summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c165
1 files changed, 141 insertions, 24 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 0710b0b2e4b6..dfc10b1baea0 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -125,24 +125,39 @@ MODULE_FIRMWARE("amdgpu/fiji_mec2.bin");
MODULE_FIRMWARE("amdgpu/fiji_rlc.bin");
MODULE_FIRMWARE("amdgpu/polaris11_ce.bin");
+MODULE_FIRMWARE("amdgpu/polaris11_ce_2.bin");
MODULE_FIRMWARE("amdgpu/polaris11_pfp.bin");
+MODULE_FIRMWARE("amdgpu/polaris11_pfp_2.bin");
MODULE_FIRMWARE("amdgpu/polaris11_me.bin");
+MODULE_FIRMWARE("amdgpu/polaris11_me_2.bin");
MODULE_FIRMWARE("amdgpu/polaris11_mec.bin");
+MODULE_FIRMWARE("amdgpu/polaris11_mec_2.bin");
MODULE_FIRMWARE("amdgpu/polaris11_mec2.bin");
+MODULE_FIRMWARE("amdgpu/polaris11_mec2_2.bin");
MODULE_FIRMWARE("amdgpu/polaris11_rlc.bin");
MODULE_FIRMWARE("amdgpu/polaris10_ce.bin");
+MODULE_FIRMWARE("amdgpu/polaris10_ce_2.bin");
MODULE_FIRMWARE("amdgpu/polaris10_pfp.bin");
+MODULE_FIRMWARE("amdgpu/polaris10_pfp_2.bin");
MODULE_FIRMWARE("amdgpu/polaris10_me.bin");
+MODULE_FIRMWARE("amdgpu/polaris10_me_2.bin");
MODULE_FIRMWARE("amdgpu/polaris10_mec.bin");
+MODULE_FIRMWARE("amdgpu/polaris10_mec_2.bin");
MODULE_FIRMWARE("amdgpu/polaris10_mec2.bin");
+MODULE_FIRMWARE("amdgpu/polaris10_mec2_2.bin");
MODULE_FIRMWARE("amdgpu/polaris10_rlc.bin");
MODULE_FIRMWARE("amdgpu/polaris12_ce.bin");
+MODULE_FIRMWARE("amdgpu/polaris12_ce_2.bin");
MODULE_FIRMWARE("amdgpu/polaris12_pfp.bin");
+MODULE_FIRMWARE("amdgpu/polaris12_pfp_2.bin");
MODULE_FIRMWARE("amdgpu/polaris12_me.bin");
+MODULE_FIRMWARE("amdgpu/polaris12_me_2.bin");
MODULE_FIRMWARE("amdgpu/polaris12_mec.bin");
+MODULE_FIRMWARE("amdgpu/polaris12_mec_2.bin");
MODULE_FIRMWARE("amdgpu/polaris12_mec2.bin");
+MODULE_FIRMWARE("amdgpu/polaris12_mec2_2.bin");
MODULE_FIRMWARE("amdgpu/polaris12_rlc.bin");
static const struct amdgpu_gds_reg_offset amdgpu_gds_reg_offset[] =
@@ -918,8 +933,17 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
BUG();
}
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
- err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+ if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp_2.bin", chip_name);
+ err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+ if (err == -ENOENT) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
+ err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+ }
+ } else {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
+ err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+ }
if (err)
goto out;
err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
@@ -929,8 +953,17 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.pfp_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
- err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+ if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me_2.bin", chip_name);
+ err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+ if (err == -ENOENT) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
+ err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+ }
+ } else {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
+ err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+ }
if (err)
goto out;
err = amdgpu_ucode_validate(adev->gfx.me_fw);
@@ -941,8 +974,17 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
- err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+ if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce_2.bin", chip_name);
+ err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+ if (err == -ENOENT) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
+ err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+ }
+ } else {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
+ err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+ }
if (err)
goto out;
err = amdgpu_ucode_validate(adev->gfx.ce_fw);
@@ -1012,8 +1054,17 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
for (i = 0 ; i < (rlc_hdr->reg_list_size_bytes >> 2); i++)
adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
- err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+ if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec_2.bin", chip_name);
+ err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+ if (err == -ENOENT) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
+ err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+ }
+ } else {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
+ err = request_firmware(&adev->gfx.mec_fw, fw_name, adev->dev);
+ }
if (err)
goto out;
err = amdgpu_ucode_validate(adev->gfx.mec_fw);
@@ -1025,8 +1076,17 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
if ((adev->asic_type != CHIP_STONEY) &&
(adev->asic_type != CHIP_TOPAZ)) {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
- err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+ if (adev->asic_type >= CHIP_POLARIS10 && adev->asic_type <= CHIP_POLARIS12) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2_2.bin", chip_name);
+ err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+ if (err == -ENOENT) {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
+ err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+ }
+ } else {
+ snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
+ err = request_firmware(&adev->gfx.mec2_fw, fw_name, adev->dev);
+ }
if (!err) {
err = amdgpu_ucode_validate(adev->gfx.mec2_fw);
if (err)
@@ -2053,6 +2113,7 @@ static int gfx_v8_0_sw_fini(void *handle)
amdgpu_gfx_compute_mqd_sw_fini(adev);
amdgpu_gfx_kiq_free_ring(&adev->gfx.kiq.ring, &adev->gfx.kiq.irq);
amdgpu_gfx_kiq_fini(adev);
+ amdgpu_bo_free_kernel(&adev->virt.csa_obj, &adev->virt.csa_vmid0_addr, NULL);
gfx_v8_0_mec_fini(adev);
gfx_v8_0_rlc_fini(adev);
@@ -3707,6 +3768,8 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev)
ELEMENT_SIZE, 1);
sh_static_mem_cfg = REG_SET_FIELD(sh_static_mem_cfg, SH_STATIC_MEM_CONFIG,
INDEX_STRIDE, 3);
+ WREG32(mmSH_STATIC_MEM_CONFIG, sh_static_mem_cfg);
+
mutex_lock(&adev->srbm_mutex);
for (i = 0; i < adev->vm_manager.id_mgr[0].num_ids; i++) {
vi_srbm_select(adev, 0, 0, 0, i);
@@ -3730,7 +3793,6 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev)
WREG32(mmSH_MEM_APE1_BASE, 1);
WREG32(mmSH_MEM_APE1_LIMIT, 0);
- WREG32(mmSH_STATIC_MEM_CONFIG, sh_static_mem_cfg);
}
vi_srbm_select(adev, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
@@ -4576,12 +4638,10 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring)
mqd->compute_static_thread_mgmt_se2 = 0xffffffff;
mqd->compute_static_thread_mgmt_se3 = 0xffffffff;
mqd->compute_misc_reserved = 0x00000003;
- if (!(adev->flags & AMD_IS_APU)) {
- mqd->dynamic_cu_mask_addr_lo = lower_32_bits(ring->mqd_gpu_addr
- + offsetof(struct vi_mqd_allocation, dyamic_cu_mask));
- mqd->dynamic_cu_mask_addr_hi = upper_32_bits(ring->mqd_gpu_addr
- + offsetof(struct vi_mqd_allocation, dyamic_cu_mask));
- }
+ mqd->dynamic_cu_mask_addr_lo = lower_32_bits(ring->mqd_gpu_addr
+ + offsetof(struct vi_mqd_allocation, dynamic_cu_mask));
+ mqd->dynamic_cu_mask_addr_hi = upper_32_bits(ring->mqd_gpu_addr
+ + offsetof(struct vi_mqd_allocation, dynamic_cu_mask));
eop_base_addr = ring->eop_gpu_addr >> 8;
mqd->cp_hqd_eop_base_addr_lo = eop_base_addr;
mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
@@ -4752,7 +4812,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
gfx_v8_0_kiq_setting(ring);
- if (adev->gfx.in_reset) { /* for GPU_RESET case */
+ if (adev->in_sriov_reset) { /* for GPU_RESET case */
/* reset MQD to a clean status */
if (adev->gfx.mec.mqd_backup[mqd_idx])
memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
@@ -4767,8 +4827,8 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)
mutex_unlock(&adev->srbm_mutex);
} else {
memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation));
- ((struct vi_mqd_allocation *)mqd)->dyamic_cu_mask = 0xFFFFFFFF;
- ((struct vi_mqd_allocation *)mqd)->dyamic_rb_mask = 0xFFFFFFFF;
+ ((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
+ ((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
mutex_lock(&adev->srbm_mutex);
vi_srbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
gfx_v8_0_mqd_init(ring);
@@ -4789,10 +4849,10 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
struct vi_mqd *mqd = ring->mqd_ptr;
int mqd_idx = ring - &adev->gfx.compute_ring[0];
- if (!adev->gfx.in_reset && !adev->gfx.in_suspend) {
+ if (!adev->in_sriov_reset && !adev->gfx.in_suspend) {
memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation));
- ((struct vi_mqd_allocation *)mqd)->dyamic_cu_mask = 0xFFFFFFFF;
- ((struct vi_mqd_allocation *)mqd)->dyamic_rb_mask = 0xFFFFFFFF;
+ ((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
+ ((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
mutex_lock(&adev->srbm_mutex);
vi_srbm_select(adev, ring->me, ring->pipe, ring->queue, 0);
gfx_v8_0_mqd_init(ring);
@@ -4801,7 +4861,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
if (adev->gfx.mec.mqd_backup[mqd_idx])
memcpy(adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(struct vi_mqd_allocation));
- } else if (adev->gfx.in_reset) { /* for GPU_RESET case */
+ } else if (adev->in_sriov_reset) { /* for GPU_RESET case */
/* reset MQD to a clean status */
if (adev->gfx.mec.mqd_backup[mqd_idx])
memcpy(mqd, adev->gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));
@@ -4974,12 +5034,69 @@ static int gfx_v8_0_hw_init(void *handle)
return r;
}
+static int gfx_v8_0_kcq_disable(struct amdgpu_ring *kiq_ring,struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = kiq_ring->adev;
+ uint32_t scratch, tmp = 0;
+ int r, i;
+
+ r = amdgpu_gfx_scratch_get(adev, &scratch);
+ if (r) {
+ DRM_ERROR("Failed to get scratch reg (%d).\n", r);
+ return r;
+ }
+ WREG32(scratch, 0xCAFEDEAD);
+
+ r = amdgpu_ring_alloc(kiq_ring, 10);
+ if (r) {
+ DRM_ERROR("Failed to lock KIQ (%d).\n", r);
+ amdgpu_gfx_scratch_free(adev, scratch);
+ return r;
+ }
+
+ /* unmap queues */
+ amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_UNMAP_QUEUES, 4));
+ amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+ PACKET3_UNMAP_QUEUES_ACTION(1) | /* RESET_QUEUES */
+ PACKET3_UNMAP_QUEUES_QUEUE_SEL(0) |
+ PACKET3_UNMAP_QUEUES_ENGINE_SEL(0) |
+ PACKET3_UNMAP_QUEUES_NUM_QUEUES(1));
+ amdgpu_ring_write(kiq_ring, PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index));
+ amdgpu_ring_write(kiq_ring, 0);
+ amdgpu_ring_write(kiq_ring, 0);
+ amdgpu_ring_write(kiq_ring, 0);
+ /* write to scratch for completion */
+ amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+ amdgpu_ring_write(kiq_ring, (scratch - PACKET3_SET_UCONFIG_REG_START));
+ amdgpu_ring_write(kiq_ring, 0xDEADBEEF);
+ amdgpu_ring_commit(kiq_ring);
+
+ for (i = 0; i < adev->usec_timeout; i++) {
+ tmp = RREG32(scratch);
+ if (tmp == 0xDEADBEEF)
+ break;
+ DRM_UDELAY(1);
+ }
+ if (i >= adev->usec_timeout) {
+ DRM_ERROR("KCQ disabled failed (scratch(0x%04X)=0x%08X)\n", scratch, tmp);
+ r = -EINVAL;
+ }
+ amdgpu_gfx_scratch_free(adev, scratch);
+ return r;
+}
+
static int gfx_v8_0_hw_fini(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ int i;
amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
+
+ /* disable KCQ to avoid CPC touch memory not valid anymore */
+ for (i = 0; i < adev->gfx.num_compute_rings; i++)
+ gfx_v8_0_kcq_disable(&adev->gfx.kiq.ring, &adev->gfx.compute_ring[i]);
+
if (amdgpu_sriov_vf(adev)) {
pr_debug("For SRIOV client, shouldn't do anything.\n");
return 0;
generated by cgit v1.2.3-70-g09d2 (git 2.48.1) at 2025-05-03 05:49:38 +0000