<hno>
oliv3r, u-boot reset does not seem to work now..
bsdfox\ has quit [Read error: Connection reset by peer]
bsdfox has joined #linux-sunxi
bsdfox has quit [Changing host]
bsdfox has joined #linux-sunxi
ssvb has joined #linux-sunxi
* rz2k
checks out a20 kernel from allwinner
<rz2k>
classic allwinner, sata driver doesnt even compile, to compile the pm support, you need to run another script other than make..
ssvb has quit [Ping timeout: 256 seconds]
soul has quit [Quit: No Ping reply in 180 seconds.]
soul has joined #linux-sunxi
BJfreeman has quit [Quit: had a good time]
rzk has joined #linux-sunxi
rz2k has quit [Ping timeout: 264 seconds]
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
ganbold_ has quit [Ping timeout: 252 seconds]
eebrah|away is now known as eebrah
eebrah is now known as Guest53383
Guest53383 is now known as eebrah|away
rzk has quit []
<hno>
oliv3r, rewritten watchdog_reset() almost completely so it works again..
<hno>
I said you should not look at cmd_watchdog, and also you need to read up a bit on operator precedence. Also made use of that shiny register value array you added...
<hno>
phew.. finally took the jump and updated u-boot sunxi branch.
<oliv3r>
and yeah, i think i forgot to use the table :S
BJfreeman has quit [Quit: had a good time]
<oliv3r>
i like your way better btw :)
piyushverma has quit [Read error: Connection reset by peer]
piyushverma has joined #linux-sunxi
paulk-desktop has joined #linux-sunxi
n01 has joined #linux-sunxi
ganbold_ has quit [Remote host closed the connection]
IggyKoopa has quit [Quit: Lost terminal]
eebrah|away is now known as eebrah
eebrah is now known as Guest17689
ganbold_ has joined #linux-sunxi
Soru_ has joined #linux-sunxi
soul has quit [Ping timeout: 276 seconds]
tzafrir has quit [Read error: Connection reset by peer]
tzafrir_laptop has joined #linux-sunxi
Guest17689 is now known as eebrah|away
sanka has joined #linux-sunxi
<Turl>
oliv3r: you insist on using goto huh? :P
<Turl>
woot, i2c merged for 3.11 :)
eebrah|away is now known as eebrah
eebrah is now known as Guest34330
\\Mr_C\\ has quit []
Guest34330 is now known as eebrah|away
\\Mr_C\\ has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
Soru_ has quit [Remote host closed the connection]
cyp has joined #linux-sunxi
<cyp>
hi
soul has joined #linux-sunxi
<cyp>
I have marsboard I successfully boot it on µSD card but I can't install on nand
<cyp>
anyone a done it succesfully ?
<Turl>
how are you installing to nand?
<cyp>
I try several ways
<cyp>
dd my sd card
<Turl>
that's not gonna work
<Turl>
at all
<cyp>
and now I get it now
<Turl>
now you'll have to livesuit an image to get the bootloader and partitions in place again, if you really raw dd'd your sdcard to nand it'll be all broken
<cyp>
Turl: behind repair ? or just need to be repair ?
<Turl>
cyp: you just need to livesuit an image to repair
<cyp>
ok
sanka has quit [Quit: Leaving]
<soul>
Hello, I'm searching the way to uninstall Android and install a GNU/Linux distribution on a Hyundai A7 (http://paste.debian.net/plain/10564) Am I on the right place?
rellla has quit [Remote host closed the connection]
Turl has quit [Remote host closed the connection]
Turl has joined #linux-sunxi
Turl has quit [Excess Flood]
Turl has joined #linux-sunxi
ganbold__ has joined #linux-sunxi
ganbold_ has quit [Ping timeout: 252 seconds]
wingrime has joined #linux-sunxi
aexl has joined #linux-sunxi
<aexl>
lo. did allwinner ever have a cortex-a5?
<Turl>
aexl: I've never heard of one so far
<aexl>
Turl: hm. i guess they confuse it with telechips ...
<soul>
Sorry for disturbing, someone red what I said? x)
<soul>
aexl: I would like to help but I'm not a programmer and I don't know anything about electronics... It took me a year more or less to find this channel (thanks to the #debian-arm channel). I thought it was impossible to find information about this, because everyone said "buy a laptop if you don't want Android then". But I want to install Debian, i.e because I don't use the tablet with Android but If I could find more uses with
<soul>
Debian i.e
<rz2k>
hno: what machine id did you set for sun7i in a20-olinuxino/eoma68 support patch?
<rz2k>
because default kernel from amery's git has machine id that doesnt match one defined in your wip/a20 u-boot branch
<aexl>
soul: i mean just infos for the wiki. format doesn't matter.
<soul>
aexl: What do you mean? Sorry If I make you tired haha
<soul>
One thing, I have a Hyundai A7, not A7HD. It will work for me? Because my device is previous
<aexl>
soul: i mean if you find errors or can elaborate on the wiki articles.
<aexl>
soul: difference is the display resolution and ram. so you probably need your own u-boot (and script.bin in any case).
<soul>
aexl: Ok thank you, I'm still reading. I will talk a lot here haha
tinti has quit [Read error: Connection reset by peer]
<aexl>
soul: np
ganbold__ has quit [Remote host closed the connection]
ganbold__ has joined #linux-sunxi
<aexl>
what is the correct way to write dram.c for my tab? my script.bin is mostly empty especially density and width ...
<oliv3r>
, "Minimize the number of returns in each routine", "use returns judiciously".
sanka has joined #linux-sunxi
<oliv3r>
but if i can please you both; i'll change my ways for you fuys!
<mripard>
Turl: some guidelines enforce a single output point
<Turl>
oliv3r: "Use a return when it enhances readability", "Use guard clauses (early returns or exits) to simplify error processing"
<Turl>
:P
<mripard>
the one I can think of right now is the one used in the french high speed train software
<mripard>
the argument is that it's easier to get where it exits
<oliv3r>
yeah I've read several other disgn documents/practises
<mripard>
the downside being that you have a lot of imbricated if statements
<Turl>
mripard: yeah but if you have that problem then your function is too big IMO
<oliv3r>
Turl: absolutly true
<mripard>
Turl: it's another problem, here, the guideline enforce you to use a single return, no matter if your function is big or not, it's just to be consistent
<oliv3r>
well the kernel uses goto to exit when there's stuff to cleanup; but not otherwise, which I find inconcistent
<mripard>
why?
<Turl>
oliv3r: readability > consistency if you wish
<mripard>
goto is convenient to factorise the exit path
<mripard>
if there's no exit path, why bother using goto?
<oliv3r>
so its concistent with other times :
<oliv3r>
:D
<mripard>
then, it's like saying that you assign a variable like this
<mripard>
a = (1);
<mripard>
just to be consistent with the cases where the braces are indeed needed :)
<oliv3r>
lol
<oliv3r>
touche
<oliv3r>
i've never seen it with numbres; But I probably would :)
<oliv3r>
I guess the reason you could argue for these more 'optimized' ways, is that you can assume anybody writing and reviewing code is high in their skill :)
<oliv3r>
Turl: yes i would do it like that except
<oliv3r>
i would let printf(); /* fall through */
<oliv3r>
:D
<Turl>
:p
<oliv3r>
i think a few of the reasons why you have certain code regulations
<techn_>
I hate goto's and return middle of function.. also recursive if's are bad :p
<oliv3r>
Turl: all 3 are true
<oliv3r>
techn_: *
<oliv3r>
techn_: so how do you do it?
<Turl>
oliv3r: he writes in asm, no ifs there ;)
<oliv3r>
Turl: :p
<oliv3r>
Turl: no i don't like mid function return! :p
<techn_>
oliv3r: better desing :)
<Turl>
ifs are more like conditional goto
leowt has joined #linux-sunxi
<oliv3r>
anyway, i think why there's these rules/recommendations, as you often have unintersted developers, who write code for money, but don't care.
<leowt>
hello
<leowt>
pplz
<oliv3r>
so quality is an issue and mistakes get made. code doesn't get reviewed etc
<Turl>
howdy leowt
<Turl>
oliv3r: also less lines -> less space used -> faster compile ;)
<oliv3r>
Turl: compile time should not be a reason to write certain code
<oliv3r>
while slightly an issue in 1960
<Turl>
oliv3r: I know, just kidding there :)
<oliv3r>
lol
<leowt>
Turl: found serial on that tablet i told you, got adb working through network. Mayb there is a bug in the usb
<techn_>
leowt: which device?
<leowt>
at the end of next week i will get a a20 and a31 tablet
<Turl>
leowt: :)
<leowt>
an generic chinese rk3168 tablet
<oliv3r>
Turl: anyway, to wrap things up; goto/return isn't handled in the kernel coding styles :) but it should really be added
<Turl>
leowt: were you using 3.4?
<Turl>
ah, rk, nvm :p
<leowt>
Turl: it is rockchip ;)
<leowt>
since the price is very similar, and linux-sunxi has been maintaining allwinner kernel code *cof*
<leowt>
is better to have allwinner devices around
<leowt>
=)
<leowt>
thing is that rockchip is very well placed in power consumption in the last chips
<Turl>
leowt: how's performance and code availability?
<leowt>
performance is good, didng test it deeply but all seems fluid with android
<leowt>
code availability is what has been bugging me
<leowt>
only kernel is somewhere
<leowt>
but for 3188
<rz2k>
hno: did you encounter problems with bootconsole on a20?
<leowt>
but it seems that 3168 shares the same kernel
<oliv3r>
how do I grep for [a]?
<hno>
rz2k, not really. Why?
<oliv3r>
i thought a simple \[a\] would do it
<rz2k>
hno: kernel loads to serial initialization and then spits out only first characters of the each text line, then panic and reboot
<rz2k>
this is with earlyprintk
<rz2k>
without earlyprintk it doesnt print anything to serial and silently reboots (probably same panic somewhere(
<hno>
oliv3r, I am using 'WDT_TIMEOUT_OFF, just wrong version pushed to my own repository. Right version in linux-sunxi repo..
<oliv3r>
hno: sorry for the crap i sent you lsat night; i re-read it and it was really bad :p
<oliv3r>
but i was really tiredto be fair :)
<hno>
rz2k, I have only tried to boot the shipped Android kernel so far, and it boots but hangs during initialization. But earlyprintk seems to output just fine.
<oliv3r>
my a20 doesn't boot when powerd with USB-OTG
<oliv3r>
it hangs after the first line of output
<hno>
I power via 12V.
<oliv3r>
this is with power from my monitor USB hub, which probably doens't do more then 400 mA
<oliv3r>
yeah now I do too
<oliv3r>
but onyl because my scanner brick fit
<oliv3r>
i have a fitting 5V 2.5A charger and that seemed to work too
<hno>
There probably should be a jumper or the like for enabling power via USB. You really need a good USB power source for it to work.
<oliv3r>
but i wouln't be supprised if allwinner developers only 'guess' at some of the things they do, due to lack of proper docs etc
<oliv3r>
'oh it works now, odn't touh it anymore'
<oliv3r>
yep clock_standby actually uses the real watchdog
<oliv3r>
but i dn't see them changing the wdt_ctrl register
<oliv3r>
hno: i've sent you 2 other cleanup patches a few days ago 'fix headers' and a nother one, did you see them?
<mripard>
Turl: nope, allwinner said "use the A13 one" when I asked
<oliv3r>
mripard: did you get a20 docs
<mripard>
neither
<oliv3r>
'use a13 docs'?
<rz2k>
olimex has a20 sdk
<oliv3r>
no usermanual with it i bet
<Turl>
mripard: but it has HDMI and A13 doesn't o.O
<oliv3r>
'use a10 usernamual'
<mripard>
Turl: yes, and the muxing, clocks, etc. are different as well
<Turl>
mripard: is there source for A10S I can use?
<oliv3r>
mripard: did you receive an olimex-a20 yet?
<mripard>
oliv3r: yes, last week
<Turl>
mripard: did you ask for it?
<mripard>
source of what?
<mripard>
ask for what?
<Turl>
A20
<mripard>
ETOOMANYQUESTIONS :)
<Turl>
:)
<hno>
oliv3r, Allwinner never touch the control register.. did not mean that we should not.
<mripard>
nope, didn't ask, just received it
<mripard>
it was a nice surprise :)
<oliv3r>
hno: i wonder how they re-arm the wdt though; don't you need ctrl reg for it :)
<Turl>
mripard: I bet I'll get one in a month or so then, judging by the A10S shipping speed :p
<oliv3r>
i wonder when cubieboards arrive
<hno>
oliv3r, they don't. They rely on the wdt to be fully stopped so it rearms itself when they try to reset the CPU..
<Turl>
oliv3r: did they ship?
<hno>
haven't heard about them ship yet.
<hno>
oliv3r, I might have overlooked those two. Can you resubmit them?
<mripard>
Turl: I don't know, I had them something like one week apart
<mripard>
but sending boards to argentina seems to be a pretty... interesting experience :)
<oliv3r>
Turl: tom said they would the 'next day' so i assume so yes
<oliv3r>
hno: sure
<hno>
I don't trust "next day" statements in such context. Only "have been shipped"
<oliv3r>
hno: but this is in 'clock standby' anyway, how does it rearm itself without the key?
<hno>
oliv3r, if you disable the WDT, wait for the timer to expire and then write a new timeout then it's rearmed.
<hno>
Most times even works writing a new value while it's counting, but the timeout is then the current counter value not what you write, and have sometimes seen that the wdt hangs (or at least never times out) when doing so.
<hno>
and with the wdt normally disabled it's not hard to rearm it.
<oliv3r>
oh you enclose brackets in brackets to grep for them
<hno>
correction, you only need to wait one timer tick, not a full timeout.
<hno>
oliv3r, \[ is better.
BJfreeman has quit [Ping timeout: 252 seconds]
<oliv3r>
hno: doesn't work
<hno>
works for me.
<hno>
what did you try?
<oliv3r>
i did grep \[nand
<oliv3r>
it complained about unbalanced ['s
<oliv3r>
grep for \[nand\] and it searches for anything maching n, a, n, d
<hno>
you need quotes. grep '\[nand', if not the shell eats the \
<oliv3r>
yeah, took me a while to find, either quote it, or double bracket it :)
<oliv3r>
hno: anyway, n01 did some wdt tests and confirmed you NEED the key to rearm it
<oliv3r>
but their trick I suppose works too
<oliv3r>
atleast we should have figured out pretty much how the wdt works by now
<hno>
you need the key to rearm it while it's counting.
<oliv3r>
ah, see, we know it pretty wlel :)
<oliv3r>
i'll add that to the wiki
<oliv3r>
volatile brain is volatile
<hno>
It is in the register guide.
<oliv3r>
hno: btw, what did you mean with tyour tip toeing around registers comment? generally speaking that is
<oliv3r>
ah, it is allready? good
<hno>
what tip toeing?
<hno>
oliv3r, and you copied it to he wiki version as well I think.
<Turl>
mripard: as long as you don't use DHL or friends there's not many issues, just the slowness :p
<oliv3r>
hno: it doesn't specially say it works thatway, we only have 'a10 might not use the key'
<hno>
a10 uses the key.
<hno>
no idea where you got the idea it does not from.
<oliv3r>
Turl: Warning: apc_store(): Potential cache slam aerted for key 'sunxi-mw:user:id:454' in /srv/http/sunxi/linux..... on line 59
<hno>
sun5i did not exist when the watchdog drivers was implemented.
<hno>
The A10 user manual do not document the wdt control register at all, and the a13 user guide have the right bit fields, but forgot to mention the key..
<oliv3r>
the documentation is a horrible mess
_BJFreeman is now known as BJfreeman
<hno>
rz2k, is uart0 correctly defined in your script.bin? Looks like it gets lost somewhere when the UART is probed and configured (earlycon disabled).
<oliv3r>
rz2k: why does yours only have 512 MB; IIRC mine has 1g
<rz2k>
i have weird script.bin
<rz2k>
i extracted it from android nanda
<oliv3r>
hno: yours has 1G aswell right?
<oliv3r>
should have, caus emnemoc wanted to double it to 2g
<hno>
There is only one A10(/A20) MICRO board layout.
<hno>
same layout for both chips.
<oliv3r>
and they should have 1 G
<hno>
yes.
<oliv3r>
so your spl seems to be wrong to start with
<rz2k>
wait I google up the ddr's
<oliv3r>
wrong freq. too, 960 MHz is weird, should be either 912 MHz or 1080 MHzs (if its some weird sun4i bootloader)
<oliv3r>
'memphis' parts
<rz2k>
i have two hynix h5tq2g63bfr
<rz2k>
hmm
<rz2k>
H5TQ2G63BFR-*xxC
<rz2k>
128Mx16
<rz2k>
...
<rz2k>
it is 256MB chips
<rz2k>
ffs Tsvetan, why...
<oliv3r>
i think i have 'memphis'
<hno>
Hm.. I also get 960MHz.
<oliv3r>
but we saw it set to 912 Mhz remember
<hno>
Yes, but have no pll1 config entry for 912MHz so code picks 960MHz.
<rz2k>
oliv3r: hno: on your a20 boards, how are ddr3's named?
<hno>
two bugs there.. it should pick the next lower, not higher, and pll1 table not updated for a20.
<hno>
rz2k, some MEMPHIS chips.
<hno>
rz2k, you have NAND on the board?
<rz2k>
yes
<hno>
rz2k, so you got one of their experiments...
<hno>
time for bed, and hopefully u-boot JTAG debugging tomorrow. Got OpenOCD running again with a little handwaving (segfaults unless careful)
<oliv3r>
ouch
<oliv3r>
i just setup a jtag rig too, put my bus blaster/pirate to use
<hno>
I have some Olimex JTAG probes. But the problem seem to be more related to the sun4i than the JTAG probe.
<hno>
anyhow, night.
<oliv3r>
nn hno :)
<rz2k>
really disappointed in olimex actions. this will split the board community in two parts, the part where hno can fix/test stuff and part where 512mb users will sit and wait till someone figures out the problem.
<rz2k>
for example this uart problem seems to be somewhat connected to that
<rz2k>
I will try to remember where amery fixed something like it (it was on a10 around half a year or so)
<rz2k>
there where irq's conflict iirc
<rz2k>
but atleast it didnt panic that time and spit out random
sanka has quit [Quit: Leaving]
_BJFreeman has joined #linux-sunxi
BJfreeman is now known as Guest42590
_BJFreeman is now known as BJfreeman
Guest42590 has quit [Ping timeout: 252 seconds]
vincenzo has joined #linux-sunxi
BJfreeman has quit [Ping timeout: 252 seconds]
\\Mr_C\\ has joined #linux-sunxi
_BJFreeman has joined #linux-sunxi
_BJFreeman is now known as BJfreeman
\\Mr_C\\ has quit []
<vincenzo>
Hello, I have a tablet I thought to be a 907c, but running the hwpack on it does not work at all.
<vincenzo>
I tried copying the script.bin from android nanda
<vincenzo>
but I get no boot (or no lcd/hdmi video out).
<vincenzo>
How can I investigate in order to find what my device neesd to boot properly?
ganbold has quit [Read error: Connection reset by peer]