diff options
Diffstat (limited to 'drivers/gpu/drm/radeon/r100.c')
| -rw-r--r-- | drivers/gpu/drm/radeon/r100.c | 627 |
1 files changed, 345 insertions, 282 deletions
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index be51c5f7d0f6..161094c07d94 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -32,6 +32,9 @@ #include "radeon_reg.h" #include "radeon.h" #include "r100d.h" +#include "rs100d.h" +#include "rv200d.h" +#include "rv250d.h" #include <linux/firmware.h> #include <linux/platform_device.h> @@ -60,18 +63,7 @@ MODULE_FIRMWARE(FIRMWARE_R520); /* This files gather functions specifics to: * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 - * - * Some of these functions might be used by newer ASICs. */ -int r200_init(struct radeon_device *rdev); -void r100_hdp_reset(struct radeon_device *rdev); -void r100_gpu_init(struct radeon_device *rdev); -int r100_gui_wait_for_idle(struct radeon_device *rdev); -int r100_mc_wait_for_idle(struct radeon_device *rdev); -void r100_gpu_wait_for_vsync(struct radeon_device *rdev); -void r100_gpu_wait_for_vsync2(struct radeon_device *rdev); -int r100_debugfs_mc_info_init(struct radeon_device *rdev); - /* * PCI GART @@ -152,136 +144,6 @@ void r100_pci_gart_fini(struct radeon_device *rdev) radeon_gart_fini(rdev); } - -/* - * MC - */ -void r100_mc_disable_clients(struct radeon_device *rdev) -{ - uint32_t ov0_scale_cntl, crtc_ext_cntl, crtc_gen_cntl, crtc2_gen_cntl; - - /* FIXME: is this function correct for rs100,rs200,rs300 ? */ - if (r100_gui_wait_for_idle(rdev)) { - printk(KERN_WARNING "Failed to wait GUI idle while " - "programming pipes. Bad things might happen.\n"); - } - - /* stop display and memory access */ - ov0_scale_cntl = RREG32(RADEON_OV0_SCALE_CNTL); - WREG32(RADEON_OV0_SCALE_CNTL, ov0_scale_cntl & ~RADEON_SCALER_ENABLE); - crtc_ext_cntl = RREG32(RADEON_CRTC_EXT_CNTL); - WREG32(RADEON_CRTC_EXT_CNTL, crtc_ext_cntl | RADEON_CRTC_DISPLAY_DIS); - crtc_gen_cntl = RREG32(RADEON_CRTC_GEN_CNTL); - - r100_gpu_wait_for_vsync(rdev); - - WREG32(RADEON_CRTC_GEN_CNTL, - (crtc_gen_cntl & ~(RADEON_CRTC_CUR_EN | RADEON_CRTC_ICON_EN)) | - RADEON_CRTC_DISP_REQ_EN_B | RADEON_CRTC_EXT_DISP_EN); - - if (!(rdev->flags & RADEON_SINGLE_CRTC)) { - crtc2_gen_cntl = RREG32(RADEON_CRTC2_GEN_CNTL); - - r100_gpu_wait_for_vsync2(rdev); - WREG32(RADEON_CRTC2_GEN_CNTL, - (crtc2_gen_cntl & - ~(RADEON_CRTC2_CUR_EN | RADEON_CRTC2_ICON_EN)) | - RADEON_CRTC2_DISP_REQ_EN_B); - } - - udelay(500); -} - -void r100_mc_setup(struct radeon_device *rdev) -{ - uint32_t tmp; - int r; - - r = r100_debugfs_mc_info_init(rdev); - if (r) { - 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.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); - - /* Enable bus mastering */ - tmp = RREG32(RADEON_BUS_CNTL) & ~RADEON_BUS_MASTER_DIS; - WREG32(RADEON_BUS_CNTL, tmp); - - if (rdev->flags & RADEON_IS_AGP) { - tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; - tmp = REG_SET(RADEON_MC_AGP_TOP, tmp >> 16); - tmp |= REG_SET(RADEON_MC_AGP_START, rdev->mc.gtt_location >> 16); - WREG32(RADEON_MC_AGP_LOCATION, tmp); - WREG32(RADEON_AGP_BASE, rdev->mc.agp_base); - } else { - WREG32(RADEON_MC_AGP_LOCATION, 0x0FFFFFFF); - WREG32(RADEON_AGP_BASE, 0); - } - - tmp = RREG32(RADEON_HOST_PATH_CNTL) & RADEON_HDP_APER_CNTL; - tmp |= (7 << 28); - WREG32(RADEON_HOST_PATH_CNTL, tmp | RADEON_HDP_SOFT_RESET | RADEON_HDP_READ_BUFFER_INVALIDATE); - (void)RREG32(RADEON_HOST_PATH_CNTL); - WREG32(RADEON_HOST_PATH_CNTL, tmp); - (void)RREG32(RADEON_HOST_PATH_CNTL); -} - -int r100_mc_init(struct radeon_device *rdev) -{ - int r; - - if (r100_debugfs_rbbm_init(rdev)) { - DRM_ERROR("Failed to register debugfs file for RBBM !\n"); - } - - r100_gpu_init(rdev); - /* Disable gart which also disable out of gart access */ - r100_pci_gart_disable(rdev); - - /* Setup GPU memory space */ - rdev->mc.gtt_location = 0xFFFFFFFFUL; - if (rdev->flags & RADEON_IS_AGP) { - r = radeon_agp_init(rdev); - if (r) { - printk(KERN_WARNING "[drm] Disabling AGP\n"); - rdev->flags &= ~RADEON_IS_AGP; - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; - } else { - rdev->mc.gtt_location = rdev->mc.agp_base; - } - } - r = radeon_mc_setup(rdev); - if (r) { - return r; - } - - r100_mc_disable_clients(rdev); - if (r100_mc_wait_for_idle(rdev)) { - printk(KERN_WARNING "Failed to wait MC idle while " - "programming pipes. Bad things might happen.\n"); - } - - r100_mc_setup(rdev); - return 0; -} - -void r100_mc_fini(struct radeon_device *rdev) -{ -} - - -/* - * Interrupts - */ int r100_irq_set(struct radeon_device *rdev) { uint32_t tmp = 0; @@ -358,10 +220,6 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc) return RREG32(RADEON_CRTC2_CRNT_FRAME); } - -/* - * Fence emission - */ void r100_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { @@ -377,10 +235,6 @@ void r100_fence_ring_emit(struct radeon_device *rdev, radeon_ring_write(rdev, RADEON_SW_INT_FIRE); } - -/* - * Writeback - */ int r100_wb_init(struct radeon_device *rdev) { int r; @@ -504,10 +358,6 @@ int r100_copy_blit(struct radeon_device *rdev, return r; } - -/* - * CP - */ static int r100_cp_wait_for_idle(struct radeon_device *rdev) { unsigned i; @@ -612,6 +462,7 @@ static int r100_cp_init_microcode(struct radeon_device *rdev) } return err; } + static void r100_cp_load_microcode(struct radeon_device *rdev) { const __be32 *fw_data; @@ -863,13 +714,11 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p, void r100_cs_dump_packet(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt) { - struct radeon_cs_chunk *ib_chunk; volatile uint32_t *ib; unsigned i; unsigned idx; ib = p->ib->ptr; - ib_chunk = &p->chunks[p->chunk_ib_idx]; idx = pkt->idx; for (i = 0; i <= (pkt->count + 1); i++, idx++) { DRM_INFO("ib[%d]=0x%08X\n", idx, ib[idx]); @@ -896,7 +745,7 @@ int r100_cs_packet_parse(struct radeon_cs_parser *p, idx, ib_chunk->length_dw); return -EINVAL; } - header = ib_chunk->kdata[idx]; + header = radeon_get_ib_value(p, idx); pkt->idx = idx; pkt->type = CP_PACKET_GET_TYPE(header); pkt->count = CP_PACKET_GET_COUNT(header); @@ -939,7 +788,6 @@ int r100_cs_packet_parse(struct radeon_cs_parser *p, */ 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; @@ -947,8 +795,9 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) int crtc_id; int r; uint32_t header, h_idx, reg; + volatile uint32_t *ib; - ib_chunk = &p->chunks[p->chunk_ib_idx]; + ib = p->ib->ptr; /* parse the wait until */ r = r100_cs_packet_parse(p, &waitreloc, p->idx); @@ -963,24 +812,24 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) return r; } - if (ib_chunk->kdata[waitreloc.idx + 1] != RADEON_WAIT_CRTC_VLINE) { + if (radeon_get_ib_value(p, 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); + r = r100_cs_packet_parse(p, &p3reloc, p->idx + waitreloc.count + 2); if (r) return r; h_idx = p->idx - 2; - p->idx += waitreloc.count; - p->idx += p3reloc.count; + p->idx += waitreloc.count + 2; + p->idx += p3reloc.count + 2; - header = ib_chunk->kdata[h_idx]; - crtc_id = ib_chunk->kdata[h_idx + 5]; - reg = ib_chunk->kdata[h_idx] >> 2; + header = radeon_get_ib_value(p, h_idx); + crtc_id = radeon_get_ib_value(p, h_idx + 5); + reg = CP_PACKET0_GET_REG(header); mutex_lock(&p->rdev->ddev->mode_config.mutex); obj = drm_mode_object_find(p->rdev->ddev, crtc_id, DRM_MODE_OBJECT_CRTC); if (!obj) { @@ -994,16 +843,16 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) 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); + ib[h_idx + 2] = PACKET2(0); + ib[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 &= ~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 &= ~R300_CP_PACKET0_REG_MASK; header |= RADEON_CRTC2_GUI_TRIG_VLINE >> 2; break; default: @@ -1011,8 +860,8 @@ int r100_cs_packet_parse_vline(struct radeon_cs_parser *p) r = -EINVAL; goto out; } - ib_chunk->kdata[h_idx] = header; - ib_chunk->kdata[h_idx + 3] |= RADEON_ENG_DISPLAY_SELECT_CRTC1; + ib[h_idx] = header; + ib[h_idx + 3] |= RADEON_ENG_DISPLAY_SELECT_CRTC1; } out: mutex_unlock(&p->rdev->ddev->mode_config.mutex); @@ -1033,7 +882,6 @@ out: int r100_cs_packet_next_reloc(struct radeon_cs_parser *p, struct radeon_cs_reloc **cs_reloc) { - struct radeon_cs_chunk *ib_chunk; struct radeon_cs_chunk *relocs_chunk; struct radeon_cs_packet p3reloc; unsigned idx; @@ -1044,7 +892,6 @@ int r100_cs_packet_next_reloc(struct radeon_cs_parser *p, return -EINVAL; } *cs_reloc = NULL; - ib_chunk = &p->chunks[p->chunk_ib_idx]; relocs_chunk = &p->chunks[p->chunk_relocs_idx]; r = r100_cs_packet_parse(p, &p3reloc, p->idx); if (r) { @@ -1057,7 +904,7 @@ int r100_cs_packet_next_reloc(struct radeon_cs_parser *p, r100_cs_dump_packet(p, &p3reloc); return -EINVAL; } - idx = ib_chunk->kdata[p3reloc.idx + 1]; + idx = radeon_get_ib_value(p, p3reloc.idx + 1); if (idx >= relocs_chunk->length_dw) { DRM_ERROR("Relocs at %d after relocations chunk end %d !\n", idx, relocs_chunk->length_dw); @@ -1126,7 +973,6 @@ static int r100_packet0_check(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt, unsigned idx, unsigned reg) { - struct radeon_cs_chunk *ib_chunk; struct radeon_cs_reloc *reloc; struct r100_cs_track *track; volatile uint32_t *ib; @@ -1134,11 +980,13 @@ static int r100_packet0_check(struct radeon_cs_parser *p, int r; int i, face; u32 tile_flags = 0; + u32 idx_value; ib = p->ib->ptr; - ib_chunk = &p->chunks[p->chunk_ib_idx]; track = (struct r100_cs_track *)p->track; + idx_value = radeon_get_ib_value(p, idx); + switch (reg) { case RADEON_CRTC_GUI_TRIG_VLINE: r = r100_cs_packet_parse_vline(p); @@ -1166,8 +1014,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->zb.robj = reloc->robj; - track->zb.offset = ib_chunk->kdata[idx]; - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + track->zb.offset = idx_value; + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_RB3D_COLOROFFSET: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1178,8 +1026,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return r; } track->cb[0].robj = reloc->robj; - track->cb[0].offset = ib_chunk->kdata[idx]; - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + track->cb[0].offset = idx_value; + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_PP_TXOFFSET_0: case RADEON_PP_TXOFFSET_1: @@ -1192,7 +1040,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].robj = reloc->robj; break; case RADEON_PP_CUBIC_OFFSET_T0_0: @@ -1208,8 +1056,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - track->textures[0].cube_info[i].offset = ib_chunk->kdata[idx]; - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + 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; break; case RADEON_PP_CUBIC_OFFSET_T1_0: @@ -1225,8 +1073,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - track->textures[1].cube_info[i].offset = ib_chunk->kdata[idx]; - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + 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; break; case RADEON_PP_CUBIC_OFFSET_T2_0: @@ -1242,12 +1090,12 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - track->textures[2].cube_info[i].offset = ib_chunk->kdata[idx]; - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + 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; break; case RADEON_RE_WIDTH_HEIGHT: - track->maxy = ((ib_chunk->kdata[idx] >> 16) & 0x7FF); + track->maxy = ((idx_value >> 16) & 0x7FF); break; case RADEON_RB3D_COLORPITCH: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1263,17 +1111,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p, if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; - tmp = ib_chunk->kdata[idx] & ~(0x7 << 16); + tmp = idx_value & ~(0x7 << 16); tmp |= tile_flags; ib[idx] = tmp; - track->cb[0].pitch = ib_chunk->kdata[idx] & RADEON_COLORPITCH_MASK; + track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; break; case RADEON_RB3D_DEPTHPITCH: - track->zb.pitch = ib_chunk->kdata[idx] & RADEON_DEPTHPITCH_MASK; + track->zb.pitch = idx_value & RADEON_DEPTHPITCH_MASK; break; case RADEON_RB3D_CNTL: - switch ((ib_chunk->kdata[idx] >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) { + switch ((idx_value >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) { case 7: case 8: case 9: @@ -1291,13 +1139,13 @@ static int r100_packet0_check(struct radeon_cs_parser *p, break; default: DRM_ERROR("Invalid color buffer format (%d) !\n", - ((ib_chunk->kdata[idx] >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f)); + ((idx_value >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f)); return -EINVAL; } - track->z_enabled = !!(ib_chunk->kdata[idx] & RADEON_Z_ENABLE); + track->z_enabled = !!(idx_value & RADEON_Z_ENABLE); break; case RADEON_RB3D_ZSTENCILCNTL: - switch (ib_chunk->kdata[idx] & 0xf) { + switch (idx_value & 0xf) { case 0: track->zb.cpp = 2; break; @@ -1321,44 +1169,44 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_PP_CNTL: { - uint32_t temp = ib_chunk->kdata[idx] >> 4; + uint32_t temp = idx_value >> 4; for (i = 0; i < track->num_texture; i++) track->textures[i].enabled = !!(temp & (1 << i)); } break; case RADEON_SE_VF_CNTL: - track->vap_vf_cntl = ib_chunk->kdata[idx]; + track->vap_vf_cntl = idx_value; break; case RADEON_SE_VTX_FMT: - track->vtx_size = r100_get_vtx_size(ib_chunk->kdata[idx]); + track->vtx_size = r100_get_vtx_size(idx_value); break; case RADEON_PP_TEX_SIZE_0: case RADEON_PP_TEX_SIZE_1: case RADEON_PP_TEX_SIZE_2: i = (reg - RADEON_PP_TEX_SIZE_0) / 8; - track->textures[i].width = (ib_chunk->kdata[idx] & RADEON_TEX_USIZE_MASK) + 1; - track->textures[i].height = ((ib_chunk->kdata[idx] & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1; + 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; 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 = ib_chunk->kdata[idx] + 32; + track->textures[i].pitch = idx_value + 32; break; case RADEON_PP_TXFILTER_0: case RADEON_PP_TXFILTER_1: case RADEON_PP_TXFILTER_2: i = (reg - RADEON_PP_TXFILTER_0) / 24; - track->textures[i].num_levels = ((ib_chunk->kdata[idx] & RADEON_MAX_MIP_LEVEL_MASK) + track->textures[i].num_levels = ((idx_value & RADEON_MAX_MIP_LEVEL_MASK) >> RADEON_MAX_MIP_LEVEL_SHIFT); - tmp = (ib_chunk->kdata[idx] >> 23) & 0x7; + tmp = (idx_value >> 23) & 0x7; if (tmp == 2 || tmp == 6) track->textures[i].roundup_w = false; - tmp = (ib_chunk->kdata[idx] >> 27) & 0x7; + tmp = (idx_value >> 27) & 0x7; if (tmp == 2 || tmp == 6) track->textures[i].roundup_h = false; break; @@ -1366,16 +1214,16 @@ static int r100_packet0_check(struct radeon_cs_parser *p, case RADEON_PP_TXFORMAT_1: case RADEON_PP_TXFORMAT_2: i = (reg - RADEON_PP_TXFORMAT_0) / 24; - if (ib_chunk->kdata[idx] & RADEON_TXFORMAT_NON_POWER2) { + if (idx_value & RADEON_TXFORMAT_NON_POWER2) { track->textures[i].use_pitch = 1; } else { track->textures[i].use_pitch = 0; - track->textures[i].width = 1 << ((ib_chunk->kdata[idx] >> RADEON_TXFORMAT_WIDTH_SHIFT) & RADEON_TXFORMAT_WIDTH_MASK); - track->textures[i].height = 1 << ((ib_chunk->kdata[idx] >> RADEON_TXFORMAT_HEIGHT_SHIFT) & RADEON_TXFORMAT_HEIGHT_MASK); + track->textures[i].width = 1 << ((idx_value >> RADEON_TXFORMAT_WIDTH_SHIFT) & RADEON_TXFORMAT_WIDTH_MASK); + track->textures[i].height = 1 << ((idx_value >> RADEON_TXFORMAT_HEIGHT_SHIFT) & RADEON_TXFORMAT_HEIGHT_MASK); } - if (ib_chunk->kdata[idx] & RADEON_TXFORMAT_CUBIC_MAP_ENABLE) + if (idx_value & RADEON_TXFORMAT_CUBIC_MAP_ENABLE) track->textures[i].tex_coord_type = 2; - switch ((ib_chunk->kdata[idx] & RADEON_TXFORMAT_FORMAT_MASK)) { + switch ((idx_value & RADEON_TXFORMAT_FORMAT_MASK)) { case RADEON_TXFORMAT_I8: case RADEON_TXFORMAT_RGB332: case RADEON_TXFORMAT_Y8: @@ -1402,13 +1250,13 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[i].cpp = 4; break; } - track->textures[i].cube_info[4].width = 1 << ((ib_chunk->kdata[idx] >> 16) & 0xf); - track->textures[i].cube_info[4].height = 1 << ((ib_chunk->kdata[idx] >> 20) & 0xf); + track->textures[i].cube_info[4].width = 1 << ((idx_value >> 16) & 0xf); + track->textures[i].cube_info[4].height = 1 << ((idx_value >> 20) & 0xf); break; case RADEON_PP_CUBIC_FACES_0: case RADEON_PP_CUBIC_FACES_1: case RADEON_PP_CUBIC_FACES_2: - tmp = ib_chunk->kdata[idx]; + tmp = idx_value; i = (reg - RADEON_PP_CUBIC_FACES_0) / 4; for (face = 0; face < 4; face++) { track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf); @@ -1427,15 +1275,14 @@ int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt, struct radeon_object *robj) { - struct radeon_cs_chunk *ib_chunk; unsigned idx; - - ib_chunk = &p->chunks[p->chunk_ib_idx]; + u32 value; idx = pkt->idx + 1; - if ((ib_chunk->kdata[idx+2] + 1) > radeon_object_size(robj)) { + value = radeon_get_ib_value(p, idx + 2); + if ((value + 1) > radeon_object_size(robj)) { DRM_ERROR("[drm] Buffer too small for PACKET3 INDX_BUFFER " "(need %u have %lu) !\n", - ib_chunk->kdata[idx+2] + 1, + value + 1, radeon_object_size(robj)); return -EINVAL; } @@ -1445,59 +1292,20 @@ int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p, static int r100_packet3_check(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt) { - struct radeon_cs_chunk *ib_chunk; struct radeon_cs_reloc *reloc; struct r100_cs_track *track; unsigned idx; - unsigned i, c; volatile uint32_t *ib; int r; ib = p->ib->ptr; - ib_chunk = &p->chunks[p->chunk_ib_idx]; idx = pkt->idx + 1; track = (struct r100_cs_track *)p->track; switch (pkt->opcode) { case PACKET3_3D_LOAD_VBPNTR: - c = ib_chunk->kdata[idx++]; - track->num_arrays = c; - for (i = 0; i < (c - 1); i += 2, idx += 3) { - r = r100_cs_packet_next_reloc(p, &reloc); - if (r) { - DRM_ERROR("No reloc for packet3 %d\n", - pkt->opcode); - r100_cs_dump_packet(p, pkt); - return r; - } - ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); - track->arrays[i + 0].robj = reloc->robj; - track->arrays[i + 0].esize = ib_chunk->kdata[idx] >> 8; - track->arrays[i + 0].esize &= 0x7F; - r = r100_cs_packet_next_reloc(p, &reloc); - if (r) { - DRM_ERROR("No reloc for packet3 %d\n", - pkt->opcode); - r100_cs_dump_packet(p, pkt); - return r; - } - ib[idx+2] = ib_chunk->kdata[idx+2] + ((u32)reloc->lobj.gpu_offset); - track->arrays[i + 1].robj = reloc->robj; - track->arrays[i + 1].esize = ib_chunk->kdata[idx] >> 24; - track->arrays[i + 1].esize &= 0x7F; - } - if (c & 1) { - r = r100_cs_packet_next_reloc(p, &reloc); - if (r) { - DRM_ERROR("No reloc for packet3 %d\n", - pkt->opcode); - r100_cs_dump_packet(p, pkt); - return r; - } - ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); - track->arrays[i + 0].robj = reloc->robj; - track->arrays[i + 0].esize = ib_chunk->kdata[idx] >> 8; - track->arrays[i + 0].esize &= 0x7F; - } + r = r100_packet3_load_vbpntr(p, pkt, idx); + if (r) + return r; break; case PACKET3_INDX_BUFFER: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1506,7 +1314,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + ib[idx+1] = radeon_get_ib_value(p, idx+1) + ((u32)reloc->lobj.gpu_offset); r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); if (r) { return r; @@ -1520,27 +1328,27 @@ static int r100_packet3_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + ib[idx] = radeon_get_ib_value(p, idx) + ((u32)reloc->lobj.gpu_offset); track->num_arrays = 1; - track->vtx_size = r100_get_vtx_size(ib_chunk->kdata[idx+2]); + track->vtx_size = r100_get_vtx_size(radeon_get_ib_value(p, idx + 2)); track->arrays[0].robj = reloc->robj; track->arrays[0].esize = track->vtx_size; - track->max_indx = ib_chunk->kdata[idx+1]; + track->max_indx = radeon_get_ib_value(p, idx+1); - track->vap_vf_cntl = ib_chunk->kdata[idx+3]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx+3); track->immd_dwords = pkt->count - 1; r = r100_cs_track_check(p->rdev, track); if (r) return r; break; case PACKET3_3D_DRAW_IMMD: - if (((ib_chunk->kdata[idx+1] >> 4) & 0x3) != 3) { + if (((radeon_get_ib_value(p, idx + 1) >> 4) & 0x3) != 3) { DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n"); return -EINVAL; } - track->vap_vf_cntl = ib_chunk->kdata[idx+1]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx + 1); track->immd_dwords = pkt->count - 1; r = r100_cs_track_check(p->rdev, track); if (r) @@ -1548,11 +1356,11 @@ static int r100_packet3_check(struct radeon_cs_parser *p, break; /* triggers drawing using in-packet vertex data */ case PACKET3_3D_DRAW_IMMD_2: - if (((ib_chunk->kdata[idx] >> 4) & 0x3) != 3) { + if (((radeon_get_ib_value(p, idx) >> 4) & 0x3) != 3) { DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n"); return -EINVAL; } - track->vap_vf_cntl = ib_chunk->kdata[idx]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx); track->immd_dwords = pkt->count; r = r100_cs_track_check(p->rdev, track); if (r) @@ -1560,28 +1368,28 @@ static int r100_packet3_check(struct radeon_cs_parser *p, break; /* triggers drawing using in-packet vertex data */ case PACKET3_3D_DRAW_VBUF_2: - track->vap_vf_cntl = ib_chunk->kdata[idx]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx); r = r100_cs_track_check(p->rdev, track); if (r) return r; break; /* triggers drawing of vertex buffers setup elsewhere */ case PACKET3_3D_DRAW_INDX_2: - track->vap_vf_cntl = ib_chunk->kdata[idx]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx); r = r100_cs_track_check(p->rdev, track); if (r) return r; break; /* triggers drawing using indices to vertex buffer */ case PACKET3_3D_DRAW_VBUF: - track->vap_vf_cntl = ib_chunk->kdata[idx + 1]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx + 1); r = r100_cs_track_check(p->rdev, track); if (r) return r; break; /* triggers drawing of vertex buffers setup elsewhere */ case PACKET3_3D_DRAW_INDX: - track->vap_vf_cntl = ib_chunk->kdata[idx + 1]; + track->vap_vf_cntl = radeon_get_ib_value(p, idx + 1); r = r100_cs_track_check(p->rdev, track); if (r) return r; @@ -2033,7 +1841,7 @@ void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) r100_pll_errata_after_data(rdev); } -int r100_init(struct radeon_device *rdev) +void r100_set_safe_registers(struct radeon_device *rdev) { if (ASIC_IS_RN50(rdev)) { rdev->config.r100.reg_safe_bm = rn50_reg_safe_bm; @@ -2042,9 +1850,8 @@ int r100_init(struct radeon_device *rdev) rdev->config.r100.reg_safe_bm = r100_reg_safe_bm; rdev->config.r100.reg_safe_bm_size = ARRAY_SIZE(r100_reg_safe_bm); } else { - return r200_init(rdev); + r200_set_safe_registers(rdev); } - return 0; } /* @@ -2342,9 +2149,11 @@ void r100_bandwidth_update(struct radeon_device *rdev) mode1 = &rdev->mode_info.crtcs[0]->base.mode; pixel_bytes1 = rdev->mode_info.crtcs[0]->base.fb->bits_per_pixel / 8; } - if (rdev->mode_info.crtcs[1]->base.enabled) { - mode2 = &rdev->mode_info.crtcs[1]->base.mode; - pixel_bytes2 = rdev->mode_info.crtcs[1]->base.fb->bits_per_pixel / 8; + if (!(rdev->flags & RADEON_SINGLE_CRTC)) { + if (rdev->mode_info.crtcs[1]->base.enabled) { + mode2 = &rdev->mode_info.crtcs[1]->base.mode; + pixel_bytes2 = rdev->mode_info.crtcs[1]->base.fb->bits_per_pixel / 8; + } } min_mem_eff.full = rfixed_const_8(0); @@ -3157,7 +2966,7 @@ void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save) WREG32(R_000740_CP_CSQ_CNTL, 0); /* Save few CRTC registers */ - save->GENMO_WT = RREG32(R_0003C0_GENMO_WT); + save->GENMO_WT = RREG8(R_0003C2_GENMO_WT); save->CRTC_EXT_CNTL = RREG32(R_000054_CRTC_EXT_CNTL); save->CRTC_GEN_CNTL = RREG32(R_000050_CRTC_GEN_CNTL); save->CUR_OFFSET = RREG32(R_000260_CUR_OFFSET); @@ -3167,7 +2976,7 @@ void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save) } /* Disable VGA aperture access */ - WREG32(R_0003C0_GENMO_WT, C_0003C0_VGA_RAM_EN & save->GENMO_WT); + WREG8(R_0003C2_GENMO_WT, C_0003C2_VGA_RAM_EN & save->GENMO_WT); /* Disable cursor, overlay, crtc */ WREG32(R_000260_CUR_OFFSET, save->CUR_OFFSET | S_000260_CUR_LOCK(1)); WREG32(R_000054_CRTC_EXT_CNTL, save->CRTC_EXT_CNTL | @@ -3199,10 +3008,264 @@ void r100_mc_resume(struct radeon_device *rdev, struct r100_mc_save *save) rdev->mc.vram_location); } /* Restore CRTC registers */ - WREG32(R_0003C0_GENMO_WT, save->GENMO_WT); + WREG8(R_0003C2_GENMO_WT, save->GENMO_WT); WREG32(R_000054_CRTC_EXT_CNTL, save->CRTC_EXT_CNTL); WREG32(R_000050_CRTC_GEN_CNTL, save->CRTC_GEN_CNTL); if (!(rdev->flags & RADEON_SINGLE_CRTC)) { WREG32(R_0003F8_CRTC2_GEN_CNTL, save->CRTC2_GEN_CNTL); } } + +void r100_vga_render_disable(struct radeon_device *rdev) +{ + u32 tmp; + + tmp = RREG8(R_0003C2_GENMO_WT); + WREG8(R_0003C2_GENMO_WT, C_0003C2_VGA_RAM_EN & tmp); +} + +static void r100_debugfs(struct radeon_device *rdev) +{ + int r; + + r = r100_debugfs_mc_info_init(rdev); + if (r) + dev_warn(rdev->dev, "Failed to create r100_mc debugfs file.\n"); +} + +static void r100_mc_program(struct radeon_device *rdev) +{ + struct r100_mc_save save; + + /* Stops all mc clients */ + r100_mc_stop(rdev, &save); + if (rdev->flags & RADEON_IS_AGP) { + WREG32(R_00014C_MC_AGP_LOCATION, + S_00014C_MC_AGP_START(rdev->mc.gtt_start >> 16) | + S_00014C_MC_AGP_TOP(rdev->mc.gtt_end >> 16)); + WREG32(R_000170_AGP_BASE, lower_32_bits(rdev->mc.agp_base)); + if (rdev->family > CHIP_RV200) + WREG32(R_00015C_AGP_BASE_2, + upper_32_bits(rdev->mc.agp_base) & 0xff); + } else { + WREG32(R_00014C_MC_AGP_LOCATION, 0x0FFFFFFF); + WREG32(R_000170_AGP_BASE, 0); + if (rdev->family > CHIP_RV200) + WREG32(R_00015C_AGP_BASE_2, 0); + } + /* Wait for mc idle */ + if (r100_mc_wait_for_idle(rdev)) + dev_warn(rdev->dev, "Wait for MC idle timeout.\n"); + /* Program MC, should be a 32bits limited address space */ + WREG32(R_000148_MC_FB_LOCATION, + S_000148_MC_FB_START(rdev->mc.vram_start >> 16) | + S_000148_MC_FB_TOP(rdev->mc.vram_end >> 16)); + r100_mc_resume(rdev, &save); +} + +void r100_clock_startup(struct radeon_device *rdev) +{ + u32 tmp; + + if (radeon_dynclks != -1 && radeon_dynclks) + radeon_legacy_set_clock_gating(rdev, 1); + /* We need to force on some of the block */ + tmp = RREG32_PLL(R_00000D_SCLK_CNTL); + tmp |= S_00000D_FORCE_CP(1) | S_00000D_FORCE_VIP(1); + if ((rdev->family == CHIP_RV250) || (rdev->family == CHIP_RV280)) + tmp |= S_00000D_FORCE_DISP1(1) | S_00000D_FORCE_DISP2(1); + WREG32_PLL(R_00000D_SCLK_CNTL, tmp); +} + +static int r100_startup(struct radeon_device *rdev) +{ + int r; + + 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) */ + if (rdev->flags & RADEON_IS_PCI) { + r = r100_pci_gart_enable(rdev); + if (r) + return r; + } + /* Enable IRQ */ + rdev->irq.sw_int = true; + r100_irq_set(rdev); + /* 1M ring buffer */ + r = r100_cp_init(rdev, 1024 * 1024); + if (r) { + dev_err(rdev->dev, "failled initializing CP (%d).\n", r); + return r; + } + r = r100_wb_init(rdev); + if (r) + dev_err(rdev->dev, "failled initializing WB (%d).\n", r); + r = r100_ib_init(rdev); + if (r) { + dev_err(rdev->dev, "failled initializing IB (%d).\n", r); + return r; + } + return 0; +} + +int r100_resume(struct radeon_device *rdev) +{ + /* Make sur GART are not working */ + if (rdev->flags & RADEON_IS_PCI) + r100_pci_gart_disable(rdev); + /* Resume clock before doing reset */ + r100_clock_startup(rdev); + /* Reset gpu before posting otherwise ATOM will enter infinite loop */ + if (radeon_gpu_reset(rdev)) { + dev_warn(rdev->dev, "GPU reset failed ! (0xE40=0x%08X, 0x7C0=0x%08X)\n", + RREG32(R_000E40_RBBM_STATUS), + RREG32(R_0007C0_CP_STAT)); + } + /* post */ + radeon_combios_asic_init(rdev->ddev); + /* Resume clock after posting */ + r100_clock_startup(rdev); + return r100_startup(rdev); +} + +int r100_suspend(struct radeon_device *rdev) +{ + r100_cp_disable(rdev); + r100_wb_disable(rdev); + r100_irq_disable(rdev); + if (rdev->flags & RADEON_IS_PCI) + r100_pci_gart_disable(rdev); + return 0; +} + +void r100_fini(struct radeon_device *rdev) +{ + r100_suspend(rdev); + r100_cp_fini(rdev); + r100_wb_fini(rdev); + r100_ib_fini(rdev); + radeon_gem_fini(rdev); + if (rdev->flags & RADEON_IS_PCI) + r100_pci_gart_fini(rdev); + radeon_irq_kms_fini(rdev); + radeon_fence_driver_fini(rdev); + radeon_object_fini(rdev); + radeon_atombios_fini(rdev); + kfree(rdev->bios); + rdev->bios = NULL; +} + +int r100_mc_init(struct radeon_device *rdev) +{ + int r; + u32 tmp; + + /* Setup GPU memory space */ + rdev->mc.vram_location = 0xFFFFFFFFUL; + rdev->mc.gtt_location = 0xFFFFFFFFUL; + if (rdev->flags & RADEON_IS_IGP) { + tmp = G_00015C_MC_FB_START(RREG32(R_00015C_NB_TOM)); + rdev->mc.vram_location = tmp << 16; + } + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + printk(KERN_WARNING "[drm] Disabling AGP\n"); + rdev->flags &= ~RADEON_IS_AGP; + rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; + } else { + rdev->mc.gtt_location = rdev->mc.agp_base; + } + } + r = radeon_mc_setup(rdev); + if (r) + return r; + return 0; +} + +int r100_init(struct radeon_device *rdev) +{ + int r; + + /* Register debugfs file specific to this group of asics */ + r100_debugfs(rdev); + /* Disable VGA */ + r100_vga_render_disable(rdev); + /* Initialize scratch registers */ + radeon_scratch_init(rdev); + /* Initialize surface registers */ + radeon_surface_init(rdev); + /* TODO: disable VGA need to use VGA request */ + /* BIOS*/ + if (!radeon_get_bios(rdev)) { + if (ASIC_IS_AVIVO(rdev)) + return -EINVAL; + } + if (rdev->is_atom_bios) { + dev_err(rdev->dev, "Expecting combios for RS400/RS480 GPU\n"); + return -EINVAL; + } else { + r = radeon_combios_init(rdev); + if (r) + return r; + } + /* Reset gpu before posting otherwise ATOM will enter infinite loop */ + if (radeon_gpu_reset(rdev)) { + dev_warn(rdev->dev, + "GPU reset failed ! (0xE40=0x%08X, 0x7C0=0x%08X)\n", + RREG32(R_000E40_RBBM_STATUS), + RREG32(R_0007C0_CP_STAT)); + } + /* check if cards are posted or not */ + if (!radeon_card_posted(rdev) && rdev->bios) { + DRM_INFO("GPU not posted. posting now...\n"); + radeon_combios_asic_init(rdev->ddev); + } + /* Set asic errata */ + r100_errata(rdev); + /* Initialize clocks */ + radeon_get_clock_info(rdev->ddev); + /* Get vram informations */ + r100_vram_info(rdev); + /* Initialize memory controller (also test AGP) */ + r = r100_mc_init(rdev); + if (r) + return r; + /* Fence driver */ + r = radeon_fence_driver_init(rdev); + if (r) + return r; + r = radeon_irq_kms_init(rdev); + if (r) + return r; + /* Memory manager */ + r = radeon_object_init(rdev); + if (r) + return r; + if (rdev->flags & RADEON_IS_PCI) { + r = r100_pci_gart_init(rdev); + if (r) + return r; + } + r100_set_safe_registers(rdev); + rdev->accel_working = true; + r = r100_startup(rdev); + if (r) { + /* Somethings want wront with the accel init stop accel */ + dev_err(rdev->dev, "Disabling GPU acceleration\n"); + r100_suspend(rdev); + r100_cp_fini(rdev); + r100_wb_fini(rdev); + r100_ib_fini(rdev); + if (rdev->flags & RADEON_IS_PCI) + r100_pci_gart_fini(rdev); + radeon_irq_kms_fini(rdev); + rdev->accel_working = false; + } + return 0; +} |