[11:52] <hverkuil> hi all
[11:57] <hverkuil> Feedback for this RFC patch would be appreciated: https://patchwork.linuxtv.org/patch/65318/
[11:58] <syoung> hi all
[12:00] <hverkuil> syoung: when testing on a Allwinner A10 cubieboard this week I noticed that the kernel log was spammed with "IR event FIFO is full!".
[12:00] <hverkuil> not sure if dev_err is the right error level for that message.
[12:00] <hverkuil> or what the cause is.
[12:01] <hverkuil> Otherwise no updates from me other than that I'm behind with patch review and hope to go through pending patches tomorrow or Monday.
[12:04] <syoung> hverkuil: that's interesting, I've never seen this message being spammed
[12:05] <syoung> Was the device receiving IR or did it start spamming all on its own?
[12:06] <hverkuil> Started right after booting it. I didn't do anything with the IR.
[12:07] <hverkuil> I can debug if you give some pointers.
[12:13] <syoung> hverkuil: thanks!
[12:15] <syoung> So this is the sunxi-cir.c driver. The driver reports IR activity from the interrupt handler, using ir_raw_event_store_with_filter(). That IR is then processed from a separate thread, which is scheduled from ir_raw_event_handle().
[12:17] <syoung> ir_raw_event_store_with_filter() fills a kfifo of size 512 which is read from the thread.
[12:18] <syoung> Considering there is no IR at all, the driver should not be reporting anything, except for a superious read which can happen at startup or broken IR receives which detect sunlight as infrared signal (it should have a carrier so this does not usually happen)
[12:19] <syoung> For some reason the sunxi-cir is calling ir_raw_event_store_with_filter() when it shouldn't, basically.
[12:21] <syoung> This driver was originally written for the a10 and then got support added for later generations. I wouldn't be surprised if later commits broke a10.
[12:22] <syoung> I don't have this hardware (any generation) so I haven't looked at the driver much, I'm afraid.
[12:30] <mchehab> hi all... sorry for being late
[12:30] <mchehab> got sidetracked by a meeting
[12:30] <hverkuil> n/p
[12:30] <mchehab> scheduled at teh very last minute
[12:31] <mchehab> I'm also very busy those days... I should be doing some media related work probably tomorrow...
[12:32] <mchehab> merging the atomisp patches so far to the fixes branch and handling any pending PRs
[12:36] <hverkuil> pinchartl: do you think the cleanup/refactor part of the ti-vpe cal series is ready to be merged? or are you waiting for reviews of v2? I'd like to get all the cleanup/refactor stuff done soonish.
[12:38] <hverkuil> syoung: hmm, I tried to revert some promising patches, but no luck.
[12:42] <hverkuil> I keep getting interrupts. I guess I shouldn't get any interrupts if there is no IR activity?
[13:19] <pinchartl> hverkuil: in my opinion it's ready, but Tomi is on vacation this month
[17:43] <syoung> hverkuil: yes, it shouldn't be generating interrupts
[17:43] <syoung> what model do you have exactly?
[18:30] <hverkuil> The original cubieboard with an A10 SoC (arch/arm/boot/dts/sun4i-a10-cubieboard.dts)