10:04 <ukleinek> krzk: I got your mail about your pgp-key, thanks for reacting. Might I lure you into OFTC/#sequoia where people are that are more competent than me to fix that?

10:07 <ukleinek> krzk: that would be great input because that feedback probably helps to improve the suggested todo for the affected people

10:56 <arnd> I'm trying to make sense of the pgprot_noncached definition. which uses MT_DEVICE_nGnRnE on arm64, unlike pgprot_device, which is MT_DEVICE_nGnRE as expected

11:01 <arnd> On all non-arm architectures pgprot_noncached is the same as pgprot_device, and uses the same as ioremap(). I think all the portable drivers really want the pgprot_device behavior here, though there are probably some that want MT_NORMAL_NC, which on arm64 is pgprot_writecombine

12:13 <mrutland> arnd: I think that's a historical accident, and I suspect we can make pgprot_noncached() use MT_DEVICE_nGnRE

13:04 <LeSpocky> I was adding a second node to a dt binding example and get a "duplicate label" error. how can I make those two examples independent from another and avoid that error/warning when running dt_binding_check?

13:19 <ukleinek> LeSpocky: gpio/aspeed,ast2400-gpio.yaml has several examples. I would expect that being separated by "- |" doesn't make duplicate label names a problem?

13:21 <LeSpocky> ukleinek: that one has no duplicate labels, it has [none], gpio, gpio0, gpio1 … but I will just omit the label for one node in my case, I think it get's clear enough for the reader

13:30 <krzk> LeSpocky: usually aadvice is: just don't add unused labels

13:31 <krzk> and the example itself must be significantly different, not just by one property, to be worth adding and maintaining

13:32 <LeSpocky> krzk: I think it is, we'll see, thanks for your input :-(

13:32 <LeSpocky> I mean :-)

13:32 <LeSpocky> oops

16:55 <LeSpocky> *sigh*

17:00 <krzk> LeSpocky: so you sent exactly what I asked not to - difference in one property: compatible :/