topic Re: i.MX6UL/optee: How is linux getting out of 'idle' in i.MX Processors

i.MX6UL/optee: How is linux getting out of 'idle'

vsiles — Thu, 19 Apr 2018 14:24:27 GMT

(The linux I'm using is from Linaro Security Working Group · GitHub but the same code can be found in Linux's master or NXP repo)

Hi,

I'm running Linux in the normal world along side a secure OS (optee or another secure blob) without the SMP activated, so only 1 core is running. At some point, linux is deciding to put the core into idle, and is running the function imx6sx_enter_wait (see linux/cpuidle-imx6sx.c at optee · linaro-swg/linux · GitHub ). For what I understood and gathered, this function will completely powerdown the core, which will be restarted by some other device, and set to run the v7_cpu_resume function.

My question is to find how does the core (single core, remember) is reset and restored out of IDLE mode ? I tried to use the SNVS RTC interrupt to do so, but it doesn't seem to work.

Could anyone explain to me how is the core restored after an IDLE, in uniprocessor setup ? A link to the relevant code would be great !

Best,

Re: i.MX6UL/optee: How is linux getting out of 'idle'

igorpadykov — Tue, 24 Apr 2018 01:35:30 GMT

Hi Vincent

for code examples one can look on linux low power driver described

in attached Linux Manual Chapter 21 Low-level Power Management (PM) Driver

and power_mode_switch.c example from MX6UL SDK2.2.

Board Support Packages (9)
SDK2.2_iMX6UL_WIN
i.MX6UltraLite Evaluation Kit|NXP

Best regards
igor
-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

Re: i.MX6UL/optee: How is linux getting out of 'idle'

vsiles — Tue, 24 Apr 2018 07:08:57 GMT

Thank you. I was missing the unmasking of interrupt in the GPC...

Now I can resume my CPU "almost" correctly (sometimes, the CPU resets completely instead of being resumed to the address I wrote in the SRC), but it's up to me to fix that :smileyhappy: