From 56b53c0b5aa5de49747351b2ad323fd36089eb52 Mon Sep 17 00:00:00 2001 From: Dennis Li Date: Wed, 10 Mar 2021 17:20:45 +0800 Subject: drm/amdgpu: add codes to capture invalid hardware access when recovery MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When recovery thread has begun GPU reset, there should be not other threads to access hardware, otherwise system randomly hang. v2 (chk): rewritten from scratch, use trylock and lockdep instead of hand wiring the logic. v3: add in_irq check v4: change to check in_task Signed-off-by: Dennis Li Signed-off-by: Christian König Reviewed-by: Christian König Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index a037c223c251..4c2fabe93d59 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1390,6 +1390,8 @@ void amdgpu_pci_resume(struct pci_dev *pdev); bool amdgpu_device_cache_pci_state(struct pci_dev *pdev); bool amdgpu_device_load_pci_state(struct pci_dev *pdev); +bool amdgpu_device_skip_hw_access(struct amdgpu_device *adev); + #include "amdgpu_object.h" static inline bool amdgpu_is_tmz(struct amdgpu_device *adev) -- cgit v1.2.3-70-g09d2 From b98c6299ef992660f5ca4392287a11ea2439c664 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 10 Mar 2021 00:43:35 -0500 Subject: drm/amdgpu: disentangle HG systems from vgaswitcheroo There's no need to keep vgaswitcheroo around for HG systems. They don't use muxes and their power control is handled via ACPI. Reviewed-by: Evan Quan Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 3 ++- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 38 +++++++++++++----------------- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 34 +++++++++++--------------- drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 9 +++---- 4 files changed, 35 insertions(+), 49 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 4c2fabe93d59..7da8d26f96b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1275,8 +1275,9 @@ void amdgpu_device_program_register_sequence(struct amdgpu_device *adev, const u32 *registers, const u32 array_size); -bool amdgpu_device_supports_atpx(struct drm_device *dev); int amdgpu_device_mode1_reset(struct amdgpu_device *adev); +bool amdgpu_device_supports_atpx(struct drm_device *dev); +bool amdgpu_device_supports_px(struct drm_device *dev); bool amdgpu_device_supports_boco(struct drm_device *dev); bool amdgpu_device_supports_baco(struct drm_device *dev); bool amdgpu_device_is_peer_accessible(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 7db60edfb5d1..1c3044f2e767 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -212,18 +212,18 @@ static DEVICE_ATTR(serial_number, S_IRUGO, amdgpu_device_get_serial_number, NULL); /** - * amdgpu_device_supports_atpx - Is the device a dGPU with HG/PX power control + * amdgpu_device_supports_px - Is the device a dGPU with ATPX power control * * @dev: drm_device pointer * - * Returns true if the device is a dGPU with HG/PX power control, + * Returns true if the device is a dGPU with ATPX power control, * otherwise return false. */ -bool amdgpu_device_supports_atpx(struct drm_device *dev) +bool amdgpu_device_supports_px(struct drm_device *dev) { struct amdgpu_device *adev = drm_to_adev(dev); - if (adev->flags & AMD_IS_PX) + if ((adev->flags & AMD_IS_PX) && !amdgpu_is_atpx_hybrid()) return true; return false; } @@ -233,14 +233,15 @@ bool amdgpu_device_supports_atpx(struct drm_device *dev) * * @dev: drm_device pointer * - * Returns true if the device is a dGPU with HG/PX power control, + * Returns true if the device is a dGPU with ACPI power control, * otherwise return false. */ bool amdgpu_device_supports_boco(struct drm_device *dev) { struct amdgpu_device *adev = drm_to_adev(dev); - if (adev->has_pr3) + if (adev->has_pr3 || + ((adev->flags & AMD_IS_PX) && amdgpu_is_atpx_hybrid())) return true; return false; } @@ -1420,7 +1421,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, struct drm_device *dev = pci_get_drvdata(pdev); int r; - if (amdgpu_device_supports_atpx(dev) && state == VGA_SWITCHEROO_OFF) + if (amdgpu_device_supports_px(dev) && state == VGA_SWITCHEROO_OFF) return; if (state == VGA_SWITCHEROO_ON) { @@ -3226,7 +3227,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, struct drm_device *ddev = adev_to_drm(adev); struct pci_dev *pdev = adev->pdev; int r, i; - bool atpx = false; + bool px = false; u32 max_MBps; adev->shutdown = false; @@ -3388,16 +3389,12 @@ int amdgpu_device_init(struct amdgpu_device *adev, if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA) vga_client_register(adev->pdev, adev, NULL, amdgpu_device_vga_set_decode); - if (amdgpu_device_supports_atpx(ddev)) - atpx = true; - if (amdgpu_has_atpx() && - (amdgpu_is_atpx_hybrid() || - amdgpu_has_atpx_dgpu_power_cntl()) && - !pci_is_thunderbolt_attached(adev->pdev)) + if (amdgpu_device_supports_px(ddev)) { + px = true; vga_switcheroo_register_client(adev->pdev, - &amdgpu_switcheroo_ops, atpx); - if (atpx) + &amdgpu_switcheroo_ops, px); vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain); + } if (amdgpu_emu_mode == 1) { /* post the asic on emulation mode */ @@ -3604,7 +3601,7 @@ release_ras_con: failed: amdgpu_vf_error_trans_all(adev); - if (atpx) + if (px) vga_switcheroo_fini_domain_pm_ops(adev->dev); failed_unmap: @@ -3664,13 +3661,10 @@ void amdgpu_device_fini(struct amdgpu_device *adev) kfree(adev->bios); adev->bios = NULL; - if (amdgpu_has_atpx() && - (amdgpu_is_atpx_hybrid() || - amdgpu_has_atpx_dgpu_power_cntl()) && - !pci_is_thunderbolt_attached(adev->pdev)) + if (amdgpu_device_supports_px(adev_to_drm(adev))) { vga_switcheroo_unregister_client(adev->pdev); - if (amdgpu_device_supports_atpx(adev_to_drm(adev))) vga_switcheroo_fini_domain_pm_ops(adev->dev); + } if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA) vga_client_register(adev->pdev, NULL, NULL, NULL); iounmap(adev->rmmio); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 1508da68cfbb..daaacf9067b8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -1411,9 +1411,7 @@ static int amdgpu_pmops_prepare(struct device *dev) /* Return a positive number here so * DPM_FLAG_SMART_SUSPEND works properly */ - if ((amdgpu_device_supports_atpx(drm_dev) && - amdgpu_is_atpx_hybrid()) || - amdgpu_device_supports_boco(drm_dev)) + if (amdgpu_device_supports_boco(drm_dev)) return pm_runtime_suspended(dev) && pm_suspend_via_firmware(); @@ -1502,7 +1500,7 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev) } adev->in_runpm = true; - if (amdgpu_device_supports_atpx(drm_dev)) + if (amdgpu_device_supports_px(drm_dev)) drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; ret = amdgpu_device_suspend(drm_dev, false); @@ -1511,16 +1509,14 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev) return ret; } - if (amdgpu_device_supports_atpx(drm_dev)) { + if (amdgpu_device_supports_px(drm_dev)) { /* Only need to handle PCI state in the driver for ATPX * PCI core handles it for _PR3. */ - if (!amdgpu_is_atpx_hybrid()) { - amdgpu_device_cache_pci_state(pdev); - pci_disable_device(pdev); - pci_ignore_hotplug(pdev); - pci_set_power_state(pdev, PCI_D3cold); - } + amdgpu_device_cache_pci_state(pdev); + pci_disable_device(pdev); + pci_ignore_hotplug(pdev); + pci_set_power_state(pdev, PCI_D3cold); drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF; } else if (amdgpu_device_supports_baco(drm_dev)) { amdgpu_device_baco_enter(drm_dev); @@ -1539,19 +1535,17 @@ static int amdgpu_pmops_runtime_resume(struct device *dev) if (!adev->runpm) return -EINVAL; - if (amdgpu_device_supports_atpx(drm_dev)) { + if (amdgpu_device_supports_px(drm_dev)) { drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; /* Only need to handle PCI state in the driver for ATPX * PCI core handles it for _PR3. */ - if (!amdgpu_is_atpx_hybrid()) { - pci_set_power_state(pdev, PCI_D0); - amdgpu_device_load_pci_state(pdev); - ret = pci_enable_device(pdev); - if (ret) - return ret; - } + pci_set_power_state(pdev, PCI_D0); + amdgpu_device_load_pci_state(pdev); + ret = pci_enable_device(pdev); + if (ret) + return ret; pci_set_master(pdev); } else if (amdgpu_device_supports_boco(drm_dev)) { /* Only need to handle PCI state in the driver for ATPX @@ -1562,7 +1556,7 @@ static int amdgpu_pmops_runtime_resume(struct device *dev) amdgpu_device_baco_exit(drm_dev); } ret = amdgpu_device_resume(drm_dev, false); - if (amdgpu_device_supports_atpx(drm_dev)) + if (amdgpu_device_supports_px(drm_dev)) drm_dev->switch_power_state = DRM_SWITCH_POWER_ON; adev->in_runpm = false; return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 23c8f4cb0ac9..8844f650b17f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -159,7 +159,7 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags) goto out; } - if (amdgpu_device_supports_atpx(dev) && + if (amdgpu_device_supports_px(dev) && (amdgpu_runtime_pm != 0)) { /* enable runpm by default for atpx */ adev->runpm = true; dev_info(adev->dev, "Using ATPX for runtime pm\n"); @@ -200,13 +200,10 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags) if (adev->runpm) { /* only need to skip on ATPX */ - if (amdgpu_device_supports_atpx(dev) && - !amdgpu_is_atpx_hybrid()) + if (amdgpu_device_supports_px(dev)) dev_pm_set_driver_flags(dev->dev, DPM_FLAG_NO_DIRECT_COMPLETE); /* we want direct complete for BOCO */ - if ((amdgpu_device_supports_atpx(dev) && - amdgpu_is_atpx_hybrid()) || - amdgpu_device_supports_boco(dev)) + if (amdgpu_device_supports_boco(dev)) dev_pm_set_driver_flags(dev->dev, DPM_FLAG_SMART_PREPARE | DPM_FLAG_SMART_SUSPEND | DPM_FLAG_MAY_SKIP_RESUME); -- cgit v1.2.3-70-g09d2 From 62498733d4c4fde8bc15215c5502923ff8224f86 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 12 Mar 2021 15:22:36 -0500 Subject: drm/amdgpu: rework S3/S4/S0ix state handling Set flags at the top level pmops callbacks to track state. This cleans up the current set of flags and properly handles S4 on S0ix capable systems. Reviewed-by: Evan Quan Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 10 +++------- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 12 +++--------- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 30 ++++++++++++++++++------------ drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c | 2 +- 4 files changed, 25 insertions(+), 29 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 7da8d26f96b8..3316618f809d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1030,13 +1030,9 @@ struct amdgpu_device { /* s3/s4 mask */ bool in_suspend; - bool in_hibernate; - - /* - * The combination flag in_poweroff_reboot_com used to identify the poweroff - * and reboot opt in the s0i3 system-wide suspend. - */ - bool in_poweroff_reboot_com; + bool in_s3; + bool in_s4; + bool in_s0ix; atomic_t in_gpu_reset; enum pp_mp1_state mp1_state; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 87ebf67086e0..2359449567be 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2685,8 +2685,7 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev) { int i, r; - if (adev->in_poweroff_reboot_com || adev->in_hibernate || - !amdgpu_acpi_is_s0ix_supported(adev) || amdgpu_in_reset(adev)) { + if (!adev->in_s0ix || amdgpu_in_reset(adev)) { amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE); amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE); } @@ -3766,12 +3765,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) amdgpu_fence_driver_suspend(adev); - /* - * TODO: Need figure out the each GNB IP idle off dependency and then - * improve the AMDGPU suspend/resume sequence for system-wide Sx entry/exit. - */ - if (adev->in_poweroff_reboot_com || adev->in_hibernate || - !amdgpu_acpi_is_s0ix_supported(adev) || amdgpu_in_reset(adev)) + if (!adev->in_s0ix || amdgpu_in_reset(adev)) r = amdgpu_device_ip_suspend_phase2(adev); else amdgpu_gfx_state_change_set(adev, sGpuChangeState_D3Entry); @@ -3805,7 +3799,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon) if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; - if (amdgpu_acpi_is_s0ix_supported(adev)) + if (adev->in_s0ix) amdgpu_gfx_state_change_set(adev, sGpuChangeState_D0Entry); /* post card */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index daaacf9067b8..02b75fa64d7d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -1335,9 +1335,7 @@ amdgpu_pci_shutdown(struct pci_dev *pdev) */ if (!amdgpu_passthrough(adev)) adev->mp1_state = PP_MP1_STATE_UNLOAD; - adev->in_poweroff_reboot_com = true; amdgpu_device_ip_suspend(adev); - adev->in_poweroff_reboot_com = false; adev->mp1_state = PP_MP1_STATE_NONE; } @@ -1426,15 +1424,28 @@ static void amdgpu_pmops_complete(struct device *dev) static int amdgpu_pmops_suspend(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(drm_dev); + int r; - return amdgpu_device_suspend(drm_dev, true); + if (amdgpu_acpi_is_s0ix_supported(adev)) + adev->in_s0ix = true; + adev->in_s3 = true; + r = amdgpu_device_suspend(drm_dev, true); + adev->in_s3 = false; + + return r; } static int amdgpu_pmops_resume(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); + struct amdgpu_device *adev = drm_to_adev(drm_dev); + int r; - return amdgpu_device_resume(drm_dev, true); + r = amdgpu_device_resume(drm_dev, true); + if (amdgpu_acpi_is_s0ix_supported(adev)) + adev->in_s0ix = false; + return r; } static int amdgpu_pmops_freeze(struct device *dev) @@ -1443,9 +1454,9 @@ static int amdgpu_pmops_freeze(struct device *dev) struct amdgpu_device *adev = drm_to_adev(drm_dev); int r; - adev->in_hibernate = true; + adev->in_s4 = true; r = amdgpu_device_suspend(drm_dev, true); - adev->in_hibernate = false; + adev->in_s4 = false; if (r) return r; return amdgpu_asic_reset(adev); @@ -1461,13 +1472,8 @@ static int amdgpu_pmops_thaw(struct device *dev) static int amdgpu_pmops_poweroff(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); - struct amdgpu_device *adev = drm_to_adev(drm_dev); - int r; - adev->in_poweroff_reboot_com = true; - r = amdgpu_device_suspend(drm_dev, true); - adev->in_poweroff_reboot_com = false; - return r; + return amdgpu_device_suspend(drm_dev, true); } static int amdgpu_pmops_restore(struct device *dev) diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index e722adcf2f53..cfcac110ed84 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -1337,7 +1337,7 @@ static int smu_disable_dpms(struct smu_context *smu) bool use_baco = !smu->is_apu && ((amdgpu_in_reset(adev) && (amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_BACO)) || - ((adev->in_runpm || adev->in_hibernate) && amdgpu_asic_supports_baco(adev))); + ((adev->in_runpm || adev->in_s4) && amdgpu_asic_supports_baco(adev))); /* * For custom pptable uploading, skip the DPM features -- cgit v1.2.3-70-g09d2 From 32c811b0972ca10a49217c35080771c64b7ddca6 Mon Sep 17 00:00:00 2001 From: Wan Jiabing Date: Mon, 22 Mar 2021 20:02:25 +0800 Subject: drivers: gpu: Remove duplicate include of amdgpu_hdp.h MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit amdgpu_hdp.h has been included at line 91, so remove the duplicate include. Acked-by: Christian König Signed-off-by: Wan Jiabing Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 3316618f809d..6434a15789f6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -107,7 +107,6 @@ #include "amdgpu_gfxhub.h" #include "amdgpu_df.h" #include "amdgpu_smuio.h" -#include "amdgpu_hdp.h" #define MAX_GPU_INSTANCE 16 -- cgit v1.2.3-70-g09d2 From e071dce38f0a3acb2f6d019c6bb7933fd20d3357 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Tue, 16 Mar 2021 20:19:06 +0800 Subject: drm/amdgpu: Add reset control to amdgpu_device v1: Add generic amdgpu_reset_control to handle different types of resets. It may be added at device, hive or ip level. Each reset control has a list of handlers associated with it to handle different types of reset. Reset control is responsible for choosing the right handler given a particular reset context. Handler objects may implement a set of functions on how to handle a particular type of reset. prepare_env = Prepare environment/software context (not used currently). prepare_hwcontext = Prepare hardware context for the reset. perform_reset = Perform the type of reset. restore_hwcontext = Restore the hw context after reset. restore_env = Restore the environment after reset (not used currently). Reset context carries the context of reset, as of now this is based on the parameters used for current set of resets. v2: Fix coding style Signed-off-by: Lijo Lazar Reviewed-by: Feifei Xu Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/Makefile | 5 ++ drivers/gpu/drm/amd/amdgpu/amdgpu.h | 3 ++ drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c | 82 +++++++++++++++++++++++++++++ drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h | 85 +++++++++++++++++++++++++++++++ 4 files changed, 175 insertions(+) create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c create mode 100644 drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index 741b68874e53..a0a5fd1788b5 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -179,9 +179,14 @@ amdgpu-y += \ smuio_v11_0_6.o \ smuio_v13_0.o +# add reset block +amdgpu-y += \ + amdgpu_reset.o + # add amdkfd interfaces amdgpu-y += amdgpu_amdkfd.o + ifneq ($(CONFIG_HSA_AMD),) AMDKFD_PATH := ../amdkfd include $(FULL_AMD_PATH)/amdkfd/Makefile diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 6434a15789f6..4b72fb32f22c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -270,6 +270,7 @@ struct amdgpu_bo_va_mapping; struct amdgpu_atif; struct kfd_vm_fault_info; struct amdgpu_hive_info; +struct amdgpu_reset_control; enum amdgpu_cp_irq { AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP = 0, @@ -588,6 +589,7 @@ struct amdgpu_allowed_register_entry { }; enum amd_reset_method { + AMD_RESET_METHOD_NONE = -1, AMD_RESET_METHOD_LEGACY = 0, AMD_RESET_METHOD_MODE0, AMD_RESET_METHOD_MODE1, @@ -1073,6 +1075,7 @@ struct amdgpu_device { bool in_pci_err_recovery; struct pci_saved_state *pci_state; + struct amdgpu_reset_control *reset_cntl; }; static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c new file mode 100644 index 000000000000..5bc94b8320ba --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c @@ -0,0 +1,82 @@ +/* + * Copyright 2021 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include "amdgpu_reset.h" +#include "aldebaran.h" + +int amdgpu_reset_add_handler(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_handler *handler) +{ + /* TODO: Check if handler exists? */ + list_add_tail(&handler->handler_list, &reset_ctl->reset_handlers); + return 0; +} + +int amdgpu_reset_init(struct amdgpu_device *adev) +{ + int ret = 0; + + return ret; +} + +int amdgpu_reset_fini(struct amdgpu_device *adev) +{ + int ret = 0; + + return ret; +} + +int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev, + struct amdgpu_reset_context *reset_context) +{ + struct amdgpu_reset_handler *reset_handler = NULL; + + if (adev->reset_cntl && adev->reset_cntl->get_reset_handler) + reset_handler = adev->reset_cntl->get_reset_handler( + adev->reset_cntl, reset_context); + if (!reset_handler) + return -ENOSYS; + + return reset_handler->prepare_hwcontext(adev->reset_cntl, + reset_context); +} + +int amdgpu_reset_perform_reset(struct amdgpu_device *adev, + struct amdgpu_reset_context *reset_context) +{ + int ret; + struct amdgpu_reset_handler *reset_handler = NULL; + + if (adev->reset_cntl) + reset_handler = adev->reset_cntl->get_reset_handler( + adev->reset_cntl, reset_context); + if (!reset_handler) + return -ENOSYS; + + ret = reset_handler->perform_reset(adev->reset_cntl, reset_context); + if (ret) + return ret; + + return reset_handler->restore_hwcontext(adev->reset_cntl, + reset_context); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h new file mode 100644 index 000000000000..dc84d871fe72 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h @@ -0,0 +1,85 @@ +/* + * Copyright 2021 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __AMDUGPU_RESET_H__ +#define __AMDGPU_RESET_H__ + +#include "amdgpu.h" + +enum AMDGPU_RESET_FLAGS { + + AMDGPU_NEED_FULL_RESET = 0, + AMDGPU_SKIP_HW_RESET = 1, +}; + +struct amdgpu_reset_context { + enum amd_reset_method method; + struct amdgpu_device *reset_req_dev; + struct amdgpu_job *job; + struct amdgpu_hive_info *hive; + unsigned long flags; +}; + +struct amdgpu_reset_handler { + enum amd_reset_method reset_method; + struct list_head handler_list; + int (*prepare_env)(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_context *context); + int (*prepare_hwcontext)(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_context *context); + int (*perform_reset)(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_context *context); + int (*restore_hwcontext)(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_context *context); + int (*restore_env)(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_context *context); + + int (*do_reset)(struct amdgpu_device *adev); +}; + +struct amdgpu_reset_control { + void *handle; + struct work_struct reset_work; + struct mutex reset_lock; + struct list_head reset_handlers; + atomic_t in_reset; + enum amd_reset_method active_reset; + struct amdgpu_reset_handler *(*get_reset_handler)( + struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_context *context); + void (*async_reset)(struct work_struct *work); +}; + +int amdgpu_reset_init(struct amdgpu_device *adev); +int amdgpu_reset_fini(struct amdgpu_device *adev); + +int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev, + struct amdgpu_reset_context *reset_context); + +int amdgpu_reset_perform_reset(struct amdgpu_device *adev, + struct amdgpu_reset_context *reset_context); + +int amdgpu_reset_add_handler(struct amdgpu_reset_control *reset_ctl, + struct amdgpu_reset_handler *handler); + +#endif -- cgit v1.2.3-70-g09d2 From 04442bf70debb197d4ed4e850aa77213e685b352 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Tue, 16 Mar 2021 20:31:51 +0800 Subject: drm/amdgpu: Add reset control handling to reset workflow This prefers reset control based handling if it's implemented for a particular ASIC. If not, it takes the legacy path. It uses the legacy method of preparing environment (job, scheduler tasks) and restoring environment. v2: remove unused variable (Alex) Signed-off-by: Lijo Lazar Reviewed-by: Feifei Xu Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 11 ++- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 108 +++++++++++++++++++++-------- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 17 ++++- 3 files changed, 97 insertions(+), 39 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 4b72fb32f22c..ec21a71c74eb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -270,6 +270,7 @@ struct amdgpu_bo_va_mapping; struct amdgpu_atif; struct kfd_vm_fault_info; struct amdgpu_hive_info; +struct amdgpu_reset_context; struct amdgpu_reset_control; enum amdgpu_cp_irq { @@ -1075,6 +1076,7 @@ struct amdgpu_device { bool in_pci_err_recovery; struct pci_saved_state *pci_state; + struct amdgpu_reset_control *reset_cntl; }; @@ -1127,13 +1129,10 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type); bool amdgpu_device_has_dc_support(struct amdgpu_device *adev); int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, - struct amdgpu_job *job, - bool *need_full_reset_arg); + struct amdgpu_reset_context *reset_context); -int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, - struct list_head *device_list_handle, - bool *need_full_reset_arg, - bool skip_hw_reset); +int amdgpu_do_asic_reset(struct list_head *device_list_handle, + struct amdgpu_reset_context *reset_context); int emu_soc_asic_init(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 3cbe8137a6af..a51f470631d4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -65,6 +65,7 @@ #include "amdgpu_ras.h" #include "amdgpu_pmu.h" #include "amdgpu_fru_eeprom.h" +#include "amdgpu_reset.h" #include #include @@ -3421,6 +3422,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, goto fence_driver_init; } + amdgpu_reset_init(adev); + /* detect if we are with an SRIOV vbios */ amdgpu_device_detect_sriov_bios(adev); @@ -3671,6 +3674,9 @@ void amdgpu_device_fini(struct amdgpu_device *adev) release_firmware(adev->firmware.gpu_info_fw); adev->firmware.gpu_info_fw = NULL; adev->accel_working = false; + + amdgpu_reset_fini(adev); + /* free i2c buses */ if (!amdgpu_device_has_dc_support(adev)) amdgpu_i2c_fini(adev); @@ -4239,11 +4245,15 @@ int amdgpu_device_mode1_reset(struct amdgpu_device *adev) } int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, - struct amdgpu_job *job, - bool *need_full_reset_arg) + struct amdgpu_reset_context *reset_context) { int i, r = 0; - bool need_full_reset = *need_full_reset_arg; + struct amdgpu_job *job = NULL; + bool need_full_reset = + test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + + if (reset_context->reset_req_dev == adev) + job = reset_context->job; /* no need to dump if device is not in good state during probe period */ if (!adev->gmc.xgmi.pending_reset) @@ -4268,6 +4278,10 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if(job) drm_sched_increase_karma(&job->base); + r = amdgpu_reset_prepare_hwcontext(adev, reset_context); + if (r != -ENOSYS) + return r; + /* Don't suspend on bare metal if we are not going to HW reset the ASIC */ if (!amdgpu_sriov_vf(adev)) { @@ -4286,22 +4300,36 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if (need_full_reset) r = amdgpu_device_ip_suspend(adev); - - *need_full_reset_arg = need_full_reset; + if (need_full_reset) + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + else + clear_bit(AMDGPU_NEED_FULL_RESET, + &reset_context->flags); } return r; } -int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, - struct list_head *device_list_handle, - bool *need_full_reset_arg, - bool skip_hw_reset) +int amdgpu_do_asic_reset(struct list_head *device_list_handle, + struct amdgpu_reset_context *reset_context) { struct amdgpu_device *tmp_adev = NULL; - bool need_full_reset = *need_full_reset_arg, vram_lost = false; + bool need_full_reset, skip_hw_reset, vram_lost = false; int r = 0; + /* Try reset handler method first */ + tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device, + reset_list); + r = amdgpu_reset_perform_reset(tmp_adev, reset_context); + + if (r != -ENOSYS) + return r; + + /* Reset handler not implemented, use the default method */ + need_full_reset = + test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + skip_hw_reset = test_bit(AMDGPU_SKIP_HW_RESET, &reset_context->flags); + /* * ASIC reset has to be done on all XGMI hive nodes ASAP * to allow proper links negotiation in FW (within 1 sec) @@ -4385,7 +4413,8 @@ int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, */ amdgpu_register_gpu_instance(tmp_adev); - if (!hive && tmp_adev->gmc.xgmi.num_physical_nodes > 1) + if (!reset_context->hive && + tmp_adev->gmc.xgmi.num_physical_nodes > 1) amdgpu_xgmi_add_device(tmp_adev); r = amdgpu_device_ip_late_init(tmp_adev); @@ -4413,8 +4442,10 @@ int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, } /* Update PSP FW topology after reset */ - if (hive && tmp_adev->gmc.xgmi.num_physical_nodes > 1) - r = amdgpu_xgmi_update_topology(hive, tmp_adev); + if (reset_context->hive && + tmp_adev->gmc.xgmi.num_physical_nodes > 1) + r = amdgpu_xgmi_update_topology( + reset_context->hive, tmp_adev); } } @@ -4438,7 +4469,10 @@ out: } end: - *need_full_reset_arg = need_full_reset; + if (need_full_reset) + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + else + clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); return r; } @@ -4575,10 +4609,9 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev) return 0; } -void amdgpu_device_recheck_guilty_jobs(struct amdgpu_device *adev, - struct amdgpu_hive_info *hive, - struct list_head *device_list_handle, - bool *need_full_reset) +void amdgpu_device_recheck_guilty_jobs( + struct amdgpu_device *adev, struct list_head *device_list_handle, + struct amdgpu_reset_context *reset_context) { int i, r = 0; @@ -4614,8 +4647,10 @@ retry: if (r) adev->asic_reset_res = r; } else { - r = amdgpu_do_asic_reset(hive, device_list_handle, - need_full_reset, false); + clear_bit(AMDGPU_SKIP_HW_RESET, + &reset_context->flags); + r = amdgpu_do_asic_reset(device_list_handle, + reset_context); if (r && r == -EAGAIN) goto retry; } @@ -4657,7 +4692,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, struct amdgpu_job *job) { struct list_head device_list, *device_list_handle = NULL; - bool need_full_reset = false; bool job_signaled = false; struct amdgpu_hive_info *hive = NULL; struct amdgpu_device *tmp_adev = NULL; @@ -4665,6 +4699,9 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, bool need_emergency_restart = false; bool audio_suspended = false; int tmp_vram_lost_counter; + struct amdgpu_reset_context reset_context; + + memset(&reset_context, 0, sizeof(reset_context)); /* * Special case: RAS triggered and full reset isn't supported @@ -4705,6 +4742,12 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, mutex_lock(&hive->hive_lock); } + reset_context.method = AMD_RESET_METHOD_NONE; + reset_context.reset_req_dev = adev; + reset_context.job = job; + reset_context.hive = hive; + clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + /* * lock the device before we try to operate the linked list * if didn't get the device lock, don't touch the linked list since @@ -4805,9 +4848,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, retry: /* Rest of adevs pre asic reset from XGMI hive. */ list_for_each_entry(tmp_adev, device_list_handle, reset_list) { - r = amdgpu_device_pre_asic_reset(tmp_adev, - (tmp_adev == adev) ? job : NULL, - &need_full_reset); + r = amdgpu_device_pre_asic_reset(tmp_adev, &reset_context); /*TODO Should we stop ?*/ if (r) { dev_err(tmp_adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ", @@ -4824,7 +4865,7 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ if (r) adev->asic_reset_res = r; } else { - r = amdgpu_do_asic_reset(hive, device_list_handle, &need_full_reset, false); + r = amdgpu_do_asic_reset(device_list_handle, &reset_context); if (r && r == -EAGAIN) goto retry; } @@ -4843,8 +4884,8 @@ skip_hw_reset: */ if (amdgpu_gpu_recovery == 2 && !(tmp_vram_lost_counter < atomic_read(&adev->vram_lost_counter))) - amdgpu_device_recheck_guilty_jobs(tmp_adev, hive, - device_list_handle, &need_full_reset); + amdgpu_device_recheck_guilty_jobs( + tmp_adev, device_list_handle, &reset_context); for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = tmp_adev->rings[i]; @@ -5189,12 +5230,14 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) struct drm_device *dev = pci_get_drvdata(pdev); struct amdgpu_device *adev = drm_to_adev(dev); int r, i; - bool need_full_reset = true; + struct amdgpu_reset_context reset_context; u32 memsize; struct list_head device_list; DRM_INFO("PCI error: slot reset callback!!\n"); + memset(&reset_context, 0, sizeof(reset_context)); + INIT_LIST_HEAD(&device_list); list_add_tail(&adev->reset_list, &device_list); @@ -5217,13 +5260,18 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) goto out; } + reset_context.method = AMD_RESET_METHOD_NONE; + reset_context.reset_req_dev = adev; + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags); + adev->in_pci_err_recovery = true; - r = amdgpu_device_pre_asic_reset(adev, NULL, &need_full_reset); + r = amdgpu_device_pre_asic_reset(adev, &reset_context); adev->in_pci_err_recovery = false; if (r) goto out; - r = amdgpu_do_asic_reset(NULL, &device_list, &need_full_reset, true); + r = amdgpu_do_asic_reset(&device_list, &reset_context); out: if (!r) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 0e10c3958f94..d8f131ed10cb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -47,6 +47,7 @@ #include "amdgpu_ras.h" #include "amdgpu_xgmi.h" +#include "amdgpu_reset.h" /* * KMS wrapper. @@ -1349,7 +1350,9 @@ static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work) struct list_head device_list; struct amdgpu_device *adev; int i, r; - bool need_full_reset = true; + struct amdgpu_reset_context reset_context; + + memset(&reset_context, 0, sizeof(reset_context)); mutex_lock(&mgpu_info.mutex); if (mgpu_info.pending_reset == true) { @@ -1359,9 +1362,14 @@ static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work) mgpu_info.pending_reset = true; mutex_unlock(&mgpu_info.mutex); + /* Use a common context, just need to make sure full reset is done */ + reset_context.method = AMD_RESET_METHOD_NONE; + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + for (i = 0; i < mgpu_info.num_dgpu; i++) { adev = mgpu_info.gpu_ins[i].adev; - r = amdgpu_device_pre_asic_reset(adev, NULL, &need_full_reset); + reset_context.reset_req_dev = adev; + r = amdgpu_device_pre_asic_reset(adev, &reset_context); if (r) { dev_err(adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ", r, adev_to_drm(adev)->unique); @@ -1388,7 +1396,10 @@ static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work) list_for_each_entry(adev, &device_list, reset_list) amdgpu_unregister_gpu_instance(adev); - r = amdgpu_do_asic_reset(NULL, &device_list, &need_full_reset, true); + /* Use a common context, just need to make sure full reset is done */ + set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags); + r = amdgpu_do_asic_reset(&device_list, &reset_context); + if (r) { DRM_ERROR("reinit gpus failure"); return; -- cgit v1.2.3-70-g09d2 From 5d89bb2d2f53b27460f77c89756597dee507cd34 Mon Sep 17 00:00:00 2001 From: Lijo Lazar Date: Tue, 16 Mar 2021 21:14:40 +0800 Subject: drm/amdgpu: Make set PG/CG state functions public Expose PG/CG set states functions for other clients Signed-off-by: Lijo Lazar Reviewed-by: Feifei Xu Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 5 +++++ drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 7 ++++--- 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index ec21a71c74eb..ce13b2c8db90 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1390,6 +1390,11 @@ bool amdgpu_device_load_pci_state(struct pci_dev *pdev); bool amdgpu_device_skip_hw_access(struct amdgpu_device *adev); +int amdgpu_device_set_cg_state(struct amdgpu_device *adev, + enum amd_clockgating_state state); +int amdgpu_device_set_pg_state(struct amdgpu_device *adev, + enum amd_powergating_state state); + #include "amdgpu_object.h" static inline bool amdgpu_is_tmz(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 a51f470631d4..2c66fb144d55 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2362,8 +2362,8 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev) * Returns 0 on success, negative error code on failure. */ -static int amdgpu_device_set_cg_state(struct amdgpu_device *adev, - enum amd_clockgating_state state) +int amdgpu_device_set_cg_state(struct amdgpu_device *adev, + enum amd_clockgating_state state) { int i, j, r; @@ -2398,7 +2398,8 @@ static int amdgpu_device_set_cg_state(struct amdgpu_device *adev, return 0; } -static int amdgpu_device_set_pg_state(struct amdgpu_device *adev, enum amd_powergating_state state) +int amdgpu_device_set_pg_state(struct amdgpu_device *adev, + enum amd_powergating_state state) { int i, j, r; -- cgit v1.2.3-70-g09d2 From f066af882b3755c5cdd2574e860433750c6bce1e Mon Sep 17 00:00:00 2001 From: Jude Shih Date: Thu, 1 Apr 2021 11:10:41 +0800 Subject: drm/amdgpu: add DMUB outbox event IRQ source define/complete/debug flag [Why & How] We use outbox interrupt that allows us to do the AUX via DMUB Therefore, we need to add some irq source related definition in the header files; Signed-off-by: Jude Shih Reviewed-by: Harry Wentland Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 + drivers/gpu/drm/amd/include/ivsrcid/dcn/irqsrcs_dcn_1_0.h | 2 ++ 2 files changed, 3 insertions(+) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu.h') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index ce13b2c8db90..dc3a69296321 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -922,6 +922,7 @@ struct amdgpu_device { struct amdgpu_irq_src pageflip_irq; struct amdgpu_irq_src hpd_irq; struct amdgpu_irq_src dmub_trace_irq; + struct amdgpu_irq_src dmub_outbox_irq; /* rings */ u64 fence_context; diff --git a/drivers/gpu/drm/amd/include/ivsrcid/dcn/irqsrcs_dcn_1_0.h b/drivers/gpu/drm/amd/include/ivsrcid/dcn/irqsrcs_dcn_1_0.h index e2bffcae273a..754170a86ea4 100644 --- a/drivers/gpu/drm/amd/include/ivsrcid/dcn/irqsrcs_dcn_1_0.h +++ b/drivers/gpu/drm/amd/include/ivsrcid/dcn/irqsrcs_dcn_1_0.h @@ -1132,5 +1132,7 @@ #define DCN_1_0__SRCID__DMCUB_OUTBOX_HIGH_PRIORITY_READY_INT 0x68 #define DCN_1_0__CTXID__DMCUB_OUTBOX_HIGH_PRIORITY_READY_INT 6 +#define DCN_1_0__SRCID__DMCUB_OUTBOX_LOW_PRIORITY_READY_INT 0x68 // DMCUB_IHC_outbox1_ready_int IHC_DMCUB_outbox1_ready_int_ack DMCUB_OUTBOX_LOW_PRIORITY_READY_INTERRUPT DISP_INTERRUPT_STATUS_CONTINUE24 Level/Pulse +#define DCN_1_0__CTXID__DMCUB_OUTBOX_LOW_PRIORITY_READY_INT 8 #endif // __IRQSRCS_DCN_1_0_H__ -- cgit v1.2.3-70-g09d2