>From 8f47a2a76ceb0638cfd3b053e7cb2bf3dc1f8b4a Mon Sep 17 00:00:00 2001 From: Alex Deucher <[email protected]> Date: Thu, 18 Feb 2010 13:39:36 -0500 Subject: [PATCH] drm/radeon/r7xx: fixes to gfx init
- RV740 requires a special backend map - updated swizzle modes for backend map setup - fix programming of a few gfx regs This fixes occulusion queries and rendering errors on RV740. Signed-off-by: Alex Deucher <[email protected]> --- drivers/gpu/drm/radeon/r600_cp.c | 185 +++++++++++++++++++++++++++---------- drivers/gpu/drm/radeon/rv770.c | 189 +++++++++++++++++++++++++++----------- 2 files changed, 272 insertions(+), 102 deletions(-) diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c index d9712a1..b90d9e6 100644 --- a/drivers/gpu/drm/radeon/r600_cp.c +++ b/drivers/gpu/drm/radeon/r600_cp.c @@ -1162,7 +1162,8 @@ static void r600_gfx_init(struct drm_device *dev, } -static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, +static u32 r700_get_tile_pipe_to_backend_map(drm_radeon_private_t *dev_priv, + u32 num_tile_pipes, u32 num_backends, u32 backend_disable_mask) { @@ -1173,6 +1174,7 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, u32 swizzle_pipe[R7XX_MAX_PIPES]; u32 cur_backend; u32 i; + bool force_no_swizzle; if (num_tile_pipes > R7XX_MAX_PIPES) num_tile_pipes = R7XX_MAX_PIPES; @@ -1202,6 +1204,18 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, if (enabled_backends_count != num_backends) num_backends = enabled_backends_count; + switch (dev_priv->flags & RADEON_FAMILY_MASK) { + case CHIP_RV770: + case CHIP_RV730: + force_no_swizzle = false; + break; + case CHIP_RV710: + case CHIP_RV740: + default: + force_no_swizzle = true; + break; + } + memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * R7XX_MAX_PIPES); switch (num_tile_pipes) { case 1: @@ -1212,49 +1226,100 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, swizzle_pipe[1] = 1; break; case 3: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 1; + } break; case 4: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 3; - swizzle_pipe[3] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 3; + swizzle_pipe[3] = 1; + } break; case 5: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 1; - swizzle_pipe[4] = 3; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 1; + swizzle_pipe[4] = 3; + } break; case 6: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 5; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 5; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + } break; case 7: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 5; + } break; case 8: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 7; - swizzle_pipe[7] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + swizzle_pipe[7] = 7; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 7; + swizzle_pipe[7] = 5; + } break; } @@ -1275,8 +1340,10 @@ static void r700_gfx_init(struct drm_device *dev, drm_radeon_private_t *dev_priv) { int i, j, num_qd_pipes; + u32 ta_aux_cntl; u32 sx_debug_1; u32 smx_dc_ctl0; + u32 db_debug3; u32 num_gs_verts_per_thread; u32 vgt_gs_per_es; u32 gs_prim_buffer_depth = 0; @@ -1439,9 +1506,15 @@ static void r700_gfx_init(struct drm_device *dev, gb_tiling_config |= R600_BANK_SWAPS(1); - backend_map = r700_get_tile_pipe_to_backend_map(dev_priv->r600_max_tile_pipes, - dev_priv->r600_max_backends, - (0xff << dev_priv->r600_max_backends) & 0xff); + /* RV740 has a special backend map */ + if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV740) + backend_map = 0x28; + else + backend_map = r700_get_tile_pipe_to_backend_map(dev_priv, + dev_priv->r600_max_tile_pipes, + dev_priv->r600_max_backends, + ((R7XX_MAX_BACKENDS_MASK << dev_priv->r600_max_backends) + & R7XX_MAX_BACKENDS_MASK)); gb_tiling_config |= R600_BACKEND_MAP(backend_map); cc_gc_shader_pipe_config = @@ -1488,10 +1561,8 @@ static void r700_gfx_init(struct drm_device *dev, RADEON_WRITE(R600_CP_MEQ_THRESHOLDS, R700_STQ_SPLIT(0x30)); - RADEON_WRITE(R600_TA_CNTL_AUX, (R600_DISABLE_CUBE_ANISO | - R600_SYNC_GRADIENT | - R600_SYNC_WALKER | - R600_SYNC_ALIGNER)); + ta_aux_cntl = RADEON_READ(R600_TA_CNTL_AUX); + RADEON_WRITE(R600_TA_CNTL_AUX, ta_aux_cntl | R600_DISABLE_CUBE_ANISO); sx_debug_1 = RADEON_READ(R700_SX_DEBUG_1); sx_debug_1 |= R700_ENABLE_NEW_SMX_ADDRESS; @@ -1502,14 +1573,28 @@ static void r700_gfx_init(struct drm_device *dev, smx_dc_ctl0 |= R700_CACHE_DEPTH((dev_priv->r700_sx_num_of_sets * 64) - 1); RADEON_WRITE(R600_SMX_DC_CTL0, smx_dc_ctl0); - RADEON_WRITE(R700_SMX_EVENT_CTL, (R700_ES_FLUSH_CTL(4) | - R700_GS_FLUSH_CTL(4) | - R700_ACK_FLUSH_CTL(3) | - R700_SYNC_FLUSH_CTL)); + if ((dev_priv->flags & RADEON_FAMILY_MASK) != CHIP_RV740) + RADEON_WRITE(R700_SMX_EVENT_CTL, (R700_ES_FLUSH_CTL(4) | + R700_GS_FLUSH_CTL(4) | + R700_ACK_FLUSH_CTL(3) | + R700_SYNC_FLUSH_CTL)); - if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV770) - RADEON_WRITE(R700_DB_DEBUG3, R700_DB_CLK_OFF_DELAY(0x1f)); - else { + db_debug3 = RADEON_READ(R700_DB_DEBUG3); + db_debug3 &= ~R700_DB_CLK_OFF_DELAY(0x1f); + switch (dev_priv->flags & RADEON_FAMILY_MASK) { + case CHIP_RV770: + case CHIP_RV740: + db_debug3 |= R700_DB_CLK_OFF_DELAY(0x1f); + break; + case CHIP_RV710: + case CHIP_RV730: + default: + db_debug3 |= R700_DB_CLK_OFF_DELAY(2); + break; + } + RADEON_WRITE(R700_DB_DEBUG3, db_debug3); + + if ((dev_priv->flags & RADEON_FAMILY_MASK) != CHIP_RV770) { db_debug4 = RADEON_READ(RV700_DB_DEBUG4); db_debug4 |= RV700_DISABLE_TILE_COVERED_FOR_PS_ITER; RADEON_WRITE(RV700_DB_DEBUG4, db_debug4); @@ -1538,10 +1623,10 @@ static void r700_gfx_init(struct drm_device *dev, R600_ALU_UPDATE_FIFO_HIWATER(0x8)); switch (dev_priv->flags & RADEON_FAMILY_MASK) { case CHIP_RV770: - sq_ms_fifo_sizes |= R600_FETCH_FIFO_HIWATER(0x1); - break; case CHIP_RV730: case CHIP_RV710: + sq_ms_fifo_sizes |= R600_FETCH_FIFO_HIWATER(0x1); + break; case CHIP_RV740: default: sq_ms_fifo_sizes |= R600_FETCH_FIFO_HIWATER(0x4); diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index fbec052..3fce1e7 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -274,9 +274,10 @@ static int rv770_cp_load_microcode(struct radeon_device *rdev) /* * Core functions */ -static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, - u32 num_backends, - u32 backend_disable_mask) +static u32 r700_get_tile_pipe_to_backend_map(struct radeon_device *rdev, + u32 num_tile_pipes, + u32 num_backends, + u32 backend_disable_mask) { u32 backend_map = 0; u32 enabled_backends_mask; @@ -285,6 +286,7 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, u32 swizzle_pipe[R7XX_MAX_PIPES]; u32 cur_backend; u32 i; + bool force_no_swizzle; if (num_tile_pipes > R7XX_MAX_PIPES) num_tile_pipes = R7XX_MAX_PIPES; @@ -314,6 +316,18 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, if (enabled_backends_count != num_backends) num_backends = enabled_backends_count; + switch (rdev->family) { + case CHIP_RV770: + case CHIP_RV730: + force_no_swizzle = false; + break; + case CHIP_RV710: + case CHIP_RV740: + default: + force_no_swizzle = true; + break; + } + memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * R7XX_MAX_PIPES); switch (num_tile_pipes) { case 1: @@ -324,49 +338,100 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, swizzle_pipe[1] = 1; break; case 3: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 1; + } break; case 4: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 3; - swizzle_pipe[3] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 3; + swizzle_pipe[3] = 1; + } break; case 5: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 1; - swizzle_pipe[4] = 3; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 1; + swizzle_pipe[4] = 3; + } break; case 6: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 5; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 5; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + } break; case 7: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 5; + } break; case 8: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 7; - swizzle_pipe[7] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + swizzle_pipe[7] = 7; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 7; + swizzle_pipe[7] = 5; + } break; } @@ -386,8 +451,10 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, static void rv770_gpu_init(struct radeon_device *rdev) { int i, j, num_qd_pipes; + u32 ta_aux_cntl; u32 sx_debug_1; u32 smx_dc_ctl0; + u32 db_debug3; u32 num_gs_verts_per_thread; u32 vgt_gs_per_es; u32 gs_prim_buffer_depth = 0; @@ -556,9 +623,15 @@ static void rv770_gpu_init(struct radeon_device *rdev) gb_tiling_config |= BANK_SWAPS(1); - backend_map = r700_get_tile_pipe_to_backend_map(rdev->config.rv770.max_tile_pipes, - rdev->config.rv770.max_backends, - (0xff << rdev->config.rv770.max_backends) & 0xff); + /* RV740 has a special backend map */ + if (rdev->family == CHIP_RV740) + backend_map = 0x28; + else + backend_map = r700_get_tile_pipe_to_backend_map(rdev, + rdev->config.rv770.max_tile_pipes, + rdev->config.rv770.max_backends, + ((R7XX_MAX_BACKENDS_MASK << rdev->config.rv770.max_backends) + & R7XX_MAX_BACKENDS_MASK)); gb_tiling_config |= BACKEND_MAP(backend_map); cc_gc_shader_pipe_config = @@ -594,10 +667,8 @@ static void rv770_gpu_init(struct radeon_device *rdev) WREG32(CP_MEQ_THRESHOLDS, STQ_SPLIT(0x30)); - WREG32(TA_CNTL_AUX, (DISABLE_CUBE_ANISO | - SYNC_GRADIENT | - SYNC_WALKER | - SYNC_ALIGNER)); + ta_aux_cntl = RREG32(TA_CNTL_AUX); + WREG32(TA_CNTL_AUX, ta_aux_cntl | DISABLE_CUBE_ANISO); sx_debug_1 = RREG32(SX_DEBUG_1); sx_debug_1 |= ENABLE_NEW_SMX_ADDRESS; @@ -608,14 +679,28 @@ static void rv770_gpu_init(struct radeon_device *rdev) smx_dc_ctl0 |= CACHE_DEPTH((rdev->config.rv770.sx_num_of_sets * 64) - 1); WREG32(SMX_DC_CTL0, smx_dc_ctl0); - WREG32(SMX_EVENT_CTL, (ES_FLUSH_CTL(4) | - GS_FLUSH_CTL(4) | - ACK_FLUSH_CTL(3) | - SYNC_FLUSH_CTL)); + if (rdev->family != CHIP_RV740) + WREG32(SMX_EVENT_CTL, (ES_FLUSH_CTL(4) | + GS_FLUSH_CTL(4) | + ACK_FLUSH_CTL(3) | + SYNC_FLUSH_CTL)); - if (rdev->family == CHIP_RV770) - WREG32(DB_DEBUG3, DB_CLK_OFF_DELAY(0x1f)); - else { + db_debug3 = RREG32(DB_DEBUG3); + db_debug3 &= ~DB_CLK_OFF_DELAY(0x1f); + switch (rdev->family) { + case CHIP_RV770: + case CHIP_RV740: + db_debug3 |= DB_CLK_OFF_DELAY(0x1f); + break; + case CHIP_RV710: + case CHIP_RV730: + default: + db_debug3 |= DB_CLK_OFF_DELAY(2); + break; + } + WREG32(DB_DEBUG3, db_debug3); + + if (rdev->family != CHIP_RV770) { db_debug4 = RREG32(DB_DEBUG4); db_debug4 |= DISABLE_TILE_COVERED_FOR_PS_ITER; WREG32(DB_DEBUG4, db_debug4); @@ -644,10 +729,10 @@ static void rv770_gpu_init(struct radeon_device *rdev) ALU_UPDATE_FIFO_HIWATER(0x8)); switch (rdev->family) { case CHIP_RV770: - sq_ms_fifo_sizes |= FETCH_FIFO_HIWATER(0x1); - break; case CHIP_RV730: case CHIP_RV710: + sq_ms_fifo_sizes |= FETCH_FIFO_HIWATER(0x1); + break; case CHIP_RV740: default: sq_ms_fifo_sizes |= FETCH_FIFO_HIWATER(0x4); -- 1.5.6.3
0001-drm-radeon-r7xx-fixes-to-gfx-init.patch
Description: application/mbox
------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev
-- _______________________________________________ Dri-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dri-devel
