aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/radeon/si.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/radeon/si.c')
-rw-r--r--drivers/gpu/drm/radeon/si.c1236
1 files changed, 653 insertions, 583 deletions
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index 6ca904673a4..9e854fd016d 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -39,35 +39,43 @@ MODULE_FIRMWARE("radeon/TAHITI_pfp.bin");
MODULE_FIRMWARE("radeon/TAHITI_me.bin");
MODULE_FIRMWARE("radeon/TAHITI_ce.bin");
MODULE_FIRMWARE("radeon/TAHITI_mc.bin");
+MODULE_FIRMWARE("radeon/TAHITI_mc2.bin");
MODULE_FIRMWARE("radeon/TAHITI_rlc.bin");
MODULE_FIRMWARE("radeon/TAHITI_smc.bin");
MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin");
MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_mc2.bin");
MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin");
MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin");
MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
MODULE_FIRMWARE("radeon/VERDE_me.bin");
MODULE_FIRMWARE("radeon/VERDE_ce.bin");
MODULE_FIRMWARE("radeon/VERDE_mc.bin");
+MODULE_FIRMWARE("radeon/VERDE_mc2.bin");
MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
MODULE_FIRMWARE("radeon/VERDE_smc.bin");
MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
MODULE_FIRMWARE("radeon/OLAND_me.bin");
MODULE_FIRMWARE("radeon/OLAND_ce.bin");
MODULE_FIRMWARE("radeon/OLAND_mc.bin");
+MODULE_FIRMWARE("radeon/OLAND_mc2.bin");
MODULE_FIRMWARE("radeon/OLAND_rlc.bin");
MODULE_FIRMWARE("radeon/OLAND_smc.bin");
MODULE_FIRMWARE("radeon/HAINAN_pfp.bin");
MODULE_FIRMWARE("radeon/HAINAN_me.bin");
MODULE_FIRMWARE("radeon/HAINAN_ce.bin");
MODULE_FIRMWARE("radeon/HAINAN_mc.bin");
+MODULE_FIRMWARE("radeon/HAINAN_mc2.bin");
MODULE_FIRMWARE("radeon/HAINAN_rlc.bin");
MODULE_FIRMWARE("radeon/HAINAN_smc.bin");
+static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh);
static void si_pcie_gen3_enable(struct radeon_device *rdev);
static void si_program_aspm(struct radeon_device *rdev);
+extern void sumo_rlc_fini(struct radeon_device *rdev);
+extern int sumo_rlc_init(struct radeon_device *rdev);
extern int r600_ih_ring_alloc(struct radeon_device *rdev);
extern void r600_ih_ring_fini(struct radeon_device *rdev);
extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
@@ -76,6 +84,13 @@ extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_
extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
extern bool evergreen_is_display_hung(struct radeon_device *rdev);
+static void si_enable_gui_idle_interrupt(struct radeon_device *rdev,
+ bool enable);
+static void si_init_pg(struct radeon_device *rdev);
+static void si_init_cg(struct radeon_device *rdev);
+static void si_fini_pg(struct radeon_device *rdev);
+static void si_fini_cg(struct radeon_device *rdev);
+static void si_rlc_stop(struct radeon_device *rdev);
static const u32 verde_rlc_save_restore_register_list[] =
{
@@ -1453,41 +1468,38 @@ static const u32 hainan_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = {
};
/* ucode loading */
-static int si_mc_load_microcode(struct radeon_device *rdev)
+int si_mc_load_microcode(struct radeon_device *rdev)
{
const __be32 *fw_data;
u32 running, blackout = 0;
u32 *io_mc_regs;
- int i, ucode_size, regs_size;
+ int i, regs_size, ucode_size;
if (!rdev->mc_fw)
return -EINVAL;
+ ucode_size = rdev->mc_fw->size / 4;
+
switch (rdev->family) {
case CHIP_TAHITI:
io_mc_regs = (u32 *)&tahiti_io_mc_regs;
- ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_PITCAIRN:
io_mc_regs = (u32 *)&pitcairn_io_mc_regs;
- ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_VERDE:
default:
io_mc_regs = (u32 *)&verde_io_mc_regs;
- ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_OLAND:
io_mc_regs = (u32 *)&oland_io_mc_regs;
- ucode_size = OLAND_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_HAINAN:
io_mc_regs = (u32 *)&hainan_io_mc_regs;
- ucode_size = OLAND_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
}
@@ -1543,7 +1555,7 @@ static int si_init_microcode(struct radeon_device *rdev)
const char *chip_name;
const char *rlc_chip_name;
size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size;
- size_t smc_req_size;
+ size_t smc_req_size, mc2_req_size;
char fw_name[30];
int err;
@@ -1558,6 +1570,7 @@ static int si_init_microcode(struct radeon_device *rdev)
ce_req_size = SI_CE_UCODE_SIZE * 4;
rlc_req_size = SI_RLC_UCODE_SIZE * 4;
mc_req_size = SI_MC_UCODE_SIZE * 4;
+ mc2_req_size = TAHITI_MC_UCODE_SIZE * 4;
smc_req_size = ALIGN(TAHITI_SMC_UCODE_SIZE, 4);
break;
case CHIP_PITCAIRN:
@@ -1568,6 +1581,7 @@ static int si_init_microcode(struct radeon_device *rdev)
ce_req_size = SI_CE_UCODE_SIZE * 4;
rlc_req_size = SI_RLC_UCODE_SIZE * 4;
mc_req_size = SI_MC_UCODE_SIZE * 4;
+ mc2_req_size = PITCAIRN_MC_UCODE_SIZE * 4;
smc_req_size = ALIGN(PITCAIRN_SMC_UCODE_SIZE, 4);
break;
case CHIP_VERDE:
@@ -1578,6 +1592,7 @@ static int si_init_microcode(struct radeon_device *rdev)
ce_req_size = SI_CE_UCODE_SIZE * 4;
rlc_req_size = SI_RLC_UCODE_SIZE * 4;
mc_req_size = SI_MC_UCODE_SIZE * 4;
+ mc2_req_size = VERDE_MC_UCODE_SIZE * 4;
smc_req_size = ALIGN(VERDE_SMC_UCODE_SIZE, 4);
break;
case CHIP_OLAND:
@@ -1587,7 +1602,7 @@ static int si_init_microcode(struct radeon_device *rdev)
me_req_size = SI_PM4_UCODE_SIZE * 4;
ce_req_size = SI_CE_UCODE_SIZE * 4;
rlc_req_size = SI_RLC_UCODE_SIZE * 4;
- mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+ mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4;
smc_req_size = ALIGN(OLAND_SMC_UCODE_SIZE, 4);
break;
case CHIP_HAINAN:
@@ -1597,7 +1612,7 @@ static int si_init_microcode(struct radeon_device *rdev)
me_req_size = SI_PM4_UCODE_SIZE * 4;
ce_req_size = SI_CE_UCODE_SIZE * 4;
rlc_req_size = SI_RLC_UCODE_SIZE * 4;
- mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+ mc_req_size = mc2_req_size = OLAND_MC_UCODE_SIZE * 4;
smc_req_size = ALIGN(HAINAN_SMC_UCODE_SIZE, 4);
break;
default: BUG();
@@ -1650,22 +1665,33 @@ static int si_init_microcode(struct radeon_device *rdev)
err = -EINVAL;
}
- snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+ snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc2.bin", chip_name);
err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
- if (err)
- goto out;
- if (rdev->mc_fw->size != mc_req_size) {
+ if (err) {
+ snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+ err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
+ if (err)
+ goto out;
+ }
+ if ((rdev->mc_fw->size != mc_req_size) &&
+ (rdev->mc_fw->size != mc2_req_size)) {
printk(KERN_ERR
"si_mc: Bogus length %zu in firmware \"%s\"\n",
rdev->mc_fw->size, fw_name);
err = -EINVAL;
}
+ DRM_INFO("%s: %zu bytes\n", fw_name, rdev->mc_fw->size);
snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
- if (err)
- goto out;
- if (rdev->smc_fw->size != smc_req_size) {
+ if (err) {
+ printk(KERN_ERR
+ "smc: error loading firmware \"%s\"\n",
+ fw_name);
+ release_firmware(rdev->smc_fw);
+ rdev->smc_fw = NULL;
+ err = 0;
+ } else if (rdev->smc_fw->size != smc_req_size) {
printk(KERN_ERR
"si_smc: Bogus length %zu in firmware \"%s\"\n",
rdev->smc_fw->size, fw_name);
@@ -1700,7 +1726,8 @@ static u32 dce6_line_buffer_adjust(struct radeon_device *rdev,
struct drm_display_mode *mode,
struct drm_display_mode *other_mode)
{
- u32 tmp;
+ u32 tmp, buffer_alloc, i;
+ u32 pipe_offset = radeon_crtc->crtc_id * 0x20;
/*
* Line Buffer Setup
* There are 3 line buffers, each one shared by 2 display controllers.
@@ -1715,16 +1742,30 @@ static u32 dce6_line_buffer_adjust(struct radeon_device *rdev,
* non-linked crtcs for maximum line buffer allocation.
*/
if (radeon_crtc->base.enabled && mode) {
- if (other_mode)
+ if (other_mode) {
tmp = 0; /* 1/2 */
- else
+ buffer_alloc = 1;
+ } else {
tmp = 2; /* whole */
- } else
+ buffer_alloc = 2;
+ }
+ } else {
tmp = 0;
+ buffer_alloc = 0;
+ }
WREG32(DC_LB_MEMORY_SPLIT + radeon_crtc->crtc_offset,
DC_LB_MEMORY_CONFIG(tmp));
+ WREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset,
+ DMIF_BUFFERS_ALLOCATED(buffer_alloc));
+ for (i = 0; i < rdev->usec_timeout; i++) {
+ if (RREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset) &
+ DMIF_BUFFERS_ALLOCATED_COMPLETED)
+ break;
+ udelay(1);
+ }
+
if (radeon_crtc->base.enabled && mode) {
switch (tmp) {
case 0:
@@ -2784,7 +2825,7 @@ static void si_setup_spi(struct radeon_device *rdev,
}
static u32 si_get_rb_disabled(struct radeon_device *rdev,
- u32 max_rb_num, u32 se_num,
+ u32 max_rb_num_per_se,
u32 sh_per_se)
{
u32 data, mask;
@@ -2798,14 +2839,14 @@ static u32 si_get_rb_disabled(struct radeon_device *rdev,
data >>= BACKEND_DISABLE_SHIFT;
- mask = si_create_bitmask(max_rb_num / se_num / sh_per_se);
+ mask = si_create_bitmask(max_rb_num_per_se / sh_per_se);
return data & mask;
}
static void si_setup_rb(struct radeon_device *rdev,
u32 se_num, u32 sh_per_se,
- u32 max_rb_num)
+ u32 max_rb_num_per_se)
{
int i, j;
u32 data, mask;
@@ -2815,19 +2856,21 @@ static void si_setup_rb(struct radeon_device *rdev,
for (i = 0; i < se_num; i++) {
for (j = 0; j < sh_per_se; j++) {
si_select_se_sh(rdev, i, j);
- data = si_get_rb_disabled(rdev, max_rb_num, se_num, sh_per_se);
+ data = si_get_rb_disabled(rdev, max_rb_num_per_se, sh_per_se);
disabled_rbs |= data << ((i * sh_per_se + j) * TAHITI_RB_BITMAP_WIDTH_PER_SH);
}
}
si_select_se_sh(rdev, 0xffffffff, 0xffffffff);
mask = 1;
- for (i = 0; i < max_rb_num; i++) {
+ for (i = 0; i < max_rb_num_per_se * se_num; i++) {
if (!(disabled_rbs & mask))
enabled_rbs |= mask;
mask <<= 1;
}
+ rdev->config.si.backend_enable_mask = enabled_rbs;
+
for (i = 0; i < se_num; i++) {
si_select_se_sh(rdev, i, 0xffffffff);
data = 0;
@@ -2858,7 +2901,7 @@ static void si_gpu_init(struct radeon_device *rdev)
u32 sx_debug_1;
u32 hdp_host_path_cntl;
u32 tmp;
- int i, j;
+ int i, j, k;
switch (rdev->family) {
case CHIP_TAHITI:
@@ -3056,6 +3099,14 @@ static void si_gpu_init(struct radeon_device *rdev)
rdev->config.si.max_sh_per_se,
rdev->config.si.max_cu_per_sh);
+ for (i = 0; i < rdev->config.si.max_shader_engines; i++) {
+ for (j = 0; j < rdev->config.si.max_sh_per_se; j++) {
+ for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) {
+ rdev->config.si.active_cus +=
+ hweight32(si_get_cu_active_bitmap(rdev, i, j));
+ }
+ }
+ }
/* set HW defaults for 3D engine */
WREG32(CP_QUEUE_THRESHOLDS, (ROQ_IB1_START(0x16) |
@@ -3144,7 +3195,7 @@ void si_fence_ring_emit(struct radeon_device *rdev,
/* EVENT_WRITE_EOP - flush caches, send int */
radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5));
- radeon_ring_write(ring, addr & 0xffffffff);
+ radeon_ring_write(ring, lower_32_bits(addr));
radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2));
radeon_ring_write(ring, fence->seq);
radeon_ring_write(ring, 0);
@@ -3177,7 +3228,7 @@ void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
radeon_ring_write(ring, (1 << 8));
radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
- radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
+ radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr));
radeon_ring_write(ring, next_rptr);
}
@@ -3218,7 +3269,8 @@ static void si_cp_enable(struct radeon_device *rdev, bool enable)
if (enable)
WREG32(CP_ME_CNTL, 0);
else {
- radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+ if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX)
+ radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT));
WREG32(SCRATCH_UMSK, 0);
rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
@@ -3360,16 +3412,7 @@ static int si_cp_resume(struct radeon_device *rdev)
u32 rb_bufsz;
int r;
- /* Reset cp; if cp is reset, then PA, SH, VGT also need to be reset */
- WREG32(GRBM_SOFT_RESET, (SOFT_RESET_CP |
- SOFT_RESET_PA |
- SOFT_RESET_VGT |
- SOFT_RESET_SPI |
- SOFT_RESET_SX));
- RREG32(GRBM_SOFT_RESET);
- mdelay(15);
- WREG32(GRBM_SOFT_RESET, 0);
- RREG32(GRBM_SOFT_RESET);
+ si_enable_gui_idle_interrupt(rdev, false);
WREG32(CP_SEM_WAIT_TIMER, 0x0);
WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
@@ -3383,8 +3426,8 @@ static int si_cp_resume(struct radeon_device *rdev)
/* ring 0 - compute and gfx */
/* Set ring buffer size */
ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
- rb_bufsz = drm_order(ring->ring_size / 8);
- tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+ rb_bufsz = order_base_2(ring->ring_size / 8);
+ tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
#ifdef __BIG_ENDIAN
tmp |= BUF_SWAP_32BIT;
#endif
@@ -3411,13 +3454,11 @@ static int si_cp_resume(struct radeon_device *rdev)
WREG32(CP_RB0_BASE, ring->gpu_addr >> 8);
- ring->rptr = RREG32(CP_RB0_RPTR);
-
/* ring1 - compute only */
/* Set ring buffer size */
ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
- rb_bufsz = drm_order(ring->ring_size / 8);
- tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+ rb_bufsz = order_base_2(ring->ring_size / 8);
+ tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
#ifdef __BIG_ENDIAN
tmp |= BUF_SWAP_32BIT;
#endif
@@ -3437,13 +3478,11 @@ static int si_cp_resume(struct radeon_device *rdev)
WREG32(CP_RB1_BASE, ring->gpu_addr >> 8);
- ring->rptr = RREG32(CP_RB1_RPTR);
-
/* ring2 - compute only */
/* Set ring buffer size */
ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
- rb_bufsz = drm_order(ring->ring_size / 8);
- tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+ rb_bufsz = order_base_2(ring->ring_size / 8);
+ tmp = (order_base_2(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
#ifdef __BIG_ENDIAN
tmp |= BUF_SWAP_32BIT;
#endif
@@ -3463,8 +3502,6 @@ static int si_cp_resume(struct radeon_device *rdev)
WREG32(CP_RB2_BASE, ring->gpu_addr >> 8);
- ring->rptr = RREG32(CP_RB2_RPTR);
-
/* start the rings */
si_cp_start(rdev);
rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true;
@@ -3486,10 +3523,15 @@ static int si_cp_resume(struct radeon_device *rdev)
rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
}
+ si_enable_gui_idle_interrupt(rdev, true);
+
+ if (rdev->asic->copy.copy_ring_index == RADEON_RING_TYPE_GFX_INDEX)
+ radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
+
return 0;
}
-static u32 si_gpu_check_soft_reset(struct radeon_device *rdev)
+u32 si_gpu_check_soft_reset(struct radeon_device *rdev)
{
u32 reset_mask = 0;
u32 tmp;
@@ -3587,6 +3629,13 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
dev_info(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+ /* disable PG/CG */
+ si_fini_pg(rdev);
+ si_fini_cg(rdev);
+
+ /* stop the rlc */
+ si_rlc_stop(rdev);
+
/* Disable CP parsing/prefetching */
WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
@@ -3695,6 +3744,106 @@ static void si_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
evergreen_print_gpu_status_regs(rdev);
}
+static void si_set_clk_bypass_mode(struct radeon_device *rdev)
+{
+ u32 tmp, i;
+
+ tmp = RREG32(CG_SPLL_FUNC_CNTL);
+ tmp |= SPLL_BYPASS_EN;
+ WREG32(CG_SPLL_FUNC_CNTL, tmp);
+
+ tmp = RREG32(CG_SPLL_FUNC_CNTL_2);
+ tmp |= SPLL_CTLREQ_CHG;
+ WREG32(CG_SPLL_FUNC_CNTL_2, tmp);
+
+ for (i = 0; i < rdev->usec_timeout; i++) {
+ if (RREG32(SPLL_STATUS) & SPLL_CHG_STATUS)
+ break;
+ udelay(1);
+ }
+
+ tmp = RREG32(CG_SPLL_FUNC_CNTL_2);
+ tmp &= ~(SPLL_CTLREQ_CHG | SCLK_MUX_UPDATE);
+ WREG32(CG_SPLL_FUNC_CNTL_2, tmp);
+
+ tmp = RREG32(MPLL_CNTL_MODE);
+ tmp &= ~MPLL_MCLK_SEL;
+ WREG32(MPLL_CNTL_MODE, tmp);
+}
+
+static void si_spll_powerdown(struct radeon_device *rdev)
+{
+ u32 tmp;
+
+ tmp = RREG32(SPLL_CNTL_MODE);
+ tmp |= SPLL_SW_DIR_CONTROL;
+ WREG32(SPLL_CNTL_MODE, tmp);
+
+ tmp = RREG32(CG_SPLL_FUNC_CNTL);
+ tmp |= SPLL_RESET;
+ WREG32(CG_SPLL_FUNC_CNTL, tmp);
+
+ tmp = RREG32(CG_SPLL_FUNC_CNTL);
+ tmp |= SPLL_SLEEP;
+ WREG32(CG_SPLL_FUNC_CNTL, tmp);
+
+ tmp = RREG32(SPLL_CNTL_MODE);
+ tmp &= ~SPLL_SW_DIR_CONTROL;
+ WREG32(SPLL_CNTL_MODE, tmp);
+}
+
+static void si_gpu_pci_config_reset(struct radeon_device *rdev)
+{
+ struct evergreen_mc_save save;
+ u32 tmp, i;
+
+ dev_info(rdev->dev, "GPU pci config reset\n");
+
+ /* disable dpm? */
+
+ /* disable cg/pg */
+ si_fini_pg(rdev);
+ si_fini_cg(rdev);
+
+ /* Disable CP parsing/prefetching */
+ WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
+ /* dma0 */
+ tmp = RREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET);
+ tmp &= ~DMA_RB_ENABLE;
+ WREG32(DMA_RB_CNTL + DMA0_REGISTER_OFFSET, tmp);
+ /* dma1 */
+ tmp = RREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET);
+ tmp &= ~DMA_RB_ENABLE;
+ WREG32(DMA_RB_CNTL + DMA1_REGISTER_OFFSET, tmp);
+ /* XXX other engines? */
+
+ /* halt the rlc, disable cp internal ints */
+ si_rlc_stop(rdev);
+
+ udelay(50);
+
+ /* disable mem access */
+ evergreen_mc_stop(rdev, &save);
+ if (evergreen_mc_wait_for_idle(rdev)) {
+ dev_warn(rdev->dev, "Wait for MC idle timed out !\n");
+ }
+
+ /* set mclk/sclk to bypass */
+ si_set_clk_bypass_mode(rdev);
+ /* powerdown spll */
+ si_spll_powerdown(rdev);
+ /* disable BM */
+ pci_clear_master(rdev->pdev);
+ /* reset */
+ radeon_pci_config_reset(rdev);
+ /* wait for asic to come out of reset */
+ for (i = 0; i < rdev->usec_timeout; i++) {
+ if (RREG32(CONFIG_MEMSIZE) != 0xffffffff)
+ break;
+ udelay(1);
+ }
+}
+
int si_asic_reset(struct radeon_device *rdev)
{
u32 reset_mask;
@@ -3704,10 +3853,17 @@ int si_asic_reset(struct radeon_device *rdev)
if (reset_mask)
r600_set_bios_scratch_engine_hung(rdev, true);
+ /* try soft reset */
si_gpu_soft_reset(rdev, reset_mask);
reset_mask = si_gpu_check_soft_reset(rdev);
+ /* try pci config reset */
+ if (reset_mask && radeon_hard_reset)
+ si_gpu_pci_config_reset(rdev);
+
+ reset_mask = si_gpu_check_soft_reset(rdev);
+
if (!reset_mask)
r600_set_bios_scratch_engine_hung(rdev, false);
@@ -3730,39 +3886,9 @@ bool si_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
if (!(reset_mask & (RADEON_RESET_GFX |
RADEON_RESET_COMPUTE |
RADEON_RESET_CP))) {
- radeon_ring_lockup_update(ring);
+ radeon_ring_lockup_update(rdev, ring);
return false;
}
- /* force CP activities */
- radeon_ring_force_activity(rdev, ring);
- return radeon_ring_test_lockup(rdev, ring);
-}
-
-/**
- * si_dma_is_lockup - Check if the DMA engine is locked up
- *
- * @rdev: radeon_device pointer
- * @ring: radeon_ring structure holding ring information
- *
- * Check if the async DMA engine is locked up.
- * Returns true if the engine appears to be locked up, false if not.
- */
-bool si_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
-{
- u32 reset_mask = si_gpu_check_soft_reset(rdev);
- u32 mask;
-
- if (ring->idx == R600_RING_TYPE_DMA_INDEX)
- mask = RADEON_RESET_DMA;
- else
- mask = RADEON_RESET_DMA1;
-
- if (!(reset_mask & mask)) {
- radeon_ring_lockup_update(ring);
- return false;
- }
- /* force ring activities */
- radeon_ring_force_activity(rdev, ring);
return radeon_ring_test_lockup(rdev, ring);
}
@@ -3883,8 +4009,15 @@ static int si_mc_init(struct radeon_device *rdev)
rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0);
rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0);
/* size in MB on si */
- rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL;
- rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024ULL * 1024ULL;
+ tmp = RREG32(CONFIG_MEMSIZE);
+ /* some boards may have garbage in the upper 16 bits */
+ if (tmp & 0xffff0000) {
+ DRM_INFO("Probable bad vram size: 0x%08x\n", tmp);
+ if (tmp & 0xffff)
+ tmp &= 0xffff;
+ }
+ rdev->mc.mc_vram_size = tmp * 1024ULL * 1024ULL;
+ rdev->mc.real_vram_size = rdev->mc.mc_vram_size;
rdev->mc.visible_vram_size = rdev->mc.aper_size;
si_vram_gtt_location(rdev, &rdev->mc);
radeon_update_bandwidth_info(rdev);
@@ -3920,18 +4053,21 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
WREG32(MC_VM_MX_L1_TLB_CNTL,
(0xA << 7) |
ENABLE_L1_TLB |
+ ENABLE_L1_FRAGMENT_PROCESSING |
SYSTEM_ACCESS_MODE_NOT_IN_SYS |
ENABLE_ADVANCED_DRIVER_MODEL |
SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
/* Setup L2 cache */
WREG32(VM_L2_CNTL, ENABLE_L2_CACHE |
+ ENABLE_L2_FRAGMENT_PROCESSING |
ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
EFFECTIVE_L2_QUEUE_SIZE(7) |
CONTEXT1_IDENTITY_ACCESS_MODE(1));
WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE);
WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
- L2_CACHE_BIGK_FRAGMENT_SIZE(0));
+ BANK_SELECT(4) |
+ L2_CACHE_BIGK_FRAGMENT_SIZE(4));
/* setup context0 */
WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12);
WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12);
@@ -3968,6 +4104,7 @@ static int si_pcie_gart_enable(struct radeon_device *rdev)
(u32)(rdev->dummy_page.addr >> 12));
WREG32(VM_CONTEXT1_CNTL2, 4);
WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) |
+ PAGE_TABLE_BLOCK_SIZE(radeon_vm_block_size - 9) |
RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
RANGE_PROTECTION_FAULT_ENABLE_DEFAULT |
DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
@@ -4079,13 +4216,64 @@ static int si_vm_packet3_ce_check(struct radeon_device *rdev,
return 0;
}
+static int si_vm_packet3_cp_dma_check(u32 *ib, u32 idx)
+{
+ u32 start_reg, reg, i;
+ u32 command = ib[idx + 4];
+ u32 info = ib[idx + 1];
+ u32 idx_value = ib[idx];
+ if (command & PACKET3_CP_DMA_CMD_SAS) {
+ /* src address space is register */
+ if (((info & 0x60000000) >> 29) == 0) {
+ start_reg = idx_value << 2;
+ if (command & PACKET3_CP_DMA_CMD_SAIC) {
+ reg = start_reg;
+ if (!si_vm_reg_valid(reg)) {
+ DRM_ERROR("CP DMA Bad SRC register\n");
+ return -EINVAL;
+ }
+ } else {
+ for (i = 0; i < (command & 0x1fffff); i++) {
+ reg = start_reg + (4 * i);
+ if (!si_vm_reg_valid(reg)) {
+ DRM_ERROR("CP DMA Bad SRC register\n");
+ return -EINVAL;
+ }
+ }
+ }
+ }
+ }
+ if (command & PACKET3_CP_DMA_CMD_DAS) {
+ /* dst address space is register */
+ if (((info & 0x00300000) >> 20) == 0) {
+ start_reg = ib[idx + 2];
+ if (command & PACKET3_CP_DMA_CMD_DAIC) {
+ reg = start_reg;
+ if (!si_vm_reg_valid(reg)) {
+ DRM_ERROR("CP DMA Bad DST register\n");
+ return -EINVAL;
+ }
+ } else {
+ for (i = 0; i < (command & 0x1fffff); i++) {
+ reg = start_reg + (4 * i);
+ if (!si_vm_reg_valid(reg)) {
+ DRM_ERROR("CP DMA Bad DST register\n");
+ return -EINVAL;
+ }
+ }
+ }
+ }
+ }
+ return 0;
+}
+
static int si_vm_packet3_gfx_check(struct radeon_device *rdev,
u32 *ib, struct radeon_cs_packet *pkt)
{
+ int r;
u32 idx = pkt->idx + 1;
u32 idx_value = ib[idx];
u32 start_reg, end_reg, reg, i;
- u32 command, info;
switch (pkt->opcode) {
case PACKET3_NOP:
@@ -4186,50 +4374,9 @@ static int si_vm_packet3_gfx_check(struct radeon_device *rdev,
}
break;
case PACKET3_CP_DMA:
- command = ib[idx + 4];
- info = ib[idx + 1];
- if (command & PACKET3_CP_DMA_CMD_SAS) {
- /* src address space is register */
- if (((info & 0x60000000) >> 29) == 0) {
- start_reg = idx_value << 2;
- if (command & PACKET3_CP_DMA_CMD_SAIC) {
- reg = start_reg;
- if (!si_vm_reg_valid(reg)) {
- DRM_ERROR("CP DMA Bad SRC register\n");
- return -EINVAL;
- }
- } else {
- for (i = 0; i < (command & 0x1fffff); i++) {
- reg = start_reg + (4 * i);
- if (!si_vm_reg_valid(reg)) {
- DRM_ERROR("CP DMA Bad SRC register\n");
- return -EINVAL;
- }
- }
- }
- }
- }
- if (command & PACKET3_CP_DMA_CMD_DAS) {
- /* dst address space is register */
- if (((info & 0x00300000) >> 20) == 0) {
- start_reg = ib[idx + 2];
- if (command & PACKET3_CP_DMA_CMD_DAIC) {
- reg = start_reg;
- if (!si_vm_reg_valid(reg)) {
- DRM_ERROR("CP DMA Bad DST register\n");
- return -EINVAL;
- }
- } else {
- for (i = 0; i < (command & 0x1fffff); i++) {
- reg = start_reg + (4 * i);
- if (!si_vm_reg_valid(reg)) {
- DRM_ERROR("CP DMA Bad DST register\n");
- return -EINVAL;
- }
- }
- }
- }
- }
+ r = si_vm_packet3_cp_dma_check(ib, idx);
+ if (r)
+ return r;
break;
default:
DRM_ERROR("Invalid GFX packet3: 0x%x\n", pkt->opcode);
@@ -4241,6 +4388,7 @@ static int si_vm_packet3_gfx_check(struct radeon_device *rdev,
static int si_vm_packet3_compute_check(struct radeon_device *rdev,
u32 *ib, struct radeon_cs_packet *pkt)
{
+ int r;
u32 idx = pkt->idx + 1;
u32 idx_value = ib[idx];
u32 start_reg, reg, i;
@@ -4313,6 +4461,11 @@ static int si_vm_packet3_compute_check(struct radeon_device *rdev,
return -EINVAL;
}
break;
+ case PACKET3_CP_DMA:
+ r = si_vm_packet3_cp_dma_check(ib, idx);
+ if (r)
+ return r;
+ break;
default:
DRM_ERROR("Invalid Compute packet3: 0x%x\n", pkt->opcode);
return -EINVAL;
@@ -4653,112 +4806,6 @@ static void si_vm_decode_fault(struct radeon_device *rdev,
block, mc_id);
}
-/**
- * si_vm_set_page - update the page tables using the CP
- *
- * @rdev: radeon_device pointer
- * @ib: indirect buffer to fill with commands
- * @pe: addr of the page entry
- * @addr: dst addr to write into pe
- * @count: number of page entries to update
- * @incr: increase next addr by incr bytes
- * @flags: access flags
- *
- * Update the page tables using the CP (SI).
- */
-void si_vm_set_page(struct radeon_device *rdev,
- struct radeon_ib *ib,
- uint64_t pe,
- uint64_t addr, unsigned count,
- uint32_t incr, uint32_t flags)
-{
- uint32_t r600_flags = cayman_vm_page_flags(rdev, flags);
- uint64_t value;
- unsigned ndw;
-
- if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) {
- while (count) {
- ndw = 2 + count * 2;
- if (ndw > 0x3FFE)
- ndw = 0x3FFE;
-
- ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw);
- ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) |
- WRITE_DATA_DST_SEL(1));
- ib->ptr[ib->length_dw++] = pe;
- ib->ptr[ib->length_dw++] = upper_32_bits(pe);
- for (; ndw > 2; ndw -= 2, --count, pe += 8) {
- if (flags & RADEON_VM_PAGE_SYSTEM) {
- value = radeon_vm_map_gart(rdev, addr);
- value &= 0xFFFFFFFFFFFFF000ULL;
- } else if (flags & RADEON_VM_PAGE_VALID) {
- value = addr;
- } else {
- value = 0;
- }
- addr += incr;
- value |= r600_flags;
- ib->ptr[ib->length_dw++] = value;
- ib->ptr[ib->length_dw++] = upper_32_bits(value);
- }
- }
- } else {
- /* DMA */
- if (flags & RADEON_VM_PAGE_SYSTEM) {
- while (count) {
- ndw = count * 2;
- if (ndw > 0xFFFFE)
- ndw = 0xFFFFE;
-
- /* for non-physically contiguous pages (system) */
- ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_WRITE, 0, 0, 0, ndw);
- ib->ptr[ib->length_dw++] = pe;
- ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
- for (; ndw > 0; ndw -= 2, --count, pe += 8) {
- if (flags & RADEON_VM_PAGE_SYSTEM) {
- value = radeon_vm_map_gart(rdev, addr);
- value &= 0xFFFFFFFFFFFFF000ULL;
- } else if (flags & RADEON_VM_PAGE_VALID) {
- value = addr;
- } else {
- value = 0;
- }
- addr += incr;
- value |= r600_flags;
- ib->ptr[ib->length_dw++] = value;
- ib->ptr[ib->length_dw++] = upper_32_bits(value);
- }
- }
- } else {
- while (count) {
- ndw = count * 2;
- if (ndw > 0xFFFFE)
- ndw = 0xFFFFE;
-
- if (flags & RADEON_VM_PAGE_VALID)
- value = addr;
- else
- value = 0;
- /* for physically contiguous pages (vram) */
- ib->ptr[ib->length_dw++] = DMA_PTE_PDE_PACKET(ndw);
- ib->ptr[ib->length_dw++] = pe; /* dst addr */
- ib->ptr[ib->length_dw++] = upper_32_bits(pe) & 0xff;
- ib->ptr[ib->length_dw++] = r600_flags; /* mask */
- ib->ptr[ib->length_dw++] = 0;
- ib->ptr[ib->length_dw++] = value; /* value */
- ib->ptr[ib->length_dw++] = upper_32_bits(value);
- ib->ptr[ib->length_dw++] = incr; /* increment size */
- ib->ptr[ib->length_dw++] = 0;
- pe += ndw * 4;
- addr += (ndw / 2) * incr;
- count -= ndw / 2;
- }
- }
- while (ib->length_dw & 0x7)
- ib->ptr[ib->length_dw++] = DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0);
- }
-}
-
void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
{
struct radeon_ring *ring = &rdev->ring[ridx];
@@ -4802,32 +4849,6 @@ void si_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
radeon_ring_write(ring, 0x0);
}
-void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
-{
- struct radeon_ring *ring = &rdev->ring[ridx];
-
- if (vm == NULL)
- return;
-
- radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
- if (vm->id < 8) {
- radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2));
- } else {
- radeon_ring_write(ring, (0xf << 16) | ((VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2));
- }
- radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
-
- /* flush hdp cache */
- radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
- radeon_ring_write(ring, (0xf << 16) | (HDP_MEM_COHERENCY_FLUSH_CNTL >> 2));
- radeon_ring_write(ring, 1);
-
- /* bits 0-7 are the VM contexts0-7 */
- radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_SRBM_WRITE, 0, 0, 0, 0));
- radeon_ring_write(ring, (0xf << 16) | (VM_INVALIDATE_REQUEST >> 2));
- radeon_ring_write(ring, 1 << vm->id);
-}
-
/*
* Power and clock gating
*/
@@ -4895,7 +4916,7 @@ static void si_set_uvd_dcm(struct radeon_device *rdev,
WREG32_UVD_CTX(UVD_CGC_CTRL2, tmp2);
}
-static void si_init_uvd_internal_cg(struct radeon_device *rdev)
+void si_init_uvd_internal_cg(struct radeon_device *rdev)
{
bool hw_mode = true;
@@ -4938,7 +4959,7 @@ static void si_enable_dma_pg(struct radeon_device *rdev, bool enable)
u32 data, orig;
orig = data = RREG32(DMA_PG);
- if (enable)
+ if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_SDMA))
data |= PG_CNTL_ENABLE;
else
data &= ~PG_CNTL_ENABLE;
@@ -4962,7 +4983,7 @@ static void si_enable_gfx_cgpg(struct radeon_device *rdev,
{
u32 tmp;
- if (enable) {
+ if (enable && (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG)) {
tmp = RLC_PUD(0x10) | RLC_PDD(0x10) | RLC_TTPD(0x10) | RLC_MSD(0x10);
WREG32(RLC_TTOP_D, tmp);
@@ -5065,9 +5086,9 @@ static void si_enable_cgcg(struct radeon_device *rdev,
orig = data = RREG32(RLC_CGCG_CGLS_CTRL);
- si_enable_gui_idle_interrupt(rdev, enable);
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CGCG)) {
+ si_enable_gui_idle_interrupt(rdev, true);
- if (enable) {
WREG32(RLC_GCPM_GENERAL_3, 0x00000080);
tmp = si_halt_rlc(rdev);
@@ -5084,6 +5105,8 @@ static void si_enable_cgcg(struct radeon_device *rdev,
data |= CGCG_EN | CGLS_EN;
} else {
+ si_enable_gui_idle_interrupt(rdev, false);
+
RREG32(CB_CGTT_SCLK_CTRL);
RREG32(CB_CGTT_SCLK_CTRL);
RREG32(CB_CGTT_SCLK_CTRL);
@@ -5101,16 +5124,18 @@ static void si_enable_mgcg(struct radeon_device *rdev,
{
u32 data, orig, tmp = 0;
- if (enable) {
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_MGCG)) {
orig = data = RREG32(CGTS_SM_CTRL_REG);
data = 0x96940200;
if (orig != data)
WREG32(CGTS_SM_CTRL_REG, data);
- orig = data = RREG32(CP_MEM_SLP_CNTL);
- data |= CP_MEM_LS_EN;
- if (orig != data)
- WREG32(CP_MEM_SLP_CNTL, data);
+ if (rdev->cg_flags & RADEON_CG_SUPPORT_GFX_CP_LS) {
+ orig = data = RREG32(CP_MEM_SLP_CNTL);
+ data |= CP_MEM_LS_EN;
+ if (orig != data)
+ WREG32(CP_MEM_SLP_CNTL, data);
+ }
orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
data &= 0xffffffc0;
@@ -5155,7 +5180,7 @@ static void si_enable_uvd_mgcg(struct radeon_device *rdev,
{
u32 orig, data, tmp;
- if (enable) {
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_UVD_MGCG)) {
tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL);
tmp |= 0x3fff;
WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp);
@@ -5203,7 +5228,7 @@ static void si_enable_mc_ls(struct radeon_device *rdev,
for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
orig = data = RREG32(mc_cg_registers[i]);
- if (enable)
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_LS))
data |= MC_LS_ENABLE;
else
data &= ~MC_LS_ENABLE;
@@ -5212,226 +5237,301 @@ static void si_enable_mc_ls(struct radeon_device *rdev,
}
}
-
-static void si_init_cg(struct radeon_device *rdev)
+static void si_enable_mc_mgcg(struct radeon_device *rdev,
+ bool enable)
{
- si_enable_mgcg(rdev, true);
- si_enable_cgcg(rdev, false);
- /* disable MC LS on Tahiti */
- if (rdev->family == CHIP_TAHITI)
- si_enable_mc_ls(rdev, false);
- if (rdev->has_uvd) {
- si_enable_uvd_mgcg(rdev, true);
- si_init_uvd_internal_cg(rdev);
+ int i;
+ u32 orig, data;
+
+ for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
+ orig = data = RREG32(mc_cg_registers[i]);
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_MC_MGCG))
+ data |= MC_CG_ENABLE;
+ else
+ data &= ~MC_CG_ENABLE;
+ if (data != orig)
+ WREG32(mc_cg_registers[i], data);
}
}
-static void si_fini_cg(struct radeon_device *rdev)
+static void si_enable_dma_mgcg(struct radeon_device *rdev,
+ bool enable)
{
- if (rdev->has_uvd)
- si_enable_uvd_mgcg(rdev, false);
- si_enable_cgcg(rdev, false);
- si_enable_mgcg(rdev, false);
-}
+ u32 orig, data, offset;
+ int i;
-static void si_init_pg(struct radeon_device *rdev)
-{
- bool has_pg = false;
-#if 0
- /* only cape verde supports PG */
- if (rdev->family == CHIP_VERDE)
- has_pg = true;
-#endif
- if (has_pg) {
- si_init_ao_cu_mask(rdev);
- si_init_dma_pg(rdev);
- si_enable_dma_pg(rdev, true);
- si_init_gfx_cgpg(rdev);
- si_enable_gfx_cgpg(rdev, true);
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_SDMA_MGCG)) {
+ for (i = 0; i < 2; i++) {
+ if (i == 0)
+ offset = DMA0_REGISTER_OFFSET;
+ else
+ offset = DMA1_REGISTER_OFFSET;
+ orig = data = RREG32(DMA_POWER_CNTL + offset);
+ data &= ~MEM_POWER_OVERRIDE;
+ if (data != orig)
+ WREG32(DMA_POWER_CNTL + offset, data);
+ WREG32(DMA_CLK_CTRL + offset, 0x00000100);
+ }
} else {
- WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
- WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+ for (i = 0; i < 2; i++) {
+ if (i == 0)
+ offset = DMA0_REGISTER_OFFSET;
+ else
+ offset = DMA1_REGISTER_OFFSET;
+ orig = data = RREG32(DMA_POWER_CNTL + offset);
+ data |= MEM_POWER_OVERRIDE;
+ if (data != orig)
+ WREG32(DMA_POWER_CNTL + offset, data);
+
+ orig = data = RREG32(DMA_CLK_CTRL + offset);
+ data = 0xff000000;
+ if (data != orig)
+ WREG32(DMA_CLK_CTRL + offset, data);
+ }
}
}
-static void si_fini_pg(struct radeon_device *rdev)
+static void si_enable_bif_mgls(struct radeon_device *rdev,
+ bool enable)
{
- bool has_pg = false;
+ u32 orig, data;
- /* only cape verde supports PG */
- if (rdev->family == CHIP_VERDE)
- has_pg = true;
+ orig = data = RREG32_PCIE(PCIE_CNTL2);
- if (has_pg) {
- si_enable_dma_pg(rdev, false);
- si_enable_gfx_cgpg(rdev, false);
- }
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_BIF_LS))
+ data |= SLV_MEM_LS_EN | MST_MEM_LS_EN |
+ REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN;
+ else
+ data &= ~(SLV_MEM_LS_EN | MST_MEM_LS_EN |
+ REPLAY_MEM_LS_EN | SLV_MEM_AGGRESSIVE_LS_EN);
+
+ if (orig != data)
+ WREG32_PCIE(PCIE_CNTL2, data);
}
-/*
- * RLC
- */
-void si_rlc_fini(struct radeon_device *rdev)
+static void si_enable_hdp_mgcg(struct radeon_device *rdev,
+ bool enable)
{
- int r;
-
- /* save restore block */
- if (rdev->rlc.save_restore_obj) {
- r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
- if (unlikely(r != 0))
- dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r);
- radeon_bo_unpin(rdev->rlc.save_restore_obj);
- radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+ u32 orig, data;
- radeon_bo_unref(&rdev->rlc.save_restore_obj);
- rdev->rlc.save_restore_obj = NULL;
- }
+ orig = data = RREG32(HDP_HOST_PATH_CNTL);
- /* clear state block */
- if (rdev->rlc.clear_state_obj) {
- r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
- if (unlikely(r != 0))
- dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r);
- radeon_bo_unpin(rdev->rlc.clear_state_obj);
- radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_MGCG))
+ data &= ~CLOCK_GATING_DIS;
+ else
+ data |= CLOCK_GATING_DIS;
- radeon_bo_unref(&rdev->rlc.clear_state_obj);
- rdev->rlc.clear_state_obj = NULL;
- }
+ if (orig != data)
+ WREG32(HDP_HOST_PATH_CNTL, data);
}
-#define RLC_CLEAR_STATE_END_MARKER 0x00000001
-
-int si_rlc_init(struct radeon_device *rdev)
+static void si_enable_hdp_ls(struct radeon_device *rdev,
+ bool enable)
{
- volatile u32 *dst_ptr;
- u32 dws, data, i, j, k, reg_num;
- u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index;
- u64 reg_list_mc_addr;
- const struct cs_section_def *cs_data = si_cs_data;
- int r;
+ u32 orig, data;
- /* save restore block */
- if (rdev->rlc.save_restore_obj == NULL) {
- r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
- RADEON_GEM_DOMAIN_VRAM, NULL,
- &rdev->rlc.save_restore_obj);
- if (r) {
- dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
- return r;
+ orig = data = RREG32(HDP_MEM_POWER_LS);
+
+ if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_HDP_LS))
+ data |= HDP_LS_ENABLE;
+ else
+ data &= ~HDP_LS_ENABLE;
+
+ if (orig != data)
+ WREG32(HDP_MEM_POWER_LS, data);
+}
+
+static void si_update_cg(struct radeon_device *rdev,
+ u32 block, bool enable)
+{
+ if (block & RADEON_CG_BLOCK_GFX) {
+ si_enable_gui_idle_interrupt(rdev, false);
+ /* order matters! */
+ if (enable) {
+ si_enable_mgcg(rdev, true);
+ si_enable_cgcg(rdev, true);
+ } else {
+ si_enable_cgcg(rdev, false);
+ si_enable_mgcg(rdev, false);
}
+ si_enable_gui_idle_interrupt(rdev, true);
}
- r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
- if (unlikely(r != 0)) {
- si_rlc_fini(rdev);
- return r;
+ if (block & RADEON_CG_BLOCK_MC) {
+ si_enable_mc_mgcg(rdev, enable);
+ si_enable_mc_ls(rdev, enable);
}
- r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM,
- &rdev->rlc.save_restore_gpu_addr);
- if (r) {
- radeon_bo_unreserve(rdev->rlc.save_restore_obj);
- dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r);
- si_rlc_fini(rdev);
- return r;
+
+ if (block & RADEON_CG_BLOCK_SDMA) {
+ si_enable_dma_mgcg(rdev, enable);
}
- if (rdev->family == CHIP_VERDE) {
- r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr);
- if (r) {
- dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r);
- si_rlc_fini(rdev);
- return r;
- }
- /* write the sr buffer */
- dst_ptr = rdev->rlc.sr_ptr;
- for (i = 0; i < ARRAY_SIZE(verde_rlc_save_restore_register_list); i++) {
- dst_ptr[i] = verde_rlc_save_restore_register_list[i];
- }
- radeon_bo_kunmap(rdev->rlc.save_restore_obj);
+ if (block & RADEON_CG_BLOCK_BIF) {
+ si_enable_bif_mgls(rdev, enable);
}
- radeon_bo_unreserve(rdev->rlc.save_restore_obj);
- /* clear state block */
- reg_list_num = 0;
- dws = 0;
- for (i = 0; cs_data[i].section != NULL; i++) {
- for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
- reg_list_num++;
- dws += cs_data[i].section[j].reg_count;
+ if (block & RADEON_CG_BLOCK_UVD) {
+ if (rdev->has_uvd) {
+ si_enable_uvd_mgcg(rdev, enable);
}
}
- reg_list_blk_index = (3 * reg_list_num + 2);
- dws += reg_list_blk_index;
- if (rdev->rlc.clear_state_obj == NULL) {
- r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
- RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
- if (r) {
- dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
- si_rlc_fini(rdev);
- return r;
- }
+ if (block & RADEON_CG_BLOCK_HDP) {
+ si_enable_hdp_mgcg(rdev, enable);
+ si_enable_hdp_ls(rdev, enable);
}
- r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
- if (unlikely(r != 0)) {
- si_rlc_fini(rdev);
- return r;
+}
+
+static void si_init_cg(struct radeon_device *rdev)
+{
+ si_update_cg(rdev, (RADEON_CG_BLOCK_GFX |
+ RADEON_CG_BLOCK_MC |
+ RADEON_CG_BLOCK_SDMA |
+ RADEON_CG_BLOCK_BIF |
+ RADEON_CG_BLOCK_HDP), true);
+ if (rdev->has_uvd) {
+ si_update_cg(rdev, RADEON_CG_BLOCK_UVD, true);
+ si_init_uvd_internal_cg(rdev);
}
- r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM,
- &rdev->rlc.clear_state_gpu_addr);
- if (r) {
+}
- radeon_bo_unreserve(rdev->rlc.clear_state_obj);
- dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r);
- si_rlc_fini(rdev);
- return r;
+static void si_fini_cg(struct radeon_device *rdev)
+{
+ if (rdev->has_uvd) {
+ si_update_cg(rdev, RADEON_CG_BLOCK_UVD, false);
}
- r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr);
- if (r) {
- dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r);
- si_rlc_fini(rdev);
- return r;
+ si_update_cg(rdev, (RADEON_CG_BLOCK_GFX |
+ RADEON_CG_BLOCK_MC |
+ RADEON_CG_BLOCK_SDMA |
+ RADEON_CG_BLOCK_BIF |
+ RADEON_CG_BLOCK_HDP), false);
+}
+
+u32 si_get_csb_size(struct radeon_device *rdev)
+{
+ u32 count = 0;
+ const struct cs_section_def *sect = NULL;
+ const struct cs_extent_def *ext = NULL;
+
+ if (rdev->rlc.cs_data == NULL)
+ return 0;
+
+ /* begin clear state */
+ count += 2;
+ /* context control state */
+ count += 3;
+
+ for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) {
+ for (ext = sect->section; ext->extent != NULL; ++ext) {
+ if (sect->id == SECT_CONTEXT)
+ count += 2 + ext->reg_count;
+ else
+ return 0;
+ }
}
- /* set up the cs buffer */
- dst_ptr = rdev->rlc.cs_ptr;
- reg_list_hdr_blk_index = 0;
- reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4);
- data = upper_32_bits(reg_list_mc_addr);
- dst_ptr[reg_list_hdr_blk_index] = data;
- reg_list_hdr_blk_index++;
- for (i = 0; cs_data[i].section != NULL; i++) {
- for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
- reg_num = cs_data[i].section[j].reg_count;
- data = reg_list_mc_addr & 0xffffffff;
- dst_ptr[reg_list_hdr_blk_index] = data;
- reg_list_hdr_blk_index++;
-
- data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff;
- dst_ptr[reg_list_hdr_blk_index] = data;
- reg_list_hdr_blk_index++;
-
- data = 0x08000000 | (reg_num * 4);
- dst_ptr[reg_list_hdr_blk_index] = data;
- reg_list_hdr_blk_index++;
-
- for (k = 0; k < reg_num; k++) {
- data = cs_data[i].section[j].extent[k];
- dst_ptr[reg_list_blk_index + k] = data;
+ /* pa_sc_raster_config */
+ count += 3;
+ /* end clear state */
+ count += 2;
+ /* clear state */
+ count += 2;
+
+ return count;
+}
+
+void si_get_csb_buffer(struct radeon_device *rdev, volatile u32 *buffer)
+{
+ u32 count = 0, i;
+ const struct cs_section_def *sect = NULL;
+ const struct cs_extent_def *ext = NULL;
+
+ if (rdev->rlc.cs_data == NULL)
+ return;
+ if (buffer == NULL)
+ return;
+
+ buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+ buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+ buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+ buffer[count++] = cpu_to_le32(0x80000000);
+ buffer[count++] = cpu_to_le32(0x80000000);
+
+ for (sect = rdev->rlc.cs_data; sect->section != NULL; ++sect) {
+ for (ext = sect->section; ext->extent != NULL; ++ext) {
+ if (sect->id == SECT_CONTEXT) {
+ buffer[count++] =
+ cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
+ buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000);
+ for (i = 0; i < ext->reg_count; i++)
+ buffer[count++] = cpu_to_le32(ext->extent[i]);
+ } else {
+ return;
}
- reg_list_mc_addr += reg_num * 4;
- reg_list_blk_index += reg_num;
}
}
- dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER;
- radeon_bo_kunmap(rdev->rlc.clear_state_obj);
- radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+ buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+ buffer[count++] = cpu_to_le32(PA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START);
+ switch (rdev->family) {
+ case CHIP_TAHITI:
+ case CHIP_PITCAIRN:
+ buffer[count++] = cpu_to_le32(0x2a00126a);
+ break;
+ case CHIP_VERDE:
+ buffer[count++] = cpu_to_le32(0x0000124a);
+ break;
+ case CHIP_OLAND:
+ buffer[count++] = cpu_to_le32(0x00000082);
+ break;
+ case CHIP_HAINAN:
+ buffer[count++] = cpu_to_le32(0x00000000);
+ break;
+ default:
+ buffer[count++] = cpu_to_le32(0x00000000);
+ break;
+ }
- return 0;
+ buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+ buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+ buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
+ buffer[count++] = cpu_to_le32(0);
}
-static void si_rlc_reset(struct radeon_device *rdev)
+static void si_init_pg(struct radeon_device *rdev)
+{
+ if (rdev->pg_flags) {
+ if (rdev->pg_flags & RADEON_PG_SUPPORT_SDMA) {
+ si_init_dma_pg(rdev);
+ }
+ si_init_ao_cu_mask(rdev);
+ if (rdev->pg_flags & RADEON_PG_SUPPORT_GFX_PG) {
+ si_init_gfx_cgpg(rdev);
+ } else {
+ WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+ WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+ }
+ si_enable_dma_pg(rdev, true);
+ si_enable_gfx_cgpg(rdev, true);
+ } else {
+ WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+ WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+ }
+}
+
+static void si_fini_pg(struct radeon_device *rdev)
+{
+ if (rdev->pg_flags) {
+ si_enable_dma_pg(rdev, false);
+ si_enable_gfx_cgpg(rdev, false);
+ }
+}
+
+/*
+ * RLC
+ */
+void si_rlc_reset(struct radeon_device *rdev)
{
u32 tmp = RREG32(GRBM_SOFT_RESET);
@@ -5561,7 +5661,9 @@ static void si_disable_interrupt_state(struct radeon_device *rdev)
{
u32 tmp;
- WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+ tmp = RREG32(CP_INT_CNTL_RING0) &
+ (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+ WREG32(CP_INT_CNTL_RING0, tmp);
WREG32(CP_INT_CNTL_RING1, 0);
WREG32(CP_INT_CNTL_RING2, 0);
tmp = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
@@ -5596,7 +5698,7 @@ static void si_disable_interrupt_state(struct radeon_device *rdev)
}
if (!ASIC_IS_NODCE(rdev)) {
- WREG32(DACA_AUTODETECT_INT_CONTROL, 0);
+ WREG32(DAC_AUTODETECT_INT_CONTROL, 0);
tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY;
WREG32(DC_HPD1_INT_CONTROL, tmp);
@@ -5647,7 +5749,7 @@ static int si_irq_init(struct radeon_device *rdev)
WREG32(INTERRUPT_CNTL, interrupt_cntl);
WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8);
- rb_bufsz = drm_order(rdev->ih.ring_size / 4);
+ rb_bufsz = order_base_2(rdev->ih.ring_size / 4);
ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE |
IH_WPTR_OVERFLOW_CLEAR |
@@ -5686,12 +5788,11 @@ static int si_irq_init(struct radeon_device *rdev)
int si_irq_set(struct radeon_device *rdev)
{
- u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE;
+ u32 cp_int_cntl;
u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0;
u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
u32 hpd1 = 0, hpd2 = 0, hpd3 = 0, hpd4 = 0, hpd5 = 0, hpd6 = 0;
u32 grbm_int_cntl = 0;
- u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
u32 dma_cntl, dma_cntl1;
u32 thermal_int = 0;
@@ -5707,6 +5808,9 @@ int si_irq_set(struct radeon_device *rdev)
return 0;
}
+ cp_int_cntl = RREG32(CP_INT_CNTL_RING0) &
+ (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+
if (!ASIC_IS_NODCE(rdev)) {
hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN;
hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN;
@@ -5827,16 +5931,22 @@ int si_irq_set(struct radeon_device *rdev)
}
if (rdev->num_crtc >= 2) {
- WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1);
- WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2);
+ WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET,
+ GRPH_PFLIP_INT_MASK);
+ WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET,
+ GRPH_PFLIP_INT_MASK);
}
if (rdev->num_crtc >= 4) {
- WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3);
- WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4);
+ WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET,
+ GRPH_PFLIP_INT_MASK);
+ WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET,
+ GRPH_PFLIP_INT_MASK);
}
if (rdev->num_crtc >= 6) {
- WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5);
- WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6);
+ WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET,
+ GRPH_PFLIP_INT_MASK);
+ WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET,
+ GRPH_PFLIP_INT_MASK);
}
if (!ASIC_IS_NODCE(rdev)) {
@@ -5993,6 +6103,7 @@ static inline u32 si_get_ih_wptr(struct radeon_device *rdev)
tmp = RREG32(IH_RB_CNTL);
tmp |= IH_WPTR_OVERFLOW_CLEAR;
WREG32(IH_RB_CNTL, tmp);
+ wptr &= ~RB_OVERFLOW;
}
return (wptr & rdev->ih.ptr_mask);
}
@@ -6054,7 +6165,7 @@ restart_ih:
wake_up(&rdev->irq.vblank_queue);
}
if (atomic_read(&rdev->irq.pflip[0]))
- radeon_crtc_handle_flip(rdev, 0);
+ radeon_crtc_handle_vblank(rdev, 0);
rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT;
DRM_DEBUG("IH: D1 vblank\n");
}
@@ -6080,7 +6191,7 @@ restart_ih:
wake_up(&rdev->irq.vblank_queue);
}
if (atomic_read(&rdev->irq.pflip[1]))
- radeon_crtc_handle_flip(rdev, 1);
+ radeon_crtc_handle_vblank(rdev, 1);
rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT;
DRM_DEBUG("IH: D2 vblank\n");
}
@@ -6106,7 +6217,7 @@ restart_ih:
wake_up(&rdev->irq.vblank_queue);
}
if (atomic_read(&rdev->irq.pflip[2]))
- radeon_crtc_handle_flip(rdev, 2);
+ radeon_crtc_handle_vblank(rdev, 2);
rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT;
DRM_DEBUG("IH: D3 vblank\n");
}
@@ -6132,7 +6243,7 @@ restart_ih:
wake_up(&rdev->irq.vblank_queue);
}
if (atomic_read(&rdev->irq.pflip[3]))
- radeon_crtc_handle_flip(rdev, 3);
+ radeon_crtc_handle_vblank(rdev, 3);
rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT;
DRM_DEBUG("IH: D4 vblank\n");
}
@@ -6158,7 +6269,7 @@ restart_ih:
wake_up(&rdev->irq.vblank_queue);
}
if (atomic_read(&rdev->irq.pflip[4]))
- radeon_crtc_handle_flip(rdev, 4);
+ radeon_crtc_handle_vblank(rdev, 4);
rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT;
DRM_DEBUG("IH: D5 vblank\n");
}
@@ -6184,7 +6295,7 @@ restart_ih:
wake_up(&rdev->irq.vblank_queue);
}
if (atomic_read(&rdev->irq.pflip[5]))
- radeon_crtc_handle_flip(rdev, 5);
+ radeon_crtc_handle_vblank(rdev, 5);
rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT;
DRM_DEBUG("IH: D6 vblank\n");
}
@@ -6200,6 +6311,15 @@ restart_ih:
break;
}
break;
+ case 8: /* D1 page flip */
+ case 10: /* D2 page flip */
+ case 12: /* D3 page flip */
+ case 14: /* D4 page flip */
+ case 16: /* D5 page flip */
+ case 18: /* D6 page flip */
+ DRM_DEBUG("IH: D%d flip\n", ((src_id - 8) >> 1) + 1);
+ radeon_crtc_handle_flip(rdev, (src_id - 8) >> 1);
+ break;
case 42: /* HPD hotplug */
switch (src_data) {
case 0:
@@ -6249,18 +6369,24 @@ restart_ih:
break;
}
break;
+ case 124: /* UVD */
+ DRM_DEBUG("IH: UVD int: 0x%08x\n", src_data);
+ radeon_fence_process(rdev, R600_RING_TYPE_UVD_INDEX);
+ break;
case 146:
case 147:
addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR);
status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS);
+ /* reset addr and status */
+ WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
+ if (addr == 0x0 && status == 0x0)
+ break;
dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n",
addr);
dev_err(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
status);
si_vm_decode_fault(rdev, status, addr);
- /* reset addr and status */
- WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
break;
case 176: /* RINGID0 CP_INT */
radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
@@ -6331,80 +6457,6 @@ restart_ih:
return IRQ_HANDLED;
}
-/**
- * si_copy_dma - copy pages using the DMA engine
- *
- * @rdev: radeon_device pointer
- * @src_offset: src GPU address
- * @dst_offset: dst GPU address
- * @num_gpu_pages: number of GPU pages to xfer
- * @fence: radeon fence object
- *
- * Copy GPU paging using the DMA engine (SI).
- * Used by the radeon ttm implementation to move pages if
- * registered as the asic copy callback.
- */
-int si_copy_dma(struct radeon_device *rdev,
- uint64_t src_offset, uint64_t dst_offset,
- unsigned num_gpu_pages,
- struct radeon_fence **fence)
-{
- struct radeon_semaphore *sem = NULL;
- int ring_index = rdev->asic->copy.dma_ring_index;
- struct radeon_ring *ring = &rdev->ring[ring_index];
- u32 size_in_bytes, cur_size_in_bytes;
- int i, num_loops;
- int r = 0;
-
- r = radeon_semaphore_create(rdev, &sem);
- if (r) {
- DRM_ERROR("radeon: moving bo (%d).\n", r);
- return r;
- }
-
- size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
- num_loops = DIV_ROUND_UP(size_in_bytes, 0xfffff);
- r = radeon_ring_lock(rdev, ring, num_loops * 5 + 11);
- if (r) {
- DRM_ERROR("radeon: moving bo (%d).\n", r);
- radeon_semaphore_free(rdev, &sem, NULL);
- return r;
- }
-
- if (radeon_fence_need_sync(*fence, ring->idx)) {
- radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring,
- ring->idx);
- radeon_fence_note_sync(*fence, ring->idx);
- } else {
- radeon_semaphore_free(rdev, &sem, NULL);
- }
-
- for (i = 0; i < num_loops; i++) {
- cur_size_in_bytes = size_in_bytes;
- if (cur_size_in_bytes > 0xFFFFF)
- cur_size_in_bytes = 0xFFFFF;
- size_in_bytes -= cur_size_in_bytes;
- radeon_ring_write(ring, DMA_PACKET(DMA_PACKET_COPY, 1, 0, 0, cur_size_in_bytes));
- radeon_ring_write(ring, dst_offset & 0xffffffff);
- radeon_ring_write(ring, src_offset & 0xffffffff);
- radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff);
- radeon_ring_write(ring, upper_32_bits(src_offset) & 0xff);
- src_offset += cur_size_in_bytes;
- dst_offset += cur_size_in_bytes;
- }
-
- r = radeon_fence_emit(rdev, fence, ring->idx);
- if (r) {
- radeon_ring_unlock_undo(rdev, ring);
- return r;
- }
-
- radeon_ring_unlock_commit(rdev, ring);
- radeon_semaphore_free(rdev, &sem, *fence);
-
- return r;
-}
-
/*
* startup/shutdown callbacks
*/
@@ -6418,33 +6470,34 @@ static int si_startup(struct radeon_device *rdev)
/* enable aspm */
si_program_aspm(rdev);
- if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
- !rdev->rlc_fw || !rdev->mc_fw) {
- r = si_init_microcode(rdev);
- if (r) {
- DRM_ERROR("Failed to load firmware!\n");
- return r;
- }
- }
-
- r = si_mc_load_microcode(rdev);
- if (r) {
- DRM_ERROR("Failed to load MC firmware!\n");
- return r;
- }
-
+ /* scratch needs to be initialized before MC */
r = r600_vram_scratch_init(rdev);
if (r)
return r;
si_mc_program(rdev);
+
+ if (!rdev->pm.dpm_enabled) {
+ r = si_mc_load_microcode(rdev);
+ if (r) {
+ DRM_ERROR("Failed to load MC firmware!\n");
+ return r;
+ }
+ }
+
r = si_pcie_gart_enable(rdev);
if (r)
return r;
si_gpu_init(rdev);
/* allocate rlc buffers */
- r = si_rlc_init(rdev);
+ if (rdev->family == CHIP_VERDE) {
+ rdev->rlc.reg_list = verde_rlc_save_restore_register_list;
+ rdev->rlc.reg_list_size =
+ (u32)ARRAY_SIZE(verde_rlc_save_restore_register_list);
+ }
+ rdev->rlc.cs_data = si_cs_data;
+ r = sumo_rlc_init(rdev);
if (r) {
DRM_ERROR("Failed to init rlc BOs!\n");
return r;
@@ -6486,7 +6539,7 @@ static int si_startup(struct radeon_device *rdev)
}
if (rdev->has_uvd) {
- r = rv770_uvd_resume(rdev);
+ r = uvd_v2_2_resume(rdev);
if (!r) {
r = radeon_fence_driver_start_ring(rdev,
R600_RING_TYPE_UVD_INDEX);
@@ -6514,38 +6567,31 @@ static int si_startup(struct radeon_device *rdev)
ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
- CP_RB0_RPTR, CP_RB0_WPTR,
- 0, 0xfffff, RADEON_CP_PACKET2);
+ RADEON_CP_PACKET2);
if (r)
return r;
ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET,
- CP_RB1_RPTR, CP_RB1_WPTR,
- 0, 0xfffff, RADEON_CP_PACKET2);
+ RADEON_CP_PACKET2);
if (r)
return r;
ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET,
- CP_RB2_RPTR, CP_RB2_WPTR,
- 0, 0xfffff, RADEON_CP_PACKET2);
+ RADEON_CP_PACKET2);
if (r)
return r;
ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET,
- DMA_RB_RPTR + DMA0_REGISTER_OFFSET,
- DMA_RB_WPTR + DMA0_REGISTER_OFFSET,
- 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0));
+ DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0));
if (r)
return r;
ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET,
- DMA_RB_RPTR + DMA1_REGISTER_OFFSET,
- DMA_RB_WPTR + DMA1_REGISTER_OFFSET,
- 2, 0x3fffc, DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0));
+ DMA_PACKET(DMA_PACKET_NOP, 0, 0, 0, 0));
if (r)
return r;
@@ -6563,12 +6609,10 @@ static int si_startup(struct radeon_device *rdev)
if (rdev->has_uvd) {
ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
if (ring->ring_size) {
- r = radeon_ring_init(rdev, ring, ring->ring_size,
- R600_WB_UVD_RPTR_OFFSET,
- UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
- 0, 0xfffff, RADEON_CP_PACKET2);
+ r = radeon_ring_init(rdev, ring, ring->ring_size, 0,
+ RADEON_CP_PACKET2);
if (!r)
- r = r600_uvd_init(rdev);
+ r = uvd_v1_0_init(rdev);
if (r)
DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
}
@@ -6586,6 +6630,10 @@ static int si_startup(struct radeon_device *rdev)
return r;
}
+ r = dce6_audio_init(rdev);
+ if (r)
+ return r;
+
return 0;
}
@@ -6603,6 +6651,9 @@ int si_resume(struct radeon_device *rdev)
/* init golden registers */
si_init_golden_registers(rdev);
+ if (rdev->pm.pm_method == PM_METHOD_DPM)
+ radeon_pm_resume(rdev);
+
rdev->accel_working = true;
r = si_startup(rdev);
if (r) {
@@ -6617,13 +6668,17 @@ int si_resume(struct radeon_device *rdev)
int si_suspend(struct radeon_device *rdev)
{
+ radeon_pm_suspend(rdev);
+ dce6_audio_fini(rdev);
radeon_vm_manager_fini(rdev);
si_cp_enable(rdev, false);
cayman_dma_stop(rdev);
if (rdev->has_uvd) {
- r600_uvd_rbc_stop(rdev);
+ uvd_v1_0_fini(rdev);
radeon_uvd_suspend(rdev);
}
+ si_fini_pg(rdev);
+ si_fini_cg(rdev);
si_irq_suspend(rdev);
radeon_wb_disable(rdev);
si_pcie_gart_disable(rdev);
@@ -6687,6 +6742,18 @@ int si_init(struct radeon_device *rdev)
if (r)
return r;
+ if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
+ !rdev->rlc_fw || !rdev->mc_fw) {
+ r = si_init_microcode(rdev);
+ if (r) {
+ DRM_ERROR("Failed to load firmware!\n");
+ return r;
+ }
+ }
+
+ /* Initialize power management */
+ radeon_pm_init(rdev);
+
ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
ring->ring_obj = NULL;
r600_ring_init(rdev, ring, 1024 * 1024);
@@ -6730,7 +6797,7 @@ int si_init(struct radeon_device *rdev)
si_cp_fini(rdev);
cayman_dma_fini(rdev);
si_irq_fini(rdev);
- si_rlc_fini(rdev);
+ sumo_rlc_fini(rdev);
radeon_wb_fini(rdev);
radeon_ib_pool_fini(rdev);
radeon_vm_manager_fini(rdev);
@@ -6753,18 +6820,21 @@ int si_init(struct radeon_device *rdev)
void si_fini(struct radeon_device *rdev)
{
+ radeon_pm_fini(rdev);
si_cp_fini(rdev);
cayman_dma_fini(rdev);
- si_irq_fini(rdev);
- si_rlc_fini(rdev);
- si_fini_cg(rdev);
si_fini_pg(rdev);
+ si_fini_cg(rdev);
+ si_irq_fini(rdev);
+ sumo_rlc_fini(rdev);
radeon_wb_fini(rdev);
radeon_vm_manager_fini(rdev);
radeon_ib_pool_fini(rdev);
radeon_irq_kms_fini(rdev);
- if (rdev->has_uvd)
+ if (rdev->has_uvd) {
+ uvd_v1_0_fini(rdev);
radeon_uvd_fini(rdev);
+ }
si_pcie_gart_fini(rdev);
r600_vram_scratch_fini(rdev);
radeon_gem_fini(rdev);
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-10 20:30:44 +0000