On Wed, Aug 21, 2019 at 6:38 AM Tony Lindgren <tony@xxxxxxxxxxx> wrote: > > * Adam Ford <aford173@xxxxxxxxx> [190821 00:06]: > > Tony, > > > > I noticed the AM3517-evm gets a bunch of splat when loading the wlcore > > and 80211 stuff. It appears to be limited to your for-next branch. I > > haven not bisected it, but before I did, I wanted to see if you were > > aware of anything. If not, I'll work on bisecting it. I just didn't > > want to waste time if you had any ideas first. > > Hmm not sure why anything in my for-next would affect am3517? Maybe > something is broken in v5.3-rc1 that my for-next is based on. The problem exists in v5.3-rc1 (not just your for-next), but it appears to be resolved in 5.3-rc5, so I'm going hold back any further testing. I applied the changes in for-next to rc5, and I am still not seeing an issue. > > Also, please test (again) with level wlcore interrupt instead of > edge. We recent gpio-omap fixes it's starting to look that we can > still have edge interrupts lost in retention idle with no status > to be seen in the gpio controller. So anything that can be treated > as level interrupt probaby should be. Currently, it it setup as interrupts = <10 IRQ_TYPE_EDGE_RISING>; /* gpio_170 */ Are you suggesting we go back to interrupts = <10 IRQ_TYPE_LEVEL_HIGH>; /* gpio_170 */ adam > > Regards, > > Tony
