libv changed the topic of #linux-sunxi to: Allwinner/sunxi development - Did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait for an answer! - This channel is logged at https://oftc.irclog.whitequark.org/linux-sunxi
Mangy_Dog has quit [Ping timeout: 480 seconds]
montjoie_ has joined #linux-sunxi
montjoie has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
gnarface has quit [Remote host closed the connection]
tuxd3v has quit [Ping timeout: 480 seconds]
gnarface has joined #linux-sunxi
<smaeul> for clock-names, are we preferring "bus" or "apb"/"ahb" now? I seem to remember it is "bus", but I want to make sure before giving bad patch suggestions
pmp-p has quit [Ping timeout: 480 seconds]
pmp-p has joined #linux-sunxi
sunshavi has quit [Remote host closed the connection]
sunshavi has joined #linux-sunxi
Net147 has quit [Ping timeout: 480 seconds]
Net147 has joined #linux-sunxi
Danct12 has joined #linux-sunxi
<jernej> smaeul: it's bus/mod now
apritzel has joined #linux-sunxi
hlauer has joined #linux-sunxi
cmeerw has joined #linux-sunxi
Net147 has quit [Ping timeout: 480 seconds]
Net147 has joined #linux-sunxi
ftg has joined #linux-sunxi
cmeerw has quit [Ping timeout: 480 seconds]
apritzel has quit [Ping timeout: 480 seconds]
prefixcactus has joined #linux-sunxi
apritzel has joined #linux-sunxi
tnovotny has joined #linux-sunxi
jernej_ has joined #linux-sunxi
jernej has quit [Read error: Connection reset by peer]
jernej_ has quit []
jernej has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
throwaway123 has joined #linux-sunxi
choozy has joined #linux-sunxi
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
jernej has joined #linux-sunxi
Turl has quit [Quit: . o O ( why am I quitting? )]
throwaway123 has quit []
Turl has joined #linux-sunxi
choozy has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
tnovotny has quit []
<apritzel> mripard: wens: so what are the chances of some H616 support still making it into 5.14?
<apritzel> with probably removing the RTC for now?
<mripard> apritzel: the PR have been sent last week
<mripard> so it won't make it for 5.14
<apritzel> bummer!
<apritzel> and there is no chance for a 2nd PR?
<mripard> Linus wants every commit in a PR to be in next for two weeks, so arm-soc has the same cut-out
<mripard> for fixes or small stuff, it can be negociated, but not for a series that large
<arnd> mripard: I would probably still consider merging it, since the changes in the series (I just had a look after you mentioning "arm-soc" summoned me ;-)) are either trivial or new files with no regression potential
<arnd> but it's up to you (and olofj, since he's dealing with the 5.14 merges)
<arnd> I guess the phy driver changes aren't quite trivial, which may also be a show-stopper
jernej_ has joined #linux-sunxi
jernej has quit [Ping timeout: 480 seconds]
<mripard> Lee also hasn't reviewed the MFD parts yet
<mripard> davem didn't either for the net part (even though that's not really a huge concern)
<mripard> I'd really prefer to target 5.15 for this
<arnd> makes sense
<arnd> I was only looking at the arch/arm/ changes, which should be uncontroversial in this case, but depend on the drivers also getting merged
<apritzel> yeah, thanks to AW we need small changes in every subsystem when a new SoC appears ...
<apritzel> I am just trying to get this off my back
hlauer has quit [Ping timeout: 480 seconds]
cmeerw has joined #linux-sunxi
<apritzel> mripard: btw: Lee actually took the AXP patch already (in v6), and it's in -next
<apritzel> and we don't need the net part at the moment, that's just for the second EMAC, for which we require AC200 support first anyway
Danct12 has quit [Remote host closed the connection]
Danct12 has joined #linux-sunxi
martinayotte has joined #linux-sunxi
jernej_ has quit []
jernej has joined #linux-sunxi
<mripard> apritzel: generally speaking, it's easier to get series like this merged if they are split
<mripard> some parts could have been merged way earlier but didn't because you were doing everything at once
<mripard> the DTSI with the UART, SPI and MMC have been fairly stable for a while for example
<mripard> but we never could merged it because it also had the USB bits that were controversial
<apritzel> I know, that's why I dropped USB once (in v5, I believe)
<apritzel> but USB is quite useful, especially for those TV boxes
<apritzel> but yeah, originally I anticipated much less changes, so thought we can get some more peripherals supported on the first go
<mripard> sure, I didn't say you should abandon USB support :)
<apritzel> and splitting has the disadvantage of making the dependencies more complicated
<apritzel> I could make a minimal v8: no USB, no RTC, no net, and the AXP patch being merged would reduce it to the .dtsi and .dts patches, I think
<apritzel> which has little danger of regressions, I'd say
<apritzel> but would allow the other parts (RTC, USB, HDMI, ...) to become independent from each other
Daanct12 has joined #linux-sunxi
<jernej> apritzel: I played with 7 segment display on T95
<jernej> fun fact, it's weird I2C protocol
Danct12 has quit [Ping timeout: 480 seconds]
<jernej> 1. no address per se - register address is effectively I2C address
<apritzel> yeah, I found this as well
<apritzel> I think the BSP bitbangs this?
<jernej> 2. it never acknowledges, even if it is successfull
<jernej> yes
<apritzel> and I found there is some framework that supports that chip/protocol
<jernej> but i2c-gpio works just fine, once you start ignoring ACKs
<apritzel> can you do that without hacks?
<jernej> where? I couldn't find any 7 segment led framework, only several RFC series which were all turned down for various reasons
<apritzel> jernej: did you play with this OpenVFD?
<jernej> nope and I don't plan to
<apritzel> Google turned that up when I searched for the chip
<apritzel> it looks like userland?
<apritzel> I haven't time to look into this, just made a note
<jernej> there is some history with LE but it was ripped off from build system
<jernej> no, it's both, kernel and userspace
<jernej> kernel side is just driver with custom interface
<jernej> anyway, it's good for documentation
<apritzel> I see, I thought it would be bitbanging GPIOs, or using I2C (with some tweaks, maybe)
<jernej> driver does bitbanging
<jernej> but that seems redundant if kernel supports that
<apritzel> you mean with i2c-gpio?
<jernej> for example
<jernej> but maybe it's better to reuse just i2c-algo-bit
<jernej> afaik, i2c-gpio would mean to treat as normal i2c bus, which is not really
<apritzel> yeah, I was wondering about that
<apritzel> so I guess you can't tweak that from userland, for instance to ignore ACKs?
<jernej> not sure, I plan to check that
<jernej> if acks depend on userspace, it could be useful just to enable i2c-gpio and let userspace deal with actual device
<jernej> I mean, if userspace can ignore acks
<apritzel> I agree
<apritzel> (as I was also looking for a 7 segment kernel framework, and came to the same conclusions as you)
<jernej> but long term, if 7 seg. display interface ever become a thing, driver could be easily made, it's extremely simple device
<apritzel> you mean like: echo 17:43 > /dev/7seg0 ?
<jernej> for example
prefixcactus has quit [Ping timeout: 480 seconds]
<jernej> but I think only numbers would be covered
<jernej> not all displays have ":" and in my case, it's driven separately
<apritzel> that's why the driver could sort that out (and drop it, if needed)
<jernej> oh, there is I2C_M_IGNORE_NAK flag
<jernej> which must be supported by the HW/driver, but in i2c-gpio case it is
<jernej> apritzel: this snippet works fine, without any kernel hack: https://pastebin.com/raw/bWvhNVEx
<apritzel> jernej: very nice!
ftg has quit [Read error: Connection reset by peer]
<jernej> it's just extremely annoying that register locations act as addresses
<apritzel> and i2c-gpio must be described via DT? Or can you instantiate this on the fly, from userland?
<jernej> i2c-gpio node is of course needed in DT
<jernej> but this is as basic as it gets: https://pastebin.com/raw/j1grYKg2
vagrantc has joined #linux-sunxi
Daanct12 has quit [Quit: Quitting]
Danct12 has joined #linux-sunxi
megi has quit [Quit: WeeChat 3.1]
megi has joined #linux-sunxi
Luke-Jr has quit [Ping timeout: 480 seconds]
hlauer has joined #linux-sunxi
<apritzel> jernej: did you find a comprehensive protocol description for this LED chip? Or is that borrowed from OpenVFD?
<jernej> apritzel: I decompiled original driver
<apritzel> from BSP?
<jernej> no, there is no source
<jernej> I just extracted Image file, convert it to ELF and used ghidra
<apritzel> ouch
<apritzel> you have high pain tolerance
<jernej> well, Linux kernel most of the times contains symbol table
<jernej> in order to load kernel modules
<jernej> but guess what, there are symbols for *all* symbols, not just exported
<jernej> so ghidra can create very comprehensible code
<jernej> anyway, register 0x48 (divide by 2 for I2C address) is control register
<jernej> bit 0 -> power on/off bit1-3 - brigthness (0 - most bright)
Luke-Jr has joined #linux-sunxi
<jernej> and then you have output registers at 0x66 - 0x69
<jernej> bit 0-6 are for each of 7 segments
<jernej> that's all there is
<apritzel> nice
<jernej> you can write output registers in bulk or separately
<apritzel> and you are taking away my fun task I was saving for after the H616 support series was merged ;-)
<jernej> which driver IC is used on Y96?
<jernej> *X96
<apritzel> not sure, but most I saw were using the same
<apritzel> there was one simpler model, though, IIRC
<apritzel> ah, I remember I made sure it's readable on the PCB picture: TW1628
<apritzel> is that the same?
<jernej> no
<jernej> that uses 3-wire SPI :)
<jernej> so you have something to tinker still :)
<apritzel> \o/
<jernej> but I imagine you can use same approach with spi-gpio
<jernej> and spidev
<apritzel> I remember Google turned something up in OpenVFD, so I will just borrow something from there
<jernej> yeah, commands
<apritzel> so is this the T95s, with the H616?
<jernej> and logic analyzer is of great help
<apritzel> I think the T95 is the model with the H6?
<jernej> yes
<jernej> hm... let me check
<jernej> box says only T95, but on PCB it says T95MAX
<apritzel> lol, I have an H6 box which says "T95 max" on the plastic enclosure
<apritzel> and this one does not have a display at all ;-)
cmeerw has quit [Ping timeout: 480 seconds]
hlauer has quit [Ping timeout: 480 seconds]
pmp-p_ has joined #linux-sunxi
pmp-p is now known as Guest2453
pmp-p_ is now known as pmp-p
Guest2453 has quit [Ping timeout: 480 seconds]
Danct12 is now known as melt
melt has quit []
Danct12 has joined #linux-sunxi
vagrantc has quit [Quit: leaving]