search

andy-shev / linux

Linux kernel source tree
Other
25 stars 11 forks source link

Eds acpi 5.12 rc8 #33

Closed htot closed 3 years ago

htot commented 3 years ago

I rebased eds-acpi on 5.12-rc8 as you always do, added 3 patches from usb-next + 3 of mine dwc3 fixes. And just tested this on edison. Works nicely.

As I wrote here https://github.com/andy-shev/linux/issues/31#issuecomment-825927788 to prevent port resets we disable LPM entirely, which makes some of the patches (BESL, START_TRANSFER) redundant. I added them in case someone wants to enable LPM again, and resolve the port resets in another way.

As Thinh Nguyen wrote on ML:

Nothing obvious stands out as a problem from the dwc3 driver or the controller. I see a (port) reset after 30 seconds of inactivity, which is a typical timeout and recovery mechanism in the upperlayer from host.

We may need to trace the host (PC) end to see why it times out. But for me, it's good enough without LPM as is.

If you would have a look a my commit messages I would be grateful!

htot commented 3 years ago

Similar to extcon, i.e. add a blank line, before your SoB, followed by Fixes tag: Fixes: 066c09593454 ("usb: dwc3: pci: Enable extcon driver for Intel Merrifield")

I think this is for the "usb: dwc3: pci: Enable usb2-gadget-lpm-disable for Intel Merrifield" patch? (GH broke the View changes button, after I rebased. Grr). I don't know which patch to reference here. It's not caused by "usb: dwc3: pci: Enable extcon driver for Intel Merrifield". LPM in dwc3 causes the problem, but I don't know if that can be traced to a specific commit.

htot commented 3 years ago

Sent to ml, thanks for reviewing.