Skip to content
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

Jump to bottom

Temperature reporting fix for AMD Zen CPUs #1176

Draft
wants to merge 8 commits into
base: main
Choose a base branch
from
Draft

Temperature reporting fix for AMD Zen CPUs #1176

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
81fb542
Temperature update
JohnsonGlenT Jan 23, 2023
95b42ac
Update AMD Zen detection for temperature reporting
JohnsonGlenT Jan 24, 2023
72f044f
updated to use actual core values, not average
JohnsonGlenT Jan 24, 2023
9f7e8f2
added debug line and refactored
JohnsonGlenT Jan 27, 2023
ab066ed
fixed typo
JohnsonGlenT Jan 27, 2023
8fe2754
AMD Zen CPU temperature k10temp reporting fix
JohnsonGlenT Jan 27, 2023
f8ee8ba
Updated to fix issues for other platforms
JohnsonGlenT Jan 28, 2023
df319c2
Fixed static analyzers issue and added fallback
JohnsonGlenT Feb 4, 2023
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
52 changes: 44 additions & 8 deletions linux/LibSensors.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -243,24 +243,60 @@ void LibSensors_getCPUTemperatures(CPUData* cpus, unsigned int existingCPUs, uns
/* Check for further adjustments */
}

/* Only temperature for core 0, maybe Ryzen - copy to all other cores */
if (coreTempCount == 1 && !isnan(data[1])) {
for (unsigned int i = 2; i <= existingCPUs; i++)
data[i] = data[1];
/* Half the temperatures, probably HT/SMT - copy to second half */
const unsigned int delta = activeCPUs / 2;
if (coreTempCount == delta) {
memcpy(&data[delta + 1], &data[1], delta * sizeof(*data));

/* No further adjustments */
goto out;
}

/* Half the temperatures, probably HT/SMT - copy to second half */
const unsigned int delta = activeCPUs / 2;
if (coreTempCount == delta) {
memcpy(&data[delta + 1], &data[1], delta * sizeof(*data));
/* Check AMD Zen CPUs packages, and mirror temps across packages */
if (coreTempCount > 0 && coreTempCount != existingCPUs) {
double temp[coreTempCount];
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is left partially uninitialized if data contains NAN values.

unsigned int count = 0;
unsigned int coresInDie = existingCPUs / coreTempCount;

/* Find package temps */
temp[0] = NAN;
for (unsigned int i = 1; i <= existingCPUs; i++) {
if (!isnan(data[i])) {
temp[count] = data[i];
count++;
if (count == coreTempCount) {
break;
}
}
}

/* Returns to conditional checking if no temp is found */
if (!isnan(temp[0])) {
// Set Temperature
data[0] = temp[0];
for (unsigned int die = 0; die < coreTempCount; die++) {
for (unsigned int i = 1; i <= coresInDie; i++) {
data[(die * coresInDie) + i] = temp[die];
}
}

/* No further adjustments */
goto out;
Comment on lines +282 to +284
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This should be one level up, because the special-casing should end here regardless of a global temperature found or not.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think that the goto statement should stay here and a fallback conditional should be created for the case where coreTempCount is 1 and data[0] has the only value in it.

This check can be added into the current conditional statement, however, from looking at all the other conditional statements, we should have that moved out into its own check so that it remains a fallback and is not coupled with the section specified for AMD Zen CPU temperature reporting. (change idea added below)

I may be needlessly worried about a impossible or niche hardware/platform config.

Suggested change
/* No further adjustments */
goto out;
/* No further adjustments */
goto out;
}
/* AMD Zen CPU can report coreTempCount as 1 - use fallback conditional */
}
/* Fallback for single temperature count reporting */
if (coreTempCount == 1 && !isnan(data[0])) {
for (unsigned int i = 1; i <= existingCPUs; i++) {
data[i] = data[0];
/* No further adjustments */
goto out;
}

}
/* AMD Zen CPU can report coreTempCount as 1 - returns if temps are not reported as expected */
}

/* Fallback for single temperature count reporting */
if (coreTempCount == 1 && !isnan(data[0])) {
for (unsigned int i = 1; i <= existingCPUs; i++) {
data[i] = data[0];
}

/* No further adjustments */
goto out;
}


out:
for (unsigned int i = 0; i <= existingCPUs; i++)
cpus[i].temperature = data[i];
Expand Down