<hno>
slapin, you got 6 and the chip have 6 byte ID. But here is 8 if that makes more sense to you but last two is just repetition of the first two (and continues if you read more). ad d7 94 da 74 c3 ad d7
Mehhh has quit [Quit: Page closed]
FreddyAV has joined #arm-netbook
<slapin>
hno: this is what I want to hear, will check that today. We probably jut need to add hynix vendor ID to appropriate place
<slapin>
good morning, all!
ibrah has joined #arm-netbook
<slapin>
hno: ping
<slapin>
hno: do you have datasheet for your flash?
<slapin>
hno: anyway, drivers/mtd/nand/nand_base.c:2700 contains code for Samsung package, which works for me, please compare to HYNIX one, if matches, just add NAND_MFR_SAMSUNG check.
<slapin>
hno: else path there assumes 8-byte IDs hence, various data like data bus width are missing from 6-byte IDs.
<slapin>
hno: at least tell your p/n again
<slapin>
hno: can't find datasheet for it, probably it is samsung clone, or something
<empat0>
Graffiti: i had a short look. will try to compile it and see. gimli for sure has a deeper knowledge about dvdplayer as i have. what i do not understand is that everybody is re-inventing libcedarv, also the guy who has done the vlc port.
<empat0>
has someone tried vlc with xvid? does it work with xvid?
<empat0>
or does it only support the media types xbmca10 supports?
<slapin>
hno: pull my last patch, does it work for you?
ZaEarl has quit [Ping timeout: 252 seconds]
graffiti has quit [Ping timeout: 245 seconds]
eFfeM has quit [Quit: Leaving.]
eFfeM has joined #arm-netbook
empat0 has quit [Quit: Verlassend]
kaspter has quit [Ping timeout: 246 seconds]
ibrah has quit [Ping timeout: 250 seconds]
tinti has joined #arm-netbook
Max_nl has joined #arm-netbook
<Max_nl>
Hi guys
<Max_nl>
Anyone around that understands the internals of u-boot sunxi_mmc.c?
<Max_nl>
Get a "spl: mmc blk read err - 0" on a board with a low dram_clk/PLL5 speed (288 mhz)
<Max_nl>
Does work if I raise the clock. But prefer to use the settings supplied by vendor, that do work under Android
focus_that has joined #arm-netbook
SouL_ has joined #arm-netbook
focus_that has quit [Quit: Leaving]
Max_nl has quit [Quit: Page closed]
rz2k has quit [Read error: Connection reset by peer]
QingPei has joined #arm-netbook
cat_x301 has quit [Ping timeout: 252 seconds]
grimes99 has joined #arm-netbook
<slapin>
awwwwwww
* slapin
hates watchdogs he can't disable :(
SouL_ has left #arm-netbook [#arm-netbook]
Max_nl has joined #arm-netbook
<slapin>
Max_nl: try to figure-out MMC clock calculation and try reducing it.
<Max_nl>
one moment, will try playing with the divider
slash_random has joined #arm-netbook
Max_nl_ has joined #arm-netbook
<Max_nl_>
does work with divider =2 instead of 3
<Max_nl_>
so increasing the MMC clock speed
gzamboni has quit [Read error: Operation timed out]
Max_nl has quit [Ping timeout: 245 seconds]
<slapin>
Max_nl: the best approach is to really calculate clock based on PLL5 value, look at mtd branch in hno's uboot-allwinner repository (file drivers/mtd/nand/sunxi_nand.) for approach and give it some thought.
<slapin>
Max_nl: don't produce hardcoded values, when one can be calculated
<Max_nl_>
well, I'm not sure if the mtd calculation is ok either
<Max_nl_>
anyway, I will report my findings to the mailinglist, and hope someone will come up with the ultimate calculation code. I do not really have much of a hardware background, so cannot help much with that.
gzamboni has joined #arm-netbook
<slapin>
Max_nl_: I said, it needs thought, so not to be used exacly like it is. But it is taken from original code, so might be close. And also, NAND works on 3MHZ clock, MMC can have much bigger (like 50MHz) clock.
<slapin>
Max_nl_: just give your math skills a try
hg_5 has joined #arm-netbook
<rm>
hmm so maybe there's hope for my 'bad' MK802
<rm>
where uboot-SPL does not detect MMC to load uboot
rsalveti has quit [Ping timeout: 265 seconds]
rsalveti has joined #arm-netbook
<mnemoc>
rm: does that happen with recent u-boot-sunxi too?
<rm>
chrisb, yeah it's a good overview, but I was particularly surprised how it repeats many of the points I made in my http://romanrm.ru/en/a10/mk802-server
<chrisb>
rm: ah, ok
hg_5 has quit [Ping timeout: 252 seconds]
rellla has quit [Remote host closed the connection]
SouL_ has joined #arm-netbook
FreddyAV has quit [Ping timeout: 250 seconds]
revident has joined #arm-netbook
hg_5 has joined #arm-netbook
rsalveti has quit [Read error: No route to host]
SouL_ has left #arm-netbook [#arm-netbook]
rsalveti has joined #arm-netbook
SouL_ has joined #arm-netbook
lerc_ has joined #arm-netbook
lerc has quit [Ping timeout: 260 seconds]
SouL_ has left #arm-netbook [#arm-netbook]
<hno>
slapin, will try tonight. Been sleeping all day.
<hno>
and still dead tired.
slash_random has joined #arm-netbook
<hno>
Max_nl_, don't know why I need to reduce PLL5 speed for our MTD driver to detect the nand. The NAND clock is the same (or actually slightly faster) with the reduced PLL5.
<hno>
And do work fine with PLL5 @480MHz if I first let Allwinner driver access the nand and then reinitialize the NAND controller using our code.
grimes99 has quit []
<hno>
shuld make myself a needlepad for NAND spying.. not in a mood for soldering wires on the cubie nand.
mikey_w has joined #arm-netbook
gimli has joined #arm-netbook
<Max_nl_>
And do work fine with PLL5 @480MHz if I first let Allwinner driver access the nand and then reinitialize the NAND controller using our code. <- So some odd settings persist even if resetting the nfc and chips?
vinifm has joined #arm-netbook
<vinifm>
mnemoc?
<L84Supper>
hno: did you ever find out if the A10 GPIO is fast enough for EPP type IO? To drive things like stepper motor pulses or bit-banging to 1 MHz
<vinifm>
mnemoc: i could use: "#define UART_GET_CSR(port)__raw_readl(UARTx_BASE(1) + UART_LSR)"
vinifm has quit [Read error: Connection timed out]
vinifm has joined #arm-netbook
hg_5_ has joined #arm-netbook
chrisb has quit [Ping timeout: 240 seconds]
<hno>
L84Supper, for stepper motor pulses I would use one of the PWM counters.
hg_5 has quit [Ping timeout: 264 seconds]
hg_5_ is now known as hg_5
<hno>
L84Supper, I am prettu sure bitbanging can be done at fairly high frequency, but you don't want Linux in the middle if targeting those frequencies.
<hno>
Max_nl_, does not seem to be a setting. More likely a timing of something.
<brooc>
I would like to get into kernel programming and bootloader
<Turl>
hno works on the bootloader :)
<brooc>
Does that mean it is taken?
<brooc>
Or that he is the one to talk to about this?
<hno>
The latter.
<brooc>
:)
<brooc>
Do you all own a dev board?
<Max_nl_>
hno: Yes, read your post
<hno>
Max_nl_, did it make sense?
<Max_nl_>
hno: problem with "divisor = pll5_clk / desired_freq;" is that it is not even clear to me what desired frequency is
<Max_nl_>
don't know what a MMC normally runs at
<hno>
somewhere close to what 408MHz gives with normal divisor I think.
<hno>
There is some upper bound one must not exceed.
vinifm has quit [Read error: Connection timed out]
<hno>
anything lower is ok, but if lower than needed then performance suffer.
vinifm has joined #arm-netbook
<brooc>
hno - where can I start?
<hno>
Max_nl_, maybe the kernel driver have some hint? It should set up clocks differently.
<Max_nl_>
then why didn't it work with divider 3, but does work with 2, if lower frequency is ok?
<brooc>
Who is doing the kernel modules?
<hno>
Hmm.. good question. Maybe low frequency is not OK in MMC/SD mode and only SPI mide.
<Max_nl_>
that's what I find difficult
<Max_nl_>
I have no idea of what is the normal min. max. frequency range
<hno>
brooc, several people involved with different drivers.
<hno>
Max_nl_, the kernel driver almost certainly have a exact frequency it's trying to use.
<Max_nl_>
I now just added to the u-boot I ship with my universal image: if lower then 300 mhz then use 2
<hno>
kernel clock driver works on frequencies, not divisors.
<Max_nl_>
which solves the problem for my problem board
<Max_nl_>
but am afraid that tomorrow someone may try a different board, and find out it doesn't work
<Max_nl_>
like my stuff to work everywhere
<hno>
That change is very unlikely to cause harm to anyone.
<Max_nl_>
yeah, figured there may be few other boards that have lower then 300 mhz
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
<brooc>
hno, do I need to buy a dev board or can I start without one?
<hno>
brooc, you don't need a dev board to do useful stuff, but usually more interesting to actually try running what you have done.
<hno>
so helps having some board with an appropriate CPU.
<hno>
brooc, where to start is the question I asked you above. What stuff you like doing most?
<brooc>
I am not sure yet
<brooc>
I would like to try kernel programming and the bootloader
<brooc>
both of which I think Ill enjoy
<brooc>
I have a RPi which is an ARM11 based SoC
<hno>
Yes, kernel and u-boot share a lot in programming style.
<hno>
but quite different in runtme environment.
<hno>
And both are almost 100% C.
<brooc>
I work in C and I like the language
<hno>
Good
<brooc>
I also enjoy asemblly
slash_random has quit [Ping timeout: 240 seconds]
<brooc>
How should I begin
<mnemoc>
if you enjoy asm you might want to RE the gps driver ;-)
<hno>
In u-boot we currently have 2.5 projects. 1. Mainlining the changes. 2. NAND MTD driver. 3. Figuring out how to handle the variety of boards we have.
<hno>
mnemoc, we have no hardware to test that on.
<mnemoc>
fair point :(
<hno>
Hmm.. do the A13 have GPS pins?
<brooc>
What would you suggest for a beginner to look into?
<brooc>
Where can I read about all this?
<brooc>
The NAND MTD driver sounds interesting
<hno>
As you have no board then trying to port a driver to Linux-3.7 might be a good challenge.
<hno>
For the current MTD work you basically need a board + logic analyser.
<brooc>
And I have neither
<brooc>
Wouldn't qemu work?
<hno>
We have no documentation on the NAND controller so we need to experiement a bit and see what happens.
<hno>
QEMU works fine for generic ARM stuff, but not for A10 specific drivers.
<hno>
i.e. none of the work we do in u-boot or kernel.
<vinifm>
mnemoc?
<brooc>
How much does a board cost?
<brooc>
Where can I get one?
<mnemoc>
vinifm: yes?
<vinifm>
mnemoc: i could use: "#define UART_GET_CSR(port)__raw_readl(UARTx_BASE(1) + UART_LSR)"
<techn>
all of those should be good for dev work, but it's ofcourse easier if board has serial port already soldered.. like dedicated dev boards or mele's
<techn>
hno: couple row abowe :)
<hno>
brooc, recently a minimal port to linux-3.7 was done. This is still almost an empty shell and needs to be fleshed out with more drivers.
<vinifm>
the logic is right? BASE + SOME_REGISTER_UART
<hno>
and in the process those drivers needs updating to use device tree instead of hardwired configuration.
<hno>
vinifm, yes.
<hno>
with shift << 2 to word-align registers.
<mnemoc>
vinifm: almost everything in this world is readl/writel a base + offset ...
<vinifm>
if it works, it will be very cool :)
<brooc>
Just to make sure I understand you mean that there is a basic port of the linux3.7 kernel for the A10 core. Is that correct?
<hno>
brooc, yes.
<brooc>
If so what driver would you say is a good place to start with?
<mnemoc>
did arnd merge it already?
<brooc>
How do I setup my dev environment?
<brooc>
Where can I start reading?
<hno>
mnemoc, not sure.
<techn>
brooc: A13 core is also mainlined
ibrah has joined #arm-netbook
<hno>
mnemoc, you have any suggestion on which driver to start on?
<mnemoc>
techn: arm-soc != master
<mnemoc>
hno: i2c maybe?
<hno>
Yes, that would be a good one, and extending the clock driver a bit too as needed.
<mnemoc>
but i haven't been able to follow the last weeks... trying to salvage a project :|
<hno>
neither have i, mostly seen the bed.
<brooc>
Where can I start reading?
<hno>
brooc, sorry for the confused view here. It's just that a lot of things are happening.
<mnemoc>
work on the clock driver needs more knowledge
<brooc>
I am in the clock recovery team where I work
<brooc>
Not sure what you refer to as clock driver
<Max_nl_>
Where can I get one? <- Can get A10 tablets at a lot of regular stores as well. Sold by electronic stores, toy stores and even a pharmacy chain here in NL.
<brooc>
How come Olimex do not sell A10-OLinuXino
<hno>
brooc, drivers/clocksource/ in the kernel tree.
<hno>
brooc, because it's not ready yet.
<hno>
or.. not sure any more.
<hno>
about clock driver. No that's not the right one.
<brooc>
hno, what should I start reading?
<brooc>
I have never ported a driver
<hno>
brooc, better start doing, and read bits you need along the way.
<brooc>
How do I setup a dev env?
<brooc>
I guess I'll need the sources from somewhere
<hno>
brooc, install a suitable cross-compiler. Pick an editor you like. And grab the sources using git.
* mnemoc
puzzled
<hno>
mnemoc?
<mnemoc>
brooc confused me :)
<hno>
brooc, github.com/linux-sunxi/ has our repositories. But not sure where the 3.7 tree is currently.
<brooc>
How so?
<mnemoc>
you don't have a sunxi device or know where the sources are, and want to start porting drivers...
* hno
is not pussled about that.
<techn>
berryboot installed for android.. lol :D
gzamboni has quit [Ping timeout: 252 seconds]
<mnemoc>
most people comes after having messed with a device in one or another way
<hno>
and some come to learn.
<mnemoc>
sure
<brooc>
Well, I posted on linkedin a post saying I would like to join an opensource group and was spointed here
<brooc>
I am trying to learn
<mnemoc>
brooc: aha! that was the missing link :)
<mnemoc>
brooc: now I understand
<brooc>
I have not heard of this project before
<brooc>
So I am working my way around this...
<mnemoc>
that was exactly my puzzling bit
<mnemoc>
the wiki is a good place to get familiar with the platform. and github.com/linux-sunxi/ for sources
<brooc>
I have built my own custom microchip based dev board with a WiFi device and Ethernet port
<techn>
.. waiting for cubie and christmas holidays ;)
<RaYmAn>
Seems like someone who should get a cubie sooner rather than later ;)
<brooc>
What do you think of the hackberry?
<hno>
prefer a board with some I/O extensibility like Cubieboard or OlinuXino
hg_5 has quit [Changing host]
hg_5 has joined #arm-netbook
<brooc>
Both the hackberry and the Cubieboard are very similar in functions to the Raspberry Pi
<Max_nl_>
What do you think of the hackberry? <- Have one, but like my Cubie better.
<Max_nl_>
One of the very basic but annoying issues I have with my hackberry is that connectors are close to each other and all on one side
gzamboni has joined #arm-netbook
<Max_nl_>
if you insert a wide HDMI to DVI convertor thingy it blocks a USB port and your network cable connection that are next to it
<brooc>
Not a good design...
<brooc>
I see that buying is not available for the Cubieboard at the moment
hg_5 has quit [Quit: ChatZilla 0.9.89-rdmsoft [XULRunner 1.9.0.17/2009122204]]
vinifm has quit [Read error: Connection timed out]
vinifm has joined #arm-netbook
hg_5 has joined #arm-netbook
<brooc>
What is the price difference between 1g and 512M?
<hno>
brooc, of cubieboard?
<brooc>
Yes
<mnemoc>
in the cubieboard? 512M was $4 cheaper in the first batch. second and third batches were 1GB-only
<brooc>
Do you know when the next batch will be?
<hno>
january/february some time.
<hno>
a guess
ibrah has quit [Ping timeout: 244 seconds]
<brooc>
So not in the next few weeks
<mnemoc>
a new campaign to collect funds in indiegogo might start in the next few weeks... but not for shipping this year
<hno>
dust need to settle a little from the current (third) batch first.
<brooc>
What do you mean?
<brooc>
Were ther problems?
<brooc>
there*
<brooc>
So there is no dev board that is selling in mass production?
<mnemoc>
the a13-olinuxino
<mnemoc>
and if you don't need extension pins, a mele
<hno>
brooc, there is always problems when doing something for the first time.
<mnemoc>
brooc: the money for the 3rd batch was just wired yesterday from indiegogo... they at least need to send these 1.5k pieces to the factory before thinking in the nex
grimes99 has joined #arm-netbook
<brooc>
mnemoc, will I be able to develope on an a13-olinuxino?
<mnemoc>
sure
<brooc>
Wouldn't buying an A13 core beat the point?
* hno
did most u-boot development on A13 OlinuXino
<hno>
brooc, A13 is the little sister of A10.
<hno>
almost the same, but very reduced I/O pins.
<brooc>
same architecture?
<hno>
yes.
<brooc>
What did they take out?
<hno>
same ARM Cortex A8 CPU core, mostly the same I/O blocks.
<hno>
Some 350 or so pins.
<hno>
less than A10.
freakazoid0223 has joined #arm-netbook
<hno>
What people notice most is that A13 do not have HDMI.
<brooc>
Why the reduction?
<mnemoc>
$$
<hno>
to compete in the market of very cheap tablets.
<hno>
slightly cheaper CPU. Much chaper board production thans to different CPU package format.
<hno>
for kernel & u-boot dev the A13 is very fine. For use as media box not.
<brooc>
Yeah
<brooc>
Well I have the RPi as media box
<hno>
Cubieboard combines the best of both. HDMI, SATA, LAN, penty of GPIO, 1GB, etc etc.
<brooc>
It really does
<brooc>
Wish I could buy one
<hno>
And A10 OlinuXino will be a perfect dev board when it arrives.
<hno>
having everything Cubieboard misses, save for small size.
<mnemoc>
it feels like tsvetan gave up on that one... due to the cubieboard
<hno>
He haven't.
<hno>
but not rushing it.
<mnemoc>
:)
<hno>
A13 OlinuXino MICRO is next.
<mnemoc>
got a prototype?
<hno>
hope to get one next week.
<mnemoc>
great
<hno>
hehopmajieh_ sent u-boot support for the micro today, cleaned and merged a while ago.
<brooc>
Is there anything I can do without an actual board?
<brooc>
Like booting on qemu?
<mnemoc>
brooc: you can learn about the platform completing A10/PIO and A13/PIO pages in the wiki :p
<hno>
brooc, you can't run anything that is Allwinner A1X specific in qemu.
<hno>
it lacks knowledge of any A1X hardware details.
ZaEarl has quit [Ping timeout: 252 seconds]
<brooc>
You said I should start with porting a driver like the I2C to linux 3.7, say I understand how to do that (which at the moment I don't), how will I test it?
<hno>
You could send it to linux-sunxi mailinglist for review & test.
<hno>
The sunxi serial ports do support DMA, but have not seen any in action.
<brooc>
I guess I'll have to come back once I get a board
<slapin>
hno: anything new?
* slapin
is all burned out with wavecom stuff
* slapin
hates modem telling it can't connect to GPRS at random, and sometimes can't connect half a day, while the same SIM card works perfectly with another one :(
<vinifm>
can I comment the lines about dma?
<slapin>
vinifm: DMA is good if you want to connect modem there