hw_info.group_bytes = 512;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.group_bytes = 256;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
return 0;
}
static int r6_surface_init_linear(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign;
unsigned i;
/* compute alignment */
if (!start_level) {
surf->bo_alignment = MAX2(256, surf_man->hw_info.group_bytes);
}
/* the 32 alignment is for scanout, cb or db but to allow texture to be
* easily bound as such we force this alignment to all surface
*/
radeon: fix pitch alignment for scanout buffer Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-13 18:45:53 -07:00
xalign = MAX2(1, surf_man->hw_info.group_bytes / surf->bpe);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
yalign = 1;
zalign = 1;
radeon: fix pitch alignment for scanout buffer Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-13 18:45:53 -07:00
if (surf->flags & RADEON_SURF_SCANOUT) {
xalign = MAX2((surf->bpe == 1) ? 64 : 32, xalign);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
surf->level[i].mode = RADEON_SURF_MODE_LINEAR;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf_minify(surf, surf->level+i, surf->bpe, i, xalign, yalign, zalign, offset);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
}
return 0;
}
static int r6_surface_init_linear_aligned(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign;
unsigned i;
/* compute alignment */
if (!start_level) {
surf->bo_alignment = MAX2(256, surf_man->hw_info.group_bytes);
}
xalign = MAX2(64, surf_man->hw_info.group_bytes / surf->bpe);
yalign = 1;
zalign = 1;
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
surf->level[i].mode = RADEON_SURF_MODE_LINEAR_ALIGNED;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf_minify(surf, surf->level+i, surf->bpe, i, xalign, yalign, zalign, offset);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
}
return 0;
}
static int r6_surface_init_1d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign, tilew;
unsigned i;
/* compute alignment */
tilew = 8;
xalign = surf_man->hw_info.group_bytes / (tilew * surf->bpe * surf->nsamples);
xalign = MAX2(tilew, xalign);
yalign = tilew;
zalign = 1;
radeon: fix pitch alignment for scanout buffer Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-13 18:45:53 -07:00
if (surf->flags & RADEON_SURF_SCANOUT) {
xalign = MAX2((surf->bpe == 1) ? 64 : 32, xalign);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
if (!start_level) {
surf->bo_alignment = MAX2(256, surf_man->hw_info.group_bytes);
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
surf->level[i].mode = RADEON_SURF_MODE_1D;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf_minify(surf, surf->level+i, surf->bpe, i, xalign, yalign, zalign, offset);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
}
return 0;
}
static int r6_surface_init_2d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign, tilew;
unsigned i;
/* compute alignment */
tilew = 8;
zalign = 1;
xalign = (surf_man->hw_info.group_bytes * surf_man->hw_info.num_banks) /
(tilew * surf->bpe * surf->nsamples);
xalign = MAX2(tilew * surf_man->hw_info.num_banks, xalign);
radeon: align r600/700 fmask to 128 X blocks. After much searching and empricial testing, and reading of things I've no justifcation for this fix, other than it really appears this is what the hw is doing or close enough. It makes sense that each entry in the FMASK corresponds to an entry in the CMASKm and the CMASK is organised into 128x128 blocks, but I can't find anything in any of the docs/info from AMD. But I've spent a lot of time on this, and this seems to be the simplest fix, in that we don't over allocate things too much, once this fix in place we can nuke the extra multiplier in mesa. Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2015-01-08 20:34:41 -07:00
if (surf->flags & RADEON_SURF_FMASK)
xalign = MAX2(128, xalign);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
yalign = tilew * surf_man->hw_info.num_pipes;
radeon: fix pitch alignment for scanout buffer Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-13 18:45:53 -07:00
if (surf->flags & RADEON_SURF_SCANOUT) {
xalign = MAX2((surf->bpe == 1) ? 64 : 32, xalign);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
if (!start_level) {
surf->bo_alignment =
MAX2(surf_man->hw_info.num_pipes *
surf_man->hw_info.num_banks *
radeon: align r600 msaa buffers to a multiple of macrotile size * num samples I am not sure whether this is needed, but better be safe than sorry.
2012-08-20 09:15:15 -06:00
surf->nsamples * surf->bpe * 64,
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
xalign * yalign * surf->nsamples * surf->bpe);
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
surf->level[i].mode = RADEON_SURF_MODE_2D;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf_minify(surf, surf->level+i, surf->bpe, i, xalign, yalign, zalign, offset);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
if (surf->level[i].mode == RADEON_SURF_MODE_1D) {
return r6_surface_init_1d(surf_man, surf, offset, i);
}
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
}
return 0;
}
static int r6_surface_init(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned mode;
int r;
radeon: force 2D tiling for MSAA surfaces Reviewed-by: Jerome Glisse <jglisse@redhat.com>
2012-08-07 14:50:39 -06:00
/* MSAA surfaces support the 2D mode only. */
if (surf->nsamples > 1) {
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_2D, MODE);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
radeon: simplify ZS buffer checking on r600 Setting those flags has no effect anywhere else. Reviewed-by: Christian König <christian.koenig@amd.com>
2012-07-29 06:10:07 -06:00
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER)) {
radeon: force 1D array mode for z/stencil surface On r6xx or evergreen z/stencil surface don't support linear or linear aligned surface, force 1D tiled mode for those. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-06-12 16:09:55 -06:00
/* zbuffer only support 1D or 2D tiled surface */
switch (mode) {
case RADEON_SURF_MODE_1D:
case RADEON_SURF_MODE_2D:
break;
default:
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_1D, MODE);
break;
}
}
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
/* force 1d on kernel that can't do 2d */
if (!surf_man->hw_info.allow_2d && mode > RADEON_SURF_MODE_1D) {
radeon: force 2D tiling for MSAA surfaces Reviewed-by: Jerome Glisse <jglisse@redhat.com>
2012-08-07 14:50:39 -06:00
if (surf->nsamples > 1) {
fprintf(stderr, "radeon: Cannot use 2D tiling for an MSAA surface (%i).\n", __LINE__);
return -EFAULT;
}
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(mode, MODE);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* check surface dimension */
if (surf->npix_x > 8192 || surf->npix_y > 8192 || surf->npix_z > 8192) {
return -EINVAL;
}
/* check mipmap last_level */
if (surf->last_level > 14) {
return -EINVAL;
}
/* check tiling mode */
switch (mode) {
case RADEON_SURF_MODE_LINEAR:
r = r6_surface_init_linear(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_LINEAR_ALIGNED:
r = r6_surface_init_linear_aligned(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_1D:
r = r6_surface_init_1d(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_2D:
r = r6_surface_init_2d(surf_man, surf, 0, 0);
break;
default:
return -EINVAL;
}
return r;
}
static int r6_surface_best(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
/* no value to optimize for r6xx/r7xx */
return 0;
}
/* ===========================================================================
* evergreen family
*/
static int eg_init_hw_info(struct radeon_surface_manager *surf_man)
{
uint32_t tiling_config;
drmVersionPtr version;
int r;
r = radeon_get_value(surf_man->fd, RADEON_INFO_TILING_CONFIG,
&tiling_config);
if (r) {
return r;
}
surf_man->hw_info.allow_2d = 0;
version = drmGetVersion(surf_man->fd);
radeon: enabled 2D tiling for evergreen only on fixed kernel Due to a kernel bug, enabled 2D tiling for evergreen only on newer fixed kernel. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-06-11 13:04:45 -06:00
if (version && version->version_minor >= 16) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
surf_man->hw_info.allow_2d = 1;
}
radeon/surface: free version after using it. fixes leak in valgrind. Signed-off-by: Dave Airlie <airlied@redhat.com>
2012-06-17 02:18:03 -06:00
drmFreeVersion(version);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
switch (tiling_config & 0xf) {
case 0:
surf_man->hw_info.num_pipes = 1;
break;
case 1:
surf_man->hw_info.num_pipes = 2;
break;
case 2:
surf_man->hw_info.num_pipes = 4;
break;
case 3:
surf_man->hw_info.num_pipes = 8;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.num_pipes = 8;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
switch ((tiling_config & 0xf0) >> 4) {
case 0:
surf_man->hw_info.num_banks = 4;
break;
case 1:
surf_man->hw_info.num_banks = 8;
break;
case 2:
surf_man->hw_info.num_banks = 16;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.num_banks = 8;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
switch ((tiling_config & 0xf00) >> 8) {
case 0:
surf_man->hw_info.group_bytes = 256;
break;
case 1:
surf_man->hw_info.group_bytes = 512;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.group_bytes = 256;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
switch ((tiling_config & 0xf000) >> 12) {
case 0:
surf_man->hw_info.row_size = 1024;
break;
case 1:
surf_man->hw_info.row_size = 2048;
break;
case 2:
surf_man->hw_info.row_size = 4096;
break;
default:
radeon: fall back to 1D tiling only with broken kernels Certain cards report the the wrong bank setup which causes surface init to fail in the ddx and leads to no accel. If we hit an invalid tiling parameter, just set a default value and disable 2D tiling. Should fix: https://bugs.freedesktop.org/show_bug.cgi?id=43448 Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2012-06-06 07:12:37 -06:00
surf_man->hw_info.row_size = 4096;
surf_man->hw_info.allow_2d = 0;
break;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
return 0;
}
static void eg_surf_minify(struct radeon_surface *surf,
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
struct radeon_surface_level *surflevel,
unsigned bpe,
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
unsigned level,
unsigned slice_pt,
unsigned mtilew,
unsigned mtileh,
unsigned mtileb,
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
uint64_t offset)
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
{
unsigned mtile_pr, mtile_ps;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surflevel->npix_x = mip_minify(surf->npix_x, level);
surflevel->npix_y = mip_minify(surf->npix_y, level);
surflevel->npix_z = mip_minify(surf->npix_z, level);
surflevel->nblk_x = (surflevel->npix_x + surf->blk_w - 1) / surf->blk_w;
surflevel->nblk_y = (surflevel->npix_y + surf->blk_h - 1) / surf->blk_h;
surflevel->nblk_z = (surflevel->npix_z + surf->blk_d - 1) / surf->blk_d;
radeon: add RADEON_SURF_FMASK flag which disables 2D->1D tiling transition Signed-off-by: Marek Olšák <maraeo@gmail.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-04-24 12:39:45 -06:00
if (surf->nsamples == 1 && surflevel->mode == RADEON_SURF_MODE_2D &&
!(surf->flags & RADEON_SURF_FMASK)) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
if (surflevel->nblk_x < mtilew || surflevel->nblk_y < mtileh) {
surflevel->mode = RADEON_SURF_MODE_1D;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
return;
}
}
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surflevel->nblk_x = ALIGN(surflevel->nblk_x, mtilew);
surflevel->nblk_y = ALIGN(surflevel->nblk_y, mtileh);
surflevel->nblk_z = ALIGN(surflevel->nblk_z, 1);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* macro tile per row */
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
mtile_pr = surflevel->nblk_x / mtilew;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* macro tile per slice */
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
mtile_ps = (mtile_pr * surflevel->nblk_y) / mtileh;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surflevel->offset = offset;
radeon: Always multiply pitch_bytes by nsamples, not by slice_pt slice_pt is tileb[0] / tile_split, which isn't directly related to the pitch. This caused pitch_bytes to be too large in some cases. [0] Tile size in bytes Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2014-09-26 03:19:12 -06:00
surflevel->pitch_bytes = surflevel->nblk_x * bpe * surf->nsamples;
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
surflevel->slice_size = (uint64_t)mtile_ps * mtileb * slice_pt;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf->bo_size = offset + surflevel->slice_size * surflevel->nblk_z * surf->array_size;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
static int eg_surface_init_1d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
struct radeon_surface_level *level,
unsigned bpe,
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign, tilew;
unsigned i;
/* compute alignment */
tilew = 8;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
xalign = surf_man->hw_info.group_bytes / (tilew * bpe * surf->nsamples);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
xalign = MAX2(tilew, xalign);
yalign = tilew;
zalign = 1;
radeon: fix pitch alignment for scanout buffer Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-13 18:45:53 -07:00
if (surf->flags & RADEON_SURF_SCANOUT) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
xalign = MAX2((bpe == 1) ? 64 : 32, xalign);
radeon: fix pitch alignment for scanout buffer Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-13 18:45:53 -07:00
}
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
if (!start_level) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
unsigned alignment = MAX2(256, surf_man->hw_info.group_bytes);
surf->bo_alignment = MAX2(surf->bo_alignment, alignment);
if (offset) {
offset = ALIGN(offset, alignment);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
level[i].mode = RADEON_SURF_MODE_1D;
surf_minify(surf, level+i, bpe, i, xalign, yalign, zalign, offset);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
}
return 0;
}
static int eg_surface_init_2d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
struct radeon_surface_level *level,
unsigned bpe, unsigned tile_split,
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
uint64_t offset, unsigned start_level)
{
unsigned tilew, tileh, tileb;
unsigned mtilew, mtileh, mtileb;
unsigned slice_pt;
unsigned i;
/* compute tile values */
tilew = 8;
tileh = 8;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
tileb = tilew * tileh * bpe * surf->nsamples;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* slices per tile */
slice_pt = 1;
radeon: avoid possible divide by 0 in surface manager Some users report hitting a divide by 0 with the tile split in certain apps. Tile_split shouldn't ever be 0 unless the surface structure was not properly initialized. I think there may be some cases where mesa uses an improperly initialized surface struct, but I haven't had time to track it down. Bug: https://bugs.freedesktop.org/show_bug.cgi?id=72425 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2013-12-10 10:38:26 -07:00
if (tileb > tile_split && tile_split) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
slice_pt = tileb / tile_split;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
tileb = tileb / slice_pt;
/* macro tile width & height */
mtilew = (tilew * surf->bankw * surf_man->hw_info.num_pipes) * surf->mtilea;
mtileh = (tileh * surf->bankh * surf_man->hw_info.num_banks) / surf->mtilea;
/* macro tile bytes */
mtileb = (mtilew / tilew) * (mtileh / tileh) * tileb;
if (!start_level) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
unsigned alignment = MAX2(256, mtileb);
surf->bo_alignment = MAX2(surf->bo_alignment, alignment);
if (offset) {
offset = ALIGN(offset, alignment);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
level[i].mode = RADEON_SURF_MODE_2D;
eg_surf_minify(surf, level+i, bpe, i, slice_pt, mtilew, mtileh, mtileb, offset);
if (level[i].mode == RADEON_SURF_MODE_1D) {
return eg_surface_init_1d(surf_man, surf, level, bpe, offset, i);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
}
return 0;
}
static int eg_surface_sanity(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
unsigned mode)
{
unsigned tileb;
/* check surface dimension */
if (surf->npix_x > 16384 || surf->npix_y > 16384 || surf->npix_z > 16384) {
return -EINVAL;
}
/* check mipmap last_level */
if (surf->last_level > 15) {
return -EINVAL;
}
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
/* force 1d on kernel that can't do 2d */
if (!surf_man->hw_info.allow_2d && mode > RADEON_SURF_MODE_1D) {
radeon: force 2D tiling for MSAA surfaces Reviewed-by: Jerome Glisse <jglisse@redhat.com>
2012-08-07 14:50:39 -06:00
if (surf->nsamples > 1) {
fprintf(stderr, "radeon: Cannot use 2D tiling for an MSAA surface (%i).\n", __LINE__);
return -EFAULT;
}
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(mode, MODE);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* check tile split */
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
if (mode == RADEON_SURF_MODE_2D) {
switch (surf->tile_split) {
case 64:
case 128:
case 256:
case 512:
case 1024:
case 2048:
case 4096:
break;
default:
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
return -EINVAL;
}
radeon: surface fix macro -> micro tile fallback We need to force 1D tiling only on old kernel the fallback was broken along the way. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-02 12:17:10 -07:00
switch (surf->mtilea) {
case 1:
case 2:
case 4:
case 8:
break;
default:
return -EINVAL;
}
/* check aspect ratio */
if (surf_man->hw_info.num_banks < surf->mtilea) {
return -EINVAL;
}
/* check bank width */
switch (surf->bankw) {
case 1:
case 2:
case 4:
case 8:
break;
default:
return -EINVAL;
}
/* check bank height */
switch (surf->bankh) {
case 1:
case 2:
case 4:
case 8:
break;
default:
return -EINVAL;
}
tileb = MIN2(surf->tile_split, 64 * surf->bpe * surf->nsamples);
if ((tileb * surf->bankh * surf->bankw) < surf_man->hw_info.group_bytes) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
return -EINVAL;
}
}
return 0;
}
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
static int eg_surface_init_1d_miptrees(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned zs_flags = RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER;
int r, is_depth_stencil = (surf->flags & zs_flags) == zs_flags;
drm: rename libdrm{,_macros}.h Provide a more meaningful name, considering what it does. Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
2015-04-05 08:51:59 -06:00
/* Old libdrm_macros.headers didn't have stencil_level in it. This prevents crashes. */
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
struct radeon_surface_level tmp[RADEON_SURF_MAX_LEVEL];
struct radeon_surface_level *stencil_level =
(surf->flags & RADEON_SURF_HAS_SBUFFER_MIPTREE) ? surf->stencil_level : tmp;
r = eg_surface_init_1d(surf_man, surf, surf->level, surf->bpe, 0, 0);
if (r)
return r;
if (is_depth_stencil) {
r = eg_surface_init_1d(surf_man, surf, stencil_level, 1,
surf->bo_size, 0);
surf->stencil_offset = stencil_level[0].offset;
}
return r;
}
static int eg_surface_init_2d_miptrees(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned zs_flags = RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER;
int r, is_depth_stencil = (surf->flags & zs_flags) == zs_flags;
drm: rename libdrm{,_macros}.h Provide a more meaningful name, considering what it does. Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
2015-04-05 08:51:59 -06:00
/* Old libdrm_macros.headers didn't have stencil_level in it. This prevents crashes. */
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
struct radeon_surface_level tmp[RADEON_SURF_MAX_LEVEL];
struct radeon_surface_level *stencil_level =
(surf->flags & RADEON_SURF_HAS_SBUFFER_MIPTREE) ? surf->stencil_level : tmp;
r = eg_surface_init_2d(surf_man, surf, surf->level, surf->bpe,
surf->tile_split, 0, 0);
if (r)
return r;
if (is_depth_stencil) {
r = eg_surface_init_2d(surf_man, surf, stencil_level, 1,
surf->stencil_tile_split, surf->bo_size, 0);
surf->stencil_offset = stencil_level[0].offset;
}
return r;
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
static int eg_surface_init(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned mode;
int r;
radeon: force 2D tiling for MSAA surfaces Reviewed-by: Jerome Glisse <jglisse@redhat.com>
2012-08-07 14:50:39 -06:00
/* MSAA surfaces support the 2D mode only. */
if (surf->nsamples > 1) {
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_2D, MODE);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
radeon: optimize allocation for depth w/o stencil and stencil w/o depth on EG If we don't need stencil, don't allocate it. If we need only stencil (like PIPE_FORMAT_S8_UINT), don't allocate depth. v2: actually do it correctly Reviewed-by: Christian König <christian.koenig@amd.com>
2012-07-29 07:20:15 -06:00
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER)) {
radeon: force 1D array mode for z/stencil surface On r6xx or evergreen z/stencil surface don't support linear or linear aligned surface, force 1D tiled mode for those. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-06-12 16:09:55 -06:00
/* zbuffer only support 1D or 2D tiled surface */
switch (mode) {
case RADEON_SURF_MODE_1D:
case RADEON_SURF_MODE_2D:
break;
default:
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_1D, MODE);
break;
}
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
r = eg_surface_sanity(surf_man, surf, mode);
if (r) {
return r;
}
radeon: always properly initialize stencil_offset field Reported-by: Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-06-11 13:01:12 -06:00
surf->stencil_offset = 0;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf->bo_alignment = 0;
radeon: always properly initialize stencil_offset field Reported-by: Vadim Girlin <vadimgirlin@gmail.com> Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-06-11 13:01:12 -06:00
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* check tiling mode */
switch (mode) {
case RADEON_SURF_MODE_LINEAR:
r = r6_surface_init_linear(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_LINEAR_ALIGNED:
r = r6_surface_init_linear_aligned(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_1D:
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
r = eg_surface_init_1d_miptrees(surf_man, surf);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
break;
case RADEON_SURF_MODE_2D:
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
r = eg_surface_init_2d_miptrees(surf_man, surf);
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
break;
default:
return -EINVAL;
}
return r;
}
static unsigned log2_int(unsigned x)
{
unsigned l;
if (x < 2) {
return 0;
}
for (l = 2; ; l++) {
if ((unsigned)(1 << l) > x) {
return l - 1;
}
}
return 0;
}
/* compute best tile_split, bankw, bankh, mtilea
* depending on surface
*/
static int eg_surface_best(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned mode, tileb, h_over_w;
int r;
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
/* set some default value to avoid sanity check choking on them */
surf->tile_split = 1024;
surf->bankw = 1;
surf->bankh = 1;
surf->mtilea = surf_man->hw_info.num_banks;
tileb = MIN2(surf->tile_split, 64 * surf->bpe * surf->nsamples);
for (; surf->bankh <= 8; surf->bankh *= 2) {
if ((tileb * surf->bankh * surf->bankw) >= surf_man->hw_info.group_bytes) {
break;
}
}
if (surf->mtilea > 8) {
surf->mtilea = 8;
}
r = eg_surface_sanity(surf_man, surf, mode);
if (r) {
return r;
}
if (mode != RADEON_SURF_MODE_2D) {
/* nothing to do for non 2D tiled surface */
return 0;
}
radeon: tweak TILE_SPLIT for MSAA surfaces Reviewed-by: Jerome Glisse <jglisse@redhat.com>
2012-08-07 15:38:19 -06:00
/* Tweak TILE_SPLIT for performance here. */
if (surf->nsamples > 1) {
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER)) {
switch (surf->nsamples) {
case 2:
surf->tile_split = 128;
break;
case 4:
surf->tile_split = 128;
break;
case 8:
surf->tile_split = 256;
break;
case 16: /* cayman only */
surf->tile_split = 512;
break;
default:
fprintf(stderr, "radeon: Wrong number of samples %i (%i)\n",
surf->nsamples, __LINE__);
return -EINVAL;
}
surf->stencil_tile_split = 64;
} else {
radeon: use SAMPLE_SPLIT=2 for better MSAA perf on EG/CM Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2016-05-26 06:33:12 -06:00
/* tile split must be >= 256 for colorbuffer surfaces,
* SAMPLE_SPLIT = tile_split / (bpe * 64), the optimal value is 2
*/
surf->tile_split = MAX2(2 * surf->bpe * 64, 256);
radeon: fix tile_split of 128-bit surface formats with 8x MSAA The calculation led to the number 8192, which is too high. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-10-15 18:08:02 -06:00
if (surf->tile_split > 4096)
surf->tile_split = 4096;
radeon: tweak TILE_SPLIT for MSAA surfaces Reviewed-by: Jerome Glisse <jglisse@redhat.com>
2012-08-07 15:38:19 -06:00
}
} else {
/* set tile split to row size */
surf->tile_split = surf_man->hw_info.row_size;
surf->stencil_tile_split = surf_man->hw_info.row_size / 2;
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* bankw or bankh greater than 1 increase alignment requirement, not
* sure if it's worth using smaller bankw & bankh to stick with 2D
* tiling on small surface rather than falling back to 1D tiling.
libdrm: random typo fixes Just some trivial boring typo fixes all over the tree. READMEs and comments only. Signed-off-by: Grazvydas Ignotas <notasas@gmail.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
2016-11-20 11:25:46 -07:00
* Use recommended value based on tile size for now.
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
*
* fmask buffer has different optimal value figure them out once we
* use it.
*/
radeon: don't take the stencil-specific codepath for buffers without stencil Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-29 07:10:33 -06:00
if (surf->flags & RADEON_SURF_SBUFFER) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* assume 1 bytes for stencil, we optimize for stencil as stencil
* and depth shares surface values
*/
tileb = MIN2(surf->tile_split, 64 * surf->nsamples);
} else {
tileb = MIN2(surf->tile_split, 64 * surf->bpe * surf->nsamples);
}
/* use bankw of 1 to minimize width alignment, might be interesting to
* increase it for large surface
*/
surf->bankw = 1;
switch (tileb) {
case 64:
surf->bankh = 4;
break;
case 128:
case 256:
surf->bankh = 2;
break;
default:
surf->bankh = 1;
break;
}
/* double check the constraint */
for (; surf->bankh <= 8; surf->bankh *= 2) {
if ((tileb * surf->bankh * surf->bankw) >= surf_man->hw_info.group_bytes) {
break;
}
}
h_over_w = (((surf->bankh * surf_man->hw_info.num_banks) << 16) /
(surf->bankw * surf_man->hw_info.num_pipes)) >> 16;
surf->mtilea = 1 << (log2_int(h_over_w) >> 1);
return 0;
}
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
/* ===========================================================================
* Southern Islands family
*/
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
#define SI__GB_TILE_MODE__PIPE_CONFIG(x) (((x) >> 6) & 0x1f)
#define SI__PIPE_CONFIG__ADDR_SURF_P2 0
#define SI__PIPE_CONFIG__ADDR_SURF_P4_8x16 4
#define SI__PIPE_CONFIG__ADDR_SURF_P4_16x16 5
#define SI__PIPE_CONFIG__ADDR_SURF_P4_16x32 6
#define SI__PIPE_CONFIG__ADDR_SURF_P4_32x32 7
#define SI__PIPE_CONFIG__ADDR_SURF_P8_16x16_8x16 8
#define SI__PIPE_CONFIG__ADDR_SURF_P8_16x32_8x16 9
#define SI__PIPE_CONFIG__ADDR_SURF_P8_32x32_8x16 10
#define SI__PIPE_CONFIG__ADDR_SURF_P8_16x32_16x16 11
#define SI__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x16 12
#define SI__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x32 13
#define SI__PIPE_CONFIG__ADDR_SURF_P8_32x64_32x32 14
#define SI__GB_TILE_MODE__TILE_SPLIT(x) (((x) >> 11) & 0x7)
#define SI__TILE_SPLIT__64B 0
#define SI__TILE_SPLIT__128B 1
#define SI__TILE_SPLIT__256B 2
#define SI__TILE_SPLIT__512B 3
#define SI__TILE_SPLIT__1024B 4
#define SI__TILE_SPLIT__2048B 5
#define SI__TILE_SPLIT__4096B 6
#define SI__GB_TILE_MODE__BANK_WIDTH(x) (((x) >> 14) & 0x3)
#define SI__BANK_WIDTH__1 0
#define SI__BANK_WIDTH__2 1
#define SI__BANK_WIDTH__4 2
#define SI__BANK_WIDTH__8 3
#define SI__GB_TILE_MODE__BANK_HEIGHT(x) (((x) >> 16) & 0x3)
#define SI__BANK_HEIGHT__1 0
#define SI__BANK_HEIGHT__2 1
#define SI__BANK_HEIGHT__4 2
#define SI__BANK_HEIGHT__8 3
#define SI__GB_TILE_MODE__MACRO_TILE_ASPECT(x) (((x) >> 18) & 0x3)
#define SI__MACRO_TILE_ASPECT__1 0
#define SI__MACRO_TILE_ASPECT__2 1
#define SI__MACRO_TILE_ASPECT__4 2
#define SI__MACRO_TILE_ASPECT__8 3
#define SI__GB_TILE_MODE__NUM_BANKS(x) (((x) >> 20) & 0x3)
#define SI__NUM_BANKS__2_BANK 0
#define SI__NUM_BANKS__4_BANK 1
#define SI__NUM_BANKS__8_BANK 2
#define SI__NUM_BANKS__16_BANK 3
static void si_gb_tile_mode(uint32_t gb_tile_mode,
unsigned *num_pipes,
unsigned *num_banks,
uint32_t *macro_tile_aspect,
uint32_t *bank_w,
uint32_t *bank_h,
uint32_t *tile_split)
{
if (num_pipes) {
switch (SI__GB_TILE_MODE__PIPE_CONFIG(gb_tile_mode)) {
case SI__PIPE_CONFIG__ADDR_SURF_P2:
default:
*num_pipes = 2;
break;
case SI__PIPE_CONFIG__ADDR_SURF_P4_8x16:
case SI__PIPE_CONFIG__ADDR_SURF_P4_16x16:
case SI__PIPE_CONFIG__ADDR_SURF_P4_16x32:
case SI__PIPE_CONFIG__ADDR_SURF_P4_32x32:
*num_pipes = 4;
break;
case SI__PIPE_CONFIG__ADDR_SURF_P8_16x16_8x16:
case SI__PIPE_CONFIG__ADDR_SURF_P8_16x32_8x16:
case SI__PIPE_CONFIG__ADDR_SURF_P8_32x32_8x16:
case SI__PIPE_CONFIG__ADDR_SURF_P8_16x32_16x16:
case SI__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x16:
case SI__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x32:
case SI__PIPE_CONFIG__ADDR_SURF_P8_32x64_32x32:
*num_pipes = 8;
break;
}
}
if (num_banks) {
switch (SI__GB_TILE_MODE__NUM_BANKS(gb_tile_mode)) {
default:
case SI__NUM_BANKS__2_BANK:
*num_banks = 2;
break;
case SI__NUM_BANKS__4_BANK:
*num_banks = 4;
break;
case SI__NUM_BANKS__8_BANK:
*num_banks = 8;
break;
case SI__NUM_BANKS__16_BANK:
*num_banks = 16;
break;
}
}
if (macro_tile_aspect) {
switch (SI__GB_TILE_MODE__MACRO_TILE_ASPECT(gb_tile_mode)) {
default:
case SI__MACRO_TILE_ASPECT__1:
*macro_tile_aspect = 1;
break;
case SI__MACRO_TILE_ASPECT__2:
*macro_tile_aspect = 2;
break;
case SI__MACRO_TILE_ASPECT__4:
*macro_tile_aspect = 4;
break;
case SI__MACRO_TILE_ASPECT__8:
*macro_tile_aspect = 8;
break;
}
}
if (bank_w) {
switch (SI__GB_TILE_MODE__BANK_WIDTH(gb_tile_mode)) {
default:
case SI__BANK_WIDTH__1:
*bank_w = 1;
break;
case SI__BANK_WIDTH__2:
*bank_w = 2;
break;
case SI__BANK_WIDTH__4:
*bank_w = 4;
break;
case SI__BANK_WIDTH__8:
*bank_w = 8;
break;
}
}
if (bank_h) {
switch (SI__GB_TILE_MODE__BANK_HEIGHT(gb_tile_mode)) {
default:
case SI__BANK_HEIGHT__1:
*bank_h = 1;
break;
case SI__BANK_HEIGHT__2:
*bank_h = 2;
break;
case SI__BANK_HEIGHT__4:
*bank_h = 4;
break;
case SI__BANK_HEIGHT__8:
*bank_h = 8;
break;
}
}
if (tile_split) {
switch (SI__GB_TILE_MODE__TILE_SPLIT(gb_tile_mode)) {
default:
case SI__TILE_SPLIT__64B:
*tile_split = 64;
break;
case SI__TILE_SPLIT__128B:
*tile_split = 128;
break;
case SI__TILE_SPLIT__256B:
*tile_split = 256;
break;
case SI__TILE_SPLIT__512B:
*tile_split = 512;
break;
case SI__TILE_SPLIT__1024B:
*tile_split = 1024;
break;
case SI__TILE_SPLIT__2048B:
*tile_split = 2048;
break;
case SI__TILE_SPLIT__4096B:
*tile_split = 4096;
break;
}
}
}
static int si_init_hw_info(struct radeon_surface_manager *surf_man)
{
uint32_t tiling_config;
drmVersionPtr version;
int r;
r = radeon_get_value(surf_man->fd, RADEON_INFO_TILING_CONFIG,
&tiling_config);
if (r) {
return r;
}
surf_man->hw_info.allow_2d = 0;
version = drmGetVersion(surf_man->fd);
if (version && version->version_minor >= 33) {
if (!radeon_get_value(surf_man->fd, RADEON_INFO_SI_TILE_MODE_ARRAY, surf_man->hw_info.tile_mode_array)) {
surf_man->hw_info.allow_2d = 1;
}
}
drmFreeVersion(version);
switch (tiling_config & 0xf) {
case 0:
surf_man->hw_info.num_pipes = 1;
break;
case 1:
surf_man->hw_info.num_pipes = 2;
break;
case 2:
surf_man->hw_info.num_pipes = 4;
break;
case 3:
surf_man->hw_info.num_pipes = 8;
break;
default:
surf_man->hw_info.num_pipes = 8;
surf_man->hw_info.allow_2d = 0;
break;
}
switch ((tiling_config & 0xf0) >> 4) {
case 0:
surf_man->hw_info.num_banks = 4;
break;
case 1:
surf_man->hw_info.num_banks = 8;
break;
case 2:
surf_man->hw_info.num_banks = 16;
break;
default:
surf_man->hw_info.num_banks = 8;
surf_man->hw_info.allow_2d = 0;
break;
}
switch ((tiling_config & 0xf00) >> 8) {
case 0:
surf_man->hw_info.group_bytes = 256;
break;
case 1:
surf_man->hw_info.group_bytes = 512;
break;
default:
surf_man->hw_info.group_bytes = 256;
surf_man->hw_info.allow_2d = 0;
break;
}
switch ((tiling_config & 0xf000) >> 12) {
case 0:
surf_man->hw_info.row_size = 1024;
break;
case 1:
surf_man->hw_info.row_size = 2048;
break;
case 2:
surf_man->hw_info.row_size = 4096;
break;
default:
surf_man->hw_info.row_size = 4096;
surf_man->hw_info.allow_2d = 0;
break;
}
return 0;
}
static int si_surface_sanity(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
unsigned mode, unsigned *tile_mode, unsigned *stencil_tile_mode)
{
uint32_t gb_tile_mode;
/* check surface dimension */
if (surf->npix_x > 16384 || surf->npix_y > 16384 || surf->npix_z > 16384) {
return -EINVAL;
}
/* check mipmap last_level */
if (surf->last_level > 15) {
return -EINVAL;
}
/* force 1d on kernel that can't do 2d */
if (mode > RADEON_SURF_MODE_1D &&
(!surf_man->hw_info.allow_2d || !(surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX))) {
if (surf->nsamples > 1) {
fprintf(stderr, "radeon: Cannot use 1D tiling for an MSAA surface (%i).\n", __LINE__);
return -EFAULT;
}
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(mode, MODE);
}
if (surf->nsamples > 1 && mode != RADEON_SURF_MODE_2D) {
return -EINVAL;
}
if (!surf->tile_split) {
/* default value */
surf->mtilea = 1;
surf->bankw = 1;
radeon: Fix surf->bankh init by default value when surf->tile_split == 0 Signed-off-by: Maks Naumov <maksqwe1@ukr.net> Signed-off-by: Marek Olšák <marek.olsak@amd.com>
2014-08-18 13:59:35 -06:00
surf->bankh = 1;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
surf->tile_split = 64;
surf->stencil_tile_split = 64;
}
switch (mode) {
case RADEON_SURF_MODE_2D:
if (surf->flags & RADEON_SURF_SBUFFER) {
switch (surf->nsamples) {
case 1:
*stencil_tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D;
break;
case 2:
*stencil_tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D_2AA;
break;
case 4:
*stencil_tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D_4AA;
break;
case 8:
*stencil_tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D_8AA;
break;
default:
return -EINVAL;
}
/* retrieve tiling mode value */
gb_tile_mode = surf_man->hw_info.tile_mode_array[*stencil_tile_mode];
si_gb_tile_mode(gb_tile_mode, NULL, NULL, NULL, NULL, NULL, &surf->stencil_tile_split);
}
if (surf->flags & RADEON_SURF_ZBUFFER) {
switch (surf->nsamples) {
case 1:
*tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D;
break;
case 2:
*tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D_2AA;
break;
case 4:
*tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D_4AA;
break;
case 8:
*tile_mode = SI_TILE_MODE_DEPTH_STENCIL_2D_8AA;
break;
default:
return -EINVAL;
}
} else if (surf->flags & RADEON_SURF_SCANOUT) {
switch (surf->bpe) {
case 2:
*tile_mode = SI_TILE_MODE_COLOR_2D_SCANOUT_16BPP;
break;
case 4:
*tile_mode = SI_TILE_MODE_COLOR_2D_SCANOUT_32BPP;
break;
default:
return -EINVAL;
}
} else {
switch (surf->bpe) {
case 1:
*tile_mode = SI_TILE_MODE_COLOR_2D_8BPP;
break;
case 2:
*tile_mode = SI_TILE_MODE_COLOR_2D_16BPP;
break;
case 4:
*tile_mode = SI_TILE_MODE_COLOR_2D_32BPP;
break;
case 8:
case 16:
*tile_mode = SI_TILE_MODE_COLOR_2D_64BPP;
break;
default:
return -EINVAL;
}
}
/* retrieve tiling mode value */
gb_tile_mode = surf_man->hw_info.tile_mode_array[*tile_mode];
si_gb_tile_mode(gb_tile_mode, NULL, NULL, &surf->mtilea, &surf->bankw, &surf->bankh, &surf->tile_split);
break;
case RADEON_SURF_MODE_1D:
if (surf->flags & RADEON_SURF_SBUFFER) {
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
*stencil_tile_mode = SI_TILE_MODE_DEPTH_STENCIL_1D;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
}
if (surf->flags & RADEON_SURF_ZBUFFER) {
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
*tile_mode = SI_TILE_MODE_DEPTH_STENCIL_1D;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
} else if (surf->flags & RADEON_SURF_SCANOUT) {
*tile_mode = SI_TILE_MODE_COLOR_1D_SCANOUT;
} else {
*tile_mode = SI_TILE_MODE_COLOR_1D;
}
break;
case RADEON_SURF_MODE_LINEAR_ALIGNED:
default:
*tile_mode = SI_TILE_MODE_COLOR_LINEAR_ALIGNED;
}
return 0;
}
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
static void si_surf_minify(struct radeon_surface *surf,
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
struct radeon_surface_level *surflevel,
unsigned bpe, unsigned level,
uint32_t xalign, uint32_t yalign, uint32_t zalign,
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
uint32_t slice_align, uint64_t offset)
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
{
radeon: fix pitch alignment for non-power-of-two mipmaps on SI This fixes VM protection faults. I have a new piglit test which can iterate over all possible widths, heights, and depths (including NPOT) and tests mipmapping with various texture targets. After this is committed, I'll make a new release of libdrm and bump the libdrm version requirement in Mesa.
2013-09-19 06:19:02 -06:00
if (level == 0) {
surflevel->npix_x = surf->npix_x;
} else {
surflevel->npix_x = mip_minify(next_power_of_two(surf->npix_x), level);
}
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surflevel->npix_y = mip_minify(surf->npix_y, level);
surflevel->npix_z = mip_minify(surf->npix_z, level);
radeon: Memory footprint of SI mipmap base level is padded to powers of two. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-04 10:53:55 -06:00
if (level == 0 && surf->last_level > 0) {
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surflevel->nblk_x = (next_power_of_two(surflevel->npix_x) + surf->blk_w - 1) / surf->blk_w;
surflevel->nblk_y = (next_power_of_two(surflevel->npix_y) + surf->blk_h - 1) / surf->blk_h;
surflevel->nblk_z = (next_power_of_two(surflevel->npix_z) + surf->blk_d - 1) / surf->blk_d;
radeon: Memory footprint of SI mipmap base level is padded to powers of two. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-04 10:53:55 -06:00
} else {
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surflevel->nblk_x = (surflevel->npix_x + surf->blk_w - 1) / surf->blk_w;
surflevel->nblk_y = (surflevel->npix_y + surf->blk_h - 1) / surf->blk_h;
surflevel->nblk_z = (surflevel->npix_z + surf->blk_d - 1) / surf->blk_d;
radeon: Memory footprint of SI mipmap base level is padded to powers of two. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-04 10:53:55 -06:00
}
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surflevel->nblk_y = ALIGN(surflevel->nblk_y, yalign);
radeon: Sampling pitch for non-mipmaps seems padded to slice alignment on SI. Another corner case that isn't well-explained yet. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-05 10:44:45 -06:00
/* XXX: Texture sampling uses unexpectedly large pitches in some cases,
* these are just guesses for the rules behind those
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
*/
radeon: Sampling pitch for non-mipmaps seems padded to slice alignment on SI. Another corner case that isn't well-explained yet. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-05 10:44:45 -06:00
if (level == 0 && surf->last_level == 0)
/* Non-mipmap pitch padded to slice alignment */
radeon: don't overallocate stencil by 4 on SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>
2013-11-18 03:40:08 -07:00
/* Using just bpe here breaks stencil blitting; surf->bpe works. */
radeon: Sampling pitch for non-mipmaps seems padded to slice alignment on SI. Another corner case that isn't well-explained yet. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-05 10:44:45 -06:00
xalign = MAX2(xalign, slice_align / surf->bpe);
radeon: fix pitch alignment for non-power-of-two mipmaps on SI This fixes VM protection faults. I have a new piglit test which can iterate over all possible widths, heights, and depths (including NPOT) and tests mipmapping with various texture targets. After this is committed, I'll make a new release of libdrm and bump the libdrm version requirement in Mesa.
2013-09-19 06:19:02 -06:00
else if (surflevel->mode == RADEON_SURF_MODE_LINEAR_ALIGNED)
radeon: Sampling pitch for non-mipmaps seems padded to slice alignment on SI. Another corner case that isn't well-explained yet. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-09-05 10:44:45 -06:00
/* Small rows evenly distributed across slice */
radeon: don't overallocate stencil by 4 on SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>
2013-11-18 03:40:08 -07:00
xalign = MAX2(xalign, slice_align / bpe / surflevel->nblk_y);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surflevel->nblk_x = ALIGN(surflevel->nblk_x, xalign);
surflevel->nblk_z = ALIGN(surflevel->nblk_z, zalign);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surflevel->offset = offset;
radeon: don't overallocate stencil by 4 on SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com>
2013-11-18 03:40:08 -07:00
surflevel->pitch_bytes = surflevel->nblk_x * bpe * surf->nsamples;
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
surflevel->slice_size = ALIGN((uint64_t)surflevel->pitch_bytes * surflevel->nblk_y,
(uint64_t)slice_align);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
surf->bo_size = offset + surflevel->slice_size * surflevel->nblk_z * surf->array_size;
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
static void si_surf_minify_2d(struct radeon_surface *surf,
struct radeon_surface_level *surflevel,
unsigned bpe, unsigned level, unsigned slice_pt,
uint32_t xalign, uint32_t yalign, uint32_t zalign,
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
unsigned mtileb, uint64_t offset)
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
{
unsigned mtile_pr, mtile_ps;
radeon: fix pitch alignment for non-power-of-two mipmaps on SI This fixes VM protection faults. I have a new piglit test which can iterate over all possible widths, heights, and depths (including NPOT) and tests mipmapping with various texture targets. After this is committed, I'll make a new release of libdrm and bump the libdrm version requirement in Mesa.
2013-09-19 06:19:02 -06:00
if (level == 0) {
surflevel->npix_x = surf->npix_x;
} else {
surflevel->npix_x = mip_minify(next_power_of_two(surf->npix_x), level);
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
surflevel->npix_y = mip_minify(surf->npix_y, level);
surflevel->npix_z = mip_minify(surf->npix_z, level);
if (level == 0 && surf->last_level > 0) {
surflevel->nblk_x = (next_power_of_two(surflevel->npix_x) + surf->blk_w - 1) / surf->blk_w;
surflevel->nblk_y = (next_power_of_two(surflevel->npix_y) + surf->blk_h - 1) / surf->blk_h;
surflevel->nblk_z = (next_power_of_two(surflevel->npix_z) + surf->blk_d - 1) / surf->blk_d;
} else {
surflevel->nblk_x = (surflevel->npix_x + surf->blk_w - 1) / surf->blk_w;
surflevel->nblk_y = (surflevel->npix_y + surf->blk_h - 1) / surf->blk_h;
surflevel->nblk_z = (surflevel->npix_z + surf->blk_d - 1) / surf->blk_d;
}
radeon: add RADEON_SURF_FMASK flag which disables 2D->1D tiling transition Signed-off-by: Marek Olšák <maraeo@gmail.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-04-24 12:39:45 -06:00
if (surf->nsamples == 1 && surflevel->mode == RADEON_SURF_MODE_2D &&
!(surf->flags & RADEON_SURF_FMASK)) {
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
if (surflevel->nblk_x < xalign || surflevel->nblk_y < yalign) {
surflevel->mode = RADEON_SURF_MODE_1D;
return;
}
}
surflevel->nblk_x = ALIGN(surflevel->nblk_x, xalign);
surflevel->nblk_y = ALIGN(surflevel->nblk_y, yalign);
surflevel->nblk_z = ALIGN(surflevel->nblk_z, zalign);
/* macro tile per row */
mtile_pr = surflevel->nblk_x / xalign;
/* macro tile per slice */
mtile_ps = (mtile_pr * surflevel->nblk_y) / yalign;
surflevel->offset = offset;
radeon: Always multiply pitch_bytes by nsamples, not by slice_pt slice_pt is tileb[0] / tile_split, which isn't directly related to the pitch. This caused pitch_bytes to be too large in some cases. [0] Tile size in bytes Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2014-09-26 03:19:12 -06:00
surflevel->pitch_bytes = surflevel->nblk_x * bpe * surf->nsamples;
radeon: Handle surface offsets exceeding 32 bits correctly The slice_size and bo_size fields were getting truncated to 32 bits. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-09-08 00:03:55 -06:00
surflevel->slice_size = (uint64_t)mtile_ps * mtileb * slice_pt;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
surf->bo_size = offset + surflevel->slice_size * surflevel->nblk_z * surf->array_size;
}
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
static int si_surface_init_linear_aligned(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
unsigned tile_mode,
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign, slice_align;
unsigned i;
/* compute alignment */
if (!start_level) {
surf->bo_alignment = MAX2(256, surf_man->hw_info.group_bytes);
}
xalign = MAX2(8, 64 / surf->bpe);
yalign = 1;
zalign = 1;
slice_align = MAX2(64 * surf->bpe, surf_man->hw_info.group_bytes);
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
surf->level[i].mode = RADEON_SURF_MODE_LINEAR_ALIGNED;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
si_surf_minify(surf, surf->level+i, surf->bpe, i, xalign, yalign, zalign, slice_align, offset);
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
offset = ALIGN(offset, surf->bo_alignment);
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
if (surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX) {
surf->tiling_index[i] = tile_mode;
}
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
}
return 0;
}
static int si_surface_init_1d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
struct radeon_surface_level *level,
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
unsigned bpe, unsigned tile_mode,
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
uint64_t offset, unsigned start_level)
{
uint32_t xalign, yalign, zalign, slice_align;
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
unsigned alignment = MAX2(256, surf_man->hw_info.group_bytes);
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
unsigned i;
/* compute alignment */
xalign = 8;
yalign = 8;
zalign = 1;
slice_align = surf_man->hw_info.group_bytes;
if (surf->flags & RADEON_SURF_SCANOUT) {
xalign = MAX2((bpe == 1) ? 64 : 32, xalign);
}
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
if (start_level <= 1) {
surf->bo_alignment = MAX2(surf->bo_alignment, alignment);
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
if (offset) {
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
offset = ALIGN(offset, alignment);
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
}
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
level[i].mode = RADEON_SURF_MODE_1D;
si_surf_minify(surf, level+i, bpe, i, xalign, yalign, zalign, slice_align, offset);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
/* level0 and first mipmap need to have alignment */
offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
offset = ALIGN(offset, alignment);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
if (surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX) {
if (surf->level == level) {
surf->tiling_index[i] = tile_mode;
/* it's ok because stencil is done after */
surf->stencil_tiling_index[i] = tile_mode;
} else {
surf->stencil_tiling_index[i] = tile_mode;
}
}
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
}
return 0;
}
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
static int si_surface_init_1d_miptrees(struct radeon_surface_manager *surf_man,
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
struct radeon_surface *surf,
unsigned tile_mode, unsigned stencil_tile_mode)
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
{
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
int r;
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
r = si_surface_init_1d(surf_man, surf, surf->level, surf->bpe, tile_mode, 0, 0);
if (r) {
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
return r;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
}
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
if (surf->flags & RADEON_SURF_SBUFFER) {
r = si_surface_init_1d(surf_man, surf, surf->stencil_level, 1, stencil_tile_mode, surf->bo_size, 0);
surf->stencil_offset = surf->stencil_level[0].offset;
}
return r;
}
static int si_surface_init_2d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
struct radeon_surface_level *level,
unsigned bpe, unsigned tile_mode,
unsigned num_pipes, unsigned num_banks,
unsigned tile_split,
uint64_t offset,
unsigned start_level)
{
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
uint64_t aligned_offset = offset;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
unsigned tilew, tileh, tileb;
unsigned mtilew, mtileh, mtileb;
unsigned slice_pt;
unsigned i;
/* compute tile values */
tilew = 8;
tileh = 8;
tileb = tilew * tileh * bpe * surf->nsamples;
/* slices per tile */
slice_pt = 1;
radeon: avoid possible divide by 0 in surface manager Some users report hitting a divide by 0 with the tile split in certain apps. Tile_split shouldn't ever be 0 unless the surface structure was not properly initialized. I think there may be some cases where mesa uses an improperly initialized surface struct, but I haven't had time to track it down. Bug: https://bugs.freedesktop.org/show_bug.cgi?id=72425 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2013-12-10 10:38:26 -07:00
if (tileb > tile_split && tile_split) {
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
slice_pt = tileb / tile_split;
}
tileb = tileb / slice_pt;
/* macro tile width & height */
mtilew = (tilew * surf->bankw * num_pipes) * surf->mtilea;
mtileh = (tileh * surf->bankh * num_banks) / surf->mtilea;
/* macro tile bytes */
mtileb = (mtilew / tilew) * (mtileh / tileh) * tileb;
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
if (start_level <= 1) {
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
unsigned alignment = MAX2(256, mtileb);
surf->bo_alignment = MAX2(surf->bo_alignment, alignment);
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
if (aligned_offset) {
aligned_offset = ALIGN(aligned_offset, alignment);
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
}
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
level[i].mode = RADEON_SURF_MODE_2D;
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
si_surf_minify_2d(surf, level+i, bpe, i, slice_pt, mtilew, mtileh, 1, mtileb, aligned_offset);
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
if (level[i].mode == RADEON_SURF_MODE_1D) {
switch (tile_mode) {
case SI_TILE_MODE_COLOR_2D_8BPP:
case SI_TILE_MODE_COLOR_2D_16BPP:
case SI_TILE_MODE_COLOR_2D_32BPP:
case SI_TILE_MODE_COLOR_2D_64BPP:
tile_mode = SI_TILE_MODE_COLOR_1D;
break;
case SI_TILE_MODE_COLOR_2D_SCANOUT_16BPP:
case SI_TILE_MODE_COLOR_2D_SCANOUT_32BPP:
tile_mode = SI_TILE_MODE_COLOR_1D_SCANOUT;
break;
case SI_TILE_MODE_DEPTH_STENCIL_2D:
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
tile_mode = SI_TILE_MODE_DEPTH_STENCIL_1D;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
break;
default:
return -EINVAL;
}
return si_surface_init_1d(surf_man, surf, level, bpe, tile_mode, offset, i);
}
/* level0 and first mipmap need to have alignment */
radeon: Update unaligned offset for 2D->1D tiling transition on SI Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71983 Tested-by: Arek Ruśniak <arek.rusi@gmail.com>
2013-11-26 02:16:03 -07:00
aligned_offset = offset = surf->bo_size;
radeon: Remove superfluous parentheses. Signed-off-by: Thomas Klausner <wiz@NetBSD.org>
2014-07-15 11:22:52 -06:00
if (i == 0) {
radeon: fix mipmap level 0 and 1 alignment for SI and CIK Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-14 08:40:30 -07:00
aligned_offset = ALIGN(aligned_offset, surf->bo_alignment);
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
}
if (surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX) {
if (surf->level == level) {
surf->tiling_index[i] = tile_mode;
/* it's ok because stencil is done after */
surf->stencil_tiling_index[i] = tile_mode;
} else {
surf->stencil_tiling_index[i] = tile_mode;
}
}
}
return 0;
}
static int si_surface_init_2d_miptrees(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
unsigned tile_mode, unsigned stencil_tile_mode)
{
unsigned num_pipes, num_banks;
uint32_t gb_tile_mode;
int r;
/* retrieve tiling mode value */
gb_tile_mode = surf_man->hw_info.tile_mode_array[tile_mode];
si_gb_tile_mode(gb_tile_mode, &num_pipes, &num_banks, NULL, NULL, NULL, NULL);
r = si_surface_init_2d(surf_man, surf, surf->level, surf->bpe, tile_mode, num_pipes, num_banks, surf->tile_split, 0, 0);
if (r) {
return r;
}
if (surf->flags & RADEON_SURF_SBUFFER) {
r = si_surface_init_2d(surf_man, surf, surf->stencil_level, 1, stencil_tile_mode, num_pipes, num_banks, surf->stencil_tile_split, surf->bo_size, 0);
surf->stencil_offset = surf->stencil_level[0].offset;
radeon: Fix 1D tiling layout on SI. Very similar to Evergreen, but slightly different rules for tile / slice alignment. Fortunately, these map quite naturally onto the previous fixes for linear aligned layout on SI. 2D tiling still needs more work here and possibly in the kernel. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
2013-01-17 11:03:42 -07:00
}
return r;
}
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
static int si_surface_init(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
unsigned mode, tile_mode, stencil_tile_mode;
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
int r;
/* MSAA surfaces support the 2D mode only. */
if (surf->nsamples > 1) {
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_2D, MODE);
}
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER)) {
/* zbuffer only support 1D or 2D tiled surface */
switch (mode) {
case RADEON_SURF_MODE_1D:
case RADEON_SURF_MODE_2D:
break;
default:
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_1D, MODE);
break;
}
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
r = si_surface_sanity(surf_man, surf, mode, &tile_mode, &stencil_tile_mode);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
if (r) {
return r;
}
surf->stencil_offset = 0;
radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI This allows texturing with depth-stencil buffers directly without the copy to CB. The separate miptree description for stencil is added, because the stencil mipmap offsets are not really depth offsets/4 (at least for the texture units). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2012-09-30 11:20:04 -06:00
surf->bo_alignment = 0;
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
/* check tiling mode */
switch (mode) {
case RADEON_SURF_MODE_LINEAR:
r = r6_surface_init_linear(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_LINEAR_ALIGNED:
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
r = si_surface_init_linear_aligned(surf_man, surf, tile_mode, 0, 0);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
break;
case RADEON_SURF_MODE_1D:
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
r = si_surface_init_1d_miptrees(surf_man, surf, tile_mode, stencil_tile_mode);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
break;
case RADEON_SURF_MODE_2D:
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
r = si_surface_init_2d_miptrees(surf_man, surf, tile_mode, stencil_tile_mode);
radeon: Fix layout of linear aligned mipmaps on SI. Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
2012-08-31 11:29:33 -06:00
break;
default:
return -EINVAL;
}
return r;
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
/*
* depending on surface
*/
static int si_surface_best(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned mode, tile_mode, stencil_tile_mode;
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER) &&
!(surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX)) {
/* depth/stencil force 1d tiling for old mesa */
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_1D, MODE);
}
return si_surface_sanity(surf_man, surf, mode, &tile_mode, &stencil_tile_mode);
}
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
/* ===========================================================================
* Sea Islands family
*/
#define CIK__GB_TILE_MODE__PIPE_CONFIG(x) (((x) >> 6) & 0x1f)
#define CIK__PIPE_CONFIG__ADDR_SURF_P2 0
#define CIK__PIPE_CONFIG__ADDR_SURF_P4_8x16 4
#define CIK__PIPE_CONFIG__ADDR_SURF_P4_16x16 5
#define CIK__PIPE_CONFIG__ADDR_SURF_P4_16x32 6
#define CIK__PIPE_CONFIG__ADDR_SURF_P4_32x32 7
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_16x16_8x16 8
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_16x32_8x16 9
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_32x32_8x16 10
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_16x32_16x16 11
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x16 12
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x32 13
#define CIK__PIPE_CONFIG__ADDR_SURF_P8_32x64_32x32 14
radeon: handle P16 pipe configs for Hawaii
2013-11-22 16:34:45 -07:00
#define CIK__PIPE_CONFIG__ADDR_SURF_P16_32X32_8X16 16
#define CIK__PIPE_CONFIG__ADDR_SURF_P16_32X32_16X16 17
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
#define CIK__GB_TILE_MODE__TILE_SPLIT(x) (((x) >> 11) & 0x7)
#define CIK__TILE_SPLIT__64B 0
#define CIK__TILE_SPLIT__128B 1
#define CIK__TILE_SPLIT__256B 2
#define CIK__TILE_SPLIT__512B 3
#define CIK__TILE_SPLIT__1024B 4
#define CIK__TILE_SPLIT__2048B 5
#define CIK__TILE_SPLIT__4096B 6
#define CIK__GB_TILE_MODE__SAMPLE_SPLIT(x) (((x) >> 25) & 0x3)
#define CIK__SAMPLE_SPLIT__1 0
#define CIK__SAMPLE_SPLIT__2 1
#define CIK__SAMPLE_SPLIT__4 2
#define CIK__SAMPLE_SPLIT__8 3
#define CIK__GB_MACROTILE_MODE__BANK_WIDTH(x) ((x) & 0x3)
#define CIK__BANK_WIDTH__1 0
#define CIK__BANK_WIDTH__2 1
#define CIK__BANK_WIDTH__4 2
#define CIK__BANK_WIDTH__8 3
#define CIK__GB_MACROTILE_MODE__BANK_HEIGHT(x) (((x) >> 2) & 0x3)
#define CIK__BANK_HEIGHT__1 0
#define CIK__BANK_HEIGHT__2 1
#define CIK__BANK_HEIGHT__4 2
#define CIK__BANK_HEIGHT__8 3
#define CIK__GB_MACROTILE_MODE__MACRO_TILE_ASPECT(x) (((x) >> 4) & 0x3)
#define CIK__MACRO_TILE_ASPECT__1 0
#define CIK__MACRO_TILE_ASPECT__2 1
#define CIK__MACRO_TILE_ASPECT__4 2
#define CIK__MACRO_TILE_ASPECT__8 3
#define CIK__GB_MACROTILE_MODE__NUM_BANKS(x) (((x) >> 6) & 0x3)
#define CIK__NUM_BANKS__2_BANK 0
#define CIK__NUM_BANKS__4_BANK 1
#define CIK__NUM_BANKS__8_BANK 2
#define CIK__NUM_BANKS__16_BANK 3
static void cik_get_2d_params(struct radeon_surface_manager *surf_man,
unsigned bpe, unsigned nsamples, bool is_color,
unsigned tile_mode,
uint32_t *num_pipes,
uint32_t *tile_split_ptr,
uint32_t *num_banks,
uint32_t *macro_tile_aspect,
uint32_t *bank_w,
uint32_t *bank_h)
{
uint32_t gb_tile_mode = surf_man->hw_info.tile_mode_array[tile_mode];
unsigned tileb_1x, tileb;
unsigned gb_macrotile_mode;
unsigned macrotile_index;
unsigned tile_split, sample_split;
if (num_pipes) {
switch (CIK__GB_TILE_MODE__PIPE_CONFIG(gb_tile_mode)) {
case CIK__PIPE_CONFIG__ADDR_SURF_P2:
default:
*num_pipes = 2;
break;
case CIK__PIPE_CONFIG__ADDR_SURF_P4_8x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P4_16x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P4_16x32:
case CIK__PIPE_CONFIG__ADDR_SURF_P4_32x32:
*num_pipes = 4;
break;
case CIK__PIPE_CONFIG__ADDR_SURF_P8_16x16_8x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P8_16x32_8x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P8_32x32_8x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P8_16x32_16x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x16:
case CIK__PIPE_CONFIG__ADDR_SURF_P8_32x32_16x32:
case CIK__PIPE_CONFIG__ADDR_SURF_P8_32x64_32x32:
*num_pipes = 8;
break;
radeon: handle P16 pipe configs for Hawaii
2013-11-22 16:34:45 -07:00
case CIK__PIPE_CONFIG__ADDR_SURF_P16_32X32_8X16:
case CIK__PIPE_CONFIG__ADDR_SURF_P16_32X32_16X16:
*num_pipes = 16;
break;
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
}
}
switch (CIK__GB_TILE_MODE__TILE_SPLIT(gb_tile_mode)) {
default:
case CIK__TILE_SPLIT__64B:
tile_split = 64;
break;
case CIK__TILE_SPLIT__128B:
tile_split = 128;
break;
case CIK__TILE_SPLIT__256B:
tile_split = 256;
break;
case CIK__TILE_SPLIT__512B:
tile_split = 512;
break;
case CIK__TILE_SPLIT__1024B:
tile_split = 1024;
break;
case CIK__TILE_SPLIT__2048B:
tile_split = 2048;
break;
case CIK__TILE_SPLIT__4096B:
tile_split = 4096;
break;
}
switch (CIK__GB_TILE_MODE__SAMPLE_SPLIT(gb_tile_mode)) {
default:
case CIK__SAMPLE_SPLIT__1:
sample_split = 1;
break;
case CIK__SAMPLE_SPLIT__2:
radeon: fix typo in sample split / fixes MSAA on Hawaii Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2014-07-26 12:07:33 -06:00
sample_split = 2;
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
break;
case CIK__SAMPLE_SPLIT__4:
sample_split = 4;
break;
case CIK__SAMPLE_SPLIT__8:
sample_split = 8;
break;
}
/* Adjust the tile split. */
tileb_1x = 8 * 8 * bpe;
if (is_color) {
tile_split = MAX2(256, sample_split * tileb_1x);
}
tile_split = MIN2(surf_man->hw_info.row_size, tile_split);
/* Determine the macrotile index. */
tileb = MIN2(tile_split, nsamples * tileb_1x);
for (macrotile_index = 0; tileb > 64; macrotile_index++) {
tileb >>= 1;
}
gb_macrotile_mode = surf_man->hw_info.macrotile_mode_array[macrotile_index];
if (tile_split_ptr) {
*tile_split_ptr = tile_split;
}
if (num_banks) {
switch (CIK__GB_MACROTILE_MODE__NUM_BANKS(gb_macrotile_mode)) {
default:
case CIK__NUM_BANKS__2_BANK:
*num_banks = 2;
break;
case CIK__NUM_BANKS__4_BANK:
*num_banks = 4;
break;
case CIK__NUM_BANKS__8_BANK:
*num_banks = 8;
break;
case CIK__NUM_BANKS__16_BANK:
*num_banks = 16;
break;
}
}
if (macro_tile_aspect) {
switch (CIK__GB_MACROTILE_MODE__MACRO_TILE_ASPECT(gb_macrotile_mode)) {
default:
case CIK__MACRO_TILE_ASPECT__1:
*macro_tile_aspect = 1;
break;
case CIK__MACRO_TILE_ASPECT__2:
*macro_tile_aspect = 2;
break;
case CIK__MACRO_TILE_ASPECT__4:
*macro_tile_aspect = 4;
break;
case CIK__MACRO_TILE_ASPECT__8:
*macro_tile_aspect = 8;
break;
}
}
if (bank_w) {
switch (CIK__GB_MACROTILE_MODE__BANK_WIDTH(gb_macrotile_mode)) {
default:
case CIK__BANK_WIDTH__1:
*bank_w = 1;
break;
case CIK__BANK_WIDTH__2:
*bank_w = 2;
break;
case CIK__BANK_WIDTH__4:
*bank_w = 4;
break;
case CIK__BANK_WIDTH__8:
*bank_w = 8;
break;
}
}
if (bank_h) {
switch (CIK__GB_MACROTILE_MODE__BANK_HEIGHT(gb_macrotile_mode)) {
default:
case CIK__BANK_HEIGHT__1:
*bank_h = 1;
break;
case CIK__BANK_HEIGHT__2:
*bank_h = 2;
break;
case CIK__BANK_HEIGHT__4:
*bank_h = 4;
break;
case CIK__BANK_HEIGHT__8:
*bank_h = 8;
break;
}
}
}
static int cik_init_hw_info(struct radeon_surface_manager *surf_man)
{
uint32_t tiling_config;
drmVersionPtr version;
int r;
r = radeon_get_value(surf_man->fd, RADEON_INFO_TILING_CONFIG,
&tiling_config);
if (r) {
return r;
}
surf_man->hw_info.allow_2d = 0;
version = drmGetVersion(surf_man->fd);
if (version && version->version_minor >= 35) {
if (!radeon_get_value(surf_man->fd, RADEON_INFO_SI_TILE_MODE_ARRAY, surf_man->hw_info.tile_mode_array) &&
!radeon_get_value(surf_man->fd, RADEON_INFO_CIK_MACROTILE_MODE_ARRAY, surf_man->hw_info.macrotile_mode_array)) {
surf_man->hw_info.allow_2d = 1;
}
}
drmFreeVersion(version);
switch (tiling_config & 0xf) {
case 0:
surf_man->hw_info.num_pipes = 1;
break;
case 1:
surf_man->hw_info.num_pipes = 2;
break;
case 2:
surf_man->hw_info.num_pipes = 4;
break;
case 3:
surf_man->hw_info.num_pipes = 8;
break;
default:
surf_man->hw_info.num_pipes = 8;
surf_man->hw_info.allow_2d = 0;
break;
}
switch ((tiling_config & 0xf0) >> 4) {
case 0:
surf_man->hw_info.num_banks = 4;
break;
case 1:
surf_man->hw_info.num_banks = 8;
break;
case 2:
surf_man->hw_info.num_banks = 16;
break;
default:
surf_man->hw_info.num_banks = 8;
surf_man->hw_info.allow_2d = 0;
break;
}
switch ((tiling_config & 0xf00) >> 8) {
case 0:
surf_man->hw_info.group_bytes = 256;
break;
case 1:
surf_man->hw_info.group_bytes = 512;
break;
default:
surf_man->hw_info.group_bytes = 256;
surf_man->hw_info.allow_2d = 0;
break;
}
switch ((tiling_config & 0xf000) >> 12) {
case 0:
surf_man->hw_info.row_size = 1024;
break;
case 1:
surf_man->hw_info.row_size = 2048;
break;
case 2:
surf_man->hw_info.row_size = 4096;
break;
default:
surf_man->hw_info.row_size = 4096;
surf_man->hw_info.allow_2d = 0;
break;
}
return 0;
}
static int cik_surface_sanity(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
unsigned mode, unsigned *tile_mode, unsigned *stencil_tile_mode)
{
/* check surface dimension */
if (surf->npix_x > 16384 || surf->npix_y > 16384 || surf->npix_z > 16384) {
return -EINVAL;
}
/* check mipmap last_level */
if (surf->last_level > 15) {
return -EINVAL;
}
/* force 1d on kernel that can't do 2d */
if (mode > RADEON_SURF_MODE_1D &&
(!surf_man->hw_info.allow_2d || !(surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX))) {
if (surf->nsamples > 1) {
fprintf(stderr, "radeon: Cannot use 1D tiling for an MSAA surface (%i).\n", __LINE__);
return -EFAULT;
}
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(mode, MODE);
}
if (surf->nsamples > 1 && mode != RADEON_SURF_MODE_2D) {
return -EINVAL;
}
if (!surf->tile_split) {
/* default value */
surf->mtilea = 1;
surf->bankw = 1;
radeon: Fix surf->bankh init by default value when surf->tile_split == 0 Signed-off-by: Maks Naumov <maksqwe1@ukr.net> Signed-off-by: Marek Olšák <marek.olsak@amd.com>
2014-08-18 13:59:35 -06:00
surf->bankh = 1;
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
surf->tile_split = 64;
surf->stencil_tile_split = 64;
}
switch (mode) {
case RADEON_SURF_MODE_2D: {
if (surf->flags & RADEON_SURF_Z_OR_SBUFFER) {
switch (surf->nsamples) {
case 1:
*tile_mode = CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_64;
break;
case 2:
case 4:
*tile_mode = CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_128;
break;
case 8:
*tile_mode = CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_256;
break;
default:
return -EINVAL;
}
if (surf->flags & RADEON_SURF_SBUFFER) {
*stencil_tile_mode = *tile_mode;
cik_get_2d_params(surf_man, 1, surf->nsamples, false,
*stencil_tile_mode, NULL,
&surf->stencil_tile_split,
NULL, NULL, NULL, NULL);
}
} else if (surf->flags & RADEON_SURF_SCANOUT) {
*tile_mode = CIK_TILE_MODE_COLOR_2D_SCANOUT;
} else {
*tile_mode = CIK_TILE_MODE_COLOR_2D;
}
/* retrieve tiling mode values */
cik_get_2d_params(surf_man, surf->bpe, surf->nsamples,
!(surf->flags & RADEON_SURF_Z_OR_SBUFFER), *tile_mode,
NULL, &surf->tile_split, NULL, &surf->mtilea,
&surf->bankw, &surf->bankh);
break;
}
case RADEON_SURF_MODE_1D:
if (surf->flags & RADEON_SURF_SBUFFER) {
*stencil_tile_mode = CIK_TILE_MODE_DEPTH_STENCIL_1D;
}
if (surf->flags & RADEON_SURF_ZBUFFER) {
*tile_mode = CIK_TILE_MODE_DEPTH_STENCIL_1D;
} else if (surf->flags & RADEON_SURF_SCANOUT) {
*tile_mode = SI_TILE_MODE_COLOR_1D_SCANOUT;
} else {
*tile_mode = SI_TILE_MODE_COLOR_1D;
}
break;
case RADEON_SURF_MODE_LINEAR_ALIGNED:
default:
radeon: fix missing stencil_tile_mode initialisation in the linear/fallback case ../radeon/radeon_surface.c:1611:13: error: 'stencil_tile_mode' may be used uninitialized [-Werror=maybe-uninitialized] 1611 | r = si_surface_init_1d(surf_man, surf, surf->stencil_level, 1, stencil_tile_mode, surf->bo_size, 0); | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2024-01-01 00:41:14 -07:00
*stencil_tile_mode = SI_TILE_MODE_COLOR_LINEAR_ALIGNED;
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
*tile_mode = SI_TILE_MODE_COLOR_LINEAR_ALIGNED;
}
return 0;
}
static int cik_surface_init_2d(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
struct radeon_surface_level *level,
unsigned bpe, unsigned tile_mode,
unsigned tile_split,
unsigned num_pipes, unsigned num_banks,
uint64_t offset,
unsigned start_level)
{
uint64_t aligned_offset = offset;
unsigned tilew, tileh, tileb_1x, tileb;
unsigned mtilew, mtileh, mtileb;
unsigned slice_pt;
unsigned i;
/* compute tile values */
tilew = 8;
tileh = 8;
tileb_1x = tilew * tileh * bpe;
tile_split = MIN2(surf_man->hw_info.row_size, tile_split);
tileb = surf->nsamples * tileb_1x;
/* slices per tile */
slice_pt = 1;
radeon: avoid possible divide by 0 in surface manager Some users report hitting a divide by 0 with the tile split in certain apps. Tile_split shouldn't ever be 0 unless the surface structure was not properly initialized. I think there may be some cases where mesa uses an improperly initialized surface struct, but I haven't had time to track it down. Bug: https://bugs.freedesktop.org/show_bug.cgi?id=72425 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com>
2013-12-10 10:38:26 -07:00
if (tileb > tile_split && tile_split) {
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
slice_pt = tileb / tile_split;
tileb = tileb / slice_pt;
}
/* macro tile width & height */
mtilew = (tilew * surf->bankw * num_pipes) * surf->mtilea;
mtileh = (tileh * surf->bankh * num_banks) / surf->mtilea;
/* macro tile bytes */
mtileb = (mtilew / tilew) * (mtileh / tileh) * tileb;
if (start_level <= 1) {
unsigned alignment = MAX2(256, mtileb);
surf->bo_alignment = MAX2(surf->bo_alignment, alignment);
if (aligned_offset) {
aligned_offset = ALIGN(aligned_offset, alignment);
}
}
/* build mipmap tree */
for (i = start_level; i <= surf->last_level; i++) {
level[i].mode = RADEON_SURF_MODE_2D;
si_surf_minify_2d(surf, level+i, bpe, i, slice_pt, mtilew, mtileh, 1, mtileb, aligned_offset);
if (level[i].mode == RADEON_SURF_MODE_1D) {
switch (tile_mode) {
case CIK_TILE_MODE_COLOR_2D:
tile_mode = SI_TILE_MODE_COLOR_1D;
break;
case CIK_TILE_MODE_COLOR_2D_SCANOUT:
tile_mode = SI_TILE_MODE_COLOR_1D_SCANOUT;
break;
case CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_64:
case CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_128:
case CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_256:
case CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_512:
case CIK_TILE_MODE_DEPTH_STENCIL_2D_TILESPLIT_ROW_SIZE:
tile_mode = CIK_TILE_MODE_DEPTH_STENCIL_1D;
break;
default:
return -EINVAL;
}
return si_surface_init_1d(surf_man, surf, level, bpe, tile_mode, offset, i);
}
/* level0 and first mipmap need to have alignment */
aligned_offset = offset = surf->bo_size;
if (i == 0) {
aligned_offset = ALIGN(aligned_offset, surf->bo_alignment);
}
if (surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX) {
if (surf->level == level) {
surf->tiling_index[i] = tile_mode;
/* it's ok because stencil is done after */
surf->stencil_tiling_index[i] = tile_mode;
} else {
surf->stencil_tiling_index[i] = tile_mode;
}
}
}
return 0;
}
static int cik_surface_init_2d_miptrees(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
unsigned tile_mode, unsigned stencil_tile_mode)
{
int r;
uint32_t num_pipes, num_banks;
cik_get_2d_params(surf_man, surf->bpe, surf->nsamples,
!(surf->flags & RADEON_SURF_Z_OR_SBUFFER), tile_mode,
&num_pipes, NULL, &num_banks, NULL, NULL, NULL);
r = cik_surface_init_2d(surf_man, surf, surf->level, surf->bpe, tile_mode,
surf->tile_split, num_pipes, num_banks, 0, 0);
if (r) {
return r;
}
if (surf->flags & RADEON_SURF_SBUFFER) {
r = cik_surface_init_2d(surf_man, surf, surf->stencil_level, 1, stencil_tile_mode,
surf->stencil_tile_split, num_pipes, num_banks,
surf->bo_size, 0);
surf->stencil_offset = surf->stencil_level[0].offset;
}
return r;
}
static int cik_surface_init(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned mode, tile_mode, stencil_tile_mode;
int r;
/* MSAA surfaces support the 2D mode only. */
if (surf->nsamples > 1) {
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_2D, MODE);
}
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER)) {
/* zbuffer only support 1D or 2D tiled surface */
switch (mode) {
case RADEON_SURF_MODE_1D:
case RADEON_SURF_MODE_2D:
break;
default:
mode = RADEON_SURF_MODE_1D;
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_1D, MODE);
break;
}
}
r = cik_surface_sanity(surf_man, surf, mode, &tile_mode, &stencil_tile_mode);
if (r) {
return r;
}
surf->stencil_offset = 0;
surf->bo_alignment = 0;
/* check tiling mode */
switch (mode) {
case RADEON_SURF_MODE_LINEAR:
r = r6_surface_init_linear(surf_man, surf, 0, 0);
break;
case RADEON_SURF_MODE_LINEAR_ALIGNED:
r = si_surface_init_linear_aligned(surf_man, surf, tile_mode, 0, 0);
break;
case RADEON_SURF_MODE_1D:
r = si_surface_init_1d_miptrees(surf_man, surf, tile_mode, stencil_tile_mode);
break;
case RADEON_SURF_MODE_2D:
r = cik_surface_init_2d_miptrees(surf_man, surf, tile_mode, stencil_tile_mode);
break;
default:
return -EINVAL;
}
return r;
}
/*
* depending on surface
*/
static int cik_surface_best(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
{
unsigned mode, tile_mode, stencil_tile_mode;
/* tiling mode */
mode = (surf->flags >> RADEON_SURF_MODE_SHIFT) & RADEON_SURF_MODE_MASK;
if (surf->flags & (RADEON_SURF_ZBUFFER | RADEON_SURF_SBUFFER) &&
!(surf->flags & RADEON_SURF_HAS_TILE_MODE_INDEX)) {
/* depth/stencil force 1d tiling for old mesa */
surf->flags = RADEON_SURF_CLR(surf->flags, MODE);
surf->flags |= RADEON_SURF_SET(RADEON_SURF_MODE_1D, MODE);
}
return cik_surface_sanity(surf_man, surf, mode, &tile_mode, &stencil_tile_mode);
}
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
/* ===========================================================================
* public API
*/
radeon: annotate public functions while read sym; do read f func line _ <<<$(cscope -d -L -1 $sym) if [ ! -z "$f" ]; then sed -i "${line}s/^/drm_public /" $f fi done < /tmp/a.txt In which /tmp/a.txt contains the public symbols from radeon-symbol-check. The idea here will be to switch the default visibility to hidden so we don't export symbols we shouldn't. Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> Acked-by: Eric Engestrom <eric.engestrom@intel.com>
2018-09-13 17:05:15 -06:00
drm_public struct radeon_surface_manager *
radeon: Use symbol visibility. All the bof_* symbols are now no longer exported. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2014-07-31 07:39:15 -06:00
radeon_surface_manager_new(int fd)
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
{
struct radeon_surface_manager *surf_man;
surf_man = calloc(1, sizeof(struct radeon_surface_manager));
if (surf_man == NULL) {
return NULL;
}
surf_man->fd = fd;
if (radeon_get_value(fd, RADEON_INFO_DEVICE_ID, &surf_man->device_id)) {
goto out_err;
}
if (radeon_get_family(surf_man)) {
goto out_err;
}
if (surf_man->family <= CHIP_RV740) {
if (r6_init_hw_info(surf_man)) {
goto out_err;
}
surf_man->surface_init = &r6_surface_init;
surf_man->surface_best = &r6_surface_best;
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
} else if (surf_man->family <= CHIP_ARUBA) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
if (eg_init_hw_info(surf_man)) {
goto out_err;
}
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
surf_man->surface_init = &eg_surface_init;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
surf_man->surface_best = &eg_surface_best;
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
} else if (surf_man->family < CHIP_BONAIRE) {
radeon: add si tiling support v5 v2: Only writte tile index if flags for it is set v3: Remove useless allow2d scanout flags v4: Split radeon_drm.h update to its own patch v5: update against lastest next tree for radeon Signed-off-by: Jerome Glisse <jglisse@redhat.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2013-04-08 11:35:37 -06:00
if (si_init_hw_info(surf_man)) {
goto out_err;
}
surf_man->surface_init = &si_surface_init;
surf_man->surface_best = &si_surface_best;
radeon: implement 2D tiling for CIK Bug fixes and simplification by Marek. We have to use the tile index of 0 for non-MSAA depth-stencil after all. Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-and-Tested-by: Michel Dänzer <michel.daenzer@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2013-11-22 15:31:13 -07:00
} else {
if (cik_init_hw_info(surf_man)) {
goto out_err;
}
surf_man->surface_init = &cik_surface_init;
surf_man->surface_best = &cik_surface_best;
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
}
return surf_man;
out_err:
free(surf_man);
return NULL;
}
radeon: annotate public functions while read sym; do read f func line _ <<<$(cscope -d -L -1 $sym) if [ ! -z "$f" ]; then sed -i "${line}s/^/drm_public /" $f fi done < /tmp/a.txt In which /tmp/a.txt contains the public symbols from radeon-symbol-check. The idea here will be to switch the default visibility to hidden so we don't export symbols we shouldn't. Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> Acked-by: Eric Engestrom <eric.engestrom@intel.com>
2018-09-13 17:05:15 -06:00
drm_public void
radeon: Use symbol visibility. All the bof_* symbols are now no longer exported. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2014-07-31 07:39:15 -06:00
radeon_surface_manager_free(struct radeon_surface_manager *surf_man)
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
{
free(surf_man);
}
static int radeon_surface_sanity(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf,
unsigned type,
unsigned mode)
{
if (surf_man == NULL || surf_man->surface_init == NULL || surf == NULL) {
return -EINVAL;
}
/* all dimension must be at least 1 ! */
if (!surf->npix_x || !surf->npix_y || !surf->npix_z) {
return -EINVAL;
}
radeon: fix surface API for good before anyone start relying on it The mipmap level computation was wrong, we need to know the block width, height, depth of compressed texture to properly compute this. Change API to provide block width, height, depth instead of nblk_x, nblk_y, nblk_z. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2012-02-03 10:22:11 -07:00
if (!surf->blk_w || !surf->blk_h || !surf->blk_d) {
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
return -EINVAL;
}
if (!surf->array_size) {
return -EINVAL;
}
/* array size must be a power of 2 */
surf->array_size = next_power_of_two(surf->array_size);
switch (surf->nsamples) {
case 1:
case 2:
case 4:
case 8:
break;
default:
return -EINVAL;
}
/* check type */
switch (type) {
case RADEON_SURF_TYPE_1D:
if (surf->npix_y > 1) {
return -EINVAL;
}
radeon: add fallthrough annotation GCC 7 started warning when a switch case has neither a `break` nor a "fallthrough" comment. Let's be explicit that we meant to fall through here. Signed-off-by: Eric Engestrom <eric@engestrom.ch> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2017-07-30 14:34:16 -06:00
/* fallthrough */
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
case RADEON_SURF_TYPE_2D:
if (surf->npix_z > 1) {
return -EINVAL;
}
break;
case RADEON_SURF_TYPE_CUBEMAP:
if (surf->npix_z > 1) {
return -EINVAL;
}
/* deal with cubemap as they were texture array */
if (surf_man->family >= CHIP_RV770) {
surf->array_size = 8;
} else {
surf->array_size = 6;
}
break;
case RADEON_SURF_TYPE_3D:
break;
case RADEON_SURF_TYPE_1D_ARRAY:
if (surf->npix_y > 1) {
return -EINVAL;
}
case RADEON_SURF_TYPE_2D_ARRAY:
break;
default:
return -EINVAL;
}
return 0;
}
radeon: annotate public functions while read sym; do read f func line _ <<<$(cscope -d -L -1 $sym) if [ ! -z "$f" ]; then sed -i "${line}s/^/drm_public /" $f fi done < /tmp/a.txt In which /tmp/a.txt contains the public symbols from radeon-symbol-check. The idea here will be to switch the default visibility to hidden so we don't export symbols we shouldn't. Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> Acked-by: Eric Engestrom <eric.engestrom@intel.com>
2018-09-13 17:05:15 -06:00
drm_public int
radeon: Use symbol visibility. All the bof_* symbols are now no longer exported. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2014-07-31 07:39:15 -06:00
radeon_surface_init(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
{
unsigned mode, type;
int r;
type = RADEON_SURF_GET(surf->flags, TYPE);
mode = RADEON_SURF_GET(surf->flags, MODE);
r = radeon_surface_sanity(surf_man, surf, type, mode);
if (r) {
return r;
}
return surf_man->surface_init(surf_man, surf);
}
radeon: annotate public functions while read sym; do read f func line _ <<<$(cscope -d -L -1 $sym) if [ ! -z "$f" ]; then sed -i "${line}s/^/drm_public /" $f fi done < /tmp/a.txt In which /tmp/a.txt contains the public symbols from radeon-symbol-check. The idea here will be to switch the default visibility to hidden so we don't export symbols we shouldn't. Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com> Acked-by: Eric Engestrom <eric.engestrom@intel.com>
2018-09-13 17:05:15 -06:00
drm_public int
radeon: Use symbol visibility. All the bof_* symbols are now no longer exported. Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2014-07-31 07:39:15 -06:00
radeon_surface_best(struct radeon_surface_manager *surf_man,
struct radeon_surface *surf)
radeon: add surface allocator helper v10 The surface allocator is able to build complete miptree when allocating surface for r600/r700/evergreen/northern islands GPU family. It also compute bo size and alignment for render buffer, depth buffer and scanout buffer. v2 fix r6xx/r7xx 2D tiling width align computation v3 add tile split support and fix 1d texture alignment v4 rework to more properly support compressed format, split surface pixel size and surface element size in separate fields v5 support texture array (still issue on r6xx) v6 split surface value computation and mipmap tree building, rework eg and newer computation v7 add a check for tile split and 2d tiled v8 initialize mode value before testing it in all case, reenable 2D macro tile mode on r6xx for cubemap and array. Fix cubemap to force array size to the number of face. v9 fix handling of stencil buffer on evergreen v10 on evergreen depth buffer need to have enough room for a stencil buffer just after depth one Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2011-12-09 19:07:15 -07:00
{
unsigned mode, type;
int r;
type = RADEON_SURF_GET(surf->flags, TYPE);
mode = RADEON_SURF_GET(surf->flags, MODE);
r = radeon_surface_sanity(surf_man, surf, type, mode);
if (r) {
return r;
}
return surf_man->surface_best(surf_man, surf);
}