Skip to contentSkip to conversation

FAILED: patch "[PATCH] drm/amd/pm: fix torn gpu metrics reads" failed to apply to 6.12-stable tree

PATCH New [email protected] 2 participants 4 messages Aug 5 – Aug 12

Subsystems (4)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 6.12-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.12.y
git checkout FETCH_HEAD
git cherry-pick -x 048f4541b71fb19645fb79d6e62e6e4da23a4035
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026080524-joystick-anagram-33dc@gregkh' --subject-prefix 'PATCH 6.12.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 048f4541b71fb19645fb79d6e62e6e4da23a4035 Mon Sep 17 00:00:00 2001
From: Yang Wang <[email protected]>
Date: Thu, 23 Jul 2026 23:18:33 +0800
Subject: [PATCH] drm/amd/pm: fix torn gpu metrics reads

amdgpu_dpm_get_gpu_metrics() returns a pointer to the shared metrics cache after dropping adev->pm.mutex. The sysfs path then copies from that pointer. Another reader can refresh the cache in place during the copy and return a snapshot containing data from two generations.

Pass caller-provided storage through the DPM interface and copy the metrics while the mutex is held. This keeps the cache pointer private and makes each sysfs read observe one complete sample.

Fixes 25c933b1c4fc ("drm/amd/powerplay: add new sysfs interface for retrieving gpu metrics(V2)")
(cherry picked from commit 862333bb48693ecafcae25af0c9d9ec31015ac77)
drivers/gpu/drm/amd/pm/amdgpu_dpm.c +9 -3
index c787e772b3cc..6d1ad4d5b8f0 100644
@@ -1434,17 +1434,23 @@ int amdgpu_dpm_set_power_profile_mode(struct amdgpu_device *adev,
 	return ret;
 }
 
-int amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void **table)
+ssize_t amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void *buf,
+				   size_t size)
 {
 	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
-	int ret = 0;
+	void *table;
+	ssize_t ret;
 
 	if (!pp_funcs->get_gpu_metrics)
 		return 0;
 
 	mutex_lock(&adev->pm.mutex);
 	ret = pp_funcs->get_gpu_metrics(adev->powerplay.pp_handle,
-					table);
+					&table);
+	if (ret > 0) {
+		ret = min_t(ssize_t, ret, size);
+		memcpy(buf, table, ret);
+	}
 	mutex_unlock(&adev->pm.mutex);
 
 	return ret;
drivers/gpu/drm/amd/pm/amdgpu_pm.c +1 -7
index d94ea54c33c3..763947fbf16f 100644
@@ -1774,7 +1774,6 @@ static ssize_t amdgpu_get_gpu_metrics(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	void *gpu_metrics;
 	ssize_t size = 0;
 	int ret;
 
@@ -1782,15 +1781,10 @@ static ssize_t amdgpu_get_gpu_metrics(struct device *dev,
 	if (ret)
 		return ret;
 
-	size = amdgpu_dpm_get_gpu_metrics(adev, &gpu_metrics);
+	size = amdgpu_dpm_get_gpu_metrics(adev, buf, PAGE_SIZE - 1);
 	if (size <= 0)
 		goto out;
 
-	if (size >= PAGE_SIZE)
-		size = PAGE_SIZE - 1;
-
-	memcpy(buf, gpu_metrics, size);
-
 out:
 	amdgpu_pm_put_access(adev);
 
drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h +2 -1
index e95cd22a31cf..8d1b097a3c3c 100644
@@ -518,7 +518,8 @@ int amdgpu_dpm_get_power_profile_mode(struct amdgpu_device *adev,
 				      char *buf);
 int amdgpu_dpm_set_power_profile_mode(struct amdgpu_device *adev,
 				      long *input, uint32_t size);
-int amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void **table);
+ssize_t amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void *buf,
+				   size_t size);
 ssize_t amdgpu_dpm_get_xcp_metrics(struct amdgpu_device *adev, int xcp_id,
 				   void *table);
 ssize_t amdgpu_dpm_get_temp_metrics(struct amdgpu_device *adev,
maintainer
#1 lore raw

PATCH 1/3 drm/amd/pm: Use same metric table for APU

From: Asad Kamal <[email protected]>

[ Upstream commit 56cbb36696407a27cfbfdbc2dee0b1e206178cec ]

Use same metric table for APU and Non APU systems for smu_v_13_0_6 to get metric data based on newer pmfw versions

v2: Use inline func to check for unified metrics support

Stable-dep-of: 048f4541b71f ("drm/amd/pm: fix torn gpu metrics reads")

---
 .../drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c  | 102 ++++++++++--------
 1 file changed, 55 insertions(+), 47 deletions(-)
drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +55 -47
index 9ac694c4f1f7a..b79a2b66d30d0 100644
@@ -102,6 +102,12 @@ MODULE_FIRMWARE("amdgpu/smu_13_0_14.bin");
 #define MCA_BANK_IPID(_ip, _hwid, _type) \
 	[AMDGPU_MCA_IP_##_ip] = { .hwid = _hwid, .mcatype = _type, }
 
+static inline bool smu_v13_0_6_is_unified_metrics(struct smu_context *smu)
+{
+	return (smu->adev->flags & AMD_IS_APU) &&
+		smu->smc_fw_version <= 0x4556900;
+}
+
 struct mca_bank_ipid {
 	enum amdgpu_mca_ip ip;
 	uint16_t hwid;
@@ -253,7 +259,7 @@ struct PPTable_t {
 #define SMUQ10_TO_UINT(x) ((x) >> 10)
 #define SMUQ10_FRAC(x) ((x) & 0x3ff)
 #define SMUQ10_ROUND(x) ((SMUQ10_TO_UINT(x)) + ((SMUQ10_FRAC(x)) >= 0x200))
-#define GET_METRIC_FIELD(field) ((adev->flags & AMD_IS_APU) ?\
+#define GET_METRIC_FIELD(field, flag) ((flag) ?\
 		(metrics_a->field) : (metrics_x->field))
 
 struct smu_v13_0_6_dpm_map {
@@ -584,7 +590,7 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu)
 	MetricsTableA_t *metrics_a = (MetricsTableA_t *)smu_table->metrics_table;
 	struct PPTable_t *pptable =
 		(struct PPTable_t *)smu_table->driver_pptable;
-	struct amdgpu_device *adev = smu->adev;
+	bool flag = smu_v13_0_6_is_unified_metrics(smu);
 	int ret, i, retry = 100;
 	uint32_t table_version;
 
@@ -596,7 +602,7 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu)
 				return ret;
 
 			/* Ensure that metrics have been updated */
-			if (GET_METRIC_FIELD(AccumulationCounter))
+			if (GET_METRIC_FIELD(AccumulationCounter, flag))
 				break;
 
 			usleep_range(1000, 1100);
@@ -613,29 +619,29 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu)
 			table_version;
 
 		pptable->MaxSocketPowerLimit =
-			SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketPowerLimit));
+			SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketPowerLimit, flag));
 		pptable->MaxGfxclkFrequency =
-			SMUQ10_ROUND(GET_METRIC_FIELD(MaxGfxclkFrequency));
+			SMUQ10_ROUND(GET_METRIC_FIELD(MaxGfxclkFrequency, flag));
 		pptable->MinGfxclkFrequency =
-			SMUQ10_ROUND(GET_METRIC_FIELD(MinGfxclkFrequency));
+			SMUQ10_ROUND(GET_METRIC_FIELD(MinGfxclkFrequency, flag));
 
 		for (i = 0; i < 4; ++i) {
 			pptable->FclkFrequencyTable[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(FclkFrequencyTable)[i]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(FclkFrequencyTable, flag)[i]);
 			pptable->UclkFrequencyTable[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequencyTable)[i]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequencyTable, flag)[i]);
 			pptable->SocclkFrequencyTable[i] = SMUQ10_ROUND(
-				GET_METRIC_FIELD(SocclkFrequencyTable)[i]);
+				GET_METRIC_FIELD(SocclkFrequencyTable, flag)[i]);
 			pptable->VclkFrequencyTable[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequencyTable)[i]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequencyTable, flag)[i]);
 			pptable->DclkFrequencyTable[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequencyTable)[i]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequencyTable, flag)[i]);
 			pptable->LclkFrequencyTable[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(LclkFrequencyTable)[i]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(LclkFrequencyTable, flag)[i]);
 		}
 
 		/* use AID0 serial number by default */
-		pptable->PublicSerialNumber_AID = GET_METRIC_FIELD(PublicSerialNumber_AID)[0];
+		pptable->PublicSerialNumber_AID = GET_METRIC_FIELD(PublicSerialNumber_AID, flag)[0];
 
 		pptable->Init = true;
 	}
@@ -958,6 +964,7 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu,
 	struct smu_table_context *smu_table = &smu->smu_table;
 	MetricsTableX_t *metrics_x = (MetricsTableX_t *)smu_table->metrics_table;
 	MetricsTableA_t *metrics_a = (MetricsTableA_t *)smu_table->metrics_table;
+	bool flag = smu_v13_0_6_is_unified_metrics(smu);
 	struct amdgpu_device *adev = smu->adev;
 	int ret = 0;
 	int xcc_id;
@@ -972,50 +979,50 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu,
 	case METRICS_AVERAGE_GFXCLK:
 		if (smu->smc_fw_version >= 0x552F00) {
 			xcc_id = GET_INST(GC, 0);
-			*value = SMUQ10_ROUND(GET_METRIC_FIELD(GfxclkFrequency)[xcc_id]);
+			*value = SMUQ10_ROUND(GET_METRIC_FIELD(GfxclkFrequency, flag)[xcc_id]);
 		} else {
 			*value = 0;
 		}
 		break;
 	case METRICS_CURR_SOCCLK:
 	case METRICS_AVERAGE_SOCCLK:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(SocclkFrequency)[0]);
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(SocclkFrequency, flag)[0]);
 		break;
 	case METRICS_CURR_UCLK:
 	case METRICS_AVERAGE_UCLK:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequency));
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequency, flag));
 		break;
 	case METRICS_CURR_VCLK:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequency)[0]);
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequency, flag)[0]);
 		break;
 	case METRICS_CURR_DCLK:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequency)[0]);
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequency, flag)[0]);
 		break;
 	case METRICS_CURR_FCLK:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(FclkFrequency));
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(FclkFrequency, flag));
 		break;
 	case METRICS_AVERAGE_GFXACTIVITY:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusy));
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusy, flag));
 		break;
 	case METRICS_AVERAGE_MEMACTIVITY:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilization));
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilization, flag));
 		break;
 	case METRICS_CURR_SOCKETPOWER:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(SocketPower)) << 8;
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(SocketPower, flag)) << 8;
 		break;
 	case METRICS_TEMPERATURE_HOTSPOT:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature)) *
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature, flag)) *
 			 SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_TEMPERATURE_MEM:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxHbmTemperature)) *
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxHbmTemperature, flag)) *
 			 SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	/* This is the max of all VRs and not just SOC VR.
 	 * No need to define another data type for the same.
 	 */
 	case METRICS_TEMPERATURE_VRSOC:
-		*value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxVrTemperature)) *
+		*value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxVrTemperature, flag)) *
 			 SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	default:
@@ -2303,6 +2310,7 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 	struct smu_table_context *smu_table = &smu->smu_table;
 	struct gpu_metrics_v1_5 *gpu_metrics =
 		(struct gpu_metrics_v1_5 *)smu_table->gpu_metrics_table;
+	bool flag = smu_v13_0_6_is_unified_metrics(smu);
 	struct amdgpu_device *adev = smu->adev;
 	int ret = 0, xcc_id, inst, i, j;
 	MetricsTableX_t *metrics_x;
@@ -2321,50 +2329,50 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 5);
 
 	gpu_metrics->temperature_hotspot =
-		SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature));
+		SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature, flag));
 	/* Individual HBM stack temperature is not reported */
 	gpu_metrics->temperature_mem =
-		SMUQ10_ROUND(GET_METRIC_FIELD(MaxHbmTemperature));
+		SMUQ10_ROUND(GET_METRIC_FIELD(MaxHbmTemperature, flag));
 	/* Reports max temperature of all voltage rails */
 	gpu_metrics->temperature_vrsoc =
-		SMUQ10_ROUND(GET_METRIC_FIELD(MaxVrTemperature));
+		SMUQ10_ROUND(GET_METRIC_FIELD(MaxVrTemperature, flag));
 
 	gpu_metrics->average_gfx_activity =
-		SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusy));
+		SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusy, flag));
 	gpu_metrics->average_umc_activity =
-		SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilization));
+		SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilization, flag));
 
 	gpu_metrics->curr_socket_power =
-		SMUQ10_ROUND(GET_METRIC_FIELD(SocketPower));
+		SMUQ10_ROUND(GET_METRIC_FIELD(SocketPower, flag));
 	/* Energy counter reported in 15.259uJ (2^-16) units */
-	gpu_metrics->energy_accumulator = GET_METRIC_FIELD(SocketEnergyAcc);
+	gpu_metrics->energy_accumulator = GET_METRIC_FIELD(SocketEnergyAcc, flag);
 
 	for (i = 0; i < MAX_GFX_CLKS; i++) {
 		xcc_id = GET_INST(GC, i);
 		if (xcc_id >= 0)
 			gpu_metrics->current_gfxclk[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(GfxclkFrequency)[xcc_id]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(GfxclkFrequency, flag)[xcc_id]);
 
 		if (i < MAX_CLKS) {
 			gpu_metrics->current_socclk[i] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(SocclkFrequency)[i]);
+				SMUQ10_ROUND(GET_METRIC_FIELD(SocclkFrequency, flag)[i]);
 			inst = GET_INST(VCN, i);
 			if (inst >= 0) {
 				gpu_metrics->current_vclk0[i] =
-					SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequency)[inst]);
+					SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequency, flag)[inst]);
 				gpu_metrics->current_dclk0[i] =
-					SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequency)[inst]);
+					SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequency, flag)[inst]);
 			}
 		}
 	}
 
-	gpu_metrics->current_uclk = SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequency));
+	gpu_metrics->current_uclk = SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequency, flag));
 
 	/* Throttle status is not reported through metrics now */
 	gpu_metrics->throttle_status = 0;
 
 	/* Clock Lock Status. Each bit corresponds to each GFXCLK instance */
-	gpu_metrics->gfxclk_lock_status = GET_METRIC_FIELD(GfxLockXCDMak) >> GET_INST(GC, 0);
+	gpu_metrics->gfxclk_lock_status = GET_METRIC_FIELD(GfxLockXCDMak, flag) >> GET_INST(GC, 0);
 
 	if (!(adev->flags & AMD_IS_APU)) {
 		/*Check smu version, PCIE link speed and width will be reported from pmfw metric
@@ -2405,22 +2413,22 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
 
 	gpu_metrics->gfx_activity_acc =
-		SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusyAcc));
+		SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusyAcc, flag));
 	gpu_metrics->mem_activity_acc =
-		SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilizationAcc));
+		SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilizationAcc, flag));
 
 	for (i = 0; i < NUM_XGMI_LINKS; i++) {
 		gpu_metrics->xgmi_read_data_acc[i] =
-			SMUQ10_ROUND(GET_METRIC_FIELD(XgmiReadDataSizeAcc)[i]);
+			SMUQ10_ROUND(GET_METRIC_FIELD(XgmiReadDataSizeAcc, flag)[i]);
 		gpu_metrics->xgmi_write_data_acc[i] =
-			SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWriteDataSizeAcc)[i]);
+			SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWriteDataSizeAcc, flag)[i]);
 	}
 
 	for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) {
 		inst = GET_INST(JPEG, i);
 		for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) {
 			gpu_metrics->jpeg_activity[(i * adev->jpeg.num_jpeg_rings) + j] =
-				SMUQ10_ROUND(GET_METRIC_FIELD(JpegBusy)
+				SMUQ10_ROUND(GET_METRIC_FIELD(JpegBusy, flag)
 				[(inst * adev->jpeg.num_jpeg_rings) + j]);
 		}
 	}
@@ -2428,13 +2436,13 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 	for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
 		inst = GET_INST(VCN, i);
 		gpu_metrics->vcn_activity[i] =
-			SMUQ10_ROUND(GET_METRIC_FIELD(VcnBusy)[inst]);
+			SMUQ10_ROUND(GET_METRIC_FIELD(VcnBusy, flag)[inst]);
 	}
 
-	gpu_metrics->xgmi_link_width = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWidth));
-	gpu_metrics->xgmi_link_speed = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiBitrate));
+	gpu_metrics->xgmi_link_width = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWidth, flag));
+	gpu_metrics->xgmi_link_speed = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiBitrate, flag));
 
-	gpu_metrics->firmware_timestamp = GET_METRIC_FIELD(Timestamp);
+	gpu_metrics->firmware_timestamp = GET_METRIC_FIELD(Timestamp, flag);
 
 	*table = (void *)gpu_metrics;
 	kfree(metrics_x);
Signature · 1 line
2.53.0
maintainer
#2 lore raw

PATCH 2/3 drm/amd/pm: Use macro to initialize metrics table

From: Lijo Lazar <[email protected]>

[ Upstream commit 1327d8f4061c08c29ea8ce7bb89e209d3c1e8b29 ]

Helps to keep a build time check about usage of right datatype and avoids maintenance as new versions get added.

Stable-dep-of: 048f4541b71f ("drm/amd/pm: fix torn gpu metrics reads")

---
 drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c | 58 --------------------------
 drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h | 14 ++++++-
 2 files changed, 12 insertions(+), 60 deletions(-)
drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c +0 -58
index d2f11d82312f0..c1a41610036dc 100644
@@ -1051,64 +1051,6 @@ int smu_cmn_get_combo_pptable(struct smu_context *smu)
 				    false);
 }
 
-void smu_cmn_init_soft_gpu_metrics(void *table, uint8_t frev, uint8_t crev)
-{
-	struct metrics_table_header *header = (struct metrics_table_header *)table;
-	uint16_t structure_size;
-
-#define METRICS_VERSION(a, b)	((a << 16) | b)
-
-	switch (METRICS_VERSION(frev, crev)) {
-	case METRICS_VERSION(1, 0):
-		structure_size = sizeof(struct gpu_metrics_v1_0);
-		break;
-	case METRICS_VERSION(1, 1):
-		structure_size = sizeof(struct gpu_metrics_v1_1);
-		break;
-	case METRICS_VERSION(1, 2):
-		structure_size = sizeof(struct gpu_metrics_v1_2);
-		break;
-	case METRICS_VERSION(1, 3):
-		structure_size = sizeof(struct gpu_metrics_v1_3);
-		break;
-	case METRICS_VERSION(1, 4):
-		structure_size = sizeof(struct gpu_metrics_v1_4);
-		break;
-	case METRICS_VERSION(1, 5):
-		structure_size = sizeof(struct gpu_metrics_v1_5);
-		break;
-	case METRICS_VERSION(2, 0):
-		structure_size = sizeof(struct gpu_metrics_v2_0);
-		break;
-	case METRICS_VERSION(2, 1):
-		structure_size = sizeof(struct gpu_metrics_v2_1);
-		break;
-	case METRICS_VERSION(2, 2):
-		structure_size = sizeof(struct gpu_metrics_v2_2);
-		break;
-	case METRICS_VERSION(2, 3):
-		structure_size = sizeof(struct gpu_metrics_v2_3);
-		break;
-	case METRICS_VERSION(2, 4):
-		structure_size = sizeof(struct gpu_metrics_v2_4);
-		break;
-	case METRICS_VERSION(3, 0):
-		structure_size = sizeof(struct gpu_metrics_v3_0);
-		break;
-	default:
-		return;
-	}
-
-#undef METRICS_VERSION
-
-	memset(header, 0xFF, structure_size);
-
-	header->format_revision = frev;
-	header->content_revision = crev;
-	header->structure_size = structure_size;
-
-}
-
 int smu_cmn_set_mp1_state(struct smu_context *smu,
 			  enum pp_mp1_state mp1_state)
 {
drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h +12 -2
index a020277dec3e9..ade36a8ffa668 100644
@@ -40,6 +40,18 @@
 #define SMU_IH_INTERRUPT_CONTEXT_ID_FAN_ABNORMAL        0x8
 #define SMU_IH_INTERRUPT_CONTEXT_ID_FAN_RECOVERY        0x9
 
+#define smu_cmn_init_soft_gpu_metrics(ptr, frev, crev)         \
+	do {                                                   \
+		typecheck(struct gpu_metrics_v##frev##_##crev, \
+			  typeof(*(ptr)));                     \
+		struct metrics_table_header *header =          \
+			(struct metrics_table_header *)(ptr);  \
+		memset(header, 0xFF, sizeof(*(ptr)));          \
+		header->format_revision = frev;                \
+		header->content_revision = crev;               \
+		header->structure_size = sizeof(*(ptr));       \
+	} while (0)
+
 extern const int link_speed[];
 
 /* Helper to Convert from PCIE Gen 1/2/3/4/5/6 to 0.1 GT/s speed units */
@@ -125,8 +137,6 @@ int smu_cmn_get_metrics_table(struct smu_context *smu,
 
 int smu_cmn_get_combo_pptable(struct smu_context *smu);
 
-void smu_cmn_init_soft_gpu_metrics(void *table, uint8_t frev, uint8_t crev);
-
 int smu_cmn_set_mp1_state(struct smu_context *smu,
 			  enum pp_mp1_state mp1_state);
 
Signature · 1 line
2.53.0
maintainer
#3 lore raw

PATCH 3/3 drm/amd/pm: fix torn gpu metrics reads

In reply to Sasha Levin

From: Yang Wang <[email protected]>

[ Upstream commit 048f4541b71fb19645fb79d6e62e6e4da23a4035 ]

amdgpu_dpm_get_gpu_metrics() returns a pointer to the shared metrics cache after dropping adev->pm.mutex. The sysfs path then copies from that pointer. Another reader can refresh the cache in place during the copy and return a snapshot containing data from two generations.

Pass caller-provided storage through the DPM interface and copy the metrics while the mutex is held. This keeps the cache pointer private and makes each sysfs read observe one complete sample.

Fixes 25c933b1c4fc ("drm/amd/powerplay: add new sysfs interface for retrieving gpu metrics(V2)")
(cherry picked from commit 862333bb48693ecafcae25af0c9d9ec31015ac77)
---
 drivers/gpu/drm/amd/pm/amdgpu_dpm.c     | 12 +++++++++---
 drivers/gpu/drm/amd/pm/amdgpu_pm.c      |  8 +-------
 drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h |  3 ++-
 3 files changed, 12 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/pm/amdgpu_dpm.c +9 -3
index 9dc82f4d7c937..bab0b2e92b6b4 100644
@@ -1318,17 +1318,23 @@ int amdgpu_dpm_set_power_profile_mode(struct amdgpu_device *adev,
 	return ret;
 }
 
-int amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void **table)
+ssize_t amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void *buf,
+				   size_t size)
 {
 	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
-	int ret = 0;
+	void *table;
+	ssize_t ret;
 
 	if (!pp_funcs->get_gpu_metrics)
 		return 0;
 
 	mutex_lock(&adev->pm.mutex);
 	ret = pp_funcs->get_gpu_metrics(adev->powerplay.pp_handle,
-					table);
+					&table);
+	if (ret > 0) {
+		ret = min_t(ssize_t, ret, size);
+		memcpy(buf, table, ret);
+	}
 	mutex_unlock(&adev->pm.mutex);
 
 	return ret;
drivers/gpu/drm/amd/pm/amdgpu_pm.c +1 -7
index 2252c054c7ddd..af196800196c2 100644
@@ -1889,7 +1889,6 @@ static ssize_t amdgpu_get_gpu_metrics(struct device *dev,
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(ddev);
-	void *gpu_metrics;
 	ssize_t size = 0;
 	int ret;
 
@@ -1904,15 +1903,10 @@ static ssize_t amdgpu_get_gpu_metrics(struct device *dev,
 		return ret;
 	}
 
-	size = amdgpu_dpm_get_gpu_metrics(adev, &gpu_metrics);
+	size = amdgpu_dpm_get_gpu_metrics(adev, buf, PAGE_SIZE - 1);
 	if (size <= 0)
 		goto out;
 
-	if (size >= PAGE_SIZE)
-		size = PAGE_SIZE - 1;
-
-	memcpy(buf, gpu_metrics, size);
-
 out:
 	pm_runtime_mark_last_busy(ddev->dev);
 	pm_runtime_put_autosuspend(ddev->dev);
drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h +2 -1
index f5bf41f21c412..63074a95c4277 100644
@@ -513,7 +513,8 @@ int amdgpu_dpm_get_power_profile_mode(struct amdgpu_device *adev,
 				      char *buf);
 int amdgpu_dpm_set_power_profile_mode(struct amdgpu_device *adev,
 				      long *input, uint32_t size);
-int amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void **table);
+ssize_t amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void *buf,
+				   size_t size);
 
 /**
  * @get_pm_metrics: Get one snapshot of power management metrics from PMFW. The
Signature · 1 line
2.53.0