* Jerome Glisse
*/
#include <linux/seq_file.h>
+#include <linux/slab.h>
#include "drmP.h"
#include "drm.h"
#include "radeon_reg.h"
#include "rv350d.h"
#include "r300_reg_safe.h"
-/* This files gather functions specifics to: r300,r350,rv350,rv370,rv380 */
+/* This files gather functions specifics to: r300,r350,rv350,rv370,rv380
+ *
+ * GPU Errata:
+ * - HOST_PATH_CNTL: r300 family seems to dislike write to HOST_PATH_CNTL
+ * using MMIO to flush host path read cache, this lead to HARDLOCKUP.
+ * However, scheduling such write to the ring seems harmless, i suspect
+ * the CP read collide with the flush somehow, or maybe the MC, hard to
+ * tell. (Jerome Glisse)
+ */
/*
* rv370,rv380 PCIE GART
r = radeon_gart_table_vram_pin(rdev);
if (r)
return r;
+ radeon_gart_restore(rdev);
/* discard memory request outside of configured range */
tmp = RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp);
- WREG32_PCIE(RADEON_PCIE_TX_GART_START_LO, rdev->mc.gtt_location);
- tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - RADEON_GPU_PAGE_SIZE;
+ WREG32_PCIE(RADEON_PCIE_TX_GART_START_LO, rdev->mc.gtt_start);
+ tmp = rdev->mc.gtt_end & ~RADEON_GPU_PAGE_MASK;
WREG32_PCIE(RADEON_PCIE_TX_GART_END_LO, tmp);
WREG32_PCIE(RADEON_PCIE_TX_GART_START_HI, 0);
WREG32_PCIE(RADEON_PCIE_TX_GART_END_HI, 0);
table_addr = rdev->gart.table_addr;
WREG32_PCIE(RADEON_PCIE_TX_GART_BASE, table_addr);
/* FIXME: setup default page */
- WREG32_PCIE(RADEON_PCIE_TX_DISCARD_RD_ADDR_LO, rdev->mc.vram_location);
+ WREG32_PCIE(RADEON_PCIE_TX_DISCARD_RD_ADDR_LO, rdev->mc.vram_start);
WREG32_PCIE(RADEON_PCIE_TX_DISCARD_RD_ADDR_HI, 0);
/* Clear error */
WREG32_PCIE(0x18, 0);
void rv370_pcie_gart_disable(struct radeon_device *rdev)
{
- uint32_t tmp;
+ u32 tmp;
+ int r;
tmp = RREG32_PCIE(RADEON_PCIE_TX_GART_CNTL);
tmp |= RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp & ~RADEON_PCIE_TX_GART_EN);
if (rdev->gart.table.vram.robj) {
- radeon_object_kunmap(rdev->gart.table.vram.robj);
- radeon_object_unpin(rdev->gart.table.vram.robj);
+ r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+ if (likely(r == 0)) {
+ radeon_bo_kunmap(rdev->gart.table.vram.robj);
+ radeon_bo_unpin(rdev->gart.table.vram.robj);
+ radeon_bo_unreserve(rdev->gart.table.vram.robj);
+ }
}
}
/* Who ever call radeon_fence_emit should call ring_lock and ask
* for enough space (today caller are ib schedule and buffer move) */
/* Write SC register so SC & US assert idle */
- radeon_ring_write(rdev, PACKET0(0x43E0, 0));
+ radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_TL, 0));
radeon_ring_write(rdev, 0);
- radeon_ring_write(rdev, PACKET0(0x43E4, 0));
+ radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_BR, 0));
radeon_ring_write(rdev, 0);
/* Flush 3D cache */
- radeon_ring_write(rdev, PACKET0(0x4E4C, 0));
- radeon_ring_write(rdev, (2 << 0));
- radeon_ring_write(rdev, PACKET0(0x4F18, 0));
- radeon_ring_write(rdev, (1 << 0));
+ radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
+ radeon_ring_write(rdev, R300_RB3D_DC_FLUSH);
+ radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
+ radeon_ring_write(rdev, R300_ZC_FLUSH);
/* Wait until IDLE & CLEAN */
- radeon_ring_write(rdev, PACKET0(0x1720, 0));
- radeon_ring_write(rdev, (1 << 17) | (1 << 16) | (1 << 9));
+ radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+ radeon_ring_write(rdev, (RADEON_WAIT_3D_IDLECLEAN |
+ RADEON_WAIT_2D_IDLECLEAN |
+ RADEON_WAIT_DMA_GUI_IDLE));
+ radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+ radeon_ring_write(rdev, rdev->config.r300.hdp_cntl |
+ RADEON_HDP_READ_BUFFER_INVALIDATE);
+ radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+ radeon_ring_write(rdev, rdev->config.r300.hdp_cntl);
/* Emit fence sequence & fire IRQ */
radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0));
radeon_ring_write(rdev, fence->seq);
radeon_ring_write(rdev, RADEON_SW_INT_FIRE);
}
-int r300_copy_dma(struct radeon_device *rdev,
- uint64_t src_offset,
- uint64_t dst_offset,
- unsigned num_pages,
- struct radeon_fence *fence)
-{
- uint32_t size;
- uint32_t cur_size;
- int i, num_loops;
- int r = 0;
-
- /* radeon pitch is /64 */
- size = num_pages << PAGE_SHIFT;
- num_loops = DIV_ROUND_UP(size, 0x1FFFFF);
- r = radeon_ring_lock(rdev, num_loops * 4 + 64);
- if (r) {
- DRM_ERROR("radeon: moving bo (%d).\n", r);
- return r;
- }
- /* Must wait for 2D idle & clean before DMA or hangs might happen */
- radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0 ));
- radeon_ring_write(rdev, (1 << 16));
- for (i = 0; i < num_loops; i++) {
- cur_size = size;
- if (cur_size > 0x1FFFFF) {
- cur_size = 0x1FFFFF;
- }
- size -= cur_size;
- radeon_ring_write(rdev, PACKET0(0x720, 2));
- radeon_ring_write(rdev, src_offset);
- radeon_ring_write(rdev, dst_offset);
- radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30));
- src_offset += cur_size;
- dst_offset += cur_size;
- }
- radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
- radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE);
- if (fence) {
- r = radeon_fence_emit(rdev, fence);
- }
- radeon_ring_unlock_commit(rdev);
- return r;
-}
-
void r300_ring_start(struct radeon_device *rdev)
{
unsigned gb_tile_config;
radeon_ring_write(rdev,
RADEON_WAIT_2D_IDLECLEAN |
RADEON_WAIT_3D_IDLECLEAN);
- radeon_ring_write(rdev, PACKET0(0x170C, 0));
- radeon_ring_write(rdev, 1 << 31);
+ radeon_ring_write(rdev, PACKET0(R300_DST_PIPE_CONFIG, 0));
+ radeon_ring_write(rdev, R300_PIPE_AUTO_CONFIG);
radeon_ring_write(rdev, PACKET0(R300_GB_SELECT, 0));
radeon_ring_write(rdev, 0);
radeon_ring_write(rdev, PACKET0(R300_GB_ENABLE, 0));
for (i = 0; i < rdev->usec_timeout; i++) {
/* read MC_STATUS */
- tmp = RREG32(0x0150);
- if (tmp & (1 << 4)) {
+ tmp = RREG32(RADEON_MC_STATUS);
+ if (tmp & R300_MC_IDLE) {
return 0;
}
DRM_UDELAY(1);
"programming pipes. Bad things might happen.\n");
}
- tmp = RREG32(0x170C);
- WREG32(0x170C, tmp | (1 << 31));
+ tmp = RREG32(R300_DST_PIPE_CONFIG);
+ WREG32(R300_DST_PIPE_CONFIG, tmp | R300_PIPE_AUTO_CONFIG);
WREG32(R300_RB2D_DSTCACHE_MODE,
R300_DC_AUTOFLUSH_ENABLE |
/* GA still busy soft reset it */
WREG32(0x429C, 0x200);
WREG32(R300_VAP_PVS_STATE_FLUSH_REG, 0);
- WREG32(0x43E0, 0);
- WREG32(0x43E4, 0);
+ WREG32(R300_RE_SCISSORS_TL, 0);
+ WREG32(R300_RE_SCISSORS_BR, 0);
WREG32(0x24AC, 0);
}
/* Wait to prevent race in RBBM_STATUS */
}
/* Check if GPU is idle */
status = RREG32(RADEON_RBBM_STATUS);
- if (status & (1 << 31)) {
+ if (status & RADEON_RBBM_ACTIVE) {
DRM_ERROR("Failed to reset GPU (RBBM_STATUS=0x%08X)\n", status);
return -1;
}
/*
* r300,r350,rv350,rv380 VRAM info
*/
-void r300_vram_info(struct radeon_device *rdev)
+void r300_mc_init(struct radeon_device *rdev)
{
- uint32_t tmp;
+ u64 base;
+ u32 tmp;
/* DDR for all card after R300 & IGP */
rdev->mc.vram_is_ddr = true;
tmp = RREG32(RADEON_MEM_CNTL);
- if (tmp & R300_MEM_NUM_CHANNELS_MASK) {
- rdev->mc.vram_width = 128;
- } else {
- rdev->mc.vram_width = 64;
+ tmp &= R300_MEM_NUM_CHANNELS_MASK;
+ switch (tmp) {
+ case 0: rdev->mc.vram_width = 64; break;
+ case 1: rdev->mc.vram_width = 128; break;
+ case 2: rdev->mc.vram_width = 256; break;
+ default: rdev->mc.vram_width = 128; break;
}
-
r100_vram_init_sizes(rdev);
+ base = rdev->mc.aper_base;
+ if (rdev->flags & RADEON_IS_IGP)
+ base = (RREG32(RADEON_NB_TOM) & 0xffff) << 16;
+ radeon_vram_location(rdev, &rdev->mc, base);
+ if (!(rdev->flags & RADEON_IS_AGP))
+ radeon_gtt_location(rdev, &rdev->mc);
}
void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes)
}
+int rv370_get_pcie_lanes(struct radeon_device *rdev)
+{
+ u32 link_width_cntl;
+
+ if (rdev->flags & RADEON_IS_IGP)
+ return 0;
+
+ if (!(rdev->flags & RADEON_IS_PCIE))
+ return 0;
+
+ /* FIXME wait for idle */
+
+ if (rdev->family < CHIP_R600)
+ link_width_cntl = RREG32_PCIE(RADEON_PCIE_LC_LINK_WIDTH_CNTL);
+ else
+ link_width_cntl = RREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL);
+
+ switch ((link_width_cntl & RADEON_PCIE_LC_LINK_WIDTH_RD_MASK) >> RADEON_PCIE_LC_LINK_WIDTH_RD_SHIFT) {
+ case RADEON_PCIE_LC_LINK_WIDTH_X0:
+ return 0;
+ case RADEON_PCIE_LC_LINK_WIDTH_X1:
+ return 1;
+ case RADEON_PCIE_LC_LINK_WIDTH_X2:
+ return 2;
+ case RADEON_PCIE_LC_LINK_WIDTH_X4:
+ return 4;
+ case RADEON_PCIE_LC_LINK_WIDTH_X8:
+ return 8;
+ case RADEON_PCIE_LC_LINK_WIDTH_X16:
+ default:
+ return 16;
+ }
+}
+
#if defined(CONFIG_DEBUG_FS)
static int rv370_debugfs_pcie_gart_info(struct seq_file *m, void *data)
{
r100_cs_dump_packet(p, pkt);
return r;
}
- ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
+
+ if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+ tile_flags |= R300_TXO_MACRO_TILE;
+ if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+ tile_flags |= R300_TXO_MICRO_TILE;
+ else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
+ tile_flags |= R300_TXO_MICRO_TILE_SQUARE;
+
+ tmp = idx_value + ((u32)reloc->lobj.gpu_offset);
+ tmp |= tile_flags;
+ ib[idx] = tmp;
track->textures[i].robj = reloc->robj;
break;
/* Tracked registers */
tile_flags |= R300_COLOR_TILE_ENABLE;
if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
tile_flags |= R300_COLOR_MICROTILE_ENABLE;
+ else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
+ tile_flags |= R300_COLOR_MICROTILE_SQUARE_ENABLE;
tmp = idx_value & ~(0x7 << 16);
tmp |= tile_flags;
if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
tile_flags |= R300_DEPTHMACROTILE_ENABLE;
if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
- tile_flags |= R300_DEPTHMICROTILE_TILED;;
+ tile_flags |= R300_DEPTHMICROTILE_TILED;
+ else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
+ tile_flags |= R300_DEPTHMICROTILE_TILED_SQUARE;
tmp = idx_value & ~(0x7 << 16);
tmp |= tile_flags;
case R300_TX_FORMAT_Z6Y5X5:
case R300_TX_FORMAT_W4Z4Y4X4:
case R300_TX_FORMAT_W1Z5Y5X5:
- case R300_TX_FORMAT_DXT1:
case R300_TX_FORMAT_D3DMFT_CxV8U8:
case R300_TX_FORMAT_B8G8_B8G8:
case R300_TX_FORMAT_G8R8_G8B8:
case 0x17:
case R300_TX_FORMAT_FL_I32:
case 0x1e:
- case R300_TX_FORMAT_DXT3:
- case R300_TX_FORMAT_DXT5:
track->textures[i].cpp = 4;
break;
case R300_TX_FORMAT_W16Z16Y16X16:
case R300_TX_FORMAT_FL_R32G32B32A32:
track->textures[i].cpp = 16;
break;
+ case R300_TX_FORMAT_DXT1:
+ track->textures[i].cpp = 1;
+ track->textures[i].compress_format = R100_TRACK_COMP_DXT1;
+ break;
+ case R300_TX_FORMAT_ATI2N:
+ if (p->rdev->family < CHIP_R420) {
+ DRM_ERROR("Invalid texture format %u\n",
+ (idx_value & 0x1F));
+ return -EINVAL;
+ }
+ /* The same rules apply as for DXT3/5. */
+ /* Pass through. */
+ case R300_TX_FORMAT_DXT3:
+ case R300_TX_FORMAT_DXT5:
+ track->textures[i].cpp = 1;
+ track->textures[i].compress_format = R100_TRACK_COMP_DXT35;
+ break;
default:
DRM_ERROR("Invalid texture format %u\n",
(idx_value & 0x1F));
track->textures[i].width_11 = tmp;
tmp = ((idx_value >> 16) & 1) << 11;
track->textures[i].height_11 = tmp;
+
+ /* ATI1N */
+ if (idx_value & (1 << 14)) {
+ /* The same rules apply as for DXT1. */
+ track->textures[i].compress_format =
+ R100_TRACK_COMP_DXT1;
+ }
+ } else if (idx_value & (1 << 14)) {
+ DRM_ERROR("Forbidden bit TXFORMAT_MSB\n");
+ return -EINVAL;
}
break;
case 0x4480:
}
ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
break;
+ case 0x4e0c:
+ /* RB3D_COLOR_CHANNEL_MASK */
+ track->color_channel_mask = idx_value;
+ break;
+ case 0x4d1c:
+ /* ZB_BW_CNTL */
+ track->fastfill = !!(idx_value & (1 << 2));
+ break;
+ case 0x4e04:
+ /* RB3D_BLENDCNTL */
+ track->blend_read_enable = !!(idx_value & (1 << 2));
+ break;
case 0x4be8:
/* valid register only on RV530 */
if (p->rdev->family == CHIP_RV530)
{
int r;
+ /* set common regs */
+ r100_set_common_regs(rdev);
+ /* program mc */
r300_mc_program(rdev);
/* Resume clock */
r300_clock_startup(rdev);
return r;
}
/* Enable IRQ */
- rdev->irq.sw_int = true;
r100_irq_set(rdev);
+ rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
/* 1M ring buffer */
r = r100_cp_init(rdev, 1024 * 1024);
if (r) {
radeon_combios_asic_init(rdev->ddev);
/* Resume clock after posting */
r300_clock_startup(rdev);
+ /* Initialize surface registers */
+ radeon_surface_init(rdev);
return r300_startup(rdev);
}
void r300_fini(struct radeon_device *rdev)
{
- r300_suspend(rdev);
r100_cp_fini(rdev);
r100_wb_fini(rdev);
r100_ib_fini(rdev);
rv370_pcie_gart_fini(rdev);
if (rdev->flags & RADEON_IS_PCI)
r100_pci_gart_fini(rdev);
+ radeon_agp_fini(rdev);
radeon_irq_kms_fini(rdev);
radeon_fence_driver_fini(rdev);
- radeon_object_fini(rdev);
+ radeon_bo_fini(rdev);
radeon_atombios_fini(rdev);
kfree(rdev->bios);
rdev->bios = NULL;
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);
- }
+ if (radeon_boot_test_post_card(rdev) == false)
+ return -EINVAL;
/* Set asic errata */
r300_errata(rdev);
/* Initialize clocks */
radeon_get_clock_info(rdev->ddev);
- /* Get vram informations */
- r300_vram_info(rdev);
- /* Initialize memory controller (also test AGP) */
- r = r420_mc_init(rdev);
- if (r)
- return r;
+ /* Initialize power management */
+ radeon_pm_init(rdev);
+ /* initialize AGP */
+ if (rdev->flags & RADEON_IS_AGP) {
+ r = radeon_agp_init(rdev);
+ if (r) {
+ radeon_agp_disable(rdev);
+ }
+ }
+ /* initialize memory controller */
+ r300_mc_init(rdev);
/* Fence driver */
r = radeon_fence_driver_init(rdev);
if (r)
if (r)
return r;
/* Memory manager */
- r = radeon_object_init(rdev);
+ r = radeon_bo_init(rdev);
if (r)
return r;
if (rdev->flags & RADEON_IS_PCIE) {
if (r) {
/* Somethings want wront with the accel init stop accel */
dev_err(rdev->dev, "Disabling GPU acceleration\n");
- r300_suspend(rdev);
r100_cp_fini(rdev);
r100_wb_fini(rdev);
r100_ib_fini(rdev);
+ radeon_irq_kms_fini(rdev);
if (rdev->flags & RADEON_IS_PCIE)
rv370_pcie_gart_fini(rdev);
if (rdev->flags & RADEON_IS_PCI)
r100_pci_gart_fini(rdev);
- radeon_irq_kms_fini(rdev);
+ radeon_agp_fini(rdev);
rdev->accel_working = false;
}
return 0;
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff