Turl changed the topic of #linux-sunxi to: Allwinner/sunxi /development discussion - did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait! - https://github.com/linux-sunxi/ - Logs at http://irclog.whitequark.org/linux-sunxi
naobsd has joined #linux-sunxi
flyhorse_cvVjP has joined #linux-sunxi
flyhorse has quit [Read error: Connection reset by peer]
Akagi201 has joined #linux-sunxi
lemonzest has quit [Quit: Leaving]
AnticitizenOne has joined #linux-sunxi
AnticitizenOne has quit [Client Quit]
ninolein has joined #linux-sunxi
ninolein_ has quit [Ping timeout: 244 seconds]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
kaspter has joined #linux-sunxi
Akagi201_ has joined #linux-sunxi
Akagi201 has quit [Ping timeout: 240 seconds]
p1u3sch1 has quit [Ping timeout: 264 seconds]
p1u3sch1 has joined #linux-sunxi
Akagi201 has joined #linux-sunxi
Akagi201_ has quit [Ping timeout: 255 seconds]
TheSeven has quit [Ping timeout: 265 seconds]
TheSeven has joined #linux-sunxi
p1u3sch1 has quit [Ping timeout: 256 seconds]
p1u3sch1 has joined #linux-sunxi
naobsd has quit [Ping timeout: 250 seconds]
JohnDoe_71Rus has joined #linux-sunxi
gzamboni has quit [Ping timeout: 255 seconds]
gzamboni has joined #linux-sunxi
reinforce has joined #linux-sunxi
HeHoPMaJIeH has joined #linux-sunxi
simosx has joined #linux-sunxi
kaspter has quit [Remote host closed the connection]
kaspter has joined #linux-sunxi
domidumont has joined #linux-sunxi
sehraf has joined #linux-sunxi
domidumont has quit [Remote host closed the connection]
domidumont has joined #linux-sunxi
kaspter has quit [Remote host closed the connection]
lemonzest has joined #linux-sunxi
dothebart has quit [Remote host closed the connection]
reinforce has quit [Quit: Leaving.]
_massi has joined #linux-sunxi
reinforce has joined #linux-sunxi
hansg has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
ricardocrudo has joined #linux-sunxi
<hansg> RSpliet, moving the sunxi talk here :) Note that in my kernel sunxi-wip branch I also have a patch to get the ecc strength / size from the extended nand id, so that we do not need to add an entry for each nand to the nand-ids table. This is for Samsung only for now, once I get things working on the A10 I'll start looking into hynix nand support.
<RSpliet> hansg: from what I can tell, parsing the ID is pretty standard
<RSpliet> the NAND framework already has similar support
<RSpliet> also, ECC strength in the ID table only specifies a recommended minimum strength, OBB size dictates a maximum, the user gets to choose everything between no ECC and whatever fits physically
<RSpliet> BROM places some additional constraints :-P
_massi has quit [Remote host closed the connection]
<hansg> AFAICT the id parsing is vendor specific, at least the kernel has specific code for each vendor (for things like determining the oobsiize
<hansg> I know the strength from the nand id is only a minimum, but it seems that sofar that minimum is what the sunxi code is using, and it was using it by harcoding it per nand chip, so it seemed better to get it from the id :)
<hansg> I'm worried a bit about the BROM constraints, if we ever encounter a nand where the id recommended ecc settings are not supported by the BROM we will need to teach the sunxi kernel code to pick the first stronger ecc which is supported
<RSpliet> sunxi-kernel is using DT specifications I think
prz has joined #linux-sunxi
<RSpliet> or... Boris' driver is :-)
<a1d3s> hansg your sound codec in sunxi-wip works on bananapro also
_massi has joined #linux-sunxi
<jackdaniel> o/
<hansg> RSpliet, it allows setting the ecc strength / size in dts that is through, but for the Cubietruck Boris is relying on this (in sunxi-nand-next): https://github.com/bbrezillon/linux-sunxi/commit/355dc1fd9d5341ff5389e4df00be2bf254e9043e
<hansg> And I personally believe that auto-detecting settings is better then relying on people to specify things manually, be it in nand-ids.c or in dts, so I've written this:
<hansg> Down side of my solution is that it may come up with settings which are not supported by the BROM, but so far that is not a problem, and for those cases we can always override things in dts (or teach the kernel to pick the next strongest setting which is supported)
<jackdaniel> hansg: to not use old kernel - your sunxi-wip branch https://github.com/jwrdegoede/linux-sunxi/tree/sunxi-wip is capable of writing mtd0 so brom reads it?
<hansg> jackdaniel, yes (on cubieboard2 and cubietruck only atm)
<RSpliet> hansg: hmm... that does not correspond with the Hynix chip I've seen
<jackdaniel> I have only cubieboard2, so not a problem for me
<hansg> RSpliet, right this is samsung only, it is in an existing samsung only block
<hansg> As I said id decoding is not as straight forward as I would like it to be ...
<hansg> s/block/code block/
<RSpliet> meh... well, maybe we can turn this into a hashtable with {strength, step-size} pairs, so it can be easily interchanged with other vendors without having to re-write switch-cases :-)
<hansg> We'll see. I'm all in favor of baby steps here, for now we can set the ecc settings in dts for hynix chips
<RSpliet> my primary concern now is U-boot, so I tend to agree :-)
<orly_owl> why not barebox? (:
<orly_owl> #barebox
<jackdaniel> may I flip a question - why barebox?
<bbrezillon> hansg, RSpliet, jackdaniel: yep, the whole chip capabilities detection is a pain, I thought about adding vendor specific init functions (stored in drivers/mtd/nand/<vendor>_nand.c files) to fill the nand_chip struct appropriately
<jackdaniel> hansg: is this branch by accident support also audio?
<hansg> jackdaniel, yes
<a1d3s> jackdaniel, and it works ;)
<jackdaniel> that's great :-) that should save me from mangling kernels
<a1d3s> i copied the &codec thing from cubie to banana and it works :)
<hansg> Note if any you make any dts additions to enable audio on more boards I would welcome (clean, with S-o-b) patches to merge into my sunxi-wip branch. mripard is currently working on upstreaming the audio bits
<a1d3s> first time i forgot to enable dma
<jackdaniel> S-o-b?
<hansg> Oh and my sunxi-wip branch also has usb-otg support (on some boards)
<jackdaniel> what does S-o-b stand for?
<RSpliet> hansg: did I spot correctly that the A83t uses the "old" CCU register definitions again, even though the A80 (both sun9i) has a fresh new register lay-out?
<hansg> And then specifically the section titled "11) Sign your work", note this is not gpg signing it is just a single line in your commit message, all very easy todo.
<jackdaniel> hansg: yes, I know it since your first mail on sunxi ml, just couldn't resolve acronym
<hansg> RSpliet, I've a mail with which block comes from what for the A83, let me check.
<RSpliet> hansg: I observed this in the A83t user manual
<jackdaniel> RSpliet: I'll start work from your v3
<RSpliet> jackdaniel: on what?
<RSpliet> would it help if I obtain push rights to your gitlab tree?
<jackdaniel> patch v3
<jackdaniel> yes, do you have gitlab account?
<RSpliet> I... tried making one a few days ago
awe00 has joined #linux-sunxi
<hansg> RSpliet, the mail I have says that the CCU is more like the A80 though ..., so we likely need to compare every register and may need to define a whole new clock_sun9i_a83.[c|h] for this
<RSpliet> and failed because I typo'd my e-mail addrsss
<RSpliet> hansg: I was afraid you were going to say that... I think it corresponds pretty well with the sun4i definition
<RSpliet> unless the A83t manual is just plain wrong
<hansg> RSpliet, if we can make the sun4i files work with a few #ifdefs that is fine with me. Have you checked the clk files in the released kernel sources ?
<jackdaniel> RSpliet: then please send me your public key, gitlab has "deploy key" option and I think it's about push privigiles for people from outside of gitlab
<RSpliet> jackdaniel: I got my account fixed now :-)
<RSpliet> just add the user UltiSpliet
<jackdaniel> ok, second
awe00 has quit [Quit: WeeChat 1.2]
<wens> hansg: i must say you are working on a lot of things :)
<RSpliet> hansg: not yet... how well is A83 supported there?
<RSpliet> wens: he just joined the nouveau team too :-D
<wens> does anyone even have A83 hardware yet?
<jackdaniel> RSpliet: should work now
<a1d3s> wens i have a31s hardware
<RSpliet> jackdaniel: ok, here comes the current work
<wens> RSpliet: wow
<jackdaniel> that's nice, was going to apply patches from ml by hand :)
<RSpliet> give me a bit of time, it's based on the sunxi custodians tree from a week ago or so, and internet here is slow as mad
<hansg> RSpliet, I mean the kernel tree here: https://github.com/allwinner-zh/linux-3.4-sunxi/tree/A83T
<RSpliet> hansg: thanks
<jackdaniel> sure, are you going to rebase it?
awe00 has joined #linux-sunxi
<hansg> Those are the official allwinner A83 kernel sources, so when in doubt about the user-manual those should be right (although we have had cases where the user-manual actually was correct and the kernel code worked by accident)(
dothebart has joined #linux-sunxi
<RSpliet> jackdaniel: it's a simple push --force
<jackdaniel> sure
<jackdaniel> s/sure/ok/
<dothebart> re.
<jackdaniel> o/
<RSpliet> hansg: thanks, I'll see if I can find the defines in there
iamfrankenstein has quit [Quit: iamfrankenstein]
<RSpliet> hansg: oh got it, there's just a minor mistake on the linux-sunxi wiki... A83t is a Sun8i
<hansg> Right, that makes sense (soemwhat) since it only uses Cortex A7 cores
<RSpliet> for as much as it can make sense yes :-)
<hansg> So it should probably be using clock_sun6i.[c|h]
<RSpliet> it does :-)
<jackdaniel> is powering cubieboard2 with mini-USB a bug or a feature?
<RSpliet> I'd like to find out if we can change some variable names in the A80 just enough to make the NAND driver work without ifdefs
<sunxi_fan> hansg: as we were talking about sunxi-wip audio support (last commit four days ago..), i'm wondering if you think the porting of I2S linux-sunxi 3.4 driver can be brought on top of that commit or if it's a completely a different task. can you give me an educated guess?
<hansg> jackdaniel, both, if there were to be done via the axp209 vbus input it is a feature, but it seems that the 5v of the plug is directly connected to the 5v of the board, meaning that we cannot not supply v5 when we've another power supply and we could end up tying the 5v out of an external adapter to that of the pc the cubieboard is plugged into
<hansg> jackdaniel, this also makes it hard to do otg reliable on the cubieboard
<sunxi_fan> i happen to have an olimex lime and i've got a dirty cheap I2S external board with UDA1380 I2S codec (from dx.com..), i'd like to use it as an educational step for introducing into the wonders of digital audio! :-)
<hansg> sunxi_fan, i2c is a different functional block then the build in dac, so it would mostly be a new driver, but the dma engine code also in sunxi-wip certainly helps a lot and you can use the internal dac driver to see how to best port the i2s driver from sunxi-3.4 to mainline
<RSpliet> jackdaniel: pushed, but beware I might overwrite changes again
<jackdaniel> hm, I think it'd be good to have just sunxi-nand-v3 branch, and sunxi-nand-wip (where stuff might change)
<jackdaniel> in sense of history°
<RSpliet> well, v3 is superseded... so well, perhaps sunxi-nand-wip might have been good
<RSpliet> I'll go and push there then
<jackdaniel> great, thanks :-)
<hansg> RSpliet, v3 is superseeded? What has changed? Also do you've a link to the gitlab repo you're usign ?
<sunxi_fan> hansg: thanx.. i'll give a look and test your kernel tree and see if the i2s porting is feasible (for my skill level, i mean..)
<sunxi_fan> hansg: is the DmaEngine patch already folded in your tree?
<mripard> sunxi_fan: it's also something on my todo list, but if you have some i2s codec (which I don't yet), go ahead :)
<RSpliet> I changed some bits in the various clock definitions, and made the nand_init explicitly enable DMA too (plus using the right defines there)
<jackdaniel> RSpliet: so we'll work on sunxi-nand-* branches, ok? I'd like to keep master synced with u-boot upstream as a mirror
<RSpliet> sunxi-nand-wip it is :-)
<jackdaniel> thanks, I'll re-push master then
<sunxi_fan> mripard: thanx i'll give a look at the MOD patch (is that just a "rework" of the "traditional" linux-sunxi 3.4?); BTW i've got a couple "codec" from dx.com just for fun. one is the "typical" raspi SPDIF extension and than there's this UDA1380 analog. both are I2S for audio stream and I2C bus for controlling/configuring the audio chip..
<mripard> sunxi_fan: yeah, it sounds pretty standard
<mripard> and yes, it's a 3.4-based driver
<mripard> so with a lot of useless stuff...
<mripard> one thing you need to do first is probably to make sure there's a driver for your codec already
<mripard> so that you know that it works, and can focus on the i2s driver itself
<sunxi_fan> mripard: yes, now i was testing if the codecs were working. with A20 I2C driver i can already "configure" pretty easily the audio chip, with i2cset, for example..
<sunxi_fan> mripard: then i was wondering if i could "drive" the I2S module straight from /dev/mem, memory mapping the right IO space and reading/writing to registers..
<mripard> sunxi_fan: that's not what I meant.
<mripard> what I meant was that you should check first that there is already a driver written for that codec
<mripard> that will configure the codec properly
<sunxi_fan> mripard: i was testing with the typical call: i2s_regs = mmap(0, PageSize*2, PROT_READ|PROT_WRITE, MAP_SHARED, fd, i2s_start);
<sunxi_fan> after having opened /dev/mem with: fd = open("/dev/mem", O_RDWR);
<sunxi_fan> ..that i'm using to accessing the PIO registers (where it works..) but here if i dump the first 0x100 32 bit regs, i get back always a 0x00
<mripard> and you can't really do that
<mripard> you have no way to set the clock rates from userspace, trigger DMA transfers, get an interrupt, etc. using /dev/mem
<mripard> UIO can help for the interrupts, but you'll still have the clocks and DMA issues
<mripard> and what you're seeing is precisely because the clocks are not enabled.
<RSpliet> hansg: I'll send the (final) V4 patches to the mailing lists as soon as we figured out your timeouts btw...
<hansg> RSpliet, ok, currently I'm working on $dayjob stuff (nouveau) I'll look into the timeouts tonight.
<RSpliet> thanks!
iamfrankenstein has joined #linux-sunxi
<RSpliet> the u-boot tree I linked you explicitly enables DMA in the CCU, I assume it doesn't, but it might help
<sunxi_fan> mripard: i understand the "long term" solution is really a proper setup into the audio ASOC subsystem, but this was just my approach to become more confident that everything was under control.. i was trying to check at least the MCLK was getting out of the proper pin of the olimex board using my trusty Openbench logic analyzer.. that's the point where the fun is.. :-)
<sunxi_fan> moreover the sunxi do have pretty deep FIFO on I2S tx so i was supposing i could feed that TX fifo in a "timed" way just to hear something on the earphone jack. 128 bytes at 48KHz are 1msec if i'm not mistaken..
<sunxi_fan> BTW the UDA1380 codec is already there: http://lxr.free-electrons.com/source/sound/soc/codecs/uda1380.c
<mripard> sunxi_fan: this is not the long term solution
<mripard> it's the only viable solution.
<sunxi_fan> mripard: just an hint; as far as i can understand, the UDA1380 codec is present in .../soc/codecs as it's used in a HTC PDA called Magician, ref:
<sunxi_fan> do you think the use case of the codec as an optional addon of A20 based card (and so of eventuall many other "devel cards" with I2S/I2C busses), is going to impact someway the uda1380.c and the magician.c files?
<sunxi_fan> i'm asking because i'm trying to make myself more clear about the steps to get to some "audible" sign of life (without breaking previous working setups..) .-)
prz has quit [Ping timeout: 264 seconds]
simosx has quit [Quit: Leaving]
Andy-D has quit [Ping timeout: 256 seconds]
FR^2 has joined #linux-sunxi
Andy-D has joined #linux-sunxi
arossdotme has quit [Quit: Ex-Chat]
afaerber_ has quit [Quit: Verlassend]
hansg has quit [Quit: Leaving]
arossdotme has joined #linux-sunxi
ricardocrudo has quit [Ping timeout: 240 seconds]
afaerber has joined #linux-sunxi
whirm has joined #linux-sunxi
<whirm> hi all
<whirm> Is it better to use upstream's u-boot or sunxi's to boot an mainline kernel (4.0) in a cubieboard 2?
<whirm> In the wiki it says I shouldn't use an old version but It doesn't say what an old version is nor if the sunxi fork is considered old
<jackdaniel> whirm: use upstream u-boot
<whirm> jackdaniel: Ok, I'll update the recipes then. Thanks!
<whirm> jackdaniel: is that USB1+2 bug mentioned in the wiki still present?
<jackdaniel> I don't know
<whirm> ok, thanks!
da3m0n22 has joined #linux-sunxi
<jackdaniel> :)
<da3m0n22> hello, dealing with HDMI problem, display is correctly set up by uEnv.txt/script.bin and after fbset/xinit/apps work as intended, however after 10-20 minutes HDMI output stops, and cannot be turned on again. It's not blanking (display is not just blank, but shows "no signal" error message), and dpms is set off, cannot be woken by usb mouse/keyboard/...
<da3m0n22> no error in dmesg, syslog, "mali" module loaded on boot
<da3m0n22> also ssh on given machine still works, i'm just not able to restart X server (dies without error message) on display :0 (hdmi, framebuffer display1), using Bananian 15.04 r01
<TheLinuxBug> which board
kaspter has joined #linux-sunxi
<TheLinuxBug> I had same issue on BPi with some kernels
<TheLinuxBug> I would say find a different kernel
<TheLinuxBug> when I would see issues it would be high CPU usage right before and then HDMI would shut off like your suggesting
<TheLinuxBug> Was using Igor's kernels and kept having that issue on the BPi
<TheLinuxBug> went to an arch dist and kernel and its been much more stable
<da3m0n22> TheLinuxBug: I guess standard BananaPi, how do i find out? i'm using currently 3.4.104-bananian
<TheLinuxBug> I have had good luck with that dist
<TheLinuxBug> if your not afraid of Arch
<TheLinuxBug> hes built a very stable kernel with all the video support
<TheLinuxBug> and has precompiled and setup smplayer and XBMC with hardware accelerated drivers
<da3m0n22> TheLinuxBug: i'll watch for cpu usage, however I'm only running image slideshow through EOG/QIV, nothing CPU or GPU intensive, just on large flat screen in D:1920x1080p-60 mode
<TheLinuxBug> yeah all I know is I would see the same thing after a period of time when CPU was under some load, all sudden hdmi would die, hdmi_proc would remain running but no output on the scrreen
<da3m0n22> TheLinuxBug: cool, on Arch this doesn't happen at all or it's more rare?
<TheLinuxBug> I haven't seen it in my use -- thats not to say it can't/won't happen
kaspter has quit [Remote host closed the connection]
<TheLinuxBug> but I have used it for a good period of time without issue
<TheLinuxBug> (4-5 hours)
<da3m0n22> TheLinuxBug: great, i'll give it a chance, thank you
<TheLinuxBug> Though they say these things can handle a reasonable ambient temp I think that the main issue is heat for it.. for some reason it may hit a heat point that causes the chip to stop till full reset
Andy-D has quit [Ping timeout: 276 seconds]
<TheLinuxBug> an some kernels don't know how to handle the event
ricardocrudo has joined #linux-sunxi
kaspter has joined #linux-sunxi
<mripard> sunxi_fan: no, it shouldn't
<mripard> and the whole point is that you can even assume it just works
Akagi201 has quit [Remote host closed the connection]
kaspter has quit [Remote host closed the connection]
Kasreyn has joined #linux-sunxi
<sunxi_fan> mripard: thanx for the reassuring words! :-) it's becoming a compelling experience, given all the stuff already in place.. :-)
domidumont has quit [Read error: Connection reset by peer]
Andy-D has joined #linux-sunxi
<mripard> sunxi_fan: the only thing that really needs some work will be the i2s driver itself
Black_Horseman has quit [Read error: Connection reset by peer]
<Kasreyn> hello.. i got a BPi running Bananian (so i know the hardware works).. however.. nothing but red LED following "Installing from an SD card image" instructions @ https://wiki.debian.org/InstallingDebianOn/Allwinner .. anyone else have this problem?
<mripard> and modifying your board DT to use simple-card probably, and do the link between the i2s controller and the codec
<mripard> Kasreyn: do you have something out of the UART? or really nothing ?
<Kasreyn> mripard: if i had it here with me i could check that right away.. but i don't.. it only lit up the RED led as far as i could tell..
<Kasreyn> i can try again later on today
<Kasreyn> they are daily builds after all?
<mripard> Kasreyn: by default, we don't have any display support in the kernel
<mripard> we use simplefb, but that requires some cooperation from u-boot
<Kasreyn> i'm ok with that.. will use it as a server :)
<mripard> so if you didn't update u-boot, or use a rather old version, possibly, you will see nothing
<Kasreyn> ah ok thanks :)
flyhorse has joined #linux-sunxi
<Kasreyn> i'll hook it up to USB UART adapter later on today and check
flyhorse_cvVjP has quit [Read error: Connection reset by peer]
<markvandenborre> any way to boot the orange pi plus (Allwinner H3 based) from eMMC?
<sunxi_fan> mripard: "simple-card"? i suppose you are talking about this stuff: https://lkml.org/lkml/2015/1/1/12
<sunxi_fan> never heard about it.. :-) it's not so infrequent.. in this domain. do you think it's a matter of multiple codecs linked on the same DAI (DA0) in my A20, i suppose..
<sunxi_fan> looking at "6.9.2. Digital Audio Interface Block Diagram" of A20 user man, i think the internal analog codec and the I2S are not sharing hardware AFAICS..
<wens> nope
naobsd has joined #linux-sunxi
a1d3s_mobile has joined #linux-sunxi
<mripard> sunxi_fan: this patch modifies simple-card
<mripard> to handle multiple DAIs
<mripard> but you don't have to
<mripard> it can be used if you have a single DAI connected to a single codec
<mripard> which seems to be your use case right now.
<sunxi_fan> mripard: i see... i'm going to take some time studying and testing this stuff. i'll come back when i'll have some result to share or just more questions.. thanx for the time
a1d3s_mobile has quit [Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )]
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
prz has joined #linux-sunxi
<a1d3s> Kasreyn, did ssh works?
bjoernb has left #linux-sunxi [#linux-sunxi]
diego_r has joined #linux-sunxi
<Kasreyn> a1d3s: i only checked HDMI output + red/green LEDs.. with no light on green at all..
<Kasreyn> difficult to check SSH without knowing the IP.. but as i said.. i will continue later on.. see if it prints anything to debug tty
<markvandenborre> Kasreyn: nmap is your friend...
<markvandenborre> nmap -sP <your ip>/24
<Kasreyn> i rather not nmap range some public ips ;) (we get external IPs from our DHCP server)
<markvandenborre> yeah, in that case :-)
<markvandenborre> maybe you can ask the dhcp server very nicely?
<markvandenborre> :p
<Kasreyn> yes.. it's silly when i keep a server running with the same IP 24/7
<markvandenborre> I guess you might have access to its logs...
<mripard> arpwatch ? :)
<Kasreyn> aha.. that's a tool i havn't used before
reinforce has quit [Quit: Leaving.]
reinforce has joined #linux-sunxi
romain_____ has joined #linux-sunxi
wickwire_ has joined #linux-sunxi
wickwire__ has joined #linux-sunxi
<romain_____> Hi, I'm wondering where the L2 cache get initialized during Linux bootstrap on a A20 SoC ? My understanding so far is that the Cortex-A7 (http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0503h/CHDCBBIE.html) use a integrated cache controller, no more an external one like the pl310, etc. Therefore the cache controller is no more described in the dts. I can't see any infos about L2 during boot messages. I need to
wickwire has quit [Ping timeout: 258 seconds]
<TheLinuxBug> sup rm
<mripard> (so much suspense, it's killing me.... :))
<TheLinuxBug> cd ..
wickwire_ has quit [Ping timeout: 255 seconds]
<mripard> romain_____: your message was truncated.
<romain_____> oh, at which point?
<mripard> "L2 during boot messages. I need to"
<mripard> and then nothing
<mripard> :)
<romain_____> ok thanks, here is the end : [...] L2 during boot messages. I need to split L2 cache between two cores, thanks!
kaspter has joined #linux-sunxi
<mripard> IIRC, the L2 cache is enabled as soon as the MMU is
<mripard> I'm not really sure to what extent you can control it
<romain_____> So the only control I may have is by tuning some SCU registers?
kaspter has quit [Remote host closed the connection]
kaspter has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
gianMOD has joined #linux-sunxi
<mripard> romain_____: probably yeah
Akagi201 has joined #linux-sunxi
<romain_____> @mripard : too bad, but thanks for your help!
jinzo has joined #linux-sunxi
<mripard> romain_____: to be honest, I never looked very much into it since it was just working
<mripard> so I might very well be wrong on this
kaspter has quit [Remote host closed the connection]
<romain_____> but this sound correct as the A7 has now an architected cache controller coupled with the SCU (according to this http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0464e/BABJECJI.html)
<wens> l2 is enabled as part of the boot process i think
<romain_____> But even with the SCU on top of it?
<romain_____> I have to admit that I'm a bit lost with SCU , l2 cache controller integrated/not integrated
<RSpliet> romain____: risking to sound silly, are you sure you want to do that? You'll likely lose coherence
<romain_____> the arm documentation writes about the A7 : "Optional tightly-coupled L2 cache that includes"
<romain_____> can't figure out if it integrated by sunxi as the a20 manual is not very clear about it
<romain_____> RSpliet: yes indeed but the goal is to run one Linux on each core without any interactions between them
<romain_____> therefore i'd like to give half the L2 cache size to one Linux and the other half to the second Linux
TheSeven has quit [Quit: No Ping reply in 180 seconds.]
TheSeven has joined #linux-sunxi
<wens> how would you share peripherals and memory?
<mripard> you don't
<mripard> you split the memory in half, and dedicate peripherals to one instance or the other
<mripard> it's something "commonly" used when you want to make a RT system
<mripard> it would be easier to do that with the A31+ hwspinlocks though
<mripard> since you can share the devices between the two OSes
<wens> hmm...
domidumont has joined #linux-sunxi
<jackdaniel> ls
<lukas2511> rm -rf /
<jackdaniel> /quit
<romain_____> we use paravirualization technics do it and it's already running
<romain_____> but we'd like to split the L2 cache to improve performances
<romain_____> if you're interested it's called EmbX (formerly EmbeddedXen) : http://en.wikipedia.org/wiki/EmbeddedXEN
<romain_____> but everything on the internet is outdated, we will soon release a new version (with the new name :-p)
romain__1 has joined #linux-sunxi
leviathanch has quit [Ping timeout: 265 seconds]
leviathanch has joined #linux-sunxi
romain__1 has quit [Ping timeout: 246 seconds]
kaspter has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
HeHoPMaJIeH has quit [Quit: Konversation terminated!]
JohnDoe_71Rus has joined #linux-sunxi
prz has quit [Ping timeout: 264 seconds]
blsd has quit [Quit: Leaving]
<RSpliet> ijc: I typo'd your e-mail address in the patches I just sent you, sorry. I don't think it justifies a re-send, but please check patchwork for my "final" sunxi nand spl patches
<jackdaniel> RSpliet: does kernel from Hans detect ecc for you, or you specify it explicitly in dts?
<ijc> RSpliet: Actually, you had it twice, once wrongly in To: and once correctly in Cc: , so I did get it
<ijc> Actually, just the first one was CCd. Agreed wrt not resending though
<ijc> I'll dig it out of the list
Gerwin_J has joined #linux-sunxi
domidumont has quit [Read error: Connection reset by peer]
gzamboni has quit [Ping timeout: 250 seconds]
adj_ has quit [Quit: Leaving]
<RSpliet> ijc: I'm guessing git added you as well
<RSpliet> jackdaniel: no, I've been explicit about my ECC settings
gianMOD has quit [Remote host closed the connection]
gianMOD has joined #linux-sunxi
gzamboni has joined #linux-sunxi
gianMOD has quit [Ping timeout: 264 seconds]
imcsk8 has quit [Remote host closed the connection]
iamfrankenstein has quit [Quit: iamfrankenstein]
gzamboni has quit [Ping timeout: 256 seconds]
reinforce has joined #linux-sunxi
Akagi201 has quit [Remote host closed the connection]
gzamboni has joined #linux-sunxi
TheLinuxBug has quit [Ping timeout: 256 seconds]
TheLinuxBug has joined #linux-sunxi
bonbons has joined #linux-sunxi
leviathanch has quit [Ping timeout: 265 seconds]
khuey|away is now known as khuey
gianMOD has joined #linux-sunxi
leviathanch has joined #linux-sunxi
gianMOD has quit [Ping timeout: 265 seconds]
awe00 has quit [Ping timeout: 265 seconds]
ricardocrudo has quit [Remote host closed the connection]
Akagi201 has joined #linux-sunxi
kaspter has quit [Ping timeout: 246 seconds]
diego_r has quit [Ping timeout: 264 seconds]
Akagi201 has quit [Ping timeout: 265 seconds]
_massi has quit [Quit: Leaving]
HeavyMetal has quit [Ping timeout: 256 seconds]
HeavyMetal has joined #linux-sunxi
gianMOD has joined #linux-sunxi
dev1990 has joined #linux-sunxi
gianMOD has quit [Ping timeout: 264 seconds]
khuey is now known as khuey|away
<Kasreyn> BPi fails to boot u-boot (tty debug) with the SD card i created "zcat firmware.<board_name>.img.gz partition.img.gz > /dev/SDCARD_DEVICE" method
afaerber has quit [Quit: Verlassend]
<Kasreyn> (i used the same SD card)
<Kasreyn> either the instructions or the images i downloaded.. something must be wrong
<mripard> I don't know who wrote these instructions
<mripard> ijc maybe ?
<Kasreyn> well.. installing debian netboot style from SD card would be pretty sweet.. that's what everyone wants, right? :)
RSpliet has left #linux-sunxi [#linux-sunxi]
<a1d3s> Kasreyn, do you need a server or desktop? pirojects.net there are some images who work
<Kasreyn> i don't even see the partition on the computer i use to create the SD card with "zcat firmware.<board_name>.img.gz partition.img.gz > /dev/SDCARD_DEVICE"
<Kasreyn> but i do get a working u-boot from "zcat bananapi.sdcard.img.gz > /dev/sdb"..
<Kasreyn> a1d3s: i need server.. i have bananian running on it.. just exploring my options right now
<Kasreyn> and well.. reporting on potential mistakes in online documentation..
domidumont has joined #linux-sunxi
DagoRed has quit [Ping timeout: 272 seconds]
<ijc> mripard: I wrote a fair bit ofthat page, but not he sdcard stuff, chap called Karsten Merker did that, but he seems to not be on the channel right now.
<ijc> Kasreyn: Perhaps try debian-arm@lists.debian.org?
Akagi201 has joined #linux-sunxi
<Kasreyn> ok ijc :)
<ijc> Kasreyn: Actually first, try the images from http://ftp.uk.debian.org/debian/dists/jessie/main/installer-armhf/current/images/ rather than the daily builds, they can be in flux sometimes
<ijc> You most likely want jessie anyway, but if you want sid you can just upgrade right after installing
<Kasreyn> aha ok
Netlynx has joined #linux-sunxi
* ijc updates the wiki assuming that to be good advice...
Akagi201 has quit [Ping timeout: 244 seconds]
<Kasreyn> i'll download those images and try again in a minute..
ricardocrudo has joined #linux-sunxi
khuey|away is now known as khuey
awe00 has joined #linux-sunxi
RSpliet has joined #linux-sunxi
<a1d3s> Kasreyn, if u have banananien then you have only dd bs=4M if=x.img of=/dev/SDCARD
<a1d3s> Kasreyn, all images on lemaker.org page have working u-boot included
DagoRed has joined #linux-sunxi
<a1d3s> Kasreyn, there is no need to extract it and then write u-boot without the whole OS
gianMOD has joined #linux-sunxi
<Kasreyn> not sure why there's "Error reading cluster" reading vmlinuz..
gianMOD_ has joined #linux-sunxi
<Kasreyn> i assume it needs vmlinuz to start the installer..
gianMOD has quit [Read error: Connection reset by peer]
<a1d3s> Kasreyn, yes
<a1d3s> Kasreyn, but you can use the bananien image with : dd bs=4M if=bananien.img of=/dev/SDCARD
<Kasreyn> yes.. but i want to use the debian installer ;)
<Kasreyn> maybe you can start the installer from bananian..?
orly_owl has quit [Ping timeout: 256 seconds]
JannF has joined #linux-sunxi
<JannF> hello people :)
<lemonzest> hello
<JannF> Is there someone who can help me port a new device?
<JannF> I will finish the wiki page finally sometime this night
<JannF> It's this device
<JannF> I had problem some time ago, because I was not able to go to FEL and NAND was inaccesible.
<JannF> *problems
ricardocrudo has quit [Ping timeout: 256 seconds]
<JannF> now I finally have a script.bin and already extracted the script.fex
<JannF> I want to build a uboot for myself now to verify it's all working and to document everything.
<Kasreyn> aha a tablet :)
<JannF> yes, I got it from a friend and it was dead then.
<JannF> now I found out that when I send "1" over UART ist would mount the system partition and data from android and expose it over USB
<JannF> so I was able to get the script.bin and found out that sending "2" over UART would lead me to FEL
<JannF> unfortunately this was the wrong boot1, so I tried several button combinations and found that the "J" key combined with the power key would lead to the right FEL and I was able to flash a partly functional LiveSUit Image for a similar Image.
domidumont has quit [Ping timeout: 246 seconds]
<JannF> now I cloned the the u-boot-sunxi repo and don't know how to create a defconfig or if there is a script to create one from the scrip.fex
<JannF> that's the point where I need your help ^^ :>
JohnDoe_71Rus has quit [Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/]
Netlynx has quit [Quit: Leaving]
gianMOD_ has quit [Remote host closed the connection]
gianMOD has joined #linux-sunxi
Akagi201 has joined #linux-sunxi
gianMOD has quit [Ping timeout: 244 seconds]
Akagi201 has quit [Ping timeout: 250 seconds]
gianMOD has joined #linux-sunxi
<JannF> no one to help out there?
orly_owl has joined #linux-sunxi
gianMOD has quit [Remote host closed the connection]
reinforce has quit [Quit: Leaving.]
<whirm> JannF: I'm a total newbie but maybe: https://linux-sunxi.org/Sunxi-tools#bin2fex
ricardocrudo has joined #linux-sunxi
<JannF> whirm: that's the way I've already done, it converts the script.bin to a script.fex (which is readable and not a binary)
physis has quit [Read error: Connection reset by peer]
<whirm> ah, sorry I didn't read it right :) no idea then :)
physis has joined #linux-sunxi
hansg has joined #linux-sunxi
FR^2 has quit [Quit: Connection reset by peer]
kaplan2539 has joined #linux-sunxi
kaplan2539 has left #linux-sunxi [#linux-sunxi]
dev1990 has quit [Quit: Konversation terminated!]
Black_Horseman has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
Gerwin_J has joined #linux-sunxi
kaplan2539 has joined #linux-sunxi
kaplan2539 has left #linux-sunxi [#linux-sunxi]
physis has quit [Read error: Connection reset by peer]
physis has joined #linux-sunxi
kaplan2539 has joined #linux-sunxi
bonbons has quit [Quit: Leaving]
<hansg> Hi kaplan2539
<kaplan2539> Hi hansg
adj_ has joined #linux-sunxi
<hansg> kaplan2539, I've just done a (forced) push to my sunxi-wip u-boot branch adding support for the SPL reading u-boot from nand
<hansg> kaplan2539, to enable this make this change:
<hansg> --- a/configs/A13-OLinuXino_defconfig
<hansg> +++ b/configs/A13-OLinuXino_defconfig
<hansg> @@ -13,4 +13,5 @@ CONFIG_VIDEO_LCD_POWER="AXP0-0"
<hansg> CONFIG_VIDEO_LCD_BL_PWM="PB2"
<hansg> CONFIG_DEFAULT_DEVICE_TREE="sun5i-a13-olinuxino"
<hansg> CONFIG_SPL=y
<hansg> +CONFIG_SPL_NAND_SUPPORT=y
<hansg> CONFIG_SYS_EXTRA_OPTIONS="CONS_INDEX=2,AXP209_POWER,USB_EHCI"
<hansg> And then do:
<hansg> make -j4 CROSS_COMPILE=arm-linux-gnu- A13-OLinuXino_defconfig
<hansg> make -j4 CROSS_COMPILE=arm-linux-gnu-
<kaplan2539> hansg: thanks alot! Just finished to compile the kernel from sunxi-wip
<hansg> And then take thye build u-boot-sunxi-with-spl.bin and on an A13 OLinuxIno booted from sdcard with my sunxi-wip kernel branch + .config I send you, do:
<hansg> flash_erase /dev/mtd0 0 0
<hansg> nandwrite -p /dev/mtd0 /mnt/u-boot-sunxi-with-spl.bin
<hansg> And then cleanly shutdown the linux booted from the sdcard (poweroff), then remove the sdcard, and power up the board without it. Now you should see the SPL load and then end up in the u-boot environment. At this point things stop since we do not yet support reading the kernel from the nand. Roy is working on adding support for that
<kaplan2539> sounds promising! I'll compile u-boot now
Akagi201 has joined #linux-sunxi
JannF has quit [Quit: Page closed]
Akagi201 has quit [Ping timeout: 272 seconds]
Andy-D has quit [Ping timeout: 272 seconds]
sehraf has quit [Read error: Connection reset by peer]
<hansg> kaplan2539, bedtime for me, if you've any questions drop me a mail.
<kaplan2539> hansg: thanks alot! I was able to boot linux but had trouble compiling u-boot
hansg has quit [Quit: Leaving]
Andy-D has joined #linux-sunxi
VargaD_ has joined #linux-sunxi
Andy-D has quit [Ping timeout: 256 seconds]
VargaD has quit [Ping timeout: 276 seconds]
VargaD_ is now known as VargaD
ricardocrudo has quit [Ping timeout: 252 seconds]
Black_Horseman has quit [Read error: Connection reset by peer]
Andy-D has joined #linux-sunxi
Black_Horseman has joined #linux-sunxi
fredy has quit [Excess Flood]
lemonzest has quit [Quit: Leaving]
awe00 has quit [Ping timeout: 244 seconds]
jinzo has quit [Quit: Leaving]
fredy has joined #linux-sunxi
kaplan2539 has left #linux-sunxi [#linux-sunxi]
naobsd has quit [Quit: naobsd]
flyhorse_PvTFu has joined #linux-sunxi
flyhorse has quit [Read error: Connection reset by peer]
sehraf has joined #linux-sunxi
Akagi201 has joined #linux-sunxi
Akagi201 has quit [Ping timeout: 264 seconds]
sehraf has quit [Read error: Connection reset by peer]
p1u3sch1 has quit [Ping timeout: 256 seconds]
gianMOD has joined #linux-sunxi
whirm has quit [Quit: WeeChat 1.2]
nicksydney has joined #linux-sunxi
vishnup has joined #linux-sunxi