Skip to content

interface: prevent divide-by-zero for missing power limit #489

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
SevenCircle9 wants to merge 1 commit into
base: master
Choose a base branch
from
+48 −15
Open

interface: prevent divide-by-zero for missing power limit #489

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 34 additions & 7 deletions src/extract_gpuinfo_amdgpu.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,6 +122,7 @@ struct gpu_info_amdgpu {
FILE *fanSpeedFILE; // FILE* for this device current fan speed
FILE *PCIeBW; // FILE* for this device PCIe bandwidth over one second
FILE *powerCap; // FILE* for this device power cap
FILE *powerCapDefault;

nvtop_device *amdgpuDevice; // The AMDGPU driver device
nvtop_device *hwmonDevice; // The AMDGPU driver hwmon device
Expand Down Expand Up @@ -241,6 +242,8 @@ static void gpuinfo_amdgpu_shutdown(void) {
fclose(gpu_info->PCIeBW);
if (gpu_info->powerCap)
fclose(gpu_info->powerCap);
if (gpu_info->powerCapDefault)
fclose(gpu_info->powerCapDefault);
nvtop_device_unref(gpu_info->amdgpuDevice);
nvtop_device_unref(gpu_info->hwmonDevice);
_drmFreeVersion(gpu_info->drmVersion);
Expand Down Expand Up @@ -363,8 +366,18 @@ static void initDeviceSysfsPaths(struct gpu_info_amdgpu *gpu_info) {
// Open the power cap file for dynamic info gathering
gpu_info->powerCap = NULL;
int powerCapFD = openat(hwmonFD, "power1_cap", O_RDONLY);
if (powerCapFD) {
gpu_info->powerCap = fdopen(powerCapFD, "r");
if (powerCapFD >= 0) {
gpu_info->powerCap = fdopen(powerCapFD, "r");
if (!gpu_info->powerCap)
close(powerCapFD);
}

gpu_info->powerCapDefault = NULL;
int powerCapDefaultFD = openat(hwmonFD, "power1_cap_default", O_RDONLY);
if (powerCapDefaultFD >= 0) {
gpu_info->powerCapDefault = fdopen(powerCapDefaultFD, "r");
if (!gpu_info->powerCapDefault)
close(powerCapDefaultFD);
}
close(hwmonFD);
}
Expand Down Expand Up @@ -773,14 +786,28 @@ static void gpuinfo_amdgpu_refresh_dynamic_info(struct gpu_info *_gpu_info) {
}
}

unsigned powerCap = 0;
bool havePowerCap = false;

/* First try the current enforced power cap */
if (gpu_info->powerCap) {
// The power cap in microwatts
unsigned powerCap;
int NreadPatterns = rewindAndReadPattern(gpu_info->powerCap, "%u", &powerCap);
if (NreadPatterns == 1) {
SET_GPUINFO_DYNAMIC(dynamic_info, power_draw_max, powerCap / 1000);
if (rewindAndReadPattern(gpu_info->powerCap, "%u", &powerCap) == 1 &&
powerCap > 0) {
havePowerCap = true;
}
}

/* If unavailable or reported as 0, fall back to the default cap */
if (!havePowerCap && gpu_info->powerCapDefault) {
if (rewindAndReadPattern(gpu_info->powerCapDefault, "%u", &powerCap) == 1 &&
powerCap > 0) {
havePowerCap = true;
}
}

if (havePowerCap) {
SET_GPUINFO_DYNAMIC(dynamic_info, power_draw_max, powerCap / 1000);
}
}

static const char drm_amdgpu_pdev_old[] = "pdev";
Expand Down
22 changes: 14 additions & 8 deletions src/interface.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1706,14 +1706,20 @@ void save_current_data_to_ring(struct list_head *devices, struct nvtop_interface
data_val = 100u;
}
break;
case plot_gpu_power_draw_rate:
if (GPUINFO_DYNAMIC_FIELD_VALID(&device->dynamic_info, power_draw) &&
GPUINFO_DYNAMIC_FIELD_VALID(&device->dynamic_info, power_draw_max)) {
data_val = device->dynamic_info.power_draw * 100 / device->dynamic_info.power_draw_max;
if (data_val > 100)
data_val = 100u;
}
break;
case plot_gpu_power_draw_rate:
if (GPUINFO_DYNAMIC_FIELD_VALID(&device->dynamic_info, power_draw) &&
GPUINFO_DYNAMIC_FIELD_VALID(&device->dynamic_info, power_draw_max) &&
device->dynamic_info.power_draw_max > 0) {

data_val = device->dynamic_info.power_draw * 100 /
device->dynamic_info.power_draw_max;

if (data_val > 100)
data_val = 100u;
} else {
data_val = 0;
}
break;
case plot_fan_speed:
if (GPUINFO_DYNAMIC_FIELD_VALID(&device->dynamic_info, fan_speed)) {
data_val = device->dynamic_info.fan_speed;
Expand Down