#asahi · 2021-03-26 09:14
marcanyeah, let me double check but I think they are NFC#asahi · 2021-03-26 09:12
marcanI'm guessing this was largely automated then, after he sent those tags even though there were a couple nits left#asahi · 2021-03-26 09:12
marcanyeah, they got merged verbatim, just with Krzysztof's tags#asahi · 2021-03-26 09:10
marcanbut I can re-diff and split#asahi · 2021-03-26 09:10
marcanonly annoying thing now is I already squashed all of that in :-)#asahi · 2021-03-26 09:09
marcanin which case what's left is andy's feedback being too late#asahi · 2021-03-26 09:08
marcankrzysztof reviewed everything but had a couple caveats, but I think they only related to a spurious whitespace change and commit messages; nothing is actionable there any more if they already got merged#asahi · 2021-03-26 09:08
marcanlet me see what gregkh actually merged#asahi · 2021-03-26 09:08
marcanyeah#asahi · 2021-03-26 09:05
marcanyeah, that was my thinking, I can rebase on top of a common point like that#asahi · 2021-03-26 09:03
marcanI'm honestly a bit confused as to how the pipeline with gregkh and krzysztof works here :)#asahi · 2021-03-26 09:01
marcanin that case the changes I have queued for v4 should go in fixup patches I guess#asahi · 2021-03-26 09:00
marcanI was hoping he would just review them, not merge them#asahi · 2021-03-26 09:00
marcanyeah, I noticed that... before they were finally reviewed too, not sure if that will be a problem now since those changed#asahi · 2021-03-26 08:59
marcanbut I think we agree on the approach#asahi · 2021-03-26 08:59
marcanI think the nonposted stuff is also pretty much OK, robher just needs to review my final, simpler version#asahi · 2021-03-26 08:59
marcaner, on v3 (for v4)#asahi · 2021-03-26 08:58
marcanarnd: I think we're pretty much there, I just sent the last feedback email on v4#asahi · 2021-03-26 04:09
marcanthe kernel release schedule has hard deadlines and missing them means months of wait until the next version, so I really want to make sure we make 5.13#asahi · 2021-03-26 04:08
marcanmy priorities right now are v4 first, then that#asahi · 2021-03-26 04:08
marcandon't worry, that's coming soon#asahi · 2021-03-26 04:08
marcanyeah, m1n1 hasn't been updated#asahi · 2021-03-26 02:48
marcanthat's because most of the past month on the Linux branch has been review feedback, not new code :)#asahi · 2021-03-26 02:48
marcanso all the rebasing and squashing I've been doing does not get reflected in those dates#asahi · 2021-03-26 02:48
marcancommit date reflects the date that commit was *created*, not *revised*#asahi · 2021-03-26 02:47
marcanbut the linux branch is much newer than february#asahi · 2021-03-26 02:47
marcanm1n1 is going to get the more recent changes merged in, don't worry#asahi · 2021-03-26 02:47
marcanamw: you are confusing commit date with branch age#asahi · 2021-03-26 00:37
marcanI guess the AIC driver needs to just accept both types then#asahi · 2021-03-26 00:35
marcaneither way, for all intents and purposes that makes it edge triggered to us, implicitly#asahi · 2021-03-26 00:35
marcankettenis: cute, so either it's latched by AIC itself, or by the PCIe core with some kind of feedback of when it is acknowledged#asahi · 2021-03-25 18:52
marcangoing to get some sleep, will check the backlog tomorrow :)#asahi · 2021-03-25 18:50
marcan(if the MSI offset is 32-aligned then all you need to read is a single 32-bit word)#asahi · 2021-03-25 18:49
marcanI guess something like read it unconditionally before the reason, then if it was an MSI, read it again after the reason and actually print both values#asahi · 2021-03-25 18:48
marcanespecially before and after the AIC event read#asahi · 2021-03-25 18:48
marcanso yeah, I'm interested in what HW_STATE says during MSI handling#asahi · 2021-03-25 18:48
marcanmakes sense#asahi · 2021-03-25 18:47
marcanright?#asahi · 2021-03-25 18:46
marcanI mean the MSI data value is just the index#asahi · 2021-03-25 18:43
marcanthat thing (it's a big bitfield, one bit per IRQ indicates level state)#asahi · 2021-03-25 18:42
marcanAIC_HW_STATE = AIC + 0x4200#asahi · 2021-03-25 18:42
marcanand if it's high then goes low at some point#asahi · 2021-03-25 18:42
marcanit might be interesting looking at the hardware state register at various points in the handler, see what level it clames those lines are at#asahi · 2021-03-25 18:41
marcanand yeah, if it all "magically" works then it must be edge-triggered#asahi · 2021-03-25 18:41
marcankettenis: where the MSI value just maps to the index in AIC I presume#asahi · 2021-03-25 18:31
marcanif it's implicit in the ack in AIC (whether internally or via some feedback mechanism), then for all intents and purposes they are edge-triggered#asahi · 2021-03-25 18:30
marcanthe question is what clears the AIC IRQ then#asahi · 2021-03-25 16:26
marcanI think kettenis might know?#asahi · 2021-03-25 16:26
marcanmaybe the PCIe controller does some kind of latching?#asahi · 2021-03-25 16:26
marcanmaz: I haven't seen any register to that effect; I'm not sure how MSIs are supposed to work