<BorgCuba>
I defined a "phy-names = "usb2-phy";" in the dts "usb_otg" node, but that does not solve anything
<BorgCuba>
from rk3xxx.dtsi: usb_otg: usb@10180000 { compatible = "rockchip,rk3066-usb", "snps,dwc2";
<Astralix>
Looks like there is no phy driver for the usb.
GriefNorth has quit [Quit: WeeChat 1.0.1]
<Astralix>
AFAAIK there is a new RK3288 USB PHY driver on the ramp for linux next
Astralix has quit [Remote host closed the connection]
<rperier>
there is no phy usb driver for RK3188/RK3066
<AstralixNB1>
still not?
<rperier>
no, we should patch the one for rk3288 and add a devicetree variant for RK3188 and RK3066, and add support to the dts
<AstralixNB1>
did someone test the rk3288 version?
<rperier>
does not look so complicated, so after have solved a crash in dwc2 I could write this variant in phy-usb-rockchip before switch back to drm-rockchip
BorgCuba has quit [Ping timeout: 250 seconds]
<rperier>
AstralixNB1: don't know, but you could :)
jas-hacks has joined #linux-rockchip
<AstralixNB1>
I have a look to it this night
<rperier>
in this case don't remember to add "Tested-by: your name <your email>" to the corresponding patch on the ML :)
<AstralixNB1>
put it as nr 1 on the list
<rperier>
don't forget *
<AstralixNB1>
no, already have it on the screen
<AstralixNB1>
wasnt the driver rejected because of the cryptic email address?
VargaD has quit [Ping timeout: 250 seconds]
VargaD has joined #linux-rockchip
BorgCuba has joined #linux-rockchip
<jas-hacks>
for RK3288 it is possible to get u-boot to load dtb/kernel from a FAT or EXT partition?
naobsd has joined #linux-rockchip
<AstralixNB1>
jas-hacks: afaik currently not
<AstralixNB1>
Adding all the usual but mussing features to uboot should have hi prios. This would make mainlining and general kernel work much more easy.
<AstralixNB1>
bootp and tftp loading is most important, then to load kernel and dtbs as files from a normal filesystem
<naobsd>
for trying new kernel/dts w/o flashing on rk3288
<naobsd>
try fastboot
<naobsd>
"fastboot boot"
gb_master has joined #linux-rockchip
cnxsoft has joined #linux-rockchip
jas-hacks has left #linux-rockchip [#linux-rockchip]
field^Mop has joined #linux-rockchip
gb_master has quit [Remote host closed the connection]
field^Mop has quit [Ping timeout: 264 seconds]
cnxsoft has quit [Quit: cnxsoft]
<rperier>
mmind00: If I pass console=uart8250,iomem,<phys_addr>,115200n8 as argument to the kernel, would it be possible to logs all kernel logs, including earlyprintk to uart0 ? if the corresponding uart is not initialized by the bootloader , the kernel can do initialization itself ? I can logs some of these logs to ttyS0 but it is done once the console subsystem is initialized, most of the time it is too late
<rperier>
as my uart2 is not really deterministic I would like to use another uart just in case (which is very stable when I tested it yesterday)
<mmind00>
rperier: earlyprintk relys completely on the bootloader initializing the console ... an alternative is earlycon, which might work as a replacement of earlyprintk using the real console driver
<rperier>
I am trying earlycon... a replacement of earlyprintk you said... mhhh
<mmind00>
rperier: the dw_serial seems to support earlycon already, but I haven't looked at this yet at all
<rperier>
I see
<rperier>
I will try
<mmind00>
great :-)
<rperier>
the mmio address would be the physical one right ?
<BorgCuba>
AstralixNB1, so you say another usb phy driver is needed for the otg driver (s3c-hsotg) to work?
<BorgCuba>
the rk3066 has integrated phy as far as I could see (from the ds and a schematic)
<BorgCuba>
an
<BorgCuba>
btw, why is the otg driver called "s3c-hsotg"? reminds me of some samsung soc
<mmind00>
rperier: I'd think so
<mmind00>
BorgCuba: historically, only the gadget-part of the dwc2 was implemented ... as s3c-hsotg ... probably because the first implementer didn't know it was a dw IP [which is normally not communicated in manuals etc]
<mmind00>
BorgCuba: while the phy is integrated into the soc, it is of course not part of the dwc2 IP, so the dwc2 driver of course would need to handle to the phy component
<mmind00>
BorgCuba: and of course the integration of the s3c-hsotg into the dwc2 codebase is still progressing and not finished
<rperier>
with "earlycon=uart8250,mmio32,0X10124000,115200n8" does not seem to work...
<BorgCuba>
mmind00, so you say the otg part (as a device / gadget) is not yet functional?
<mmind00>
BorgCuba: no ... I'm simply not sure what the current state is
<mmind00>
BorgCuba: there is already a thread on dwc2 improvements from just today :-)
<BorgCuba>
thanks! I'll take a look
bengal has joined #linux-rockchip
<mmind00>
BorgCuba: on wednesday there was a patch about having the initial mode as gadget ([PATCH] usb: dwc2: add support for initial state for dual-role mode) ... in general there is quite some activity on the dwc2 :-)
<BorgCuba>
thats nice to hear
<BorgCuba>
mmind00, btw, how many people are working on rockchip mainlining? Are there any drivers nobody is yet working on?
<BorgCuba>
I really like this kernel compared to the rockchip kernels, so far I was able to get things to work without any hacks (only menuconfig and dts mods)
akaizen has quit [Read error: Connection reset by peer]
JohnDoe_71Rus has joined #linux-rockchip
JohnDoe_71Rus has quit [Changing host]
JohnDoe_71Rus has joined #linux-rockchip
<karlp>
BorgCuba: sorry, I don't remember anymore what the sdmmc stuff was, but I know I needed "rootwait" to get the sd card to be detected, there were definitely problems with the power tree stuff, but most of it just got deferred, it's been a few months now though sorry
<karlp>
rperier: you got your uart worked out right? what you said sounded reasonable, but I don't have a rock :)
akaizen has joined #linux-rockchip
<karlp>
dlezcano_: what do you mean by a "broadcast timer" ?
<rperier>
karlp: my uart debug (uart2) works randomly, sometimes it works perfectly fine, like today sometimes I have nothing, sometime I have unreadable output, I think it has to do with my usb ttl
<rperier>
(my usb ttl converters I mean)
<rperier>
so in case I would just like to be able to use another uart, that all :)
<rperier>
I can use the uart once the console subsystem is initiliazed,and in this case using ttyS0 works fine
<rperier>
however, I cannot use earlycon
<karlp>
hrm, yeah, I was kinda curious why yo uwant to swtich to uart1, it will probably behave just like uart2 :)
<rperier>
I'm surprised that the kernel cannot do what a bootloader does before activate paging and MMU... I mean, it has just to setup uart... if the driver itself handles this setup... it would just work...
<rperier>
I am still investigating, except these points my RR pro works fine :)
<BorgCuba>
karlp, the mmc is working (mmind00 was helping me out yesterday). Now I am facing otg issues, I am currently reading the mailing list. There seem to be a lot of patches that may apply here
<mmind00>
BorgCuba: I don't regularly count heads :-) There are of course a big number of Rockchip people contributing drivers but of course mainly for rk3288.
<BorgCuba>
mmind00, yes, the phy patch from the mailing list seems to be for the rk3288 (at least rk3288-evb was mentioned as the test platform)
<BorgCuba>
btw, do I have to copy, paste and apply the diffs from the mailing list if I want to test these patches or is there a more convenient way to do this?
<mmind00>
BorgCuba: regular mail clients provide a way to safe a mbox of a mail, then you can use "git am file.mbox" to apply the patch
<BorgCuba>
okay, currently I am using the web interface ...
dlezcano_ has joined #linux-rockchip
<dlezcano_>
karlp: a broadcast timer is a backup timer when the local timer is shutdown at idle time in the cpuidle driver