diff options
Diffstat (limited to 'drivers/gpu/drm/radeon')
40 files changed, 3437 insertions, 1510 deletions
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index 5fae1e074b4b..013d38059943 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -13,7 +13,8 @@ radeon-$(CONFIG_DRM_RADEON_KMS) += radeon_device.o radeon_kms.o \ radeon_encoders.o radeon_display.o radeon_cursor.o radeon_i2c.o \ radeon_clocks.o radeon_fb.o radeon_gem.o radeon_ring.o radeon_irq_kms.o \ radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \ - rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o + rs400.o rs600.o rs690.o rv515.o r520.o r600.o rs780.o rv770.o \ + radeon_test.o radeon-$(CONFIG_COMPAT) += radeon_ioc32.o diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index c0080cc9bf8d..74d034f77c6b 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -31,6 +31,132 @@ #include "atom.h" #include "atom-bits.h" +static void atombios_overscan_setup(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + SET_CRTC_OVERSCAN_PS_ALLOCATION args; + int index = GetIndexIntoMasterTable(COMMAND, SetCRTC_OverScan); + int a1, a2; + + memset(&args, 0, sizeof(args)); + + args.usOverscanRight = 0; + args.usOverscanLeft = 0; + args.usOverscanBottom = 0; + args.usOverscanTop = 0; + args.ucCRTC = radeon_crtc->crtc_id; + + 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; + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + 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; + } 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; + } + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + break; + case RMX_FULL: + default: + args.usOverscanRight = 0; + args.usOverscanLeft = 0; + args.usOverscanBottom = 0; + args.usOverscanTop = 0; + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + break; + } +} + +static void atombios_scaler_setup(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + ENABLE_SCALER_PS_ALLOCATION args; + int index = GetIndexIntoMasterTable(COMMAND, EnableScaler); + /* fixme - fill in enc_priv for atom dac */ + enum radeon_tv_std tv_std = TV_STD_NTSC; + + if (!ASIC_IS_AVIVO(rdev) && radeon_crtc->crtc_id) + return; + + memset(&args, 0, sizeof(args)); + + args.ucScaler = radeon_crtc->crtc_id; + + if (radeon_crtc->devices & (ATOM_DEVICE_TV_SUPPORT)) { + switch (tv_std) { + case TV_STD_NTSC: + default: + args.ucTVStandard = ATOM_TV_NTSC; + break; + case TV_STD_PAL: + args.ucTVStandard = ATOM_TV_PAL; + break; + case TV_STD_PAL_M: + args.ucTVStandard = ATOM_TV_PALM; + break; + case TV_STD_PAL_60: + args.ucTVStandard = ATOM_TV_PAL60; + break; + case TV_STD_NTSC_J: + args.ucTVStandard = ATOM_TV_NTSCJ; + break; + case TV_STD_SCART_PAL: + args.ucTVStandard = ATOM_TV_PAL; /* ??? */ + break; + case TV_STD_SECAM: + args.ucTVStandard = ATOM_TV_SECAM; + break; + case TV_STD_PAL_CN: + args.ucTVStandard = ATOM_TV_PALCN; + break; + } + args.ucEnable = SCALER_ENABLE_MULTITAP_MODE; + } else if (radeon_crtc->devices & (ATOM_DEVICE_CV_SUPPORT)) { + args.ucTVStandard = ATOM_TV_CV; + args.ucEnable = SCALER_ENABLE_MULTITAP_MODE; + } else { + switch (radeon_crtc->rmx_type) { + case RMX_FULL: + args.ucEnable = ATOM_SCALER_EXPANSION; + break; + case RMX_CENTER: + args.ucEnable = ATOM_SCALER_CENTER; + break; + case RMX_ASPECT: + args.ucEnable = ATOM_SCALER_EXPANSION; + break; + default: + if (ASIC_IS_AVIVO(rdev)) + args.ucEnable = ATOM_SCALER_DISABLE; + else + args.ucEnable = ATOM_SCALER_CENTER; + break; + } + } + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + if (radeon_crtc->devices & (ATOM_DEVICE_CV_SUPPORT | ATOM_DEVICE_TV_SUPPORT) + && rdev->family >= CHIP_RV515 && rdev->family <= CHIP_RV570) { + atom_rv515_force_tv_scaler(rdev); + } +} + static void atombios_lock_crtc(struct drm_crtc *crtc, int lock) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); @@ -203,6 +329,12 @@ void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) if (ASIC_IS_AVIVO(rdev)) { uint32_t ss_cntl; + if ((rdev->family == CHIP_RS600) || + (rdev->family == CHIP_RS690) || + (rdev->family == CHIP_RS740)) + pll_flags |= (RADEON_PLL_USE_FRAC_FB_DIV | + RADEON_PLL_PREFER_CLOSEST_LOWER); + if (ASIC_IS_DCE32(rdev) && mode->clock > 200000) /* range limits??? */ pll_flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; else @@ -321,7 +453,7 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, struct drm_gem_object *obj; struct drm_radeon_gem_object *obj_priv; uint64_t fb_location; - uint32_t fb_format, fb_pitch_pixels; + uint32_t fb_format, fb_pitch_pixels, tiling_flags; if (!crtc->fb) return -EINVAL; @@ -358,7 +490,14 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, return -EINVAL; } - /* TODO tiling */ + radeon_object_get_tiling_flags(obj->driver_private, + &tiling_flags, NULL); + if (tiling_flags & RADEON_TILING_MACRO) + fb_format |= AVIVO_D1GRPH_MACRO_ADDRESS_MODE; + + if (tiling_flags & RADEON_TILING_MICRO) + fb_format |= AVIVO_D1GRPH_TILED; + if (radeon_crtc->crtc_id == 0) WREG32(AVIVO_D1VGA_CONTROL, 0); else @@ -509,6 +648,9 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc, radeon_crtc_set_base(crtc, x, y, old_fb); radeon_legacy_atom_set_surface(crtc); } + atombios_overscan_setup(crtc, mode, adjusted_mode); + atombios_scaler_setup(crtc); + radeon_bandwidth_update(rdev); return 0; } @@ -516,6 +658,8 @@ static bool atombios_crtc_mode_fixup(struct drm_crtc *crtc, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { + if (!radeon_crtc_scaling_mode_fixup(crtc, mode, adjusted_mode)) + return false; return true; } @@ -548,148 +692,3 @@ void radeon_atombios_init_crtc(struct drm_device *dev, AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL; drm_crtc_helper_add(&radeon_crtc->base, &atombios_helper_funcs); } - -void radeon_init_disp_bw_avivo(struct drm_device *dev, - struct drm_display_mode *mode1, - uint32_t pixel_bytes1, - struct drm_display_mode *mode2, - uint32_t pixel_bytes2) -{ - struct radeon_device *rdev = dev->dev_private; - fixed20_12 min_mem_eff; - fixed20_12 peak_disp_bw, mem_bw, pix_clk, pix_clk2, temp_ff; - fixed20_12 sclk_ff, mclk_ff; - uint32_t dc_lb_memory_split, temp; - - min_mem_eff.full = rfixed_const_8(0); - if (rdev->disp_priority == 2) { - uint32_t mc_init_misc_lat_timer = 0; - if (rdev->family == CHIP_RV515) - mc_init_misc_lat_timer = - RREG32_MC(RV515_MC_INIT_MISC_LAT_TIMER); - else if (rdev->family == CHIP_RS690) - mc_init_misc_lat_timer = - RREG32_MC(RS690_MC_INIT_MISC_LAT_TIMER); - - mc_init_misc_lat_timer &= - ~(R300_MC_DISP1R_INIT_LAT_MASK << - R300_MC_DISP1R_INIT_LAT_SHIFT); - mc_init_misc_lat_timer &= - ~(R300_MC_DISP0R_INIT_LAT_MASK << - R300_MC_DISP0R_INIT_LAT_SHIFT); - - if (mode2) - mc_init_misc_lat_timer |= - (1 << R300_MC_DISP1R_INIT_LAT_SHIFT); - if (mode1) - mc_init_misc_lat_timer |= - (1 << R300_MC_DISP0R_INIT_LAT_SHIFT); - - if (rdev->family == CHIP_RV515) - WREG32_MC(RV515_MC_INIT_MISC_LAT_TIMER, - mc_init_misc_lat_timer); - else if (rdev->family == CHIP_RS690) - WREG32_MC(RS690_MC_INIT_MISC_LAT_TIMER, - mc_init_misc_lat_timer); - } - - /* - * determine is there is enough bw for current mode - */ - temp_ff.full = rfixed_const(100); - mclk_ff.full = rfixed_const(rdev->clock.default_mclk); - mclk_ff.full = rfixed_div(mclk_ff, temp_ff); - sclk_ff.full = rfixed_const(rdev->clock.default_sclk); - sclk_ff.full = rfixed_div(sclk_ff, temp_ff); - - temp = (rdev->mc.vram_width / 8) * (rdev->mc.vram_is_ddr ? 2 : 1); - temp_ff.full = rfixed_const(temp); - mem_bw.full = rfixed_mul(mclk_ff, temp_ff); - mem_bw.full = rfixed_mul(mem_bw, min_mem_eff); - - pix_clk.full = 0; - pix_clk2.full = 0; - peak_disp_bw.full = 0; - if (mode1) { - temp_ff.full = rfixed_const(1000); - pix_clk.full = rfixed_const(mode1->clock); /* convert to fixed point */ - pix_clk.full = rfixed_div(pix_clk, temp_ff); - temp_ff.full = rfixed_const(pixel_bytes1); - peak_disp_bw.full += rfixed_mul(pix_clk, temp_ff); - } - if (mode2) { - temp_ff.full = rfixed_const(1000); - pix_clk2.full = rfixed_const(mode2->clock); /* convert to fixed point */ - pix_clk2.full = rfixed_div(pix_clk2, temp_ff); - temp_ff.full = rfixed_const(pixel_bytes2); - peak_disp_bw.full += rfixed_mul(pix_clk2, temp_ff); - } - - if (peak_disp_bw.full >= mem_bw.full) { - DRM_ERROR - ("You may not have enough display bandwidth for current mode\n" - "If you have flickering problem, try to lower resolution, refresh rate, or color depth\n"); - printk("peak disp bw %d, mem_bw %d\n", - rfixed_trunc(peak_disp_bw), rfixed_trunc(mem_bw)); - } - - /* - * Line Buffer Setup - * There is a single line buffer shared by both display controllers. - * DC_LB_MEMORY_SPLIT controls how that line buffer is shared between the display - * controllers. The paritioning can either be done manually or via one of four - * preset allocations specified in bits 1:0: - * 0 - line buffer is divided in half and shared between each display controller - * 1 - D1 gets 3/4 of the line buffer, D2 gets 1/4 - * 2 - D1 gets the whole buffer - * 3 - D1 gets 1/4 of the line buffer, D2 gets 3/4 - * Setting bit 2 of DC_LB_MEMORY_SPLIT controls switches to manual allocation mode. - * In manual allocation mode, D1 always starts at 0, D1 end/2 is specified in bits - * 14:4; D2 allocation follows D1. - */ - - /* is auto or manual better ? */ - dc_lb_memory_split = - RREG32(AVIVO_DC_LB_MEMORY_SPLIT) & ~AVIVO_DC_LB_MEMORY_SPLIT_MASK; - dc_lb_memory_split &= ~AVIVO_DC_LB_MEMORY_SPLIT_SHIFT_MODE; -#if 1 - /* auto */ - if (mode1 && mode2) { - if (mode1->hdisplay > mode2->hdisplay) { - if (mode1->hdisplay > 2560) - dc_lb_memory_split |= - AVIVO_DC_LB_MEMORY_SPLIT_D1_3Q_D2_1Q; - else - dc_lb_memory_split |= - AVIVO_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF; - } else if (mode2->hdisplay > mode1->hdisplay) { - if (mode2->hdisplay > 2560) - dc_lb_memory_split |= - AVIVO_DC_LB_MEMORY_SPLIT_D1_1Q_D2_3Q; - else - dc_lb_memory_split |= - AVIVO_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF; - } else - dc_lb_memory_split |= - AVIVO_DC_LB_MEMORY_SPLIT_D1HALF_D2HALF; - } else if (mode1) { - dc_lb_memory_split |= AVIVO_DC_LB_MEMORY_SPLIT_D1_ONLY; - } else if (mode2) { - dc_lb_memory_split |= AVIVO_DC_LB_MEMORY_SPLIT_D1_1Q_D2_3Q; - } -#else - /* manual */ - dc_lb_memory_split |= AVIVO_DC_LB_MEMORY_SPLIT_SHIFT_MODE; - dc_lb_memory_split &= - ~(AVIVO_DC_LB_DISP1_END_ADR_MASK << - AVIVO_DC_LB_DISP1_END_ADR_SHIFT); - if (mode1) { - dc_lb_memory_split |= - ((((mode1->hdisplay / 2) + 64) & AVIVO_DC_LB_DISP1_END_ADR_MASK) - << AVIVO_DC_LB_DISP1_END_ADR_SHIFT); - } else if (mode2) { - dc_lb_memory_split |= (0 << AVIVO_DC_LB_DISP1_END_ADR_SHIFT); - } -#endif - WREG32(AVIVO_DC_LB_MEMORY_SPLIT, dc_lb_memory_split); -} diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index c550932a108f..f1ba8ff41130 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -110,7 +110,7 @@ int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) if (i < 0 || i > rdev->gart.num_gpu_pages) { return -EINVAL; } - rdev->gart.table.ram.ptr[i] = cpu_to_le32((uint32_t)addr); + rdev->gart.table.ram.ptr[i] = cpu_to_le32(lower_32_bits(addr)); return 0; } @@ -173,8 +173,12 @@ void r100_mc_setup(struct radeon_device *rdev) DRM_ERROR("Failed to register debugfs file for R100 MC !\n"); } /* Write VRAM size in case we are limiting it */ - WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); - tmp = rdev->mc.vram_location + rdev->mc.vram_size - 1; + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); + /* Novell bug 204882 for RN50/M6/M7 with 8/16/32MB VRAM, + * if the aperture is 64MB but we have 32MB VRAM + * we report only 32MB VRAM but we have to set MC_FB_LOCATION + * to 64MB, otherwise the gpu accidentially dies */ + tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1; tmp = REG_SET(RADEON_MC_FB_TOP, tmp >> 16); tmp |= REG_SET(RADEON_MC_FB_START, rdev->mc.vram_location >> 16); WREG32(RADEON_MC_FB_LOCATION, tmp); @@ -215,7 +219,6 @@ int r100_mc_init(struct radeon_device *rdev) r100_pci_gart_disable(rdev); /* Setup GPU memory space */ - rdev->mc.vram_location = 0xFFFFFFFFUL; rdev->mc.gtt_location = 0xFFFFFFFFUL; if (rdev->flags & RADEON_IS_AGP) { r = radeon_agp_init(rdev); @@ -719,13 +722,14 @@ int r100_cs_packet_parse(struct radeon_cs_parser *p, unsigned idx) { struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx]; - uint32_t header = ib_chunk->kdata[idx]; + uint32_t header; if (idx >= ib_chunk->length_dw) { DRM_ERROR("Can not parse packet at %d after CS end %d !\n", idx, ib_chunk->length_dw); return -EINVAL; } + header = ib_chunk->kdata[idx]; pkt->idx = idx; pkt->type = CP_PACKET_GET_TYPE(header); pkt->count = CP_PACKET_GET_COUNT(header); @@ -753,6 +757,102 @@ int r100_cs_packet_parse(struct radeon_cs_parser *p, } /** + * r100_cs_packet_next_vline() - parse userspace VLINE packet + * @parser: parser structure holding parsing context. + * + * Userspace sends a special sequence for VLINE waits. + * PACKET0 - VLINE_START_END + value + * PACKET0 - WAIT_UNTIL +_value + * RELOC (P3) - crtc_id in reloc. + * + * This function parses this and relocates the VLINE START END + * and WAIT UNTIL packets to the correct crtc. + * It also detects a switched off crtc and nulls out the + * wait in that case. + */ +int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) +{ + struct radeon_cs_chunk *ib_chunk; + struct drm_mode_object *obj; + struct drm_crtc *crtc; + struct radeon_crtc *radeon_crtc; + struct radeon_cs_packet p3reloc, waitreloc; + int crtc_id; + int r; + uint32_t header, h_idx, reg; + + ib_chunk = &p->chunks[p->chunk_ib_idx]; + + /* parse the wait until */ + r = r100_cs_packet_parse(p, &waitreloc, p->idx); + if (r) + return r; + + /* check its a wait until and only 1 count */ + if (waitreloc.reg != RADEON_WAIT_UNTIL || + waitreloc.count != 0) { + DRM_ERROR("vline wait had illegal wait until segment\n"); + r = -EINVAL; + return r; + } + + if (ib_chunk->kdata[waitreloc.idx + 1] != RADEON_WAIT_CRTC_VLINE) { + DRM_ERROR("vline wait had illegal wait until\n"); + r = -EINVAL; + return r; + } + + /* jump over the NOP */ + r = r100_cs_packet_parse(p, &p3reloc, p->idx); + if (r) + return r; + + h_idx = p->idx - 2; + p->idx += waitreloc.count; + p->idx += p3reloc.count; + + header = ib_chunk->kdata[h_idx]; + crtc_id = ib_chunk->kdata[h_idx + 5]; + reg = ib_chunk->kdata[h_idx] >> 2; + mutex_lock(&p->rdev->ddev->mode_config.mutex); + obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC); + if (!obj) { + DRM_ERROR("cannot find crtc %d\n", crtc_id); + r = -EINVAL; + goto out; + } + crtc = obj_to_crtc(obj); + radeon_crtc = to_radeon_crtc(crtc); + crtc_id = radeon_crtc->crtc_id; + + if (!crtc->enabled) { + /* if the CRTC isn't enabled - we need to nop out the wait until */ + ib_chunk->kdata[h_idx + 2] = PACKET2(0); + ib_chunk->kdata[h_idx + 3] = PACKET2(0); + } else if (crtc_id == 1) { + switch (reg) { + case AVIVO_D1MODE_VLINE_START_END: + header &= R300_CP_PACKET0_REG_MASK; + header |= AVIVO_D2MODE_VLINE_START_END >> 2; + break; + case RADEON_CRTC_GUI_TRIG_VLINE: + header &= R300_CP_PACKET0_REG_MASK; + header |= RADEON_CRTC2_GUI_TRIG_VLINE >> 2; + break; + default: + DRM_ERROR("unknown crtc reloc\n"); + r = -EINVAL; + goto out; + } + ib_chunk->kdata[h_idx] = header; + ib_chunk->kdata[h_idx + 3] |= RADEON_ENG_DISPLAY_SELECT_CRTC1; + } +out: + mutex_unlock(&p->rdev->ddev->mode_config.mutex); + return r; +} + +/** * r100_cs_packet_next_reloc() - parse next packet which should be reloc packet3 * @parser: parser structure holding parsing context. * @data: pointer to relocation data @@ -814,6 +914,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, unsigned idx; bool onereg; int r; + u32 tile_flags = 0; ib = p->ib->ptr; ib_chunk = &p->chunks[p->chunk_ib_idx]; @@ -825,6 +926,15 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } for (i = 0; i <= pkt->count; i++, idx++, reg += 4) { switch (reg) { + case RADEON_CRTC_GUI_TRIG_VLINE: + r = r100_cs_packet_parse_vline(p); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + r100_cs_dump_packet(p, pkt); + return r; + } + break; /* FIXME: only allow PACKET3 blit? easier to check for out of * range access */ case RADEON_DST_PITCH_OFFSET: @@ -838,7 +948,20 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } tmp = ib_chunk->kdata[idx] & 0x003fffff; tmp += (((u32)reloc->lobj.gpu_offset) >> 10); - ib[idx] = (ib_chunk->kdata[idx] & 0xffc00000) | tmp; + + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_DST_TILE_MACRO; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + if (reg == RADEON_SRC_PITCH_OFFSET) { + DRM_ERROR("Cannot src blit from microtiled surface\n"); + r100_cs_dump_packet(p, pkt); + return -EINVAL; + } + tile_flags |= RADEON_DST_TILE_MICRO; + } + + tmp |= tile_flags; + ib[idx] = (ib_chunk->kdata[idx] & 0x3fc00000) | tmp; break; case RADEON_RB3D_DEPTHOFFSET: case RADEON_RB3D_COLOROFFSET: @@ -869,6 +992,11 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case R300_TX_OFFSET_0+52: case R300_TX_OFFSET_0+56: case R300_TX_OFFSET_0+60: + /* rn50 has no 3D engine so fail on any 3d setup */ + if (ASIC_IS_RN50(p->rdev)) { + DRM_ERROR("attempt to use RN50 3D engine failed\n"); + return -EINVAL; + } r = r100_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", @@ -878,6 +1006,25 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); break; + case R300_RB3D_COLORPITCH0: + case RADEON_RB3D_COLORPITCH: + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + r100_cs_dump_packet(p, pkt); + return r; + } + + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_COLOR_TILE_ENABLE; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; + + tmp = ib_chunk->kdata[idx] & ~(0x7 << 16); + tmp |= tile_flags; + ib[idx] = tmp; + break; default: /* FIXME: we don't want to allow anyothers packet */ break; @@ -1256,29 +1403,100 @@ static void r100_vram_get_type(struct radeon_device *rdev) } } -void r100_vram_info(struct radeon_device *rdev) +static u32 r100_get_accessible_vram(struct radeon_device *rdev) { - r100_vram_get_type(rdev); + u32 aper_size; + u8 byte; + + aper_size = RREG32(RADEON_CONFIG_APER_SIZE); + + /* Set HDP_APER_CNTL only on cards that are known not to be broken, + * that is has the 2nd generation multifunction PCI interface + */ + if (rdev->family == CHIP_RV280 || + rdev->family >= CHIP_RV350) { + WREG32_P(RADEON_HOST_PATH_CNTL, RADEON_HDP_APER_CNTL, + ~RADEON_HDP_APER_CNTL); + DRM_INFO("Generation 2 PCI interface, using max accessible memory\n"); + return aper_size * 2; + } + + /* Older cards have all sorts of funny issues to deal with. First + * check if it's a multifunction card by reading the PCI config + * header type... Limit those to one aperture size + */ + pci_read_config_byte(rdev->pdev, 0xe, &byte); + if (byte & 0x80) { + DRM_INFO("Generation 1 PCI interface in multifunction mode\n"); + DRM_INFO("Limiting VRAM to one aperture\n"); + return aper_size; + } + + /* Single function older card. We read HDP_APER_CNTL to see how the BIOS + * have set it up. We don't write this as it's broken on some ASICs but + * we expect the BIOS to have done the right thing (might be too optimistic...) + */ + if (RREG32(RADEON_HOST_PATH_CNTL) & RADEON_HDP_APER_CNTL) + return aper_size * 2; + return aper_size; +} + +void r100_vram_init_sizes(struct radeon_device *rdev) +{ + u64 config_aper_size; + u32 accessible; + + config_aper_size = RREG32(RADEON_CONFIG_APER_SIZE); if (rdev->flags & RADEON_IS_IGP) { uint32_t tom; /* read NB_TOM to get the amount of ram stolen for the GPU */ tom = RREG32(RADEON_NB_TOM); - rdev->mc.vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16); - WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); + rdev->mc.real_vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16); + /* for IGPs we need to keep VRAM where it was put by the BIOS */ + rdev->mc.vram_location = (tom & 0xffff) << 16; + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); + rdev->mc.mc_vram_size = rdev->mc.real_vram_size; } else { - rdev->mc.vram_size = RREG32(RADEON_CONFIG_MEMSIZE); + rdev->mc.real_vram_size = RREG32(RADEON_CONFIG_MEMSIZE); /* Some production boards of m6 will report 0 * if it's 8 MB */ - if (rdev->mc.vram_size == 0) { - rdev->mc.vram_size = 8192 * 1024; - WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); + if (rdev->mc.real_vram_size == 0) { + rdev->mc.real_vram_size = 8192 * 1024; + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); } + /* let driver place VRAM */ + rdev->mc.vram_location = 0xFFFFFFFFUL; + /* Fix for RN50, M6, M7 with 8/16/32(??) MBs of VRAM - + * Novell bug 204882 + along with lots of ubuntu ones */ + if (config_aper_size > rdev->mc.real_vram_size) + rdev->mc.mc_vram_size = config_aper_size; + else + rdev->mc.mc_vram_size = rdev->mc.real_vram_size; } + /* work out accessible VRAM */ + accessible = r100_get_accessible_vram(rdev); + rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); + + if (accessible > rdev->mc.aper_size) + accessible = rdev->mc.aper_size; + + if (rdev->mc.mc_vram_size > rdev->mc.aper_size) + rdev->mc.mc_vram_size = rdev->mc.aper_size; + + if (rdev->mc.real_vram_size > rdev->mc.aper_size) + rdev->mc.real_vram_size = rdev->mc.aper_size; +} + +void r100_vram_info(struct radeon_device *rdev) +{ + r100_vram_get_type(rdev); + + r100_vram_init_sizes(rdev); } @@ -1533,3 +1751,530 @@ int r100_debugfs_mc_info_init(struct radeon_device *rdev) return 0; #endif } + +int r100_set_surface_reg(struct radeon_device *rdev, int reg, + uint32_t tiling_flags, uint32_t pitch, + uint32_t offset, uint32_t obj_size) +{ + int surf_index = reg * 16; + int flags = 0; + + /* r100/r200 divide by 16 */ + if (rdev->family < CHIP_R300) + flags = pitch / 16; + else + flags = pitch / 8; + + if (rdev->family <= CHIP_RS200) { + if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) + == (RADEON_TILING_MACRO|RADEON_TILING_MICRO)) + flags |= RADEON_SURF_TILE_COLOR_BOTH; + if (tiling_flags & RADEON_TILING_MACRO) + flags |= RADEON_SURF_TILE_COLOR_MACRO; + } else if (rdev->family <= CHIP_RV280) { + if (tiling_flags & (RADEON_TILING_MACRO)) + flags |= R200_SURF_TILE_COLOR_MACRO; + if (tiling_flags & RADEON_TILING_MICRO) + flags |= R200_SURF_TILE_COLOR_MICRO; + } else { + if (tiling_flags & RADEON_TILING_MACRO) + flags |= R300_SURF_TILE_MACRO; + if (tiling_flags & RADEON_TILING_MICRO) + flags |= R300_SURF_TILE_MICRO; + } + + DRM_DEBUG("writing surface %d %d %x %x\n", reg, flags, offset, offset+obj_size-1); + WREG32(RADEON_SURFACE0_INFO + surf_index, flags); + WREG32(RADEON_SURFACE0_LOWER_BOUND + surf_index, offset); + WREG32(RADEON_SURFACE0_UPPER_BOUND + surf_index, offset + obj_size - 1); + return 0; +} + +void r100_clear_surface_reg(struct radeon_device *rdev, int reg) +{ + int surf_index = reg * 16; + WREG32(RADEON_SURFACE0_INFO + surf_index, 0); +} + +void r100_bandwidth_update(struct radeon_device *rdev) +{ + fixed20_12 trcd_ff, trp_ff, tras_ff, trbs_ff, tcas_ff; + fixed20_12 sclk_ff, mclk_ff, sclk_eff_ff, s |