drm/amdgpu: Adjust removal control flow for smu v13_0_2

Adjust removal control flow for smu v13_0_2:
During amdgpu uninstallation, when removing the first
device, the kernel needs to first send a mode1reset message
to all gpu devices. Otherwise, smu initialization will fail
the next time amdgpu is installed.

V2:
1. Update commit comments.
2. Remove the global variable amdgpu_device_remove_cnt
and add a variable to the structure amdgpu_hive_info.
3. Use hive to detect the first removed device instead of
a global variable.

V3:
1. Update commit comments.
2. Split a patch into multiple patches.
3. The current patch does:
a. Add a work mode of AMDGPU_RESET_FOR_DEVICE_REMOVE into
the existing gpu recover path, which make all devices
in hive list only have HW reset but no resume (except
the base IP).
b. Call AMDGPU_RESET_FOR_DEVICE_REMOVE and
AMDGPU_NEED_FULL_RESET mode of amdgpu_device_gpu_recover
in amdgpu_pci_remove when removing the first device in
hive list.
c. When removing the first device, the IP blocks keyword
function call sequence is as follows:
.suspend->mode1reset->.resume(basic ip)->.hw_fini->.early_fini->.sw_fini.
^ |
|-<----------<---------<----|
The first three sequences are because of a call to
amdgpu_device_gpu_recover. The three sequences will be
executed in a loop until all devices in the hive list
are iterated.
The sequences starting from .hw_fini only apply to the
first device. Since .suspend has been called before,
except the resumed phase1 basic ip blocks, all other ip
blocks .hw_fini of current device will do nothing.
d. When removing other devices, the calling sequences is the
same as legacy:
.hw_fini -> .early_fini -> .sw_fini.
Since .suspend has been called when removing the first device,
except the resumed phase1 basic ip blocks, all of other ip
blocks .hw_fini of current device will do nothing.

Signed-off-by: YiPeng Chai <[e-mail]>
Reviewed-by: Hawking Zhang <[e-mail]>
Signed-off-by: Alex Deucher <[e-mail]>

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

···
 	struct amdgpu_device *tmp_adev = NULL;
 	bool need_full_reset, skip_hw_reset, vram_lost = false;
 	int r = 0;
+	bool gpu_reset_for_dev_remove = 0;
 
 	/* Try reset handler method first */
 	tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device,
···
 	need_full_reset =
 		test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 	skip_hw_reset = test_bit(AMDGPU_SKIP_HW_RESET, &reset_context->flags);
+
+	gpu_reset_for_dev_remove =
+		test_bit(AMDGPU_RESET_FOR_DEVICE_REMOVE, &reset_context->flags) &&
+			test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 
 	/*
 	 * ASIC reset has to be done on all XGMI hive nodes ASAP
···
 		}
 
 		amdgpu_ras_intr_cleared();
+	}
+
+	/* Since the mode1 reset affects base ip blocks, the
+	 * phase1 ip blocks need to be resumed. Otherwise there
+	 * will be a BIOS signature error and the psp bootloader
+	 * can't load kdb on the next amdgpu install.
+	 */
+	if (gpu_reset_for_dev_remove) {
+		list_for_each_entry(tmp_adev, device_list_handle, reset_list)
+			amdgpu_device_ip_resume_phase1(tmp_adev);
+
+		goto end;
 	}
 
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
···
 	bool need_emergency_restart = false;
 	bool audio_suspended = false;
 	int tmp_vram_lost_counter;
+	bool gpu_reset_for_dev_remove = false;
+
+	gpu_reset_for_dev_remove =
+			test_bit(AMDGPU_RESET_FOR_DEVICE_REMOVE, &reset_context->flags) &&
+				test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 
 	/*
 	 * Special case: RAS triggered and full reset isn't supported
···
 
 retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+		if (gpu_reset_for_dev_remove) {
+			/* Workaroud for ASICs need to disable SMC first */
+			amdgpu_device_smu_fini_early(tmp_adev);
+		}
 		r = amdgpu_device_pre_asic_reset(tmp_adev, reset_context);
 		/*TODO Should we stop ?*/
 		if (r) {
···
 			adev->asic_reset_res = 0;
 			goto retry;
 		}
+
+		if (!r && gpu_reset_for_dev_remove)
+			goto recover_end;
 	}
 
 skip_hw_reset:
···
 		amdgpu_device_unset_mp1_state(tmp_adev);
 	}
 
+recover_end:
 	tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device,
 					    reset_list);
 	amdgpu_device_unlock_reset_domain(tmp_adev->reset_domain);

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

···
 		pm_runtime_forbid(dev->dev);
 	}
 
+	if (adev->ip_versions[MP1_HWIP][0] == IP_VERSION(13, 0, 2)) {
+		bool need_to_reset_gpu = false;
+
+		if (adev->gmc.xgmi.num_physical_nodes > 1) {
+			struct amdgpu_hive_info *hive;
+
+			hive = amdgpu_get_xgmi_hive(adev);
+			if (hive->device_remove_count == 0)
+				need_to_reset_gpu = true;
+			hive->device_remove_count++;
+			amdgpu_put_xgmi_hive(hive);
+		} else {
+			need_to_reset_gpu = true;
+		}
+
+		/* Workaround for ASICs need to reset SMU.
+		 * Called only when the first device is removed.
+		 */
+		if (need_to_reset_gpu) {
+			struct amdgpu_reset_context reset_context;
+
+			memset(&reset_context, 0, sizeof(reset_context));
+			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_RESET_FOR_DEVICE_REMOVE, &reset_context.flags);
+			amdgpu_device_gpu_recover(adev, NULL, &reset_context);
+		}
+	}
+
 	amdgpu_driver_unload_kms(dev);
 
 	drm_dev_unplug(dev);

drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h

···
 	AMDGPU_NEED_FULL_RESET = 0,
 	AMDGPU_SKIP_HW_RESET = 1,
 	AMDGPU_SKIP_MODE2_RESET = 2,
+	AMDGPU_RESET_FOR_DEVICE_REMOVE = 3,
 };
 
 struct amdgpu_reset_context {

drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h

···
 	} pstate;
 
 	struct amdgpu_reset_domain *reset_domain;
+	uint32_t device_remove_count;
 };
 
 struct amdgpu_pcs_ras_field {