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
mzki has joined #linux-sunxi
phipli has quit [Ping timeout: 250 seconds]
<naobsd> about NES classic, currently I cannot get info by thins way
<naobsd> I can see some DRAM info on console http://linux-sunxi.org/Nintendo_NES_Classic_Edition#Stock_U-Boot
<naobsd> is it enough for u-boot spl?
atsampson has quit [Ping timeout: 258 seconds]
<naobsd> (looking info in NAND...)
mzki has quit [Ping timeout: 260 seconds]
atsampson has joined #linux-sunxi
victhor_ has joined #linux-sunxi
<MoeIcenowy> naobsd: U-Boot SPL uses only a few parameters
<MoeIcenowy> could I build one for you to test?
<naobsd> enough?: CONFIG_DRAM_CLK=600 CONFIG_DRAM_ZQ=15291 CONFIG_ODT_EN=y
<MoeIcenowy> I think so
<naobsd> these values are same as parrot R16
<MoeIcenowy> I think you have UART access now?
<naobsd> for now, temporary (not soldered ;)
<naobsd> but u-boot/kernel is still stock, very restricted
<MoeIcenowy> but mainline kernel have poor NAND support...
<naobsd> for u-boot I should get regulator info?
<naobsd> well, my current goal is adding very very initial support ;)
<naobsd> where is common location for script.bin on A33 devices?
<MoeIcenowy> After U-Boot...
futabachannel has joined #linux-sunxi
<MoeIcenowy> or find it in stock U-Boot in memory at 0x43000000
<MoeIcenowy> made an initial u-boot-sunxi-with-spl.bin for NES mini
<naobsd> well
<naobsd> I know mainline spl works, how to build spl
<naobsd> compatible = "allwinner,parrot",
<naobsd> are regulator settings right?
<naobsd> anyway, what I want to do is collection info
<naobsd> 0x43000000 doesn't have valid header :(
<naobsd> 00000000 4e 00 00 00 d0 a3 00 00 01 00 00 00 02 00 00 00 |N...............| 00000010 70 72 6f 64 75 63 74 00 00 00 00 00 00 00 00 00 |product.........|
<naobsd> it might be better to build modified u-boot :(
vagrantc has joined #linux-sunxi
jernej has quit [Ping timeout: 260 seconds]
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Ping timeout: 264 seconds]
<MoeIcenowy> naobsd: I think it's a script.bin.
<MoeIcenowy> as there's "product" here
<naobsd> yeah, many params are follow,
<naobsd> but bin2fex didn't work...
Colani1210 has joined #linux-sunxi
<MoeIcenowy> what's the output of bin2fex?
<naobsd> E: fexc-bin: Malformed data: version 41936.1.2.
Colani1200 has quit [Ping timeout: 260 seconds]
<MoeIcenowy> ok use the newest git master
<naobsd> it should be new...
<naobsd> grrrr
<naobsd> works now :(
<naobsd> mmm it's updated a little...anyway thanks
perr has quit [Quit: Leaving]
<naobsd> fexc-bin: script.bin: version: 1.2 fexc-bin: script.bin: size: 131072 (78 sections), header value: 41936 E: fexc-bin: script.bin: empty entry in section: csi0
<naobsd> is last E: ok?
<naobsd> machine = "parrot" lol
<MoeIcenowy> maybe it's ok...
<MoeIcenowy> I do not know
<MoeIcenowy> could you fork sunxi-boards on github, then add your fex file to A33 folder, then pull request?
fire219 has quit [Ping timeout: 256 seconds]
<naobsd> yes of course
<naobsd> I know what should be done, what I don't know well is about A33 ;)
apritzel has quit [Ping timeout: 246 seconds]
terra854 has joined #linux-sunxi
<naobsd> sys_config/a33? or sys_config/r16?
whaf has joined #linux-sunxi
fire219 has joined #linux-sunxi
whaf has quit [Ping timeout: 248 seconds]
<naobsd> u-boot-sunxi mirror/next is for upstreaming, right?
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
berenm has quit [Remote host closed the connection]
berenm has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Ping timeout: 264 seconds]
ninolein has quit [Ping timeout: 260 seconds]
ninolein has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
<wens> naobsd: it's a mirror of the sunxi custodian repository
<naobsd> wens: http://git.denx.de/?p=u-boot/u-boot-sunxi.git;a=shortlog;h=refs/heads/next this should be used for base?
<wens> naobsd: given that sunxi u-boot is currently in limbo, i can't say for sure :|
<naobsd> I see
<naobsd> oops
<naobsd> I was thinkig that use OTG as HOST
<naobsd> I pulled USB cable after running fel uboot
<naobsd> power is lost, of course ;)
whaf has joined #linux-sunxi
<naobsd> MMC need to be disabled to boot properly...
<ssvb> naobsd: so which steps were needed to extract the script.bin data?
victhor_ has quit [Ping timeout: 245 seconds]
victhor__ has joined #linux-sunxi
<naobsd> ssvb: update sunxi-tools... sunxi-fel read 0x43000000 0x10000 script.bin is fine
Gerwin_J has joined #linux-sunxi
<ssvb> naobsd: I see, and was it so that you read the leftover data in DRAM after resetting the device when running the stock firmware?
<ssvb> naobsd: I mean, if you enter the FEL mode after turning the device off and waiting some tens of seconds, is the script.bin data there?
<naobsd> send 's' to UART -> run 'fastboot' -> FEL mode with DRAM data
<naobsd> ssvb: I never tried that method
<naobsd> I'll try later
<ssvb> OK, I see, so the stock firmware initializes the DRAM and switches into FEL mode itself
<ssvb> this makes everything way too easy :-)
<naobsd> if you send 's' to UART, yes
<naobsd> (RESET button method and send '2' method don't run boot1, DRAM is not initialized)
Gerwin_J has quit [Ping timeout: 264 seconds]
<ssvb> so the current instructions from the linux-sunxi wiki about extracting script.bin actually work? or do they need to be updated?
<naobsd> http://linux-sunxi.org/Retrieving_device_information#Retrieving_data_over_USB_in_FEL_mode "./sunxi-fel read 0x43000000 0x20000 script.bin" this works
<naobsd> "./sunxi-fel read 0x42400000 0x82d0 boot1.header" didn't work
huawei has quit [Quit: ZNC 1.6.3 - http://znc.in]
huawei has joined #linux-sunxi
IgorPec has joined #linux-sunxi
pg12 has quit [Ping timeout: 250 seconds]
pg12 has joined #linux-sunxi
berenm has quit [Remote host closed the connection]
havoc_ has quit [Ping timeout: 260 seconds]
berenm has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
havoc_ has joined #linux-sunxi
Gerwin_J has quit [Client Quit]
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Client Quit]
victhor__ has quit [Ping timeout: 256 seconds]
Gerwin_J has joined #linux-sunxi
<naobsd> oops
futabachannel has quit [Ping timeout: 240 seconds]
TheSeven has quit [Disconnected by services]
[7] has joined #linux-sunxi
Putti has quit [Quit: Leaving]
vagrantc has quit [Quit: leaving]
JohnDoe_71Rus has joined #linux-sunxi
tsuggs has quit [Remote host closed the connection]
zzeroo has joined #linux-sunxi
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
alexxei has quit [Read error: Connection reset by peer]
jernej has joined #linux-sunxi
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
alexxei has joined #linux-sunxi
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
scream has joined #linux-sunxi
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
jernej has quit [Ping timeout: 260 seconds]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
<wens> codekipper: my opi pc struggles with "sunxi-mmc 1c0f000.mmc: fatal err update clk timeout" with your patches :/
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
reinforce has joined #linux-sunxi
iamfrankenstein1 has joined #linux-sunxi
BigFellow has quit [Client Quit]
jernej has joined #linux-sunxi
alexxei has quit [Ping timeout: 260 seconds]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has joined #linux-sunxi
jernej has quit [Ping timeout: 256 seconds]
BigFellow has quit [Client Quit]
jstein_ has joined #linux-sunxi
jstein_ has quit [Remote host closed the connection]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
scream has quit [Remote host closed the connection]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
BigFellow has joined #linux-sunxi
BigFellow has quit [Client Quit]
<wens> codekipper: seems like the pinctrl patches are at fault
f0xx has joined #linux-sunxi
<wens> codekipper: so, h3's spdif controller has a different register map compared to a20
<wens> codekipper: specifically, the TX FIFO was moved
<wens> codekipper: looks like it was swapped with the interrupt status register
<wens> codekipper: in addition, the dma controller is subtly different from a31, in that the dma burst length register is offset by 1
<wens> so setting a burst of 1 still gives 1, setting a burst of 8 actually gives a burst of 4
<wens> mripard: bad news, my opi pc fails to boot with the new pinctrl stuff
<wens> mripard: apparently mmc needs pull-ups during initial handshaking
<wens> and we have NO_PULL set for the mmc pins
<mripard> wens: and it needs to be disabled after initialisation?
Gerwin_J has quit [Quit: Gerwin_J]
<wens> mripard: i guess you can leave it on afterwards since it has a resistor, no big deal
premoboss has joined #linux-sunxi
<wens> some boards have external pull-ups on all pins, some only on the cmd pin :/
codekipper has joined #linux-sunxi
<codekipper> wens: thanks for looking at it...I'll dig out the specs again and check....the i2s behaved in the same manner.
<codekipper> mripard: Just seen that you've delivered capture into i2s. Did you test that using the uda1380?
<mripard> codekipper: no, using a wm8978
<mripard> wens: can you send a patch then?
massi has joined #linux-sunxi
ErwinH has joined #linux-sunxi
perr has joined #linux-sunxi
perr has quit [Changing host]
perr has joined #linux-sunxi
<longsleep> looks interesting if the hardware works properly
<beeble> wens: mripard: DAT lines are specified as high when idle. so pull-ups enabled for the whole time are fine
paulk-collins has joined #linux-sunxi
fkluknav has joined #linux-sunxi
<KotCzarny> 2GB DDR3 SDRAM with 733MHz.
<KotCzarny> lol
<KotCzarny> i wonder if they really cruise with that speed
premoboss has quit [Ping timeout: 268 seconds]
leviathanch has joined #linux-sunxi
<wens> mripard: i'm worried if other boards would fail, what about your NO_PULL removal patch?
<mripard> wens: what about it ?
<mripard> if other boards fail, we will fix them too
<wens> yeah, i guess they're separate issues
<longsleep> I wonder what the BananaPi M64 might cost, anyone able to find a link where to buy?
tkaiser has joined #linux-sunxi
<tkaiser> longsleep: They told Jean-Luc that they will charger $35 for it (same with BPi M2 Ultra) so just check real price for the latter and you know it (10 bucks more)
<longsleep> tkaiser: ok, thats not too bad then
<KotCzarny> not too bad in specs, but they didnt deliver good quality since original bpi-m1
<tkaiser> BPi M2 Ultra is sold for $45, so expect the A64 in the same range (it's more or less the same board)
<tkaiser> longsleep: Also http://bananapiwarehouse.com -- but the last time I dealt with them they did not even know that they don't distribute LeMaker stuff... be warned! ;)
<mripard> wens: I'm not really sure when you mentionned the NO_PULL removal though :/
<wens> mripard: the DT conversion patches you originally sent with the generic pinconf bindings series
<tkaiser> longsleep: BTW: apritzel has one on his desk so for you it's just asking Nora Lee to get a few ;)
<mripard> wens: yes, I know, but why did you mention it ? :)
ScrumpyJack has quit [Ping timeout: 245 seconds]
phipli has joined #linux-sunxi
alexxei has joined #linux-sunxi
florianH has joined #linux-sunxi
phipli has quit [Ping timeout: 240 seconds]
<wens> mripard: i assume it would preserve whatever behavior we are counting on now :)
<mripard> aaaah
<mripard> indeed
<mripard> I was planning on sending it as a late PR so that they merge it after the pinctrl stuff
<wens> i think that would be a proper fix wrt the new pinconf patches
<wens> we can then send patches for individual board that we actually identify are having problems
<mripard> well, the proper fix would really be to have the pull-ups enabled in the first place
<mripard> in the DT
pekka10 has quit [Ping timeout: 260 seconds]
Andy-D has joined #linux-sunxi
berenm has quit [Quit: berenm]
petr has quit [Ping timeout: 268 seconds]
petr has joined #linux-sunxi
f0xx has quit [Ping timeout: 256 seconds]
tkaiser has quit [Quit: jIRCii - http://www.oldschoolirc.com]
Christos_ has joined #linux-sunxi
<Christos_> hmm.. H5 Linux SDK --> https://github.com/OrangePiLibra
<wens> mripard: i agree
marcan has joined #linux-sunxi
Andy-D has quit [Ping timeout: 260 seconds]
Christos_ has quit [Client Quit]
<buZz> oo h5 is also a53 , nice
<buZz> ah thats old H64?
<buZz> memory scramble? is that anti-rowhammer?
matthias_bgg has joined #linux-sunxi
perr has quit [Quit: Leaving]
gumblex has joined #linux-sunxi
<wens> mripard: what do you think, do it per board, or just change the default in the .dtsi?
<wens> afaik u-boot sets the pull-ups, which is why NO_PULL (i.e. no-op) worked before
f0xx has joined #linux-sunxi
IgorPec has quit [Ping timeout: 256 seconds]
<mripard> I think it would make more sense that it's per board
premoboss has joined #linux-sunxi
leviathanch has quit [Remote host closed the connection]
tkaiser has joined #linux-sunxi
<tkaiser> Christos_: Well, this is more or less a copy of longsleep's simpleimage stuff (without crediting him), Allwinner's BSP source drop and still no documentation :(
diego_r has joined #linux-sunxi
<KotCzarny> hmm
<KotCzarny> maybe longsleep should add trojan to his image, that will detect when ran from allwinner/xunlong hq it will upload everything to github?
<KotCzarny> ;)
<jelle> lol
<KotCzarny> or look for sources?
<longsleep> hehe yeah
<longsleep> but its all bash scripts, so would need to be in plain sight
<KotCzarny> ehehe
<KotCzarny> that's the trick
<KotCzarny> they dont bother looking at those scripts
<KotCzarny> ;)
<tkaiser> The less wheels reinvented the better
<terra854> Hey guys
<KotCzarny> grabbing someone's work without crediting or contacting first isnt good
<KotCzarny> especially that wens speaks chinese
<KotCzarny> so its not about language barrier
<terra854> Wait... where did this happen?
<tkaiser> KotCzarny: Depends: SinoVoip for example uses stuff from Armbian, then cripples it so that it's only crap anymore and then provides 'Armbian for BPi M3'. That is even more worse.
<KotCzarny> uh, did they actually named it 'armbian' ?
<tkaiser> KotCzarny: Same with longsleep's name in Pine64 forum. He gets credited for a lot of crappy OS images there ;)
<longsleep> yeah that bothers me much more than someone taking my crappy scripting and doing whatever with it
yann-kaelig has joined #linux-sunxi
<KotCzarny> longsleep, add your github url to the bash scripts
<longsleep> nah i intentionally did not add anything there
<KotCzarny> and write something like 'if you didnt obtain this software from ... and used it unmodified, please dont credit me'
<longsleep> i should add a wtfpl license file or something
<longsleep> ah thats a good idea
<longsleep> actually its also kind of an experiment - i did not expect for my pine64 images to stay around that long
<KotCzarny> there are no better
<KotCzarny> so.
<KotCzarny> ;)
<terra854> Well, it's better than nothing
<terra854> Without you, the Pine can only run Android from Allwinner BSP
<KotCzarny> nope. its better than sinovoip/clueless folks on different forums crap
<KotCzarny> s/its/nothing/
<longsleep> yes but still, the gear is intentionally under-documented and not meant to be used by anyone else other than developers - i mean i feel sorry for everyone who has to use my scripts to produce images which are better than the ones which would be produced without :)
<tkaiser> terra854: Please don't forget about the foundation. There's so much work done here the average Pine64 user is not aware of.
<terra854> These corporate folks really need to know how to build good quality software instead of churning out crippled crap
<terra854> tkaiser: What foundation?
<tkaiser> linux-sunxi community. These guys here doing so much work and getting no credits for. Just for fun ;)
<tkaiser> Fortunately longsleep was the first being able to assemble something that booted BSP kernel with a linux rootfs so everyone else relied on this stuff. Otherwise it would really look different
ErwinH has quit [Ping timeout: 260 seconds]
<terra854> Yeah. Doing all these with little documentation
<longsleep> the stuff i did is not complicated
<tkaiser> terra854: Since he added stuff to update the important parts of whatever OS image later already (scripts to update u-boot/kernel).
<tkaiser> Then came Pine Inc folks, crippled images, introduced bugs and shared them. And the unfortunate pine64.pro site
<longsleep> i was just too lazy to do the kernel packaging and needed a way to update my kernels, turned out to be a good idea to be lazy :)
<terra854> Yeah. And the maintainer of the site gone AWOL
<terra854> Yeah. And the maintainer of the site has gone AWOL
<tkaiser> longsleep: Definitely. BTW: spotted your u-boot/kernel update scripts also in "H5 SDK" before. :)
<KotCzarny> hehe
<longsleep> tkaiser: oh really, did they change at least the GPG key?
<KotCzarny> HEHE
<KotCzarny> unlikely
<tkaiser> longsleep: Just saw that something is prepared. Maybe I spent more time opening an issue there and asking for the H5 user manual
<longsleep> tkaiser: if they didnt change the URLs then it would just download and apply the latest pine64 kernel from my site
<tkaiser> longsleep: Sure, but that would be already the SinoVoip way ;)
<KotCzarny> which is a nice way to modify motd ;)
bugzc has quit [Ping timeout: 260 seconds]
<tkaiser> They provide OS images for one of their boards, rely on an Armbian rootfs, forget to clear sources.list and the next time the user does an apt upgrade wrong u-boot+spl will brick the image
<longsleep> oh nice
<tkaiser> Even nicer: they don't adjust that even if their forum is flooded with stuff like that :)
<KotCzarny> and went 'that armbian stuff is crap' ?
<KotCzarny> seriously. armbian should add similar disclaimer too
mzki has joined #linux-sunxi
<KotCzarny> to boot script
<tkaiser> BTW: I tried to boot an unmodified Pine64 image on OPi PC 2 but no signs of live. Seems the first boot code already checks SoC ID?
<longsleep> tkaiser: more likely the device tree does not fit
<longsleep> tkaiser: or do you mean you do not even see boot0 output?
<tkaiser> longsleep: Not even a 'BOOT0' appeared, nothing
<longsleep> tkaiser: well then replace boot0 with the one from the android image
<KotCzarny> i think yesterday apritzel was fiddling with it
<KotCzarny> and something changed about boot0
<terra854> Speaking of boot0, is there any effort to get rid of it?
<tkaiser> longsleep: Sure, it worked the same way with H8 vs. A83T. But there a few lines were printed to console claiming 'wrong chipid'. With A64/H5 simply nothing
<terra854> As in, anyone cooking up alternatives?
perr has joined #linux-sunxi
<wens> apritzel: ^
perr has quit [Changing host]
perr has joined #linux-sunxi
<longsleep> terra854: upstream u-boot does not use boot0, but there is the problem with dram settings being closed source
<wens> he's not here atm
<terra854> So we only rely on boot0 for that?
<longsleep> terra854: the bsp u-boot relies on it
<terra854> the dram settings?
<longsleep> terra854: boot0
<terra854> As in relying boot0 for the dram?
<longsleep> among other things yes
<longsleep> uboot spl does not use boot0, but needs to initialize dram which right now cannot be done without libdram
<terra854> Any plans on reverse engineering or getting rid of boot0?
<longsleep> no idea if anyone has plans to get into this
<beeble> we have one
<wens> beeble: is there any reason to have external pull-ups on only the mmc cmd line?
<wens> beeble: and not all of them?
<wens> (nice to have a hardware engineer here)
<beeble> wens: the cmd line is required since opendrain (think of i2c)
<tkaiser> longsleep: With A64 it's still 'dd if=sdcard-image.bin of=boot0.bin bs=1024 skip=8 count=32' to extract boot0?
hojnikb has joined #linux-sunxi
<wens> beeble: but not the clk or dat pins? puzzling
<longsleep> tkaiser: let me check
<beeble> so a64 uboot spl is working
<beeble> but the patches are not yet published since they need some cleanup
<longsleep> tkaiser: yes, boot0 is at 8KiB for a64
<tkaiser> longsleep: Thanks, would be a simple test to exchange boot0 and .dtb of a Pine64 image and look what happens on a H5 device.
<wens> with the latest pinconf patches now honor SUNXI_PINCTRL_NO_PULL, so the internal pull-ups get disabled
<longsleep> tkaiser: yes, though u-boot has a device tree too, not sure how that will work
<wens> and now i'm getting intermittent mmc time out messages on my bpi-m1-plus
<longsleep> tkaiser: but at least you should get output on uart0 with the correct boot0
<tkaiser> longsleep: Ah, but that would be also easy using your gear to assemble uboot+spl. Might try that later.
<longsleep> tkaiser: yes, but i somewhat doubt that the pine64 uboot and atf works on the h5
<longsleep> tkaiser: does the h5 have an arisc too?
<tkaiser> longsleep: Nope
<tkaiser> longsleep: Sorry, it has
<longsleep> then, no idea what might happen if you load the firmware blob from pine64 into one with a h5 chip :)
<tkaiser> longsleep: I have to try. Since I already partially damaged my OPi Zero ;)
<longsleep> damaged with software?
<beeble> wens: are talking to a sd card or emmc?
IgorPec has joined #linux-sunxi
Putti has joined #linux-sunxi
<tkaiser> longsleep: More probably by accident fiddling around physically. Wi-Fi isn't accessible through SDIO any more.
<hojnikb> that sux
<hojnikb> guys
<hojnikb> are there any quotes on how much chips like h2 and h5 cost in bulk
<hojnikb> i wonder whats the cost difference between something like h2+ and h3
<hojnikb> i suspect not much, given how both probably use the same die
<tkaiser> hojnikb: Most probably it's just like that: test the stuff coming from the wafer, it it shows stability problems regarding GBit Ethernet MAC, HDMI output or when clocked higher than 1.4GHz then it's an H2+ afterwards otherswise H3
<hojnikb> tkaiser: oh, so h2+ doesn't have gbit mac ?
<tkaiser> Yeah, and limited to 1080p HDMI output according to product brief
<hojnikb> nothing really mayor if you ask me. Xunlong could introduce a new PC with H2+ instead of H3... For the same price they could include something like wifi on the money saved for using h2+
<hojnikb> enter Orange Pi PC Wifi or something :)
<hojnikb> as for h5/pc2, were you able to do some power consumption tests ?
<tkaiser> hojnikb: Hmm... it scares my that people consider devices like an OPi PC *expensive*. And no, no consumption measurements done since useless. No voltage regulation currently working and based on looking at the DT for OPi PC 2 this is just more or less Allwinner defaults so tweaking here and there might also be necessary
<hojnikb> on a single unit yes, it's really inexpensive. But if you need something like 1k units, then every $ counts :) And having a well balanced board with all the necessery features is definetly a plus :)
<tkaiser> hojnikb: Also a question of the use case. It seems people buy OPi PC 2 since they think this could be a PC replacement (thinking of surfing the web, watching video, doing real work). This consumption stuff is more important for SBC that are used like that: low power thingies collecting data, switching relais and stuff like that
<hojnikb> well consumption is somewhat important, since it directly relates to heat; more power consumtion, more heat to dissapate :)
<hojnikb> so it's desirable to have as little power consumtion as possible, because then you can have passive cooling without a problem, even for desktop use
<tkaiser> hojnikb: Sure but this is where dvfs jumps in. And that's not working. At least I managed to exceed 100¡C yesterday running cpuburn-a53 so everything as expected regarding CPU heat.
<tkaiser> hojnikb: So it's a misunderstanding anyway. The consumption tests I make cover idle consumption. You're speaking of 'normal use' and there Mali450 could add significantly. No one knows now and I won't test it anyway (those use cases like using an SBC as desktop replacement)
<hojnikb> i'm guessing thats without heatsink and at 1.1V ?
<tkaiser> Yes
<KotCzarny> uhoh, 1GB for a pc replacement?
<KotCzarny> even 2GB is not enough
<hojnikb> It's plenty
<hojnikb> I'm running daily a z3735f with 2g of ram and win10
* KotCzarny checks firefox mem usage. 550MB atm, now add os, video buffers other stuff
<hojnikb> it's all about memory managment
<tkaiser> KotCzarny: I don't understand those guys. But they exist. Same with Orange Pi One. People started this thing like crazy just to complain that the solder pads to add the 2 missing USB ports are too tiny. Instead of spending just $5 more to get an OPi PC
<hojnikb> web browsers use 90% of their memory for caching stuff
<KotCzarny> well, if you count heavy browsing out, limit yourself to some simple wm, sure, its *doable*, dont expect smooth sailing though
<hojnikb> obviously not. But it's enough for most people.
<hojnikb> with a lightweight distro and a properly tuned web browser, even 1GB is doable
<hojnikb> obviously not 100 tabs with 90 of them filled with pornhub :)
<KotCzarny> nope. web is ugly nowadays
<KotCzarny> hd jpegs everywhere, or more
<KotCzarny> now add flash, fucktillion of js and forget about 1GB
<hojnikb> flash is pretty much dead
<KotCzarny> sure, html5 ads are all the reage
<KotCzarny> hehe
<KotCzarny> *rage
<hojnikb> uBlock :)
<KotCzarny> which doesnt help much
<JohnDoe_71Rus> yep. even on "big" PC 1Gb is nothing now days
<KotCzarny> there is unblock on the way that promises working around most of adblockers
<hojnikb> it's really pathetic how memory is treated nowadays
<hojnikb> i mean yes it's cheap
<hojnikb> but still
<hojnikb> thats just lazy programming
<KotCzarny> 640kB should be enough for everyone?
<hojnikb> although dram prices are on the rise if anyone noticed :)
<hojnikb> i bet this was part of the reason why pc2 is 20$ not 15
<KotCzarny> firefox switched from nativ ui into dom ui around 44 i think, then things went downhill for it
<hojnikb> you should try palemoon
<hojnikb> it even has arm v7 optimized version
<wens> beeble: micro sd card
* KotCzarny googles
<hojnikb> although last time i checked, it performed slower on octane v2 than firefox
<hojnikb> for some odd reason
<KotCzarny> wth is goanna?
<hojnikb> although i don't know how well firefox is optimized for arm anyway
<KotCzarny> firefox ui sucks since dom change
<KotCzarny> its noticable even on x86 (when using anything slower than i5)
<hojnikb> thats horrible
<hojnikb> i personally switched to opera some time ago
<hojnikb> on desktop
<KotCzarny> in a way its nice, because you can modify everything with a css, but its heavy on resources
<JohnDoe_71Rus> opera is chrome now
<hojnikb> yeah i'm aware of that.
<hojnikb> still ok. little heavy on the ram as usual for chrome
IgorPec has quit [Ping timeout: 260 seconds]
<hojnikb> anyone knows why webkit1 browsers like Web and midori are unstable on arm ?
<hojnikb> i always get segfaults after some time
<JohnDoe_71Rus> nobody whant fix it
OtakuNekoP has joined #linux-sunxi
<KotCzarny> anything in dmesg?
<wens> mripard: so i grepped through sunxi-boards, and almost all of them have pullups enabled for all the mmc0 pins
<wens> mripard: except for 3 A33 q8 tablets, which don't have it enabled _only_ on the dat3 pin
<hojnikb> JohnDoe_71Rus: oh that makes sense.
<wens> mripard: would that sway you? :)
<KotCzarny> hojnikb: where is that armv7 optimized version?
Putti has quit [Ping timeout: 252 seconds]
<KotCzarny> thx
<hojnikb> you have to compile it yourself though
<hojnikb> no packages as far as i'm aware
<KotCzarny> i wonder whats' the difference from normal source build
<hojnikb> there is a good chance palemoon didn't switch to dom ui just yet
<hojnikb> check on palemoon forums, they explain it well.
<KotCzarny> it even has some egl support
<KotCzarny> maybe it would be the first real use case for mali
<hojnikb> it's worth giving it a shot
<hojnikb> i personally just tested octane on opi pc
<hojnikb> and scores were lower on almost everything
<hojnikb> than on firefox
<JohnDoe_71Rus> i have situations. one site + html5 work for me in firefox but don't work in palemoon. for another person contrary
<hojnikb> could be a compatability thing
<hojnikb> maybe website has hardcoded version checks
<hojnikb> palemoon still reports itself as firefox 26 i believe
<JohnDoe_71Rus> one version palemoon
<KotCzarny> did you check misc ff exploits on it?
<hojnikb> no i haven't
<hojnikb> honestly i haven't used it in a long time
<hojnikb> just cought my eye a week ago, since i saw an arm build
<hojnikb> so i gave it try
<hojnikb> as far as security goes, i believe they patch firefox exploits pretty quickly.
<mripard> wens: hmmm, yes, go ahead
<hojnikb> finally, my pc2 left china... :D
Ntemis has joined #linux-sunxi
Net147 has quit [Quit: Quit]
tkaiser has quit [Quit: jIRCii - http://www.oldschoolirc.com]
<jelle> mine should be arriving today =)
<hojnikb> ^^ i'm totally jealous :D
<KotCzarny> hojnikb: unless you are going to do devel work, dont be, it doesnt even boot yet
<MoeIcenowy> mine arrived 6 days ago ;-)
<MoeIcenowy> since I'm in China ;-)
hojnikb has quit [Remote host closed the connection]
hojnikb has joined #linux-sunxi
<KotCzarny> uh, did i hurt his feelings? :>
<hojnikb> :D
<KotCzarny> i know it might be shocking ;)
<hojnikb> i thought we already have a booting image for h5
<jelle> KotCzarny: obv. bought it for devel stuff, what else do I have to do with my spare time :P
<Ntemis> i found a fix for orange pi soc and unstable memories
<Ntemis> i think i can make it up to 816mhz
<hojnikb> care to share ?
<KotCzarny> ntemis: what tool did you use to confirm its stable?
<Ntemis> vs 624mhz right now
<hojnikb> well, dram should be stable at 800Mhz, given how modules are 1600MT
<hojnikb> so it's really a matter for memory controller not playing nice above 667
<Ntemis> KotCzarny: is the SY8008B the issue here
<Ntemis> it cant feed the ram than needs 1.35v
<Ntemis> it only can output 1.2v
<hojnikb> 1.2V for ddr3
<KotCzarny> Ntemis: but what tool did you use to confirm it works stably
<hojnikb> thats a miracle it even boots
paulk-collins has quit [Remote host closed the connection]
<Ntemis> replace this with LTC3250
<Ntemis> and it will boot
<Ntemis> and memory can fly
<hojnikb> and the dram will clock higher becase of this ?
<Ntemis> yeap
<Ntemis> you can set it higher yes
<Ntemis> 816mhz
<hojnikb> but stock can only go to 1.2V and no more, correct ?
<Ntemis> aha
fkluknav has quit [Ping timeout: 246 seconds]
<Ntemis> but memory pdf says it clearly it can fuction stably at 1.35v at 1600mhz
<Ntemis> but SY8008B only gives 1.2v max and guess what that the max
<hojnikb> at 1.2V it's still severely undervolted.
<Ntemis> so could be lower
<Ntemis> severely yes
<hojnikb> it this a regulator limiation or is something else at play ?
<KotCzarny> Ntemis: ok, so that's a hw mod. care to run limamemtester at that speed?
<Ntemis> sure
<KotCzarny> only then it can be announced as a success
<Ntemis> blueprints specs cant lie
<Ntemis> it will be
<hojnikb> blueprint specs ?
<KotCzarny> remmeber to connect hdmi, you have to visually confirm spinning cube on *gray* bg
<KotCzarny> if it survives 30-40 minutes, yay
<Ntemis> ok i will for sure
<Ntemis> but for fact wrong Step-Down DC/DC Converter was selected for those socs
<hojnikb> is LTC3250 pin and software compatabile to sy80008 ?
<Ntemis> hojnikb: check pdf of SY8008B and K4B4G1646Q-HYK0
<Ntemis> you will see the problem right there
<Ntemis> and yes they are :)
<Ntemis> pin yes
<hojnikb> is desolder sy and solder ltc
<hojnikb> so*
<Ntemis> aha
<Ntemis> sec
<hojnikb> and dvfs will work ?
<Ntemis> check it out
tkaiser has joined #linux-sunxi
<Ntemis> use 1.5v ofc not 1.2
<Ntemis> there are two
<tkaiser> Ntemis: So you used the test points on your OPi PC and measured voltage?
<Ntemis> even better
<Ntemis> found out what the issue was
<tkaiser> Ntemis: So you did *NOT* measure, right?
<Ntemis> right
<tkaiser> Ntemis: There are people who did that.
<Ntemis> aha
<hojnikb> it's really odd, that sy can't supply dram with 1.5V, but it can the cpu cores
<beeble> thats an adjustable regulator
<tkaiser> Ntemis: *surprise*: it's 1.35V on OPi PC and 1.5V on the smaller Oranges (DDR3 and not DDR3L). I sent you the link to these results yesterday or the day before.
<beeble> so it depends on the voltage divider
<hojnikb> so it could be a board issue ?
<Ntemis> tkaiser: fact SY8008B cant supply more than 1.2v
<hojnikb> explain than, why boards are overheating using stock dvfs settings
<hojnikb> at 1.2V they shouldn't
<KotCzarny> dvfs manages soc, not dram, isnt it?
<tkaiser> Ntemis: Did you use multimeter?
vishnup has joined #linux-sunxi
<hojnikb> yes, but i'm guessing the same voltage regulator is used for dram and soc
<hojnikb> is that correct ?
<tkaiser> Nope, that's BS
<tkaiser> The whole discussion is BS
<tkaiser> As yesterday and the day before
<hojnikb> i don't know, ddr3 at 1.2V would be pretty unstable imo
<Ntemis> yeap
<tkaiser> hojnikb: Ok, I claim it's 0.8V. What's now. My claims are backed by the same stuff like Ntemis: nothing
<hojnikb> so something else must at play here
<tkaiser> OMG
tkaiser has quit [Quit: jIRCii - http://www.oldschoolirc.com]
<Ntemis> am not claiming anything that cant be proofed
<Ntemis> so proof me wrong
<KotCzarny> Ntemis: just do the software test
<hojnikb> is there even any testpoints on the board to test dram voltage directly ?
<Ntemis> that is exactly why i havent done wny tests yet
<Ntemis> that is exactly why i havent done any tests yet
<Ntemis> i have to find those
<Ntemis> and am not an expert am just a plain user
<NiteHawk> hojnikb: yup, "1.5V"
<hojnikb> is dram supplied directly via voltage regulator or is something in between ?
fkluknav has joined #linux-sunxi
<hojnikb> if so it would be pretty straightforward to the the voltage :)
<Ntemis> there are a bunch of resistors and capacitors
<NiteHawk> errr... hold on - that is assuming we're still taking "orange pi pc"?
<Ntemis> and thats it
<hojnikb> yes
<Ntemis> aha
<NiteHawk> right, that page shows the TP
<Ntemis> but as it seems others are even wrorst place needing 1.5v for the ram vs 1.35v as pi pc
<Ntemis> no way SY8008B can deliver that voltage
<NiteHawk> section "D3", right next to VCC-DRAM
<KotCzarny> then measure point marked as '1_5V'
<NiteHawk> the schematics page states that you could/should replace Rd2 (R14) for 1.5V (substituting 100K for 120K)
<hojnikb> looks like sy8008b inly supplies dram
<Ntemis> and supplies it with less voltage
<KotCzarny> there is another sy8008b supplying 3.3V apparently
<KotCzarny> avcc
<hojnikb> thats RT9050-33GB
<KotCzarny> on schematics its U5
<hojnikb> guess they try to save every cent on the worst bottom barrel vrms
victhor__ has joined #linux-sunxi
<hojnikb> i'm really surprised those SoC don't have somekind of integraded voltage regulator
<hojnikb> would make things somewhat cheaper
<hojnikb> and better controller
<hojnikb> control*
<MoeIcenowy> it's the work of AXP
<MoeIcenowy> X-Powers have nearly the same ownership with Allwinner
<KotCzarny> MoeIcenowy: there is no axp for h3
<MoeIcenowy> yes
<hojnikb> no i mean, integrate axp features directly to the soc
<hojnikb> so no need for an extra chip and traces on the pcb
<KotCzarny> hojnikb: maybe you need bigger elements for those amps?
tkaiser has joined #linux-sunxi
<wens> hojnikb: dcdc converter is kind of noisy, isn't it?
<hojnikb> i guess
<mdsrv> i cannot hear it
<mdsrv> so not
<hojnikb> but i could be done
<KotCzarny> mdsrv: you cannot see/hear wireless or electricity, but you use it
<hojnikb> but i guess they just rather sell axp chips separately :)
<mdsrv> but its not noisy
<hojnikb> and use up older fabs :)
<KotCzarny> electricity and wireless are noisy as hell
<mripard> I guess it also heats a lot
<mdsrv> no, because u cannot hear them
<mripard> and the SoC already have enough components that generate heat
<hojnikb> maybe stack power managment on top of the soc itself
<beeble> multi die package does not make it cheaper :)
<wens> hojnikb: that doesn't help the thermal issue :p
OtakuNekoP has quit [Remote host closed the connection]
<hojnikb> thats true i guess
<beeble> power is also a different process as the soc. so it makes no sense to do it in 28nm tsmc
<KotCzarny> someone should upload better quality pic of opipc
<KotCzarny> the one on wiki page is blurry, cant read board markings
popolon has quit [Ping timeout: 252 seconds]
<hojnikb> not necessarily
<hojnikb> you could do power managment on 28nm
<hojnikb> if intel can do it at 14nm :)
<beeble> intel has its own fabs
<beeble> so doing it inhouse makes sense
<wens> even so, if they have more valuable stuff to build, power management might get kicked aside
<beeble> buying your manufacturing capacity external you want to use the high tech waver space for as much socs as you can
<hojnikb> i guess so
<beeble> axp can be done in china mainland
<hojnikb> and you want as high yields as possible
<hojnikb> and axp could be fabbed on a cheaper node
<miasma> KotCzarny: i could take one with a dslr :)
<hojnikb> what are those made
<hojnikb> 130nm
<hojnikb> 90nm ?
<KotCzarny> miasma, make sure lighting is good :)
<tkaiser> Ntemis: Yesterday or the day before, check IRC log
paulk-minnie has joined #linux-sunxi
<tkaiser> Ntemis: linux-sunxi collects knowledge also in a wiki: http://linux-sunxi.org/H3#Documentation -- please open H3 user manual and do a simple search for these three digits: 672
<tkaiser> Ntemis: And then use a multimeter, check the test point and if that's anything other than 1.35V on your board *then* it gets interesting.
<Ntemis> hmm
<Ntemis> interesting indeed
<tkaiser> Ntemis: You mean 'reality'?
<beeble> hojnikb: depends. but pmics in 250 and 180nm would not be surprising to see
<Ntemis> no i mean did those measures where under load?
<paulk-minnie> anyone got a close look at that NES board?
<paulk-minnie> is there exposed UART?
<Ntemis> can it feed 1.35v sastained
<jelle> paulk-minnie: yes
<Ntemis> constantly
<hojnikb> Ntemis: nothign a heatsink can't fix :D
<KotCzarny> find the test point and check?
<tkaiser> Ntemis: Please do us a favour and do this test yourself and only then start to spread FUD again
<beeble> paulk-minnie: yes, uart0 is exposed
<Ntemis> tkaiser: clear
<hojnikb> i wonder why Nintendo used slc flash :D
<tkaiser> jelle: this community here is really strange.
pg12 has quit [Ping timeout: 268 seconds]
<jelle> tkaiser: who?
<jelle> paulk-minnie: oh wait not there :p
<tkaiser> jelle: linux-sunxi community ignoring linux-sunxi wiki all the time!
<KotCzarny> hojnikb: fewer returns ?
<NiteHawk> paulk-minnie: we have links/preliminary docs at http://linux-sunxi.org/Nintendo_NES_Classic_Edition
<jelle> tkaiser: oh my fault :P
<hojnikb> KotCzarny: mlc flash would be just fine for the purpose and would be somewhat cheaper.
<tkaiser> jelle: np but maybe half of 'open questions' are already answered and available in the wiki.
<KotCzarny> you can use mlc flash as slc ;)
<hojnikb> yes you can :)
<hojnikb> i guess using slc means this piece of kit won't be useless in 5-10 years
pg12 has joined #linux-sunxi
<jelle> tkaiser: seems I keep forgetting ;-)
<hojnikb> using a newer tlc for example would lead to data curruption within a year or so
vishnup has quit [Ping timeout: 240 seconds]
<jelle> tkaiser: hey, but it's better then the amlogic wiki!
<beeble> also they are only using a 512mbit flash. you can't get that as anything else then slc
* jelle has no clue about the u-boot status for his S905X
<beeble> mlc starts at usually at 16gbit
<hojnikb> beeble: i guess you're right.
<hojnikb> flash dies nowadays are usually 128Gbit or more
<hojnikb> smaller devices are usually just using faulty 128Gbit dies
<paulk-minnie> nice, thanks
<hojnikb> just fyi
<hojnikb> on the subject of flash
<hojnikb> i bought a cheap tlc flash drive in early 2016
<hojnikb> copied a testfile on it for fun
<hojnikb> left it unpowerred for 7 months
<tkaiser> paulk-minnie: What's so special with an expensive device featuring almost no DRAM and storage and lacking input/output capabilities?
<hojnikb> well i did checks every few weeks or so
<hojnikb> well curruption started on 7th month
<hojnikb> so there's tlc flash for ya
<KotCzarny> there is tlc flash and cheap tlc flash
vishnup has joined #linux-sunxi
<KotCzarny> uh, i was going to think im blind and couldnt find 1_5V test point. its on the bottom side ;)
<hojnikb> even the most expensive tlc flash will be curruption within a few years of being unpowered.
<hojnikb> thats just the sad state of flash nowadays
<paulk-minnie> tkaiser, indeed
<KotCzarny> sadder thing is that reading disrupts flash too
<paulk-minnie> tkaiser, not a very interesting machine
<paulk-minnie> tkaiser, and no mcc support→
<paulk-minnie> …
<beeble> hojnikb: jeded defines data retention only for a year
<beeble> at 30 degrees for consumer grade stuff
<KotCzarny> beeble: that's braindead.
<hojnikb> thats for exhausted p/e
<beeble> *jedec
<beeble> page 7
<hojnikb> i know about jedec
<hojnikb> i also know toshiba 15nm doesn't comply with this
<hojnikb> :)
<beeble> the thing is
<beeble> there are not a lot of electrons left in the cell
<hojnikb> yeah electrons leak over time
<beeble> physics is a bitch :)
<NiteHawk> ^ this :D
<KotCzarny> those electrons! its all their fault.
<hojnikb> there are only a few eletroncs inside a 15nm flash cells
<hojnikb> crazy
<beeble> just always remember that NAND is only working at all because they put a lot of ECC in
<hojnikb> NAND is _shit_ without ecc :)
<beeble> and it's required all the time
<hojnikb> i still strongly believe 2d tlc shoudln't be used without ldpc
popolon has joined #linux-sunxi
<hojnikb> most flash drives and cards still use bch and tl
<hojnikb> c
<hojnikb> it's crazy
<hojnikb> and it gets even wrose
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<KotCzarny> hmm, palemoon is nice in a way that it uses gstreamer to play html5 videos
pg12 has quit [Ping timeout: 245 seconds]
<hojnikb> thats pretty cool
<hojnikb> firefox doens't do that ?
ElBarto has joined #linux-sunxi
pg12 has joined #linux-sunxi
<KotCzarny> oh, it does too
<miasma> KotCzarny: my lights seem to suck. they reflect too much from the chips so you can't read the text, but what do you think? more readable? http://users.utu.fi/jmjmak/tmp/DSCF9904.JPG
<miasma> it's mostly ambient light
<KotCzarny> yup. almost perfect
<miasma> i've another one for the v1.2 version
<miasma> ok so i'll upload them
<KotCzarny> please do, as they are different
<KotCzarny> imo, you can replace the ones that are already present (blurry/dark ones)
<miasma> thanks :)
IgorPec has joined #linux-sunxi
<tkaiser> miasma: Though I would prefer this http://kaiser-edv.de/tmp/oBO5Ru/OPi_PC_Rev_1.3.jpg ;)
<wens> tkaiser: that's cheating! :p
<KotCzarny> overexposed :P
thc013 has joined #linux-sunxi
<thc013> what is te login for the debian server h5 img
<KotCzarny> thc013: who knows what you are talking about?
<miasma> what os image?
<KotCzarny> i think he is talking about xunlong's 'debian server' img
<tkaiser> thc013: the one from yesterday had debian/debian, the one from today root/orangepi
<thc013> the debian server image that xulong posted yesterday and what looks from somebody else
<KotCzarny> if one cant hack into local img and get root, one is not worthy!
Putti has joined #linux-sunxi
<miasma> maybe i'll upgrade to studio light after christmas. then, sunxi will have better product pictures than the original vendors ;P
<miasma> *lights
<thc013> ow oke if it is hackable i gonna search and yes i'm noob but learned to first search and then ask
<tkaiser> miasma: If you improve exposure a little bit it would already be great ;)
<KotCzarny> thc013: i would search on orange pi forums
<miasma> tkaiser: yea'll i'll tune them a bit in lightroom
<tkaiser> KotCzarny: why?
<KotCzarny> tkaiser: why what?
<thc013> i did and ony found orangepi but that didn't work
<thc013> thnx tkaiser had the one from yesterday
fkluknav has quit [Ping timeout: 260 seconds]
cnxsoft1 has joined #linux-sunxi
<miasma> tkaiser: does any h3 orange pi board need heatsink in normal use if you use armbian kernels?
<tkaiser> miasma: Only the small ones if you use mainline kernel and there a branch without ths/throttling and if you plan to run insanely demanding stuff. Otherwise not
<miasma> right
cnxsoft has quit [Ping timeout: 265 seconds]
cnxsoft1 is now known as cnxsoft
<tkaiser> miasma: I couldn't do a heatsink better
<KotCzarny> o.O
<KotCzarny> i hope they put some stress underneath too
thc013 has quit [Quit: Page closed]
cnxsoft has quit [Client Quit]
codekipper has quit [Quit: Page closed]
afaerber has joined #linux-sunxi
afaerber has quit [Ping timeout: 260 seconds]
afaerber has joined #linux-sunxi
<hojnikb> orange pi pc v1.2 and 1.3
<hojnikb> only dram has changed
<hojnikb> something else ?
<miasma> i don't think so
<hojnikb> at least they are using samsung ram
<hojnikb> instead of the crap elpida and hynix put out
fkluknav has joined #linux-sunxi
<ssvb> hojnikb: dram changed from what exactly and to what?
<hojnikb> different part number
<hojnikb> still samsung
<ssvb> ok, I see that somebody has documented this already - https://linux-sunxi.org/Orange_Pi_PC
pekka10 has joined #linux-sunxi
f0xx has quit [Ping timeout: 246 seconds]
<ssvb> so the dram reliability may be a bit different in Orange Pi PC revision 1.3
<hojnikb> yep
<hojnikb> maybe samsung stopped supplying that part and they decided to switch
perr has quit [Remote host closed the connection]
<NiteHawk> did someone take a look at the sources yet that Nintento provides for the NES classic - anything interesting in there? https://www.mail-archive.com/linux-sunxi@googlegroups.com/msg19625.html
vishnup has quit [Ping timeout: 256 seconds]
<tkaiser> IIRC yesterday someone reported here spotting on Orange Pi Zero (?) schematics that SD card can also be driven with 1.8V and would allow for higher speeds. Unfortunately I'm too dumb to read schematics: https://linux-sunxi.org/File:Orange-Pi-Zero-Schanetics-v1_11.pdf
<miasma> schanetics :)
<tkaiser> NiteHawk: Has been discussed yesterday here. For whatever reasons a lot of blobs while at the same time in tinalinux SDK the same stuff has been released open
<NiteHawk> a mix of schematics and shenigans?
<NiteHawk> tkaiser: i see, thanks
<willmore> miasma, that heatsink scares me.
silviop has joined #linux-sunxi
<silviop> Hi
<willmore> Anyone with a Opi zero: can you see if R445 and C80 are populated on the board? They should be by the NOR flash slot.
<silviop> My A33 tablet seems not switch on VBUS1 s, i din't see wifi-bt adapter(only VBUS2 otg in host mode bus), does someone know what i need to add to dts to force switch on ?
<plaes> silviop: what tablet?
f0xx has joined #linux-sunxi
<silviop> script.fex is: [wifi_para]
<silviop> wifi_used = 1
<silviop> wifi_sdc_id = 1
<silviop> wifi_usbc_id = 1
<silviop> wifi_usbc_type = 1
<silviop> wifi_mod_sel = 7
<silviop> wifi_power = "axp22_dldo1"
<silviop> wifi_power_ext1 = "axp22_dldo2"
<silviop> wifi_power_ext2 = ""
<silviop> rda5990_wl_host_wake = port:PL07<4><default><default><0>
<silviop> rda5990_bt_host_wake = port:PL06<4><default><default><0>
<silviop> rda5990_lpo_use_apclk = 1
<silviop> q8 generic tabòet 1280600
<beeble> tkaiser: i don't see any way to drive the sd card with 1.8v. but i couldn't find a h3 electical datasheet to see the port to voltage rail mapping. but since there is only a 3.3v rail i'm not sure how it would work with 1.8v signaling if there isn't a buildin regulator
JohnDoe_71Rus has joined #linux-sunxi
premoboss has quit [Ping timeout: 250 seconds]
<silviop> i added host mode only for USB2 and work like charm, now i need to bringup wifi
<KotCzarny> willmore: its populated on this pic: http://linux-sunxi.org/images/4/46/OPi_Zero_Bottom.jpg
<silviop> on BUS1
<KotCzarny> if i look right its on the bottem in the centre
<KotCzarny> (C80)
<willmore> Looking at the power chain for the uSD card: DC_IN->DC_5V->Vcc_IO->Vcc3V3_SDC->Vdd on TF slot. So, no, it's fixed 3.3v.
<KotCzarny> R445 too (this one is nearby flash)
<tkaiser> beeble: willmore: thanks
<willmore> KotCzarny, thanks. That's the support chips for the NOR flash. I was hoping they would be populated even if the NOR wasn't.
<KotCzarny> but it would be nice to confirm with other boards too
<willmore> tkaiser, these schematics are a nightmare, but I'm familar with they way they organized it now, so tracing other would be easier. Any other questions?
<willmore> That's for the Opi zero. Other boards may differ.
<KotCzarny> i meant, other opi zero boards
reinforce has quit [Quit: Leaving.]
<willmore> KotCzarny, understood. I saw beeble's H3 reference, so I wanted to make the context of my voltage comment clear.
<tkaiser> willmore: You answered one already. So NOR flash requires soldering a few more components. We can only hope that they start to ship out the next batch fully equipped
<willmore> tkaiser, from what KotCzarny said, the R445 and C80 are already populated. You should be able to put on the little NOR chip and be good to go.
fvogt has joined #linux-sunxi
<willmore> Supposedly, I have a tube of 16Mb NOR chips in the mail from China. Let's see if they get here before my Opi 0.
<willmore> Do we have PC 2 schematics?
<tkaiser> willmore: Sure
<willmore> Okay, I'll go look.
<tkaiser> Unfortunately schanetics still missing ;)
<willmore> The PC2 page has some. Looks like the uSD slot only gets 3.3V on that as well, so don't count on any faster storage I/O.
<tkaiser> willmore: So let's see what Olimex did on their A64-OLinuXino if/when it will be available. This board would be then the first to also implement 1.8V
young_ice has joined #linux-sunxi
young_ice has left #linux-sunxi [#linux-sunxi]
<wens> only a few chips have separate VCC-PF pins
<willmore> tkaiser, schematics for that available, yet?
<hojnikb> What does it mean that sd card is powered with 1.8V ?
<willmore> PF, wens?
<wens> willmore: PF pins are for mmc0
<willmore> hojnikb, uSD cards require 1.8V for the higher speed modes. If they are stuck at 3.3V, they can't go into the high speed modes.
<wens> the internal buffers must also run at 1.8V
<hojnikb> oh
<tkaiser> willmore: Somewhere below https://github.com/OLIMEX/OLINUXINO/ as usual?
<hojnikb> that makes sense
<willmore> wens, okay, makes sense.
<hojnikb> whats the maks speed for 3.3v then ?
<hojnikb> 25MB/s ?
<wens> hojnikb: 50 MHz 4bit SDR, 25 MB/s or so
<willmore> tkaiser, sorry, I have never touched any olimex hardware and I am unfamiliar with their stuff. I'll go look.
<wens> yup
<tkaiser> willmore: Not needed.
<hojnikb> that sux
<hojnikb> why cant xunlong wire sd cards to the faster mmc controller
<hojnikb> h5 apperently supports up to 150Mhz
<tkaiser> hojnikb: In reality it maxes out at ~23 MB/s on all sunxi boards.
<KotCzarny> because no one told them so?
alexxei has quit [Ping timeout: 268 seconds]
<hojnikb> i believe even h3 supports faster modes
<willmore> tkaiser, the A64-OlinuXino is fixed 3.3V.
<tkaiser> willmore: O, then it's just eMMC? Combined with an ultra slow industrial eMMC chip :)
<hojnikb> h3 supports ddr 50Mhz @ 8 bits
<willmore> The eMMC on that board can be at 3.3V or 1.8V, but the uSD is stuck at 3.3. Silly.
alexxei has joined #linux-sunxi
simosx has joined #linux-sunxi
<hojnikb> i don't know why that can't be used for sd cards as well
<willmore> Because SD cards are 4 bit?
simosx has quit [Remote host closed the connection]
f0xx has quit [Ping timeout: 258 seconds]
<hojnikb> what about ddr then ?
<hojnikb> that could be good for 50MB/s
<hojnikb> nevermind, sdcards don't support ddr either
<hojnikb> silly me
<beeble> with uhs it supports ddr
<willmore> UHS2... :)
* willmore is not holding his breath
<hojnikb> oh
<beeble> even uhs-i should support ddr
<willmore> So, for all boards I've looked at 25MB/s seems the max.
<hojnikb> basically sd reader on allwinner sux
<beeble> iirc
<hojnikb> they can't even provide uhs-i
<willmore> beeble, Doesn't UHS-2 do 8 bit?
<beeble> i'm not sure. haven't looked very deep into uhs-2
* willmore has to run. later.
<hojnikb> i was alway under the impression that emmc is just a socketed sd card
<beeble> never used sd card primary storage
<willmore> hojnikb, nope. Not for a while.
<hojnikb> and that same protocols apply
<hojnikb> guess i'm wrong :)
<beeble> emmc is way cooler
<beeble> boot block, trim, health status etc
nove has joined #linux-sunxi
<tkaiser> hojnikb: https://en.wikipedia.org/wiki/Secure_Digital#Transfer_modes (maybe the answer why cheap SBC do not implement anything other than 50 MHz / 4 bit)
<tkaiser> Ah no, wrong
<hojnikb> i still think the whole sdio/mmc interface is dumb
<hojnikb> i mean sd cards and emmc already have flash controllers
<hojnikb> so why bother with crappy mmc interface
<hojnikb> use the same formfactor but something like pcie or usb3 instead
<miasma> the sd interface also supports other devices
<hojnikb> still
gumblex has quit [Ping timeout: 245 seconds]
<hojnikb> you can run other stuff via pcie and usb just as well
<miasma> i guess emmc is still cheaper than the pcie ssd modules
Pepe has joined #linux-sunxi
<tkaiser> hojnikb: USB3 and PCIe on a $7 device? Really?
<hojnikb> why not
<hojnikb> :)
<hojnikb> i'm not talking about cheap arm devices
<hojnikb> but devices in general
<hojnikb> mmc prodocol needs to die
<hojnikb> there are better replacements out there
<miasma> so buy the up board
<miasma> ok it also uses mmc, but it has usb3 :)
<hojnikb> :)
<hojnikb> thats not the point
<hojnikb> i'm trying to make
IgorPec has quit [Ping timeout: 260 seconds]
<hojnikb> i'm just trying to point out, that sd cards could use usb3 instead of uhs-i or ii
<tkaiser> miasma: USB3 is inefficient. If we don't get at least NVM Express on $7 SBC I don't need any storage!!1!
<miasma> :P
<hojnikb> you don't need pcie for nvme :)
<KotCzarny> usb sucks
<KotCzarny> too fragile
<miasma> tkaiser: you can double the flash bandwidth by using a filesystem with a transparent lz4 compression :) the cpu requirements for lz4 and the speed of the flash usually go hand in hand
<hojnikb> mmc sux harder :)
<tkaiser> miasma: I know and I use this extensively.
<hojnikb> i'm talking about usb as a protocol not as a connector
<miasma> tkaiser: cool, squashfs or btrfs?
<tkaiser> miasma: In productive use mostly ZFS (Solaris, FreeBSD), on Linux btrfs
<miasma> ah
<hojnikb> no love for f2fs ?
<hojnikb> :D
<miasma> tkaiser: might be harder to achieve any bandwidth improvements with lz4 and zfs :) unless the system has 10G ethernet
lamer14792248312 has joined #linux-sunxi
lamer14792248312 has quit [Client Quit]
<hojnikb> just fyi
<hojnikb> sd cards actually support erase command, which is somewhat similar to trim
tkaiser has quit [Ping timeout: 260 seconds]
<TheLinuxBug> KotCzarny: per your "all other aps either crashed or were unavailable" this is most of the time caused because the chinese users always install miracast and some other app which if you don't install terminal and kill them off manually in their Android versions and then either remove them they plauge you with a very slow machine. I honestly have NO idea why they package these apps as they are a complete waste of resources..
<hojnikb> they get paid to bundle such apps
<hojnikb> same thing with pc manufacturers
<TheLinuxBug> yeah and the apps are unusable and basicly crash Android
<TheLinuxBug> then a lot of people think it works like shit
<TheLinuxBug> they are pretty dumb to package that stuff imo
<TheLinuxBug> BananaPi did that as well in their Android Images
<TheLinuxBug> pissed me off
<TheLinuxBug> cause even if you don't use miracast for some reason its in the abckground using 100% cpu resources
<KotCzarny> TheLinuxBug: im just talking about first-time experience, will probably try to build something decent
<hojnikb> everything to get their margins up
<hojnikb> thats one way to do it
<TheLinuxBug> yeah well the baord vendors fuck that up by installing garbage
<hojnikb> they dont care about user experience
<KotCzarny> if you have any easy guide i wonldnt mind
<TheLinuxBug> unfortunately it ruinss the out of box experience
<hojnikb> buy us fuck you
<hojnikb> thats their attitide :)
<TheLinuxBug> lol
<TheLinuxBug> KotCzarny: use that crap app store they provide and download the terminal app
<TheLinuxBug> then you can use ps or top
<TheLinuxBug> and see it running and kill it
tkaiser has joined #linux-sunxi
<TheLinuxBug> then go search for the apps file and del it
<TheLinuxBug> so it doesn't boot at start
paulk-minnie has quit [Ping timeout: 256 seconds]
<TheLinuxBug> but again, I have a vendetta against Xunlong
<TheLinuxBug> cause they are assholes and spent no time with the SDK at all
<TheLinuxBug> they just built and packaged in their garbage
<hojnikb> they are a hw manufacturer
<hojnikb> they dont give fuck about software
<TheLinuxBug> leaving the selinux garbage in which makes it impossible to root
<TheLinuxBug> which means no google store
<TheLinuxBug> which means their Android images are total trash
<TheLinuxBug> and also they spent so little time they didn't even fix DRAM settings correct
<TheLinuxBug> so if you try 1080p on their stock android
<TheLinuxBug> the board will freeze
<tkaiser> TheLinuxBug: Well known so why complaining. Use any of the Zido or whatever Android images for H3 devices and be happy. After exchanging sys_config.fex
<hojnikb> anyone tried android on h5
<TheLinuxBug> I truly wish Xunlong would stop putting out so many shit boards and spend some time actually developing good software stacks for their stuff
<hojnikb> i think xunlong should just outsource software development
<TheLinuxBug> tkaiser: if you know of an already working image or a page that shows the correct way to make those changes you insist are so easy I will do so
<hojnikb> pay armbian folks to do it right
<TheLinuxBug> I spent like 10 hours TRYing to do that
<TheLinuxBug> it is not as easy as you think
<TheLinuxBug> well at least for me it wasn't
<hojnikb> i don't know
<TheLinuxBug> maybe with your skillz its easy
<tkaiser> hojnikb: I was interested in F2FS some time ago and did also some tests with good SD cards. No benefits so why choosing it? If I go away from ext4 then due to data integrity and then it's btrfs. See Armbian forum, Free section and the pinned 'SD card performance' thread. For the first Samsung cards there I always also tested F2FS. Not impressive
<hojnikb> zidoo image works fine for me
<hojnikb> even on 1080p
<TheLinuxBug> on Pi Plus 2e?
<TheLinuxBug> I got a working image for OPi PC
<TheLinuxBug> sure
<TheLinuxBug> but there isn't one for the Plus 2E
<TheLinuxBug> not that works out of the box
<TheLinuxBug> and I spent time trying to change the fex and such to no avail
<TheLinuxBug> If someone puts together a working Android for Pi Plus 2E and send me a link or posts it on their forum and sends me a link I will send you $25 for yout time
<TheLinuxBug> has to be rootable
<TheLinuxBug> and has to be able to run google store
<TheLinuxBug> and DRAM settings need to be correct so it doesn't overheat
<tkaiser> Hmm... there's something wrong today...
tkaiser has quit [Quit: jIRCii - http://www.oldschoolirc.com]
<TheLinuxBug> if it was easy I would have already accomplished it my self, spent lots of hours on it trying different things
<TheLinuxBug> only thing I haven't done which if I find the motiviation I may do is build my own kernel and build my own Android from SDK but thats a whole other animal
reinforce has joined #linux-sunxi
<hojnikb> oh 2e
<hojnikb> haven tried that one
<hojnikb> only have opi pc
<TheLinuxBug> ya
<TheLinuxBug> and some people have made third part images for the Pi PC
<TheLinuxBug> and several other boards where they have hacked up set top box manufactures images to work
<TheLinuxBug> but in my attempts I couldn't get them working on the Pi Plus 2E
<TheLinuxBug> also last I checked the link to the Android SDK on Xunlongs site for orange Pi Plus 2E didn't work either
<hojnikb> that sux
<hojnikb> you get an more expensive board just to figure out its actually crap
<hojnikb> :)
<TheLinuxBug> well its made to be crap cause Xunlong is a bunch of lazy people
<TheLinuxBug> who spent no time on any images they made for the devices
<hojnikb> buy us fuck u :)
<TheLinuxBug> it is one major reason I respect IgorPec and tkaiser so much, they actually go out of their way to make a useable dist for these boards
<hojnikb> that seems to be their attitude .)
<hojnikb> yeah much respect to armbian team
<hojnikb> i would throw my opi pcs to trash otherwise
<hojnikb> or made some nice keychains or something
<hojnikb> out of them
<hojnikb> good work of armbian also made openelec better
<hojnikb> and retroorangepi
<hojnikb> nowadays i almost exlusively use oranges for retro games
<KotCzarny> TheLinuxBug: i will try to make things using their sdk (from h2 actually but we will see)
premoboss has joined #linux-sunxi
<ssvb> TheLinuxBug: Xunlong people are making good hardware, and they are able to design it and push to market really fast
<KotCzarny> if i succeed, i'll let you know (but probably it will fail to compile, he he)
<hojnikb> yup
<miasma> hojnikb: nowadays you should be able to use other arm distros too, if you just compile and patch the kernel yourself
<TheLinuxBug> yeah but they spend no time at all on their software and its obvious, a good example is KotCzarny's experience with Android which they burn to emmc and out of box experience
<hojnikb> miasma: so i can make something like ubuntu mate using armbian patches ?
<ssvb> TheLinuxBug: yes, it's just that the software is not their strong point, but they are not charging premium price either :-)
<TheLinuxBug> it makes the board seem like crap
<miasma> hojnikb: if it runs armv7 binaries and you don't need X, you can use whatever armv7 distro
<TheLinuxBug> I would pay the extra 10$ on the board if it meant good working software
<KotCzarny> TheLinuxBug: i've tried it by mistake, when sd card didnt boot. but filemanager actually was working fine (mounted my nfs shares and actually played the media)
<hojnikb> i need x :)
<ssvb> TheLinuxBug: well, looks like there is some business opportunity for software developers :-)
<hojnikb> 10$ a pop times 10k
<hojnikb> software development costs money
<KotCzarny> is it legal to charge for android build?
<KotCzarny> or it's 'software consulting' ? ;)
<hojnikb> i'm guessing its much much cheaper to design a board with off the shelf components
<ssvb> I think it's legal to charge for hours of work
<KotCzarny> hojnikb: unrealistic, once you sell 1-2 copies, they will show up on the internet ;)
<hojnikb> than making quality software for such board
fkluknav has quit [Ping timeout: 244 seconds]
<hojnikb> KotCzarny: you could lock the image to a certain hw id
<hojnikb> similar to what raspberry fundation does with mpeg codecs
<KotCzarny> usually it works via donations, if people recognize your work
<hojnikb> install a hidden cryptominer and get your revenue that way >:D
<hojnikb> jk
<ssvb> if someone were to collect 10$ from every devboard user who wants better software and sponsored the lima project, then open source drivers for mali could have been a reality by now :-)
<TheLinuxBug> what you need to do is make that known
<TheLinuxBug> cause then people would donate
<hojnikb> 100k is not enough to make a good gpu driver
<hojnikb> not even close
<TheLinuxBug> but there isn't really much talk about it in mainstream anywhere
<hojnikb> what is the purpose of lima anyway ?
<ssvb> hojnikb: it's a bit more than 100k
<hojnikb> aren't there gpu drivers for mali already ?
<ssvb> yes, buggy closed source drivers for OpenGL ES only (no real OpenGL), and also with a weird license forbidding redistribution of the blobs until they changed it in 2016
<ssvb> anyway, this was just an example
<ssvb> and people seem to be complaining about it the GPU support most :-)
<ssvb> there are surely a lot of other things that could be improved
<hojnikb> could utgard even do regular open gl ?
<ssvb> well, there is the glshim project - https://github.com/lunixbochs/glshim
<jelle> hmm does anyone know where apritzel's WIP H5 u-boot stuff is located?
paulk-collins has joined #linux-sunxi
<miasma> hojnikb: i thought the desktops are moving towards opengl es
<miasma> it's basically sufficient for most tasks
<hojnikb> you mean wayland
<ssvb> miasma: how so? the open source drivers are all implementing full OpenGL
<wens> jelle: check his github repo
<hojnikb> is there even a gpu with good open source drivers ?
<hojnikb> for arm
<ssvb> miasma: it's one of the very rare exceptions
<ssvb> you can't generalize
fvogt has quit [Quit: Konversation terminated!]
<jelle> wens: thanks
<wens> jelle: i think he just updated it today
<ssvb> basically, OpenGL ES support is very poor in linux, and rather than porting application to OpenGL ES, the drivers are developed to support full OpenGL
<hojnikb> ssvb: that makes sense
simosx has joined #linux-sunxi
<hojnikb> i'm guessing arm will never bother making opengl drivers
whaf has quit [Quit: Leaving.]
<hojnikb> although i wonder why they bothered with es in the first place
whaf has joined #linux-sunxi
<ssvb> this makes the proprietary binary graphics drivers for arm third class citizens in linux
<miasma> ssvb: technically I don't really understand why linux is so stuck with traditional OpenGL. the latest ES is pretty powerful and most desktop applications don't do any magic tricks with the GPU
<ssvb> es makes sense in iOS and Android
<ssvb> because people are making money developing proprietary games
<jelle> hmm seems getting it to work for the h5 requires some effort hrrm
whaf has quit [Client Quit]
<miasma> also most applications depend on gui toolkits, not directly on opengl
fkluknav has joined #linux-sunxi
<ssvb> if somebody sponsors porting Linux applications to OpenGL ES, this may happen too
whaf has joined #linux-sunxi
<hojnikb> what does it mean exactly now that pc2 has spi flash ?
<hojnikb> we can put u-boot there
<hojnikb> would that make things easier ?
<miasma> ssvb: clutter also supports opengl es? https://wiki.gnome.org/Projects/Clutter
<ssvb> hojnikb: yes, because Xunlong can eventually start providing U-Boot on their boards out of the box
<hojnikb> so having u-boot onboard means no need for specific images/kernels for each board ?
<KotCzarny> last time i've checked about:gpu in chrome on thinkpad t500 (amd), it was using gles2.0
<ssvb> hojnikb: this will be a lot easier for noobs, who expect that they can plug a hdmi cable, usb keyboard, a power plug and get some signs of life on their monitor or tv :-)
<hojnikb> :D
Gerwin_J has joined #linux-sunxi
<ssvb> miasma: in theory yes, but in practice you will encounter a ton of bugs and problems
<ssvb> miasma: because almost nobody is testing this stuff
<KotCzarny> but it would allow booting from usb/sata/network etc
<ssvb> KotCzarny: yes, this too
<KotCzarny> maybe even doing iot things if someone fits in 2MB with metal banging
<hojnikb> is there only 1MB of flash ?
jernej has joined #linux-sunxi
<KotCzarny> it depends what chip they/you use
<KotCzarny> xunlong advertised 8Mbit chips, but apparently sent boards with 16Mbit ones
<hojnikb> hahahaha :D
<hojnikb> good guy xunlong :D
<KotCzarny> nah, probably just it was cheaper ;)
<hojnikb> :)
<hojnikb> probably
<hojnikb> 2MB is enough to fix some basic openwrt or something
<miasma> how would the xunlong provided u-boot work in practice? would it have some hard coded locations for dtb/kernel/initramfs?
<hojnikb> fit*
<KotCzarny> miasma: its up to us mostly
<KotCzarny> as they dont know a thing about software and just piece things from what they scrap on forums
<miasma> but you'd need to update the dtb with a new kernel right?
<ssvb> advertising 8Mbit makes sense because they can use 8Mbit in later revisions and avoid receiving hate mail from the users :-)
<hojnikb> :D :D :D
<ssvb> it's more like >=8Mbit
<hojnikb> or maybe they just like the number 8
<KotCzarny> miasma, maybe by reading uenv.txt from boot source?
<KotCzarny> :P
<hojnikb> chinese folks have some wierd hardon on number 8
<hojnikb> :)
<ssvb> also they don't advertise the CPU clock speed for H5 boards, seems like they learned their lesson
<miasma> the spi could also contain all the schanetics and manuals :)
<miasma> since the web is so slow in china
<miasma> maybe rot13 encoded if the government is worried about IPR
<KotCzarny> hojnikb: but 64Mbit is 8*8!
<KotCzarny> ;)
<KotCzarny> someone hint xunlong about that, quick
<ssvb> some people really don't like when they bought a supposedly 1.6GHz board, which can only work at 1.2GHz or 1.3GHz in practice
<wens> plaes: updated both mainline kernel and u-boot progress pages
<hojnikb> 64Mbit would eat into their margins though :)
vagrantc has joined #linux-sunxi
<KotCzarny> hojnikb: wouldnt you pay a bit more for 8MB of boot space?
<KotCzarny> its enough for uboot,kernel and minimalistic rootfs
<hojnikb> yes i would
<hojnikb> 50c tops though :)
<KotCzarny> are there socketed spi chips?
<ssvb> somebody mentioned that 16Mbit of SPI flash is common on x86 computers
<ssvb> so probably it's the cheapest option because of the high volume
<ssvb> it might be that the 8Mbit chips are harder to find and they cost the same or more :-)
IgorPec has joined #linux-sunxi
<hojnikb> now they actually advertise 2MB on their site
<hojnikb> so i guess 16Mbit is here to stay
<ssvb> ok, makes sense
IgorPec has quit [Client Quit]
<victhor__> hi. I'm wondering, is it possible to use TVDAC to output arbitrary values, from, say, DMA? instead of only being able to output encoded video from the framebuffer. I don't think I have been able to find TVDAC documentation, on the H3 datasheet at least...
Andy-D has joined #linux-sunxi
IgorPec has joined #linux-sunxi
hojnikb has quit [Quit: Page closed]
miasma has quit [Ping timeout: 245 seconds]
miasma has joined #linux-sunxi
IgorPec has quit [Ping timeout: 250 seconds]
whaf has quit [Quit: Leaving.]
whaf has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
victhor__ is now known as victhor
popolon has quit [Quit: WeeChat 1.4]
fkluknav has quit [Ping timeout: 265 seconds]
JohnDoe_71Rus has quit [Quit: KVIrc 4.9.2 Aria http://www.kvirc.net/]
JohnDoe_71Rus has joined #linux-sunxi
<premoboss> hi, in boot.sct i have parameter console=ttyS0,115200. i want to silent de debug message at boot. i tried to eliminate console=ttyS0,115200 and redo mkimage, but no way, i still get messages on debug port. my gola is to eliminate 100% debug message on ttyS0, how can i do?
<buZz> premoboss: just add 'quiet'
<buZz> also, read /Documentation in the kernel source some day ;)
<KotCzarny> recompiling without earlydebug could also help
<premoboss> i read a lot of documentation but no find right suggestion, so i came here for help.
<buZz> :)
<buZz> 'quiet' is a kernel option anyway, described in kernel docs
<premoboss> so, i must replace console=ttyS0,115200 with quit? ok.
<premoboss> quit/quiet-*
cptG_ has joined #linux-sunxi
<KotCzarny> but dont go evil way of using us to produce evil 'unhackable' blobby image
<premoboss> i must silent ttyS0 bvecause i have a device connected to it, and debug message make it not working. i must have ttyS0 silent duting boot (no debug) so i can manage he device proprerly.
<premoboss> mo board has only 2 ttyS0/1 and i need both.
<KotCzarny> um, you will probably have to check if you have earlydebug enabled
<vagrantc> won't the boot firmware also output to serial console?
<buZz> uboot aswell, wont it?
<premoboss> i use a13-olinuxino board.
<premoboss> i must silet everithinh, both kernel and uboot
cptG has quit [Ping timeout: 246 seconds]
Gerwin_J has quit [Quit: Gerwin_J]
terra854 has quit [Quit: Connection closed for inactivity]
massi has quit [Quit: Leaving]
IgorPec has joined #linux-sunxi
matthias_bgg has quit [Ping timeout: 245 seconds]
BigFellow has joined #linux-sunxi
<KotCzarny> amazingly, debian 8.6 fresh install comes with kernel 3.16.0
<KotCzarny> so, all in all allwinner is not that far behind, hehe
phipli has joined #linux-sunxi
<miasma> overlayfs introduced in 3.18 iirc :) still pretty unusable for me
<vagrantc> though debian will likely release a new version in early-mid 2017
<KotCzarny> shucks. i was too optimistic about that android package thinking 12GB would suffice
<KotCzarny> time to remove few steam games on this laptop
<KotCzarny> lets retry with 32GB
simosx has quit [Quit: Yakkety Bye!]
<diego71> miasma: you can install recent kernel from backports in debian
<diego71> if you have a serial console for uboot you can give a try (sometimes kernel from backports have bugs :) )
<miasma> i'm not really interested in the allwinner images, ty :)
<nove> my bet is unfounded rumors with nothing happen
<ssvb> nove: yeah, I don't take rumors very seriously
|Jeroen| has joined #linux-sunxi
willmore has quit [Ping timeout: 248 seconds]
barrymossel has joined #linux-sunxi
ninolein has quit [Ping timeout: 245 seconds]
ninolein has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
willmore has joined #linux-sunxi
barrymossel has quit [Quit: Page closed]
Gerwin_J has quit [Quit: Gerwin_J]
HeavyMetal has quit [Ping timeout: 260 seconds]
mzki has quit [Ping timeout: 244 seconds]
bugzc has joined #linux-sunxi
mzki has joined #linux-sunxi
nashpa has quit [Ping timeout: 244 seconds]
nashpa has joined #linux-sunxi
phipli has quit [Ping timeout: 244 seconds]
jstein_ has joined #linux-sunxi
jstein_ is now known as jstein
Ntemis has quit [Remote host closed the connection]
Mr__Anderson has joined #linux-sunxi
VargaD has quit [Ping timeout: 245 seconds]
VargaD has joined #linux-sunxi
IgorPec2 has joined #linux-sunxi
IgorPec has quit [Ping timeout: 244 seconds]
Andy-D has quit [Ping timeout: 252 seconds]
<jelle> I love the wording though: 'open-source ARM action'
bugzc has quit [Ping timeout: 260 seconds]
phipli has joined #linux-sunxi
<jelle> do I need linux-sunxi git HEAD for the H5?
florianH has quit [Quit: Connection closed for inactivity]
Andy-D has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: KVIrc 4.9.2 Aria http://www.kvirc.net/]
<jelle> ok I'm silly tonight, I obviously need aarch64 arm compiler :P
HeavyMetal has joined #linux-sunxi
HeavyMetal has quit [Changing host]
HeavyMetal has joined #linux-sunxi
jinzo has joined #linux-sunxi
afaerber has quit [Quit: Ex-Chat]
Ntemis has joined #linux-sunxi
|Jeroen| has quit [Quit: dada]
IgorPec2 has quit [Quit: Nettalk6 - www.ntalk.de]
yann-kaelig has quit [Quit: Leaving]
iamfrankenstein1 has quit [Read error: Connection reset by peer]
netlynx has quit [Quit: Ex-Chat]
premoboss has quit [Ping timeout: 240 seconds]
souther has quit [Ping timeout: 260 seconds]
bugzc has joined #linux-sunxi
souther has joined #linux-sunxi
HeavyMetal has quit [Ping timeout: 258 seconds]
HeavyMetal has joined #linux-sunxi
HeavyMetal has joined #linux-sunxi
jinzo has quit [Remote host closed the connection]
HeavyMetal has quit [Ping timeout: 260 seconds]
Mr__Anderson has quit [Ping timeout: 248 seconds]
Mr__Anderson has joined #linux-sunxi
Net147 has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
HeavyMetal has joined #linux-sunxi
HeavyMetal has joined #linux-sunxi
HeavyMetal has quit [Changing host]
<beeble> KotCzarny: a repo checkout is about 100gb clean. 150gb if build
<beeble> if you plan to build more then one time you may want to setup ccache, then you should plan for even more
<beeble> i would free at least 200gb if you want to play around
mzki has quit [Quit: leaving]
ikmaak has quit [Read error: Connection reset by peer]
ikmaak has joined #linux-sunxi
apritzel has joined #linux-sunxi
<apritzel> jelle: what do you want to build? an H5 kernel?
nove has quit [Quit: nove]
lerc has quit [Ping timeout: 265 seconds]
perr has joined #linux-sunxi
lerc has joined #linux-sunxi
Pepe has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
vbmithr_ has quit [Quit: leaving]
vbmithr has joined #linux-sunxi
BigFellow has quit [Quit: BigFellow]
Mr__Anderson has quit [Remote host closed the connection]
victhor_ has joined #linux-sunxi
victhor has quit [Ping timeout: 245 seconds]
victhor_ has quit [Ping timeout: 250 seconds]
jernej has quit [Ping timeout: 256 seconds]