diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 1382 |
1 files changed, 1086 insertions, 296 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index b719852daa07..fa2a22dfa048 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -1,5 +1,5 @@ /* - * Copyright 2018 Advanced Micro Devices, Inc. + * Copyright 2018-2024 Advanced Micro Devices, Inc. All rights reserved. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -27,20 +27,27 @@ #include "amdgpu_discovery.h" #include "soc15_hw_ip.h" #include "discovery.h" +#include "amdgpu_ras.h" #include "soc15.h" #include "gfx_v9_0.h" +#include "gfx_v9_4_3.h" #include "gmc_v9_0.h" #include "df_v1_7.h" #include "df_v3_6.h" #include "df_v4_3.h" +#include "df_v4_6_2.h" +#include "df_v4_15.h" #include "nbio_v6_1.h" #include "nbio_v7_0.h" #include "nbio_v7_4.h" +#include "nbio_v7_9.h" +#include "nbio_v7_11.h" #include "hdp_v4_0.h" #include "vega10_ih.h" #include "vega20_ih.h" #include "sdma_v4_0.h" +#include "sdma_v4_4_2.h" #include "uvd_v7_0.h" #include "vce_v4_0.h" #include "vcn_v1_0.h" @@ -49,43 +56,77 @@ #include "smuio_v9_0.h" #include "gmc_v10_0.h" #include "gmc_v11_0.h" +#include "gmc_v12_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 "nbif_v6_3_1.h" #include "hdp_v5_0.h" #include "hdp_v5_2.h" #include "hdp_v6_0.h" +#include "hdp_v7_0.h" #include "nv.h" #include "soc21.h" +#include "soc24.h" #include "navi10_ih.h" #include "ih_v6_0.h" +#include "ih_v6_1.h" +#include "ih_v7_0.h" #include "gfx_v10_0.h" #include "gfx_v11_0.h" +#include "gfx_v12_0.h" #include "sdma_v5_0.h" #include "sdma_v5_2.h" #include "sdma_v6_0.h" +#include "sdma_v7_0.h" #include "lsdma_v6_0.h" +#include "lsdma_v7_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 "vcn_v4_0_3.h" +#include "jpeg_v4_0_3.h" +#include "vcn_v4_0_5.h" +#include "jpeg_v4_0_5.h" #include "amdgpu_vkms.h" -#include "mes_v10_1.h" #include "mes_v11_0.h" +#include "mes_v12_0.h" #include "smuio_v11_0.h" #include "smuio_v11_0_6.h" #include "smuio_v13_0.h" +#include "smuio_v13_0_3.h" #include "smuio_v13_0_6.h" +#include "smuio_v14_0_2.h" +#include "vcn_v5_0_0.h" +#include "vcn_v5_0_1.h" +#include "jpeg_v5_0_0.h" +#include "jpeg_v5_0_1.h" +#include "amdgpu_ras_mgr.h" + +#include "amdgpu_vpe.h" +#if defined(CONFIG_DRM_AMD_ISP) +#include "amdgpu_isp.h" +#endif -#define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin" -MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY); - +MODULE_FIRMWARE("amdgpu/ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/vega10_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/vega12_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/vega20_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/raven_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/raven2_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/picasso_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/arcturus_ip_discovery.bin"); +MODULE_FIRMWARE("amdgpu/aldebaran_ip_discovery.bin"); + +#define mmIP_DISCOVERY_VERSION 0x16A00 #define mmRCC_CONFIG_MEMSIZE 0xde3 +#define mmMP0_SMN_C2PMSG_33 0x16061 #define mmMM_INDEX 0x0 #define mmMM_INDEX_HI 0x6 #define mmMM_DATA 0x1 @@ -167,6 +208,7 @@ static const char *hw_id_names[HW_ID_MAX] = { [XGMI_HWID] = "XGMI", [XGBE_HWID] = "XGBE", [MP0_HWID] = "MP0", + [VPE_HWID] = "VPE", }; static int hw_id_map[MAX_HWIP] = { @@ -196,37 +238,117 @@ static int hw_id_map[MAX_HWIP] = { [XGMI_HWIP] = XGMI_HWID, [DCI_HWIP] = DCI_HWID, [PCIE_HWIP] = PCIE_HWID, + [VPE_HWIP] = VPE_HWID, + [ISP_HWIP] = ISP_HWID, }; -static int amdgpu_discovery_read_binary_from_vram(struct amdgpu_device *adev, uint8_t *binary) +static int amdgpu_discovery_read_binary_from_sysmem(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; + u64 tmr_offset, tmr_size, pos; + void *discv_regn; + int ret; - amdgpu_device_vram_access(adev, pos, (uint32_t *)binary, - adev->mman.discovery_tmr_size, false); - return 0; + ret = amdgpu_acpi_get_tmr_info(adev, &tmr_offset, &tmr_size); + if (ret) + return ret; + + pos = tmr_offset + tmr_size - DISCOVERY_TMR_OFFSET; + + /* This region is read-only and reserved from system use */ + discv_regn = memremap(pos, adev->discovery.size, MEMREMAP_WC); + if (discv_regn) { + memcpy(binary, discv_regn, adev->discovery.size); + memunmap(discv_regn); + return 0; + } + + return -ENOENT; } -static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, uint8_t *binary) +#define IP_DISCOVERY_V2 2 +#define IP_DISCOVERY_V4 4 + +static int amdgpu_discovery_read_binary_from_mem(struct amdgpu_device *adev, + uint8_t *binary) { - const struct firmware *fw; - const char *fw_name; - int r; + bool sz_valid = true; + uint64_t vram_size; + int i, ret = 0; + u32 msg; + + if (!amdgpu_sriov_vf(adev)) { + /* It can take up to two second for IFWI init to complete on some dGPUs, + * but generally it should be in the 60-100ms range. Normally this starts + * as soon as the device gets power so by the time the OS loads this has long + * completed. However, when a card is hotplugged via e.g., USB4, we need to + * wait for this to complete. Once the C2PMSG is updated, we can + * continue. + */ - 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; + for (i = 0; i < 2000; i++) { + msg = RREG32(mmMP0_SMN_C2PMSG_33); + if (msg & 0x80000000) + break; + msleep(1); + } } - r = request_firmware(&fw, fw_name, adev->dev); + vram_size = RREG32(mmRCC_CONFIG_MEMSIZE); + if (!vram_size || vram_size == U32_MAX) + sz_valid = false; + else + vram_size <<= 20; + + /* + * If in VRAM, discovery TMR is marked for reservation. If it is in system mem, + * then it is not required to be reserved. + */ + if (sz_valid) { + if (amdgpu_sriov_vf(adev) && adev->virt.is_dynamic_crit_regn_enabled) { + /* For SRIOV VFs with dynamic critical region enabled, + * we will get the IPD binary via below call. + * If dynamic critical is disabled, fall through to normal seq. + */ + if (amdgpu_virt_get_dynamic_data_info(adev, + AMD_SRIOV_MSG_IPD_TABLE_ID, binary, + &adev->discovery.size)) { + dev_err(adev->dev, + "failed to read discovery info from dynamic critical region."); + ret = -EINVAL; + goto exit; + } + } else { + uint64_t pos = vram_size - DISCOVERY_TMR_OFFSET; + + amdgpu_device_vram_access(adev, pos, (uint32_t *)binary, + adev->discovery.size, false); + adev->discovery.reserve_tmr = true; + } + } else { + ret = amdgpu_discovery_read_binary_from_sysmem(adev, binary); + } + + if (ret) + dev_err(adev->dev, + "failed to read discovery info from memory, vram size read: %llx", + vram_size); +exit: + return ret; +} + +static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, + uint8_t *binary, + const char *fw_name) +{ + const struct firmware *fw; + int r; + + r = firmware_request_nowarn(&fw, fw_name, adev->dev); if (r) { - dev_err(adev->dev, "can't load firmware \"%s\"\n", - fw_name); + if (amdgpu_discovery == 2) + dev_err(adev->dev, "can't load firmware \"%s\"\n", fw_name); + else + drm_info(&adev->ddev, "Optional firmware \"%s\" was not found\n", fw_name); return r; } @@ -267,8 +389,8 @@ 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))) { + if ((amdgpu_ip_version(adev, UVD_HWIP, 1) == IP_VERSION(3, 0, 1)) && + (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(10, 3, 2))) { switch (adev->pdev->revision) { case 0xC1: case 0xC2: @@ -278,6 +400,7 @@ static void amdgpu_discovery_harvest_config_quirk(struct amdgpu_device *adev) case 0xCF: case 0xDF: adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + adev->vcn.inst_mask &= ~AMDGPU_VCN_HARVEST_VCN1; break; default: break; @@ -285,58 +408,118 @@ static void amdgpu_discovery_harvest_config_quirk(struct amdgpu_device *adev) } } +static int amdgpu_discovery_verify_npsinfo(struct amdgpu_device *adev, + struct binary_header *bhdr) +{ + uint8_t *discovery_bin = adev->discovery.bin; + struct table_info *info; + uint16_t checksum; + uint16_t offset; + + info = &bhdr->table_list[NPS_INFO]; + offset = le16_to_cpu(info->offset); + checksum = le16_to_cpu(info->checksum); + + struct nps_info_header *nhdr = + (struct nps_info_header *)(discovery_bin + offset); + + if (le32_to_cpu(nhdr->table_id) != NPS_INFO_TABLE_ID) { + dev_dbg(adev->dev, "invalid ip discovery nps info table id\n"); + return -EINVAL; + } + + if (!amdgpu_discovery_verify_checksum(discovery_bin + offset, + le32_to_cpu(nhdr->size_bytes), + checksum)) { + dev_dbg(adev->dev, "invalid nps info data table checksum\n"); + return -EINVAL; + } + + return 0; +} + +static const char *amdgpu_discovery_get_fw_name(struct amdgpu_device *adev) +{ + if (amdgpu_discovery == 2) { + /* Assume there is valid discovery TMR in VRAM even if binary is sideloaded */ + adev->discovery.reserve_tmr = true; + return "amdgpu/ip_discovery.bin"; + } + + switch (adev->asic_type) { + case CHIP_VEGA10: + return "amdgpu/vega10_ip_discovery.bin"; + case CHIP_VEGA12: + return "amdgpu/vega12_ip_discovery.bin"; + case CHIP_RAVEN: + if (adev->apu_flags & AMD_APU_IS_RAVEN2) + return "amdgpu/raven2_ip_discovery.bin"; + else if (adev->apu_flags & AMD_APU_IS_PICASSO) + return "amdgpu/picasso_ip_discovery.bin"; + else + return "amdgpu/raven_ip_discovery.bin"; + case CHIP_VEGA20: + return "amdgpu/vega20_ip_discovery.bin"; + case CHIP_ARCTURUS: + return "amdgpu/arcturus_ip_discovery.bin"; + case CHIP_ALDEBARAN: + return "amdgpu/aldebaran_ip_discovery.bin"; + default: + return NULL; + } +} + static int amdgpu_discovery_init(struct amdgpu_device *adev) { struct table_info *info; struct binary_header *bhdr; + uint8_t *discovery_bin; + const char *fw_name; uint16_t offset; uint16_t size; uint16_t checksum; int r; - adev->mman.discovery_tmr_size = DISCOVERY_TMR_SIZE; - adev->mman.discovery_bin = kzalloc(adev->mman.discovery_tmr_size, GFP_KERNEL); - if (!adev->mman.discovery_bin) + adev->discovery.bin = kzalloc(DISCOVERY_TMR_SIZE, GFP_KERNEL); + if (!adev->discovery.bin) return -ENOMEM; + adev->discovery.size = DISCOVERY_TMR_SIZE; + adev->discovery.debugfs_blob.data = adev->discovery.bin; + adev->discovery.debugfs_blob.size = adev->discovery.size; + + discovery_bin = adev->discovery.bin; + /* Read from file if it is the preferred option */ + fw_name = amdgpu_discovery_get_fw_name(adev); + if (fw_name != NULL) { + drm_dbg(&adev->ddev, "use ip discovery information from file"); + r = amdgpu_discovery_read_binary_from_file(adev, discovery_bin, + fw_name); + if (r) + goto out; + } else { + drm_dbg(&adev->ddev, "use ip discovery information from memory"); + r = amdgpu_discovery_read_binary_from_mem(adev, discovery_bin); + if (r) + goto out; + } - 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"); + /* check the ip discovery binary signature */ + if (!amdgpu_discovery_verify_binary_signature(discovery_bin)) { + dev_err(adev->dev, + "get invalid ip discovery binary signature\n"); r = -EINVAL; goto out; } - if (!amdgpu_discovery_verify_binary_signature(adev->mman.discovery_bin) || amdgpu_discovery == 2) { - /* ignore the discovery binary from vram if discovery=2 in kernel module parameter */ - if (amdgpu_discovery == 2) - dev_info(adev->dev,"force read ip discovery binary from file"); - else - 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) { - 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; + bhdr = (struct binary_header *)discovery_bin; offset = offsetof(struct binary_header, binary_checksum) + sizeof(bhdr->binary_checksum); size = le16_to_cpu(bhdr->binary_size) - offset; checksum = le16_to_cpu(bhdr->binary_checksum); - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - size, checksum)) { + if (!amdgpu_discovery_verify_checksum(discovery_bin + offset, size, + checksum)) { dev_err(adev->dev, "invalid ip discovery binary checksum\n"); r = -EINVAL; goto out; @@ -348,15 +531,16 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (offset) { struct ip_discovery_header *ihdr = - (struct ip_discovery_header *)(adev->mman.discovery_bin + offset); + (struct ip_discovery_header *)(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)) { + if (!amdgpu_discovery_verify_checksum(discovery_bin + offset, + le16_to_cpu(ihdr->size), + checksum)) { dev_err(adev->dev, "invalid ip discovery data table checksum\n"); r = -EINVAL; goto out; @@ -369,7 +553,7 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (offset) { struct gpu_info_header *ghdr = - (struct gpu_info_header *)(adev->mman.discovery_bin + offset); + (struct gpu_info_header *)(discovery_bin + offset); if (le32_to_cpu(ghdr->table_id) != GC_TABLE_ID) { dev_err(adev->dev, "invalid ip discovery gc table id\n"); @@ -377,8 +561,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) goto out; } - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - le32_to_cpu(ghdr->size), checksum)) { + if (!amdgpu_discovery_verify_checksum(discovery_bin + offset, + le32_to_cpu(ghdr->size), + checksum)) { dev_err(adev->dev, "invalid gc data table checksum\n"); r = -EINVAL; goto out; @@ -391,7 +576,7 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (offset) { struct harvest_info_header *hhdr = - (struct harvest_info_header *)(adev->mman.discovery_bin + offset); + (struct harvest_info_header *)(discovery_bin + offset); if (le32_to_cpu(hhdr->signature) != HARVEST_TABLE_SIGNATURE) { dev_err(adev->dev, "invalid ip discovery harvest table signature\n"); @@ -399,8 +584,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) goto out; } - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - sizeof(struct harvest_table), checksum)) { + if (!amdgpu_discovery_verify_checksum( + discovery_bin + offset, + sizeof(struct harvest_table), checksum)) { dev_err(adev->dev, "invalid harvest data table checksum\n"); r = -EINVAL; goto out; @@ -413,7 +599,7 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (offset) { struct vcn_info_header *vhdr = - (struct vcn_info_header *)(adev->mman.discovery_bin + offset); + (struct vcn_info_header *)(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"); @@ -421,8 +607,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) goto out; } - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - le32_to_cpu(vhdr->size_bytes), checksum)) { + if (!amdgpu_discovery_verify_checksum( + discovery_bin + offset, + le32_to_cpu(vhdr->size_bytes), checksum)) { dev_err(adev->dev, "invalid vcn data table checksum\n"); r = -EINVAL; goto out; @@ -435,7 +622,7 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) if (0 && offset) { struct mall_info_header *mhdr = - (struct mall_info_header *)(adev->mman.discovery_bin + offset); + (struct mall_info_header *)(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"); @@ -443,8 +630,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) goto out; } - if (!amdgpu_discovery_verify_checksum(adev->mman.discovery_bin + offset, - le32_to_cpu(mhdr->size_bytes), checksum)) { + if (!amdgpu_discovery_verify_checksum( + discovery_bin + offset, + le32_to_cpu(mhdr->size_bytes), checksum)) { dev_err(adev->dev, "invalid mall data table checksum\n"); r = -EINVAL; goto out; @@ -454,9 +642,11 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev) return 0; out: - kfree(adev->mman.discovery_bin); - adev->mman.discovery_bin = NULL; - + kfree(adev->discovery.bin); + adev->discovery.bin = NULL; + if ((amdgpu_discovery != 2) && + (RREG32(mmIP_DISCOVERY_VERSION) == 4)) + amdgpu_ras_query_boot_status(adev, 4); return r; } @@ -465,20 +655,23 @@ 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; + kfree(adev->discovery.bin); + adev->discovery.bin = NULL; } -static int amdgpu_discovery_validate_ip(const struct ip *ip) +static int amdgpu_discovery_validate_ip(struct amdgpu_device *adev, + uint8_t instance, uint16_t hw_id) { - if (ip->number_instance >= HWIP_MAX_INSTANCE) { - DRM_ERROR("Unexpected number_instance (%d) from ip discovery blob\n", - ip->number_instance); + if (instance >= HWIP_MAX_INSTANCE) { + dev_err(adev->dev, + "Unexpected instance_number (%d) from ip discovery blob\n", + instance); return -EINVAL; } - if (le16_to_cpu(ip->hw_id) >= HW_ID_MAX) { - DRM_ERROR("Unexpected hw_id (%d) from ip discovery blob\n", - le16_to_cpu(ip->hw_id)); + if (hw_id >= HW_ID_MAX) { + dev_err(adev->dev, + "Unexpected hw_id (%d) from ip discovery blob\n", + hw_id); return -EINVAL; } @@ -488,49 +681,64 @@ static int amdgpu_discovery_validate_ip(const struct ip *ip) static void amdgpu_discovery_read_harvest_bit_per_ip(struct amdgpu_device *adev, uint32_t *vcn_harvest_count) { + uint8_t *discovery_bin = adev->discovery.bin; 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; + uint16_t hw_id; + uint8_t inst; 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)); + bhdr = (struct binary_header *)discovery_bin; + ihdr = (struct ip_discovery_header + *)(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); + dhdr = (struct die_header *)(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)) + ip = (struct ip *)(discovery_bin + ip_offset); + inst = ip->number_instance; + hw_id = le16_to_cpu(ip->hw_id); + if (amdgpu_discovery_validate_ip(adev, inst, hw_id)) goto next_ip; - if (le16_to_cpu(ip->harvest) == 1) { - switch (le16_to_cpu(ip->hw_id)) { + if (ip->harvest == 1) { + switch (hw_id) { case VCN_HWID: (*vcn_harvest_count)++; - if (ip->number_instance == 0) + if (inst == 0) { adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN0; - else + adev->vcn.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN0; + adev->jpeg.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN0; + } else { adev->vcn.harvest_config |= AMDGPU_VCN_HARVEST_VCN1; + adev->vcn.inst_mask &= + ~AMDGPU_VCN_HARVEST_VCN1; + adev->jpeg.inst_mask &= + ~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); + ip_offset += struct_size(ip, base_address, + ip->num_base_address); } } } @@ -539,12 +747,14 @@ static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev, uint32_t *vcn_harvest_count, uint32_t *umc_harvest_count) { + uint8_t *discovery_bin = adev->discovery.bin; struct binary_header *bhdr; struct harvest_table *harvest_info; u16 offset; int i; + uint32_t umc_harvest_config = 0; - bhdr = (struct binary_header *)adev->mman.discovery_bin; + bhdr = (struct binary_header *)discovery_bin; offset = le16_to_cpu(bhdr->table_list[HARVEST_INFO].offset); if (!offset) { @@ -552,7 +762,7 @@ static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev, return; } - harvest_info = (struct harvest_table *)(adev->mman.discovery_bin + offset); + harvest_info = (struct harvest_table *)(discovery_bin + offset); for (i = 0; i < 32; i++) { if (le16_to_cpu(harvest_info->list[i].hw_id) == 0) @@ -561,21 +771,45 @@ static void amdgpu_discovery_read_from_harvest_table(struct amdgpu_device *adev, 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; + adev->vcn.harvest_config |= + (1 << harvest_info->list[i].number_instance); + adev->jpeg.harvest_config |= + (1 << harvest_info->list[i].number_instance); + + adev->vcn.inst_mask &= + ~(1U << harvest_info->list[i].number_instance); + adev->jpeg.inst_mask &= + ~(1U << harvest_info->list[i].number_instance); break; case DMU_HWID: adev->harvest_ip_mask |= AMD_HARVEST_IP_DMU_MASK; break; case UMC_HWID: + umc_harvest_config |= + 1 << (le16_to_cpu(harvest_info->list[i].number_instance)); (*umc_harvest_count)++; break; + case GC_HWID: + adev->gfx.xcc_mask &= + ~(1U << harvest_info->list[i].number_instance); + break; + case SDMA0_HWID: + adev->sdma.sdma_mask &= + ~(1U << harvest_info->list[i].number_instance); + break; +#if defined(CONFIG_DRM_AMD_ISP) + case ISP_HWID: + adev->isp.harvest_config |= + ~(1U << harvest_info->list[i].number_instance); + break; +#endif default: break; } } + + adev->umc.active_mask = ((1 << adev->umc.node_inst_num) - 1) & + ~umc_harvest_config; } /* ================================================== */ @@ -589,7 +823,7 @@ struct ip_hw_instance { u8 harvest; int num_base_addresses; - u32 base_addr[]; + u32 base_addr[] __counted_by(num_base_addresses); }; struct ip_hw_id { @@ -705,7 +939,7 @@ static void ip_hw_instance_release(struct kobject *kobj) kfree(ip_hw_instance); } -static struct kobj_type ip_hw_instance_ktype = { +static const 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, @@ -724,7 +958,7 @@ static void ip_hw_id_release(struct kobject *kobj) kfree(ip_hw_id); } -static struct kobj_type ip_hw_id_ktype = { +static const struct kobj_type ip_hw_id_ktype = { .release = ip_hw_id_release, .sysfs_ops = &kobj_sysfs_ops, }; @@ -787,18 +1021,18 @@ static const struct sysfs_ops ip_die_entry_sysfs_ops = { .show = ip_die_entry_attr_show, }; -static struct kobj_type ip_die_entry_ktype = { +static const 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 = { +static const struct kobj_type die_kobj_ktype = { .release = die_kobj_release, .sysfs_ops = &kobj_sysfs_ops, }; -static struct kobj_type ip_discovery_ktype = { +static const struct kobj_type ip_discovery_ktype = { .release = ip_disc_release, .sysfs_ops = &kobj_sysfs_ops, }; @@ -824,15 +1058,51 @@ static void ip_disc_release(struct kobject *kobj) kobj); struct amdgpu_device *adev = ip_top->adev; - adev->ip_top = NULL; kfree(ip_top); + adev->discovery.ip_top = NULL; +} + +static uint8_t amdgpu_discovery_get_harvest_info(struct amdgpu_device *adev, + uint16_t hw_id, uint8_t inst) +{ + uint8_t harvest = 0; + + /* Until a uniform way is figured, get mask based on hwid */ + switch (hw_id) { + case VCN_HWID: + /* VCN vs UVD+VCE */ + if (!amdgpu_ip_version(adev, VCE_HWIP, 0)) + harvest = ((1 << inst) & adev->vcn.inst_mask) == 0; + break; + case DMU_HWID: + if (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK) + harvest = 0x1; + break; + case UMC_HWID: + /* TODO: It needs another parsing; for now, ignore.*/ + break; + case GC_HWID: + harvest = ((1 << inst) & adev->gfx.xcc_mask) == 0; + break; + case SDMA0_HWID: + harvest = ((1 << inst) & adev->sdma.sdma_mask) == 0; + break; + default: + break; + } + + return harvest; } 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) + const size_t _ip_offset, const int num_ips, + bool reg_base_64) { + uint8_t *discovery_bin = adev->discovery.bin; int ii, jj, kk, res; + uint16_t hw_id; + uint8_t inst; DRM_DEBUG("num_ips:%d", num_ips); @@ -844,12 +1114,14 @@ static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, size_t ip_offset = _ip_offset; for (jj = 0; jj < num_ips; jj++) { - struct ip *ip; + struct ip_v4 *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) + ip = (struct ip_v4 *)(discovery_bin + ip_offset); + inst = ip->instance_number; + hw_id = le16_to_cpu(ip->hw_id); + if (amdgpu_discovery_validate_ip(adev, inst, hw_id) || + hw_id != ii) goto next_ip; DRM_DEBUG("match:%d @ ip_offset:%zu", ii, ip_offset); @@ -895,22 +1167,35 @@ static int amdgpu_discovery_sysfs_ips(struct amdgpu_device *adev, 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->num_instance = ip->instance_number; 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->harvest = + amdgpu_discovery_get_harvest_info( + adev, ip_hw_instance->hw_id, + ip_hw_instance->num_instance); 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]; + for (kk = 0; kk < ip_hw_instance->num_base_addresses; kk++) { + if (reg_base_64) + ip_hw_instance->base_addr[kk] = + lower_32_bits(le64_to_cpu(ip->base_address_64[kk])) & 0x3FFFFFFF; + else + 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); + if (reg_base_64) + ip_offset += struct_size(ip, base_address_64, + ip->num_base_address); + else + ip_offset += struct_size(ip, base_address, + ip->num_base_address); } } @@ -919,17 +1204,20 @@ next_ip: static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev) { + struct ip_discovery_top *ip_top = adev->discovery.ip_top; + uint8_t *discovery_bin = adev->discovery.bin; struct binary_header *bhdr; struct ip_discovery_header *ihdr; struct die_header *dhdr; - struct kset *die_kset = &adev->ip_top->die_kset; + struct kset *die_kset = &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)); + bhdr = (struct binary_header *)discovery_bin; + ihdr = (struct ip_discovery_header + *)(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); @@ -938,7 +1226,7 @@ static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev) 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); + dhdr = (struct die_header *)(discovery_bin + die_offset); num_ips = le16_to_cpu(dhdr->num_ips); ip_offset = die_offset + sizeof(*dhdr); @@ -964,7 +1252,7 @@ static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev) return res; } - amdgpu_discovery_sysfs_ips(adev, ip_die_entry, ip_offset, num_ips); + amdgpu_discovery_sysfs_ips(adev, ip_die_entry, ip_offset, num_ips, !!ihdr->base_addr_64_bit); } return 0; @@ -972,27 +1260,32 @@ static int amdgpu_discovery_sysfs_recurse(struct amdgpu_device *adev) static int amdgpu_discovery_sysfs_init(struct amdgpu_device *adev) { + uint8_t *discovery_bin = adev->discovery.bin; + struct ip_discovery_top *ip_top; struct kset *die_kset; int res, ii; - adev->ip_top = kzalloc(sizeof(*adev->ip_top), GFP_KERNEL); - if (!adev->ip_top) - return -ENOMEM; + if (!discovery_bin) + return -EINVAL; - adev->ip_top->adev = adev; + ip_top = kzalloc(sizeof(*ip_top), GFP_KERNEL); + if (!ip_top) + return -ENOMEM; - res = kobject_init_and_add(&adev->ip_top->kobj, &ip_discovery_ktype, + ip_top->adev = adev; + adev->discovery.ip_top = ip_top; + res = kobject_init_and_add(&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; + die_kset = &ip_top->die_kset; kobject_set_name(&die_kset->kobj, "%s", "die"); - die_kset->kobj.parent = &adev->ip_top->kobj; + die_kset->kobj.parent = &ip_top->kobj; die_kset->kobj.ktype = &die_kobj_ktype; - res = kset_register(&adev->ip_top->die_kset); + res = kset_register(&ip_top->die_kset); if (res) { DRM_ERROR("Couldn't register die_kset"); goto Err; @@ -1006,7 +1299,7 @@ static int amdgpu_discovery_sysfs_init(struct amdgpu_device *adev) return res; Err: - kobject_put(&adev->ip_top->kobj); + kobject_put(&ip_top->kobj); return res; } @@ -1051,10 +1344,11 @@ static void amdgpu_discovery_sysfs_die_free(struct ip_die_entry *ip_die_entry) static void amdgpu_discovery_sysfs_fini(struct amdgpu_device *adev) { + struct ip_discovery_top *ip_top = adev->discovery.ip_top; struct list_head *el, *tmp; struct kset *die_kset; - die_kset = &adev->ip_top->die_kset; + die_kset = &ip_top->die_kset; spin_lock(&die_kset->list_lock); list_for_each_prev_safe(el, tmp, &die_kset->list) { list_del_init(el); @@ -1063,43 +1357,51 @@ static void amdgpu_discovery_sysfs_fini(struct amdgpu_device *adev) 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); + kobject_put(&ip_top->die_kset.kobj); + kobject_put(&ip_top->kobj); } /* ================================================== */ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) { + uint8_t num_base_address, subrev, variant; struct binary_header *bhdr; struct ip_discovery_header *ihdr; struct die_header *dhdr; - struct ip *ip; + uint8_t *discovery_bin; + struct ip_v4 *ip; uint16_t die_offset; uint16_t ip_offset; uint16_t num_dies; + uint32_t wafl_ver; uint16_t num_ips; - uint8_t num_base_address; + uint16_t hw_id; + uint8_t inst; int hw_ip; int i, j, k; int r; r = amdgpu_discovery_init(adev); - if (r) { - DRM_ERROR("amdgpu_discovery_init failed\n"); + if (r) return r; - } - - 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)); + discovery_bin = adev->discovery.bin; + wafl_ver = 0; + adev->gfx.xcc_mask = 0; + adev->sdma.sdma_mask = 0; + adev->vcn.inst_mask = 0; + adev->jpeg.inst_mask = 0; + bhdr = (struct binary_header *)discovery_bin; + ihdr = (struct ip_discovery_header + *)(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 (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); + dhdr = (struct die_header *)(discovery_bin + die_offset); num_ips = le16_to_cpu(dhdr->num_ips); ip_offset = die_offset + sizeof(*dhdr); @@ -1113,9 +1415,11 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) le16_to_cpu(dhdr->die_id), num_ips); for (j = 0; j < num_ips; j++) { - ip = (struct ip *)(adev->mman.discovery_bin + ip_offset); + ip = (struct ip_v4 *)(discovery_bin + ip_offset); - if (amdgpu_discovery_validate_ip(ip)) + inst = ip->instance_number; + hw_id = le16_to_cpu(ip->hw_id); + if (amdgpu_discovery_validate_ip(adev, inst, hw_id)) goto next_ip; num_base_address = ip->num_base_address; @@ -1123,7 +1427,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) DRM_DEBUG("%s(%d) #%d v%d.%d.%d:\n", hw_id_names[le16_to_cpu(ip->hw_id)], le16_to_cpu(ip->hw_id), - ip->number_instance, + ip->instance_number, ip->major, ip->minor, ip->revision); @@ -1134,45 +1438,88 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) * 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) + if (adev->vcn.num_vcn_inst < + AMDGPU_MAX_VCN_INSTANCES) { + adev->vcn.inst[adev->vcn.num_vcn_inst].vcn_config = + ip->revision & 0xc0; adev->vcn.num_vcn_inst++; - else + adev->vcn.inst_mask |= + (1U << ip->instance_number); + adev->jpeg.inst_mask |= + (1U << ip->instance_number); + } else { dev_err(adev->dev, "Too many VCN instances: %d vs %d\n", adev->vcn.num_vcn_inst + 1, AMDGPU_MAX_VCN_INSTANCES); + } + ip->revision &= ~0xc0; } 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) { - if (adev->sdma.num_instances < AMDGPU_MAX_SDMA_INSTANCES) + if (adev->sdma.num_instances < + AMDGPU_MAX_SDMA_INSTANCES) { adev->sdma.num_instances++; - else + adev->sdma.sdma_mask |= + (1U << ip->instance_number); + } 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) == VPE_HWID) { + if (adev->vpe.num_instances < AMDGPU_MAX_VPE_INSTANCES) + adev->vpe.num_instances++; + else + dev_err(adev->dev, "Too many VPE instances: %d vs %d\n", + adev->vpe.num_instances + 1, + AMDGPU_MAX_VPE_INSTANCES); } - if (le16_to_cpu(ip->hw_id) == UMC_HWID) + if (le16_to_cpu(ip->hw_id) == UMC_HWID) { adev->gmc.num_umc++; + adev->umc.node_inst_num++; + } + + if (le16_to_cpu(ip->hw_id) == GC_HWID) + adev->gfx.xcc_mask |= + (1U << ip->instance_number); + + if (!wafl_ver && le16_to_cpu(ip->hw_id) == WAFLC_HWID) + wafl_ver = IP_VERSION_FULL(ip->major, ip->minor, + ip->revision, 0, 0); for (k = 0; k < num_base_address; k++) { /* * convert the endianness of base addresses in place, * so that we don't need to convert them when accessing adev->reg_offset. */ - ip->base_address[k] = le32_to_cpu(ip->base_address[k]); + if (ihdr->base_addr_64_bit) + /* Truncate the 64bit base address from ip discovery + * and only store lower 32bit ip base in reg_offset[]. + * Bits > 32 follows ASIC specific format, thus just + * discard them and handle it within specific ASIC. + * By this way reg_offset[] and related helpers can + * stay unchanged. + * The base address is in dwords, thus clear the + * highest 2 bits to store. + */ + ip->base_address[k] = + lower_32_bits(le64_to_cpu(ip->base_address_64[k])) & 0x3FFFFFFF; + else + ip->base_address[k] = le32_to_cpu(ip->base_address[k]); DRM_DEBUG("\t0x%08x\n", ip->base_address[k]); } for (hw_ip = 0; hw_ip < MAX_HWIP; hw_ip++) { - if (hw_id_map[hw_ip] == le16_to_cpu(ip->hw_id)) { + if (hw_id_map[hw_ip] == le16_to_cpu(ip->hw_id) && + hw_id_map[hw_ip] != 0) { DRM_DEBUG("set register base offset for %s\n", hw_id_names[le16_to_cpu(ip->hw_id)]); - adev->reg_offset[hw_ip][ip->number_instance] = + adev->reg_offset[hw_ip][ip->instance_number] = ip->base_address; /* Instance support is somewhat inconsistent. * SDMA is a good example. Sienna cichlid has 4 total @@ -1183,80 +1530,59 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev) * example. On most chips there are multiple instances * with the same HWID. */ - adev->ip_versions[hw_ip][ip->number_instance] = - IP_VERSION(ip->major, ip->minor, ip->revision); + + if (ihdr->version < 3) { + subrev = 0; + variant = 0; + } else { + subrev = ip->sub_revision; + variant = ip->variant; + } + + adev->ip_versions[hw_ip] + [ip->instance_number] = + IP_VERSION_FULL(ip->major, + ip->minor, + ip->revision, + variant, + subrev); } } next_ip: - ip_offset += struct_size(ip, base_address, ip->num_base_address); + if (ihdr->base_addr_64_bit) + ip_offset += struct_size(ip, base_address_64, ip->num_base_address); + else + ip_offset += struct_size(ip, base_address, ip->num_base_address); } } - amdgpu_discovery_sysfs_init(adev); + if (wafl_ver && !adev->ip_versions[XGMI_HWIP][0]) + adev->ip_versions[XGMI_HWIP][0] = wafl_ver; return 0; } -int amdgpu_discovery_get_ip_version(struct amdgpu_device *adev, int hw_id, int number_instance, - int *major, int *minor, int *revision) -{ - struct binary_header *bhdr; - struct ip_discovery_header *ihdr; - struct die_header *dhdr; - struct ip *ip; - uint16_t die_offset; - uint16_t ip_offset; - uint16_t num_dies; - uint16_t num_ips; - int i, j; - - if (!adev->mman.discovery_bin) { - DRM_ERROR("ip discovery uninitialized\n"); - return -EINVAL; - } - - 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); - - 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 ((le16_to_cpu(ip->hw_id) == hw_id) && (ip->number_instance == number_instance)) { - if (major) - *major = ip->major; - if (minor) - *minor = ip->minor; - if (revision) - *revision = ip->revision; - return 0; - } - ip_offset += struct_size(ip, base_address, ip->num_base_address); - } - } - - return -EINVAL; -} - static void amdgpu_discovery_harvest_ip(struct amdgpu_device *adev) { + uint8_t *discovery_bin = adev->discovery.bin; + struct ip_discovery_header *ihdr; + struct binary_header *bhdr; int vcn_harvest_count = 0; int umc_harvest_count = 0; + uint16_t offset, ihdr_ver; + bhdr = (struct binary_header *)discovery_bin; + offset = le16_to_cpu(bhdr->table_list[IP_DISCOVERY].offset); + ihdr = (struct ip_discovery_header *)(discovery_bin + offset); + ihdr_ver = le16_to_cpu(ihdr->version); /* * 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 (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 2, 0) && + ihdr_ver <= 2) { if ((adev->pdev->device == 0x731E && (adev->pdev->revision == 0xC6 || adev->pdev->revision == 0xC7)) || @@ -1288,27 +1614,30 @@ 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_v1_3 v1_3; struct gc_info_v2_0 v2; + struct gc_info_v2_1 v2_1; }; static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) { + uint8_t *discovery_bin = adev->discovery.bin; struct binary_header *bhdr; union gc_info *gc_info; u16 offset; - if (!adev->mman.discovery_bin) { + if (!discovery_bin) { DRM_ERROR("ip discovery uninitialized\n"); return -EINVAL; } - bhdr = (struct binary_header *)adev->mman.discovery_bin; + bhdr = (struct binary_header *)discovery_bin; offset = le16_to_cpu(bhdr->table_list[GC].offset); if (!offset) return 0; - gc_info = (union gc_info *)(adev->mman.discovery_bin + offset); + gc_info = (union gc_info *)(discovery_bin + offset); switch (le16_to_cpu(gc_info->v1.header.version_major)) { case 1: @@ -1330,12 +1659,12 @@ static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) 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) { + if (le16_to_cpu(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) { + if (le16_to_cpu(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); @@ -1345,6 +1674,16 @@ static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) 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); } + if (le16_to_cpu(gc_info->v1.header.version_minor) >= 3) { + adev->gfx.config.gc_tcp_size_per_cu = le32_to_cpu(gc_info->v1_3.gc_tcp_size_per_cu); + adev->gfx.config.gc_tcp_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_tcp_cache_line_size); + adev->gfx.config.gc_instruction_cache_size_per_sqc = le32_to_cpu(gc_info->v1_3.gc_instruction_cache_size_per_sqc); + adev->gfx.config.gc_instruction_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_instruction_cache_line_size); + adev->gfx.config.gc_scalar_data_cache_size_per_sqc = le32_to_cpu(gc_info->v1_3.gc_scalar_data_cache_size_per_sqc); + adev->gfx.config.gc_scalar_data_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_scalar_data_cache_line_size); + adev->gfx.config.gc_tcc_size = le32_to_cpu(gc_info->v1_3.gc_tcc_size); + adev->gfx.config.gc_tcc_cache_line_size = le32_to_cpu(gc_info->v1_3.gc_tcc_cache_line_size); + } break; case 2: adev->gfx.config.max_shader_engines = le32_to_cpu(gc_info->v2.gc_num_se); @@ -1364,6 +1703,15 @@ static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) 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); + if (le16_to_cpu(gc_info->v2.header.version_minor) == 1) { + adev->gfx.config.gc_num_tcp_per_sa = le32_to_cpu(gc_info->v2_1.gc_num_tcp_per_sh); + adev->gfx.config.gc_tcp_size_per_cu = le32_to_cpu(gc_info->v2_1.gc_tcp_size_per_cu); + adev->gfx.config.gc_num_sdp_interface = le32_to_cpu(gc_info->v2_1.gc_num_sdp_interface); /* per XCD */ + adev->gfx.config.gc_num_cu_per_sqc = le32_to_cpu(gc_info->v2_1.gc_num_cu_per_sqc); + adev->gfx.config.gc_l1_instruction_cache_size_per_sqc = le32_to_cpu(gc_info->v2_1.gc_instruction_cache_size_per_sqc); + adev->gfx.config.gc_l1_data_cache_size_per_sqc = le32_to_cpu(gc_info->v2_1.gc_scalar_data_cache_size_per_sqc); + adev->gfx.config.gc_tcc_size = le32_to_cpu(gc_info->v2_1.gc_tcc_size); /* per XCD */ + } break; default: dev_err(adev->dev, @@ -1377,28 +1725,30 @@ static int amdgpu_discovery_get_gfx_info(struct amdgpu_device *adev) union mall_info { struct mall_info_v1_0 v1; + struct mall_info_v2_0 v2; }; static int amdgpu_discovery_get_mall_info(struct amdgpu_device *adev) { + uint8_t *discovery_bin = adev->discovery.bin; 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) { + if (!discovery_bin) { DRM_ERROR("ip discovery uninitialized\n"); return -EINVAL; } - bhdr = (struct binary_header *)adev->mman.discovery_bin; + bhdr = (struct binary_header *)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); + mall_info = (union mall_info *)(discovery_bin + offset); switch (le16_to_cpu(mall_info->v1.header.version_major)) { case 1: @@ -1415,6 +1765,11 @@ static int amdgpu_discovery_get_mall_info(struct amdgpu_device *adev) mall_size += mall_size_per_umc; } adev->gmc.mall_size = mall_size; + adev->gmc.m_half_use = half_use; + break; + case 2: + mall_size_per_umc = le32_to_cpu(mall_info->v2.mall_size_per_umc); + adev->gmc.mall_size = (uint64_t)mall_size_per_umc * adev->gmc.num_umc; break; default: dev_err(adev->dev, @@ -1432,12 +1787,13 @@ union vcn_info { static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev) { + uint8_t *discovery_bin = adev->discovery.bin; struct binary_header *bhdr; union vcn_info *vcn_info; u16 offset; int v; - if (!adev->mman.discovery_bin) { + if (!discovery_bin) { DRM_ERROR("ip discovery uninitialized\n"); return -EINVAL; } @@ -1452,13 +1808,13 @@ static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev) return -EINVAL; } - bhdr = (struct binary_header *)adev->mman.discovery_bin; + bhdr = (struct binary_header *)discovery_bin; offset = le16_to_cpu(bhdr->table_list[VCN_INFO].offset); if (!offset) return 0; - vcn_info = (union vcn_info *)(adev->mman.discovery_bin + offset); + vcn_info = (union vcn_info *)(discovery_bin + offset); switch (le16_to_cpu(vcn_info->v1.header.version_major)) { case 1: @@ -1466,7 +1822,7 @@ static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev) * so this won't overflow. */ for (v = 0; v < adev->vcn.num_vcn_inst; v++) { - adev->vcn.vcn_codec_disable_mask[v] = + adev->vcn.inst[v].vcn_codec_disable_mask = le32_to_cpu(vcn_info->v1.instance_info[v].fuse_data.all_bits); } break; @@ -1480,10 +1836,113 @@ static int amdgpu_discovery_get_vcn_info(struct amdgpu_device *adev) return 0; } +union nps_info { + struct nps_info_v1_0 v1; +}; + +static int amdgpu_discovery_refresh_nps_info(struct amdgpu_device *adev, + union nps_info *nps_data) +{ + uint64_t vram_size, pos, offset; + struct nps_info_header *nhdr; + struct binary_header bhdr; + uint16_t checksum; + + vram_size = (uint64_t)RREG32(mmRCC_CONFIG_MEMSIZE) << 20; + pos = vram_size - DISCOVERY_TMR_OFFSET; + amdgpu_device_vram_access(adev, pos, &bhdr, sizeof(bhdr), false); + + offset = le16_to_cpu(bhdr.table_list[NPS_INFO].offset); + checksum = le16_to_cpu(bhdr.table_list[NPS_INFO].checksum); + + amdgpu_device_vram_access(adev, (pos + offset), nps_data, + sizeof(*nps_data), false); + + nhdr = (struct nps_info_header *)(nps_data); + if (!amdgpu_discovery_verify_checksum((uint8_t *)nps_data, + le32_to_cpu(nhdr->size_bytes), + checksum)) { + dev_err(adev->dev, "nps data refresh, checksum mismatch\n"); + return -EINVAL; + } + + return 0; +} + +int amdgpu_discovery_get_nps_info(struct amdgpu_device *adev, + uint32_t *nps_type, + struct amdgpu_gmc_memrange **ranges, + int *range_cnt, bool refresh) +{ + uint8_t *discovery_bin = adev->discovery.bin; + struct amdgpu_gmc_memrange *mem_ranges; + struct binary_header *bhdr; + union nps_info *nps_info; + union nps_info nps_data; + u16 offset; + int i, r; + + if (!nps_type || !range_cnt || !ranges) + return -EINVAL; + + if (refresh) { + r = amdgpu_discovery_refresh_nps_info(adev, &nps_data); + if (r) + return r; + nps_info = &nps_data; + } else { + if (!discovery_bin) { + dev_err(adev->dev, + "fetch mem range failed, ip discovery uninitialized\n"); + return -EINVAL; + } + + bhdr = (struct binary_header *)discovery_bin; + offset = le16_to_cpu(bhdr->table_list[NPS_INFO].offset); + + if (!offset) + return -ENOENT; + + /* If verification fails, return as if NPS table doesn't exist */ + if (amdgpu_discovery_verify_npsinfo(adev, bhdr)) + return -ENOENT; + + nps_info = (union nps_info *)(discovery_bin + offset); + } + + switch (le16_to_cpu(nps_info->v1.header.version_major)) { + case 1: + mem_ranges = kvcalloc(nps_info->v1.count, + sizeof(*mem_ranges), + GFP_KERNEL); + if (!mem_ranges) + return -ENOMEM; + *nps_type = nps_info->v1.nps_type; + *range_cnt = nps_info->v1.count; + for (i = 0; i < *range_cnt; i++) { + mem_ranges[i].base_address = + nps_info->v1.instance_info[i].base_address; + mem_ranges[i].limit_address = + nps_info->v1.instance_info[i].limit_address; + mem_ranges[i].nid_mask = -1; + mem_ranges[i].flags = 0; + } + *ranges = mem_ranges; + break; + default: + dev_err(adev->dev, "Unhandled NPS info table %d.%d\n", + le16_to_cpu(nps_info->v1.header.version_major), + le16_to_cpu(nps_info->v1.header.version_minor)); + return -EINVAL; + } + + return 0; +} + static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) { /* what IP to use for this? */ - switch (adev->ip_versions[GC_HWIP][0]) { + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(9, 0, 1): case IP_VERSION(9, 1, 0): case IP_VERSION(9, 2, 1): @@ -1492,6 +1951,9 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(9, 4, 0): case IP_VERSION(9, 4, 1): case IP_VERSION(9, 4, 2): + case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): + case IP_VERSION(9, 5, 0): amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); break; case IP_VERSION(10, 1, 10): @@ -1514,12 +1976,20 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &soc21_common_ip_block); break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + amdgpu_device_ip_block_add(adev, &soc24_common_ip_block); + break; default: dev_err(adev->dev, "Failed to add common ip block(GC_HWIP:0x%x)\n", - adev->ip_versions[GC_HWIP][0]); + amdgpu_ip_version(adev, GC_HWIP, 0)); return -EINVAL; } return 0; @@ -1528,7 +1998,7 @@ static int amdgpu_discovery_set_common_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) { /* use GC or MMHUB IP version */ - switch (adev->ip_versions[GC_HWIP][0]) { + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(9, 0, 1): case IP_VERSION(9, 1, 0): case IP_VERSION(9, 2, 1): @@ -1537,6 +2007,9 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(9, 4, 0): case IP_VERSION(9, 4, 1): case IP_VERSION(9, 4, 2): + case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): + case IP_VERSION(9, 5, 0): amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); break; case IP_VERSION(10, 1, 10): @@ -1559,12 +2032,19 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &gmc_v11_0_ip_block); break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + amdgpu_device_ip_block_add(adev, &gmc_v12_0_ip_block); + break; default: - dev_err(adev->dev, - "Failed to add gmc ip block(GC_HWIP:0x%x)\n", - adev->ip_versions[GC_HWIP][0]); + dev_err(adev->dev, "Failed to add gmc ip block(GC_HWIP:0x%x)\n", + amdgpu_ip_version(adev, GC_HWIP, 0)); return -EINVAL; } return 0; @@ -1572,7 +2052,7 @@ static int amdgpu_discovery_set_gmc_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev) { - switch (adev->ip_versions[OSSSYS_HWIP][0]) { + switch (amdgpu_ip_version(adev, OSSSYS_HWIP, 0)) { case IP_VERSION(4, 0, 0): case IP_VERSION(4, 0, 1): case IP_VERSION(4, 1, 0): @@ -1583,6 +2063,8 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(4, 2, 0): case IP_VERSION(4, 2, 1): case IP_VERSION(4, 4, 0): + case IP_VERSION(4, 4, 2): + case IP_VERSION(4, 4, 5): amdgpu_device_ip_block_add(adev, &vega20_ih_ip_block); break; case IP_VERSION(5, 0, 0): @@ -1598,10 +2080,16 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(6, 0, 2): amdgpu_device_ip_block_add(adev, &ih_v6_0_ip_block); break; + case IP_VERSION(6, 1, 0): + amdgpu_device_ip_block_add(adev, &ih_v6_1_ip_block); + break; + case IP_VERSION(7, 0, 0): + amdgpu_device_ip_block_add(adev, &ih_v7_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add ih ip block(OSSSYS_HWIP:0x%x)\n", - adev->ip_versions[OSSSYS_HWIP][0]); + amdgpu_ip_version(adev, OSSSYS_HWIP, 0)); return -EINVAL; } return 0; @@ -1609,7 +2097,7 @@ static int amdgpu_discovery_set_ih_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) { - switch (adev->ip_versions[MP0_HWIP][0]) { + switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) { case IP_VERSION(9, 0, 0): amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); break; @@ -1627,6 +2115,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 12): case IP_VERSION(11, 0, 13): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); break; case IP_VERSION(11, 0, 8): @@ -1641,19 +2130,30 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(13, 0, 2): case IP_VERSION(13, 0, 3): case IP_VERSION(13, 0, 5): + case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 7): case IP_VERSION(13, 0, 8): case IP_VERSION(13, 0, 10): case IP_VERSION(13, 0, 11): + case IP_VERSION(13, 0, 12): + case IP_VERSION(13, 0, 14): + case IP_VERSION(14, 0, 0): + case IP_VERSION(14, 0, 1): + case IP_VERSION(14, 0, 4): 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; + case IP_VERSION(14, 0, 2): + case IP_VERSION(14, 0, 3): + case IP_VERSION(14, 0, 5): + amdgpu_device_ip_block_add(adev, &psp_v14_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add psp ip block(MP0_HWIP:0x%x)\n", - adev->ip_versions[MP0_HWIP][0]); + amdgpu_ip_version(adev, MP0_HWIP, 0)); return -EINVAL; } return 0; @@ -1661,7 +2161,7 @@ static int amdgpu_discovery_set_psp_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) { - switch (adev->ip_versions[MP1_HWIP][0]) { + switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) { case IP_VERSION(9, 0, 0): case IP_VERSION(10, 0, 0): case IP_VERSION(10, 0, 1): @@ -1675,13 +2175,17 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 5): case IP_VERSION(11, 0, 9): case IP_VERSION(11, 0, 7): - case IP_VERSION(11, 0, 8): case IP_VERSION(11, 0, 11): case IP_VERSION(11, 0, 12): case IP_VERSION(11, 0, 13): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); break; + case IP_VERSION(11, 0, 8): + if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) + amdgpu_device_ip_block_add(adev, &smu_v11_0_ip_block); + break; case IP_VERSION(12, 0, 0): case IP_VERSION(12, 0, 1): amdgpu_device_ip_block_add(adev, &smu_v12_0_ip_block); @@ -1692,16 +2196,27 @@ static int amdgpu_discovery_set_smu_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(13, 0, 3): case IP_VERSION(13, 0, 4): case IP_VERSION(13, 0, 5): + case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 7): case IP_VERSION(13, 0, 8): case IP_VERSION(13, 0, 10): case IP_VERSION(13, 0, 11): + case IP_VERSION(13, 0, 14): + case IP_VERSION(13, 0, 12): amdgpu_device_ip_block_add(adev, &smu_v13_0_ip_block); break; + case IP_VERSION(14, 0, 0): + case IP_VERSION(14, 0, 1): + case IP_VERSION(14, 0, 2): + case IP_VERSION(14, 0, 3): + case IP_VERSION(14, 0, 4): + case IP_VERSION(14, 0, 5): + amdgpu_device_ip_block_add(adev, &smu_v14_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add smu ip block(MP1_HWIP:0x%x)\n", - adev->ip_versions[MP1_HWIP][0]); + amdgpu_ip_version(adev, MP1_HWIP, 0)); return -EINVAL; } return 0; @@ -1726,8 +2241,8 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) return 0; #if defined(CONFIG_DRM_AMD_DC) - if (adev->ip_versions[DCE_HWIP][0]) { - switch (adev->ip_versions[DCE_HWIP][0]) { + if (amdgpu_ip_version(adev, DCE_HWIP, 0)) { + switch (amdgpu_ip_version(adev, DCE_HWIP, 0)) { case IP_VERSION(1, 0, 0): case IP_VERSION(1, 0, 1): case IP_VERSION(2, 0, 2): @@ -1745,6 +2260,14 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(3, 1, 6): case IP_VERSION(3, 2, 0): case IP_VERSION(3, 2, 1): + case IP_VERSION(3, 5, 0): + case IP_VERSION(3, 5, 1): + case IP_VERSION(3, 6, 0): + case IP_VERSION(4, 1, 0): + /* TODO: Fix IP version. DC code expects version 4.0.1 */ + if (adev->ip_versions[DCE_HWIP][0] == IP_VERSION(4, 1, 0)) + adev->ip_versions[DCE_HWIP][0] = IP_VERSION(4, 0, 1); + if (amdgpu_sriov_vf(adev)) amdgpu_discovery_set_sriov_display(adev); else @@ -1753,11 +2276,11 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) default: dev_err(adev->dev, "Failed to add dm ip block(DCE_HWIP:0x%x)\n", - adev->ip_versions[DCE_HWIP][0]); + amdgpu_ip_version(adev, DCE_HWIP, 0)); return -EINVAL; } - } else if (adev->ip_versions[DCI_HWIP][0]) { - switch (adev->ip_versions[DCI_HWIP][0]) { + } else if (amdgpu_ip_version(adev, DCI_HWIP, 0)) { + switch (amdgpu_ip_version(adev, DCI_HWIP, 0)) { case IP_VERSION(12, 0, 0): case IP_VERSION(12, 0, 1): case IP_VERSION(12, 1, 0): @@ -1769,7 +2292,7 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) default: dev_err(adev->dev, "Failed to add dm ip block(DCI_HWIP:0x%x)\n", - adev->ip_versions[DCI_HWIP][0]); + amdgpu_ip_version(adev, DCI_HWIP, 0)); return -EINVAL; } } @@ -1779,7 +2302,7 @@ static int amdgpu_discovery_set_display_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) { - switch (adev->ip_versions[GC_HWIP][0]) { + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(9, 0, 1): case IP_VERSION(9, 1, 0): case IP_VERSION(9, 2, 1): @@ -1790,6 +2313,11 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(9, 4, 2): amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); break; + case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): + case IP_VERSION(9, 5, 0): + amdgpu_device_ip_block_add(adev, &gfx_v9_4_3_ip_block); + break; case IP_VERSION(10, 1, 10): case IP_VERSION(10, 1, 2): case IP_VERSION(10, 1, 1): @@ -1810,12 +2338,19 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &gfx_v11_0_ip_block); break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + amdgpu_device_ip_block_add(adev, &gfx_v12_0_ip_block); + break; default: - dev_err(adev->dev, - "Failed to add gfx ip block(GC_HWIP:0x%x)\n", - adev->ip_versions[GC_HWIP][0]); + dev_err(adev->dev, "Failed to add gfx ip block(GC_HWIP:0x%x)\n", + amdgpu_ip_version(adev, GC_HWIP, 0)); return -EINVAL; } return 0; @@ -1823,7 +2358,7 @@ static int amdgpu_discovery_set_gc_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev) { - switch (adev->ip_versions[SDMA0_HWIP][0]) { + switch (amdgpu_ip_version(adev, SDMA0_HWIP, 0)) { case IP_VERSION(4, 0, 0): case IP_VERSION(4, 0, 1): case IP_VERSION(4, 1, 0): @@ -1834,6 +2369,11 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(4, 4, 0): amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); break; + case IP_VERSION(4, 4, 2): + case IP_VERSION(4, 4, 5): + case IP_VERSION(4, 4, 4): + amdgpu_device_ip_block_add(adev, &sdma_v4_4_2_ip_block); + break; case IP_VERSION(5, 0, 0): case IP_VERSION(5, 0, 1): case IP_VERSION(5, 0, 2): @@ -1854,21 +2394,44 @@ static int amdgpu_discovery_set_sdma_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(6, 0, 1): case IP_VERSION(6, 0, 2): case IP_VERSION(6, 0, 3): + case IP_VERSION(6, 1, 0): + case IP_VERSION(6, 1, 1): + case IP_VERSION(6, 1, 2): + case IP_VERSION(6, 1, 3): amdgpu_device_ip_block_add(adev, &sdma_v6_0_ip_block); break; + case IP_VERSION(7, 0, 0): + case IP_VERSION(7, 0, 1): + amdgpu_device_ip_block_add(adev, &sdma_v7_0_ip_block); + break; default: dev_err(adev->dev, "Failed to add sdma ip block(SDMA0_HWIP:0x%x)\n", - adev->ip_versions[SDMA0_HWIP][0]); + amdgpu_ip_version(adev, SDMA0_HWIP, 0)); return -EINVAL; } + + return 0; +} + +static int amdgpu_discovery_set_ras_ip_blocks(struct amdgpu_device *adev) +{ + switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) { + case IP_VERSION(13, 0, 6): + case IP_VERSION(13, 0, 12): + case IP_VERSION(13, 0, 14): + amdgpu_device_ip_block_add(adev, &ras_v1_0_ip_block); + break; + default: + break; + } return 0; } static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) { - if (adev->ip_versions[VCE_HWIP][0]) { - switch (adev->ip_versions[UVD_HWIP][0]) { + if (amdgpu_ip_version(adev, VCE_HWIP, 0)) { + switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) { case IP_VERSION(7, 0, 0): case IP_VERSION(7, 2, 0): /* UVD is not supported on vega20 SR-IOV */ @@ -1878,10 +2441,10 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) default: dev_err(adev->dev, "Failed to add uvd v7 ip block(UVD_HWIP:0x%x)\n", - adev->ip_versions[UVD_HWIP][0]); + amdgpu_ip_version(adev, UVD_HWIP, 0)); return -EINVAL; } - switch (adev->ip_versions[VCE_HWIP][0]) { + switch (amdgpu_ip_version(adev, VCE_HWIP, 0)) { case IP_VERSION(4, 0, 0): case IP_VERSION(4, 1, 0): /* VCE is not supported on vega20 SR-IOV */ @@ -1891,11 +2454,11 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) default: dev_err(adev->dev, "Failed to add VCE v4 ip block(VCE_HWIP:0x%x)\n", - adev->ip_versions[VCE_HWIP][0]); + amdgpu_ip_version(adev, VCE_HWIP, 0)); return -EINVAL; } } else { - switch (adev->ip_versions[UVD_HWIP][0]) { + switch (amdgpu_ip_version(adev, UVD_HWIP, 0)) { case IP_VERSION(1, 0, 0): case IP_VERSION(1, 0, 1): amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block); @@ -1922,7 +2485,6 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) 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); if (!amdgpu_sriov_vf(adev)) amdgpu_device_ip_block_add(adev, &jpeg_v3_0_ip_block); @@ -1934,13 +2496,29 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) 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); + amdgpu_device_ip_block_add(adev, &jpeg_v4_0_ip_block); + break; + case IP_VERSION(4, 0, 3): + amdgpu_device_ip_block_add(adev, &vcn_v4_0_3_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v4_0_3_ip_block); + break; + case IP_VERSION(4, 0, 5): + case IP_VERSION(4, 0, 6): + amdgpu_device_ip_block_add(adev, &vcn_v4_0_5_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v4_0_5_ip_block); + break; + case IP_VERSION(5, 0, 0): + amdgpu_device_ip_block_add(adev, &vcn_v5_0_0_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v5_0_0_ip_block); + break; + case IP_VERSION(5, 0, 1): + amdgpu_device_ip_block_add(adev, &vcn_v5_0_1_ip_block); + amdgpu_device_ip_block_add(adev, &jpeg_v5_0_1_ip_block); break; default: dev_err(adev->dev, "Failed to add vcn/jpeg ip block(UVD_HWIP:0x%x)\n", - adev->ip_versions[UVD_HWIP][0]); + amdgpu_ip_version(adev, UVD_HWIP, 0)); return -EINVAL; } } @@ -1949,38 +2527,94 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev) static int amdgpu_discovery_set_mes_ip_blocks(struct amdgpu_device *adev) { - switch (adev->ip_versions[GC_HWIP][0]) { - case IP_VERSION(10, 1, 10): - 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): - 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; + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(11, 0, 0): case IP_VERSION(11, 0, 1): case IP_VERSION(11, 0, 2): case IP_VERSION(11, 0, 3): case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): amdgpu_device_ip_block_add(adev, &mes_v11_0_ip_block); adev->enable_mes = true; adev->enable_mes_kiq = true; break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + amdgpu_device_ip_block_add(adev, &mes_v12_0_ip_block); + adev->enable_mes = true; + adev->enable_mes_kiq = true; + if (amdgpu_uni_mes) + adev->enable_uni_mes = true; + break; + default: + break; + } + return 0; +} + +static void amdgpu_discovery_init_soc_config(struct amdgpu_device *adev) +{ + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { + case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): + case IP_VERSION(9, 5, 0): + aqua_vanjaram_init_soc_config(adev); + break; + default: + break; + } +} + +static int amdgpu_discovery_set_vpe_ip_blocks(struct amdgpu_device *adev) +{ + switch (amdgpu_ip_version(adev, VPE_HWIP, 0)) { + case IP_VERSION(6, 1, 0): + case IP_VERSION(6, 1, 1): + case IP_VERSION(6, 1, 3): + amdgpu_device_ip_block_add(adev, &vpe_v6_1_ip_block); + break; + default: + break; + } + + return 0; +} + +static int amdgpu_discovery_set_umsch_mm_ip_blocks(struct amdgpu_device *adev) +{ + switch (amdgpu_ip_version(adev, VCN_HWIP, 0)) { + case IP_VERSION(4, 0, 5): + case IP_VERSION(4, 0, 6): + if (amdgpu_umsch_mm & 0x1) { + amdgpu_device_ip_block_add(adev, &umsch_mm_v4_0_ip_block); + adev->enable_umsch_mm = true; + } + break; + default: + break; + } + + return 0; +} + +static int amdgpu_discovery_set_isp_ip_blocks(struct amdgpu_device *adev) +{ +#if defined(CONFIG_DRM_AMD_ISP) + switch (amdgpu_ip_version(adev, ISP_HWIP, 0)) { + case IP_VERSION(4, 1, 0): + amdgpu_device_ip_block_add(adev, &isp_v4_1_0_ip_block); + break; + case IP_VERSION(4, 1, 1): + amdgpu_device_ip_block_add(adev, &isp_v4_1_1_ip_block); + break; default: break; } +#endif + return 0; } @@ -1990,9 +2624,16 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) switch (adev->asic_type) { case CHIP_VEGA10: + /* This is not fatal. We only need the discovery + * binary for sysfs. We don't need it for a + * functional system. + */ + amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; 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); @@ -2012,9 +2653,16 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 0, 0); break; case CHIP_VEGA12: + /* This is not fatal. We only need the discovery + * binary for sysfs. We don't need it for a + * functional system. + */ + amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; 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); @@ -2034,10 +2682,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 0, 1); break; case CHIP_RAVEN: + /* This is not fatal. We only need the discovery + * binary for sysfs. We don't need it for a + * functional system. + */ + amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 1; + adev->sdma.sdma_mask = 1; adev->vcn.num_vcn_inst = 1; adev->gmc.num_umc = 2; + adev->gfx.xcc_mask = 1; 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); @@ -2054,6 +2709,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 2, 2); adev->ip_versions[UVD_HWIP][0] = IP_VERSION(1, 0, 1); adev->ip_versions[DCE_HWIP][0] = IP_VERSION(1, 0, 1); + adev->ip_versions[ISP_HWIP][0] = IP_VERSION(2, 0, 0); } else { adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 1, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 1, 0); @@ -2070,12 +2726,20 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[GC_HWIP][0] = IP_VERSION(9, 1, 0); adev->ip_versions[UVD_HWIP][0] = IP_VERSION(1, 0, 0); adev->ip_versions[DCE_HWIP][0] = IP_VERSION(1, 0, 0); + adev->ip_versions[ISP_HWIP][0] = IP_VERSION(2, 0, 0); } break; case CHIP_VEGA20: + /* This is not fatal. We only need the discovery + * binary for sysfs. We don't need it for a + * functional system. + */ + amdgpu_discovery_init(adev); vega20_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 8; + adev->gfx.xcc_mask = 1; 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); @@ -2096,10 +2760,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[DCI_HWIP][0] = IP_VERSION(12, 1, 0); break; case CHIP_ARCTURUS: + /* This is not fatal. We only need the discovery + * binary for sysfs. We don't need it for a + * functional system. + */ + amdgpu_discovery_init(adev); arct_reg_base_init(adev); adev->sdma.num_instances = 8; + adev->sdma.sdma_mask = 0xff; adev->vcn.num_vcn_inst = 2; adev->gmc.num_umc = 8; + adev->gfx.xcc_mask = 1; 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); @@ -2124,10 +2795,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[UVD_HWIP][1] = IP_VERSION(2, 5, 0); break; case CHIP_ALDEBARAN: + /* This is not fatal. We only need the discovery + * binary for sysfs. We don't need it for a + * functional system. + */ + amdgpu_discovery_init(adev); aldebaran_reg_base_init(adev); adev->sdma.num_instances = 5; + adev->sdma.sdma_mask = 0x1f; adev->vcn.num_vcn_inst = 2; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; 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); @@ -2149,10 +2827,44 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->ip_versions[UVD_HWIP][1] = IP_VERSION(2, 6, 0); adev->ip_versions[XGMI_HWIP][0] = IP_VERSION(6, 1, 0); break; + case CHIP_CYAN_SKILLFISH: + if (adev->apu_flags & AMD_APU_IS_CYAN_SKILLFISH2) { + r = amdgpu_discovery_reg_base_init(adev); + if (r) + return -EINVAL; + + amdgpu_discovery_harvest_ip(adev); + amdgpu_discovery_get_gfx_info(adev); + amdgpu_discovery_get_mall_info(adev); + amdgpu_discovery_get_vcn_info(adev); + } else { + cyan_skillfish_reg_base_init(adev); + adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; + adev->gfx.xcc_mask = 1; + adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(2, 0, 3); + adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(2, 0, 3); + adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(5, 0, 1); + adev->ip_versions[HDP_HWIP][0] = IP_VERSION(5, 0, 1); + adev->ip_versions[SDMA0_HWIP][0] = IP_VERSION(5, 0, 1); + adev->ip_versions[SDMA1_HWIP][1] = IP_VERSION(5, 0, 1); + adev->ip_versions[DF_HWIP][0] = IP_VERSION(3, 5, 0); + adev->ip_versions[NBIO_HWIP][0] = IP_VERSION(2, 1, 1); + adev->ip_versions[UMC_HWIP][0] = IP_VERSION(8, 1, 1); + adev->ip_versions[MP0_HWIP][0] = IP_VERSION(11, 0, 8); + adev->ip_versions[MP1_HWIP][0] = IP_VERSION(11, 0, 8); + adev->ip_versions[THM_HWIP][0] = IP_VERSION(11, 0, 1); + adev->ip_versions[SMUIO_HWIP][0] = IP_VERSION(11, 0, 8); + adev->ip_versions[GC_HWIP][0] = IP_VERSION(10, 1, 3); + adev->ip_versions[UVD_HWIP][0] = IP_VERSION(2, 0, 3); + } + break; default: r = amdgpu_discovery_reg_base_init(adev); - if (r) - return -EINVAL; + if (r) { + drm_err(&adev->ddev, "discovery failed: %d\n", r); + return r; + } amdgpu_discovery_harvest_ip(adev); amdgpu_discovery_get_gfx_info(adev); @@ -2161,12 +2873,18 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) break; } - switch (adev->ip_versions[GC_HWIP][0]) { + amdgpu_discovery_init_soc_config(adev); + amdgpu_discovery_sysfs_init(adev); + + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(9, 0, 1): case IP_VERSION(9, 2, 1): case IP_VERSION(9, 4, 0): case IP_VERSION(9, 4, 1): case IP_VERSION(9, 4, 2): + case IP_VERSION(9, 4, 3): + case IP_VERSION(9, 4, 4): + case IP_VERSION(9, 5, 0): adev->family = AMDGPU_FAMILY_AI; break; case IP_VERSION(9, 1, 0): @@ -2207,11 +2925,21 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 4): adev->family = AMDGPU_FAMILY_GC_11_0_1; break; + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): + adev->family = AMDGPU_FAMILY_GC_11_5_0; + break; + case IP_VERSION(12, 0, 0): + case IP_VERSION(12, 0, 1): + adev->family = AMDGPU_FAMILY_GC_12_0_0; + break; default: return -EINVAL; } - switch (adev->ip_versions[GC_HWIP][0]) { + switch (amdgpu_ip_version(adev, GC_HWIP, 0)) { case IP_VERSION(9, 1, 0): case IP_VERSION(9, 2, 2): case IP_VERSION(9, 3, 0): @@ -2223,17 +2951,18 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(10, 3, 7): case IP_VERSION(11, 0, 1): case IP_VERSION(11, 0, 4): + case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 1): + case IP_VERSION(11, 5, 2): + case IP_VERSION(11, 5, 3): 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; - /* set NBIO version */ - switch (adev->ip_versions[NBIO_HWIP][0]) { + switch (amdgpu_ip_version(adev, NBIO_HWIP, 0)) { case IP_VERSION(6, 1, 0): case IP_VERSION(6, 2, 0): adev->nbio.funcs = &nbio_v6_1_funcs; @@ -2251,6 +2980,18 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->nbio.funcs = &nbio_v7_4_funcs; adev->nbio.hdp_flush_reg = &nbio_v7_4_hdp_flush_reg; break; + case IP_VERSION(7, 9, 0): + case IP_VERSION(7, 9, 1): + adev->nbio.funcs = &nbio_v7_9_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_9_hdp_flush_reg; + break; + case IP_VERSION(7, 11, 0): + case IP_VERSION(7, 11, 1): + case IP_VERSION(7, 11, 2): + case IP_VERSION(7, 11, 3): + adev->nbio.funcs = &nbio_v7_11_funcs; + adev->nbio.hdp_flush_reg = &nbio_v7_11_hdp_flush_reg; + break; case IP_VERSION(7, 2, 0): case IP_VERSION(7, 2, 1): case IP_VERSION(7, 3, 0): @@ -2283,11 +3024,15 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) adev->nbio.funcs = &nbio_v7_7_funcs; adev->nbio.hdp_flush_reg = &nbio_v7_7_hdp_flush_reg; break; + case IP_VERSION(6, 3, 1): + adev->nbio.funcs = &nbif_v6_3_1_funcs; + adev->nbio.hdp_flush_reg = &nbif_v6_3_1_hdp_flush_reg; + break; default: break; } - switch (adev->ip_versions[HDP_HWIP][0]) { + switch (amdgpu_ip_version(adev, HDP_HWIP, 0)) { case IP_VERSION(4, 0, 0): case IP_VERSION(4, 0, 1): case IP_VERSION(4, 1, 0): @@ -2296,6 +3041,8 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(4, 2, 0): case IP_VERSION(4, 2, 1): case IP_VERSION(4, 4, 0): + case IP_VERSION(4, 4, 2): + case IP_VERSION(4, 4, 5): adev->hdp.funcs = &hdp_v4_0_funcs; break; case IP_VERSION(5, 0, 0): @@ -2311,13 +3058,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) break; case IP_VERSION(6, 0, 0): case IP_VERSION(6, 0, 1): + case IP_VERSION(6, 1, 0): adev->hdp.funcs = &hdp_v6_0_funcs; break; + case IP_VERSION(7, 0, 0): + adev->hdp.funcs = &hdp_v7_0_funcs; + break; default: break; } - switch (adev->ip_versions[DF_HWIP][0]) { + switch (amdgpu_ip_version(adev, DF_HWIP, 0)) { case IP_VERSION(3, 6, 0): case IP_VERSION(3, 6, 1): case IP_VERSION(3, 6, 2): @@ -2333,11 +3084,18 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(4, 3, 0): adev->df.funcs = &df_v4_3_funcs; break; + case IP_VERSION(4, 6, 2): + adev->df.funcs = &df_v4_6_2_funcs; + break; + case IP_VERSION(4, 15, 0): + case IP_VERSION(4, 15, 1): + adev->df.funcs = &df_v4_15_funcs; + break; default: break; } - switch (adev->ip_versions[SMUIO_HWIP][0]) { + switch (amdgpu_ip_version(adev, SMUIO_HWIP, 0)) { case IP_VERSION(9, 0, 0): case IP_VERSION(9, 0, 1): case IP_VERSION(10, 0, 0): @@ -2357,6 +3115,7 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(11, 0, 10): case IP_VERSION(11, 0, 11): case IP_VERSION(11, 5, 0): + case IP_VERSION(11, 5, 2): case IP_VERSION(13, 0, 1): case IP_VERSION(13, 0, 9): case IP_VERSION(13, 0, 10): @@ -2365,21 +3124,37 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) case IP_VERSION(13, 0, 2): adev->smuio.funcs = &smuio_v13_0_funcs; break; + case IP_VERSION(13, 0, 3): + case IP_VERSION(13, 0, 11): + adev->smuio.funcs = &smuio_v13_0_3_funcs; + if (adev->smuio.funcs->get_pkg_type(adev) == AMDGPU_PKG_TYPE_APU) { + adev->flags |= AMD_IS_APU; + } + break; case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 8): + case IP_VERSION(14, 0, 0): + case IP_VERSION(14, 0, 1): adev->smuio.funcs = &smuio_v13_0_6_funcs; break; + case IP_VERSION(14, 0, 2): + adev->smuio.funcs = &smuio_v14_0_2_funcs; + break; default: break; } - switch (adev->ip_versions[LSDMA_HWIP][0]) { + switch (amdgpu_ip_version(adev, 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; + case IP_VERSION(7, 0, 0): + case IP_VERSION(7, 0, 1): + adev->lsdma.funcs = &lsdma_v7_0_funcs; + break; default: break; } @@ -2430,6 +3205,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) if (r) return r; + r = amdgpu_discovery_set_ras_ip_blocks(adev); + if (r) + return r; + 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)) { @@ -2446,6 +3225,17 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) if (r) return r; + r = amdgpu_discovery_set_vpe_ip_blocks(adev); + if (r) + return r; + + r = amdgpu_discovery_set_umsch_mm_ip_blocks(adev); + if (r) + return r; + + r = amdgpu_discovery_set_isp_ip_blocks(adev); + if (r) + return r; return 0; } |