diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c')
-rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 1338 |
1 files changed, 1202 insertions, 136 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index ea00090b3fb3..3993e6134914 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -47,27 +47,42 @@ #include "jpeg_v2_5.h" #include "smuio_v9_0.h" #include "gmc_v10_0.h" +#include "gmc_v11_0.h" #include "gfxhub_v2_0.h" #include "mmhub_v2_0.h" #include "nbio_v2_3.h" +#include "nbio_v4_3.h" #include "nbio_v7_2.h" +#include "nbio_v7_7.h" #include "hdp_v5_0.h" +#include "hdp_v5_2.h" +#include "hdp_v6_0.h" #include "nv.h" +#include "soc21.h" #include "navi10_ih.h" +#include "ih_v6_0.h" #include "gfx_v10_0.h" +#include "gfx_v11_0.h" #include "sdma_v5_0.h" #include "sdma_v5_2.h" +#include "sdma_v6_0.h" +#include "lsdma_v6_0.h" #include "vcn_v2_0.h" #include "jpeg_v2_0.h" #include "vcn_v3_0.h" #include "jpeg_v3_0.h" +#include "vcn_v4_0.h" +#include "jpeg_v4_0.h" #include "amdgpu_vkms.h" #include "mes_v10_1.h" +#include "mes_v11_0.h" #include "smuio_v11_0.h" #include "smuio_v11_0_6.h" #include "smuio_v13_0.h" +#include "smuio_v13_0_6.h" -MODULE_FIRMWARE("amdgpu/ip_discovery.bin"); +#define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin" +MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY); #define mmRCC_CONFIG_MEMSIZE 0xde3 #define mmMM_INDEX 0x0 @@ -110,6 +125,7 @@ static const char *hw_id_names[HW_ID_MAX] = { [SDMA1_HWID] = "SDMA1", [SDMA2_HWID] = "SDMA2", [SDMA3_HWID] = "SDMA3", + [LSDMA_HWID] = "LSDMA", [ISP_HWID] = "ISP", [DBGU_IO_HWID] = "DBGU_IO", [DF_HWID] = "DF", @@ -159,6 +175,7 @@ static int hw_id_map[MAX_HWIP] = { [SDMA1_HWIP] = SDMA1_HWID, [SDMA2_HWIP] = SDMA2_HWID, [SDMA3_HWIP] = SDMA3_HWID, + [LSDMA_HWIP] = LSDMA_HWID, [MMHUB_HWIP] = MMHUB_HWID, [ATHUB_HWIP] = ATHUB_HWID, [NBIO_HWIP] = NBIF_HWID, @@ -177,9 +194,10 @@ static int hw_id_map[MAX_HWIP] = { [UMC_HWIP] = UMC_HWID, [XGMI_HWIP] = XGMI_HWID, [DCI_HWIP] = DCI_HWID, + [PCIE_HWIP] = PCIE_HWID, }; -static int amdgpu_discovery_read_binary(struct amdgpu_device *adev, uint8_t *binary) +static int amdgpu_discovery_read_binary_from_vram(struct amdgpu_device *adev, uint8_t *binary) { uint64_t vram_size = (uint64_t)RREG32(mmRCC_CONFIG_MEMSIZE) << 20; uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET; @@ -189,6 +207,34 @@ static int amdgpu_discovery_read_binary(struct amdgpu_device *adev, uint8_t *bin return 0; } +static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, uint8_t *binary) +{ + const struct firmware *fw; + const char *fw_name; + int r; + + switch (amdgpu_discovery) { + case 2: + fw_name = FIRMWARE_IP_DISCOVERY; + break; + default: + dev_warn(adev->dev, "amdgpu_discovery is not set properly\n"); + return -EINVAL; + } + + r = request_firmware(&fw, fw_name, adev->dev); + if (r) { + dev_err(adev->dev, "can't load firmware \"%s\"\n", + fw_name); + return r; + } + + memcpy((u8 *)binary, (u8 *)fw->data, fw->size); + release_firmware(fw); + + return 0; +} + static uint16_t amdgpu_discovery_calculate_checksum(uint8_t *data, uint32_t size) { uint16_t checksum = 0; @@ -206,13 +252,42 @@ static inline bool amdgpu_discovery_verify_checksum(uint8_t *data, uint32_t size return !!(amdgpu_discovery_calculate_checksum(data, size) == expected); } +static inline bool amdgpu_discovery_verify_binary_signature(uint8_t *binary) +{ + struct binary_header *bhdr; + bhdr = (struct binary_header *)binary; + + return (le32_to_cpu(bhdr->binary_signature) == BINARY_SIGNATURE); +} + +static void amdgpu_discovery_harvest_config_quirk(struct amdgpu_device *adev) +{ + /* + * So far, apply this quirk only on those Navy Flounder boards which + * have a bad harvest table of VCN config. + */ + if ((adev->ip_versions[UVD_HWIP][1] == IP_VERSION(3, 0, 1)) && + (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 2))) { + switch (adev->pdev->revision) { + case 0xC1: + case 0xC2: + case 0xC3: + case 0xC5: + case 0xC7: + case 0xCF: + case 0xDF: + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + break; + default: + break; + } + } +} + static int amdgpu_discovery_init(struct amdgpu_device *adev) { struct table_info *info; struct binary_header *bhdr; - struct ip_discovery_header *ihdr; - struct gpu_info_header *ghdr; - const struct firmware *fw; uint16_t offset; uint16_t size; uint16_t checksum; @@ -223,31 +298,32 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (!adev->mman.discovery_bin) return -ENOMEM; - if (amdgpu_discovery == 2) { - r = request_firmware(&fw, "amdgpu/ip_discovery.bin", adev->dev); - if (r) - goto get_from_vram; - dev_info(adev->dev, "Using IP discovery from file\n"); - memcpy((u8 *)adev->mman.discovery_bin, (u8 *)fw->data, - adev->mman.discovery_tmr_size); - release_firmware(fw); - } else { -get_from_vram: - r = amdgpu_discovery_read_binary(adev, adev->mman.discovery_bin); + r = amdgpu_discovery_read_binary_from_vram(adev, adev->mman.discovery_bin); + if (r) { + dev_err(adev->dev, "failed to read ip discovery binary from vram\n"); + r = -EINVAL; + goto out; + } + + if (!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) { + dev_warn(adev->dev, "get invalid ip discovery binary signature from vram\n"); + /* retry read ip discovery binary from file */ + r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin); if (r) { - DRM_ERROR("failed to read ip discovery binary\n"); + dev_err(adev->dev, "failed to read ip discovery binary from file\n"); + r = -EINVAL; + goto out; + } + /* check the ip discovery binary signature */ + if(!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin)) { + dev_warn(adev->dev, "get invalid ip discovery binary signature from file\n"); + r = -EINVAL; goto out; } } bhdr = (struct binary_header *)adev->mman.discovery_bin; - if (le32_to_cpu(bhdr->binary_signature) != BINARY_SIGNATURE) { - DRM_ERROR("invalid ip discovery binary signature\n"); - r = -EINVAL; - goto out; - } - offset = offsetof(struct binary_header, binary_checksum) + sizeof(bhdr->binary_checksum); size = le16_to_cpu(bhdr->binary_size) - offset; @@ -255,7 +331,7 @@ get_from_vram: if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, size, checksum)) { - DRM_ERROR("invalid ip discovery binary checksum\n"); + dev_err(adev->dev, "invalid ip discovery binary checksum\n"); r = -EINVAL; goto out; } @@ -263,31 +339,110 @@ get_from_vram: info = &bhdr->table_list[IP_DISCOVERY]; offset = le16_to_cpu(info->offset); checksum = le16_to_cpu(info->checksum); - ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + offset); - if (le32_to_cpu(ihdr->signature) != DISCOVERY_TABLE_SIGNATURE) { - DRM_ERROR("invalid ip discovery data table signature\n"); - r = -EINVAL; - goto out; - } + if (offset) { + struct ip_discovery_header *ihdr = + (struct ip_discovery_header *)(adev->mman.discovery_bin + offset); + if (le32_to_cpu(ihdr->signature) != DISCOVERY_TABLE_SIGNATURE) { + dev_err(adev->dev, "invalid ip discovery data table signature\n"); + r = -EINVAL; + goto out; + } - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - le16_to_cpu(ihdr->size), checksum)) { - DRM_ERROR("invalid ip discovery data table checksum\n"); - r = -EINVAL; - goto out; + if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, + le16_to_cpu(ihdr->size), checksum)) { + dev_err(adev->dev, "invalid ip discovery data table checksum\n"); + r = -EINVAL; + goto out; + } } info = &bhdr->table_list[GC]; offset = le16_to_cpu(info->offset); checksum = le16_to_cpu(info->checksum); - ghdr = (struct gpu_info_header *)(adev->mman.discovery_bin + offset); - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - le32_to_cpu(ghdr->size), checksum)) { - DRM_ERROR("invalid gc data table checksum\n"); - r = -EINVAL; - goto out; + if (offset) { + struct gpu_info_header *ghdr = + (struct gpu_info_header *)(adev->mman.discovery_bin + offset); + + if (le32_to_cpu(ghdr->table_id) != GC_TABLE_ID) { + dev_err(adev->dev, "invalid ip discovery gc table id\n"); + r = -EINVAL; + goto out; + } + + if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, + le32_to_cpu(ghdr->size), checksum)) { + dev_err(adev->dev, "invalid gc data table checksum\n"); + r = -EINVAL; + goto out; + } + } + + info = &bhdr->table_list[HARVEST_INFO]; + offset = le16_to_cpu(info->offset); + checksum = le16_to_cpu(info->checksum); + + if (offset) { + struct harvest_info_header *hhdr = + (struct harvest_info_header *)(adev->mman.discovery_bin + offset); + + if (le32_to_cpu(hhdr->signature) != HARVEST_TABLE_SIGNATURE) { + dev_err(adev->dev, "invalid ip discovery harvest table signature\n"); + r = -EINVAL; + goto out; + } + + if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, + sizeof(struct harvest_table), checksum)) { + dev_err(adev->dev, "invalid harvest data table checksum\n"); + r = -EINVAL; + goto out; + } + } + + info = &bhdr->table_list[VCN_INFO]; + offset = le16_to_cpu(info->offset); + checksum = le16_to_cpu(info->checksum); + + if (offset) { + struct vcn_info_header *vhdr = + (struct vcn_info_header *)(adev->mman.discovery_bin + offset); + + if (le32_to_cpu(vhdr->table_id) != VCN_INFO_TABLE_ID) { + dev_err(adev->dev, "invalid ip discovery vcn table id\n"); + r = -EINVAL; + goto out; + } + + if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, + le32_to_cpu(vhdr->size_bytes), checksum)) { + dev_err(adev->dev, "invalid vcn data table checksum\n"); + r = -EINVAL; + goto out; + } + } + + info = &bhdr->table_list[MALL_INFO]; + offset = le16_to_cpu(info->offset); + checksum = le16_to_cpu(info->checksum); + + if (0 && offset) { + struct mall_info_header *mhdr = + (struct mall_info_header *)(adev->mman.discovery_bin + offset); + + if (le32_to_cpu(mhdr->table_id) != MALL_INFO_TABLE_ID) { + dev_err(adev->dev, "invalid ip discovery mall table id\n"); + r = -EINVAL; + goto out; + } + + if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, + le32_to_cpu(mhdr->size_bytes), checksum)) { + dev_err(adev->dev, "invalid mall data table checksum\n"); + r = -EINVAL; + goto out; + } } return 0; @@ -299,8 +454,11 @@ out: return r; } +static void amdgpu_discovery_sysfs_fini(struct amdgpu_device *adev); + void amdgpu_discovery_fini(struct amdgpu_device *adev) { + amdgpu_discovery_sysfs_fini(adev); kfree(adev->mman.discovery_bin); adev->mman.discovery_bin = NULL; } @@ -321,7 +479,591 @@ static int amdgpu_discovery_validate_ip(const struct ip *ip) return 0; } -int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) +static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, + uint32_t *vcn_harvest_count) +{ + struct binary_header *bhdr; + struct ip_discovery_header *ihdr; + struct die_header *dhdr; + struct ip *ip; + uint16_t die_offset, ip_offset, num_dies, num_ips; + int i, j; + + bhdr = (struct binary_header *)adev->mman.discovery_bin; + ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + + le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset)); + num_dies = le16_to_cpu(ihdr->num_dies); + + /* scan harvest bit of all IP data structures */ + for (i = 0; i < num_dies; i++) { + die_offset = le16_to_cpu(ihdr->die_info[i].die_offset); + dhdr = (struct die_header *)(adev->mman.discovery_bin + die_offset); + num_ips = le16_to_cpu(dhdr->num_ips); + ip_offset = die_offset + sizeof(*dhdr); + + for (j = 0; j < num_ips; j++) { + ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + + if (amdgpu_discovery_validate_ip(ip)) + goto next_ip; + + if (le16_to_cpu(ip->harvest) == 1) { + switch (le16_to_cpu(ip->hw_id)) { + case VCN_HWID: + (*vcn_harvest_count)++; + if (ip->number_instance == 0) + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; + else + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + break; + case DMU_HWID: + adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; + break; + default: + break; + } + } +next_ip: + ip_offset += struct_size(ip, base_address, ip->num_base_address); + } + } +} + +static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev, + uint32_t *vcn_harvest_count, + uint32_t *umc_harvest_count) +{ + struct binary_header *bhdr; + struct harvest_table *harvest_info; + u16 offset; + int i; + + bhdr = (struct binary_header *)adev->mman.discovery_bin; + offset = le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset); + + if (!offset) { + dev_err(adev->dev, "invalid harvest table offset\n"); + return; + } + + harvest_info = (struct harvest_table *)(adev->mman.discovery_bin + offset); + + for (i = 0; i < 32; i++) { + if (le16_to_cpu(harvest_info->list[i].hw_id) == 0) + break; + + switch (le16_to_cpu(harvest_info->list[i].hw_id)) { + case VCN_HWID: + (*vcn_harvest_count)++; + if (harvest_info->list[i].number_instance == 0) + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; + else + adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + break; + case DMU_HWID: + adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; + break; + case UMC_HWID: + (*umc_harvest_count)++; + break; + default: + break; + } + } +} + +/* ================================================== */ + +struct ip_hw_instance { + struct kobject kobj; /* ip_discovery/die/#die/#hw_id/#instance/<attrs...> */ + + int hw_id; + u8 num_instance; + u8 major, minor, revision; + u8 harvest; + + int num_base_addresses; + u32 base_addr[]; +}; + +struct ip_hw_id { + struct kset hw_id_kset; /* ip_discovery/die/#die/#hw_id/, contains ip_hw_instance */ + int hw_id; +}; + +struct ip_die_entry { + struct kset ip_kset; /* ip_discovery/die/#die/, contains ip_hw_id */ + u16 num_ips; +}; + +/* -------------------------------------------------- */ + +struct ip_hw_instance_attr { + struct attribute attr; + ssize_t (*show)(struct ip_hw_instance *ip_hw_instance, char *buf); +}; + +static ssize_t hw_id_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_hw_instance->hw_id); +} + +static ssize_t num_instance_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_hw_instance->num_instance); +} + +static ssize_t major_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_hw_instance->major); +} + +static ssize_t minor_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_hw_instance->minor); +} + +static ssize_t revision_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_hw_instance->revision); +} + +static ssize_t harvest_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "0x%01X\n", ip_hw_instance->harvest); +} + +static ssize_t num_base_addresses_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_hw_instance->num_base_addresses); +} + +static ssize_t base_addr_show(struct ip_hw_instance *ip_hw_instance, char *buf) +{ + ssize_t res, at; + int ii; + + for (res = at = ii = 0; ii < ip_hw_instance->num_base_addresses; ii++) { + /* Here we satisfy the condition that, at + size <= PAGE_SIZE. + */ + if (at + 12 > PAGE_SIZE) + break; + res = sysfs_emit_at(buf, at, "0x%08X\n", + ip_hw_instance->base_addr[ii]); + if (res <= 0) + break; + at += res; + } + + return res < 0 ? res : at; +} + +static struct ip_hw_instance_attr ip_hw_attr[] = { + __ATTR_RO(hw_id), + __ATTR_RO(num_instance), + __ATTR_RO(major), + __ATTR_RO(minor), + __ATTR_RO(revision), + __ATTR_RO(harvest), + __ATTR_RO(num_base_addresses), + __ATTR_RO(base_addr), +}; + +static struct attribute *ip_hw_instance_attrs[ARRAY_SIZE(ip_hw_attr) + 1]; +ATTRIBUTE_GROUPS(ip_hw_instance); + +#define to_ip_hw_instance(x) container_of(x, struct ip_hw_instance, kobj) +#define to_ip_hw_instance_attr(x) container_of(x, struct ip_hw_instance_attr, attr) + +static ssize_t ip_hw_instance_attr_show(struct kobject *kobj, + struct attribute *attr, + char *buf) +{ + struct ip_hw_instance *ip_hw_instance = to_ip_hw_instance(kobj); + struct ip_hw_instance_attr *ip_hw_attr = to_ip_hw_instance_attr(attr); + + if (!ip_hw_attr->show) + return -EIO; + + return ip_hw_attr->show(ip_hw_instance, buf); +} + +static const struct sysfs_ops ip_hw_instance_sysfs_ops = { + .show = ip_hw_instance_attr_show, +}; + +static void ip_hw_instance_release(struct kobject *kobj) +{ + struct ip_hw_instance *ip_hw_instance = to_ip_hw_instance(kobj); + + kfree(ip_hw_instance); +} + +static struct kobj_type ip_hw_instance_ktype = { + .release = ip_hw_instance_release, + .sysfs_ops = &ip_hw_instance_sysfs_ops, + .default_groups = ip_hw_instance_groups, +}; + +/* -------------------------------------------------- */ + +#define to_ip_hw_id(x) container_of(to_kset(x), struct ip_hw_id, hw_id_kset) + +static void ip_hw_id_release(struct kobject *kobj) +{ + struct ip_hw_id *ip_hw_id = to_ip_hw_id(kobj); + + if (!list_empty(&ip_hw_id->hw_id_kset.list)) + DRM_ERROR("ip_hw_id->hw_id_kset is not empty"); + kfree(ip_hw_id); +} + +static struct kobj_type ip_hw_id_ktype = { + .release = ip_hw_id_release, + .sysfs_ops = &kobj_sysfs_ops, +}; + +/* -------------------------------------------------- */ + +static void die_kobj_release(struct kobject *kobj); +static void ip_disc_release(struct kobject *kobj); + +struct ip_die_entry_attribute { + struct attribute attr; + ssize_t (*show)(struct ip_die_entry *ip_die_entry, char *buf); +}; + +#define to_ip_die_entry_attr(x) container_of(x, struct ip_die_entry_attribute, attr) + +static ssize_t num_ips_show(struct ip_die_entry *ip_die_entry, char *buf) +{ + return sysfs_emit(buf, "%d\n", ip_die_entry->num_ips); +} + +/* If there are more ip_die_entry attrs, other than the number of IPs, + * we can make this intro an array of attrs, and then initialize + * ip_die_entry_attrs in a loop. + */ +static struct ip_die_entry_attribute num_ips_attr = + __ATTR_RO(num_ips); + +static struct attribute *ip_die_entry_attrs[] = { + &num_ips_attr.attr, + NULL, +}; +ATTRIBUTE_GROUPS(ip_die_entry); /* ip_die_entry_groups */ + +#define to_ip_die_entry(x) container_of(to_kset(x), struct ip_die_entry, ip_kset) + +static ssize_t ip_die_entry_attr_show(struct kobject *kobj, + struct attribute *attr, + char *buf) +{ + struct ip_die_entry_attribute *ip_die_entry_attr = to_ip_die_entry_attr(attr); + struct ip_die_entry *ip_die_entry = to_ip_die_entry(kobj); + + if (!ip_die_entry_attr->show) + return -EIO; + + return ip_die_entry_attr->show(ip_die_entry, buf); +} + +static void ip_die_entry_release(struct kobject *kobj) +{ + struct ip_die_entry *ip_die_entry = to_ip_die_entry(kobj); + + if (!list_empty(&ip_die_entry->ip_kset.list)) + DRM_ERROR("ip_die_entry->ip_kset is not empty"); + kfree(ip_die_entry); +} + +static const struct sysfs_ops ip_die_entry_sysfs_ops = { + .show = ip_die_entry_attr_show, +}; + +static struct kobj_type ip_die_entry_ktype = { + .release = ip_die_entry_release, + .sysfs_ops = &ip_die_entry_sysfs_ops, + .default_groups = ip_die_entry_groups, +}; + +static struct kobj_type die_kobj_ktype = { + .release = die_kobj_release, + .sysfs_ops = &kobj_sysfs_ops, +}; + +static struct kobj_type ip_discovery_ktype = { + .release = ip_disc_release, + .sysfs_ops = &kobj_sysfs_ops, +}; + +struct ip_discovery_top { + struct kobject kobj; /* ip_discovery/ */ + struct kset die_kset; /* ip_discovery/die/, contains ip_die_entry */ + struct amdgpu_device *adev; +}; + +static void die_kobj_release(struct kobject *kobj) +{ + struct ip_discovery_top *ip_top = container_of(to_kset(kobj), + struct ip_discovery_top, + die_kset); + if (!list_empty(&ip_top->die_kset.list)) + DRM_ERROR("ip_top->die_kset is not empty"); +} + +static void ip_disc_release(struct kobject *kobj) +{ + struct ip_discovery_top *ip_top = container_of(kobj, struct ip_discovery_top, + kobj); + struct amdgpu_device *adev = ip_top->adev; + + adev->ip_top = NULL; + kfree(ip_top); +} + +static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, + struct ip_die_entry *ip_die_entry, + const size_t _ip_offset, const int num_ips) +{ + int ii, jj, kk, res; + + DRM_DEBUG("num_ips:%d", num_ips); + + /* Find all IPs of a given HW ID, and add their instance to + * #die/#hw_id/#instance/<attributes> + */ + for (ii = 0; ii < HW_ID_MAX; ii++) { + struct ip_hw_id *ip_hw_id = NULL; + size_t ip_offset = _ip_offset; + + for (jj = 0; jj < num_ips; jj++) { + struct ip *ip; + struct ip_hw_instance *ip_hw_instance; + + ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + if (amdgpu_discovery_validate_ip(ip) || + le16_to_cpu(ip->hw_id) != ii) + goto next_ip; + + DRM_DEBUG("match:%d @ ip_offset:%zu", ii, ip_offset); + + /* We have a hw_id match; register the hw + * block if not yet registered. + */ + if (!ip_hw_id) { + ip_hw_id = kzalloc(sizeof(*ip_hw_id), GFP_KERNEL); + if (!ip_hw_id) + return -ENOMEM; + ip_hw_id->hw_id = ii; + + kobject_set_name(&ip_hw_id->hw_id_kset.kobj, "%d", ii); + ip_hw_id->hw_id_kset.kobj.kset = &ip_die_entry->ip_kset; + ip_hw_id->hw_id_kset.kobj.ktype = &ip_hw_id_ktype; + res = kset_register(&ip_hw_id->hw_id_kset); + if (res) { + DRM_ERROR("Couldn't register ip_hw_id kset"); + kfree(ip_hw_id); + return res; + } + if (hw_id_names[ii]) { + res = sysfs_create_link(&ip_die_entry->ip_kset.kobj, + &ip_hw_id->hw_id_kset.kobj, + hw_id_names[ii]); + if (res) { + DRM_ERROR("Couldn't create IP link %s in IP Die:%s\n", + hw_id_names[ii], + kobject_name(&ip_die_entry->ip_kset.kobj)); + } + } + } + + /* Now register its instance. + */ + ip_hw_instance = kzalloc(struct_size(ip_hw_instance, + base_addr, + ip->num_base_address), + GFP_KERNEL); + if (!ip_hw_instance) { + DRM_ERROR("no memory for ip_hw_instance"); + return -ENOMEM; + } + ip_hw_instance->hw_id = le16_to_cpu(ip->hw_id); /* == ii */ + ip_hw_instance->num_instance = ip->number_instance; + ip_hw_instance->major = ip->major; + ip_hw_instance->minor = ip->minor; + ip_hw_instance->revision = ip->revision; + ip_hw_instance->harvest = ip->harvest; + ip_hw_instance->num_base_addresses = ip->num_base_address; + + for (kk = 0; kk < ip_hw_instance->num_base_addresses; kk++) + ip_hw_instance->base_addr[kk] = ip->base_address[kk]; + + kobject_init(&ip_hw_instance->kobj, &ip_hw_instance_ktype); + ip_hw_instance->kobj.kset = &ip_hw_id->hw_id_kset; + res = kobject_add(&ip_hw_instance->kobj, NULL, + "%d", ip_hw_instance->num_instance); +next_ip: + ip_offset += struct_size(ip, base_address, ip->num_base_address); + } + } + + return 0; +} + +static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev) +{ + struct binary_header *bhdr; + struct ip_discovery_header *ihdr; + struct die_header *dhdr; + struct kset *die_kset = &adev->ip_top->die_kset; + u16 num_dies, die_offset, num_ips; + size_t ip_offset; + int ii, res; + + bhdr = (struct binary_header *)adev->mman.discovery_bin; + ihdr = (struct ip_discovery_header *)(adev->mman.discovery_bin + + le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset)); + num_dies = le16_to_cpu(ihdr->num_dies); + + DRM_DEBUG("number of dies: %d\n", num_dies); + + for (ii = 0; ii < num_dies; ii++) { + struct ip_die_entry *ip_die_entry; + + die_offset = le16_to_cpu(ihdr->die_info[ii].die_offset); + dhdr = (struct die_header *)(adev->mman.discovery_bin + die_offset); + num_ips = le16_to_cpu(dhdr->num_ips); + ip_offset = die_offset + sizeof(*dhdr); + + /* Add the die to the kset. + * + * dhdr->die_id == ii, which was checked in + * amdgpu_discovery_reg_base_init(). + */ + + ip_die_entry = kzalloc(sizeof(*ip_die_entry), GFP_KERNEL); + if (!ip_die_entry) + return -ENOMEM; + + ip_die_entry->num_ips = num_ips; + + kobject_set_name(&ip_die_entry->ip_kset.kobj, "%d", le16_to_cpu(dhdr->die_id)); + ip_die_entry->ip_kset.kobj.kset = die_kset; + ip_die_entry->ip_kset.kobj.ktype = &ip_die_entry_ktype; + res = kset_register(&ip_die_entry->ip_kset); + if (res) { + DRM_ERROR("Couldn't register ip_die_entry kset"); + kfree(ip_die_entry); + return res; + } + + amdgpu_discovery_sysfs_ips(adev, ip_die_entry, ip_offset, num_ips); + } + + return 0; +} + +static int amdgpu_discovery_sysfs_init(struct amdgpu_device *adev) +{ + struct kset *die_kset; + int res, ii; + + adev->ip_top = kzalloc(sizeof(*adev->ip_top), GFP_KERNEL); + if (!adev->ip_top) + return -ENOMEM; + + adev->ip_top->adev = adev; + + res = kobject_init_and_add(&adev->ip_top->kobj, &ip_discovery_ktype, + &adev->dev->kobj, "ip_discovery"); + if (res) { + DRM_ERROR("Couldn't init and add ip_discovery/"); + goto Err; + } + + die_kset = &adev->ip_top->die_kset; + kobject_set_name(&die_kset->kobj, "%s", "die"); + die_kset->kobj.parent = &adev->ip_top->kobj; + die_kset->kobj.ktype = &die_kobj_ktype; + res = kset_register(&adev->ip_top->die_kset); + if (res) { + DRM_ERROR("Couldn't register die_kset"); + goto Err; + } + + for (ii = 0; ii < ARRAY_SIZE(ip_hw_attr); ii++) + ip_hw_instance_attrs[ii] = &ip_hw_attr[ii].attr; + ip_hw_instance_attrs[ii] = NULL; + + res = amdgpu_discovery_sysfs_recurse(adev); + + return res; +Err: + kobject_put(&adev->ip_top->kobj); + return res; +} + +/* -------------------------------------------------- */ + +#define list_to_kobj(el) container_of(el, struct kobject, entry) + +static void amdgpu_discovery_sysfs_ip_hw_free(struct ip_hw_id *ip_hw_id) +{ + struct list_head *el, *tmp; + struct kset *hw_id_kset; + + hw_id_kset = &ip_hw_id->hw_id_kset; + spin_lock(&hw_id_kset->list_lock); + list_for_each_prev_safe(el, tmp, &hw_id_kset->list) { + list_del_init(el); + spin_unlock(&hw_id_kset->list_lock); + /* kobject is embedded in ip_hw_instance */ + kobject_put(list_to_kobj(el)); + spin_lock(&hw_id_kset->list_lock); + } + spin_unlock(&hw_id_kset->list_lock); + kobject_put(&ip_hw_id->hw_id_kset.kobj); +} + +static void amdgpu_discovery_sysfs_die_free(struct ip_die_entry *ip_die_entry) +{ + struct list_head *el, *tmp; + struct kset *ip_kset; + + ip_kset = &ip_die_entry->ip_kset; + spin_lock(&ip_kset->list_lock); + list_for_each_prev_safe(el, tmp, &ip_kset->list) { + list_del_init(el); + spin_unlock(&ip_kset->list_lock); + amdgpu_discovery_sysfs_ip_hw_free(to_ip_hw_id(list_to_kobj(el))); + spin_lock(&ip_kset->list_lock); + } + spin_unlock(&ip_kset->list_lock); + kobject_put(&ip_die_entry->ip_kset.kobj); +} + +static void amdgpu_discovery_sysfs_fini(struct amdgpu_device *adev) +{ + struct list_head *el, *tmp; + struct kset *die_kset; + + die_kset = &adev->ip_top->die_kset; + spin_lock(&die_kset->list_lock); + list_for_each_prev_safe(el, tmp, &die_kset->list) { + list_del_init(el); + spin_unlock(&die_kset->list_lock); + amdgpu_discovery_sysfs_die_free(to_ip_die_entry(list_to_kobj(el))); + spin_lock(&die_kset->list_lock); + } + spin_unlock(&die_kset->list_lock); + kobject_put(&adev->ip_top->die_kset.kobj); + kobject_put(&adev->ip_top->kobj); +} + +/* ================================================== */ + +static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) { struct binary_header *bhdr; struct ip_discovery_header *ihdr; @@ -379,13 +1121,37 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) ip->major, ip->minor, ip->revision); - if (le16_to_cpu(ip->hw_id) == VCN_HWID) - adev->vcn.num_vcn_inst++; + if (le16_to_cpu(ip->hw_id) == VCN_HWID) { + /* Bit [5:0]: original revision value + * Bit [7:6]: en/decode capability: + * 0b00 : VCN function normally + * 0b10 : encode is disabled + * 0b01 : decode is disabled + */ + adev->vcn.vcn_config[adev->vcn.num_vcn_inst] = + ip->revision & 0xc0; + ip->revision &= ~0xc0; + if (adev->vcn.num_vcn_inst < AMDGPU_MAX_VCN_INSTANCES) + adev->vcn.num_vcn_inst++; + else + dev_err(adev->dev, "Too many VCN instances: %d vs %d\n", + adev->vcn.num_vcn_inst + 1, + AMDGPU_MAX_VCN_INSTANCES); + } if (le16_to_cpu(ip->hw_id) == SDMA0_HWID || le16_to_cpu(ip->hw_id) == SDMA1_HWID || le16_to_cpu(ip->hw_id) == SDMA2_HWID || - le16_to_cpu(ip->hw_id) == SDMA3_HWID) - adev->sdma.num_instances++; + le16_to_cpu(ip->hw_id) == SDMA3_HWID) { + if (adev->sdma.num_instances < AMDGPU_MAX_SDMA_INSTANCES) + adev->sdma.num_instances++; + else + dev_err(adev->dev, "Too many SDMA instances: %d vs %d\n", + adev->sdma.num_instances + 1, + AMDGPU_MAX_SDMA_INSTANCES); + } + + if (le16_to_cpu(ip->hw_id) == UMC_HWID) + adev->gmc.num_umc++; for (k = 0; k < num_base_address; k++) { /* @@ -417,10 +1183,12 @@ int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) } next_ip: - ip_offset += sizeof(*ip) + 4 * (ip->num_base_address - 1); + ip_offset += struct_size(ip, base_address, ip->num_base_address); } } + amdgpu_discovery_sysfs_init(adev); + return 0; } @@ -465,100 +1233,244 @@ int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int n *revision = ip->revision; return 0; } - ip_offset += sizeof(*ip) + 4 * (ip->num_base_address - 1); + ip_offset += struct_size(ip, base_address, ip->num_base_address); } } return -EINVAL; } - -int amdgpu_discovery_get_vcn_version(struct amdgpu_device *adev, int vcn_instance, - int *major, int *minor, int *revision) +static void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) { - return amdgpu_discovery_get_ip_version(adev, VCN_HWID, - vcn_instance, major, minor, revision); + int vcn_harvest_count = 0; + int umc_harvest_count = 0; + + /* + * Harvest table does not fit Navi1x and legacy GPUs, + * so read harvest bit per IP data structure to set + * harvest configuration. + */ + if (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 2, 0)) { + if ((adev->pdev->device == 0x731E && + (adev->pdev->revision == 0xC6 || + adev->pdev->revision == 0xC7)) || + (adev->pdev->device == 0x7340 && + adev->pdev->revision == 0xC9) || + (adev->pdev->device == 0x7360 && + adev->pdev->revision == 0xC7)) + amdgpu_discovery_read_harvest_bit_per_ip(adev, + &vcn_harvest_count); + } else { + amdgpu_discovery_read_from_harvest_table(adev, + &vcn_harvest_count, + &umc_harvest_count); + } + + amdgpu_discovery_harvest_config_quirk(adev); + + if (vcn_harvest_count == adev->vcn.num_vcn_inst) { + adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK; + adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK; + } + + if (umc_harvest_count < adev->gmc.num_umc) { + adev->gmc.num_umc -= umc_harvest_count; + } } -void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) +union gc_info { + struct gc_info_v1_0 v1; + struct gc_info_v1_1 v1_1; + struct gc_info_v1_2 v1_2; + struct gc_info_v2_0 v2; +}; + +static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) { struct binary_header *bhdr; - struct harvest_table *harvest_info; - int i, vcn_harvest_count = 0; + union gc_info *gc_info; + u16 offset; + + if (!adev->mman.discovery_bin) { + DRM_ERROR("ip discovery uninitialized\n"); + return -EINVAL; + } bhdr = (struct binary_header *)adev->mman.discovery_bin; - harvest_info = (struct harvest_table *)(adev->mman.discovery_bin + - le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset)); + offset = le16_to_cpu(bhdr->table_list[GC].offset); + + if (!offset) + return 0; + + gc_info = (union gc_info *)(adev->mman.discovery_bin + offset); + + switch (le16_to_cpu(gc_info->v1.header.version_major)) { + case 1: + adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v1.gc_num_se); + adev->gfx.config.max_cu_per_sh = 2 * (le32_to_cpu(gc_info->v1.gc_num_wgp0_per_sa) + + le32_to_cpu(gc_info->v1.gc_num_wgp1_per_sa)); + adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->v1.gc_num_sa_per_se); + adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->v1.gc_num_rb_per_se); + adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->v1.gc_num_gl2c); + adev->gfx.config.max_gprs = le32_to_cpu(gc_info->v1.gc_num_gprs); + adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->v1.gc_num_max_gs_thds); + adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->v1.gc_gs_table_depth); + adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->v1.gc_gsprim_buff_depth); + adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->v1.gc_double_offchip_lds_buffer); + adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->v1.gc_wave_size); + adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->v1.gc_max_waves_per_simd); + adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->v1.gc_max_scratch_slots_per_cu); + adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->v1.gc_lds_size); + adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->v1.gc_num_sc_per_se) / + le32_to_cpu(gc_info->v1.gc_num_sa_per_se); + adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->v1.gc_num_packer_per_sc); + if (gc_info->v1.header.version_minor >= 1) { + adev->gfx.config.gc_num_tcp_per_sa = le32_to_cpu(gc_info->v1_1.gc_num_tcp_per_sa); + adev->gfx.config.gc_num_sdp_interface = le32_to_cpu(gc_info->v1_1.gc_num_sdp_interface); + adev->gfx.config.gc_num_tcps = le32_to_cpu(gc_info->v1_1.gc_num_tcps); + } + if (gc_info->v1.header.version_minor >= 2) { + adev->gfx.config.gc_num_tcp_per_wpg = le32_to_cpu(gc_info->v1_2.gc_num_tcp_per_wpg); + adev->gfx.config.gc_tcp_l1_size = le32_to_cpu(gc_info->v1_2.gc_tcp_l1_size); + adev->gfx.config.gc_num_sqc_per_wgp = le32_to_cpu(gc_info->v1_2.gc_num_sqc_per_wgp); + adev->gfx.config.gc_l1_instruction_cache_size_per_sqc = le32_to_cpu(gc_info->v1_2.gc_l1_instruction_cache_size_per_sqc); + adev->gfx.config.gc_l1_data_cache_size_per_sqc = le32_to_cpu(gc_info->v1_2.gc_l1_data_cache_size_per_sqc); + adev->gfx.config.gc_gl1c_per_sa = le32_to_cpu(gc_info->v1_2.gc_gl1c_per_sa); + adev->gfx.config.gc_gl1c_size_per_instance = le32_to_cpu(gc_info->v1_2.gc_gl1c_size_per_instance); + adev->gfx.config.gc_gl2c_per_gpu = le32_to_cpu(gc_info->v1_2.gc_gl2c_per_gpu); + } + break; + case 2: + adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v2.gc_num_se); + adev->gfx.config.max_cu_per_sh = le32_to_cpu(gc_info->v2.gc_num_cu_per_sh); + adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->v2.gc_num_sh_per_se); + adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->v2.gc_num_rb_per_se); + adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->v2.gc_num_tccs); + adev->gfx.config.max_gprs = le32_to_cpu(gc_info->v2.gc_num_gprs); + adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->v2.gc_num_max_gs_thds); + adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->v2.gc_gs_table_depth); + adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->v2.gc_gsprim_buff_depth); + adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->v2.gc_double_offchip_lds_buffer); + adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->v2.gc_wave_size); + adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->v2.gc_max_waves_per_simd); + adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->v2.gc_max_scratch_slots_per_cu); + adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->v2.gc_lds_size); + adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->v2.gc_num_sc_per_se) / + le32_to_cpu(gc_info->v2.gc_num_sh_per_se); + adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->v2.gc_num_packer_per_sc); + break; + default: + dev_err(adev->dev, + "Unhandled GC info table %d.%d\n", + le16_to_cpu(gc_info->v1.header.version_major), + le16_to_cpu(gc_info->v1.header.version_minor)); + return -EINVAL; + } + return 0; +} - for (i = 0; i < 32; i++) { - if (le16_to_cpu(harvest_info->list[i].hw_id) == 0) - break; +union mall_info { + struct mall_info_v1_0 v1; +}; - switch (le16_to_cpu(harvest_info->list[i].hw_id)) { - case VCN_HWID: - vcn_harvest_count++; - if (harvest_info->list[i].number_instance == 0) - adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; +static int amdgpu_discovery_get_mall_info(struct amdgpu_device *adev) +{ + struct binary_header *bhdr; + union mall_info *mall_info; + u32 u, mall_size_per_umc, m_s_present, half_use; + u64 mall_size; + u16 offset; + + if (!adev->mman.discovery_bin) { + DRM_ERROR("ip discovery uninitialized\n"); + return -EINVAL; + } + + bhdr = (struct binary_header *)adev->mman.discovery_bin; + offset = le16_to_cpu(bhdr->table_list[MALL_INFO].offset); + + if (!offset) + return 0; + + mall_info = (union mall_info *)(adev->mman.discovery_bin + offset); + + switch (le16_to_cpu(mall_info->v1.header.version_major)) { + case 1: + mall_size = 0; + mall_size_per_umc = le32_to_cpu(mall_info->v1.mall_size_per_m); + m_s_present = le32_to_cpu(mall_info->v1.m_s_present); + half_use = le32_to_cpu(mall_info->v1.m_half_use); + for (u = 0; u < adev->gmc.num_umc; u++) { + if (m_s_present & (1 << u)) + mall_size += mall_size_per_umc * 2; + else if (half_use & (1 << u)) + mall_size += mall_size_per_umc / 2; else - adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; - break; - case DMU_HWID: - adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; - break; - default: - break; + mall_size += mall_size_per_umc; } + adev->gmc.mall_size = mall_size; + break; + default: + dev_err(adev->dev, + "Unhandled MALL info table %d.%d\n", + le16_to_cpu(mall_info->v1.header.version_major), + le16_to_cpu(mall_info->v1.header.version_minor)); + return -EINVAL; } - /* some IP discovery tables on Navy Flounder don't have this set correctly */ - if ((adev->ip_versions[UVD_HWIP][1] == IP_VERSION(3, 0, 1)) && - (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 2))) - adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; - if (vcn_harvest_count == adev->vcn.num_vcn_inst) { - adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK; - adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK; - } - if ((adev->pdev->device == 0x731E && - (adev->pdev->revision == 0xC6 || adev->pdev->revision == 0xC7)) || - (adev->pdev->device == 0x7340 && adev->pdev->revision == 0xC9) || - (adev->pdev->device == 0x7360 && adev->pdev->revision == 0xC7)) { - adev->harvest_ip_mask |= AMD_HARVEST_IP_VCN_MASK; - adev->harvest_ip_mask |= AMD_HARVEST_IP_JPEG_MASK; - } + return 0; } -int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) +union vcn_info { + struct vcn_info_v1_0 v1; +}; + +static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev) { struct binary_header *bhdr; - struct gc_info_v1_0 *gc_info; + union vcn_info *vcn_info; + u16 offset; + int v; if (!adev->mman.discovery_bin) { DRM_ERROR("ip discovery uninitialized\n"); return -EINVAL; } + /* num_vcn_inst is currently limited to AMDGPU_MAX_VCN_INSTANCES + * which is smaller than VCN_INFO_TABLE_MAX_NUM_INSTANCES + * but that may change in the future with new GPUs so keep this + * check for defensive purposes. + */ + if (adev->vcn.num_vcn_inst > VCN_INFO_TABLE_MAX_NUM_INSTANCES) { + dev_err(adev->dev, "invalid vcn instances\n"); + return -EINVAL; + } + bhdr = (struct binary_header *)adev->mman.discovery_bin; - gc_info = (struct gc_info_v1_0 *)(adev->mman.discovery_bin + - le16_to_cpu(bhdr->table_list[GC].offset)); - - adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->gc_num_se); - adev->gfx.config.max_cu_per_sh = 2 * (le32_to_cpu(gc_info->gc_num_wgp0_per_sa) + - le32_to_cpu(gc_info->gc_num_wgp1_per_sa)); - adev->gfx.config.max_sh_per_se = le32_to_cpu(gc_info->gc_num_sa_per_se); - adev->gfx.config.max_backends_per_se = le32_to_cpu(gc_info->gc_num_rb_per_se); - adev->gfx.config.max_texture_channel_caches = le32_to_cpu(gc_info->gc_num_gl2c); - adev->gfx.config.max_gprs = le32_to_cpu(gc_info->gc_num_gprs); - adev->gfx.config.max_gs_threads = le32_to_cpu(gc_info->gc_num_max_gs_thds); - adev->gfx.config.gs_vgt_table_depth = le32_to_cpu(gc_info->gc_gs_table_depth); - adev->gfx.config.gs_prim_buffer_depth = le32_to_cpu(gc_info->gc_gsprim_buff_depth); - adev->gfx.config.double_offchip_lds_buf = le32_to_cpu(gc_info->gc_double_offchip_lds_buffer); - adev->gfx.cu_info.wave_front_size = le32_to_cpu(gc_info->gc_wave_size); - adev->gfx.cu_info.max_waves_per_simd = le32_to_cpu(gc_info->gc_max_waves_per_simd); - adev->gfx.cu_info.max_scratch_slots_per_cu = le32_to_cpu(gc_info->gc_max_scratch_slots_per_cu); - adev->gfx.cu_info.lds_size = le32_to_cpu(gc_info->gc_lds_size); - adev->gfx.config.num_sc_per_sh = le32_to_cpu(gc_info->gc_num_sc_per_se) / - le32_to_cpu(gc_info->gc_num_sa_per_se); - adev->gfx.config.num_packer_per_sc = le32_to_cpu(gc_info->gc_num_packer_per_sc); + offset = le16_to_cpu(bhdr->table_list[VCN_INFO].offset); + + if (!offset) + return 0; + vcn_info = (union vcn_info *)(adev->mman.discovery_bin + offset); + + switch (le16_to_cpu(vcn_info->v1.header.version_major)) { + case 1: + /* num_vcn_inst is currently limited to AMDGPU_MAX_VCN_INSTANCES + * so this won't overflow. + */ + for (v = 0; v < adev->vcn.num_vcn_inst; v++) { + adev->vcn.vcn_codec_disable_mask[v] = + le32_to_cpu(vcn_info->v1.instance_info[v].fuse_data.all_bits); + } + break; + default: + dev_err(adev->dev, + "Unhandled VCN info table %d.%d\n", + le16_to_cpu(vcn_info->v1.header.version_major), + le16_to_cpu(vcn_info->v1.header.version_minor)); + return -EINVAL; + } return 0; } @@ -580,14 +1492,23 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 1, 1): case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 1, 4): case IP_VERSION(10, 3, 0): case IP_VERSION(10, 3, 1): case IP_VERSION(10, 3, 2): case IP_VERSION(10, 3, 3): case IP_VERSION(10, 3, 4): case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 6): + case IP_VERSION(10, 3, 7): amdgpu_device_ip_block_add(adev, &nv_common_ip_block); break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 1): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + amdgpu_device_ip_block_add(adev, &soc21_common_ip_block); + break; default: dev_err(adev->dev, "Failed to add common ip block(GC_HWIP:0x%x)\n", @@ -615,14 +1536,23 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 1, 1): case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 1, 4): case IP_VERSION(10, 3, 0): case IP_VERSION(10, 3, 1): case IP_VERSION(10, 3, 2): case IP_VERSION(10, 3, 3): case IP_VERSION(10, 3, 4): case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 6): + case IP_VERSION(10, 3, 7): amdgpu_device_ip_block_add(adev, &gmc_v10_0_ip_block); break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 1): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + amdgpu_device_ip_block_add(adev, &gmc_v11_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add gmc ip block(GC_HWIP:0x%x)\n", @@ -655,6 +1585,11 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(5, 2, 1): amdgpu_device_ip_block_add(adev, &navi10_ih_ip_block); break; + case IP_VERSION(6, 0, 0): + case IP_VERSION(6, 0, 1): + case IP_VERSION(6, 0, 2): + amdgpu_device_ip_block_add(adev, &ih_v6_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add ih ip block(OSSSYS_HWIP:0x%x)\n", @@ -693,11 +1628,19 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(12, 0, 1): amdgpu_device_ip_block_add(adev, &psp_v12_0_ip_block); break; + case IP_VERSION(13, 0, 0): case IP_VERSION(13, 0, 1): case IP_VERSION(13, 0, 2): case IP_VERSION(13, 0, 3): + case IP_VERSION(13, 0, 5): + case IP_VERSION(13, 0, 7): + case IP_VERSION(13, 0, 8): + case IP_VERSION(13, 0, 10): amdgpu_device_ip_block_add(adev, &psp_v13_0_ip_block); break; + case IP_VERSION(13, 0, 4): + amdgpu_device_ip_block_add(adev, &psp_v13_0_4_ip_block); + break; default: dev_err(adev->dev, "Failed to add psp ip block(MP0_HWIP:0x%x)\n", @@ -734,9 +1677,15 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(12, 0, 1): amdgpu_device_ip_block_add(adev, &smu_v12_0_ip_block); break; + case IP_VERSION(13, 0, 0): case IP_VERSION(13, 0, 1): case IP_VERSION(13, 0, 2): case IP_VERSION(13, 0, 3): + case IP_VERSION(13, 0, 4): + case IP_VERSION(13, 0, 5): + case IP_VERSION(13, 0, 7): + case IP_VERSION(13, 0, 8): + case IP_VERSION(13, 0, 10): amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block); break; default: @@ -752,8 +1701,14 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) { if (adev->enable_virtual_display || amdgpu_sriov_vf(adev)) { amdgpu_device_ip_block_add(adev, &amdgpu_vkms_ip_block); + return 0; + } + + if (!amdgpu_device_has_dc_support(adev)) + return 0; + #if defined(CONFIG_DRM_AMD_DC) - } else if (adev->ip_versions[DCE_HWIP][0]) { + if (adev->ip_versions[DCE_HWIP][0]) { switch (adev->ip_versions[DCE_HWIP][0]) { case IP_VERSION(1, 0, 0): case IP_VERSION(1, 0, 1): @@ -767,6 +1722,11 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(3, 0, 1): case IP_VERSION(3, 1, 2): case IP_VERSION(3, 1, 3): + case IP_VERSION(3, 1, 4): + case IP_VERSION(3, 1, 5): + case IP_VERSION(3, 1, 6): + case IP_VERSION(3, 2, 0): + case IP_VERSION(3, 2, 1): amdgpu_device_ip_block_add(adev, &dm_ip_block); break; default: @@ -788,8 +1748,8 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[DCI_HWIP][0]); return -EINVAL; } -#endif } +#endif return 0; } @@ -810,14 +1770,23 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 1): case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 1, 4): case IP_VERSION(10, 3, 0): case IP_VERSION(10, 3, 2): case IP_VERSION(10, 3, 1): case IP_VERSION(10, 3, 4): case IP_VERSION(10, 3, 5): + case IP_VERSION(10, 3, 6): case IP_VERSION(10, 3, 3): + case IP_VERSION(10, 3, 7): amdgpu_device_ip_block_add(adev, &gfx_v10_0_ip_block); break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 1): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + amdgpu_device_ip_block_add(adev, &gfx_v11_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add gfx ip block(GC_HWIP:0x%x)\n", @@ -850,10 +1819,18 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(5, 2, 2): case IP_VERSION(5, 2, 4): case IP_VERSION(5, 2, 5): + case IP_VERSION(5, 2, 6): case IP_VERSION(5, 2, 3): case IP_VERSION(5, 2, 1): + case IP_VERSION(5, 2, 7): amdgpu_device_ip_block_add(adev, &sdma_v5_2_ip_block); break; + case IP_VERSION(6, 0, 0): + case IP_VERSION(6, 0, 1): + case IP_VERSION(6, 0, 2): + case IP_VERSION(6, 0, 3): + amdgpu_device_ip_block_add(adev, &sdma_v6_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add sdma ip block(SDMA0_HWIP:0x%x)\n", @@ -917,8 +1894,8 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) break; case IP_VERSION(3, 0, 0): case IP_VERSION(3, 0, 16): - case IP_VERSION(3, 0, 64): case IP_VERSION(3, 1, 1): + case IP_VERSION(3, 1, 2): case IP_VERSION(3, 0, 2): case IP_VERSION(3, 0, 192): amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); @@ -928,6 +1905,13 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(3, 0, 33): amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block); break; + case IP_VERSION(4, 0, 0): + case IP_VERSION(4, 0, 2): + case IP_VERSION(4, 0, 4): + amdgpu_device_ip_block_add(adev, &vcn_v4_0_ip_block); + if (!amdgpu_sriov_vf(adev)) + amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add vcn/jpeg ip block(UVD_HWIP:0x%x)\n", @@ -945,16 +1929,31 @@ static int amdgpu_discovery_set_mes_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 1, 1): case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 1, 4): case IP_VERSION(10, 3, 0): case IP_VERSION(10, 3, 1): case IP_VERSION(10, 3, 2): case IP_VERSION(10, 3, 3): case IP_VERSION(10, 3, 4): case IP_VERSION(10, 3, 5): - amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block); + case IP_VERSION(10, 3, 6): + if (amdgpu_mes) { + amdgpu_device_ip_block_add(adev, &mes_v10_1_ip_block); + adev->enable_mes = true; + if (amdgpu_mes_kiq) + adev->enable_mes_kiq = true; + } + break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 1): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + amdgpu_device_ip_block_add(adev, &mes_v11_0_ip_block); + adev->enable_mes = true; + adev->enable_mes_kiq = true; break; default: - break;; + break; } return 0; } @@ -967,6 +1966,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case CHIP_VEGA10: vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->gmc.num_umc = 4; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 0, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 0, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 0); @@ -988,6 +1988,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case CHIP_VEGA12: vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->gmc.num_umc = 4; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 3, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 3, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 1); @@ -1010,6 +2011,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) vega10_reg_base_init(adev); adev->sdma.num_instances = 1; adev->vcn.num_vcn_inst = 1; + adev->gmc.num_umc = 2; if (adev->apu_flags & AMD_APU_IS_RAVEN2) { adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 2, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 2, 0); @@ -1047,6 +2049,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case CHIP_VEGA20: vega20_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->gmc.num_umc = 8; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 0); @@ -1070,6 +2073,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) arct_reg_base_init(adev); adev->sdma.num_instances = 8; adev->vcn.num_vcn_inst = 2; + adev->gmc.num_umc = 8; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 1); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 1); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 1); @@ -1097,6 +2101,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) aldebaran_reg_base_init(adev); adev->sdma.num_instances = 5; adev->vcn.num_vcn_inst = 2; + adev->gmc.num_umc = 4; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 2); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 2); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 4, 0); @@ -1124,11 +2129,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) return -EINVAL; amdgpu_discovery_harvest_ip(adev); - - if (!adev->mman.discovery_bin) { - DRM_ERROR("ip discovery uninitialized\n"); - return -EINVAL; - } + amdgpu_discovery_get_gfx_info(adev); + amdgpu_discovery_get_mall_info(adev); + amdgpu_discovery_get_vcn_info(adev); break; } @@ -1149,6 +2152,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 1, 1): case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 1, 4): case IP_VERSION(10, 3, 0): case IP_VERSION(10, 3, 2): case IP_VERSION(10, 3, 4): @@ -1161,10 +2165,41 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 3, 3): adev->family = AMDGPU_FAMILY_YC; break; + case IP_VERSION(10, 3, 6): + adev->family = AMDGPU_FAMILY_GC_10_3_6; + break; + case IP_VERSION(10, 3, 7): + adev->family = AMDGPU_FAMILY_GC_10_3_7; + break; + case IP_VERSION(11, 0, 0): + case IP_VERSION(11, 0, 2): + case IP_VERSION(11, 0, 3): + adev->family = AMDGPU_FAMILY_GC_11_0_0; + break; + case IP_VERSION(11, 0, 1): + adev->family = AMDGPU_FAMILY_GC_11_0_1; + break; default: return -EINVAL; } + switch (adev->ip_versions[GC_HWIP][0]) { + case IP_VERSION(9, 1, 0): + case IP_VERSION(9, 2, 2): + case IP_VERSION(9, 3, 0): + case IP_VERSION(10, 1, 3): + case IP_VERSION(10, 1, 4): + case IP_VERSION(10, 3, 1): + case IP_VERSION(10, 3, 3): + case IP_VERSION(10, 3, 6): + case IP_VERSION(10, 3, 7): + case IP_VERSION(11, 0, 1): + adev->flags |= AMD_IS_APU; + break; + default: + break; + } + if (adev->ip_versions[XGMI_HWIP][0] == IP_VERSION(4, 8, 0)) adev->gmc.xgmi.supported = true; @@ -1183,16 +2218,15 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) break; case IP_VERSION(7, 4, 0): case IP_VERSION(7, 4, 1): - adev->nbio.funcs = &nbio_v7_4_funcs; - adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg; - break; case IP_VERSION(7, 4, 4): adev->nbio.funcs = &nbio_v7_4_funcs; - adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg_ald; + adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg; break; case IP_VERSION(7, 2, 0): case IP_VERSION(7, 2, 1): + case IP_VERSION(7, 3, 0): case IP_VERSION(7, 5, 0): + case IP_VERSION(7, 5, 1): adev->nbio.funcs = &nbio_v7_2_funcs; adev->nbio.hdp_flush_reg = &nbio_v7_2_hdp_flush_reg; break; @@ -1200,15 +2234,24 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(2, 3, 0): case IP_VERSION(2, 3, 1): case IP_VERSION(2, 3, 2): - adev->nbio.funcs = &nbio_v2_3_funcs; - adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg; - break; case IP_VERSION(3, 3, 0): case IP_VERSION(3, 3, 1): case IP_VERSION(3, 3, 2): case IP_VERSION(3, 3, 3): adev->nbio.funcs = &nbio_v2_3_funcs; - adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg_sc; + adev->nbio.hdp_flush_reg = &nbio_v2_3_hdp_flush_reg; + break; + case IP_VERSION(4, 3, 0): + case IP_VERSION(4, 3, 1): + if (amdgpu_sriov_vf(adev)) + adev->nbio.funcs = &nbio_v4_3_sriov_funcs; + else + adev->nbio.funcs = &nbio_v4_3_funcs; + adev->nbio.hdp_flush_reg = &nbio_v4_3_hdp_flush_reg; + break; + case IP_VERSION(7, 7, 0): + adev->nbio.funcs = &nbio_v7_7_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_7_hdp_flush_reg; break; default: break; @@ -1233,6 +2276,13 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(5, 2, 0): adev->hdp.funcs = &hdp_v5_0_funcs; break; + case IP_VERSION(5, 2, 1): + adev->hdp.funcs = &hdp_v5_2_funcs; + break; + case IP_VERSION(6, 0, 0): + case IP_VERSION(6, 0, 1): + adev->hdp.funcs = &hdp_v6_0_funcs; + break; default: break; } @@ -1275,11 +2325,28 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 11): case IP_VERSION(11, 5, 0): case IP_VERSION(13, 0, 1): + case IP_VERSION(13, 0, 9): + case IP_VERSION(13, 0, 10): adev->smuio.funcs = &smuio_v11_0_6_funcs; break; case IP_VERSION(13, 0, 2): adev->smuio.funcs = &smuio_v13_0_funcs; break; + case IP_VERSION(13, 0, 6): + case IP_VERSION(13, 0, 8): + adev->smuio.funcs = &smuio_v13_0_6_funcs; + break; + default: + break; + } + + switch (adev->ip_versions[LSDMA_HWIP][0]) { + case IP_VERSION(6, 0, 0): + case IP_VERSION(6, 0, 1): + case IP_VERSION(6, 0, 2): + case IP_VERSION(6, 0, 3): + adev->lsdma.funcs = &lsdma_v6_0_funcs; + break; default: break; } @@ -1330,8 +2397,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) if (r) return r; - if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && - !amdgpu_sriov_vf(adev)) { + if ((adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT && + !amdgpu_sriov_vf(adev)) || + (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO && amdgpu_dpm == 1)) { r = amdgpu_discovery_set_smu_ip_blocks(adev); if (r) return r; @@ -1341,11 +2409,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) if (r) return r; - if (adev->enable_mes) { - r = amdgpu_discovery_set_mes_ip_blocks(adev); - if (r) - return r; - } + r = amdgpu_discovery_set_mes_ip_blocks(adev); + if (r) + return r; return 0; } |