On Tue, Feb 22, 2022 at 04:19:01PM +0100, Marc Kleine-Budde wrote: > On 22.02.2022 16:03:34, Sean Nyekjaer wrote: > > The TCAN4x5x device have 2 wake pins. One for waking the host controller > > and one for waking the tcan controller it self. > > > > We only have the wake for the tcan controller it self implemented > > right? > > Sort of. The tcan driver contains handling for the device_state_gpio. I > think should be the pin from the tcan to SoC. The handling of this pin > looks fishy. Ah, Okay so state-gpio maps to nWKRQ and wake-gpio maps to WAKE. I don't think one should wire the WAKE pin to the SoC. > > > I have a task that would enable our host to wake-on-can traffic. > > Due to lack of time, I haven't looked deeply into the sleep and wakeup > of the tcan and the wake from the tcan towards the SoC. As said above, > the device_state_gpio handling looks fishy. Probably first remove the > device_state_gpio handling from the driver, add missing power management > to the m_can and tcan driver, then add wakeup support. The task just grew, will see when I get time ... > > Have a look at this series, I haven't reviewed it in detail, but it > looks good at the first glance. > > | https://lore.kernel.org/all/20220111002109.927308-1-phil@xxxxxxxxxxxxxxx Thanks /Sean
