diff options
Diffstat (limited to 'drivers/gpu/drm/radeon')
34 files changed, 451 insertions, 200 deletions
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index d56f08d3cbd..a2199fe9fa9 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -48,29 +48,29 @@ static void atombios_overscan_setup(struct drm_crtc *crtc, switch (radeon_crtc->rmx_type) { case RMX_CENTER: - args.usOverscanTop = (adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2; - args.usOverscanBottom = (adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2; - args.usOverscanLeft = (adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2; - args.usOverscanRight = (adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2; + args.usOverscanTop = cpu_to_le16((adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2); + args.usOverscanBottom = cpu_to_le16((adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2); + args.usOverscanLeft = cpu_to_le16((adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2); + args.usOverscanRight = cpu_to_le16((adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2); break; case RMX_ASPECT: a1 = mode->crtc_vdisplay * adjusted_mode->crtc_hdisplay; a2 = adjusted_mode->crtc_vdisplay * mode->crtc_hdisplay; if (a1 > a2) { - args.usOverscanLeft = (adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2; - args.usOverscanRight = (adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2; + args.usOverscanLeft = cpu_to_le16((adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2); + args.usOverscanRight = cpu_to_le16((adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2); } else if (a2 > a1) { - args.usOverscanLeft = (adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2; - args.usOverscanRight = (adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2; + args.usOverscanLeft = cpu_to_le16((adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2); + args.usOverscanRight = cpu_to_le16((adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2); } break; case RMX_FULL: default: - args.usOverscanRight = radeon_crtc->h_border; - args.usOverscanLeft = radeon_crtc->h_border; - args.usOverscanBottom = radeon_crtc->v_border; - args.usOverscanTop = radeon_crtc->v_border; + args.usOverscanRight = cpu_to_le16(radeon_crtc->h_border); + args.usOverscanLeft = cpu_to_le16(radeon_crtc->h_border); + args.usOverscanBottom = cpu_to_le16(radeon_crtc->v_border); + args.usOverscanTop = cpu_to_le16(radeon_crtc->v_border); break; } atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); @@ -419,23 +419,23 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, memset(&args, 0, sizeof(args)); if (ASIC_IS_DCE5(rdev)) { - args.v3.usSpreadSpectrumAmountFrac = 0; + args.v3.usSpreadSpectrumAmountFrac = cpu_to_le16(0); args.v3.ucSpreadSpectrumType = ss->type; switch (pll_id) { case ATOM_PPLL1: args.v3.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V3_P1PLL; - args.v3.usSpreadSpectrumAmount = ss->amount; - args.v3.usSpreadSpectrumStep = ss->step; + args.v3.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_PPLL2: args.v3.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V3_P2PLL; - args.v3.usSpreadSpectrumAmount = ss->amount; - args.v3.usSpreadSpectrumStep = ss->step; + args.v3.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_DCPLL: args.v3.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V3_DCPLL; - args.v3.usSpreadSpectrumAmount = 0; - args.v3.usSpreadSpectrumStep = 0; + args.v3.usSpreadSpectrumAmount = cpu_to_le16(0); + args.v3.usSpreadSpectrumStep = cpu_to_le16(0); break; case ATOM_PPLL_INVALID: return; @@ -447,18 +447,18 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, switch (pll_id) { case ATOM_PPLL1: args.v2.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V2_P1PLL; - args.v2.usSpreadSpectrumAmount = ss->amount; - args.v2.usSpreadSpectrumStep = ss->step; + args.v2.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v2.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_PPLL2: args.v2.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V2_P2PLL; - args.v2.usSpreadSpectrumAmount = ss->amount; - args.v2.usSpreadSpectrumStep = ss->step; + args.v2.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v2.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_DCPLL: args.v2.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V2_DCPLL; - args.v2.usSpreadSpectrumAmount = 0; - args.v2.usSpreadSpectrumStep = 0; + args.v2.usSpreadSpectrumAmount = cpu_to_le16(0); + args.v2.usSpreadSpectrumStep = cpu_to_le16(0); break; case ATOM_PPLL_INVALID: return; @@ -538,7 +538,6 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, pll->flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; else pll->flags |= RADEON_PLL_PREFER_LOW_REF_DIV; - } list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { @@ -555,29 +554,28 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, dp_clock = dig_connector->dp_clock; } } -/* this might work properly with the new pll algo */ -#if 0 /* doesn't work properly on some laptops */ + /* use recommended ref_div for ss */ if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) { if (ss_enabled) { if (ss->refdiv) { + pll->flags |= RADEON_PLL_PREFER_MINM_OVER_MAXP; pll->flags |= RADEON_PLL_USE_REF_DIV; pll->reference_div = ss->refdiv; + if (ASIC_IS_AVIVO(rdev)) + pll->flags |= RADEON_PLL_USE_FRAC_FB_DIV; } } } -#endif + if (ASIC_IS_AVIVO(rdev)) { /* DVO wants 2x pixel clock if the DVO chip is in 12 bit mode */ if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1) adjusted_clock = mode->clock * 2; if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) pll->flags |= RADEON_PLL_PREFER_CLOSEST_LOWER; - /* rv515 needs more testing with this option */ - if (rdev->family != CHIP_RV515) { - if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) - pll->flags |= RADEON_PLL_IS_LCD; - } + if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) + pll->flags |= RADEON_PLL_IS_LCD; } else { if (encoder->encoder_type != DRM_MODE_ENCODER_DAC) pll->flags |= RADEON_PLL_NO_ODD_POST_DIV; @@ -664,10 +662,12 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, index, (uint32_t *)&args); adjusted_clock = le32_to_cpu(args.v3.sOutput.ulDispPllFreq) * 10; if (args.v3.sOutput.ucRefDiv) { + pll->flags |= RADEON_PLL_USE_FRAC_FB_DIV; pll->flags |= RADEON_PLL_USE_REF_DIV; pll->reference_div = args.v3.sOutput.ucRefDiv; } if (args.v3.sOutput.ucPostDiv) { + pll->flags |= RADEON_PLL_USE_FRAC_FB_DIV; pll->flags |= RADEON_PLL_USE_POST_DIV; pll->post_div = args.v3.sOutput.ucPostDiv; } @@ -721,14 +721,14 @@ static void atombios_crtc_set_dcpll(struct drm_crtc *crtc, * SetPixelClock provides the dividers */ args.v5.ucCRTC = ATOM_CRTC_INVALID; - args.v5.usPixelClock = dispclk; + args.v5.usPixelClock = cpu_to_le16(dispclk); args.v5.ucPpll = ATOM_DCPLL; break; case 6: /* if the default dcpll clock is specified, * SetPixelClock provides the dividers */ - args.v6.ulDispEngClkFreq = dispclk; + args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk); args.v6.ucPpll = ATOM_DCPLL; break; default: @@ -957,11 +957,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode /* adjust pixel clock as needed */ adjusted_clock = atombios_adjust_pll(crtc, mode, pll, ss_enabled, &ss); - /* rv515 seems happier with the old algo */ - if (rdev->family == CHIP_RV515) - radeon_compute_pll_legacy(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, - &ref_div, &post_div); - else if (ASIC_IS_AVIVO(rdev)) + if (ASIC_IS_AVIVO(rdev)) radeon_compute_pll_avivo(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, &ref_div, &post_div); else @@ -995,9 +991,9 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode } } -static int evergreen_crtc_do_set_base(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int x, int y, int atomic) +static int dce4_crtc_do_set_base(struct drm_crtc *crtc, + struct drm_framebuffer *fb, + int x, int y, int atomic) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct drm_device *dev = crtc->dev; @@ -1137,12 +1133,6 @@ static int evergreen_crtc_do_set_base(struct drm_crtc *crtc, WREG32(EVERGREEN_VIEWPORT_SIZE + radeon_crtc->crtc_offset, (crtc->mode.hdisplay << 16) | crtc->mode.vdisplay); - if (crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) - WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, - EVERGREEN_INTERLEAVE_EN); - else - WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, 0); - if (!atomic && fb && fb != crtc->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = gem_to_radeon_bo(radeon_fb->obj); @@ -1300,12 +1290,6 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, WREG32(AVIVO_D1MODE_VIEWPORT_SIZE + radeon_crtc->crtc_offset, (crtc->mode.hdisplay << 16) | crtc->mode.vdisplay); - if (crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) - WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, - AVIVO_D1MODE_INTERLEAVE_EN); - else - WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, 0); - if (!atomic && fb && fb != crtc->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = gem_to_radeon_bo(radeon_fb->obj); @@ -1329,7 +1313,7 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, struct radeon_device *rdev = dev->dev_private; if (ASIC_IS_DCE4(rdev)) - return evergreen_crtc_do_set_base(crtc, old_fb, x, y, 0); + return dce4_crtc_do_set_base(crtc, old_fb, x, y, 0); else if (ASIC_IS_AVIVO(rdev)) return avivo_crtc_do_set_base(crtc, old_fb, x, y, 0); else @@ -1344,7 +1328,7 @@ int atombios_crtc_set_base_atomic(struct drm_crtc *crtc, struct radeon_device *rdev = dev->dev_private; if (ASIC_IS_DCE4(rdev)) - return evergreen_crtc_do_set_base(crtc, fb, x, y, 1); + return dce4_crtc_do_set_base(crtc, fb, x, y, 1); else if (ASIC_IS_AVIVO(rdev)) return avivo_crtc_do_set_base(crtc, fb, x, y, 1); else diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index d4045223d0f..789441ed983 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -1192,7 +1192,11 @@ void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(rdev, 1); /* FIXME: implement */ radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, ib->gpu_addr & 0xFFFFFFFC); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (ib->gpu_addr & 0xFFFFFFFC)); radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF); radeon_ring_write(rdev, ib->length_dw); } @@ -1207,7 +1211,11 @@ static int evergreen_cp_load_microcode(struct radeon_device *rdev) return -EINVAL; r700_cp_stop(rdev); - WREG32(CP_RB_CNTL, RB_NO_UPDATE | (15 << 8) | (3 << 0)); + WREG32(CP_RB_CNTL, +#ifdef __BIG_ENDIAN + BUF_SWAP_32BIT | +#endif + RB_NO_UPDATE | RB_BLKSZ(15) | RB_BUFSZ(3)); fw_data = (const __be32 *)rdev->pfp_fw->data; WREG32(CP_PFP_UCODE_ADDR, 0); @@ -1326,7 +1334,11 @@ int evergreen_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_WPTR, 0); /* set the wb address wether it's enabled or not */ - WREG32(CP_RB_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC); + WREG32(CP_RB_RPTR_ADDR, +#ifdef __BIG_ENDIAN + RB_RPTR_SWAP(2) | +#endif + ((rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC)); WREG32(CP_RB_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF); WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF); @@ -2627,8 +2639,8 @@ restart_ih: while (rptr != wptr) { /* wptr/rptr are in bytes! */ ring_index = rptr / 4; - src_id = rdev->ih.ring[ring_index] & 0xff; - src_data = rdev->ih.ring[ring_index + 1] & 0xfffffff; + src_id = le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff; + src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff; switch (src_id) { case 1: /* D1 vblank/vline */ diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c index 2ed930e02f3..3218287f4c5 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_kms.c @@ -55,7 +55,7 @@ set_render_target(struct radeon_device *rdev, int format, if (h < 8) h = 8; - cb_color_info = ((format << 2) | (1 << 24)); + cb_color_info = ((format << 2) | (1 << 24) | (1 << 8)); pitch = (w / 8) - 1; slice = ((w * h) / 64) - 1; @@ -133,6 +133,9 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) /* high addr, stride */ sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8)); +#ifdef __BIG_ENDIAN + sq_vtx_constant_word2 |= (2 << 30); +#endif /* xyzw swizzles */ sq_vtx_constant_word3 = (0 << 3) | (1 << 6) | (2 << 9) | (3 << 12); @@ -173,7 +176,7 @@ set_tex_resource(struct radeon_device *rdev, sq_tex_resource_word0 = (1 << 0); /* 2D */ sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 6) | ((w - 1) << 18)); - sq_tex_resource_word1 = ((h - 1) << 0); + sq_tex_resource_word1 = ((h - 1) << 0) | (1 << 28); /* xyzw swizzles */ sq_tex_resource_word4 = (0 << 16) | (1 << 19) | (2 << 22) | (3 << 25); @@ -221,7 +224,11 @@ draw_auto(struct radeon_device *rdev) radeon_ring_write(rdev, DI_PT_RECTLIST); radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(rdev, DI_INDEX_SIZE_16_BIT); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 2) | +#endif + DI_INDEX_SIZE_16_BIT); radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0)); radeon_ring_write(rdev, 1); @@ -541,7 +548,7 @@ static inline uint32_t i2f(uint32_t input) int evergreen_blit_init(struct radeon_device *rdev) { u32 obj_size; - int r, dwords; + int i, r, dwords; void *ptr; u32 packet2s[16]; int num_packet2s = 0; @@ -557,7 +564,7 @@ int evergreen_blit_init(struct radeon_device *rdev) dwords = rdev->r600_blit.state_len; while (dwords & 0xf) { - packet2s[num_packet2s++] = PACKET2(0); + packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0)); dwords++; } @@ -598,8 +605,10 @@ int evergreen_blit_init(struct radeon_device *rdev) if (num_packet2s) memcpy_toio(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4), packet2s, num_packet2s * 4); - memcpy(ptr + rdev->r600_blit.vs_offset, evergreen_vs, evergreen_vs_size * 4); - memcpy(ptr + rdev->r600_blit.ps_offset, evergreen_ps, evergreen_ps_size * 4); + for (i = 0; i < evergreen_vs_size; i++) + *(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(evergreen_vs[i]); + for (i = 0; i < evergreen_ps_size; i++) + *(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(evergreen_ps[i]); radeon_bo_kunmap(rdev->r600_blit.shader_obj); radeon_bo_unreserve(rdev->r600_blit.shader_obj); diff --git a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c index ef1d28c07fb..3a10399e006 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c @@ -311,11 +311,19 @@ const u32 evergreen_vs[] = 0x00000000, 0x3c000000, 0x67961001, +#ifdef __BIG_ENDIAN + 0x000a0000, +#else 0x00080000, +#endif 0x00000000, 0x1c000000, 0x67961000, +#ifdef __BIG_ENDIAN + 0x00020008, +#else 0x00000008, +#endif 0x00000000, }; diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h index 21e839bd20e..9aaa3f0c937 100644 --- a/drivers/gpu/drm/radeon/evergreend.h +++ b/drivers/gpu/drm/radeon/evergreend.h @@ -98,6 +98,7 @@ #define BUF_SWAP_32BIT (2 << 16) #define CP_RB_RPTR 0x8700 #define CP_RB_RPTR_ADDR 0xC10C +#define RB_RPTR_SWAP(x) ((x) << 0) #define CP_RB_RPTR_ADDR_HI 0xC110 #define CP_RB_RPTR_WR 0xC108 #define CP_RB_WPTR 0xC114 diff --git a/drivers/gpu/drm/radeon/mkregtable.c b/drivers/gpu/drm/radeon/mkregtable.c index 607241c6a8a..5a82b6b7584 100644 --- a/drivers/gpu/drm/radeon/mkregtable.c +++ b/drivers/gpu/drm/radeon/mkregtable.c @@ -673,8 +673,10 @@ static int parser_auth(struct table *t, const char *filename) last_reg = strtol(last_reg_s, NULL, 16); do { - if (fgets(buf, 1024, file) == NULL) + if (fgets(buf, 1024, file) == NULL) { + fclose(file); return -1; + } len = strlen(buf); if (ftell(file) == end) done = 1; @@ -685,6 +687,7 @@ static int parser_auth(struct table *t, const char *filename) fprintf(stderr, "Error matching regular expression %d in %s\n", r, filename); + fclose(file); return -1; } else { buf[match[0].rm_eo] = 0; diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 5f15820efe1..93fa735c8c1 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -1427,6 +1427,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->zb.robj = reloc->robj; track->zb.offset = idx_value; + track->zb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_RB3D_COLOROFFSET: @@ -1439,6 +1440,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; + track->cb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_PP_TXOFFSET_0: @@ -1454,6 +1456,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T0_0: case RADEON_PP_CUBIC_OFFSET_T0_1: @@ -1471,6 +1474,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[0].cube_info[i].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[0].cube_info[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T1_0: case RADEON_PP_CUBIC_OFFSET_T1_1: @@ -1488,6 +1492,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[1].cube_info[i].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[1].cube_info[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T2_0: case RADEON_PP_CUBIC_OFFSET_T2_1: @@ -1505,9 +1510,12 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[2].cube_info[i].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[2].cube_info[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_RE_WIDTH_HEIGHT: track->maxy = ((idx_value >> 16) & 0x7FF); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_COLORPITCH: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1528,9 +1536,11 @@ static int r100_packet0_check(struct radeon_cs_parser *p, ib[idx] = tmp; track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; + track->cb_dirty = true; break; case RADEON_RB3D_DEPTHPITCH: track->zb.pitch = idx_value & RADEON_DEPTHPITCH_MASK; + track->zb_dirty = true; break; case RADEON_RB3D_CNTL: switch ((idx_value >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) { @@ -1555,6 +1565,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return -EINVAL; } track->z_enabled = !!(idx_value & RADEON_Z_ENABLE); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_ZSTENCILCNTL: switch (idx_value & 0xf) { @@ -1572,6 +1584,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, default: break; } + track->zb_dirty = true; break; case RADEON_RB3D_ZPASS_ADDR: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1588,6 +1601,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, uint32_t temp = idx_value >> 4; for (i = 0; i < track->num_texture; i++) track->textures[i].enabled = !!(temp & (1 << i)); + track->tex_dirty = true; } break; case RADEON_SE_VF_CNTL: @@ -1602,12 +1616,14 @@ static int r100_packet0_check(struct radeon_cs_parser *p, i = (reg - RADEON_PP_TEX_SIZE_0) / 8; track->textures[i].width = (idx_value & RADEON_TEX_USIZE_MASK) + 1; track->textures[i].height = ((idx_value & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1; + track->tex_dirty = true; break; case RADEON_PP_TEX_PITCH_0: case RADEON_PP_TEX_PITCH_1: case RADEON_PP_TEX_PITCH_2: i = (reg - RADEON_PP_TEX_PITCH_0) / 8; track->textures[i].pitch = idx_value + 32; + track->tex_dirty = true; break; case RADEON_PP_TXFILTER_0: case RADEON_PP_TXFILTER_1: @@ -1621,6 +1637,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, tmp = (idx_value >> 27) & 0x7; if (tmp == 2 || tmp == 6) track->textures[i].roundup_h = false; + track->tex_dirty = true; break; case RADEON_PP_TXFORMAT_0: case RADEON_PP_TXFORMAT_1: @@ -1673,6 +1690,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->textures[i].cube_info[4].width = 1 << ((idx_value >> 16) & 0xf); track->textures[i].cube_info[4].height = 1 << ((idx_value >> 20) & 0xf); + track->tex_dirty = true; break; case RADEON_PP_CUBIC_FACES_0: case RADEON_PP_CUBIC_FACES_1: @@ -1683,6 +1701,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf); track->textures[i].cube_info[face].height = 1 << ((tmp >> ((face * 8) + 4)) & 0xf); } + track->tex_dirty = true; break; default: printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", @@ -3318,9 +3337,9 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) unsigned long size; unsigned prim_walk; unsigned nverts; - unsigned num_cb = track->num_cb; + unsigned num_cb = track->cb_dirty ? track->num_cb : 0; - if (!track->zb_cb_clear && !track->color_channel_mask && + if (num_cb && !track->zb_cb_clear && !track->color_channel_mask && !track->blend_read_enable) num_cb = 0; @@ -3341,7 +3360,9 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) return -EINVAL; } } - if (track->z_enabled) { + track->cb_dirty = false; + + if (track->zb_dirty && track->z_enabled) { if (track->zb.robj == NULL) { DRM_ERROR("[drm] No buffer for z buffer !\n"); return -EINVAL; @@ -3358,6 +3379,28 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) return -EINVAL; } } + track->zb_dirty = false; + + if (track->aa_dirty && track->aaresolve) { + if (track->aa.robj == NULL) { + DRM_ERROR("[drm] No buffer for AA resolve buffer %d !\n", i); + return -EINVAL; + } + /* I believe the format comes from colorbuffer0. */ + size = track->aa.pitch * track->cb[0].cpp * track->maxy; + size += track->aa.offset; + if (size > radeon_bo_size(track->aa.robj)) { + DRM_ERROR("[drm] Buffer too small for AA resolve buffer %d " + "(need %lu have %lu) !\n", i, size, + radeon_bo_size(track->aa.robj)); + DRM_ERROR("[drm] AA resolve buffer %d (%u %u %u %u)\n", + i, track->aa.pitch, track->cb[0].cpp, + track->aa.offset, track->maxy); + return -EINVAL; + } + } + track->aa_dirty = false; + prim_walk = (track->vap_vf_cntl >> 4) & 0x3; if (track->vap_vf_cntl & (1 << 14)) { nverts = track->vap_alt_nverts; @@ -3417,13 +3460,23 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) prim_walk); return -EINVAL; } - return r100_cs_track_texture_check(rdev, track); + + if (track->tex_dirty) { + track->tex_dirty = false; + return r100_cs_track_texture_check(rdev, track); + } + return 0; } void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track) { unsigned i, face; + track->cb_dirty = true; + track->zb_dirty = true; + track->tex_dirty = true; + track->aa_dirty = true; + if (rdev->family < CHIP_R300) { track->num_cb = 1; if (rdev->family <= CHIP_RS200) @@ -3437,6 +3490,8 @@ void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track track->num_texture = 16; track->maxy = 4096; track->separate_cube = 0; + track->aaresolve = false; + track->aa.robj = NULL; } for (i = 0; i < track->num_cb; i++) { @@ -3746,8 +3801,6 @@ static int r100_startup(struct radeon_device *rdev) r100_mc_program(rdev); /* Resume clock */ r100_clock_startup(rdev); - /* Initialize GPU configuration (# pipes, ...) */ -// r100_gpu_init(rdev); /* Initialize GART (initialize after TTM so we can allocate * memory through TTM but finalize after TTM) */ r100_enable_bm(rdev); diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h index af65600e656..2fef9de7f36 100644 --- a/drivers/gpu/drm/radeon/r100_track.h +++ b/drivers/gpu/drm/radeon/r100_track.h @@ -52,14 +52,7 @@ struct r100_cs_track_texture { unsigned compress_format; }; -struct r100_cs_track_limits { - unsigned num_cb; - unsigned num_texture; - unsigned max_levels; -}; - struct r100_cs_track { - struct radeon_device *rdev; unsigned num_cb; unsigned num_texture; unsigned maxy; @@ -73,11 +66,17 @@ struct r100_cs_track { struct r100_cs_track_array arrays[11]; struct r100_cs_track_cb cb[R300_MAX_CB]; struct r100_cs_track_cb zb; + struct r100_cs_track_cb aa; struct r100_cs_track_texture textures[R300_TRACK_MAX_TEXTURE]; bool z_enabled; bool separate_cube; bool zb_cb_clear; bool blend_read_enable; + bool cb_dirty; + bool zb_dirty; + bool tex_dirty; + bool aa_dirty; + bool aaresolve; }; int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track); diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index d2408c39561..f2405830041 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -184,6 +184,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, } track->zb.robj = reloc->robj; track->zb.offset = idx_value; + track->zb_dirty = true; |