On Friday, 20 March 2026 15:45:31 CET Russell King (Oracle) wrote:

On Fri, Mar 20, 2026 at 10:39:10AM +0100, Romain Gantois wrote:

quoted

Originally, I implemented a runtime PM support in the SFP core. This
allowed to cut power to the cages when the attached network interface was
down, thereby saving power.

Have you measured how much power is saved by this?

Sure, here are a few measures of saved power with several different modules 
I've tested. The measures were done with a shunt resistor on the regulator 
line feeding all power pins on both SFP cages. Only one SFP cage was occupied 
during each measure and the upper network interface was down.

Module           | Power saved (uW)
SFP-10G-T-I    | 80
SFP-10G-T-30 | 359
SFP-GB-GE-T  | 191
CHAMPION ONE 1000SFPT | 188
SFP-H10GB-CU1M | 0
ES8512-3LCD05 | 248
SFP100BFXST | 351

So with both SFP cages occupied, we could see about 0.7mW of saved power by 
cutting the regulators to the cages.

Thanks,

-- 
Romain Gantois, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com