ifbizo_ has quit [Remote host closed the connection]
ifbizo_ has joined #linux-rockchip
vagrantc has joined #linux-rockchip
<mrjay>
thanks i'lll check your u-boot
ifbizo_ has quit [Remote host closed the connection]
ifbizo_ has joined #linux-rockchip
vagrantc has quit [Quit: leaving]
<anarsoul>
mrjay: it won't fix reboot issue though
<mrjay>
ah ok :-)
<anarsoul>
mmind00: any ideas what can cause linux to hang on subsequent boot?
<mrjay>
try vdd_log voltage setting in u-boot, maxcpus=4 boot option
<mmind00>
anarsoul: hmm, the only idea on short notice would cpu frequency vs. regulator voltage maybe ... aka system in some very low frequency state when rebooting
<anarsoul>
mrjay: vdd_log is already set, and I tried maxcpus
<mmind00>
anarsoul: I think uboot tries to ramp up frequencies during startup and pmic vdd_arm defaults are generally a bit higher
<anarsoul>
mmind00: I tried to set higher voltages for vdd_cpu_l and vdd_cpu_h
<anarsoul>
guess I can try setting max
<mmind00>
anarsoul: but then, it does survive uboot itself ... so that doesn't really look like a cause
<mmind00>
anarsoul: and btw ATF moved to gerrit, so they probably won't take your pull-request via github
<anarsoul>
:\
<anarsoul>
where's new location?
<anarsoul>
mmind00: u-boot doesn't bring up other CPUs. linux hangs *after* it brought up all the CPUs
<anarsoul>
(sounds like they've got some 'effective' managers)
<mmind00>
yep :-)
<mmind00>
anarsoul: interestingly I had a similar thing recently on the px30 ... which was caused by a bad cpu frequency
<mmind00>
anarsoul: what cpufreq governor are you running?
<anarsoul>
genpd_power_off_work_fn-related?
<anarsoul>
mmind00: conservative
<mmind00>
not exactly ... more like random rcu stalls, irq errors and a lot of random issues that were caused by the system somehow not liking 408MHz
<mmind00>
anarsoul: aka "apt-get install cpufrequtils" did cause the issue for me :-D ... because before it was just running with the "performance" (always max) governor, but after that switched to "ondemand"
<anarsoul>
mmind00: well, I can set CPU frequency in u-boot
<mmind00>
anarsoul: but I'm not sure uboot also sets the regulator
<anarsoul>
mmind00: it does not for rockpro64 :P
<anarsoul>
(see my github)
<anarsoul>
*now
<mmind00>
:-D
<anarsoul>
i.e. it does now :)
<anarsoul>
fan53555 driver was a mess, I'm 100% sure it wasn't even tested
<micken>
hey wizards! , I am trying to startup A72 on RK3399
<anarsoul>
micken: cool, still hacking your riscos?
<mmind00>
anarsoul: also ... what happens when you disable cpufreq in your kernel
<mmind00>
anarsoul: before and after reboot
<anarsoul>
mmind00: I haven't tried that yet
<micken>
I get ELR: "Synchronous Abort" handler, esr 0x02000000
<micken>
ELR: 7fd23e00
<micken>
LR: 7fd26204
<mmind00>
anarsoul: aka these stalls / genpd thingies are often just a "strawman" or what it is called and not the real reason
<micken>
I have tried both aarch64 code and aarch32
<micken>
same result
<anarsoul>
mmind00: 'red herring'
<anarsoul>
thanks, I'll try setting CPU frequency to some known value in u-boot
<anarsoul>
let's see if it fixes the issue
ifbizo_ has quit [Remote host closed the connection]
<anarsoul>
mmind00: looks like u-boot already sets CPU clock to 600 MHz for both cores
ifbizo_ has joined #linux-rockchip
<anarsoul>
see rkclk_init() in clk_rk3399.c
<mmind00>
anarsoul: yes I know ... still thinking that somehow there might be an undervolting of the cpu happening ... for example the fan5355 has this funny vsel gpio thingy that is not always initialized correctly
<anarsoul>
good point
<anarsoul>
I can check this gpio
<mmind00>
anarsoul: vsel-gpio selects which voltage setting register to use and I already managed to set the wrong one ;-)
<anarsoul>
let me check that
piter75 has joined #linux-rockchip
<anarsoul>
well, vsel-gpio is input and is 0
<anarsoul>
sorry, wrong gpio, it's actually 1
<anarsoul>
oops, nope, 0
<anarsoul>
darn, these symbolic names are confusing
piter75 has quit [Remote host closed the connection]
<anarsoul>
I already hate upstreaming process of ATF changes
<anarsoul>
why it has to be so complicated?!
<anarsoul>
gerrit documentation is terrible
mrjay has quit [Remote host closed the connection]
<anarsoul>
mmind00: I'm probably stupid but I just can't figure out how to submit my changes to ATF :\
<mmind00>
anarsoul: na, it's not you ... it's the darn process ;-)
<anarsoul>
so I have Change-Id (added the hook)
<anarsoul>
yet 'git push gerrit HEAD:refs/for/integration' complains " ! [remote rejected] HEAD -> refs/for/integration (branch integration not found)"
<mmind00>
anarsoul: my push example that I reuse all the time is "git push origin HEAD:refs/for/integration/cortex-a12-partid" ... with cortex-a12-partid being a new branch name
<anarsoul>
nope, ! [remote rejected] HEAD -> refs/for/integration/rk3399-fix-reset-poweroff-gpios (branch integration/rk3399-fix-reset-poweroff-gpios not found)
<anarsoul>
:\
<mmind00>
anarsoul: your gerrit-remote looks similar to "url = ssh://mmind@review.trustedfirmware.org:29418/TF-A/trusted-firmware-a"