SW workaround for i.MX8MQ e11231: USB: Clock must remain on during suspend/resume

[RayW8d]

Hi,

I want to let an LTE USB device wake up the system running on a custom i.MX8MQ board. I have searched the kernel source tree (lf-5.10.y) but failed (maybe missed) to find the patch that implements the SW workaround for ERR11231: USB: Clock must remain on during suspend/resume.

I'd like to ask where I could find that patch and is it a stable solution?

BR

Ray

[RayW8d]

Hi Igor,

I have applied the kernel patch (slightly modified as it appears to be designed for the older kernel) to the newer linux-imx source trees of both imx_5.4.3_2.00 and lf-5.10.y branches and applied the ATF patch to the imx-atf source tree of imx_5.4.3_2.0.0 (instead of any newer ones as there is some unfriendly code change for that patch) branch but have no luck yet.

What I have done is to use i.MX BSP of rev. LF5.10.35_2.0.0 as a base, switch the kernel/ATF sources and apply the patches accordingly.

I am afraid if there is anything wrong with my approach? And I'd like to ask if these patches have been reported once applied successfully to any imx8mq boards using newer i.MX BSPs?

BR,
Ray

[igorpadykov · Accepted solution]

Hi Ray

patches were sent via mail.

Best regards
igor

[martinkepplinge]

Hi ignor or anyone knowing the details here,

Does your answer mean that patches were sent to upstream linux? Can you link to them or even the commit ID if they got merged into the mainline? I can't find where this erratum has been addressed... thank you very much.

martin

[marcelziswiler]

That is indeed very generous towards anybody else looking for such a solution in this very forum. Could NXP maybe publish more information about such a solution publicly? Thanks!