liquidAcid has quit [Quit: Leaving]
zombah has quit [Ping timeout: 245 seconds]
<bgamari> Wizzup, do you have access to the manual?
<bgamari> Could someone check a register bit for me?
zombah has joined #linux-exynos
<Wizzup> bgamari: I don't
<Wizzup> at leastp I am not aware...
zombah has quit [Quit: Leaving]
<afaerber> bgamari, there was a v2 of my patchset but the two DT patches were indeed never applied and tfiga asked me to rebase the pinctrl patches, which I haven't gotten to yet
<afaerber> bgamari, feel free to reply with a Reviewed-by to push things forward :)
indy has quit [Ping timeout: 256 seconds]
indy has joined #linux-exynos
ssvb has quit [Ping timeout: 264 seconds]
liquidAcid has joined #linux-exynos
<bgamari> afaerber, oh?
<bgamari> afaerber, when was this posted?
<afaerber> bgamari, November maybe?
<bgamari> hmm
<bgamari> afaerber, which list?
<afaerber> linux-samsung-soc and others. I can try to rebase and repost tonight
<bgamari> afaerber, you might be interested in https://github.com/bgamari/linux/commits/odroid-3.19
<afaerber> bgamari, have you looked at my branch before embarking on this?
<bgamari> afaerber, I'm afraid I never found your branch
<afaerber> I have USB somewhat working, which I'm using as rootfs and thus can't easily test the posted series alone
<bgamari> hmm
<bgamari> afaerber, thanks!
afaerber_ has joined #linux-exynos
ssvb has joined #linux-exynos
<bgamari> afaerber, indeed there is a great deal of overlap here
<bgamari> oh well
<bgamari> it was a learning experience if nothing else
<afaerber_> there's still lots of things to do ;)
<afaerber_> USB had some problems rebasing with clocks and it changes the LEDs somehow... :/
<afaerber_> bbl
afaerber has quit [Ping timeout: 246 seconds]
afaerber_ is now known as afaerber
<bgamari> afaerber, do you handle the USB3503 secondary clock bit correctly?
<bgamari> afaerber, It doesn't look like it will get set as-is
<liquidAcid> bgamari, that is already upstream
<bgamari> liquidAcid, 3.19?
<liquidAcid> yes
<bgamari> liquidAcid, from my read of the code I don't see how secondary_ref_clk could possibly be set unless you specify a clock in the DT
<liquidAcid> here's the correct one
<liquidAcid> it was added when the odroid-x2/u3/u2 support landed upstream
<liquidAcid> or around that time
<bgamari> liquidAcid, right, but as far as I can tell this won't do the right thing in cases like https://github.com/afaerber/linux/blob/c55845139f5c88d41cd91fe59cf4b144285d6ab1/arch/arm/boot/dts/exynos5410-odroidxu.dts#L69
<bgamari> liquidAcid, where no clock is given
<liquidAcid> no, of course you have to provide the correct info in the dst
<liquidAcid> dts
<bgamari> liquidAcid, well, the binding claims that you should omit clock from the DT if refclk is always available
<bgamari> liquidAcid, which is what afaerber does
<bgamari> liquidAcid, there should be a way to specify refclk-frequency even if no clock is provided
<bgamari> I think
<bgamari> afaerber, liquidAcid, this is how I dealt with it, https://github.com/bgamari/linux/commit/9e00ecc760d49f2351028239b845af1406ede0d5
<liquidAcid> bgamari, i had the same discussion with marek
<liquidAcid> bgamari, i guess you should ask him again about that
<bgamari> done
<bgamari> afaerber, do you have access to documentation for this chip?
<liquidAcid> bgamari, you mean the 5410 user manual?
<bgamari> liquidAcid, yes
<liquidAcid> bgamari, did you post this question recently on the odroid forums?
<bgamari> liquidAcid, nope
<liquidAcid> hmm, maybe someone else did
<bgamari> I don't typically frequent forums
<liquidAcid> anyway, i don't think samsung ever made that one public
<bgamari> yeah, this appears to be the case
<bgamari> I was wondering if anyone in the community has it under NDA or something
<liquidAcid> probably hardkernel themselves
<liquidAcid> i guess afaerber also has it
<liquidAcid> would be nice to have another s5pv210 'leak'
<bgamari> liquidAcid, are you familiar with the samsung clock infrastructure?
<bgamari> CLK_USBD300 is never defined
<liquidAcid> bgamari, not really, i'm mostly working on exynos4412, and i try to stay away from all the clk/regulator/powerdomain confusion
<liquidAcid> lots of unexplained dependencies there, stuff that maybe some single guy from the hw team knows
<bgamari> yep
<afaerber> guys, no I don't have the reference manual, I took the 3.14 based queue from Hardkernel and fiddled around with it until it somewhat worked. that commit above is actually from Hardkernel and the next one is from me, making things work under [WIP] warning
<liquidAcid> afaerber, well, i thought you had used some of your 'suse power' or something :)
<afaerber> I have really no clue about those clocks except for interpreting errors from serial output or dmesg :)
<afaerber> but the pinctrl series I sent I'm pretty confident about
<bgamari> bah
<bgamari> my set gives me,
<bgamari> [ 1.318774] ERROR: could not get clock /soc/usb@12000000:usbdrd30(0)
<bgamari> [ 1.324575] exynos-dwc3 soc:usb@12000000: couldn't get clock
<bgamari> [ 1.330213] exynos-dwc3: probe of soc:usb@12000000 failed with error -22
<bgamari> unless I try to define the clock as,
<bgamari> /*GATE(CLK_USBH20, "usbh20", "aclk200", GATE_IP_FSYS, 18, 0, 0),
<bgamari> GATE(CLK_USBD300, "usbd300", "aclk200", GATE_IP_FSYS, 19, 0, 0),
<bgamari> in which case initialization fails
<afaerber> I have a USB disk attached next to the power supply
<bgamari> I can't figure out how our branches differ
<afaerber> and the lower one next to RJ45 worked okay for reading USB-SD adapters etc.
<bgamari> afaerber, so what fails?
<afaerber> dunno
<bgamari> afaerber, but it doesn't work?
<bgamari> afaerber, do you see any errors similar to mine?
<afaerber> bgamari, not every port, no
<bgamari> ahh
<bgamari> I see
<afaerber> bgamari, I'm not always at that device ;) I said I'd try rebasing tonight, not right now :)
<bgamari> afaerber, did you see my messages about the USB3503
<afaerber> bgamari, received it, thanks
<bgamari> that might help
<bgamari> afaerber, are the HardKernel folks typically responsive to email?
<afaerber> bgamari, yes, my series did get replies and some Signed-off-bys to proceed with it
* afaerber powered up ODROID-XU and started rebase
<bgamari> awesome
<bgamari> thanks!
<bgamari> I'm quite curious to know if you see the sort of clock issues I've been seeing
<afaerber> for now I'm on the first of three sub-branches and seeing the pinctrl conflict that tfiga mentioned
liquidAcid has quit [Quit: Leaving]
si1v3r has quit [Ping timeout: 256 seconds]
si1v3r has joined #linux-exynos
<afaerber> bgamari, btw I notice that you didn't CC the mailing lists on your patch
<bgamari> afaerber, Yes, I should have done so although I didn't really know which list would be appropriate
<bgamari> linux-usb?"?
<afaerber> you can try git config sendemail.cccmd "scripts/get_maintainer.pl --nogit-fallback"
<afaerber> but for the cover letter you still need to add them manually
<bgamari> I was mostly wondering what marek though of it
<bgamari> thought*
<bgamari> afaerber, how is it going?
<bgamari> feel free to steal patches from my tree
<afaerber> push is still ongoing and slow...
<bgamari> I think I've already worked out most of the rebasing in my tree, although who knows how good a job I did given the USB issues I'm seeing
<afaerber> bgamari, my plan is to respin my pinctrl/led series and send out the i2c series, then feel free to submit follow-up series, be it USB or something else
<bgamari> afaerber, have you been able to boot your machine?
<afaerber> bgamari, I barely finished rebasing, I only compile-tested the failing commits, not the final thing yet. nor did I boot it
<bgamari> fair enough, I'll let you work
<afaerber> three branch git git-rebase --onto plus two phone calls ;)
<bgamari> I'm quite curious to see what your clock tree looks like when you get a chance (/sys/kernel/debug/clk/clk_summary
<afaerber> on my current kernel (3.18.0-rc5+) I don't seem to have debugfs compiled in
<bgamari> I think I found my issue
<bgamari> I was using GATE_IP_FSYS instead of GATE_BUS_FSYS0
<afaerber> bgamari, it's not booting here
<bgamari> ouch
<afaerber> [ 1.255150] soc:usb@12000000 supply vdd33 not found, using dummy regulator
<afaerber> [ 1.261239] soc:usb@12000000 supply vdd10 not found, using dummy regulator
<afaerber> [ 1.670032] soc:usb@12400000 supply vdd33 not found, using dummy regulator
<afaerber> [ 1.675511] soc:usb@12400000 supply vdd10 not found, using dummy regulator
<bgamari> yeah, I see that as well
<bgamari> I'm booting but I still have no USB
<bgamari> the peripheral sort of initializes
<afaerber> now it's booted!
<bgamari> cat /sys/class/udc/12000000.dwc3/state says "not attached"
<afaerber> hm, I have neither udc nor usb under /sys/class
<bgamari> thanks
<bgamari> ahh
<bgamari> /sys/kernel/debug/12000000.dwc/mode says OTG
<bgamari> I wonder why
<afaerber> depends on enabled kernel drivers and dr_mode (?) on the .dts
<afaerber> *in the
<bgamari> hmm
zombah has joined #linux-exynos
Wizzup has quit [Ping timeout: 246 seconds]
Wizzup has joined #linux-exynos
zombah has quit [Quit: Leaving]
twitch153 has joined #linux-exynos
<sjoerd> bgamari: If the dr_mode isn't set in the DTS and you build your kernel with dual-role the dwc3 driver defaults to OTG