Black_Horseman has quit [Quit: Zwi se logou mou!!!]
jinzo has quit [Ping timeout: 250 seconds]
gzamboni has quit [Read error: Connection reset by peer]
konradoo87 has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 240 seconds]
<quitte>
bbrezillon: Hi. I'm trying to get ddc's u-boot to work. could you help me out by explaining what happened to onfi,nand-timimng-mode, nand-ecc-strength and nand-ecc-step-size in the dts?
igorpec has quit [Ping timeout: 250 seconds]
ssvb has quit [Ping timeout: 245 seconds]
p1u3sch1_ has joined #linux-sunxi
p1u3sch1 has quit [Ping timeout: 246 seconds]
p1u3sch1_ has quit [Ping timeout: 244 seconds]
p1u3sch1 has joined #linux-sunxi
rds has joined #linux-sunxi
<rds>
it seems Odroid-C1 and R-pi 2 was the last straw on Allwinners dev boards!
<wens>
huh?
<rds>
for the reading, it seems all Cubue, Banana whatever, very expensive H/W cannot compete with 4 cores 1.5G, 1G RAM at $35...
<rds>
In my view, Allwinner only choice is to agressively market the A80 dev boards at $19.99
<rds>
and really put a team of S/W developers to deliver something that works!
<quitte>
rds: for me sata was the reason to pick the cubietruck. no idea why allwinner dropped that from the a80
<rds>
It seems that after 3 years.... these Cubie/banana/pcDuino, etc none work perfectly doing what people want, playing HD video without hassle!
<quitte>
that's not what i want
<atsampson>
gigabit (and non-USB-attached) Ethernet is also still an advantage...
<wens>
rds: i don't know, if you get an aw based tablet/htpc that might work
<quitte>
i still have a bit of hope that amd will put it's arm into the "devboard" market. Fast io. mmmmmh.
<rds>
I need to select a 10.1" tablet for a project, can you recommend something that works ? I prefer if it runs Ubuntu, but Android is "ok" !
<quitte>
rds: i bought the v975i and regretted it since basically i went through a lot of trouble to put the efi bios on it. now it is effectively a w. It is for most intents and purposes just a pc
<quitte>
rds: i modified a usb hub. now it charges while I can attach usb devices. I must admit I'm using windows now. Gnome3 and touch sucked too much.
<quitte>
but the bad touch capability of desktop linux is no fault of the hardware
<rds>
a google search shows it costs $200, I need something in the range of $50-$70
<quitte>
rds: imho the 200$ is well worth it. beautiful screen and the interaction feels smooth.
<quitte>
if anyone knows a tablet that basically is the same but with 64-bit efi - please tell me.
<rds>
This is not for personal use, it will be used in "Kiosk
<rds>
mode in a mall, and needs to be running 24x7, and the budget limit cost is something around $70 bucks.
<rds>
I am just wondering if something that this community supports, running Ubuntu exists!
<quitte>
I'd be hesitant to charge a 50$ china tablet 24/7 unsupervised
hramrach_ has joined #linux-sunxi
<Turl>
why don't you use something like an olimex LCD and a lime or so?
<rds>
that is why I am running trough this group, because it has very smart people here! Any suggestions are welcome!
<quitte>
backlit printout!
lioka has quit [Quit: leaving]
<rds>
let me look at the Olimex LCD
jinzo has joined #linux-sunxi
gzamboni has joined #linux-sunxi
<quitte>
rds: capacitive touch might be a bad idea. It tends to work badly if you don't hold the device with your other hand. It might become entirely unusable while being charged. especially in a 50$ device
lioka has joined #linux-sunxi
ssvb has joined #linux-sunxi
<atsampson>
NiteHawk: if you're updating the BananaPi page, you might want to pinch the explanation of the GMAC fix that I've just added to the pcDuino3 Nano page
<NiteHawk>
sure i'll have a look at that...
<rds>
trust me, I know that. But today, is very hard to talk to a customer, and suggest an input device, because the customer goes to Amazon and see all these $40
<rds>
''s tablets there, and basically tell us.... "I saw this and that device on Amazon" for $40 bucks
<rds>
You need to go though a lot of talking to have something better for $60's ....
<NiteHawk>
atsampson: btw, do you know about the current situation regarding u-boot networking? last time i checked, u-boot-sunxi-lemaker had a fix for the bpi - while the 'official' u-boot-sunxi was unreliable (excessive TFTP retries)
konradoo77 has joined #linux-sunxi
<atsampson>
NiteHawk: I've not tried it, but doing the fix in U-boot ought to sort that out...
konradoo87 has quit [Ping timeout: 245 seconds]
<atsampson>
(i.e. AFAIK nobody's "fixed" sunxi u-boot, but upstream should be OK)
<atsampson>
erm, that probably doesn't make much sense
<wens>
mripard_: the new a80 mmc/usb clocks require the reset framework
<atsampson>
Hans' version of u-boot that's due to be merged into the next release of U-boot proper has the GMAC_TX_DELAY fix for the BananaPi and pcDuino3 Nano -- so if you're using Hans' tree then it should work properly on both boards
<wens>
mripard_: should we set those clocks to build only if MACH_SUN9I is selected?
<ssvb>
Turl: that was very interesting to read, thanks for the link :)
<NiteHawk>
atsampson: if i get it right, the GMAC fix you wrote about affects the kernel? i'm currently using the lemaker flavor (https://github.com/LeMaker/linux-sunxi.git) with gentoo, and seem to experience no problems
<bbrezillon>
don't know if those fields are available in u-boot though
<quitte>
if that was the problem readding the information in the dts should make it work. have there been more dt changes since september that i should maybe be aware of?
<quitte>
re-adding
<bbrezillon>
I just checked and ddc ported those changes too
<bbrezillon>
so this should work without adding anything specific in your DT
<quitte>
his code definately does try to read those values from the DT
<quitte>
bbrezillon: it detects a 8G chip but all nand commands fail.
<bbrezillon>
for nand-ecc-xxx properties, that's expected cause defining them is used to overload default NAND ECC requirements (those defined in the datasheet)
nicksydney has quit [Quit: No Ping reply in 180 seconds.]
<bbrezillon>
but keeping them undefined should work
nicksydney has joined #linux-sunxi
<quitte>
okay
<bbrezillon>
(unless you miss an entry for your NAND device in the nand_ids table...)
<quitte>
is there now a simple way to do the initial erase?
igorpec has quit [Ping timeout: 264 seconds]
<quitte>
if i recall correctly ddc has a cubietruck, too so that shouldn't be an issue
<bbrezillon>
nand erase ?
<quitte>
is there something to probe on the hardware with an oscilloscope to see if there is any activityß
<quitte>
yes, nand erase with the wrong initial badblocks
<quitte>
on the other hand that already is sorted out
<bbrezillon>
if your NAND chip is properly detected, this means there's actvity on the NAND signals ;-)
<quitte>
okay. so it is not just the fdt taking care of that
<bbrezillon>
nope
<quitte>
still it looks like _all_ the blocks were bad
<bbrezillon>
the chip is automatically detected thanks to the READID command (it's a standard NAND commanf)
<bbrezillon>
quitte: that's expected too, since the AW layer does not preserve the BB marker
<quitte>
aw?
<quitte>
allwinner?
<bbrezillon>
allwinner
<quitte>
that's long gone. i only have about 5 badblocks with your driver in linux
<bbrezillon>
quitte: okay, so I'd say you don't have the proper randomizer/ECC config
<bbrezillon>
IIRC ddc implemented a set of u-boot commands to configure them
<quitte>
bbrezillon: do you happen to know if u-boot keeps using the initial device tree or switches to the one loaded later?
<bbrezillon>
I think it keeps using the initial one for its own use
<bbrezillon>
AFAIR, the one you load for the kernel is tweaked but is only passed to the kernel and not use by u-boot for its devices config
<bbrezillon>
BTW, I don't see any specific command to configure the randomizer or ECC block, so it might well take what's defined in the DT
<quitte>
when it shockingly worked it worked really well. I don't think I even had mtdids set then. Maybe I should focus on fixing the environment by just assuming that it does indeed already work?
<quitte>
bbrezillon: when I access the nand by address - is it supposed to figure out the per partition settings from the dt?
<bbrezillon>
What did you change since then ?
<quitte>
bbrezillon: i have no idea. I couldn't even figure out what happened back in october/november
<bbrezillon>
I don't recall how partitions are working in u-boot
<bbrezillon>
but I'd say you have to access partitions not directly access the nand in raw mode
<bbrezillon>
not sure how you do that though :-)
<quitte>
okay. thanks a lot. maybe marex can shine some light on this.
<ssvb>
rds: as for the rpi2 alternative from allwinner, the h3 soc looks good (quad core a7 at a supposedly decent clock speed, mali, sata)
<ssvb>
rds: the soc costs 6$, so somebody might make cheaper boards eventually :)
jinzo has joined #linux-sunxi
<wens>
mripard_: the usb clock gates use it as well
igorpec has joined #linux-sunxi
<Turl>
ssvb: it's an A20 remake?
<Turl>
ah, quad A7
<ssvb>
Turl: yeah, h3 looks like a very much needed upgrade for a20
quitte has quit [Read error: Connection reset by peer]
<ssvb>
Turl: Mali400 MP4 would have been better than MP2 though
<wens>
isn't that just 2 more cores?
<ssvb>
wens: yes, and supposedly higher clock speed too (but we know that we can't trust this information in general)
leviathancn has quit [Remote host closed the connection]
<ssvb>
wens: when competitors (ex. rpi2) have an advantage in processing power, "just" doubling the number of CPU cores is not a minor improvement
quitte has joined #linux-sunxi
<Turl>
ssvb: but do they?
<Turl>
they have the same cores right?
premoboss has joined #linux-sunxi
<ssvb>
Turl: rpi2 has quad a7 @900mhz (with an overclock option to 1ghz)
<Turl>
ah, it's quad
* Turl
didn't pay too much attention
<Wizzup>
ssvb: interesting how some sources state the H3 has no SATA, but the board you linked has sata
leviathancn has joined #linux-sunxi
<a1d3s>
nitehawk i got m2 running but it alway stay on sunxi# an errror on boot command not found: part - try help , in help there is no part as option . if i type all from boot.scr then it switch to kernel boot
konradoo77 has quit [Ping timeout: 252 seconds]
<ssvb>
Wizzup: oh, you are right, this board has an external SATA chip on the back...
<NiteHawk>
a1d3s: that sounds a bit strange. are you using a u-boot that you compiled yourself?
<ssvb>
Turl: don't know, but I don't like it already :(
<a1d3s>
turl this patch is for a31 too?
<Turl>
a1d3s: it's for uboot in general
<Turl>
a1d3s: read the thread, it's the same error you're seeing
<a1d3s>
yes
afaerber_ has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
<a1d3s>
only i got an error on phy1 not found but on showenv it show eth at dwcmac.1c30000
<premoboss>
how to detemine in bash the type of SoC (A10 or A20) present on cubieboard? i could count the number of core in cat /proc/cpuinfo but if the kernel is not compiled for multiprocessor, there will be visible only one core also on A20. so how to discriminate?
afaerber has quit [Ping timeout: 264 seconds]
<ssvb>
premoboss: "CPU part : 0xc07" in /proc/cpuinfo means Cortex-A7
<ssvb>
premoboss: "CPU part : 0xc08" in /proc/cpuinfo means Cortex-A8
<premoboss>
ssvb, nice, thanks.
konradoo77 has quit [Ping timeout: 252 seconds]
<Turl>
also hardware: sun4i means A10, and sun7i is A20
<premoboss>
good
<ssvb>
Turl: I think this line differs in the sunxi-3.4 and the mainline kernel, so it is not very reliable for automatic parsing
konradoo77 has joined #linux-sunxi
<Turl>
yeah, I don't recall what mainline prints there
<Turl>
3.4 has the sunNi names though
<a1d3s>
hm after apply this patch m2 doesnt boot :/
AnticitizenOne has quit [Ping timeout: 244 seconds]
<NiteHawk>
a1d3s: have you recompiled your u-boot? the 'patch' just sets two configuration variables
<NiteHawk>
(something you could easily do yourself using "make menuconfig")
<a1d3s>
i do a make clean , then make Bananapi_defconfig , menuconfig to switch a31 and the make cross comile
<a1d3s>
clean mbr with dd /dev/zero and burn the new boot.bin
<NiteHawk>
well, if it went smooth, you simply enabled uuid support and the "part" command. so the new u-boot will be just as fine as the previous one, or actually should get further in the boot process. the erratic behaviour might actually be something you couldn't observe earlier
<NiteHawk>
(due to the "missing command" problem)
<a1d3s>
hm after patch i didnt see anything over hdmi
<a1d3s>
i will clone and patch again
libcg has joined #linux-sunxi
paulk-aldrin has quit [Remote host closed the connection]
Akagi201 has joined #linux-sunxi
<a1d3s>
hm nothing to see after patching .. i go to bed
melonipoika has quit [Remote host closed the connection]
<rds>
thanks to all folks for the tips! I am researching them!
libcg has quit [Quit: libcg]
<rds>
the odroid-c1 is quad as well for $35, this H3 board is dead on arrival costing 35% more.
libcg has joined #linux-sunxi
<ssvb>
rds: odroid-c1 is a cortex-a5 based amlogic, from what I read about it, its kernel is rather messy and very far from being mainlined
<ssvb>
rds: sunxi at least is doing a decent job on the software front
<ssvb>
rds: the raspberry pi is not too bad either, except for the closed source bootloader
<rds>
i thought their sw ( odroid-c1 ) was already mainlined, since is based on DT...
<ssvb>
oh, maybe I need to check it again
<ssvb>
and in fact it's better to just buy this board to see myself whether it is any good :)
konradoo77 has quit [Ping timeout: 256 seconds]
Zboonet has joined #linux-sunxi
iamfrankenstein has quit [Quit: iamfrankenstein]
<rds>
from their furins, it is 3.10.66. whatever that means!
konradoo77 has joined #linux-sunxi
rds has quit [Quit: Page closed]
DagoRed has joined #linux-sunxi
FDCX_ has joined #linux-sunxi
vishnup has quit [Remote host closed the connection]
jinzo_ has joined #linux-sunxi
f15h has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 240 seconds]
konradoo77 has joined #linux-sunxi
f15h has quit [Quit: Leaving]
jinzo_ has quit [Ping timeout: 265 seconds]
jinzo has quit [Ping timeout: 264 seconds]
premoboss has quit [Remote host closed the connection]
konradoo77 has quit [Ping timeout: 244 seconds]
<quitte>
setting the MTDID and MTDPARTS default did not help :(