<wens>
Dodger78: i have no idea what the scheduler does or supports
<wens>
Dodger78: the DT has passed whatever information it can, like core type and max-frequency
adj_ has quit [Ping timeout: 252 seconds]
adj_ has joined #linux-sunxi
avph has quit [Ping timeout: 240 seconds]
adj_ has quit [Ping timeout: 260 seconds]
adj_ has joined #linux-sunxi
p1u3sch1 has quit [Ping timeout: 265 seconds]
p1u3sch1 has joined #linux-sunxi
[7] has quit [Disconnected by services]
TheSeven has joined #linux-sunxi
adj_ has quit [Ping timeout: 246 seconds]
adj_ has joined #linux-sunxi
vipzrx has left #linux-sunxi [#linux-sunxi]
adj_ has quit [Ping timeout: 240 seconds]
adj_ has joined #linux-sunxi
adj_ has quit [Ping timeout: 250 seconds]
adj_ has joined #linux-sunxi
kz1 has quit [Ping timeout: 260 seconds]
adj_ has quit [Ping timeout: 250 seconds]
adj_ has joined #linux-sunxi
kz1 has joined #linux-sunxi
adj_ has quit [Ping timeout: 252 seconds]
adj_ has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
adj_ has quit [Ping timeout: 252 seconds]
adj_ has joined #linux-sunxi
_massi has joined #linux-sunxi
asmir has joined #linux-sunxi
<mripard>
arnd: hmmm, thanks!
<mripard>
you can indeed access the DT node from sdio_func
<mripard>
It looks like I missed it entirely.
<mripard>
arnd: ah, no, I recall now.
<mripard>
the point was that you need the regulators enabled to even enumerate the device in the first place
<mripard>
and so it has kind of a chicken-egg issue, since your driver will enable your regulator
<mripard>
but your driver won't probe because your device wasn't powered yet
<wens>
mmc solves it with the pwrseq bindings
<wens>
but for BT, uart doesn't have the equivalent
avph has joined #linux-sunxi
ialokin has joined #linux-sunxi
<mripard>
wens: I'm still not a fan of that pwrseq stuff... There's several limitations with it (plus the fact that it doesn't really belong in the DT, but that's another story)
kaspter has quit [Ping timeout: 264 seconds]
adj_ has quit [Ping timeout: 256 seconds]
adj_ has joined #linux-sunxi
domidumont has joined #linux-sunxi
domidumont has quit [Ping timeout: 246 seconds]
diego_r has joined #linux-sunxi
ialokin has quit [Ping timeout: 244 seconds]
clonak has quit [Quit: No Ping reply in 180 seconds.]
domidumont has joined #linux-sunxi
MaDMaLKaV has joined #linux-sunxi
adj_ has quit [Ping timeout: 272 seconds]
adj_ has joined #linux-sunxi
domidumont has quit [Remote host closed the connection]
cubeast has joined #linux-sunxi
domidumont has joined #linux-sunxi
clonak has joined #linux-sunxi
<wens>
mripard: it's a workaround for the chicken and egg problem
<wens>
more like work around it from outside the box
<mripard>
wens: I know.
<mripard>
but having something like a standard property in each of the nodes to power them up would have been working just as fine.
<wens>
pwrseq in the sub-nodes, instead of in the mmc host node?
<wens>
then core code looks at all the sub-nodes?
<mripard>
yes, for example.
FlorianH has joined #linux-sunxi
adj_ has quit [Ping timeout: 265 seconds]
adj_ has joined #linux-sunxi
avph has quit [Read error: Connection reset by peer]
adj_ has quit [Ping timeout: 240 seconds]
adj_ has joined #linux-sunxi
ialokin has joined #linux-sunxi
ricardocrudo has joined #linux-sunxi
philectro has joined #linux-sunxi
naobsd has quit [Ping timeout: 250 seconds]
naobsd has joined #linux-sunxi
ssvb has joined #linux-sunxi
nahom has joined #linux-sunxi
<nahom>
NiteHawk: Are you around? Nahom here!
ialokin has quit [Ping timeout: 245 seconds]
<nahom>
I installed sunxi on the nand of my Mele M3 and when booting, it stops on "random: nonblocking pool is initialized"
<nahom>
I got that from the serial console
Gerwin_J has joined #linux-sunxi
adj_ has quit [Ping timeout: 244 seconds]
adj_ has joined #linux-sunxi
<NiteHawk>
hi nahom
<NiteHawk>
what is that, a ubuntu-derivate boot up?
<nahom>
NiteHawk: Hello, it's been a while
<nahom>
Well, you might remember that I was having trouble putting a rootfs on the SD card that would boot
<nahom>
So, I finally was able to put Ubuntu 14.04 on the SD card via debootstrap
<nahom>
And now everything works perfectly when I boot from SD
asmir has quit [Read error: Connection reset by peer]
<NiteHawk>
googling for that error seems to indicate that it might just take a while at that point due to slow entropy collection
<NiteHawk>
dunno if that's specifically NAND-related
<nahom>
Next, while booted from the SD, I repartitioned the nand (mmcblk1) as per the instructions for the SD card
<nahom>
and I just copied everything form the card to the nand partitions... updated the references to the /dev/mmcblk1
<nahom>
and at somepoint it booted
<nahom>
but now it just hangs on that line
<nahom>
I did wait for it for a long time...like 20 minutes... but it wouldn't boot
<NiteHawk>
hmmm :/
<nahom>
Could the reason be because I copied everything (the root partition) from the working SD card?
adj_ has quit [Ping timeout: 240 seconds]
<nahom>
Maybe there are some settings and stuff written by the working OS regarding the current state from the SD card that are incompatible with the nand?
<NiteHawk>
if you properly changes all filesystem references that shouldn't be the issue - but of course there's a chance that you missed something
<nahom>
Could I purge some setting files that hopefully would be regenerated?
adj_ has joined #linux-sunxi
<wens>
ssvb: rtc works on my primo81 now, can finally send the dts patches
<nahom>
I did change everything I put in regarding the filesystem (boot.cmd, fstab, mtab)
<ssvb>
wens: that's great, thanks
<NiteHawk>
it might not even be directly related to that "nonblocking pool is initialized", maybe that's just the last message that gets displayed? e.g. i found http://ubuntuforums.org/showthread.php?t=2255475 - where it seems that mode setting was the culprit
<nahom>
Yeah...it could be...
<nahom>
Could the serial console be any use in debugging?
<NiteHawk>
not if you're dealing with an actual kernel "hang" or crash - it might just lock up in that case
<nahom>
There is one more thing... when booting from the SD, I leave the kb (usb) plugged in and no problem, but in case of the nand, it hangs on the keyboard detection line
<nahom>
"hid-generic 000.... input: USB HID v1.10 keyboard [CHICONY HP basic].... on usb ...
<NiteHawk>
if that message part of the kernel boot, or are you actually at the point of init scripts starting up? the latter might offer far better chances for doing some debugging - e.g. try to start some "failsafe" mode or disable scripts selectively
<nahom>
How do I do that?
<NiteHawk>
nahom: the usb stuff may be related to the u-boot version. older u-boots have trouble supporting USB 1.x devices at the same time as (sunxi) EHCI (= USB 2.0). Either upgrade to U-Boot 2015.07+, or try this workaround: http://linux-sunxi.org/Mainline_U-Boot#USB_1.x.2C_USB_keyboards
<NiteHawk>
it's still puzzling that you can boot from SD card without trouble... does that still work?
<nahom>
Yes! everytime without a glitch
<nahom>
Also regarding the uboot, for both the SD and Nand, I'm using the latest!
<nahom>
Which I git cloned and compiled on Aug 21st
<NiteHawk>
i see. in that case you might try to add a USB 2.0 hub between Mele and the keyboard - maybe it helps
<nahom>
yeah but even with the kb left unplugged, it's not booting.
<nahom>
Are you sure there isn't a hardware devices list I could purge...that was somehow cached somewhere
<NiteHawk>
if i were you i'd have a close look at the (serial) log of a successful boot then - what's the next message after "random: nonblocking pool is initialized"? that may give you an idea of what task/function/script runs next
<nahom>
Ok, let me reboot with the SD and try to see the difference
lerc has quit [Quit: No Ping reply in 180 seconds.]
<NiteHawk>
as you probably are aware, NAND (boot) support is pretty fresh into U-Boot - it's well possible there are still quirks and some loose ends...
atsampso1 has joined #linux-sunxi
<nahom>
Strange!
atsampson has quit [Ping timeout: 240 seconds]
adj_ has quit [Ping timeout: 259 seconds]
<nahom>
The "random: nonblocking..." message doesn't even show up in the console of the SD boot
<NiteHawk>
oh...
lerc has joined #linux-sunxi
<nahom>
Here are both the logs on pastebin
<nahom>
SD (successful): pastebin.com/6hshL8ad
<nahom>
Nand : pastebin.com/qnbYkE28
ricardocrudo has quit [Ping timeout: 265 seconds]
<NiteHawk>
the sdcard log doesn't have "nonblocking pool is initialized", but it shows:
<NiteHawk>
[ 1.757542] random: init urandom read with 8 bits of entropy available
<nahom>
What does that mean?
<NiteHawk>
anyway it happens near USB initialization, and the NAND log timestamp seems to indicate that "random: ..." only takes place after that USB keyboard detect timed out?
<NiteHawk>
[ 248.990748] random: nonblocking pool is initialized
naobsd has quit [Quit: naobsd]
<NiteHawk>
<nahom> "hid-generic 000.... input: USB HID v1.10 keyboard [CHICONY HP basic].... on usb ...
<NiteHawk>
did i get you right that this even happens with the keyboard UNPLUGGED the next boot?
<nahom>
Yes
<nahom>
Let me get you the log with the kb unplugged
<nahom>
Ok, let me edit all three files and try again... I'll report back in 3 minutes
<nahom>
Thanks!
reinforce has joined #linux-sunxi
paulk-collins has joined #linux-sunxi
viccuad has joined #linux-sunxi
<nahom>
NiteHawk: We seem to have another problem now... pastebin.com/RKnHEADL
ricardocrudo has joined #linux-sunxi
<NiteHawk>
nahom: at least you've mounted root successfully now. the problem seems to be that it's complaining about p1 (the FAT boot partition)
<nahom>
It's the exact same file structure that's on the SD, which is working just fine!
<NiteHawk>
nevertheless it gets you something that doesn't seem to happen with the SD card
<NiteHawk>
[ 2.688531] FAT-fs (mmcblk0p1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.
<nahom>
What do you suggest I try next?
<NiteHawk>
to make some progress: my understanding is that the boot partition isn't needed after u-boot has completed its work. maybe just try to edit fstab so p1 doesn't get automounted / fscked for now?
Gerwin_J has quit [Quit: Gerwin_J]
<nahom>
ok
<NiteHawk>
try to simply comment out the entire fstab line - you won't need /boot for the moment
<NiteHawk>
you'll eventually have to sort that out, but for now i'm more interested to see if the system boots up
<nahom>
:(
<nahom>
NiteHawk: Same output except that the one before last line ([ 2.688531] FAT-fs (mmcblk0p1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.) is now gone!
<NiteHawk>
hm. i'm puzzled by the long delays that seem to take place there, e.g. your last log shows
<NiteHawk>
[ 188.962489] random: nonblocking pool is initialized
<nahom>
Now it shows [ 210.97....]
<NiteHawk>
which means it took 180+ seconds (over 3 minutes) to reach the point where that was output. 210 is ~3.5 minutes...
<nahom>
Yeah...
<NiteHawk>
please power down your device, detach any USB peripherals and try again... somehow that may still be related to that USB stuff
<nahom>
I'm thinking that mtab might be the problem
<nahom>
NiteHawk: Same thing!
<NiteHawk>
drat.
<NiteHawk>
Ubuntu forums have several posts around 14.04 boot delays and "nonblocking pool is initialized", but none seems to really match your problem or offer a solution. E.g. http://ubuntuforums.org/showthread.php?t=2219277
<NiteHawk>
I also fail to understand what makes up the exact difference between your "sdcard" and "nand" use cases... :(
Gerwin_J has joined #linux-sunxi
<nahom>
NiteHawk: I have like 50 of these devices and I want to put them to good use, I don't want to buy SD cards for each. Plus the card protrudes out of the body and could accidentally be pulled out
<nahom>
NiteHawk: Isn't there anyway for me to debug the boot process from the serial console? Something like an interactive boot...
vishnup has joined #linux-sunxi
<NiteHawk>
i'm not an ubuntu expert. probably there's something like a failsafe or "maintenance" mode - i think they have a wiki that might provide some tips
<NiteHawk>
another question: did you try to "mix" to boot process, i.e. load u-boot, kernel and configuration from sd card, and have only the "root=" path pointing to the NAND?
ialokin has joined #linux-sunxi
bmeneg has joined #linux-sunxi
\\Mr_C\\ has quit [Quit: .]
ssvb has quit [Ping timeout: 244 seconds]
simosx has joined #linux-sunxi
ialokin has quit [Ping timeout: 245 seconds]
<bmeneg>
bbrezillon, hello, few days ago you recommended a .dts based on kernel 4.x with NAND support, but I lost the link, do you remember something about that? :D
<nahom>
NiteHawk: Seems like the problem is in the root partition of the NAND. I tried the mixed boot approach like you suggested. I changed the root partition in boot.scr and it froze at that same place
pstef has quit [Ping timeout: 246 seconds]
<NiteHawk>
nahom: okay, that at least lets you rule out that there's a problem with the early boot process (SPL / u-boot)
Net147 has joined #linux-sunxi
<bmeneg>
vishnup, simosx, wow, thank you both for these links! :D I didn't know about these irc logs!
pstef has joined #linux-sunxi
<nahom>
Yeah... I'm gonna recopy the root from the sd to the nand...just in case
<NiteHawk>
nahom: btw - i'm wondering if you did consider that rootfs is "read-only" by nature?
<NiteHawk>
("that" = NAND)
<nahom>
I don't understand...
<NiteHawk>
oh, wait... it isn't or? can the kernel actually write to that partition?
<nahom>
if it were readonly, would it have let me repartition it and stuff?
<arnd>
mripard: (regarding sdio/pwrseq): ok, that makes sense. I haven't looked at the details of the power sequencing implementation, but as I recall, it was really meant to solve the problem you described. If it doesn't do that (or doesn't do it well), we should try to fix it
<NiteHawk>
nahom: maybe my train of thought is completely off-tracks there... let's me peek at the logs again
<NiteHawk>
and after that there's a long pause... maybe the kernel tries to write something and fails?
<arnd>
mripard: it would probably be good to discuss it with ulfh when you next look at that issue
<nahom>
When I booted the card the first time after I was done building it, it was the same way...
<nahom>
I mean it was readonly too but booted
<nahom>
and then I went into the fstab file and added the rw option
<NiteHawk>
could you check what happens if you change it back to "ro", will that let you boot?
<nahom>
let me check
nahom has quit [Ping timeout: 246 seconds]
nahom has joined #linux-sunxi
<nahom>
NiteHawk: Yes, it booted fine, excpet that now / is readonly
<NiteHawk>
well... given the evidence we've collected so far, i'd say it's safe to assume that there's a problem with (kernel) rw support for the nand partition?
<NiteHawk>
how did you get the filesystem copied/installed in the first place?
<nahom>
I booted from the SD, then re-partitioned the nand... then dd u-boot-sunxi-with-spl.bin
<nahom>
then copied contents of the SD's boot and root partitions to their respective partitions on the nand
<nahom>
That's what I did!
<NiteHawk>
so the same kernel was able to properly access (i.e. write) the filesystem when booted from sd?
<mripard>
arnd: we have a pretty bad wifi driver to be honest, so the pwrseq will be enough
<mripard>
but I would have expected everything to just work
<mripard>
anyway.
<nahom>
Yes
<arnd>
mripard: which driver source do you use btw? I think I've seen multiple driver trees
<NiteHawk>
nahom: that's... weird. i'm afraid i'm running out of ideas what going wrong there
<nahom>
NiteHawk: You've been so helpful and you've brought me thus far. I wouldn't have made it this far if it weren't for you...
<nahom>
I'll keep trying different things.
<nahom>
If you get any ideas, please do let me know
<mripard>
arnd: I'm not sure, the wifi part has been outsourced to some other consultant
<bmeneg>
vishnup, is it normal that on sunxi_defconf there isn't any support either to kernel loadable modules and Allwinners SoC nand enabled?
<mripard>
and I've not looked into that yet
<arnd>
mripard: ok
<arnd>
mripard: I suppose it will have to go through staging, and eventually get rewritten into a subdriver of drivers/net/wireless/rtlwifi/, but that will take a while
<arnd>
we seem to have a handful of rtlxxxx drivers in staging at any given point in time
<wens>
i've tried that one, but couldn't get it to associate with our AP (WPA2 protected)
pmattern has joined #linux-sunxi
adj_ has quit [Ping timeout: 244 seconds]
popolon has quit [Ping timeout: 244 seconds]
vishnup has quit [Remote host closed the connection]
vishnup has joined #linux-sunxi
cubeast has quit [Quit: Leaving]
rZr is now known as RzR
ssvb has joined #linux-sunxi
<nahom>
NiteHawk: You there?
popolon has joined #linux-sunxi
<NiteHawk>
nahom: ?
<nahom>
I think the problem is the copying of the root partition form the SD to the nand
fredy has quit [Excess Flood]
<nahom>
I tried to do it again with -v for verbose and I'm seeing a lot of errors...
<nahom>
perhaps some crucial files are not being copied over
<nahom>
Do you know of a way I could dd without having to also dd the free part of the SD (the SD is 16 GB with most part of it empty) but the nand is only 4 GB
<NiteHawk>
are you saying you try to copy the / of a *running* system? if so, i'd expect some problems - yes
<nahom>
Can I just get an image of the part of the root partition of the SD that contains the files only?
<nahom>
Well, I'm a noob so please don't blame me :)
fredy has joined #linux-sunxi
<NiteHawk>
do you have an sdcard (installation) image for that ubuntu? you may setup loop devices with "losetup -o <n>" to access individual partitions
<nahom>
Since I needed to access the nand and the only way to do that was from the SD
<nahom>
No, I don't have an image of the SD
<nahom>
I built it up right on the SD
<NiteHawk>
where did you get the contents of your SD card / from then?
<nahom>
I compiled the kernel on my pc
<nahom>
and I directly built the root partition using debootstrap
* Wizzup
is trying linux-next on a olimex a10 lime right now, and it seems usb simply doesn't work
<Wizzup>
Figured I'd try the otg/musb patches :)
<bmeneg>
guys, I am facing a problem to boot my kernel image, when booting it stops with "Error: unrecognized/unsupported machine ID (r1 = 0x000010bb)." and "Available machine support:" just values like 0xffffffff. I am trying to boot with a .dtb of my cubietruck "bootm $kerneladdr - $dtbaddr", is there any relation among the boot error and the .dtb file?
<NiteHawk>
nahom: you could still apply the same logic/procedure to a "loop-mount" filesystem at /mnt/ - meaning you'd end up with the root inside a "container" file. If you keep the file reasonably sized (small enough), it's possible to dd it over to the NAND root partition (and resize2fs it)
<bmeneg>
ah, my system is on NAND and I am using kernel linux 4.2 with nand support and uboot 2011.09 with nand support too.
<nahom>
Let me give that loop thing a try
FR^2 has left #linux-sunxi [#linux-sunxi]
<NiteHawk>
you probably want the loop file to be slightly smaller than your intended NAND root partition, e.g. say 3.5G if your partition is ~4GB
<mripard>
bmeneg: use a u-boot that can handle the DT properly, or append you DT to the kernel
<Wizzup>
Probably misconfig
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<NiteHawk>
(to be sure you can dd it over 'safely')
<mripard>
note that you won't have quite a lot of features without a recent u-boot.
<bmeneg>
mripard, hmmm... I can append the dtb to any kind of kernel image right? either zImage and uImage?
<NiteHawk>
nahom: also, if you build it on you pc it might not fit onto the existing sd card any more - meaning you'd need some temporary storage to be able to transfer it to nand. e.g. a usb pendrive, or maybe you can mount a network share from your pc?
<mripard>
bmeneg: yes, but you need to generate the uImage after you've appended the DT
<mripard>
bmeneg: but why are you using that old u-boot ?
<bmeneg>
mripard, because I am building the system over nand, and it was the only one which worked for me with no problem with others 3.x kernels
<GreatSage>
hey guys
<GreatSage>
I'm starting a new page for a sun4i-based POS terminal
<mripard>
bmeneg: I'm not even sure both are compatible.
<mripard>
(old u-boot nand stuff with new linux nand driver)
<GreatSage>
which category would that fall into? (board/tablet/HTPC)
<bmeneg>
mripard, i am using patched version for both of them, so lets see what happens :D
domidumont has quit [Read error: Connection reset by peer]
<NiteHawk>
nahom: actually you won't event a loop mount for that. just don't debootstrap into a mounted filesystem, but use a (temporary) regular directory - and then create a tarball from that. step2 would then be that you unpack that tarball to the nand root fs.
<NiteHawk>
nahom: to save (build) time, you can simply tar up your existing sd card root (fs) - from your pc, while it's not running of course ;)
<Wizzup>
ok - it was a misconfig on my side, retrying now
<NiteHawk>
*you won't even need a loop mount
viccuad has quit [Quit: WeeChat 1.1.1]
<nahom>
NiteHawk: Can you tell me how to tar up the root fs on my pc?
<nahom>
Then also how would I "write" that tar file on the nand?
domidumont has joined #linux-sunxi
a9afs9af has joined #linux-sunxi
<NiteHawk>
nahom: you're using the web client/gateway - please check for a private chat msg / tab!
a9afs9af has left #linux-sunxi [#linux-sunxi]
FRQuadrat has joined #linux-sunxi
pirea has joined #linux-sunxi
MaDMaLKaV has quit [Quit: Leaving]
Gerwin_J has quit [Ping timeout: 264 seconds]
Gerwin_J has joined #linux-sunxi
FRQuadrat is now known as FR^2
Igorpec has quit [Ping timeout: 264 seconds]
simosx has quit [Quit: Leaving]
FR^2 has left #linux-sunxi ["Connection reset by peer"]
rubensm has joined #linux-sunxi
Gerwin_J has quit [Ping timeout: 264 seconds]
gzamboni has quit [Ping timeout: 244 seconds]
Igorpec has joined #linux-sunxi
tomboy64 has quit [Quit: Uhh ... gotta go.]
gzamboni has joined #linux-sunxi
nahom has quit [Quit: Page closed]
Nacho has quit [Ping timeout: 244 seconds]
tomboy64 has joined #linux-sunxi
Nacho has joined #linux-sunxi
adj_ has joined #linux-sunxi
ialokin has joined #linux-sunxi
ialokin has quit [Changing host]
ialokin has joined #linux-sunxi
domidumont has quit [Read error: Connection reset by peer]
afaerber has quit [Ping timeout: 264 seconds]
vishnup has quit [Ping timeout: 246 seconds]
vishnup has joined #linux-sunxi
ialokin has quit [Read error: Connection reset by peer]
ialokin has joined #linux-sunxi
<mripard>
NiteHawk: what?!
<mripard>
NAND is not represented as an mmc device...
<mripard>
eMMC is
afaerber has joined #linux-sunxi
<NiteHawk>
oh... my bad. i'll fix that of course
<mripard>
so this has nothing to do with the patches related to the NAND that are floating around
<NiteHawk>
i see. but with s/NAND/eMMC/ the basic caveat would still apply, at least it could catch unsuspecting users?