From d37a3929ca0363ed1dce02b2772cd5bc547ca66d Mon Sep 17 00:00:00 2001 From: Orlando Chamberlain Date: Fri, 3 Mar 2023 22:34:25 +1100 Subject: drm/amdgpu: register a vga_switcheroo client for MacBooks with apple-gmux Commit 3840c5bcc245 ("drm/amdgpu: disentangle runtime pm and vga_switcheroo") made amdgpu only register a vga_switcheroo client for GPU's with PX, however AMD GPUs in dual gpu Apple Macbooks do need to register, but don't have PX. Instead of AMD's PX, they use apple-gmux. Use apple_gmux_detect() to identify these gpus, and pci_is_thunderbolt_attached() to ensure eGPUs connected to Dual GPU Macbooks don't register with vga_switcheroo. Fixes: 3840c5bcc245 ("drm/amdgpu: disentangle runtime pm and vga_switcheroo") Link: https://lore.kernel.org/amd-gfx/20230210044826.9834-10-orlandoch.dev@gmail.com/ Signed-off-by: Orlando Chamberlain Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index c4a4e2fe66814..9bca7e5547c5d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -35,6 +35,7 @@ #include #include #include +#include #include #include @@ -3930,12 +3931,15 @@ fence_driver_init: if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA) vga_client_register(adev->pdev, amdgpu_device_vga_set_decode); - if (amdgpu_device_supports_px(ddev)) { - px = true; + px = amdgpu_device_supports_px(ddev); + + if (px || (!pci_is_thunderbolt_attached(adev->pdev) && + apple_gmux_detect(NULL, NULL))) vga_switcheroo_register_client(adev->pdev, &amdgpu_switcheroo_ops, px); + + if (px) vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain); - } if (adev->gmc.xgmi.pending_reset) queue_delayed_work(system_wq, &mgpu_info.delayed_reset_work, @@ -4039,6 +4043,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) void amdgpu_device_fini_sw(struct amdgpu_device *adev) { int idx; + bool px; amdgpu_fence_driver_sw_fini(adev); amdgpu_device_ip_fini(adev); @@ -4057,10 +4062,16 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev) kfree(adev->bios); adev->bios = NULL; - if (amdgpu_device_supports_px(adev_to_drm(adev))) { + + px = amdgpu_device_supports_px(adev_to_drm(adev)); + + if (px || (!pci_is_thunderbolt_attached(adev->pdev) && + apple_gmux_detect(NULL, NULL))) vga_switcheroo_unregister_client(adev->pdev); + + if (px) vga_switcheroo_fini_domain_pm_ops(adev->dev); - } + if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA) vga_client_unregister(adev->pdev); -- cgit v1.2.3 From 58265640fbd9a57bca521c3d83012fff2cd15fc6 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 7 Mar 2023 14:22:21 -0600 Subject: drm/amdgpu: Drop redundant pci_enable_pcie_error_reporting() pci_enable_pcie_error_reporting() enables the device to send ERR_* Messages. Since f26e58bf6f54 ("PCI/AER: Enable error reporting when AER is native"), the PCI core does this for all devices during enumeration, so the driver doesn't need to do it itself. Remove the redundant pci_enable_pcie_error_reporting() call from the driver. Note that this only controls ERR_* Messages from the device. An ERR_* Message may cause the Root Port to generate an interrupt, depending on the AER Root Error Command register managed by the AER service driver. Signed-off-by: Bjorn Helgaas Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 -- 2 files changed, 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 164141bc8b4ad..208cebb402321 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -50,7 +50,6 @@ #include #include #include -#include #include #include diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 9bca7e5547c5d..63570c59d6bc3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3774,8 +3774,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, } } - pci_enable_pcie_error_reporting(adev->pdev); - /* Post card if necessary */ if (amdgpu_device_need_post(adev)) { if (!adev->bios) { -- cgit v1.2.3 From 53e9d836ea7fa0bfe9950ffb92d50811a5e69f01 Mon Sep 17 00:00:00 2001 From: Guchun Chen Date: Mon, 13 Mar 2023 08:49:45 +0800 Subject: drm/amdgpu: drop pm_sysfs_en flag from amdgpu_device structure MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit pm_sysfs_en is overlapped with pm.sysfs_initialized, so drop it for simplifying code(no functional change). Signed-off-by: Guchun Chen Acked-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 9 +++------ 2 files changed, 3 insertions(+), 7 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 208cebb402321..21c5f4ae5af99 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1022,7 +1022,6 @@ struct amdgpu_device { bool in_runpm; bool has_pr3; - bool pm_sysfs_en; bool ucode_sysfs_en; bool psp_sysfs_en; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 63570c59d6bc3..a6c91794e0d6f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3863,11 +3863,8 @@ fence_driver_init: adev->mm_stats.log2_max_MBps = ilog2(max(1u, max_MBps)); r = amdgpu_pm_sysfs_init(adev); - if (r) { - adev->pm_sysfs_en = false; - DRM_ERROR("registering pm debugfs failed (%d).\n", r); - } else - adev->pm_sysfs_en = true; + if (r) + DRM_ERROR("registering pm sysfs failed (%d).\n", r); r = amdgpu_ucode_sysfs_init(adev); if (r) { @@ -4013,7 +4010,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev) if (adev->mman.initialized) drain_workqueue(adev->mman.bdev.wq); - if (adev->pm_sysfs_en) + if (adev->pm.sysfs_initialized) amdgpu_pm_sysfs_fini(adev); if (adev->ucode_sysfs_en) amdgpu_ucode_sysfs_fini(adev); -- cgit v1.2.3 From 6ab68650a10e89dc351cb12b42a89b8789126c13 Mon Sep 17 00:00:00 2001 From: Guchun Chen Date: Fri, 10 Mar 2023 12:59:32 +0800 Subject: drm/amdgpu: use drm_device pointer directly rather than convert again MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The convert from adev is redundant. Signed-off-by: Guchun Chen Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index a6c91794e0d6f..8fdd372f24378 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5588,7 +5588,7 @@ int amdgpu_device_baco_enter(struct drm_device *dev) struct amdgpu_device *adev = drm_to_adev(dev); struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); - if (!amdgpu_device_supports_baco(adev_to_drm(adev))) + if (!amdgpu_device_supports_baco(dev)) return -ENOTSUPP; if (ras && adev->ras_enabled && @@ -5604,7 +5604,7 @@ int amdgpu_device_baco_exit(struct drm_device *dev) struct amdgpu_ras *ras = amdgpu_ras_get_context(adev); int ret = 0; - if (!amdgpu_device_supports_baco(adev_to_drm(adev))) + if (!amdgpu_device_supports_baco(dev)) return -ENOTSUPP; ret = amdgpu_dpm_baco_exit(adev); -- cgit v1.2.3 From c69d51395a3bd3905a3837691ed48c7c89ea3d95 Mon Sep 17 00:00:00 2001 From: Guchun Chen Date: Thu, 9 Mar 2023 10:02:45 +0800 Subject: drm/amdgpu: move poll enabled/disable into non DC path Some amd asics having reliable hotplug support don't call drm_kms_helper_poll_init in driver init sequence. However, due to the unified suspend/resume path for all asics, because the output_poll_work->func is not set for these asics, a warning arrives when suspending. [ 90.656049] [ 90.656050] ? console_unlock+0x4d/0x100 [ 90.656053] ? __irq_work_queue_local+0x27/0x60 [ 90.656056] ? irq_work_queue+0x2b/0x50 [ 90.656057] ? __wake_up_klogd+0x40/0x60 [ 90.656059] __cancel_work_timer+0xed/0x180 [ 90.656061] drm_kms_helper_poll_disable.cold+0x1f/0x2c [drm_kms_helper] [ 90.656072] amdgpu_device_suspend+0x81/0x170 [amdgpu] [ 90.656180] amdgpu_pmops_runtime_suspend+0xb5/0x1b0 [amdgpu] [ 90.656269] pci_pm_runtime_suspend+0x61/0x1b0 drm_kms_helper_poll_enable/disable is valid when poll_init is called in amdgpu code, which is only used in non DC path. So move such codes into non-DC path code to get rid of such warnings. v1: introduce use_kms_poll flag in amdgpu as the poll stuff check v2: use dc_enabled as the flag to simply code v3: move code into non DC path instead of relying on any flag Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2411 Fixes: a4e771729a51 ("drm/probe_helper: sort out poll_running vs poll_enabled") Reported-by: Bert Karwatzki Suggested-by: Dmitry Baryshkov Suggested-by: Alex Deucher Signed-off-by: Guchun Chen Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 4 ---- drivers/gpu/drm/amd/amdgpu/amdgpu_display.c | 4 ++++ 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 8fdd372f24378..d4519fbd526f2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4151,8 +4151,6 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D3)) DRM_WARN("smart shift update failed\n"); - drm_kms_helper_poll_disable(dev); - if (fbcon) drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, true); @@ -4249,8 +4247,6 @@ exit: if (fbcon) drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, false); - drm_kms_helper_poll_enable(dev); - amdgpu_ras_resume(adev); if (adev->mode_info.num_crtc) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index 503f89a766c37..d60fe7eb5579a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -1618,6 +1618,8 @@ int amdgpu_display_suspend_helper(struct amdgpu_device *adev) struct drm_connector_list_iter iter; int r; + drm_kms_helper_poll_disable(dev); + /* turn off display hw */ drm_modeset_lock_all(dev); drm_connector_list_iter_begin(dev, &iter); @@ -1694,6 +1696,8 @@ int amdgpu_display_resume_helper(struct amdgpu_device *adev) drm_modeset_unlock_all(dev); + drm_kms_helper_poll_enable(dev); + return 0; } -- cgit v1.2.3 From 65ba96e91b689c23d6fa99c11cfd65965dcddc47 Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Mon, 6 Mar 2023 15:48:48 +0800 Subject: drm/amdgpu: Move to common indirect reg access helper MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace soc15, nv, soc21 specific callbacks with common one. so we don't need to duplicate code when introduce new asics. Signed-off-by: Hawking Zhang Reviewed-by: Likun Gao Acked-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 4 --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 32 ++++++++++--------- drivers/gpu/drm/amd/amdgpu/nv.c | 49 +++--------------------------- drivers/gpu/drm/amd/amdgpu/soc15.c | 49 +++--------------------------- drivers/gpu/drm/amd/amdgpu/soc21.c | 48 +++-------------------------- 5 files changed, 30 insertions(+), 152 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 21c5f4ae5af99..5ae400ae5fee3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1111,16 +1111,12 @@ void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value) uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset); u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr); u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr); void amdgpu_device_indirect_wreg(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr, u32 reg_data); void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr, u64 reg_data); bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index d4519fbd526f2..d35d50287a441 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -676,20 +676,20 @@ void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v) * amdgpu_device_indirect_rreg - read an indirect register * * @adev: amdgpu_device pointer - * @pcie_index: mmio register offset - * @pcie_data: mmio register offset * @reg_addr: indirect register address to read from * * Returns the value of indirect register @reg_addr */ u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr) { - unsigned long flags; - u32 r; + unsigned long flags, pcie_index, pcie_data; void __iomem *pcie_index_offset; void __iomem *pcie_data_offset; + u32 r; + + pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev); + pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev); spin_lock_irqsave(&adev->pcie_idx_lock, flags); pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4; @@ -707,20 +707,20 @@ u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev, * amdgpu_device_indirect_rreg64 - read a 64bits indirect register * * @adev: amdgpu_device pointer - * @pcie_index: mmio register offset - * @pcie_data: mmio register offset * @reg_addr: indirect register address to read from * * Returns the value of indirect register @reg_addr */ u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr) { - unsigned long flags; - u64 r; + unsigned long flags, pcie_index, pcie_data; void __iomem *pcie_index_offset; void __iomem *pcie_data_offset; + u64 r; + + pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev); + pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev); spin_lock_irqsave(&adev->pcie_idx_lock, flags); pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4; @@ -750,13 +750,15 @@ u64 amdgpu_device_indirect_rreg64(struct amdgpu_device *adev, * */ void amdgpu_device_indirect_wreg(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr, u32 reg_data) { - unsigned long flags; + unsigned long flags, pcie_index, pcie_data; void __iomem *pcie_index_offset; void __iomem *pcie_data_offset; + pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev); + pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev); + spin_lock_irqsave(&adev->pcie_idx_lock, flags); pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4; pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4; @@ -779,13 +781,15 @@ void amdgpu_device_indirect_wreg(struct amdgpu_device *adev, * */ void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev, - u32 pcie_index, u32 pcie_data, u32 reg_addr, u64 reg_data) { - unsigned long flags; + unsigned long flags, pcie_index, pcie_data; void __iomem *pcie_index_offset; void __iomem *pcie_data_offset; + pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev); + pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev); + spin_lock_irqsave(&adev->pcie_idx_lock, flags); pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4; pcie_data_offset = (void __iomem *)adev->rmmio + pcie_data * 4; diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c index 22e25ca285f80..1a441d302b349 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.c +++ b/drivers/gpu/drm/amd/amdgpu/nv.c @@ -280,47 +280,6 @@ static int nv_query_video_codecs(struct amdgpu_device *adev, bool encode, } } -/* - * Indirect registers accessor - */ -static u32 nv_pcie_rreg(struct amdgpu_device *adev, u32 reg) -{ - unsigned long address, data; - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - return amdgpu_device_indirect_rreg(adev, address, data, reg); -} - -static void nv_pcie_wreg(struct amdgpu_device *adev, u32 reg, u32 v) -{ - unsigned long address, data; - - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - amdgpu_device_indirect_wreg(adev, address, data, reg, v); -} - -static u64 nv_pcie_rreg64(struct amdgpu_device *adev, u32 reg) -{ - unsigned long address, data; - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - return amdgpu_device_indirect_rreg64(adev, address, data, reg); -} - -static void nv_pcie_wreg64(struct amdgpu_device *adev, u32 reg, u64 v) -{ - unsigned long address, data; - - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - amdgpu_device_indirect_wreg64(adev, address, data, reg, v); -} - static u32 nv_didt_rreg(struct amdgpu_device *adev, u32 reg) { unsigned long flags, address, data; @@ -738,10 +697,10 @@ static int nv_common_early_init(void *handle) } adev->smc_rreg = NULL; adev->smc_wreg = NULL; - adev->pcie_rreg = &nv_pcie_rreg; - adev->pcie_wreg = &nv_pcie_wreg; - adev->pcie_rreg64 = &nv_pcie_rreg64; - adev->pcie_wreg64 = &nv_pcie_wreg64; + adev->pcie_rreg = &amdgpu_device_indirect_rreg; + adev->pcie_wreg = &amdgpu_device_indirect_wreg; + adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64; + adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64; adev->pciep_rreg = amdgpu_device_pcie_port_rreg; adev->pciep_wreg = amdgpu_device_pcie_port_wreg; diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 2eddd7f6cd41e..70343f9d25b98 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -191,47 +191,6 @@ static int soc15_query_video_codecs(struct amdgpu_device *adev, bool encode, } } -/* - * Indirect registers accessor - */ -static u32 soc15_pcie_rreg(struct amdgpu_device *adev, u32 reg) -{ - unsigned long address, data; - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - return amdgpu_device_indirect_rreg(adev, address, data, reg); -} - -static void soc15_pcie_wreg(struct amdgpu_device *adev, u32 reg, u32 v) -{ - unsigned long address, data; - - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - amdgpu_device_indirect_wreg(adev, address, data, reg, v); -} - -static u64 soc15_pcie_rreg64(struct amdgpu_device *adev, u32 reg) -{ - unsigned long address, data; - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - return amdgpu_device_indirect_rreg64(adev, address, data, reg); -} - -static void soc15_pcie_wreg64(struct amdgpu_device *adev, u32 reg, u64 v) -{ - unsigned long address, data; - - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - amdgpu_device_indirect_wreg64(adev, address, data, reg, v); -} - static u32 soc15_uvd_ctx_rreg(struct amdgpu_device *adev, u32 reg) { unsigned long flags, address, data; @@ -936,10 +895,10 @@ static int soc15_common_early_init(void *handle) } adev->smc_rreg = NULL; adev->smc_wreg = NULL; - adev->pcie_rreg = &soc15_pcie_rreg; - adev->pcie_wreg = &soc15_pcie_wreg; - adev->pcie_rreg64 = &soc15_pcie_rreg64; - adev->pcie_wreg64 = &soc15_pcie_wreg64; + adev->pcie_rreg = &amdgpu_device_indirect_rreg; + adev->pcie_wreg = &amdgpu_device_indirect_wreg; + adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64; + adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64; adev->uvd_ctx_rreg = &soc15_uvd_ctx_rreg; adev->uvd_ctx_wreg = &soc15_uvd_ctx_wreg; adev->didt_rreg = &soc15_didt_rreg; diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c index c82b3a7ea5f08..5a5cc38fc0aed 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc21.c +++ b/drivers/gpu/drm/amd/amdgpu/soc21.c @@ -196,46 +196,6 @@ static int soc21_query_video_codecs(struct amdgpu_device *adev, bool encode, return -EINVAL; } } -/* - * Indirect registers accessor - */ -static u32 soc21_pcie_rreg(struct amdgpu_device *adev, u32 reg) -{ - unsigned long address, data; - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - return amdgpu_device_indirect_rreg(adev, address, data, reg); -} - -static void soc21_pcie_wreg(struct amdgpu_device *adev, u32 reg, u32 v) -{ - unsigned long address, data; - - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - amdgpu_device_indirect_wreg(adev, address, data, reg, v); -} - -static u64 soc21_pcie_rreg64(struct amdgpu_device *adev, u32 reg) -{ - unsigned long address, data; - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - return amdgpu_device_indirect_rreg64(adev, address, data, reg); -} - -static void soc21_pcie_wreg64(struct amdgpu_device *adev, u32 reg, u64 v) -{ - unsigned long address, data; - - address = adev->nbio.funcs->get_pcie_index_offset(adev); - data = adev->nbio.funcs->get_pcie_data_offset(adev); - - amdgpu_device_indirect_wreg64(adev, address, data, reg, v); -} static u32 soc21_didt_rreg(struct amdgpu_device *adev, u32 reg) { @@ -650,10 +610,10 @@ static int soc21_common_early_init(void *handle) adev->rmmio_remap.bus_addr = adev->rmmio_base + MMIO_REG_HOLE_OFFSET; adev->smc_rreg = NULL; adev->smc_wreg = NULL; - adev->pcie_rreg = &soc21_pcie_rreg; - adev->pcie_wreg = &soc21_pcie_wreg; - adev->pcie_rreg64 = &soc21_pcie_rreg64; - adev->pcie_wreg64 = &soc21_pcie_wreg64; + adev->pcie_rreg = &amdgpu_device_indirect_rreg; + adev->pcie_wreg = &amdgpu_device_indirect_wreg; + adev->pcie_rreg64 = &amdgpu_device_indirect_rreg64; + adev->pcie_wreg64 = &amdgpu_device_indirect_wreg64; adev->pciep_rreg = amdgpu_device_pcie_port_rreg; adev->pciep_wreg = amdgpu_device_pcie_port_wreg; -- cgit v1.2.3 From dabc114e4bac903c365bfe6d7b6e8ed7fa38f8ad Mon Sep 17 00:00:00 2001 From: Hawking Zhang Date: Mon, 6 Mar 2023 15:59:27 +0800 Subject: drm/amdgpu: Move to common helper to query soc rev_id Replace soc15, nv, soc21 get_rev_id callback with common helper so we don't need to duplicate code when introduce new asics. Signed-off-by: Hawking Zhang Reviewed-by: Likun Gao Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 12 ++++++++++++ drivers/gpu/drm/amd/amdgpu/nv.c | 7 +------ drivers/gpu/drm/amd/amdgpu/soc15.c | 7 +------ drivers/gpu/drm/amd/amdgpu/soc21.c | 7 +------ 5 files changed, 16 insertions(+), 19 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 5ae400ae5fee3..c66706242254a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1118,7 +1118,7 @@ void amdgpu_device_indirect_wreg(struct amdgpu_device *adev, u32 reg_addr, u32 reg_data); void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev, u32 reg_addr, u64 reg_data); - +u32 amdgpu_device_get_rev_id(struct amdgpu_device *adev); bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type); bool amdgpu_device_has_dc_support(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index d35d50287a441..6298e3c1de391 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -807,6 +807,18 @@ void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->pcie_idx_lock, flags); } +/** + * amdgpu_device_get_rev_id - query device rev_id + * + * @adev: amdgpu_device pointer + * + * Return device rev_id + */ +u32 amdgpu_device_get_rev_id(struct amdgpu_device *adev) +{ + return adev->nbio.funcs->get_rev_id(adev); +} + /** * amdgpu_invalid_rreg - dummy reg read function * diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c index 1a441d302b349..f158e17b25266 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.c +++ b/drivers/gpu/drm/amd/amdgpu/nv.c @@ -567,11 +567,6 @@ void nv_set_virt_ops(struct amdgpu_device *adev) adev->virt.ops = &xgpu_nv_virt_ops; } -static uint32_t nv_get_rev_id(struct amdgpu_device *adev) -{ - return adev->nbio.funcs->get_rev_id(adev); -} - static bool nv_need_full_reset(struct amdgpu_device *adev) { return true; @@ -713,7 +708,7 @@ static int nv_common_early_init(void *handle) adev->asic_funcs = &nv_asic_funcs; - adev->rev_id = nv_get_rev_id(adev); + adev->rev_id = amdgpu_device_get_rev_id(adev); adev->external_rev_id = 0xff; /* TODO: split the GC and PG flags based on the relevant IP version for which * they are relevant. diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 70343f9d25b98..49541639b1985 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -654,11 +654,6 @@ const struct amdgpu_ip_block_version vega10_common_ip_block = .funcs = &soc15_common_ip_funcs, }; -static uint32_t soc15_get_rev_id(struct amdgpu_device *adev) -{ - return adev->nbio.funcs->get_rev_id(adev); -} - static void soc15_reg_base_init(struct amdgpu_device *adev) { /* Set IP register base before any HW register access */ @@ -908,7 +903,7 @@ static int soc15_common_early_init(void *handle) adev->se_cac_rreg = &soc15_se_cac_rreg; adev->se_cac_wreg = &soc15_se_cac_wreg; - adev->rev_id = soc15_get_rev_id(adev); + adev->rev_id = amdgpu_device_get_rev_id(adev); adev->external_rev_id = 0xFF; /* TODO: split the GC and PG flags based on the relevant IP version for which * they are relevant. diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c index 5a5cc38fc0aed..b3a33eab1fde2 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc21.c +++ b/drivers/gpu/drm/amd/amdgpu/soc21.c @@ -481,11 +481,6 @@ const struct amdgpu_ip_block_version soc21_common_ip_block = .funcs = &soc21_common_ip_funcs, }; -static uint32_t soc21_get_rev_id(struct amdgpu_device *adev) -{ - return adev->nbio.funcs->get_rev_id(adev); -} - static bool soc21_need_full_reset(struct amdgpu_device *adev) { switch (adev->ip_versions[GC_HWIP][0]) { @@ -626,7 +621,7 @@ static int soc21_common_early_init(void *handle) adev->asic_funcs = &soc21_asic_funcs; - adev->rev_id = soc21_get_rev_id(adev); + adev->rev_id = amdgpu_device_get_rev_id(adev); adev->external_rev_id = 0xff; switch (adev->ip_versions[GC_HWIP][0]) { case IP_VERSION(11, 0, 0): -- cgit v1.2.3 From 80bd2de1db0f25c82c110433e92c0703e05e5ae9 Mon Sep 17 00:00:00 2001 From: Lee Jones Date: Fri, 17 Mar 2023 08:16:44 +0000 Subject: drm/amd/amdgpu/amdgpu_device: Provide missing kerneldoc entry for 'reset_context' MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fixes the following W=1 kernel build warning(s): drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:5152: warning: Function parameter or member 'reset_context' not described in 'amdgpu_device_gpu_recover' Cc: Alex Deucher Cc: "Christian König" Cc: "Pan, Xinhui" Cc: David Airlie Cc: Daniel Vetter Cc: Sumit Semwal Cc: amd-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Cc: linux-media@vger.kernel.org Cc: linaro-mm-sig@lists.linaro.org Signed-off-by: Lee Jones Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 6298e3c1de391..8dbec2e1e6da3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5161,6 +5161,7 @@ static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev) * * @adev: amdgpu_device pointer * @job: which job trigger hang + * @reset_context: amdgpu reset context pointer * * Attempt to reset the GPU if it has hung (all asics). * Attempt to do soft-reset or full-reset and reinitialize Asic -- cgit v1.2.3 From 3ad5dcfe00a22404dfa2168f37fecd500ab26a6d Mon Sep 17 00:00:00 2001 From: Kai-Heng Feng Date: Wed, 15 Mar 2023 20:07:23 +0800 Subject: drm/amdgpu/nv: Apply ASPM quirk on Intel ADL + AMD Navi S2idle resume freeze can be observed on Intel ADL + AMD WX5500. This is caused by commit 0064b0ce85bb ("drm/amd/pm: enable ASPM by default"). The root cause is still not clear for now. So extend and apply the ASPM quirk from commit e02fe3bc7aba ("drm/amdgpu: vi: disable ASPM on Intel Alder Lake based systems"), to workaround the issue on Navi cards too. Fixes: 0064b0ce85bb ("drm/amd/pm: enable ASPM by default") Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2458 Reviewed-by: Alex Deucher Signed-off-by: Kai-Heng Feng Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 15 +++++++++++++++ drivers/gpu/drm/amd/amdgpu/nv.c | 2 +- drivers/gpu/drm/amd/amdgpu/vi.c | 17 +---------------- 4 files changed, 18 insertions(+), 17 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index a19a6489b117a..8cf2cc50b3de8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1248,6 +1248,7 @@ void amdgpu_device_pci_config_reset(struct amdgpu_device *adev); int amdgpu_device_pci_reset(struct amdgpu_device *adev); bool amdgpu_device_need_post(struct amdgpu_device *adev); bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev); +bool amdgpu_device_aspm_support_quirk(void); void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev, u64 num_bytes, u64 num_vis_bytes); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 8dbec2e1e6da3..55312dc780fda 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -81,6 +81,10 @@ #include +#if IS_ENABLED(CONFIG_X86) +#include +#endif + MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin"); MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin"); MODULE_FIRMWARE("amdgpu/raven_gpu_info.bin"); @@ -1373,6 +1377,17 @@ bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev) return pcie_aspm_enabled(adev->pdev); } +bool amdgpu_device_aspm_support_quirk(void) +{ +#if IS_ENABLED(CONFIG_X86) + struct cpuinfo_x86 *c = &cpu_data(0); + + return !(c->x86 == 6 && c->x86_model == INTEL_FAM6_ALDERLAKE); +#else + return true; +#endif +} + /* if we get transitioned to only one device, take VGA back */ /** * amdgpu_device_vga_set_decode - enable/disable vga decode diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c index 15f3c6745ea90..47420b403871c 100644 --- a/drivers/gpu/drm/amd/amdgpu/nv.c +++ b/drivers/gpu/drm/amd/amdgpu/nv.c @@ -522,7 +522,7 @@ static int nv_set_vce_clocks(struct amdgpu_device *adev, u32 evclk, u32 ecclk) static void nv_program_aspm(struct amdgpu_device *adev) { - if (!amdgpu_device_should_use_aspm(adev)) + if (!amdgpu_device_should_use_aspm(adev) || !amdgpu_device_aspm_support_quirk()) return; if (!(adev->flags & AMD_IS_APU) && diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 2512b70ea9929..531f173ade2d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -81,10 +81,6 @@ #include "mxgpu_vi.h" #include "amdgpu_dm.h" -#if IS_ENABLED(CONFIG_X86) -#include -#endif - #define ixPCIE_LC_L1_PM_SUBSTATE 0x100100C6 #define PCIE_LC_L1_PM_SUBSTATE__LC_L1_SUBSTATES_OVERRIDE_EN_MASK 0x00000001L #define PCIE_LC_L1_PM_SUBSTATE__LC_PCI_PM_L1_2_OVERRIDE_MASK 0x00000002L @@ -1120,24 +1116,13 @@ static void vi_enable_aspm(struct amdgpu_device *adev) WREG32_PCIE(ixPCIE_LC_CNTL, data); } -static bool aspm_support_quirk_check(void) -{ -#if IS_ENABLED(CONFIG_X86) - struct cpuinfo_x86 *c = &cpu_data(0); - - return !(c->x86 == 6 && c->x86_model == INTEL_FAM6_ALDERLAKE); -#else - return true; -#endif -} - static void vi_program_aspm(struct amdgpu_device *adev) { u32 data, data1, orig; bool bL1SS = false; bool bClkReqSupport = true; - if (!amdgpu_device_should_use_aspm(adev) || !aspm_support_quirk_check()) + if (!amdgpu_device_should_use_aspm(adev) || !amdgpu_device_aspm_support_quirk()) return; if (adev->flags & AMD_IS_APU || -- cgit v1.2.3 From ec64350d01cd45c91916e701964ed8cba383aa35 Mon Sep 17 00:00:00 2001 From: YiPeng Chai Date: Fri, 3 Mar 2023 10:03:36 +0800 Subject: drm/amdgpu: reinit mes ip block during reset on SRIOV Reinit mes ip block during reset on SRIOV. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 55312dc780fda..3c7ff7322a1c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3181,6 +3181,7 @@ static int amdgpu_device_ip_reinit_late_sriov(struct amdgpu_device *adev) AMD_IP_BLOCK_TYPE_DCE, AMD_IP_BLOCK_TYPE_GFX, AMD_IP_BLOCK_TYPE_SDMA, + AMD_IP_BLOCK_TYPE_MES, AMD_IP_BLOCK_TYPE_UVD, AMD_IP_BLOCK_TYPE_VCE, AMD_IP_BLOCK_TYPE_VCN -- cgit v1.2.3 From 28606c4e5877cdb27ea443b3f36e3f0bf9e7a5ac Mon Sep 17 00:00:00 2001 From: YiPeng Chai Date: Tue, 14 Mar 2023 14:13:34 +0800 Subject: drm/amdgpu: resume ras for gfx v11_0_3 during reset on SRIOV Gfx v11_0_3 supports ras on SRIOV, so need to resume ras during reset. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 3c7ff7322a1c4..bc158995105dc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -5347,8 +5347,9 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ if (r) adev->asic_reset_res = r; - /* Aldebaran supports ras in SRIOV, so need resume ras during reset */ - if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2)) + /* Aldebaran and gfx_11_0_3 support ras in SRIOV, so need resume ras during reset */ + if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(9, 4, 2) || + adev->ip_versions[GC_HWIP][0] == IP_VERSION(11, 0, 3)) amdgpu_ras_resume(adev); } else { r = amdgpu_do_asic_reset(device_list_handle, reset_context); -- cgit v1.2.3 From d2cdc01451973235e2b83ea85493817705a48263 Mon Sep 17 00:00:00 2001 From: Yifan Zha Date: Tue, 28 Mar 2023 11:42:24 +0800 Subject: drm/amdgpu: Add JPEG IP block to SRIOV reinit [Why] Reset(mode1) failed as JPRG IP did not reinit under sriov. [How] Add JPEG IP block to sriov reinit function. Signed-off-by: Yifan Zha Reviewed-by: Horace Chen Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index bc158995105dc..fac9312b16959 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3184,7 +3184,8 @@ static int amdgpu_device_ip_reinit_late_sriov(struct amdgpu_device *adev) AMD_IP_BLOCK_TYPE_MES, AMD_IP_BLOCK_TYPE_UVD, AMD_IP_BLOCK_TYPE_VCE, - AMD_IP_BLOCK_TYPE_VCN + AMD_IP_BLOCK_TYPE_VCN, + AMD_IP_BLOCK_TYPE_JPEG }; for (i = 0; i < ARRAY_SIZE(ip_order); i++) { -- cgit v1.2.3 From 11f25c844e29f85abb0b3ffdb360a2f82a2c4ed0 Mon Sep 17 00:00:00 2001 From: Srinivasan Shanmugam Date: Wed, 5 Apr 2023 20:41:09 +0530 Subject: drm/amd/amdgpu: Drop the hang limit parameter MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The driver doesn't resubmit jobs on hangs any more, hence drop the hang limit parameter - amdgpu_job_hang_limit, wherever it is used. Suggested-by: Christian König Cc: Alex Deucher Cc: Mario Limonciello Cc: Kent Russell Signed-off-by: Srinivasan Shanmugam Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 - drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 8 -------- 3 files changed, 1 insertion(+), 10 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 8cf2cc50b3de8..833996291dee7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -185,7 +185,6 @@ extern char *amdgpu_disable_cu; extern char *amdgpu_virtual_display; extern uint amdgpu_pp_feature_mask; extern uint amdgpu_force_long_training; -extern int amdgpu_job_hang_limit; extern int amdgpu_lbpw; extern int amdgpu_compute_multipipe; extern int amdgpu_gpu_recovery; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index fac9312b16959..4819b3f867502 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2365,7 +2365,7 @@ static int amdgpu_device_init_schedulers(struct amdgpu_device *adev) } r = drm_sched_init(&ring->sched, &amdgpu_sched_ops, - ring->num_hw_submission, amdgpu_job_hang_limit, + ring->num_hw_submission, 0, timeout, adev->reset_domain->wq, ring->sched_score, ring->name, adev->dev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index b4189d669b542..b1ca1ab6d6ad3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -157,7 +157,6 @@ char *amdgpu_virtual_display; */ uint amdgpu_pp_feature_mask = 0xfff7bfff; uint amdgpu_force_long_training; -int amdgpu_job_hang_limit; int amdgpu_lbpw = -1; int amdgpu_compute_multipipe = -1; int amdgpu_gpu_recovery = -1; /* auto */ @@ -520,13 +519,6 @@ MODULE_PARM_DESC(virtual_display, "Enable virtual display feature (the virtual_display will be set like xxxx:xx:xx.x,x;xxxx:xx:xx.x,x)"); module_param_named(virtual_display, amdgpu_virtual_display, charp, 0444); -/** - * DOC: job_hang_limit (int) - * Set how much time allow a job hang and not drop it. The default is 0. - */ -MODULE_PARM_DESC(job_hang_limit, "how much time allow a job hang and not drop it (default 0)"); -module_param_named(job_hang_limit, amdgpu_job_hang_limit, int ,0444); - /** * DOC: lbpw (int) * Override Load Balancing Per Watt (LBPW) support (1 = enable, 0 = disable). The default is -1 (auto, enabled). -- cgit v1.2.3 From 00fa40353bf3894adb495f8cce10a8bce43cd375 Mon Sep 17 00:00:00 2001 From: Sreekant Somasekharan Date: Mon, 13 Mar 2023 18:05:41 -0400 Subject: drm/amdkfd: Check PCIe atomics support on GFX11 to set CP_HQD_HQ_STATUS0[29] CP_HQD_HQ_STATUS0[29] bit will be used by CPFW to acknowledge whether PCIe atomics are supported. The default value of this bit is set to 0. Driver will check whether PCIe atomics are supported and set the bit to 1 if supported. This will force CPFW to use real atomic ops. If the bit is not set, CPFW will default to read/modify/write using the firmware itself. This is applicable only to GFX11 RS64 CP with MEC FW >= 509. If MEC FW < 509 and for all GFX11 F32 CP, PCIe atomics needs to be supported else it will skip the device. This commit also involves moving amdgpu_amdkfd_device_probe() function call after per-IP early_init loop in amdgpu_device_ip_early_init() function so as to check for RS64 enabled device. Signed-off-by: Sreekant Somasekharan Reviewed-by: Graham Sider Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +- drivers/gpu/drm/amd/amdkfd/kfd_device.c | 8 ++++++++ drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c | 7 +++++++ 3 files changed, 16 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 4819b3f867502..efdc6b73a55c3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2184,7 +2184,6 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) adev->has_pr3 = parent ? pci_pr3_present(parent) : false; } - amdgpu_amdkfd_device_probe(adev); adev->pm.pp_feature = amdgpu_pp_feature_mask; if (amdgpu_sriov_vf(adev) || sched_policy == KFD_SCHED_POLICY_NO_HWS) @@ -2240,6 +2239,7 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) if (!total) return -ENODEV; + amdgpu_amdkfd_device_probe(adev); adev->cg_flags &= amdgpu_cg_mask; adev->pg_flags &= amdgpu_pg_mask; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 0bf8d75950f72..e12ff49d01e1c 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -204,6 +204,14 @@ static void kfd_device_info_init(struct kfd_dev *kfd, /* Navi1x+ */ if (gc_version >= IP_VERSION(10, 1, 1)) kfd->device_info.needs_pci_atomics = true; + } else if (gc_version < IP_VERSION(12, 0, 0)) { + /* + * PCIe atomics support acknowledgment in GFX11 RS64 CPFW requires + * MEC version >= 509. Prior RS64 CPFW versions (and all F32) require + * PCIe atomics support. + */ + kfd->device_info.needs_pci_atomics = true; + kfd->device_info.no_atomic_fw_version = kfd->adev->gfx.rs64_enable ? 509 : 0; } } else { kfd->device_info.doorbell_size = 4; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c index 4a9af800b1f1c..04a73324174b8 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v11.c @@ -143,6 +143,13 @@ static void init_mqd(struct mqd_manager *mm, void **mqd, 1 << CP_HQD_QUANTUM__QUANTUM_SCALE__SHIFT | 1 << CP_HQD_QUANTUM__QUANTUM_DURATION__SHIFT; + /* + * GFX11 RS64 CPFW version >= 509 supports PCIe atomics support + * acknowledgment. + */ + if (amdgpu_amdkfd_have_atomics_support(mm->dev->adev)) + m->cp_hqd_hq_status0 |= 1 << 29; + if (q->format == KFD_QUEUE_FORMAT_AQL) { m->cp_hqd_aql_control = 1 << CP_HQD_AQL_CONTROL__CONTROL0__SHIFT; -- cgit v1.2.3 From 0512e9ffebca0f9a91f6e54b0da90976dce2b025 Mon Sep 17 00:00:00 2001 From: Shashank Sharma Date: Fri, 24 Feb 2023 11:25:07 +0100 Subject: drm/amdgpu: rename num_doorbells Rename doorbell.num_doorbells to doorbell.num_kernel_doorbells to make it more readable. Cc: Alex Deucher Cc: Christian Koenig Acked-by: Christian Koenig Signed-off-by: Shashank Sharma Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 6 +++--- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 22 +++++++++++----------- drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h | 4 +++- 3 files changed, 17 insertions(+), 15 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index f99d4873bf223..0385f7f692785 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -96,7 +96,7 @@ static void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev, size_t *start_offset) { /* - * The first num_doorbells are used by amdgpu. + * The first num_kernel_doorbells are used by amdgpu. * amdkfd takes whatever's left in the aperture. */ if (adev->enable_mes) { @@ -109,11 +109,11 @@ static void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev, *aperture_base = adev->doorbell.base; *aperture_size = 0; *start_offset = 0; - } else if (adev->doorbell.size > adev->doorbell.num_doorbells * + } else if (adev->doorbell.size > adev->doorbell.num_kernel_doorbells * sizeof(u32)) { *aperture_base = adev->doorbell.base; *aperture_size = adev->doorbell.size; - *start_offset = adev->doorbell.num_doorbells * sizeof(u32); + *start_offset = adev->doorbell.num_kernel_doorbells * sizeof(u32); } else { *aperture_base = 0; *aperture_size = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index efdc6b73a55c3..97068c4dee74f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -602,7 +602,7 @@ u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index) if (amdgpu_device_skip_hw_access(adev)) return 0; - if (index < adev->doorbell.num_doorbells) { + if (index < adev->doorbell.num_kernel_doorbells) { return readl(adev->doorbell.ptr + index); } else { DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", index); @@ -625,7 +625,7 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v) if (amdgpu_device_skip_hw_access(adev)) return; - if (index < adev->doorbell.num_doorbells) { + if (index < adev->doorbell.num_kernel_doorbells) { writel(v, adev->doorbell.ptr + index); } else { DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", index); @@ -646,7 +646,7 @@ u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index) if (amdgpu_device_skip_hw_access(adev)) return 0; - if (index < adev->doorbell.num_doorbells) { + if (index < adev->doorbell.num_kernel_doorbells) { return atomic64_read((atomic64_t *)(adev->doorbell.ptr + index)); } else { DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", index); @@ -669,7 +669,7 @@ void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v) if (amdgpu_device_skip_hw_access(adev)) return; - if (index < adev->doorbell.num_doorbells) { + if (index < adev->doorbell.num_kernel_doorbells) { atomic64_set((atomic64_t *)(adev->doorbell.ptr + index), v); } else { DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", index); @@ -1060,7 +1060,7 @@ static int amdgpu_device_doorbell_init(struct amdgpu_device *adev) if (adev->asic_type < CHIP_BONAIRE) { adev->doorbell.base = 0; adev->doorbell.size = 0; - adev->doorbell.num_doorbells = 0; + adev->doorbell.num_kernel_doorbells = 0; adev->doorbell.ptr = NULL; return 0; } @@ -1075,27 +1075,27 @@ static int amdgpu_device_doorbell_init(struct amdgpu_device *adev) adev->doorbell.size = pci_resource_len(adev->pdev, 2); if (adev->enable_mes) { - adev->doorbell.num_doorbells = + adev->doorbell.num_kernel_doorbells = adev->doorbell.size / sizeof(u32); } else { - adev->doorbell.num_doorbells = + adev->doorbell.num_kernel_doorbells = min_t(u32, adev->doorbell.size / sizeof(u32), adev->doorbell_index.max_assignment+1); - if (adev->doorbell.num_doorbells == 0) + if (adev->doorbell.num_kernel_doorbells == 0) return -EINVAL; /* For Vega, reserve and map two pages on doorbell BAR since SDMA * paging queue doorbell use the second page. The * AMDGPU_DOORBELL64_MAX_ASSIGNMENT definition assumes all the * doorbells are in the first page. So with paging queue enabled, - * the max num_doorbells should + 1 page (0x400 in dword) + * the max num_kernel_doorbells should + 1 page (0x400 in dword) */ if (adev->asic_type >= CHIP_VEGA10) - adev->doorbell.num_doorbells += 0x400; + adev->doorbell.num_kernel_doorbells += 0x400; } adev->doorbell.ptr = ioremap(adev->doorbell.base, - adev->doorbell.num_doorbells * + adev->doorbell.num_kernel_doorbells * sizeof(u32)); if (adev->doorbell.ptr == NULL) return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h index 7199b6b0be814..12263986f8897 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h @@ -29,7 +29,9 @@ struct amdgpu_doorbell { resource_size_t base; resource_size_t size; u32 __iomem *ptr; - u32 num_doorbells; /* Number of doorbells actually reserved for amdgpu. */ + + /* Number of doorbells reserved for amdgpu kernel driver */ + u32 num_kernel_doorbells; }; /* Reserved doorbells for amdgpu (including multimedia). -- cgit v1.2.3 From f22067419e9683f8fba40ca3a0d56fb3106c7c6f Mon Sep 17 00:00:00 2001 From: Aaron Liu Date: Wed, 5 Apr 2023 19:22:20 +0800 Subject: drm/amdgpu: skip kfd-iommu suspend/resume for S0ix GFX is in gfxoff mode during s0ix so we shouldn't need to actually execute kfd_iommu_suspend/kfd_iommu_resume operation. Signed-off-by: Aaron Liu Acked-by: Alex Deucher Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 97068c4dee74f..a2292acf06d09 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3305,9 +3305,11 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev) { int r; - r = amdgpu_amdkfd_resume_iommu(adev); - if (r) - return r; + if (!adev->in_s0ix) { + r = amdgpu_amdkfd_resume_iommu(adev); + if (r) + return r; + } r = amdgpu_device_ip_resume_phase1(adev); if (r) -- cgit v1.2.3 From 38eecbe086a4e52f54b2bbda8feba65d44addbef Mon Sep 17 00:00:00 2001 From: Chong Li Date: Fri, 14 Apr 2023 13:51:19 +0800 Subject: drm/amdgpu: release gpu full access after "amdgpu_device_ip_late_init" [WHY] Function "amdgpu_irq_update()" called by "amdgpu_device_ip_late_init()" is an atomic context. We shouldn't access registers through KIQ since "msleep()" may be called in "amdgpu_kiq_rreg()". [HOW] Move function "amdgpu_virt_release_full_gpu()" after function "amdgpu_device_ip_late_init()", to ensure that registers be accessed through RLCG instead of KIQ. Call Trace: show_stack+0x52/0x69 dump_stack_lvl+0x49/0x6d dump_stack+0x10/0x18 __schedule_bug.cold+0x4f/0x6b __schedule+0x473/0x5d0 ? __wake_up_klogd.part.0+0x40/0x70 ? vprintk_emit+0xbe/0x1f0 schedule+0x68/0x110 schedule_timeout+0x87/0x160 ? timer_migration_handler+0xa0/0xa0 msleep+0x2d/0x50 amdgpu_kiq_rreg+0x18d/0x1f0 [amdgpu] amdgpu_device_rreg.part.0+0x59/0xd0 [amdgpu] amdgpu_device_rreg+0x3a/0x50 [amdgpu] amdgpu_sriov_rreg+0x3c/0xb0 [amdgpu] gfx_v10_0_set_gfx_eop_interrupt_state.constprop.0+0x16c/0x190 [amdgpu] gfx_v10_0_set_eop_interrupt_state+0xa5/0xb0 [amdgpu] amdgpu_irq_update+0x53/0x80 [amdgpu] amdgpu_irq_get+0x7c/0xb0 [amdgpu] amdgpu_fence_driver_hw_init+0x58/0x90 [amdgpu] amdgpu_device_init.cold+0x16b7/0x2022 [amdgpu] Signed-off-by: Chong Li Reviewed-by: JingWen.Chen2@amd.com Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 32 ++++++++++++++++-------------- 1 file changed, 17 insertions(+), 15 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index a2292acf06d09..9b1eaba85bbd8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2539,8 +2539,6 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) amdgpu_fru_get_product_info(adev); init_failed: - if (amdgpu_sriov_vf(adev)) - amdgpu_virt_release_full_gpu(adev, true); return r; } @@ -3859,18 +3857,6 @@ fence_driver_init: r = amdgpu_device_ip_init(adev); if (r) { - /* failed in exclusive mode due to timeout */ - if (amdgpu_sriov_vf(adev) && - !amdgpu_sriov_runtime(adev) && - amdgpu_virt_mmio_blocked(adev) && - !amdgpu_virt_wait_reset(adev)) { - dev_err(adev->dev, "VF exclusive mode timeout\n"); - /* Don't send request since VF is inactive. */ - adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME; - adev->virt.ops = NULL; - r = -EAGAIN; - goto release_ras_con; - } dev_err(adev->dev, "amdgpu_device_ip_init failed\n"); amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_INIT_FAIL, 0, 0); goto release_ras_con; @@ -3939,8 +3925,10 @@ fence_driver_init: msecs_to_jiffies(AMDGPU_RESUME_MS)); } - if (amdgpu_sriov_vf(adev)) + if (amdgpu_sriov_vf(adev)) { + amdgpu_virt_release_full_gpu(adev, true); flush_delayed_work(&adev->delayed_init_work); + } r = sysfs_create_files(&adev->dev->kobj, amdgpu_dev_attributes); if (r) @@ -3980,6 +3968,20 @@ fence_driver_init: return 0; release_ras_con: + if (amdgpu_sriov_vf(adev)) + amdgpu_virt_release_full_gpu(adev, true); + + /* failed in exclusive mode due to timeout */ + if (amdgpu_sriov_vf(adev) && + !amdgpu_sriov_runtime(adev) && + amdgpu_virt_mmio_blocked(adev) && + !amdgpu_virt_wait_reset(adev)) { + dev_err(adev->dev, "VF exclusive mode timeout\n"); + /* Don't send request since VF is inactive. */ + adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME; + adev->virt.ops = NULL; + r = -EAGAIN; + } amdgpu_release_ras_context(adev); failed: -- cgit v1.2.3 From 4eea7fb980dc44545a32eec92e2662053b34cd9d Mon Sep 17 00:00:00 2001 From: lyndonli Date: Sun, 23 Apr 2023 17:05:15 +0800 Subject: drm/amdgpu: Use the default reset when loading or reloading the driver Below call trace and errors are observed when reloading amdgpu driver with the module parameter reset_method=3. It should do a default reset when loading or reloading the driver, regardless of the module parameter reset_method. v2: add comments inside and modify commit messages. [ +2.180243] [drm] psp gfx command ID_LOAD_TOC(0x20) failed and response status is (0x0) [ +0.000011] [drm:psp_hw_start [amdgpu]] *ERROR* Failed to load toc [ +0.000890] [drm:psp_hw_start [amdgpu]] *ERROR* PSP tmr init failed! [ +0.020683] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off. [ +0.000003] RIP: 0010:amdgpu_bo_release_notify+0x1ef/0x210 [amdgpu] [ +0.000004] Call Trace: [ +0.000003] [ +0.000008] ttm_bo_release+0x2c4/0x330 [amdttm] [ +0.000026] amdttm_bo_put+0x3c/0x70 [amdttm] [ +0.000020] amdgpu_bo_free_kernel+0xe6/0x140 [amdgpu] [ +0.000728] psp_v11_0_ring_destroy+0x34/0x60 [amdgpu] [ +0.000826] psp_hw_init+0xe7/0x2f0 [amdgpu] [ +0.000813] amdgpu_device_fw_loading+0x1ad/0x2d0 [amdgpu] [ +0.000731] amdgpu_device_init.cold+0x108e/0x2002 [amdgpu] [ +0.001071] ? do_pci_enable_device+0xe1/0x110 [ +0.000011] amdgpu_driver_load_kms+0x1a/0x160 [amdgpu] [ +0.000729] amdgpu_pci_probe+0x179/0x3a0 [amdgpu] Signed-off-by: lyndonli Signed-off-by: Yunxiang Li Reviewed-by: Feifei Xu Reviewed-by: Kenneth Feng Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 9b1eaba85bbd8..981a9cfb63b57 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3578,6 +3578,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, int r, i; bool px = false; u32 max_MBps; + int tmp; adev->shutdown = false; adev->flags = flags; @@ -3799,7 +3800,13 @@ int amdgpu_device_init(struct amdgpu_device *adev, } } } else { + tmp = amdgpu_reset_method; + /* It should do a default reset when loading or reloading the driver, + * regardless of the module parameter reset_method. + */ + amdgpu_reset_method = AMD_RESET_METHOD_NONE; r = amdgpu_asic_reset(adev); + amdgpu_reset_method = tmp; if (r) { dev_err(adev->dev, "asic reset on init failed\n"); goto failed; -- cgit v1.2.3 From af7828fbceed4f9e503034111066a0adef3db383 Mon Sep 17 00:00:00 2001 From: Yifan Zhang Date: Thu, 27 Apr 2023 14:01:05 +0800 Subject: drm/amdgpu: set gfx9 onwards APU atomics support to be true APUs w/ gfx9 onwards doesn't reply on PCIe atomics, rather it is internal path w/ native atomic support. Set have_atomics_support to true. Signed-off-by: Yifan Zhang Reviewed-by: Lang Yu Acked-by: Felix Kuehling Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 981a9cfb63b57..e348a62ec0ec6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3757,6 +3757,12 @@ int amdgpu_device_init(struct amdgpu_device *adev, adev->have_atomics_support = ((struct amd_sriov_msg_pf2vf_info *) adev->virt.fw_reserve.p_pf2vf)->pcie_atomic_ops_support_flags == (PCI_EXP_DEVCAP2_ATOMIC_COMP32 | PCI_EXP_DEVCAP2_ATOMIC_COMP64); + /* APUs w/ gfx9 onwards doesn't reply on PCIe atomics, rather it is a + * internal path natively support atomics, set have_atomics_support to true. + */ + else if ((adev->flags & AMD_IS_APU) && + (adev->ip_versions[GC_HWIP][0] > IP_VERSION(9, 0, 0))) + adev->have_atomics_support = true; else adev->have_atomics_support = !pci_enable_atomic_ops_to_root(adev->pdev, -- cgit v1.2.3 From 6c032c37ac3ef3b7df30937c785ecc4da428edc0 Mon Sep 17 00:00:00 2001 From: "Lin.Cao" Date: Mon, 8 May 2023 17:28:41 +0800 Subject: drm/amdgpu: Fix vram recover doesn't work after whole GPU reset (v2) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit v1: Vmbo->shadow is used to back vram bo up when vram lost. So that we should set shadow as vmbo->shadow to recover vmbo->bo v2: Modify if(vmbo->shadow) shadow = vmbo->shadow as if(!vmbo->shadow) continue; Fixes: e18aaea733da ("drm/amdgpu: move shadow_list to amdgpu_bo_vm") Reviewed-by: Christian König Signed-off-by: Lin.Cao Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_device.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index e348a62ec0ec6..5c7d40873ee20 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4512,7 +4512,11 @@ static int amdgpu_device_recover_vram(struct amdgpu_device *adev) dev_info(adev->dev, "recover vram bo from shadow start\n"); mutex_lock(&adev->shadow_list_lock); list_for_each_entry(vmbo, &adev->shadow_list, shadow_list) { - shadow = &vmbo->bo; + /* If vm is compute context or adev is APU, shadow will be NULL */ + if (!vmbo->shadow) + continue; + shadow = vmbo->shadow; + /* No need to recover an evicted BO */ if (shadow->tbo.resource->mem_type != TTM_PL_TT || shadow->tbo.resource->start == AMDGPU_BO_INVALID_OFFSET || -- cgit v1.2.3