[04:53] <emaczen> what/where are the type definitions of __u8 __u32 etc...
[05:43] <LazyGrizzly> emaczen: types with __ as a prefix are kernel internal typedefs. generally __u8 == u8 == uint8_t try something like grep -rniI "typedef .* __u8;" to get a list of possible definitions
[05:55] <grohne> pinchartl: good question. I cannot use auto exposure, so I need to do exposure in software and for getting good enough results, I need to know what gain stages mean.
[06:04] <LazyGrizzly> pinchartl: can i go forward with SBGGR12SP - Bayer 12-Bit simple packed ?
[10:59] <peetaur2_> can someone help me solve this compile error? https://bpaste.net/show/3211eb5c739c  using kernel 4.14 and CentOS 7 and the git repo here git://linuxtv.org/media_build.git
[11:02] <peetaur2_> and if I hack it up with https://bpaste.net/show/6bb950324766 then I get this instead https://bpaste.net/show/896e5202d4f1
[11:05] <peetaur2_> compat.h has the usual ifndef thing so it shouldn't include twice... so what is going on?  https://bpaste.net/show/7e6e4075200a
[11:07] <grohne> pinchartl: I looked at mt9m032 and the gain control code says "a_gain = (1 + analog_mul) + (analog_gain_val + 1) / 16" but the data sheet says "a_gain = (1 + analog_mul) * (analog_gain_val / 16)"
[11:08] <grohne> pinchartl: but the data sheet contradicts itself when it comes to increments. I guess they really mean 2 ** analogue_mul
[11:17] <peetaur2_> I changed my hack it up patch to https://bpaste.net/show/1d1f2bd96c11  and then it took much longer and a new compile error:  https://bpaste.net/show/aaf83830e046
[11:20] *** LazyGrizzly has left 
[11:35] <peetaur2_> yay it worked... problem was I guess I had to do 'make distclean' (as I did, as they documented), but also you have to remove some generated stuff (my way is rm -rf *; git reset --hard) before it makes sense and uses the right files for your specific kernel
[11:37] <peetaur2_> so thank you all, rubber ducks
[11:37] *** peetaur2_ has left "Konversation terminated!"
[11:48] <pinchartl> grohne: it would make sense. doesn't the upstream mt9m032 driver round the gain ?
[11:54] <grohne> I'm looking ath the upstream driver
[11:59] <pinchartl> grohne: I think it should be a_gain = (1 + analog_mul) * (analog_gain_val + 1) / 16, indeed
[11:59] <pinchartl> and that's what the code implements I believe
[11:59] <pinchartl> but it doesn't expose the rounded gain to userspace
[12:00] <grohne> pinchartl: maybe only the comment is wrong, yes
[13:19] <lucaceresoli> Hi, is there an official way to assign a subdev name? I can't find any docs except "it must be unique".
[13:19] <lucaceresoli> In the imx274 driver: https://elixir.bootlin.com/linux/latest/source/drivers/media/i2c/imx274.c#L1648
[13:19] <lucaceresoli> it is set to DRIVER_NAME, which of course is not unique
[13:21] <lucaceresoli> v4l2_i2c_subdev_init provides a default that is unique, but it looks even better to set it to the devicetree node name, isn't it?
[13:25] <pinchartl> lucaceresoli: there's no global naming scheme that have been agreed on. we have started discussing the topic a few times in the past, unfortunately without reaching a conclusion as nobody seemed interested enough to drive the discussions to an end
[13:31] <lucaceresoli> pinchartl: OK :-)
[13:33] <lucaceresoli> pinchartl: let me put is in another way... would you prefer a patch that sets the I2C name or one that sets the DT name (with fallback on the I2C name)?
[13:37] <pinchartl> you would need to include the full DT path if you use the DT name, as you could have two identical components on two different I2C buses, and that would exceed the length of the name field
[13:37] <pinchartl> the I2C nam is thus likely our best option for now
[14:02] <lucaceresoli> pinchartl: fine, thanks
[15:21] <angelo_ts> mm i have a csi-2 4 lanes interface (imx8), feeding CSI-2 data froma a sensor subdevice. Streaming is not working. Debugging CSI-2 registers, i don't see any sign of received data, while i see the data on the CSI-2 bus. Any help, if you ever had similar issues, is welcome.
[17:27] *** Eloy has quit IRC (Remote host closed the connection)
[17:28] *** prabhakarlad has quit IRC (Ping timeout: 260 seconds)
[18:16] *** ChanServ has quit IRC (*.net *.split)
[18:22] *** weber.freenode.net sets mode: +o ChanServ
[18:25] *** Robert_Zenz has quit IRC (Quit: Robert_Zenz)
[18:54] *** andrey_utkin has quit IRC (Excess Flood)
[19:25] *** bparrot has quit IRC (Remote host closed the connection)
[19:28] *** Muzer has quit IRC (Ping timeout: 276 seconds)
[20:20] *** headless has quit IRC (Quit: Konversation terminated!)
[21:05] *** lyakh has quit IRC (Quit: thanks, bye)
[21:06] *** kbingham has quit IRC (Ping timeout: 240 seconds)
[21:09] *** paul374- has quit IRC (Ping timeout: 260 seconds)
[21:34] *** trumee has quit IRC (Ping timeout: 240 seconds)
[23:03] *** weber.freenode.net sets mode: +o ChanServ