diff options
32 files changed, 775 insertions, 313 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 4a03a24348f4..7975f8e157df 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -989,6 +989,8 @@ struct amdgpu_device { char product_number[16]; char product_name[32]; char serial[16]; + + struct amdgpu_autodump autodump; }; static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index 1a4894fa3693..d33cb344be69 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -27,7 +27,7 @@ #include <linux/pci.h> #include <linux/uaccess.h> #include <linux/pm_runtime.h> - +#include <linux/poll.h> #include <drm/drm_debugfs.h> #include "amdgpu.h" @@ -74,8 +74,82 @@ int amdgpu_debugfs_add_files(struct amdgpu_device *adev, return 0; } +int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev) +{ +#if defined(CONFIG_DEBUG_FS) + unsigned long timeout = 600 * HZ; + int ret; + + wake_up_interruptible(&adev->autodump.gpu_hang); + + ret = wait_for_completion_interruptible_timeout(&adev->autodump.dumping, timeout); + if (ret == 0) { + pr_err("autodump: timeout, move on to gpu recovery\n"); + return -ETIMEDOUT; + } +#endif + return 0; +} + #if defined(CONFIG_DEBUG_FS) +static int amdgpu_debugfs_autodump_open(struct inode *inode, struct file *file) +{ + struct amdgpu_device *adev = inode->i_private; + int ret; + + file->private_data = adev; + + mutex_lock(&adev->lock_reset); + if (adev->autodump.dumping.done) { + reinit_completion(&adev->autodump.dumping); + ret = 0; + } else { + ret = -EBUSY; + } + mutex_unlock(&adev->lock_reset); + + return ret; +} + +static int amdgpu_debugfs_autodump_release(struct inode *inode, struct file *file) +{ + struct amdgpu_device *adev = file->private_data; + + complete_all(&adev->autodump.dumping); + return 0; +} + +static unsigned int amdgpu_debugfs_autodump_poll(struct file *file, struct poll_table_struct *poll_table) +{ + struct amdgpu_device *adev = file->private_data; + + poll_wait(file, &adev->autodump.gpu_hang, poll_table); + + if (adev->in_gpu_reset) + return POLLIN | POLLRDNORM | POLLWRNORM; + + return 0; +} + +static const struct file_operations autodump_debug_fops = { + .owner = THIS_MODULE, + .open = amdgpu_debugfs_autodump_open, + .poll = amdgpu_debugfs_autodump_poll, + .release = amdgpu_debugfs_autodump_release, +}; + +static void amdgpu_debugfs_autodump_init(struct amdgpu_device *adev) +{ + init_completion(&adev->autodump.dumping); + complete_all(&adev->autodump.dumping); + init_waitqueue_head(&adev->autodump.gpu_hang); + + debugfs_create_file("amdgpu_autodump", 0600, + adev->ddev->primary->debugfs_root, + adev, &autodump_debug_fops); +} + /** * amdgpu_debugfs_process_reg_op - Handle MMIO register reads/writes * @@ -1434,6 +1508,8 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev) amdgpu_ras_debugfs_create_all(adev); + amdgpu_debugfs_autodump_init(adev); + return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_list, ARRAY_SIZE(amdgpu_debugfs_list)); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h index de12d1101526..2803884d338d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h @@ -31,6 +31,11 @@ struct amdgpu_debugfs { unsigned num_files; }; +struct amdgpu_autodump { + struct completion dumping; + struct wait_queue_head gpu_hang; +}; + int amdgpu_debugfs_regs_init(struct amdgpu_device *adev); int amdgpu_debugfs_init(struct amdgpu_device *adev); void amdgpu_debugfs_fini(struct amdgpu_device *adev); @@ -40,3 +45,4 @@ int amdgpu_debugfs_add_files(struct amdgpu_device *adev, int amdgpu_debugfs_fence_init(struct amdgpu_device *adev); int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev); int amdgpu_debugfs_gem_init(struct amdgpu_device *adev); +int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index cc41e8f5ad14..545beebcf43e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3927,6 +3927,8 @@ static int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, int i, r = 0; bool need_full_reset = *need_full_reset_arg; + amdgpu_debugfs_wait_dump(adev); + /* block all schedulers and reset given job's ring */ for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c index 0e8018c9aa8e..d2a105e3bf7c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.c @@ -1188,3 +1188,13 @@ int amdgpu_dpm_set_df_cstate(struct amdgpu_device *adev, return ret; } + +int amdgpu_dpm_allow_xgmi_power_down(struct amdgpu_device *adev, bool en) +{ + struct smu_context *smu = &adev->smu; + + if (is_support_sw_smu(adev)) + return smu_allow_xgmi_power_down(smu, en); + + return 0; +}
\ No newline at end of file diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h index 936d85aa0fbc..956f6c710670 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h @@ -538,4 +538,6 @@ int amdgpu_dpm_baco_enter(struct amdgpu_device *adev); int amdgpu_dpm_set_df_cstate(struct amdgpu_device *adev, uint32_t cstate); +int amdgpu_dpm_allow_xgmi_power_down(struct amdgpu_device *adev, bool en); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index beb35dd12964..a0e5b54b6e47 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -86,9 +86,10 @@ * - 3.35.0 - Add drm_amdgpu_info_device::tcc_disabled_mask * - 3.36.0 - Allow reading more status registers on si/cik * - 3.37.0 - L2 is invalidated before SDMA IBs, needed for correctness + * - 3.38.0 - Add AMDGPU_IB_FLAG_EMIT_MEM_SYNC */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 37 +#define KMS_DRIVER_MINOR 38 #define KMS_DRIVER_PATCHLEVEL 0 int amdgpu_vram_limit = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index c24366aacf3a..b91853fd66d3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -189,6 +189,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, dma_fence_put(tmp); } + if ((ib->flags & AMDGPU_IB_FLAG_EMIT_MEM_SYNC) && ring->funcs->emit_mem_sync) + ring->funcs->emit_mem_sync(ring); + if (ring->funcs->insert_start) ring->funcs->insert_start(ring); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index c762deb5abc7..b75362bf0742 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -154,9 +154,9 @@ int amdgpu_dpm_read_sensor(struct amdgpu_device *adev, enum amd_pp_sensors senso * */ -static ssize_t amdgpu_get_dpm_state(struct device *dev, - struct device_attribute *attr, - char *buf) +static ssize_t amdgpu_get_power_dpm_state(struct device *dev, + struct device_attribute *attr, + char *buf) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -189,10 +189,10 @@ static ssize_t amdgpu_get_dpm_state(struct device *dev, (pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance"); } -static ssize_t amdgpu_set_dpm_state(struct device *dev, - struct device_attribute *attr, - const char *buf, - size_t count) +static ssize_t amdgpu_set_power_dpm_state(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -294,9 +294,9 @@ static ssize_t amdgpu_set_dpm_state(struct device *dev, * */ -static ssize_t amdgpu_get_dpm_forced_performance_level(struct device *dev, - struct device_attribute *attr, - char *buf) +static ssize_t amdgpu_get_power_dpm_force_performance_level(struct device *dev, + struct device_attribute *attr, + char *buf) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -332,10 +332,10 @@ static ssize_t amdgpu_get_dpm_forced_performance_level(struct device *dev, "unknown"); } -static ssize_t amdgpu_set_dpm_forced_performance_level(struct device *dev, - struct device_attribute *attr, - const char *buf, - size_t count) +static ssize_t amdgpu_set_power_dpm_force_performance_level(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -873,10 +873,10 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev, * the corresponding bit from original ppfeature masks and input the * new ppfeature masks. */ -static ssize_t amdgpu_set_pp_feature_status(struct device *dev, - struct device_attribute *attr, - const char *buf, - size_t count) +static ssize_t amdgpu_set_pp_features(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -917,9 +917,9 @@ static ssize_t amdgpu_set_pp_feature_status(struct device *dev, return count; } -static ssize_t amdgpu_get_pp_feature_status(struct device *dev, - struct device_attribute *attr, - char *buf) +static ssize_t amdgpu_get_pp_features(struct device *dev, + struct device_attribute *attr, + char *buf) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -1663,9 +1663,9 @@ static ssize_t amdgpu_set_pp_power_profile_mode(struct device *dev, * The SMU firmware computes a percentage of load based on the * aggregate activity level in the IP cores. */ -static ssize_t amdgpu_get_busy_percent(struct device *dev, - struct device_attribute *attr, - char *buf) +static ssize_t amdgpu_get_gpu_busy_percent(struct device *dev, + struct device_attribute *attr, + char *buf) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -1699,9 +1699,9 @@ static ssize_t amdgpu_get_busy_percent(struct device *dev, * The SMU firmware computes a percentage of load based on the * aggregate activity level in the IP cores. */ -static ssize_t amdgpu_get_memory_busy_percent(struct device *dev, - struct device_attribute *attr, - char *buf) +static ssize_t amdgpu_get_mem_busy_percent(struct device *dev, + struct device_attribute *attr, + char *buf) { struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; @@ -1790,57 +1790,174 @@ static ssize_t amdgpu_get_unique_id(struct device *dev, return 0; } -static DEVICE_ATTR(power_dpm_state, S_IRUGO | S_IWUSR, amdgpu_get_dpm_state, amdgpu_set_dpm_state); -static DEVICE_ATTR(power_dpm_force_performance_level, S_IRUGO | S_IWUSR, - amdgpu_get_dpm_forced_performance_level, - amdgpu_set_dpm_forced_performance_level); -static DEVICE_ATTR(pp_num_states, S_IRUGO, amdgpu_get_pp_num_states, NULL); -static DEVICE_ATTR(pp_cur_state, S_IRUGO, amdgpu_get_pp_cur_state, NULL); -static DEVICE_ATTR(pp_force_state, S_IRUGO | S_IWUSR, - amdgpu_get_pp_force_state, - amdgpu_set_pp_force_state); -static DEVICE_ATTR(pp_table, S_IRUGO | S_IWUSR, - amdgpu_get_pp_table, - amdgpu_set_pp_table); -static DEVICE_ATTR(pp_dpm_sclk, S_IRUGO | S_IWUSR, - amdgpu_get_pp_dpm_sclk, - amdgpu_set_pp_dpm_sclk); -static DEVICE_ATTR(pp_dpm_mclk, S_IRUGO | S_IWUSR, - amdgpu_get_pp_dpm_mclk, - amdgpu_set_pp_dpm_mclk); -static DEVICE_ATTR(pp_dpm_socclk, S_IRUGO | S_IWUSR, - amdgpu_get_pp_dpm_socclk, - amdgpu_set_pp_dpm_socclk); -static DEVICE_ATTR(pp_dpm_fclk, S_IRUGO | S_IWUSR, - amdgpu_get_pp_dpm_fclk, - amdgpu_set_pp_dpm_fclk); -static DEVICE_ATTR(pp_dpm_dcefclk, S_IRUGO | S_IWUSR, - amdgpu_get_pp_dpm_dcefclk, - amdgpu_set_pp_dpm_dcefclk); -static DEVICE_ATTR(pp_dpm_pcie, S_IRUGO | S_IWUSR, - amdgpu_get_pp_dpm_pcie, - amdgpu_set_pp_dpm_pcie); -static DEVICE_ATTR(pp_sclk_od, S_IRUGO | S_IWUSR, - amdgpu_get_pp_sclk_od, - amdgpu_set_pp_sclk_od); -static DEVICE_ATTR(pp_mclk_od, S_IRUGO | S_IWUSR, - amdgpu_get_pp_mclk_od, - amdgpu_set_pp_mclk_od); -static DEVICE_ATTR(pp_power_profile_mode, S_IRUGO | S_IWUSR, - amdgpu_get_pp_power_profile_mode, - amdgpu_set_pp_power_profile_mode); -static DEVICE_ATTR(pp_od_clk_voltage, S_IRUGO | S_IWUSR, - amdgpu_get_pp_od_clk_voltage, - amdgpu_set_pp_od_clk_voltage); -static DEVICE_ATTR(gpu_busy_percent, S_IRUGO, - amdgpu_get_busy_percent, NULL); -static DEVICE_ATTR(mem_busy_percent, S_IRUGO, - amdgpu_get_memory_busy_percent, NULL); -static DEVICE_ATTR(pcie_bw, S_IRUGO, amdgpu_get_pcie_bw, NULL); -static DEVICE_ATTR(pp_features, S_IRUGO | S_IWUSR, - amdgpu_get_pp_feature_status, - amdgpu_set_pp_feature_status); -static DEVICE_ATTR(unique_id, S_IRUGO, amdgpu_get_unique_id, NULL); +static struct amdgpu_device_attr amdgpu_device_attrs[] = { + AMDGPU_DEVICE_ATTR_RW(power_dpm_state, ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF), + AMDGPU_DEVICE_ATTR_RW(power_dpm_force_performance_level, ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF), + AMDGPU_DEVICE_ATTR_RO(pp_num_states, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RO(pp_cur_state, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_force_state, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_table, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_dpm_sclk, ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF), + AMDGPU_DEVICE_ATTR_RW(pp_dpm_mclk, ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF), + AMDGPU_DEVICE_ATTR_RW(pp_dpm_socclk, ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF), + AMDGPU_DEVICE_ATTR_RW(pp_dpm_fclk, ATTR_FLAG_BASIC|ATTR_FLAG_ONEVF), + AMDGPU_DEVICE_ATTR_RW(pp_dpm_dcefclk, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_dpm_pcie, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_sclk_od, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_mclk_od, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_power_profile_mode, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_od_clk_voltage, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RO(gpu_busy_percent, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RO(mem_busy_percent, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RO(pcie_bw, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RW(pp_features, ATTR_FLAG_BASIC), + AMDGPU_DEVICE_ATTR_RO(unique_id, ATTR_FLAG_BASIC), +}; + +static int default_attr_update(struct amdgpu_device *adev, struct amdgpu_device_attr *attr, + uint32_t mask) +{ + struct device_attribute *dev_attr = &attr->dev_attr; + const char *attr_name = dev_attr->attr.name; + struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; + enum amd_asic_type asic_type = adev->asic_type; + + if (!(attr->flags & mask)) { + attr->states = ATTR_STATE_UNSUPPORTED; + return 0; + } + +#define DEVICE_ATTR_IS(_name) (!strcmp(attr_name, #_name)) + + if (DEVICE_ATTR_IS(pp_dpm_socclk)) { + if (asic_type <= CHIP_VEGA10) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(pp_dpm_dcefclk)) { + if (asic_type <= CHIP_VEGA10 || asic_type == CHIP_ARCTURUS) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(pp_dpm_fclk)) { + if (asic_type < CHIP_VEGA20) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(pp_dpm_pcie)) { + if (asic_type == CHIP_ARCTURUS) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(pp_od_clk_voltage)) { + attr->states = ATTR_STATE_UNSUPPORTED; + if ((is_support_sw_smu(adev) && adev->smu.od_enabled) || + (!is_support_sw_smu(adev) && hwmgr->od_enabled)) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(mem_busy_percent)) { + if (adev->flags & AMD_IS_APU || asic_type == CHIP_VEGA10) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(pcie_bw)) { + /* PCIe Perf counters won't work on APU nodes */ + if (adev->flags & AMD_IS_APU) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(unique_id)) { + if (!adev->unique_id) + attr->states = ATTR_STATE_UNSUPPORTED; + } else if (DEVICE_ATTR_IS(pp_features)) { + if (adev->flags & AMD_IS_APU || asic_type <= CHIP_VEGA10) + attr->states = ATTR_STATE_UNSUPPORTED; + } + + if (asic_type == CHIP_ARCTURUS) { + /* Arcturus does not support standalone mclk/socclk/fclk level setting */ + if (DEVICE_ATTR_IS(pp_dpm_mclk) || + DEVICE_ATTR_IS(pp_dpm_socclk) || + DEVICE_ATTR_IS(pp_dpm_fclk)) { + dev_attr->attr.mode &= ~S_IWUGO; + dev_attr->store = NULL; + } + } + +#undef DEVICE_ATTR_IS + + return 0; +} + + +static int amdgpu_device_attr_create(struct amdgpu_device *adev, + struct amdgpu_device_attr *attr, + uint32_t mask) +{ + int ret = 0; + struct device_attribute *dev_attr = &attr->dev_attr; + const char *name = dev_attr->attr.name; + int (*attr_update)(struct amdgpu_device *adev, struct amdgpu_device_attr *attr, + uint32_t mask) = default_attr_update; + + BUG_ON(!attr); + + attr_update = attr->attr_update ? attr_update : default_attr_update; + + ret = attr_update(adev, attr, mask); + if (ret) { + dev_err(adev->dev, "failed to update device file %s, ret = %d\n", + name, ret); + return ret; + } + + /* the attr->states maybe changed after call attr->attr_update function */ + if (attr->states == ATTR_STATE_UNSUPPORTED) + return 0; + + ret = device_create_file(adev->dev, dev_attr); + if (ret) { + dev_err(adev->dev, "failed to create device file %s, ret = %d\n", + name, ret); + } + + attr->states = ATTR_STATE_SUPPORTED; + + return ret; +} + +static void amdgpu_device_attr_remove(struct amdgpu_device *adev, struct amdgpu_device_attr *attr) +{ + struct device_attribute *dev_attr = &attr->dev_attr; + + if (attr->states == ATTR_STATE_UNSUPPORTED) + return; + + device_remove_file(adev->dev, dev_attr); + + attr->states = ATTR_STATE_UNSUPPORTED; +} + +static int amdgpu_device_attr_create_groups(struct amdgpu_device *adev, + struct amdgpu_device_attr *attrs, + uint32_t counts, + uint32_t mask) +{ + int ret = 0; + uint32_t i = 0; + + for (i = 0; i < counts; i++) { + ret = amdgpu_device_attr_create(adev, &attrs[i], mask); + if (ret) + goto failed; + } + + return 0; + +failed: + for (; i > 0; i--) { + amdgpu_device_attr_remove(adev, &attrs[i]); + } + + return ret; +} + +static void amdgpu_device_attr_remove_groups(struct amdgpu_device *adev, + struct amdgpu_device_attr *attrs, + uint32_t counts) +{ + uint32_t i = 0; + + for (i = 0; i < counts; i++) + amdgpu_device_attr_remove(adev, &attrs[i]); +} static ssize_t amdgpu_hwmon_show_temp(struct device *dev, struct device_attribute *attr, @@ -3241,8 +3358,8 @@ int amdgpu_pm_load_smu_firmware(struct amdgpu_device *adev, uint32_t *smu_versio int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) { - struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; int ret; + uint32_t mask = 0; if (adev->pm.sysfs_initialized) return 0; @@ -3260,168 +3377,25 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) return ret; } - ret = device_create_file(adev->dev, &dev_attr_power_dpm_state); - if (ret) { - DRM_ERROR("failed to create device file for dpm state\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_power_dpm_force_performance_level); - if (ret) { - DRM_ERROR("failed to create device file for dpm state\n"); - return ret; - } - - if (!amdgpu_sriov_vf(adev)) { - ret = device_create_file(adev->dev, &dev_attr_pp_num_states); - if (ret) { - DRM_ERROR("failed to create device file pp_num_states\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_pp_cur_state); - if (ret) { - DRM_ERROR("failed to create device file pp_cur_state\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_pp_force_state); - if (ret) { - DRM_ERROR("failed to create device file pp_force_state\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_pp_table); - if (ret) { - DRM_ERROR("failed to create device file pp_table\n"); - return ret; - } - } - - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_sclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_sclk\n"); - return ret; - } - - /* Arcturus does not support standalone mclk/socclk/fclk level setting */ - if (adev->asic_type == CHIP_ARCTURUS) { - dev_attr_pp_dpm_mclk.attr.mode &= ~S_IWUGO; - dev_attr_pp_dpm_mclk.store = NULL; - - dev_attr_pp_dpm_socclk.attr.mode &= ~S_IWUGO; - dev_attr_pp_dpm_socclk.store = NULL; - - dev_attr_pp_dpm_fclk.attr.mode &= ~S_IWUGO; - dev_attr_pp_dpm_fclk.store = NULL; - } - - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_mclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_mclk\n"); - return ret; - } - if (adev->asic_type >= CHIP_VEGA10) { - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_socclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_socclk\n"); - return ret; - } - if (adev->asic_type != CHIP_ARCTURUS) { - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_dcefclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_dcefclk\n"); - return ret; - } - } - } - if (adev->asic_type >= CHIP_VEGA20) { - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_fclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_fclk\n"); - return ret; - } - } - - /* the reset are not needed for SRIOV one vf mode */ - if (amdgpu_sriov_vf(adev)) { - adev->pm.sysfs_initialized = true; - return ret; + switch (amdgpu_virt_get_sriov_vf_mode(adev)) { + case SRIOV_VF_MODE_ONE_VF: + mask = ATTR_FLAG_ONEVF; + break; + case SRIOV_VF_MODE_MULTI_VF: + mask = 0; + break; + case SRIOV_VF_MODE_BARE_METAL: + default: + mask = ATTR_FLAG_MASK_ALL; + break; } - if (adev->asic_type != CHIP_ARCTURUS) { - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_pcie\n"); - return ret; - } - } - ret = device_create_file(adev->dev, &dev_attr_pp_sclk_od); - if (ret) { - DRM_ERROR("failed to create device file pp_sclk_od\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_pp_mclk_od); - if (ret) { - DRM_ERROR("failed to create device file pp_mclk_od\n"); - return ret; - } - ret = device_create_file(adev->dev, - &dev_attr_pp_power_profile_mode); - if (ret) { - DRM_ERROR("failed to create device file " - "pp_power_profile_mode\n"); - return ret; - } - if ((is_support_sw_smu(adev) && adev->smu.od_enabled) || - (!is_support_sw_smu(adev) && hwmgr->od_enabled)) { - ret = device_create_file(adev->dev, - &dev_attr_pp_od_clk_voltage); - if (ret) { - DRM_ERROR("failed to create device file " - "pp_od_clk_voltage\n"); - return ret; - } - } - ret = device_create_file(adev->dev, - &dev_attr_gpu_busy_percent); - if (ret) { - DRM_ERROR("failed to create device file " - "gpu_busy_level\n"); - return ret; - } - /* APU does not have its own dedicated memory */ - if (!(adev->flags & AMD_IS_APU) && - (adev->asic_type != CHIP_VEGA10)) { - ret = device_create_file(adev->dev, - &dev_attr_mem_busy_percent); - if (ret) { - DRM_ERROR("failed to create device file " - "mem_busy_percent\n"); - return ret; - } - } - /* PCIe Perf counters won't work on APU nodes */ - if (!(adev->flags & AMD_IS_APU)) { - ret = device_create_file(adev->dev, &dev_attr_pcie_bw); - if (ret) { - DRM_ERROR("failed to create device file pcie_bw\n"); - return ret; - } - } - if (adev->unique_id) - ret = device_create_file(adev->dev, &dev_attr_unique_id); - if (ret) { - DRM_ERROR("failed to create device file unique_id\n"); + ret = amdgpu_device_attr_create_groups(adev, + amdgpu_device_attrs, + ARRAY_SIZE(amdgpu_device_attrs), + mask); + if (ret) return ret; - } - - if ((adev->asic_type >= CHIP_VEGA10) && - !(adev->flags & AMD_IS_APU)) { - ret = device_create_file(adev->dev, - &dev_attr_pp_features); - if (ret) { - DRM_ERROR("failed to create device file " - "pp_features\n"); - return ret; - } - } adev->pm.sysfs_initialized = true; @@ -3430,51 +3404,15 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) { - struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; - if (adev->pm.dpm_enabled == 0) return; if (adev->pm.int_hwmon_dev) hwmon_device_unregister(adev->pm.int_hwmon_dev); - device_remove_file(adev->dev, &dev_attr_power_dpm_state); - device_remove_file(adev->dev, &dev_attr_power_dpm_force_performance_level); - - device_remove_file(adev->dev, &dev_attr_pp_num_states); - device_remove_file(adev->dev, &dev_attr_pp_cur_state); - device_remove_file(adev->dev, &dev_attr_pp_force_state); - device_remove_file(adev->dev, &dev_attr_pp_table); - - device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk); - device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk); - if (adev->asic_type >= CHIP_VEGA10) { - device_remove_file(adev->dev, &dev_attr_pp_dpm_socclk); - if (adev->asic_type != CHIP_ARCTURUS) - device_remove_file(adev->dev, &dev_attr_pp_dpm_dcefclk); - } - if (adev->asic_type != CHIP_ARCTURUS) - device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie); - if (adev->asic_type >= CHIP_VEGA20) - device_remove_file(adev->dev, &dev_attr_pp_dpm_fclk); - device_remove_file(adev->dev, &dev_attr_pp_sclk_od); - device_remove_file(adev->dev, &dev_attr_pp_mclk_od); - device_remove_file(adev->dev, - &dev_attr_pp_power_profile_mode); - if ((is_support_sw_smu(adev) && adev->smu.od_enabled) || - (!is_support_sw_smu(adev) && hwmgr->od_enabled)) - device_remove_file(adev->dev, - &dev_attr_pp_od_clk_voltage); - device_remove_file(adev->dev, &dev_attr_gpu_busy_percent); - if (!(adev->flags & AMD_IS_APU) && - (adev->asic_type != CHIP_VEGA10)) - device_remove_file(adev->dev, &dev_attr_mem_busy_percent); - if (!(adev->flags & AMD_IS_APU)) - device_remove_file(adev->dev, &dev_attr_pcie_bw); - if (adev->unique_id) - device_remove_file(adev->dev, &dev_attr_unique_id); - if ((adev->asic_type >= CHIP_VEGA10) && - !(adev->flags & AMD_IS_APU)) - device_remove_file(adev->dev, &dev_attr_pp_features); + + amdgpu_device_attr_remove_groups(adev, + amdgpu_device_attrs, + ARRAY_SIZE(amdgpu_device_attrs)); } void amdgpu_pm_compute_clocks(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.h index 5db0ef86e84c..48e8086baf33 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.h @@ -30,6 +30,52 @@ struct cg_flag_name const char *name; }; +enum amdgpu_device_attr_flags { + ATTR_FLAG_BASIC = (1 << 0), + ATTR_FLAG_ONEVF = (1 << 16), +}; + +#define ATTR_FLAG_TYPE_MASK (0x0000ffff) +#define ATTR_FLAG_MODE_MASK (0xffff0000) +#define ATTR_FLAG_MASK_ALL (0xffffffff) + +enum amdgpu_device_attr_states { + ATTR_STATE_UNSUPPORTED = 0, + ATTR_STATE_SUPPORTED, +}; + +struct amdgpu_device_attr { + struct device_attribute dev_attr; + enum amdgpu_device_attr_flags flags; + enum amdgpu_device_attr_states states; + int (*attr_update)(struct amdgpu_device *adev, + struct amdgpu_device_attr* attr, + uint32_t mask); +}; + +#define to_amdgpu_device_attr(_dev_attr) \ + container_of(_dev_attr, struct amdgpu_device_attr, dev_attr) + +#define __AMDGPU_DEVICE_ATTR(_name, _mode, _show, _store, _flags, ...) \ + { .dev_attr = __ATTR(_name, _mode, _show, _store), \ + .flags = _flags, \ + .states = ATTR_STATE_SUPPORTED, \ + ##__VA_ARGS__, } + +#define AMDGPU_DEVICE_ATTR(_name, _mode, _flags, ...) \ + __AMDGPU_DEVICE_ATTR(_name, _mode, \ + amdgpu_get_##_name, amdgpu_set_##_name, \ + _flags, ##__VA_ARGS__) + +#define AMDGPU_DEVICE_ATTR_RW(_name, _flags, ...) \ + AMDGPU_DEVICE_ATTR(_name, S_IRUGO | S_IWUSR, \ + _flags, ##__VA_ARGS__) + +#define AMDGPU_DEVICE_ATTR_RO(_name, _flags, ...) \ + __AMDGPU_DEVICE_ATTR(_name, S_IRUGO, \ + amdgpu_get_##_name, NULL, \ + _flags, ##__VA_ARGS__) + void amdgpu_pm_acpi_event_handler(struct amdgpu_device *adev); int amdgpu_pm_sysfs_init(struct amdgpu_device *adev); int amdgpu_pm_virt_sysfs_init(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 7348619253c7..50fe08bf2f72 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -811,6 +811,32 @@ int amdgpu_ras_error_query(struct amdgpu_device *adev, return 0; } +/* Trigger XGMI/WAFL error */ +int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev, + struct ta_ras_trigger_error_input *block_info) +{ + int ret; + + if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW)) + dev_warn(adev->dev, "Failed to disallow df cstate"); + + if (amdgpu_dpm_allow_xgmi_power_down(adev, false)) + dev_warn(adev->dev, "Failed to disallow XGMI power down"); + + ret = psp_ras_trigger_error(&adev->psp, block_info); + + if (amdgpu_ras_intr_triggered()) + return ret; + + if (amdgpu_dpm_allow_xgmi_power_down(adev, true)) + dev_warn(adev->dev, "Failed to allow XGMI power down"); + + if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW)) + dev_warn(adev->dev, "Failed to allow df cstate"); + + return ret; +} + /* wrapper of psp_ras_trigger_error */ int amdgpu_ras_error_inject(struct amdgpu_device *adev, struct ras_inject_if *info) @@ -844,10 +870,12 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev, break; case AMDGPU_RAS_BLOCK__UMC: case AMDGPU_RAS_BLOCK__MMHUB: - case AMDGPU_RAS_BLOCK__XGMI_WAFL: case AMDGPU_RAS_BLOCK__PCIE_BIF: ret = psp_ras_trigger_error(&adev->psp, &block_info); break; + case AMDGPU_RAS_BLOCK__XGMI_WAFL: + ret = amdgpu_ras_error_inject_xgmi(adev, &block_info); + break; default: dev_info(adev->dev, "%s error injection is not supported yet\n", ras_block_str(info->head.block)); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index 107e80063553..be218754629a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -195,6 +195,7 @@ struct amdgpu_ring_funcs { /* Try to soft recover the ring to make the fence signal */ void (*soft_recovery)(struct amdgpu_ring *ring, unsigned vmid); int (*preempt_ib)(struct amdgpu_ring *ring); + void (*emit_mem_sync)(struct amdgpu_ring *ring); }; struct amdgpu_ring { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index cbbb8d02535a..f3b38c9e04ca 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -370,3 +370,19 @@ void amdgpu_virt_disable_access_debugfs(struct amdgpu_device *adev) if (amdgpu_sriov_vf(adev)) adev->virt.caps |= AMDGPU_SRIOV_CAPS_RUNTIME; } + +enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *adev) +{ + enum amdgpu_sriov_vf_mode mode; + + if (amdgpu_sriov_vf(adev)) { + if (amdgpu_sriov_is_pp_one_vf(adev)) + mode = SRIOV_VF_MODE_ONE_VF; + else + mode = SRIOV_VF_MODE_MULTI_VF; + } else { + mode = SRIOV_VF_MODE_BARE_METAL; + } + + return mode; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h index de27308802c9..b90e822cebd7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h @@ -35,6 +35,12 @@ /* tonga/fiji use this offset */ #define mmBIF_IOV_FUNC_IDENTIFIER 0x1503 +enum amdgpu_sriov_vf_mode { + SRIOV_VF_MODE_BARE_METAL = 0, + SRIOV_VF_MODE_ONE_VF, + SRIOV_VF_MODE_MULTI_VF, +}; + struct amdgpu_mm_table { struct amdgpu_bo *bo; uint32_t *cpu_addr; @@ -323,4 +329,6 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev); bool amdgpu_virt_can_access_debugfs(struct amdgpu_device *adev); int amdgpu_virt_enable_access_debugfs(struct amdgpu_device *adev); void amdgpu_virt_disable_access_debugfs(struct amdgpu_device *adev); + +enum amdgpu_sriov_vf_mode amdgpu_virt_get_sriov_vf_mode(struct amdgpu_device *adev); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index 90610b4f2c75..e9e59bc68c9f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -441,7 +441,7 @@ out: int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_device *adev) { - int ret = -EINVAL; + int ret; /* Each psp need to set the latest topology */ ret = psp_xgmi_set_topology_info(&adev->psp, diff --git a/drivers/gpu/drm/amd/amdgpu/cikd.h b/drivers/gpu/drm/amd/amdgpu/cikd.h index cee6e8a3ad9c..5f3f6ebfb387 100644 --- a/drivers/gpu/drm/amd/amdgpu/cikd.h +++ b/drivers/gpu/drm/amd/amdgpu/cikd.h @@ -450,7 +450,7 @@ # define PACKET3_DMA_DATA_CMD_SAIC (1 << 28) # define PACKET3_DMA_DATA_CMD_DAIC (1 << 29) # define PACKET3_DMA_DATA_CMD_RAW_WAIT (1 << 30) -#define PACKET3_AQUIRE_MEM 0x58 +#define PACKET3_ACQUIRE_MEM 0x58 #define PACKET3_REWIND 0x59 #define PACKET3_LOAD_UCONFIG_REG 0x5E #define PACKET3_LOAD_SH_REG 0x5F diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c index 64080d209223..bd5dd4f64311 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c @@ -4577,13 +4577,11 @@ static int gfx_v10_0_init_csb(struct amdgpu_device *adev) adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr); /* csib */ - /* amdgpu_mm_wreg_mmio_rlc will fall back to mmio if doesn't support rlcg_write */ - amdgpu_mm_wreg_mmio_rlc(adev, SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI), - adev->gfx.rlc.clear_state_gpu_addr >> 32, 0); - amdgpu_mm_wreg_mmio_rlc(adev, SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_LO), - adev->gfx.rlc.clear_state_gpu_addr & 0xfffffffc, 0); - amdgpu_mm_wreg_mmio_rlc(adev, SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_LENGTH), - adev->gfx.rlc.clear_state_size, 0); + WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_ADDR_HI, + adev->gfx.rlc.clear_state_gpu_addr >> 32); + WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_ADDR_LO, + adev->gfx.rlc.clear_state_gpu_addr & 0xfffffffc); + WREG32_SOC15_RLC(GC, 0, mmRLC_CSIB_LENGTH, adev->gfx.rlc.clear_state_size); return 0; } @@ -5192,7 +5190,7 @@ static int gfx_v10_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable) tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_HALT, enable ? 0 : 1); tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, enable ? 0 : 1); tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_HALT, enable ? 0 : 1); - amdgpu_mm_wreg_mmio_rlc(adev, SOC15_REG_OFFSET(GC, 0, mmCP_ME_CNTL), tmp, 0); + WREG32_SOC15_RLC(GC, 0, mmCP_ME_CNTL, tmp); for (i = 0; i < adev->usec_timeout; i++) { if (RREG32_SOC15(GC, 0, mmCP_STAT) == 0) @@ -8022,6 +8020,29 @@ static int gfx_v10_0_kiq_irq(struct amdgpu_device *adev, return 0; } +static void gfx_v10_0_emit_mem_sync(struct amdgpu_ring *ring) +{ + const unsigned int gcr_cntl = + PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_INV(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_WB(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_INV(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_WB(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GL1_INV(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GLV_INV(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GLK_INV(1) | + PACKET3_ACQUIRE_MEM_GCR_CNTL_GLI_INV(1); + + /* ACQUIRE_MEM - make one or more surfaces valid for use by the subsequent operations */ + amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 6)); + amdgpu_ring_write(ring, 0); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0xffffff); /* CP_COHER_SIZE_HI */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE_HI */ + amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */ + amdgpu_ring_write(ring, gcr_cntl); /* GCR_CNTL */ +} + static const struct amd_ip_funcs gfx_v10_0_ip_funcs = { .name = "gfx_v10_0", .early_init = gfx_v10_0_early_init, @@ -8069,7 +8090,8 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = { 3 + /* CNTX_CTRL */ 5 + /* HDP_INVL */ 8 + 8 + /* FENCE x2 */ - 2, /* SWITCH_BUFFER */ + 2 + /* SWITCH_BUFFER */ + 8, /* gfx_v10_0_emit_mem_sync */ .emit_ib_size = 4, /* gfx_v10_0_ring_emit_ib_gfx */ .emit_ib = gfx_v10_0_ring_emit_ib_gfx, .emit_fence = gfx_v10_0_ring_emit_fence, @@ -8091,6 +8113,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_gfx = { .emit_reg_wait = gfx_v10_0_ring_emit_reg_wait, .emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v10_0_ring_soft_recovery, + .emit_mem_sync = gfx_v10_0_emit_mem_sync, }; static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { @@ -8110,7 +8133,8 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + 2 + /* gfx_v10_0_ring_emit_vm_flush */ - 8 + 8 + 8, /* gfx_v10_0_ring_emit_fence x3 for user fence, vm fence */ + 8 + 8 + 8 + /* gfx_v10_0_ring_emit_fence x3 for user fence, vm fence */ + 8, /* gfx_v10_0_emit_mem_sync */ .emit_ib_size = 7, /* gfx_v10_0_ring_emit_ib_compute */ .emit_ib = gfx_v10_0_ring_emit_ib_compute, .emit_fence = gfx_v10_0_ring_emit_fence, @@ -8125,6 +8149,7 @@ static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_compute = { .emit_wreg = gfx_v10_0_ring_emit_wreg, .emit_reg_wait = gfx_v10_0_ring_emit_reg_wait, .emit_reg_write_reg_wait = gfx_v10_0_ring_emit_reg_write_reg_wait, + .emit_mem_sync = gfx_v10_0_emit_mem_sync, }; static const struct amdgpu_ring_funcs gfx_v10_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c index aa1e1be852dd..79c52c7a02e3 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c @@ -3465,6 +3465,18 @@ static int gfx_v6_0_set_powergating_state(void *handle, return 0; } +static void gfx_v6_0_emit_mem_sync(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + amdgpu_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0x0000000A); /* poll interval */ +} + static const struct amd_ip_funcs gfx_v6_0_ip_funcs = { .name = "gfx_v6_0", .early_init = gfx_v6_0_early_init, @@ -3495,7 +3507,8 @@ static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_gfx = { 14 + 14 + 14 + /* gfx_v6_0_ring_emit_fence x3 for user fence, vm fence */ 7 + 4 + /* gfx_v6_0_ring_emit_pipeline_sync */ SI_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + 6 + /* gfx_v6_0_ring_emit_vm_flush */ - 3 + 2, /* gfx_v6_ring_emit_cntxcntl including vgt flush */ + 3 + 2 + /* gfx_v6_ring_emit_cntxcntl including vgt flush */ + 5, /* SURFACE_SYNC */ .emit_ib_size = 6, /* gfx_v6_0_ring_emit_ib */ .emit_ib = gfx_v6_0_ring_emit_ib, .emit_fence = gfx_v6_0_ring_emit_fence, @@ -3506,6 +3519,7 @@ static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_gfx = { .insert_nop = amdgpu_ring_insert_nop, .emit_cntxcntl = gfx_v6_ring_emit_cntxcntl, .emit_wreg = gfx_v6_0_ring_emit_wreg, + .emit_mem_sync = gfx_v6_0_emit_mem_sync, }; static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_compute = { @@ -3519,7 +3533,8 @@ static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_compute = { 5 + 5 + /* hdp flush / invalidate */ 7 + /* gfx_v6_0_ring_emit_pipeline_sync */ SI_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + /* gfx_v6_0_ring_emit_vm_flush */ - 14 + 14 + 14, /* gfx_v6_0_ring_emit_fence x3 for user fence, vm fence */ + 14 + 14 + 14 + /* gfx_v6_0_ring_emit_fence x3 for user fence, vm fence */ + 5, /* SURFACE_SYNC */ .emit_ib_size = 6, /* gfx_v6_0_ring_emit_ib */ .emit_ib = gfx_v6_0_ring_emit_ib, .emit_fence = gfx_v6_0_ring_emit_fence, @@ -3529,6 +3544,7 @@ static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_compute = { .test_ib = gfx_v6_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, .emit_wreg = gfx_v6_0_ring_emit_wreg, + .emit_mem_sync = gfx_v6_0_emit_mem_sync, }; static void gfx_v6_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index e5a88cad44cb..0cc011f9190d 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -4998,6 +4998,32 @@ static int gfx_v7_0_set_powergating_state(void *handle, return 0; } +static void gfx_v7_0_emit_mem_sync(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + amdgpu_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0x0000000A); /* poll interval */ +} + +static void gfx_v7_0_emit_mem_sync_compute(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 5)); + amdgpu_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0xff); /* CP_COHER_SIZE_HI */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE_HI */ + amdgpu_ring_write(ring, 0x0000000A); /* poll interval */ +} + static const struct amd_ip_funcs gfx_v7_0_ip_funcs = { .name = "gfx_v7_0", .early_init = gfx_v7_0_early_init, @@ -5030,7 +5056,8 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = { 12 + 12 + 12 + /* gfx_v7_0_ring_emit_fence_gfx x3 for user fence, vm fence */ 7 + 4 + /* gfx_v7_0_ring_emit_pipeline_sync */ CIK_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + 6 + /* gfx_v7_0_ring_emit_vm_flush */ - 3 + 4, /* gfx_v7_ring_emit_cntxcntl including vgt flush*/ + 3 + 4 + /* gfx_v7_ring_emit_cntxcntl including vgt flush*/ + 5, /* SURFACE_SYNC */ .emit_ib_size = 4, /* gfx_v7_0_ring_emit_ib_gfx */ .emit_ib = gfx_v7_0_ring_emit_ib_gfx, .emit_fence = gfx_v7_0_ring_emit_fence_gfx, @@ -5045,6 +5072,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = { .emit_cntxcntl = gfx_v7_ring_emit_cntxcntl, .emit_wreg = gfx_v7_0_ring_emit_wreg, .soft_recovery = gfx_v7_0_ring_soft_recovery, + .emit_mem_sync = gfx_v7_0_emit_mem_sync, }; static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { @@ -5061,7 +5089,8 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { 5 + /* hdp invalidate */ 7 + /* gfx_v7_0_ring_emit_pipeline_sync */ CIK_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + /* gfx_v7_0_ring_emit_vm_flush */ - 7 + 7 + 7, /* gfx_v7_0_ring_emit_fence_compute x3 for user fence, vm fence */ + 7 + 7 + 7 + /* gfx_v7_0_ring_emit_fence_compute x3 for user fence, vm fence */ + 7, /* gfx_v7_0_emit_mem_sync_compute */ .emit_ib_size = 7, /* gfx_v7_0_ring_emit_ib_compute */ .emit_ib = gfx_v7_0_ring_emit_ib_compute, .emit_fence = gfx_v7_0_ring_emit_fence_compute, @@ -5074,6 +5103,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, .emit_wreg = gfx_v7_0_ring_emit_wreg, + .emit_mem_sync = gfx_v7_0_emit_mem_sync_compute, }; static void gfx_v7_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 2fcf6865abba..1d4128227ffd 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -6817,6 +6817,34 @@ static int gfx_v8_0_sq_irq(struct amdgpu_device *adev, return 0; } +static void gfx_v8_0_emit_mem_sync(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + amdgpu_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA | + PACKET3_TC_WB_ACTION_ENA); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0x0000000A); /* poll interval */ +} + +static void gfx_v8_0_emit_mem_sync_compute(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 5)); + amdgpu_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA | + PACKET3_TC_WB_ACTION_ENA); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0xff); /* CP_COHER_SIZE_HI */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE_HI */ + amdgpu_ring_write(ring, 0x0000000A); /* poll interval */ +} + static const struct amd_ip_funcs gfx_v8_0_ip_funcs = { .name = "gfx_v8_0", .early_init = gfx_v8_0_early_init, @@ -6863,7 +6891,8 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = { 3 + /* CNTX_CTRL */ 5 + /* HDP_INVL */ 12 + 12 + /* FENCE x2 */ - 2, /* SWITCH_BUFFER */ + 2 + /* SWITCH_BUFFER */ + 5, /* SURFACE_SYNC */ .emit_ib_size = 4, /* gfx_v8_0_ring_emit_ib_gfx */ .emit_ib = gfx_v8_0_ring_emit_ib_gfx, .emit_fence = gfx_v8_0_ring_emit_fence_gfx, @@ -6881,6 +6910,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = { .patch_cond_exec = gfx_v8_0_ring_emit_patch_cond_exec, .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, }; static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { @@ -6897,7 +6927,8 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { 5 + /* hdp_invalidate */ 7 + /* gfx_v8_0_ring_emit_pipeline_sync */ VI_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + /* gfx_v8_0_ring_emit_vm_flush */ - 7 + 7 + 7, /* gfx_v8_0_ring_emit_fence_compute x3 for user fence, vm fence */ + 7 + 7 + 7 + /* gfx_v8_0_ring_emit_fence_compute x3 for user fence, vm fence */ + 7, /* gfx_v8_0_emit_mem_sync_compute */ .emit_ib_size = 7, /* gfx_v8_0_ring_emit_ib_compute */ .emit_ib = gfx_v8_0_ring_emit_ib_compute, .emit_fence = gfx_v8_0_ring_emit_fence_compute, @@ -6910,6 +6941,7 @@ 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, + .emit_mem_sync = gfx_v8_0_emit_mem_sync_compute, }; static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index a0988634aeaf..1573ac1f03b2 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -6634,6 +6634,25 @@ static int gfx_v9_0_query_ras_error_count(struct amdgpu_device *adev, return 0; } +static void gfx_v9_0_emit_mem_sync(struct amdgpu_ring *ring) +{ + const unsigned int cp_coher_cntl = + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_ICACHE_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TCL1_ACTION_ENA(1) | + PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_WB_ACTION_ENA(1); + + /* ACQUIRE_MEM -make one or more surfaces valid for use by the subsequent operations */ + amdgpu_ring_write(ring, PACKET3(PACKET3_ACQUIRE_MEM, 5)); + amdgpu_ring_write(ring, cp_coher_cntl); /* CP_COHER_CNTL */ + amdgpu_ring_write(ring, 0xffffffff); /* CP_COHER_SIZE */ + amdgpu_ring_write(ring, 0xffffff); /* CP_COHER_SIZE_HI */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE */ + amdgpu_ring_write(ring, 0); /* CP_COHER_BASE_HI */ + amdgpu_ring_write(ring, 0x0000000A); /* POLL_INTERVAL */ +} + static const struct amd_ip_funcs gfx_v9_0_ip_funcs = { .name = "gfx_v9_0", .early_init = gfx_v9_0_early_init, @@ -6680,7 +6699,8 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = { 3 + /* CNTX_CTRL */ 5 + /* HDP_INVL */ 8 + 8 + /* FENCE x2 */ - 2, /* SWITCH_BUFFER */ + 2 + /* SWITCH_BUFFER */ + 7, /* gfx_v9_0_emit_mem_sync */ .emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_gfx */ .emit_ib = gfx_v9_0_ring_emit_ib_gfx, .emit_fence = gfx_v9_0_ring_emit_fence, @@ -6701,6 +6721,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = { .emit_reg_wait = gfx_v9_0_ring_emit_reg_wait, .emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait, .soft_recovery = gfx_v9_0_ring_soft_recovery, + .emit_mem_sync = gfx_v9_0_emit_mem_sync, }; static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { @@ -6720,7 +6741,8 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + 2 + /* gfx_v9_0_ring_emit_vm_flush */ - 8 + 8 + 8, /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */ + 8 + 8 + 8 + /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */ + 7, /* gfx_v9_0_emit_mem_sync */ .emit_ib_size = 7, /* gfx_v9_0_ring_emit_ib_compute */ .emit_ib = gfx_v9_0_ring_emit_ib_compute, .emit_fence = gfx_v9_0_ring_emit_fence, @@ -6735,6 +6757,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { .emit_wreg = gfx_v9_0_ring_emit_wreg, .emit_reg_wait = gfx_v9_0_ring_emit_reg_wait, .emit_reg_write_reg_wait = gfx_v9_0_ring_emit_reg_write_reg_wait, + .emit_mem_sync = gfx_v9_0_emit_mem_sync, }; static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = { diff --git a/drivers/gpu/drm/amd/amdgpu/nvd.h b/drivers/gpu/drm/amd/amdgpu/nvd.h index f3d8771ebed4..fd6b58243b03 100644 --- a/drivers/gpu/drm/amd/amdgpu/nvd.h +++ b/drivers/gpu/drm/amd/amdgpu/nvd.h @@ -256,6 +256,54 @@ #define PACKET3_BLK_CNTX_UPDATE 0x53 #define PACKET3_INCR_UPDT_STATE 0x55 #define PACKET3_ACQUIRE_MEM 0x58 +/* 1. HEADER + * 2. COHER_CNTL [30:0] + * 2.1 ENGINE_SEL [31:31] + * 2. COHER_SIZE [31:0] + * 3. COHER_SIZE_HI [7:0] + * 4. COHER_BASE_LO [31:0] + * 5. COHER_BASE_HI [23:0] + * 7. POLL_INTERVAL [15:0] + * 8. GCR_CNTL [18:0] + */ +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GLI_INV(x) ((x) << 0) + /* + * 0:NOP + * 1:ALL + * 2:RANGE + * 3:FIRST_LAST + */ +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL1_RANGE(x) ((x) << 2) + /* + * 0:ALL + * 1:reserved + * 2:RANGE + * 3:FIRST_LAST + */ +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_WB(x) ((x) << 4) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GLM_INV(x) ((x) << 5) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GLK_WB(x) ((x) << 6) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GLK_INV(x) ((x) << 7) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GLV_INV(x) ((x) << 8) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL1_INV(x) ((x) << 9) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_US(x) ((x) << 10) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_RANGE(x) ((x) << 11) + /* + * 0:ALL + * 1:VOL + * 2:RANGE + * 3:FIRST_LAST + */ +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_DISCARD(x) ((x) << 13) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_INV(x) ((x) << 14) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_GL2_WB(x) ((x) << 15) +#define PACKET3_ACQUIRE_MEM_GCR_CNTL_SEQ(x) ((x) << 16) + /* + * 0: PARALLEL + * 1: FORWARD + * 2: REVERSE + */ +#define PACKET3_ACQUIRE_MEM_GCR_RANGE_IS_PA (1 << 18) #define PACKET3_REWIND 0x59 #define PACKET3_INTERRUPT 0x5A #define PACKET3_GEN_PDEPTE 0x5B diff --git a/drivers/gpu/drm/amd/amdgpu/soc15d.h b/drivers/gpu/drm/amd/amdgpu/soc15d.h index 295d68c5811d..799925d22fc8 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15d.h +++ b/drivers/gpu/drm/amd/amdgpu/soc15d.h @@ -253,7 +253,30 @@ # define PACKET3_DMA_DATA_CMD_SAIC (1 << 28) # define PACKET3_DMA_DATA_CMD_DAIC (1 << 29) # define PACKET3_DMA_DATA_CMD_RAW_WAIT (1 << 30) -#define PACKET3_AQUIRE_MEM 0x58 +#define PACKET3_ACQUIRE_MEM 0x58 +/* 1. HEADER + * 2. COHER_CNTL [30:0] + * 2.1 ENGINE_SEL [31:31] + * 3. COHER_SIZE [31:0] + * 4. COHER_SIZE_HI [7:0] + * 5. COHER_BASE_LO [31:0] + * 6. COHER_BASE_HI [23:0] + * 7. POLL_INTERVAL [15:0] + */ +/* COHER_CNTL fields for CP_COHER_CNTL */ +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_NC_ACTION_ENA(x) ((x) << 3) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_WC_ACTION_ENA(x) ((x) << 4) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_INV_METADATA_ACTION_ENA(x) ((x) << 5) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TCL1_VOL_ACTION_ENA(x) ((x) << 15) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_WB_ACTION_ENA(x) ((x) << 18) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TCL1_ACTION_ENA(x) ((x) << 22) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_TC_ACTION_ENA(x) ((x) << 23) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_CB_ACTION_ENA(x) ((x) << 25) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_DB_ACTION_ENA(x) ((x) << 26) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_ACTION_ENA(x) ((x) << 27) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_VOL_ACTION_ENA(x) ((x) << 28) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_ICACHE_ACTION_ENA(x) ((x) << 29) +#define PACKET3_ACQUIRE_MEM_CP_COHER_CNTL_SH_KCACHE_WB_ACTION_ENA(x) ((x) << 30) #define PACKET3_REWIND 0x59 #define PACKET3_LOAD_UCONFIG_REG 0x5E #define PACKET3_LOAD_SH_REG 0x5F diff --git a/drivers/gpu/drm/amd/amdgpu/vid.h b/drivers/gpu/drm/amd/amdgpu/vid.h index 19ddd2312e00..7a01e6133798 100644 --- a/drivers/gpu/drm/amd/amdgpu/vid.h +++ b/drivers/gpu/drm/amd/amdgpu/vid.h @@ -332,7 +332,7 @@ # define PACKET3_DMA_DATA_CMD_SAIC (1 << 28) # define PACKET3_DMA_DATA_CMD_DAIC (1 << 29) # define PACKET3_DMA_DATA_CMD_RAW_WAIT (1 << 30) -#define PACKET3_AQUIRE_MEM 0x58 +#define PACKET3_ACQUIRE_MEM 0x58 #define PACKET3_REWIND 0x59 #define PACKET3_LOAD_UCONFIG_REG 0x5E #define PACKET3_LOAD_SH_REG 0x5F diff --git a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c index de14542de775..8c684a6e0156 100644 --- a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c @@ -2100,6 +2100,28 @@ int smu_set_df_cstate(struct smu_context *smu, return ret; } +int smu_allow_xgmi_power_down(struct smu_context *smu, bool en) +{ + struct amdgpu_device *adev = smu->adev; + int ret = 0; + + if (!adev->pm.dpm_enabled) + return -EINVAL; + + if (!smu->ppt_funcs || !smu->ppt_funcs->allow_xgmi_power_down) + return 0; + + mutex_lock(&smu->mutex); + + ret = smu->ppt_funcs->allow_xgmi_power_down(smu, en); + if (ret) + pr_err("[AllowXgmiPowerDown] failed!\n"); + + mutex_unlock(&smu->mutex); + + return ret; +} + int smu_write_watermarks_table(struct smu_context *smu) { void *watermarks_table = smu->smu_table.watermarks_table; diff --git a/drivers/gpu/drm/amd/powerplay/arcturus_ppt.c b/drivers/gpu/drm/amd/powerplay/arcturus_ppt.c index cfae4bcaf32e..cbf70122de9b 100644 --- a/drivers/gpu/drm/amd/powerplay/arcturus_ppt.c +++ b/drivers/gpu/drm/amd/powerplay/arcturus_ppt.c @@ -128,6 +128,7 @@ static struct smu_11_0_cmn2aisc_mapping arcturus_message_map[SMU_MSG_MAX_COUNT] MSG_MAP(SetXgmiMode, PPSMC_MSG_SetXgmiMode), MSG_MAP(SetMemoryChannelEnable, PPSMC_MSG_SetMemoryChannelEnable), MSG_MAP(DFCstateControl, PPSMC_MSG_DFCstateControl), + MSG_MAP(GmiPwrDnControl, PPSMC_MSG_GmiPwrDnControl), }; static struct smu_11_0_cmn2aisc_mapping arcturus_clk_map[SMU_CLK_COUNT] = { @@ -2286,6 +2287,35 @@ static int arcturus_set_df_cstate(struct smu_context *smu, return smu_send_smc_msg_with_param(smu, SMU_MSG_DFCstateControl, state, NULL); } +static int arcturus_allow_xgmi_power_down(struct smu_context *smu, bool en) +{ + uint32_t smu_version; + int ret; + + ret = smu_get_smc_version(smu, NULL, &smu_version); + if (ret) { + pr_err("Failed to get smu version!\n"); + return ret; + } + + /* PPSMC_MSG_GmiPwrDnControl is supported by 54.23.0 and onwards */ + if (smu_version < 0x00361700) { + pr_err("XGMI power down control is only supported by PMFW 54.23.0 and onwards\n"); + return -EINVAL; + } + + if (en) + return smu_send_smc_msg_with_param(smu, + SMU_MSG_GmiPwrDnControl, + 1, + NULL); + + return smu_send_smc_msg_with_param(smu, + SMU_MSG_GmiPwrDnControl, + 0, + NULL); +} + static const struct pptable_funcs arcturus_ppt_funcs = { /* translate smu index into arcturus specific index */ .get_smu_msg_index = arcturus_get_smu_msg_index, @@ -2379,6 +2409,7 @@ static const struct pptable_funcs arcturus_ppt_funcs = { .override_pcie_parameters = smu_v11_0_override_pcie_parameters, .get_pptable_power_limit = arcturus_get_pptable_power_limit, .set_df_cstate = arcturus_set_df_cstate, + .allow_xgmi_power_down = arcturus_allow_xgmi_power_down, }; void arcturus_set_ppt_funcs(struct smu_context *smu) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c index 91b5d96db674..4279f95ba779 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu_helper.c @@ -22,6 +22,7 @@ */ #include <linux/pci.h> +#include <linux/reboot.h> #include "hwmgr.h" #include "pp_debug.h" @@ -595,37 +596,61 @@ int phm_irq_process(struct amdgpu_device *adev, uint32_t src_id = entry->src_id; if (client_id == AMDGPU_IRQ_CLIENTID_LEGACY) { - if (src_id == VISLANDS30_IV_SRCID_CG_TSS_THERMAL_LOW_TO_HIGH) + if (src_id == VISLANDS30_IV_SRCID_CG_TSS_THERMAL_LOW_TO_HIGH) { pr_warn("GPU over temperature range detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); - else if (src_id == VISLANDS30_IV_SRCID_CG_TSS_THERMAL_HIGH_TO_LOW) + /* + * SW CTF just occurred. + * Try to do a graceful shutdown to prevent further damage. + */ + dev_emerg(adev->dev, "System is going to shutdown due to SW CTF!\n"); + orderly_poweroff(true); + } else if (src_id == VISLANDS30_IV_SRCID_CG_TSS_THERMAL_HIGH_TO_LOW) pr_warn("GPU under temperature range detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); - else if (src_id == VISLANDS30_IV_SRCID_GPIO_19) + else if (src_id == VISLANDS30_IV_SRCID_GPIO_19) { pr_warn("GPU Critical Temperature Fault detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); + /* + * HW CTF just occurred. Shutdown to prevent further damage. + */ + dev_emerg(adev->dev, "System is going to shutdown due to HW CTF!\n"); + orderly_poweroff(true); + } } else if (client_id == SOC15_IH_CLIENTID_THM) { - if (src_id == 0) + if (src_id == 0) { pr_warn("GPU over temperature range detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); - else + /* + * SW CTF just occurred. + * Try to do a graceful shutdown to prevent further damage. + */ + dev_emerg(adev->dev, "System is going to shutdown due to SW CTF!\n"); + orderly_poweroff(true); + } else pr_warn("GPU under temperature range detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); - } else if (client_id == SOC15_IH_CLIENTID_ROM_SMUIO) + } else if (client_id == SOC15_IH_CLIENTID_ROM_SMUIO) { pr_warn("GPU Critical Temperature Fault detected on PCIe %d:%d.%d!\n", PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); + /* + * HW CTF just occurred. Shutdown to prevent further damage. + */ + dev_emerg(adev->dev, "System is going to shutdown due to HW CTF!\n"); + orderly_poweroff(true); + } return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h index 928eed220f93..4d1c2a44a8b6 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h +++ b/drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h @@ -491,6 +491,7 @@ struct pptable_funcs { int (*get_dpm_clk_limited)(struct smu_context *smu, enum smu_clk_type clk_type, uint32_t dpm_level, uint32_t *freq); int (*set_df_cstate)(struct smu_context *smu, enum pp_df_cstate state); + int (*allow_xgmi_power_down)(struct smu_context *smu, bool en); int (*update_pcie_parameters)(struct smu_context *smu, uint32_t pcie_gen_cap, uint32_t pcie_width_cap); int (*i2c_eeprom_init)(struct i2c_adapter *control); void (*i2c_eeprom_fini)(struct i2c_adapter *control); @@ -731,6 +732,7 @@ int smu_set_mp1_state(struct smu_context *smu, enum pp_mp1_state mp1_state); int smu_set_df_cstate(struct smu_context *smu, enum pp_df_cstate state); +int smu_allow_xgmi_power_down(struct smu_context *smu, bool en); int smu_get_max_sustainable_clocks_by_dc(struct smu_context *smu, struct pp_smu_nv_clock_table *max_clocks); diff --git a/drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h b/drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h index f736d773f9d6..e07478b6ac04 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h +++ b/drivers/gpu/drm/amd/powerplay/inc/arcturus_ppsmc.h @@ -114,7 +114,8 @@ #define PPSMC_MSG_SetNumBadHbmPagesRetired 0x3A #define PPSMC_MSG_DFCstateControl 0x3B -#define PPSMC_Message_Count 0x3C +#define PPSMC_MSG_GmiPwrDnControl 0x3D +#define PPSMC_Message_Count 0x3E typedef uint32_t PPSMC_Result; typedef uint32_t PPSMC_Msg; diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu_types.h b/drivers/gpu/drm/amd/powerplay/inc/smu_types.h index a5b4df146713..ee7dac4693d4 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/smu_types.h +++ b/drivers/gpu/drm/amd/powerplay/inc/smu_types.h @@ -170,6 +170,7 @@ __SMU_DUMMY_MAP(SetSoftMinJpeg), \ __SMU_DUMMY_MAP(SetHardMinFclkByFreq), \ __SMU_DUMMY_MAP(DFCstateControl), \ + __SMU_DUMMY_MAP(GmiPwrDnControl), \ __SMU_DUMMY_MAP(DAL_DISABLE_DUMMY_PSTATE_CHANGE), \ __SMU_DUMMY_MAP(DAL_ENABLE_DUMMY_PSTATE_CHANGE), \ diff --git a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c index d5a3487ccfac..3391d1c4bbbe 100644 --- a/drivers/gpu/drm/amd/powerplay/smu_v11_0.c +++ b/drivers/gpu/drm/amd/powerplay/smu_v11_0.c @@ -23,6 +23,7 @@ #include <linux/firmware.h> #include <linux/module.h> #include <linux/pci.h> +#include <linux/reboot.h> #define SMU_11_0_PARTIAL_PPTABLE @@ -1547,12 +1548,19 @@ static int smu_v11_0_ack_ac_dc_interrupt(struct smu_context *smu) #define THM_11_0__SRCID__THM_DIG_THERM_L2H 0 /* ASIC_TEMP > CG_THERMAL_INT.DIG_THERM_INTH */ #define THM_11_0__SRCID__THM_DIG_THERM_H2L 1 /* ASIC_TEMP < CG_THERMAL_INT.DIG_THERM_INTL */ +#define SMUIO_11_0__SRCID__SMUIO_GPIO19 83 + static int smu_v11_0_irq_process(struct amdgpu_device *adev, struct amdgpu_irq_src *source, struct amdgpu_iv_entry *entry) { uint32_t client_id = entry->client_id; uint32_t src_id = entry->src_id; + /* + * ctxid is used to distinguish different + * events for SMCToHost interrupt. + */ + uint32_t ctxid = entry->src_data[0]; if (client_id == SOC15_IH_CLIENTID_THM) { switch (src_id) { @@ -1561,6 +1569,12 @@ static int smu_v11_0_irq_process(struct amdgpu_device *adev, PCI_BUS_NUM(adev->pdev->devfn), PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); + /* + * SW CTF just occurred. + * Try to do a graceful shutdown to prevent further damage. + */ + dev_emerg(adev->dev, "System is going to shutdown due to SW CTF!\n"); + orderly_poweroff(true); break; case THM_11_0__SRCID__THM_DIG_THERM_H2L: pr_warn("GPU under temperature range detected on PCIe %d:%d.%d!\n", @@ -1575,11 +1589,30 @@ static int smu_v11_0_irq_process(struct amdgpu_device *adev, PCI_SLOT(adev->pdev->devfn), PCI_FUNC(adev->pdev->devfn)); break; - } + } else if (client_id == SOC15_IH_CLIENTID_ROM_SMUIO) { + pr_warn("GPU Critical Temperature Fault detected on PCIe %d:%d.%d!\n", + PCI_BUS_NUM(adev->pdev->devfn), + PCI_SLOT(adev->pdev->devfn), + PCI_FUNC(adev->pdev->devfn)); + /* + * HW CTF just occurred. Shutdown to prevent further damage. + */ + dev_emerg(adev->dev, "System is going to shutdown due to HW CTF!\n"); + orderly_poweroff(true); } else if (client_id == SOC15_IH_CLIENTID_MP1) { - if (src_id == 0xfe) - smu_v11_0_ack_ac_dc_interrupt(&adev->smu); + if (src_id == 0xfe) { + switch (ctxid) { + case 0x3: + dev_dbg(adev->dev, "Switched to AC mode!\n"); + smu_v11_0_ack_ac_dc_interrupt(&adev->smu); + break; + case 0x4: + dev_dbg(adev->dev, "Switched to DC mode!\n"); + smu_v11_0_ack_ac_dc_interrupt(&adev->smu); + break; + } + } } return 0; @@ -1619,6 +1652,13 @@ int smu_v11_0_register_irq_handler(struct smu_context *smu) if (ret) return ret; + /* Register CTF(GPIO_19) interrupt */ + ret = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_ROM_SMUIO, + SMUIO_11_0__SRCID__SMUIO_GPIO19, + irq_src); + if (ret) + return ret; + ret = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_MP1, 0xfe, irq_src); diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h index e01b673f0449..4e873dcbe68f 100644 --- a/include/uapi/drm/amdgpu_drm.h +++ b/include/uapi/drm/amdgpu_drm.h @@ -602,6 +602,10 @@ union drm_amdgpu_cs { */ #define AMDGPU_IB_FLAGS_SECURE (1 << 5) +/* Tell KMD to flush and invalidate caches + */ +#define AMDGPU_IB_FLAG_EMIT_MEM_SYNC (1 << 6) + struct drm_amdgpu_cs_chunk_ib { __u32 _pad; /** AMDGPU_IB_FLAG_* */ |