mnemoc changed the topic of #arm-netbook to: EOMA: Embedded Open Modular Architecture - Don't ask to ask. Just ask! - http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68 - ML arm-netbook@lists.phcomp.co.uk - Logs http://ibot.rikers.org/%23arm-netbook or http://irclog.whitequark.org/arm-netbook/ - http://rhombus-tech.net/
popolon has quit [Quit: Quitte]
Kraln has quit [Quit: No Ping reply in 180 seconds.]
Kraln has joined #arm-netbook
n6pfk has quit [Ping timeout: 245 seconds]
Triffid_Hunter has quit [Ping timeout: 245 seconds]
n6pfk has joined #arm-netbook
Triffid_Hunter has joined #arm-netbook
menomc has joined #arm-netbook
mnemoc has quit [Ping timeout: 245 seconds]
menomc is now known as mnemoc
rz2k has quit [Ping timeout: 245 seconds]
Manofthe1ea has joined #arm-netbook
rz2k has joined #arm-netbook
n6pfk has quit [Ping timeout: 245 seconds]
ManoftheSea has quit [Ping timeout: 245 seconds]
n6pfk has joined #arm-netbook
Manofthe1ea is now known as ManoftheSea
n6pfk has quit [Remote host closed the connection]
mikey_w has quit [Read error: Connection reset by peer]
ZaEarl has quit [Ping timeout: 244 seconds]
QingPei has joined #arm-netbook
Boulet has joined #arm-netbook
<Boulet> hello
<hno> hi hi
<Boulet> wow, still awake
<hno> not really
<Boulet> hehe
<hno> but some errors in test1 squashed, like parameters taken from A10 and not A13 and bus clock divisors wrongly set, but still no go on my board.
<Boulet> ah cool
<Boulet> i just replaced that &= by an =
<Boulet> you use the olimex board ?
<Boulet> all this works on my 4 different A13 PCBs
<hno> Well, test1 works, and writing/reading to DRAM works. But booting Linux fails.
<hno> https://github.com/hno/Allwinner-Info/commit/21b2809a87486f169b15795309cffdd0debd4917 is the changes made to clock setup by comparing with register rsults when u-boot is loaded from nand.
<Boulet> thanks
<Boulet> oh that DRAM_CCM_SDRAM_PLL_TUN_REG, i have no idea what it does
<hno> Neither do I, and the write does not seem to be effective either. The bits that differ are reserved according to the register file.
t0dbld|work has quit [Read error: Connection reset by peer]
<hno> Both AXI and APB0 clock divisors was wrongly set.
<Boulet> so the cpu runs at what frequency ?
<Boulet> 24*8*4/1/16 ?
<Boulet> oh wrong
<Boulet> sorry
<Boulet> 24*21*1/1/1
<Boulet> this calculation is wrong
<Boulet> seems like the CPU runs indeed at 1008 MHz after our PLL1 config
<Boulet> in the kernel mach-sun5i\clock\ccmu\pll_cfg_tbl.c
mikey_w has joined #arm-netbook
n6pfk has joined #arm-netbook
<Boulet> 21,1,0,0 corresponds to 1008MHz
<Boulet> i think the formula is 24*N*(k+1)/(M+1)/(P+1) ?
<hno> p=1/2/4/8
<hno> n=N
<hno> k=K+1
<hno> m=M+1
<Boulet> ah P can only be 1/2/4/8 ? 0->1 1->2 2->4 ... ?
<hno> yes
<Boulet> ok
<Boulet> and that K+1 explains why i saw 2x more than expected from the test1 comment
<Boulet> btw, i tried 1200MHz, seems fine too ;)
<hno> Not sure if it's really n=N. Documentatin says it is, but it's +1 on other PLLs.
<hno> but not all..
Boulet has quit [Ping timeout: 260 seconds]
<hno> Now time for bed.
Boulet has joined #arm-netbook
Kraln has quit [Quit: No Ping reply in 180 seconds.]
Kraln- has joined #arm-netbook
MMlosh has quit [Ping timeout: 260 seconds]
Gujs has quit [Ping timeout: 272 seconds]
MMlosh has joined #arm-netbook
hipboi has joined #arm-netbook
Gujs has joined #arm-netbook
lundman has quit [Ping timeout: 255 seconds]
lundman has joined #arm-netbook
<furan> whee, got my soldering microscope all set up
<hipboi> Turl: ping
<Turl> hipboi: :D thanks
<furan> how far are things from mali working on jellybean?
<Turl> furan: mali works
<furan> what doesn't?
<Turl> hipboi: you typoed android on the rar name :) minor nitpick :P
<Turl> on the folder inside the rar rather
<lundman> whats this
<Turl> lundman: latest mali props :)
<Turl> for android
<furan> so does this help us run jellybean?
<furan> I am JB obsessed
<lundman> havent even turned on a10 in 3 weeks, anything new to try
<Turl> furan: JB ran already with the old stuff
<Turl> lundman: didn't you use it as a NAS?
<furan> oh jeez, then I need to build a rom for it
<furan> turl sorry to bug you but did you start a repo?
<furan> I am about to solder a nand socket to an mk802 and upgrade the nand to different types to see if I have the way the nand table works worked out
<Turl> furan: yeah, just didn't make a local manifest for it yet though
<Turl> but it's all on the github org
ZaEarl has joined #arm-netbook
<furan> what do you mean by local manifest?
<furan> (I'm still working out android stuff)
<furan> nm googled
<Boulet> furan, interesting
<Boulet> do you have an external nand programmer also?
<furan> no, but i've gotten livesuit to the point of being able to flash it (have to modify the nand table in the files in the image that livesuit loads into memory to do the flash)
<furan> last time I tried this I just didn't have the values worked out, since they didn't match the datasheets
<furan> also I'm worried that since brom reads the first page it might have hardcoded nand support, in which case with new nand you might have to always boot from a u-boot sdcard that jumps to nand
<Boulet> what do you mean by "hard-coded" support ?
<furan> well, brom executes code that reads the first page of nand and then executes code on the nand
<Boulet> yeah
<furan> so that code in brom may have fixed support for nand, since nand can have different read commands (it generally doesn't, but it can)
<furan> e.g. the brom code may have its own table that it references to understand how to read the first page depending on nand id
<Boulet> there are some universal read commands though
<Boulet> that's possible yeah
<furan> are there? for some reason the nand driver keeps a table of 'architectures' with nand read/write/etc commands for each arch
<furan> I'm learning about talking to nand at the same time
<Boulet> i think to remember there are basic read commands that every nand support
<furan> that would be good
<Boulet> but then, there is ECC correction, and that is a different story
<Boulet> and also flashes can have different planes, can be SLC or MLC, blah blah blah
<furan> yeah.
<Boulet> but i think reading the first block should not be too difficult in an universal way
<furan> how big is the read? I suppose if it is smaller than the smallest page size it might work
<Boulet> ECC depends on the chip controller, so A10 will expect to find its own ECC scheme on the nand flash
<Boulet> i don't know
<furan> a10 nand driver keeps an 'ecc' field for each supported nand chip, I'm still confused about that one. might be num bits
<Boulet> i programmed a flash controller 5 years ago on a DSP chip, it was SLC NAND, 64 pages of 2048+64 bytes per page
<furan> cool
<furan> what is the +64, to account for bad blocks?
<furan> I noticed the page size+number in a lot of dataasheets
<furan> usually that number grew with the page size
<Boulet> those are used to store ECC values and other info that might be useful for your filesystem
<furan> I see
<Boulet> like logical to physical block translation
<furan> nods
<Boulet> some nand datasheet call this "spare", but on linux it's called oob data i think
<furan> ok
<Boulet> so when you read/write a page thru the NAND flash controller of A10, it should be able to calculate/apply ECC automatically
<Boulet> which is great, because doing ECC in software is quite costly
<Turl> hipboi: the kernel land changed quite a bit :P
<furan> nods
<hipboi> not that much
<furan> is there an mk802 android depot, or some way for me to just pull the device config for mk802?
<Boulet> you imagine, you have 2048 bytes, with a few extra bytes, you are able to locate and correct up to 8 (let's say) bits in those 2kb :) pretty cool
<Turl> hipboi: the diff is quite big, they refactored quite a bit of things
<hipboi> oh, yes
<furan> nods
<hipboi> btw, the rar is provided by my workmate
<Turl> hipboi: is he the one that sells devices with you?
<hipboi> no
<Turl> hipboi: mali did proper kernel makefiles/kbuild/kconfig this time apparently :)
<Turl> seems to work hipboi
ZaEarl has quit [Ping timeout: 244 seconds]
ibrah has quit [Ping timeout: 268 seconds]
ZaEarl has joined #arm-netbook
robws has quit [Ping timeout: 255 seconds]
robws has joined #arm-netbook
RITRedbeard has quit [Ping timeout: 272 seconds]
<rm> looks like my SD card in the MK802 is dying
<rm> it's got a lot of bad sectors
<rm> and some libs were unreadable