<kikuchan>
I think paired clock source and prescalar are too complicated for human being to use properly. Once a PWM channel is enabled with some period, the counterpart is restricted to its shared prescalar and clock source, and it looks unpredictable from user land (while period settings are the same, PWM resolution and capable range depend on enabled-order).
<kikuchan>
Changes are welcome, I hope this would be useful. Thank you.
paulk has joined #linux-sunxi
kikuchan has quit [Quit: Page closed]
<Jookia>
kikuchan: subtle bug: entire is -1 of the real cycles. but there's already a driver for this
apritzel has quit [Ping timeout: 480 seconds]
<Jookia>
i'm also working on a PWM driver for that hardware u-boot
<Jookia>
we're all doing PWM drivers it seems
<Jookia>
that is a nice looking driver though
wasutton3 has quit [Ping timeout: 480 seconds]
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
kikuchan has joined #linux-sunxi
<kikuchan>
cool, and thanks for the info. I'll analyze the waveform by scope.
montjoie_ has quit [Ping timeout: 480 seconds]
<Jookia>
my driver sets channels in pairs so it is deterministic
<Jookia>
i haven't written the code to set registers yet, only calculate timings
<kikuchan>
thank you!
<Jookia>
it works just by calculating channel 0's ideal clock, then channel 1's with channel 0's ideal common divisor, then calculates channel 0 again with channel 1's divisor
<Jookia>
the driver i'm writing in u-boot stores a list of the channels with their desired settings
<Jookia>
so i plan to set both channels at once when there's an update
<Jookia>
that should be deterministic
<kikuchan>
ah I see
<Jookia>
the proposed PWM driver for the kernel right now is a little hard to understand so if you want to try and make one better now's a good time
<Jookia>
there's also someone here who wants pwm support for the H616 which i think you have covered
KREYREN_oftc has quit [Remote host closed the connection]
<kikuchan>
ok ok. I'll try to understand your code first :)
<kikuchan>
uh btw, do you think it should set both PWM channels at once on Linux driver too?
<kikuchan>
it may be very trivial but I'm afraid of changing period causes some noise in a waveform...
<Jookia>
i'm not sure
<Jookia>
it may not be suitable
<Jookia>
for u-boot i imagine it's fine. but also who knows, maybe my code is overengineered and not u-boot material
<Jookia>
i am having fun writing it though
<kikuchan>
oh ok. thank you
<Jookia>
kikuchan: have you seen the current PWM driver proposed?
<kikuchan>
yes
<kikuchan>
I think it has a flaw for counterparts' resolution and range, that I mentioned above
<acmeplus>
I believe it's just the different display panel and disabled BT/Wifi
hexdump01 has joined #linux-sunxi
hexdump0815 has quit [Ping timeout: 480 seconds]
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
kikuchan has quit [Quit: Page closed]
acmeplus has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
triskit|away has quit [Remote host closed the connection]
dittid[m] has quit [Remote host closed the connection]
pgwipeout[m] has quit [Remote host closed the connection]
JosephWatson[m] has quit [Remote host closed the connection]
Newbyte has quit [Write error: connection closed]
Tooniis[m] has quit [Remote host closed the connection]
aerospace[m] has quit [Write error: connection closed]
KREYREN has quit [Write error: connection closed]
movedon5b2z4xywybidzannet[m] has quit [Write error: connection closed]
GrantM11235[m] has quit [Remote host closed the connection]
acmeplus has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has joined #linux-sunxi
acmeplus has joined #linux-sunxi
ity has quit [Remote host closed the connection]
ity has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
warpme has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
apritzel has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
warpme has quit []
acmeplus has joined #linux-sunxi
apritzel has quit [Ping timeout: 480 seconds]
warpme has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
Robot_ has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
KREYREN has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
bauen1_ has quit [Ping timeout: 480 seconds]
acmeplus has quit [Ping timeout: 480 seconds]
digetx has quit [Remote host closed the connection]
digetx has joined #linux-sunxi
acmeplus has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
apritzel has joined #linux-sunxi
acmeplus has quit [Ping timeout: 480 seconds]
acmeplus has joined #linux-sunxi
<apritzel>
Jookia: from what I read you are indeed over-engineering the U-Boot driver ;-)
<Jookia>
apritzel: hopefully this week i'll have the code working
<apritzel>
I appreciate the effort to write a nice generic driver, but for U-Boot I'd prefer simpler code over covering edge cases we will never need
<apritzel>
for instance: so far every PWM IP was just using the HOSC as the only clock source
<Jookia>
so only hosc support?
<apritzel>
and we survived with that, so can just ignore the APB clock source and the complexity of choosing the right one
<Jookia>
that's fine
<apritzel>
also: I guess for U-Boot we don't need to support the channel pairs independently. Just set up one channel, then just deny requests for the sister channel
<apritzel>
for U-Boot I'd really just look at the use cases, and that's probably just one backlight
<Jookia>
it's not too much extra code to support multiple channel pairs
bauen1 has joined #linux-sunxi
<Jookia>
i'll post the code up when it works
<Jookia>
then we can start arguing over what to cut :D
<Jookia>
but i would rather post a fully working driver somewhere so if someone wants to add features back in the future it's not too hard
<apritzel>
sounds good!
<apritzel>
just was a bit concerned since you guys started to go into intricate details about corner cases
<macromorgan>
so from what I read in some of the existing AXP drivers it looks like a few create an ADC driver then expose ADC channels to things like battery and whatnot, whereas others just read the registers directly in the battery and charger driver. Does that sound about right?
<macromorgan>
Would the ADC approach work best for the AXP717? I'm thinking I can do it similar to the axp20x_adc.c driver
<apritzel>
macromorgan: do you know which approach is older?
<apritzel>
there might have been changes in Linux meanwhile, that now favour one solution over the other
<macromorgan>
I'll look at it
<macromorgan>
I'm trying to wrap my head around what we really need still but I'm thinking it's a fuel gauge of some kind, a charger, and an extcon of some kind
<apritzel>
I don't have any practical experience with that either, so looking forward to your results ;-) but yeah, something like that
bauen1 has quit [Ping timeout: 480 seconds]
vickycq_ has joined #linux-sunxi
vickycq has quit [Ping timeout: 480 seconds]
JohnDoe_71Rus has joined #linux-sunxi
<macromorgan>
looks relatively recent. I think that's what I'll start on first then, getting axp20x_adc.c to support the AXP717
<macromorgan>
once that works I think we can then extend the axp20x_usb_power.c to support it
ftg has quit [Read error: Connection reset by peer]
ftg has joined #linux-sunxi
<jernej>
I think I found a bug in H616 DRAM initialization. trtp parameter is set wrong for all types and one register shouldn't be written directly, but rmw'ed
<jernej>
so if there is anyone who experience occasional memory issues on H616 & co, please ping me
electricworry has quit [Ping timeout: 480 seconds]
<apritzel>
jernej: can you share details? I don't see the memory doubling very often, but maybe I can provoke it with a boot loop
<apritzel>
and would that apply to the H6 as well? I see we see those "doubling issues" there even more often, so I'd think they are related
Robot_ has quit [Ping timeout: 480 seconds]
wasutton- has joined #linux-sunxi
wasutton3 has quit [Read error: Connection reset by peer]
wasutton3 has joined #linux-sunxi
wasutton- has quit [Ping timeout: 480 seconds]
ftg has quit [Read error: Connection reset by peer]