Re: [PATCH] thermal: gov_power_allocator: Fix incorrect calculation in divvy_up_power

From: Lukasz Luba
Date: Fri Feb 21 2025 - 05:35:43 EST




On 2/20/25 19:45, Rafael J. Wysocki wrote:
On Wed, Feb 19, 2025 at 8:07 AM Yu-Che Cheng <giver@xxxxxxxxxxxx> wrote:

divvy_up_power should use weighted_req_power instead of req_power to
calculate the granted_power. Otherwise, the granted_power may be
unexpected as the denominator total_req_power is weighted sum.

Yes, this is what's happening, to my eyes.

This is a mistake during the previous refactor.

Replace the req_power with weighted_req_power in divvy_up_power
calculation.

Fixes: 912e97c67cc3 ("thermal: gov_power_allocator: Move memory allocation out of throttle()")
Signed-off-by: Yu-Che Cheng <giver@xxxxxxxxxxxx>
---
drivers/thermal/gov_power_allocator.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/thermal/gov_power_allocator.c b/drivers/thermal/gov_power_allocator.c
index 3b644de3292e..3b626db55b2b 100644
--- a/drivers/thermal/gov_power_allocator.c
+++ b/drivers/thermal/gov_power_allocator.c
@@ -370,7 +370,7 @@ static void divvy_up_power(struct power_actor *power, int num_actors,

for (i = 0; i < num_actors; i++) {
struct power_actor *pa = &power[i];
- u64 req_range = (u64)pa->req_power * power_range;
+ u64 req_range = (u64)pa->weighted_req_power * power_range;

pa->granted_power = DIV_ROUND_CLOSEST_ULL(req_range,
total_req_power);

---

And the fix looks good to me.

Lukasz, any concerns?

Good catch! It went through since the test didn't set different weights.

Thanks for the fix!

Reviewed-by: Lukasz Luba <lukasz.luba@xxxxxxx>