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.c257
1 files changed, 173 insertions, 84 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 202ddda57f98..6a025438f9d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -914,7 +914,7 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = -EINVAL;
err2:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
err1:
amdgpu_device_wb_free(adev, index);
@@ -939,7 +939,6 @@ static void gfx_v8_0_free_microcode(struct amdgpu_device *adev)
static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
{
const char *chip_name;
- char fw_name[30];
int err;
struct amdgpu_firmware_info *info = NULL;
const struct common_firmware_header *header = NULL;
@@ -982,15 +981,18 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
}
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 = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s_pfp_2.bin", chip_name);
if (err == -ENODEV) {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_pfp.bin", chip_name);
}
} else {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_pfp.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_pfp.bin", chip_name);
}
if (err)
goto out;
@@ -999,15 +1001,18 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
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 = amdgpu_ucode_request(adev, &adev->gfx.me_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s_me_2.bin", chip_name);
if (err == -ENODEV) {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.me_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_me.bin", chip_name);
}
} else {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_me.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.me_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_me.bin", chip_name);
}
if (err)
goto out;
@@ -1017,15 +1022,18 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
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 = amdgpu_ucode_request(adev, &adev->gfx.ce_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s_ce_2.bin", chip_name);
if (err == -ENODEV) {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_ce.bin", chip_name);
}
} else {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_ce.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.ce_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_ce.bin", chip_name);
}
if (err)
goto out;
@@ -1044,8 +1052,9 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
} else
adev->virt.chained_ib_support = false;
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_rlc.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_rlc.bin", chip_name);
if (err)
goto out;
rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
@@ -1093,15 +1102,18 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
adev->gfx.rlc.register_restore[i] = le32_to_cpu(tmp[i]);
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 = amdgpu_ucode_request(adev, &adev->gfx.mec_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s_mec_2.bin", chip_name);
if (err == -ENODEV) {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mec.bin", chip_name);
}
} else {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mec.bin", chip_name);
}
if (err)
goto out;
@@ -1112,15 +1124,18 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
if ((adev->asic_type != CHIP_STONEY) &&
(adev->asic_type != CHIP_TOPAZ)) {
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 = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_OPTIONAL,
+ "amdgpu/%s_mec2_2.bin", chip_name);
if (err == -ENODEV) {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mec2.bin", chip_name);
}
} else {
- snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_mec2.bin", chip_name);
- err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw, fw_name);
+ err = amdgpu_ucode_request(adev, &adev->gfx.mec2_fw,
+ AMDGPU_UCODE_REQUIRED,
+ "amdgpu/%s_mec2.bin", chip_name);
}
if (!err) {
cp_hdr = (const struct gfx_firmware_header_v1_0 *)
@@ -1194,9 +1209,7 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
out:
if (err) {
- dev_err(adev->dev,
- "gfx8: Failed to load firmware \"%s\"\n",
- fw_name);
+ dev_err(adev->dev, "gfx8: Failed to load firmware %s gfx firmware\n", chip_name);
amdgpu_ucode_release(&adev->gfx.pfp_fw);
amdgpu_ucode_release(&adev->gfx.me_fw);
amdgpu_ucode_release(&adev->gfx.ce_fw);
@@ -1643,7 +1656,7 @@ static int gfx_v8_0_do_edc_gpr_workarounds(struct amdgpu_device *adev)
RREG32(sec_ded_counter_registers[i]);
fail:
- amdgpu_ib_free(adev, &ib, NULL);
+ amdgpu_ib_free(&ib, NULL);
dma_fence_put(f);
return r;
@@ -1897,12 +1910,12 @@ static int gfx_v8_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
static void gfx_v8_0_sq_irq_work_func(struct work_struct *work);
-static int gfx_v8_0_sw_init(void *handle)
+static int gfx_v8_0_sw_init(struct amdgpu_ip_block *ip_block)
{
int i, j, k, r, ring_id;
int xcc_id = 0;
struct amdgpu_ring *ring;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (adev->asic_type) {
case CHIP_TONGA:
@@ -2040,9 +2053,9 @@ static int gfx_v8_0_sw_init(void *handle)
return 0;
}
-static int gfx_v8_0_sw_fini(void *handle)
+static int gfx_v8_0_sw_fini(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int i;
for (i = 0; i < adev->gfx.num_gfx_rings; i++)
@@ -4263,7 +4276,7 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev)
ring->wptr = 0;
WREG32(mmCP_RB0_WPTR, lower_32_bits(ring->wptr));
- /* set the wb address wether it's enabled or not */
+ /* set the wb address whether it's enabled or not */
rptr_addr = ring->rptr_gpu_addr;
WREG32(mmCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
WREG32(mmCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
@@ -4307,9 +4320,7 @@ static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring)
tmp = RREG32(mmRLC_CP_SCHEDULERS);
tmp &= 0xffffff00;
tmp |= (ring->me << 5) | (ring->pipe << 3) | (ring->queue);
- WREG32(mmRLC_CP_SCHEDULERS, tmp);
- tmp |= 0x80;
- WREG32(mmRLC_CP_SCHEDULERS, tmp);
+ WREG32(mmRLC_CP_SCHEDULERS, tmp | 0x80);
}
static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev)
@@ -4786,10 +4797,10 @@ static void gfx_v8_0_cp_enable(struct amdgpu_device *adev, bool enable)
gfx_v8_0_cp_compute_enable(adev, enable);
}
-static int gfx_v8_0_hw_init(void *handle)
+static int gfx_v8_0_hw_init(struct amdgpu_ip_block *ip_block)
{
int r;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
gfx_v8_0_init_golden_registers(adev);
gfx_v8_0_constants_init(adev);
@@ -4826,6 +4837,13 @@ static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev)
amdgpu_ring_write(kiq_ring, 0);
amdgpu_ring_write(kiq_ring, 0);
}
+ /* Submit unmap queue packet */
+ amdgpu_ring_commit(kiq_ring);
+ /*
+ * Ring test will do a basic scratch register change check. Just run
+ * this to ensure that unmap queues that is submitted before got
+ * processed successfully before returning.
+ */
r = amdgpu_ring_test_helper(kiq_ring);
if (r)
DRM_ERROR("KCQ disable failed\n");
@@ -4868,13 +4886,13 @@ static int gfx_v8_0_wait_for_rlc_idle(void *handle)
return -ETIMEDOUT;
}
-static int gfx_v8_0_wait_for_idle(void *handle)
+static int gfx_v8_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
{
unsigned int i;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
for (i = 0; i < adev->usec_timeout; i++) {
- if (gfx_v8_0_is_idle(handle))
+ if (gfx_v8_0_is_idle(adev))
return 0;
udelay(1);
@@ -4882,9 +4900,9 @@ static int gfx_v8_0_wait_for_idle(void *handle)
return -ETIMEDOUT;
}
-static int gfx_v8_0_hw_fini(void *handle)
+static int gfx_v8_0_hw_fini(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
@@ -4900,8 +4918,9 @@ static int gfx_v8_0_hw_fini(void *handle)
pr_debug("For SRIOV client, shouldn't do anything.\n");
return 0;
}
+
amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
- if (!gfx_v8_0_wait_for_idle(adev))
+ if (!gfx_v8_0_wait_for_idle(ip_block))
gfx_v8_0_cp_enable(adev, false);
else
pr_err("cp is busy, skip halt cp\n");
@@ -4914,19 +4933,19 @@ static int gfx_v8_0_hw_fini(void *handle)
return 0;
}
-static int gfx_v8_0_suspend(void *handle)
+static int gfx_v8_0_suspend(struct amdgpu_ip_block *ip_block)
{
- return gfx_v8_0_hw_fini(handle);
+ return gfx_v8_0_hw_fini(ip_block);
}
-static int gfx_v8_0_resume(void *handle)
+static int gfx_v8_0_resume(struct amdgpu_ip_block *ip_block)
{
- return gfx_v8_0_hw_init(handle);
+ return gfx_v8_0_hw_init(ip_block);
}
-static bool gfx_v8_0_check_soft_reset(void *handle)
+static bool gfx_v8_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
u32 tmp;
@@ -4986,9 +5005,9 @@ static bool gfx_v8_0_check_soft_reset(void *handle)
}
}
-static int gfx_v8_0_pre_soft_reset(void *handle)
+static int gfx_v8_0_pre_soft_reset(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
u32 grbm_soft_reset = 0;
if ((!adev->gfx.grbm_soft_reset) &&
@@ -5027,9 +5046,9 @@ static int gfx_v8_0_pre_soft_reset(void *handle)
return 0;
}
-static int gfx_v8_0_soft_reset(void *handle)
+static int gfx_v8_0_soft_reset(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
u32 tmp;
@@ -5089,9 +5108,9 @@ static int gfx_v8_0_soft_reset(void *handle)
return 0;
}
-static int gfx_v8_0_post_soft_reset(void *handle)
+static int gfx_v8_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
u32 grbm_soft_reset = 0;
if ((!adev->gfx.grbm_soft_reset) &&
@@ -5257,9 +5276,9 @@ static const struct amdgpu_gfx_funcs gfx_v8_0_gfx_funcs = {
.select_me_pipe_q = &gfx_v8_0_select_me_pipe_q
};
-static int gfx_v8_0_early_init(void *handle)
+static int gfx_v8_0_early_init(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
adev->gfx.xcc_mask = 1;
adev->gfx.num_gfx_rings = GFX8_NUM_GFX_RINGS;
@@ -5274,9 +5293,9 @@ static int gfx_v8_0_early_init(void *handle)
return 0;
}
-static int gfx_v8_0_late_init(void *handle)
+static int gfx_v8_0_late_init(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int r;
r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
@@ -5316,7 +5335,7 @@ static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *ade
(adev->asic_type == CHIP_POLARIS12) ||
(adev->asic_type == CHIP_VEGAM))
/* Send msg to SMU via Powerplay */
- amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, enable);
+ amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_GFX, enable, 0);
WREG32_FIELD(RLC_PG_CNTL, STATIC_PER_CU_PG_ENABLE, enable ? 1 : 0);
}
@@ -5362,10 +5381,10 @@ static void cz_update_gfx_cg_power_gating(struct amdgpu_device *adev,
}
}
-static int gfx_v8_0_set_powergating_state(void *handle,
+static int gfx_v8_0_set_powergating_state(struct amdgpu_ip_block *ip_block,
enum amd_powergating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
bool enable = (state == AMD_PG_STATE_GATE);
if (amdgpu_sriov_vf(adev))
@@ -5620,8 +5639,6 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
{
uint32_t temp, data;
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
/* It is disabled by HW by default */
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) {
@@ -5715,8 +5732,6 @@ static void gfx_v8_0_update_medium_grain_clock_gating(struct amdgpu_device *adev
/* 7- wait for RLC_SERDES_CU_MASTER & RLC_SERDES_NONCU_MASTER idle */
gfx_v8_0_wait_for_rlc_serdes(adev);
}
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev,
@@ -5726,8 +5741,6 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
temp = data = RREG32(mmRLC_CGCG_CGLS_CTRL);
- amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
-
if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
temp1 = data1 = RREG32(mmRLC_CGTT_MGCG_OVERRIDE);
data1 &= ~RLC_CGTT_MGCG_OVERRIDE__CGCG_MASK;
@@ -5808,12 +5821,12 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev
}
gfx_v8_0_wait_for_rlc_serdes(adev);
-
- amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
}
static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
bool enable)
{
+ amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
+
if (enable) {
/* CGCG/CGLS should be enabled after MGCG/MGLS/TS(CG/LS)
* === MGCG + MGLS + TS(CG/LS) ===
@@ -5827,6 +5840,8 @@ static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev,
gfx_v8_0_update_coarse_grain_clock_gating(adev, enable);
gfx_v8_0_update_medium_grain_clock_gating(adev, enable);
}
+
+ amdgpu_gfx_rlc_exit_safe_mode(adev, 0);
return 0;
}
@@ -5977,10 +5992,10 @@ static int gfx_v8_0_polaris_update_gfx_clock_gating(struct amdgpu_device *adev,
return 0;
}
-static int gfx_v8_0_set_clockgating_state(void *handle,
+static int gfx_v8_0_set_clockgating_state(struct amdgpu_ip_block *ip_block,
enum amd_clockgating_state state)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
if (amdgpu_sriov_vf(adev))
return 0;
@@ -6152,6 +6167,7 @@ static void gfx_v8_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr,
{
bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+ bool exec = flags & AMDGPU_FENCE_FLAG_EXEC;
/* Workaround for cache flush problems. First send a dummy EOP
* event down the pipe with seq one below.
@@ -6175,7 +6191,8 @@ static void gfx_v8_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr,
EOP_TC_ACTION_EN |
EOP_TC_WB_ACTION_EN |
EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
- EVENT_INDEX(5)));
+ EVENT_INDEX(5) |
+ (exec ? EOP_EXEC : 0)));
amdgpu_ring_write(ring, addr & 0xfffffffc);
amdgpu_ring_write(ring, (upper_32_bits(addr) & 0xffff) |
DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0));
@@ -6383,6 +6400,34 @@ static void gfx_v8_0_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
amdgpu_ring_write(ring, val);
}
+static void gfx_v8_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
+ int mem_space, int opt, uint32_t addr0,
+ uint32_t addr1, uint32_t ref, uint32_t mask,
+ uint32_t inv)
+{
+ amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+ amdgpu_ring_write(ring,
+ /* memory (1) or register (0) */
+ (WAIT_REG_MEM_MEM_SPACE(mem_space) |
+ WAIT_REG_MEM_OPERATION(opt) | /* wait */
+ WAIT_REG_MEM_FUNCTION(3) | /* equal */
+ WAIT_REG_MEM_ENGINE(eng_sel)));
+
+ if (mem_space)
+ BUG_ON(addr0 & 0x3); /* Dword align */
+ amdgpu_ring_write(ring, addr0);
+ amdgpu_ring_write(ring, addr1);
+ amdgpu_ring_write(ring, ref);
+ amdgpu_ring_write(ring, mask);
+ amdgpu_ring_write(ring, inv); /* poll interval */
+}
+
+static void gfx_v8_0_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+ uint32_t val, uint32_t mask)
+{
+ gfx_v8_0_wait_reg_mem(ring, 0, 0, 0, reg, 0, val, mask, 0x20);
+}
+
static void gfx_v8_0_ring_soft_recovery(struct amdgpu_ring *ring, unsigned vmid)
{
struct amdgpu_device *adev = ring->adev;
@@ -6859,6 +6904,48 @@ static void gfx_v8_0_emit_wave_limit(struct amdgpu_ring *ring, bool enable)
}
+static int gfx_v8_0_reset_kgq(struct amdgpu_ring *ring, unsigned int vmid)
+{
+ struct amdgpu_device *adev = ring->adev;
+ struct amdgpu_kiq *kiq = &adev->gfx.kiq[0];
+ struct amdgpu_ring *kiq_ring = &kiq->ring;
+ unsigned long flags;
+ u32 tmp;
+ int r;
+
+ if (amdgpu_sriov_vf(adev))
+ return -EINVAL;
+
+ if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
+ return -EINVAL;
+
+ spin_lock_irqsave(&kiq->ring_lock, flags);
+
+ if (amdgpu_ring_alloc(kiq_ring, 5)) {
+ spin_unlock_irqrestore(&kiq->ring_lock, flags);
+ return -ENOMEM;
+ }
+
+ tmp = REG_SET_FIELD(0, CP_VMID_RESET, RESET_REQUEST, 1 << vmid);
+ gfx_v8_0_ring_emit_wreg(kiq_ring, mmCP_VMID_RESET, tmp);
+ amdgpu_ring_commit(kiq_ring);
+
+ spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+ r = amdgpu_ring_test_ring(kiq_ring);
+ if (r)
+ return r;
+
+ if (amdgpu_ring_alloc(ring, 7 + 12 + 5))
+ return -ENOMEM;
+ gfx_v8_0_ring_emit_fence_gfx(ring, ring->fence_drv.gpu_addr,
+ ring->fence_drv.sync_seq, AMDGPU_FENCE_FLAG_EXEC);
+ gfx_v8_0_ring_emit_reg_wait(ring, mmCP_VMID_RESET, 0, 0xffff);
+ gfx_v8_0_ring_emit_wreg(ring, mmCP_VMID_RESET, 0);
+
+ return amdgpu_ring_test_ring(ring);
+}
+
static const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
.name = "gfx_v8_0",
.early_init = gfx_v8_0_early_init,
@@ -6924,6 +7011,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = {
.emit_wreg = gfx_v8_0_ring_emit_wreg,
.soft_recovery = gfx_v8_0_ring_soft_recovery,
.emit_mem_sync = gfx_v8_0_emit_mem_sync,
+ .reset = gfx_v8_0_reset_kgq,
};
static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
@@ -6956,6 +7044,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = {
.insert_nop = amdgpu_ring_insert_nop,
.pad_ib = amdgpu_ring_generic_pad_ib,
.emit_wreg = gfx_v8_0_ring_emit_wreg,
+ .soft_recovery = gfx_v8_0_ring_soft_recovery,
.emit_mem_sync = gfx_v8_0_emit_mem_sync_compute,
.emit_wave_limit = gfx_v8_0_emit_wave_limit,
};
generated by cgit (git 2.34.1) at 2025-03-22 11:49:03 +0000