From 605147f9d9bdebbcafe4bc2036ec36f1c0053224 Mon Sep 17 00:00:00 2001 From: xinhui pan Date: Thu, 31 Jan 2019 16:55:07 +0800 Subject: [PATCH 1514/2940] drm/amdgpu: add debugfs ctrl node allow userspace enable/disable ras Signed-off-by: xinhui pan Reviewed-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 122 ++++++++++++++++++++++-- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 9 ++ 2 files changed, 121 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 802a423badf3..604762263221 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -199,6 +199,90 @@ static const struct file_operations amdgpu_ras_debugfs_ops = { .llseek = default_llseek }; +/* + * DOC: ras debugfs control interface + * + * It accepts struct ras_debug_if who has two members. + * + * First member: ras_debug_if::head or ras_debug_if::inject. + * It is used to indicate which IP block will be under control. + * Its contents are not human readable, IOW, write it by your programs. + * + * head has four members, they are block, type, sub_block_index, name. + * block: which IP will be under control. + * type: what kind of error will be enabled/disabled/injected. + * sub_block_index: some IPs have subcomponets. say, GFX, sDMA. + * name: the name of IP. + * + * inject has two more members than head, they are address, value. + * As their names indicate, inject operation will write the + * value to the address. + * + * Second member: struct ras_debug_if::op. + * It has three kinds of operations. + * 0: disable RAS on the block. Take ::head as its data. + * 1: enable RAS on the block. Take ::head as its data. + * 2: inject errors on the block. Take ::inject as its data. + * + * How to check the result? + * + * For disable/enable, please check ras features at + * /sys/class/drm/card[0/1/2...]/device/ras/features + * + * For inject, please check corresponding err count at + * /sys/class/drm/card[0/1/2...]/device/ras/[gfx/sdma/...]_err_count + * + * NOTE: operation is only allowed on blocks which are supported. + * Please check ras mask at /sys/module/amdgpu/parameters/ras_mask + */ +static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, const char __user *buf, + size_t size, loff_t *pos) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)file_inode(f)->i_private; + struct ras_debug_if data; + int ret = 0; + + if (size < sizeof(data)) + return -EINVAL; + + memset(&data, 0, sizeof(data)); + + if (*pos) + return -EINVAL; + + if (copy_from_user(&data, buf, sizeof(data))) + return -EINVAL; + + *pos = size; + + if (!amdgpu_ras_is_supported(adev, data.head.block)) + return -EINVAL; + + switch (data.op) { + case 0: + ret = amdgpu_ras_feature_enable(adev, &data.head, 0); + break; + case 1: + ret = amdgpu_ras_feature_enable(adev, &data.head, 1); + break; + case 2: + ret = amdgpu_ras_error_inject(adev, &data.inject); + break; + }; + + if (ret) + return -EINVAL; + + return size; +} + +static const struct file_operations amdgpu_ras_debugfs_ctrl_ops = { + .owner = THIS_MODULE, + .read = NULL, + .write = amdgpu_ras_debugfs_ctrl_write, + .llseek = default_llseek +}; + static ssize_t amdgpu_ras_sysfs_read(struct device *dev, struct device_attribute *attr, char *buf) { @@ -656,6 +740,31 @@ static int amdgpu_ras_sysfs_remove_all(struct amdgpu_device *adev) /* sysfs end */ /* debugfs begin */ +static int amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *adev) +{ + struct amdgpu_ras *con = amdgpu_ras_get_context(adev); + struct drm_minor *minor = adev->ddev->primary; + struct dentry *root = minor->debugfs_root, *dir; + struct dentry *ent; + + dir = debugfs_create_dir("ras", root); + if (IS_ERR(dir)) + return -EINVAL; + + con->dir = dir; + + ent = debugfs_create_file("ras_ctrl", + S_IWUGO | S_IRUGO, con->dir, + adev, &amdgpu_ras_debugfs_ctrl_ops); + if (IS_ERR(ent)) { + debugfs_remove(con->dir); + return -EINVAL; + } + + con->ent = ent; + return 0; +} + int amdgpu_ras_debugfs_create(struct amdgpu_device *adev, struct ras_fs_if *head) { @@ -708,8 +817,10 @@ static int amdgpu_ras_debugfs_remove_all(struct amdgpu_device *adev) amdgpu_ras_debugfs_remove(adev, &obj->head); } + debugfs_remove(con->ent); debugfs_remove(con->dir); con->dir = NULL; + con->ent = NULL; return 0; } @@ -719,17 +830,8 @@ static int amdgpu_ras_debugfs_remove_all(struct amdgpu_device *adev) static int amdgpu_ras_fs_init(struct amdgpu_device *adev) { - struct amdgpu_ras *con = amdgpu_ras_get_context(adev); - struct drm_minor *minor = adev->ddev->primary; - struct dentry *root = minor->debugfs_root, *dir; - - dir = debugfs_create_dir("ras", root); - if (IS_ERR(dir)) - return -EINVAL; - - con->dir = dir; - amdgpu_ras_sysfs_create_feature_node(adev); + amdgpu_ras_debugfs_create_ctrl_node(adev); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h index d6d4bbb92a42..599dcd01aec7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h @@ -86,6 +86,8 @@ struct amdgpu_ras { struct list_head head; /* debugfs */ struct dentry *dir; + /* debugfs ctrl */ + struct dentry *ent; /* sysfs */ struct device_attribute features_attr; /* block array */ @@ -135,6 +137,13 @@ struct ras_dispatch_if { struct amdgpu_iv_entry *entry; }; +struct ras_debug_if { + union { + struct ras_common_if head; + struct ras_inject_if inject; + }; + int op; +}; /* work flow * vbios * 1: ras feature enable (enabled by default) -- 2.17.1