rellla changed the topic of #linux-sunxi to: Allwinner/sunxi /development discussion - did you try looking at our wiki? - Don't ask to ask. Just ask and wait! - - Logs at - *only registered users can talk*
jstefanop has quit [Remote host closed the connection]
jstefanop has joined #linux-sunxi
jstefanop has quit [Read error: Connection reset by peer]
jstefanop has joined #linux-sunxi
dev1990 has quit [Quit: Konversation terminated!]
jstefanop has quit [Ping timeout: 255 seconds]
chomwitt has quit [Ping timeout: 245 seconds]
captainigloo has quit [Ping timeout: 250 seconds]
<lavamind> bbrezillon: alright so I've pulled in dist6/hynix patches into chip/nand
<lavamind> mtd is able to find the partitions but UBI fails with some ECC errors
<cch> curlybracket: I have tested with one board that VDD_EFUSEBP is connecting a 10uF/6.3v capacitor, using the code provided in that gist. By calling the print_eth_mac right after calling write_eth_mac, I have got 00:00:00:00:00:00
<cch> curlybracket: You can find the schematic file from the key is: kyat
<cch> curlybracket: Cloud you please help to investigate to make sure the board is ready for writing eFUSES?
<lavamind> bbrezillon: I added the H27Q to nand_ids.c and gave it the dist6 scheme but hitting the same ECC errors
<lavamind> do I need to repartition/reformat the NAND ?
<lavamind> if so, how ?
kaspter has joined #linux-sunxi
suprothunderbolt has joined #linux-sunxi
Rafael1980 has quit [Quit: Konversation terminated!]
dddddd has quit [Remote host closed the connection]
kaspter has quit [Quit: kaspter]
kaspter has joined #linux-sunxi
grosso has joined #linux-sunxi
victhor has quit [Ping timeout: 252 seconds]
nuuuciano_ has quit [Ping timeout: 250 seconds]
nuuuciano_ has joined #linux-sunxi
shfil has quit [Quit: Connection closed for inactivity]
megi has quit [Ping timeout: 252 seconds]
grosso has quit [Ping timeout: 246 seconds]
<lavamind> after adding nand-ecc-step-size = <1024>; and nand-ecc-strength = <56>; the ECC errors go away
<lavamind> but I'm left with "ubi0 error: ubi_add_to_av: two LEBs with same sequence number 4100"
<lavamind> I suppose that's the pairing scheme kicking in
grosso has joined #linux-sunxi
<lavamind> oh dear I fear I'm killing my NAND :(
<lavamind> this has started happening when I boot using 4.3.0-ntc, which used to work fine until now
<lavamind> maybe the NAND isn't dying, but my previous UBIFS mount attempts are responsible for this
lurchi_ has joined #linux-sunxi
lurchi__ has quit [Read error: Connection reset by peer]
lurchi_ has quit [Quit: Konversation terminated!]
lurchi__ has joined #linux-sunxi
<wens> clementp[m]: wdog works on my pine h64, but not my opi lite 2
<anarsoul> different soc revisions?
TheSeven has quit [Ping timeout: 240 seconds]
TheSeven has joined #linux-sunxi
<grosso> someone can explain: what's this ion_alloc / sunxi_alloc? Why and how to use?
<wens> both are h6 v200-awin, pine h64: H6448BA 7782, Opi Lite 2: H8068BA 61C2
<wens> grosso: ion allocator is the android allocator
<wens> grosso: no idea about sunxi_alloc
<wens> grosso: you seem to be using the vendor kernel, so you're pretty much on your own
<grosso> yes I'm using the vendor kernel
vagrantc has joined #linux-sunxi
<grosso> how it is supposed to use that android allocator?
<wens> no idea
agraf has quit [Ping timeout: 244 seconds]
agraf has joined #linux-sunxi
NeuroScr has quit [Quit: NeuroScr]
lurchi_ has joined #linux-sunxi
lurchi__ has quit [Ping timeout: 250 seconds]
selfbg has joined #linux-sunxi
<aalm> indeed, but i don't live there.. :]
<KotCzarny> dont worry, brave law makers around the world tend to copy ideas
<KotCzarny> if it means they will get some moneys
ildar has left #linux-sunxi ["User left"]
angelo_ts has quit [Changing host]
angelo_ts has joined #linux-sunxi
<aalm> nah, atleast here they could get way more from taxing alone already, yet things flow mostly w/o issues by customs, so i find it unlikely to get copied here(no fines like that around here)
f0xx has joined #linux-sunxi
<KotCzarny> here we have one of the biggest local ebay-like site crying to government about aliexpress
<KotCzarny> which resulted in new customs facility built just for china packages
<aalm> :/
nuuuciano_ has quit [Ping timeout: 255 seconds]
captainigloo has joined #linux-sunxi
<wens> we have the same stuff here in TW :)
reinforce has joined #linux-sunxi
<KotCzarny> the customs facility or fcc related fun?
<wens> you'll run into problems if you try to resell RF equipment without certifications
<wens> if you are importing for personal use, they still specify that you should have the stuff certified, and some people do go through with it
<KotCzarny> are sunxi boards certified in .tw ?
<wens> I think there's a limit to how many devices you can import without cert
<wens> KotCzarny: good question :p
<KotCzarny> so i guess, no
<KotCzarny> :P
<wens> the only RF thing is the AMPAK wifi module
<KotCzarny> in theory any electrical defice is rf source
<wens> KotCzarny: that's EMI regulations, separate from RF spectrum usage licensing
<KotCzarny> noiizzze
<KotCzarny> imo any device with usb3 should fall into rf spectrum usage
<KotCzarny> ;)
chomwitt has joined #linux-sunxi
<wens> I believe that classifies as incidental or unintentional radiators (under FCC rules)
t3st3r has quit [Ping timeout: 256 seconds]
t3st3r has joined #linux-sunxi
grosso has quit [Ping timeout: 255 seconds]
<suprothunderbolt> I'm attempting to get a MIPI DSI panel working and I think I'm most of the way there apart from the device tree overlay part. It currently stops the board from booting and most annoyingly I don't get any debugging to my UART, I think because the UART doesn't seem to be attached early on. Is there a way to make the UART attached early? Specifically I need UART4
<MoeIcenowy> earlycon?
<MoeIcenowy> BTW why do you use UART4 as debugging UART?
<MoeIcenowy> it's a strange configuration
<suprothunderbolt> umm... because I'm an idiot :)
<suprothunderbolt> I only routed 1 UART out on this board and didn't think of routing out the first one for debugging.
<MoeIcenowy> oops
<MoeIcenowy> sorry
<suprothunderbolt> I also didn't put a JTAG header on... next revision might have both of those
<libv> suprothunderbolt: mainline kernel howto
<libv> suprothunderbolt: early printk
<libv> if needs be, add another address
<libv> /topic says "have you looked at our wiki"
<libv> sorry, "did you try looking at our wiki"
<libv> i was even more snide back in... 2013?
<suprothunderbolt> hah, for many things I've looked at the wiki. :) I thought I might try asking a question someone knew about ;)
<libv> suprothunderbolt: early printk is where you want to be
<suprothunderbolt> thanks
<libv> i actually clarified it slightly yesterday
<suprothunderbolt> if anyone know about TDM for I2S that's my next mission after getting the panel working.
<libv> you can git grep the kernel for SUNXI_DEBUG_ to see what the addresses are
<libv> eh, bs
<MoeIcenowy> uart4 should be @ 0x01c29000
<libv> the definitions are in arch/arm/Kconfig.debug
<MoeIcenowy> if your SoC is not H6
<libv> not defined yet it eems
<suprothunderbolt> it's an A64
<suprothunderbolt> if there's an example of a MIPI DSI panel being setup using an overlay that would be very helpful
<MoeIcenowy> yes
<MoeIcenowy> it's also 0x01c29000
<suprothunderbolt> I still find the device tree stuff a little confusing
<suprothunderbolt> thanks
<MoeIcenowy> suprothunderbolt: Y do you want to use overlay?
<suprothunderbolt> MoeIcenowy, no idea. Should I avoid it?
<MoeIcenowy> and for A64 MIPI DSI you need to apply WIP patches from Jagan Teki
<MoeIcenowy> overlay should only be used when it needs to be plugged in and out
<KotCzarny> overlays are for distro developers mostly
<KotCzarny> for different hardware being easily added/removed
<MoeIcenowy> if the DSI panel is a fixed part of your device
<suprothunderbolt> oh okay so I should just edit the main device tree file?
<KotCzarny> if you have board that doesnt change, single dt is fine
<MoeIcenowy> you doesn't need to overlay
<MoeIcenowy> but edit the main DT file of the board
<suprothunderbolt> okay, none of the hardware will change / plug
<suprothunderbolt> okay. Sounds easier to navigate. I think the current problem is that I'm getting a name wrong and the compiler isn't picking it up because of the overlay magic
cmeerw has joined #linux-sunxi
<wens> there's also earlycon you can use, which is slightly easier since it's just earlycon=??? on the cmdline
clemens3_ has quit [Ping timeout: 255 seconds]
<libv> wens: with ??? being the address?
<libv> hrm, pretty extensive command line
<libv> it's almost easier to add another definition and rebuild :)
tllim has quit [Read error: Connection reset by peer]
airwind has joined #linux-sunxi
diego71 has quit [Remote host closed the connection]
<suprothunderbolt> wens, thanks for the info :)
<suprothunderbolt> just looking through the A64 MIPI patches to check things are actually working before I attempt to get my panel working
vagrantc has quit [Quit: leaving]
<wens> libv: yeah, but it initiates later than if you used earlyprintk
clemens3_ has joined #linux-sunxi
<libv> hrm, two downsides then
<suprothunderbolt> anyone know about any one getting TDM / PCM mode working for I2S? I'm attempting to get an 8 channel codec working. I'm planning to port the code from the BSP if it works but if someone has already done it that would be good too!
<bbrezillon> lavamind: yes, you have to re-format the UBI partition
<wens> libv: upside is it works with multi_v7 and you don't have to rebuild to change the address
<libv> yeah
yann has quit [Ping timeout: 250 seconds]
<suprothunderbolt> somewhat weird question: I've been building a kernel using the Armbian system and it's a bit vexing for debugging / dev with them magic patching. Can I just build it using make deb-pkg
<KotCzarny> you can build kernel without any additional tools
<KotCzarny> ie. unpack/make prepare/make something_defconfig/make bzImage etc
<KotCzarny> armbian adds some nice to have patches
<suprothunderbolt> yeah, so if I've got the cached version from the armbian tool I can just stop using their build script and run make deb-pkg in the source folder.
<KotCzarny> nah, im talking about using vanilla kernel
<beeble> don't forget to set ARCH if you cross-compile...
<beeble> (or dpkg-architecture if you are using deb-pkg)
diego71 has joined #linux-sunxi
tnovotny has joined #linux-sunxi
msimpson has joined #linux-sunxi
lurchi_ has quit [Ping timeout: 246 seconds]
lurchi_ has joined #linux-sunxi
lurchi_ is now known as lurchi__
yann has joined #linux-sunxi
Rafael1980 has joined #linux-sunxi
AneoX has joined #linux-sunxi
IgorPec has joined #linux-sunxi
jelly has quit [Read error: Connection reset by peer]
jelly-home has joined #linux-sunxi
AneoX_ has joined #linux-sunxi
AneoX has quit [Ping timeout: 250 seconds]
aballier has quit [Ping timeout: 245 seconds]
RichardG867 has quit [Quit: Keyboard not found, press F1 to continue]
RichardG867 has joined #linux-sunxi
suprothunderbolt has quit [Ping timeout: 246 seconds]
Mangy_Dog has joined #linux-sunxi
aballier has joined #linux-sunxi
victhor has joined #linux-sunxi
jelly-home is now known as jelly
BenG83 has quit [Ping timeout: 255 seconds]
\\Mr_C\\ has quit [Quit: (Read error: Connection reset by beer)]
Andy-D has joined #linux-sunxi
scroll2_ has quit [Remote host closed the connection]
<clementp[m]> wens: about wdog you test it with devmem ? Do you think it can be a sw issue ? Like Bl31 is different or secure zone are different or least possible ram size make a difference?
<clementp[m]> If not do you think use the r_wdog is ok ?
scroll2_ has joined #linux-sunxi
<clementp[m]> I didn't find anything about the hw packaging revision :(
suprothunderbolt has joined #linux-sunxi
<wens> clementp[m]: no, I test u-boot's reset command and linux reboot (which uses ATF to trigger the watchdog)
<wens> I previously (a few months back judging by the irc logs) I did test it with devmem on the Opi Lite 2
<MoeIcenowy> oh strange
<MoeIcenowy> how does U-Boot deal with DSI panel?
<clementp[m]> Ok so I will go for a hw errata introduce in a new revision. What's your opinion on using r_wdog instead ?
<MoeIcenowy> clementp[m]: I think we should find the reason
<MoeIcenowy> before introducing the errata
<clementp[m]> I'm ok for it too but if try the reset in u-boot with just trig the wdog and both soc doesn't have the same behavior
<clementp[m]> which*
<MoeIcenowy> what do you mean by "doesn't have the same behavior"?
<clementp[m]> And using the r_wdog is not a bad hack
<clementp[m]> one is stuck, the other one reboot
suprothunderbolt has quit [Read error: Connection reset by peer]
victhor has quit [Ping timeout: 252 seconds]
<wens> MoeIcenowy: any chance you can pass along the wdog issue to tllim and/or stephen and have them pass it to Allwinner?
<MoeIcenowy> I told it to Wink
<MoeIcenowy> wens: it's Steven, not Stephen
<wens> ah
<wens> my bad
\\Mr_C\\ has joined #linux-sunxi
<clementp[m]> Do you know who are the linux-sunxi wiki admin ? I would like to upload sys_config.fex of the board but files are not allowed :(
<clementp[m]> Ok so we are waiting the answer from Allwinner then...
<libv> clementp[m]: they should be uploaded to github
<clementp[m]> thanks
<clementp[m]> The number of PR is huge oO
<clementp[m]> not sure it will be accepted soon
<libv> yeah, people are preoccupied with upstream for development boards only
<libv> my time seems worthless as per usual, so i will have to sift through those
afaerber has joined #linux-sunxi
IgorPec has quit [Quit: Nettalk6 -]
willmore has quit [Remote host closed the connection]
msde has joined #linux-sunxi
<clementp[m]> Why not just accept most them, I mean it's a not critical to have bad sys_config.fex in this repo no ?
megi has joined #linux-sunxi
<libv> clementp[m]: i will
willmore has joined #linux-sunxi
disik has joined #linux-sunxi
<disik> hello, how can I test hardware crypto for h3?
<willmore> don't?
<KotCzarny> modprobe ss or sunxi_ss, then modprobe cryptodev, then openssl speed or cryptobench
<KotCzarny> or just read montjoie's page about it
<KotCzarny> generally it's slow
<willmore> It's slower than the CPU.
<willmore> And it has a lot of overhead.
<disik> so it is not recomended?
<willmore> Nope.
<willmore> The crypto units on most ARM chips are leftovers from back then single core processors were slow and couldn't do crypto fast.
<KotCzarny> willmore, marvel does it quite fast
<willmore> They're kept around because there might be some client who needs it for legacy reasons.
<disik> anyway, where can I read montjoie's page about it?
<willmore> KotCzarny, that's why I said "most".
<KotCzarny> it's nice to have ~60MB/s accel in luks
<willmore> KotCzarny, nice on single threaded. Not on quads.
<KotCzarny> willmore, the board i have it on is 2x A9
<KotCzarny> @800mhz
<disik> KotCzarny> isn't sunxi_ss fo the other hardware>
<willmore> Okay, you're marginal. :)
<KotCzarny> so useful in that case
<willmore> As long as you don't have to busy wait for it.
<KotCzarny> The Security System (SS for short) is a hardware cryptographic accelerator that supports AES/MD5/SHA1/DES/3DES/PRNG algorithms.
<KotCzarny> not to be confused with Schutzstaffel
<KotCzarny> which is nazi's special force
* willmore slaps KotCzarny
<disik> the page says sun8i-ce is for h3
<lennard> at work they're usually special-switch or storage-switch
<disik> cant't find it anywhere
<KotCzarny> although nazi's unit was also about security system
* willmore facepalms
<KotCzarny> something to think about. ;)
<KotCzarny> disik: what kernel version are you using?
<disik> 4.19.20-sunxi
<willmore> Is montjoie's work in mainline?
<KotCzarny> willmore: who's else?
<disik> it's stock armbian
<willmore> KotCzarny, vs not in mainline...
<willmore> vs in the sunxi specific kernel repo...
<willmore> vs just in an armbian patch set..
* willmore hasn't updated his armbian H3 systems recently.
<willmore> Should probably do that.
<KotCzarny> check sun4i-ss
<KotCzarny> probably got renamed at some point
<KotCzarny> hmm
<KotCzarny> nope, it's for earlier socs
<KotCzarny> maybe it's not in mainline then
wwilly has quit [Ping timeout: 258 seconds]
<disik> so is there a patch or something?
[TheBug] has quit [Ping timeout: 250 seconds]
[TheBug] has joined #linux-sunxi
<KotCzarny> yup
shfil has joined #linux-sunxi
BenG83 has joined #linux-sunxi
wwilly has joined #linux-sunxi
<clementp[m]> why nodoby upstream the mali node for H6 is there something blocking ?
<montjoie> disik: which soc do you have ?
<disik> montjoie> it's H3
<montjoie> disik: you need sun8i-ce which is not on mainline yet
<montjoie> I have some github repo that I need to update
<disik> doesn't it exist as a patch?
<montjoie> a serie of patch in that case
<disik> well i'd like to try it out, if you don't mind
<montjoie> let me update github first with it
<disik> okay thanks
wwilly_ has joined #linux-sunxi
wwilly has quit [Ping timeout: 258 seconds]
dddddd has joined #linux-sunxi
indy has quit [Remote host closed the connection]
<curlybracket> cch: Not sure how important it is, but there should also be a decoupling capacitor on VDD-EFUSE. See
\\Mr_C\\ has quit [Quit: (Read error: Connection reset by beer)]
lurchi__ is now known as lurchi_
ganbold has quit [Ping timeout: 250 seconds]
ganbold has joined #linux-sunxi
ganbold has quit [Ping timeout: 246 seconds]
\\Mr_C\\ has joined #linux-sunxi
indy has joined #linux-sunxi
ganbold has joined #linux-sunxi
ganbold has quit [Remote host closed the connection]
ganbold has joined #linux-sunxi
voodoovood2 has joined #linux-sunxi
airwind has quit [Quit: airwind]
voodoovood has quit [Ping timeout: 252 seconds]
<curlybracket> cch: also, try different offsets. Do you have an A64 board available?
msde has quit [Remote host closed the connection]
DuClare has quit [Changing host]
DuClare has joined #linux-sunxi
AneoX has joined #linux-sunxi
AneoX_ has quit [Ping timeout: 245 seconds]
<angelo_ts> hi, lookign a way to setup i2s0 on H3 as slave
<angelo_ts> cannot find any sample, any help would be great
DonkeyHotei has joined #linux-sunxi
<disik> i'll give it a try, thanks
<montjoie> disik: do not hesitate to say results of testing it
<disik> oaky :)
<disik> *okay
lurchi_ is now known as lurchi__
Andy-D has quit [Ping timeout: 252 seconds]
nuuuciano_ has joined #linux-sunxi
<MoeIcenowy> clementp[m]: yes, DT binding is blocking.
<clementp[m]> just adding a sun50i-h6-mali, t720-mali is not ok ?
<MoeIcenowy> clementp[m]: auxillary clocks and resets are needed
<clementp[m]> midgard is already here for rk why added h6 is difficult ?
<clementp[m]> ok
<clementp[m]> sorry
fkluknav has quit [Remote host closed the connection]
fkluknav has joined #linux-sunxi
<clementp[m]> is there a WIP patch ?
<clementp[m]> found it :)
<libv> clementp[m]: do we even have a sunxi-boards directory for h6?
<clementp[m]> nope
<libv> yeah
<libv> thanks for bringing this up
<libv> but it's one more of those things which induce serious eye-rolling in yours truly. seems like only very few people care about helping users help themselves and documenting boards before their vendors go bust
tllim has joined #linux-sunxi
<MoeIcenowy> libv: sunxi-boards?
<MoeIcenowy> Why do I remember it's only a FEX holder?
<clementp[m]> my beelink gs1 is shit since the first day...
<libv> it is, but the fex is an important source of information
<libv> very few devices come with schematics
<libv> and apart from cubie and olimex, schematics are a relatively recent thing for development boards
<MoeIcenowy> libv: unfortunately
<MoeIcenowy> in 3.10+ BSP era
<MoeIcenowy> FEX is no longer in memory
<MoeIcenowy> it's broken down and embedded into the DTB
<libv> and... like NTC, board makers go bust all the time
<MoeIcenowy> and at least for BSP 3.10 I didn't find any way to get DTB from working kernel
<mru> what's the point of a dev board without schematics?
<MoeIcenowy> in fact I usually purchase A33 tablets w/o schematics ;-)
<MoeIcenowy> 1. they usually use ref design 2. I can dump FEX
<clementp[m]> talking about closed tv box
<libv> mru: the reason why sunxi went anywhere, the reason why i got here
<MoeIcenowy> libv: The zone of sunxi is shrinking
selfbg has quit [Remote host closed the connection]
<libv> is because allwinner was less intelligent than the competition, and 2 vendors threw out gpled code for uboot and kernel, in accordance to the gpl
<mru> I don't expect consumer equipment to come with schematics, but dev boards really should
<libv> before that, none of the android devices adhered to the gpl
<libv> schematics are a massive jump forward from even adhering to the gpl
<libv> and that just did not happen before
<libv> do we actually have schematics for the rpi?
* mru does not care for rpi either way
<libv> ah, ok, they at least did that
<clementp[m]> I think is only available for 1, not for the others
<libv> the one i found which seemed relatively old was dated 2019
<libv> but there was a time when schematics were definitely not part of things
<libv> pin-outs, yes
<clementp[m]> Their schematics is only input/ouput and connectors
<mru> I refuse to work on customer hardware without access to schematics
<libv> and what made sunxi special was 1) kernel and uboot code, as per GPL, 2) .fex
<libv> without those two things, linux-sunxi would never have happened like it did
<clementp[m]> mru: some tv box are cheap with a good hardware, they only need a small push of their device-tree to become usable :)
<libv> MoeIcenowy: so h6 bsp has devicetree or some convoluted .fex?
<MoeIcenowy> libv: convoluted .fex
<MoeIcenowy> then injected to DT when building image
<MoeIcenowy> so final image has DTB
<libv> focus has been so singularly on mainline and on development boards, that people did not bother to work the wiki, add proper device pages (the NTC chip is the perfect example), or to write a tool to grab and convert that dtb
<mru> clementp[m]: sure, you can find hackable device, and that's fine
<mru> I'm talking about dev boards sold specifically for the purpose of, well, developing
<libv> mru: developing what?
<libv> and for whom?
<MoeIcenowy> some strange problem
<MoeIcenowy> should we add a device tree binding for a device's chassis
<MoeIcenowy> ?
<mru> huh?
<MoeIcenowy> if you're using systemd
<MoeIcenowy> you will find hostnamed maintains a device's "chassis"
<mru> is this some new meaning of the word chassis?
<MoeIcenowy> I'm not English expert
<MoeIcenowy> the word "chassis" is chosen by Systemd
<MoeIcenowy> it defines the type of the device
<MoeIcenowy> "desktop", "laptop", "convertible", "server", "tablet", "handset", "watch", and "embedded"
<MoeIcenowy> "vm" and "container"
<clementp[m]> French word taken by English because they are lacking words :)
<clementp[m]> (I'm French)
<libv> that sounds like it will become outdated very quickly
<mru> I know the meaning of the word chassis in english
<libv> and that they will stick everything under "embedded" or "misc"
<mru> that systemd application of it seems rather bizarre
<mru> "form factor" would be closer to established nomenclature
<MoeIcenowy> oh I think laptops after *lake CPU are all embedded ;-)
<MoeIcenowy> cursed
jstefanop has joined #linux-sunxi
<mru> two categories would have been enough: "things lennart has heard of" and "the remaining 99% of things"
<libv> well, a good friend and former colleague of mine has been at redhat for a few years now
JohnDoe_71Rus has joined #linux-sunxi
nuuuciano__ has joined #linux-sunxi
<libv> and at some redhat event in brno (150km away), 4-5 years ago, kai sievers (himself also former suse) with lennart next to him, stated "once kdbus is in the kernel, people will have no other option but to use systemd"
<mru> those two are made of pure arrogance
<libv> my friend is the most scientific mind i know, and the only paraphrasing is me retelling what my friend told me
nuuuciano_ has quit [Ping timeout: 246 seconds]
reinforce has quit [Quit: Leaving.]
nuuuciano__ has quit [Ping timeout: 255 seconds]
DonkeyHotei has quit [Ping timeout: 245 seconds]
danielg4 has joined #linux-sunxi
sn9 has joined #linux-sunxi
danielg4 has quit [Ping timeout: 255 seconds]
msimpson has quit [Read error: Connection reset by peer]
<MoeIcenowy> to be honest, I also do not like System -- but it doesn't prevent me from using it
<MoeIcenowy> ah a strange situation, right?
nuuuciano has joined #linux-sunxi
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #linux-sunxi
<EmilKarlson> even literal quoting is missing context and intention
disik has quit [Ping timeout: 245 seconds]
reinforce has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
netlynx has joined #linux-sunxi
t3st3r has quit [Ping timeout: 256 seconds]
t3st3r has joined #linux-sunxi
victhor has joined #linux-sunxi
shfil has quit [Quit: Connection closed for inactivity]
clemens3_ has quit [Ping timeout: 255 seconds]
<KotCzarny> MoeIcenowy: similarly genital diseases prevent you from living
<KotCzarny> until disease is so advanced and took over the host to the point host dies
<KotCzarny> which is the case of systemd and other evil software
<KotCzarny> *dont prevent
libv has quit [Read error: Connection reset by peer]
<KotCzarny> but it's only if /sys/firmware/devicetree/ is present
agraf has quit [Read error: Connection reset by peer]
libv has joined #linux-sunxi
<libv> it's sadly another one of those cases where, if you are part of the right crowd, or the right company, you will get away with murder
<libv> if an employee of say, canonical or suse would've displayed even a fraction of the disregard for users, he would've been been lynched several times over
<MoeIcenowy> KotCzarny: I of course know it
<MoeIcenowy> but on BSP 3.10 it's not present
<libv> the removal of the xmir support patches from the intel driver in 2013 was a good example of just how these things work
<KotCzarny> oh, i thought it was from the beginning. ho hum.
<libv> MoeIcenowy: is this mangled fex in dtb form in 3.10 part of the h6 bsp that i pulled from the pine site?
<MoeIcenowy> libv: it's not a part of the bsp you downloaded
<MoeIcenowy> it's built with FEX in orig FEX format when building image
<MoeIcenowy> and the DTB in kernel tree
<libv> can you point me to code which implements this?
<MoeIcenowy> oh I'm not familiar to it
<MoeIcenowy> will try
<libv> thanks :)
fkluknav has quit [Remote host closed the connection]
<MoeIcenowy> libv: thanks, the final SDK deleted source for these
<libv> update_fes1 seems interesting
<MoeIcenowy> oh they're not deleted
fkluknav has joined #linux-sunxi
<MoeIcenowy> libv: see lichee/tools/pack/pack
<MoeIcenowy> do_ini_to_dts()
<MoeIcenowy> The converter is a hacked dtc
<MoeIcenowy> omg
<MoeIcenowy> eys
<MoeIcenowy> yes
<libv> did they at least not breach copyright?
<MoeIcenowy> -F seems the key
<MoeIcenowy> and the hacked dtc is at lichee/linux-4.9/scripts/dtc/
sn9 is now known as DonkeyHotei
<MoeIcenowy> it adds a weird iniparser
<libv> master/scripts/dtc/script_parser/script.c is "Copyright (C) 2012 Alejandro Mery <>" :)
<libv> is where that parser comes from
<libv> seems like there is quite a bit of info available for anyone with some spare time to go and write the necessary tools to extract this from images and perhaps the systems dtb
cch has quit [Remote host closed the connection]
dev1990 has joined #linux-sunxi
tnovotny has quit [Quit: Leaving]
AneoX_ has joined #linux-sunxi
AneoX has quit [Ping timeout: 250 seconds]
shfil has joined #linux-sunxi
Perlovka_ has joined #linux-sunxi
Perlovka has quit [Ping timeout: 240 seconds]
<libv> anyway, i know what i can throw into the fex page while falcon heavy launches ;)
<libv> MoeIcenowy: thanks
Andy-D has joined #linux-sunxi
mknawabi has quit [Ping timeout: 268 seconds]
mknawabi has joined #linux-sunxi
t3st3rV2 has joined #linux-sunxi
t3st3r has quit [Ping timeout: 256 seconds]
afaerber has quit [Quit: Leaving]
Putti has joined #linux-sunxi
vagrantc has joined #linux-sunxi
Perlovka has joined #linux-sunxi
Perlovka_ has quit [Ping timeout: 246 seconds]
msimpson has joined #linux-sunxi
msimpson has quit [Remote host closed the connection]
xyntrix has quit [Quit: Leaving]
chewitt has joined #linux-sunxi
Putti has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
BenG83 has quit [Quit: Leaving]
BenG83 has joined #linux-sunxi
BenG83 has quit [Remote host closed the connection]
DonkeyHotei has quit [Ping timeout: 245 seconds]
nashpa_ has joined #linux-sunxi
nashpa has quit [Ping timeout: 246 seconds]
<jerbob92> Anyone here has some experience with USB gadgets and/or basic USB protocol?
msde has joined #linux-sunxi
<lennard> I've used an orange pi as a mass storage device once
<lennard> using the usb gadget driver
<libv> jerbob92: why?
<jerbob92> I'm having some specific questions about how USB works, so you probably can't answer them, but I'll ask anyway
<jerbob92> I'm trying to make a gadget that works with a very specific printer driver, some POS thermal printer driver that's configured as vendor class
<libv> jerbob92: is this a sunxi question, or do you just want someone to explain what is in the mindshare book?
<jerbob92> Not really sunxi related, but I think this is a channel where people might be users of usb gadgets
<libv> ...
<jerbob92> And also the only linux channel I'm in ;)
<libv> for me, on a a7hd with an a10 with a mali binary, it worked just fine for 2 year or so that i spent with it on the move
<jerbob92> Was just looking for someone that might had some specifics for me that could help me in PM
<libv> and the bsp kernel
<libv> jerbob92: so this would be a pass through device to a weird printer?
<jerbob92> Yeah
<jerbob92> It's working fine for printers that are actually configured to say they are a printer class
DonkeyHotei has joined #linux-sunxi
<libv> jerbob92: is it not easier to just have a small arm or mips (openwrt) machine with cups and a linux based printer driver?
<jerbob92> But this printer has 2 modes: printer class and vendor class
<jerbob92> Vendor class has some very odd USB control commmands
<libv> jerbob92: are these commands documented somewhere, or do you have to look at what flies past on a windows driver?
<jerbob92> They are not. I logged them and tweaked my gadget to mimic the responses
<jerbob92> But this driver is sending URB requests up to the IN endpoint, which sounds odd to me
<libv> ok, so you are using a gadget driver as yoru testbed, like ruslan bilovol who implemented uac3.0 :)
<libv> as you do not have constant access to the hw?
<jerbob92> I do
<jerbob92> The gadget needs to be able to mimic the printer
<jerbob92> So that the POS system thinks the printer is still connected
<jerbob92> While not actually hooked up to the printer
<libv> ah, cool
<jerbob92> And this works fine for normal printers
<jerbob92> This one is just very odd
<libv> so you probably will have to copy the printer gadget and fix bits to match your strange printer
<libv> +driver
<jerbob92> Yeah, that's what I'm working on
<jerbob92> Kinda stuck now on that IN endpoint
<jerbob92> And don't know enough about USB protocol to fix it, I gues
<libv> jerbob92: google for the mindshare book
<jerbob92> Thanks
<libv> i doubt that that is the full version
<libv> but this is one of the first things i bought when i was tasked with bringing up uac3
<jerbob92> What was I thinking when I thought I would be able to do this without diving into kernel module development and learning the USB spec :D
DonkeyHotei has quit [Read error: Connection reset by peer]
DonkeyHotei has joined #linux-sunxi
<libv> (which ended up getting stuck as i spent all my time making the unnamed vendor which used to sell graphics chips' unstable heap of .... firmware working)
<libv> jerbob92: it's a reasonably trivial task as far as REing and driver development goes tbh
<libv> jerbob92: just keep at it
<jerbob92> Not when you don't have any experience in kernel development tbh
<jerbob92> These gadget drivers are a mess
<jerbob92> But maybe all kernel code looks like a mess to me
<jerbob92> But maybe you know this one specific question I have: Is it possible for the USB host to send URB requests up the "wrong" endpoint
<jerbob92> I'll read the PDF too, maybe it's in there
<KotCzarny> doing things without learning specs.. sounds like recipe for disaster..
<KotCzarny> 'it works, kinda, ship it!'
<jerbob92> Well, I was hoping that I would be able to modify the existing gadget without having to fully understand USB and how it's implemented in the kernel
<KotCzarny> hire someone to do it?
<jerbob92> Maybe
netlynx has quit [Quit: Ex-Chat]
<jerbob92> I'm not against that, perhaps you know someone that would be able to do it and is also for hire?
lkcl has quit [Ping timeout: 245 seconds]
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #linux-sunxi
lkcl has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: KVIrc 5.0.0 Aria]
tllim has quit [Read error: Connection reset by peer]
lurchi__ is now known as lurchi_
voodoovood has joined #linux-sunxi
voodoovood2 has quit [Ping timeout: 255 seconds]
<mru> I don't think it's possible to fully understand usb
<anarsoul> hehe
Putti has quit [Ping timeout: 250 seconds]
lurchi_ is now known as lurchi__
cmeerw has quit [Quit: Leaving]
lurchi__ is now known as lurchi_
f0xx has quit [Ping timeout: 246 seconds]
<DonkeyHotei> mru: you think that's bad? look at this!
<mru> thunderbolt is more or less pcie
<mru> so that makes usb4 pretty much expresscard on a cable
NeuroScr has joined #linux-sunxi
<mru> or possibly something else entirely
<mru> I haven't read the spec
chewitt has quit [Quit: Zzz..]
chewitt has joined #linux-sunxi
msde has quit [Quit: Leaving]
MangyDog has joined #linux-sunxi
Mangy_Dog has quit [Ping timeout: 252 seconds]
jstefano_ has joined #linux-sunxi
jstefanop has quit [Ping timeout: 250 seconds]
reinforce has quit [Quit: Leaving.]
nashpa_ has quit [Ping timeout: 250 seconds]
nashpa has joined #linux-sunxi
chewitt has quit [Remote host closed the connection]
chewitt has joined #linux-sunxi
chomwitt has quit [Ping timeout: 246 seconds]
<karlp> well usb supespeed doesn't really have anything to do with usb highspeed7fullspeed either. not even the same pins. just a cable thhat looks similar
<karlp> jerbob92:there's nothing wrong with reading an IN endpoint though?
tllim has joined #linux-sunxi
nuuuciano has quit [Ping timeout: 255 seconds]
nuuuciano has joined #linux-sunxi
dev1990 has quit [Quit: Konversation terminated!]
vagrantc has quit [Ping timeout: 252 seconds]
vagrantc has joined #linux-sunxi
chewitt has quit [Quit: Zzz..]
gnufan_home has joined #linux-sunxi
chewitt has joined #linux-sunxi
<vagrantc> so, i've been using the pinebook daily, but I occasionally get cpu stalls: kernel:[20497.731441] watchdog: BUG: soft lockup - CPU#3 stuck for 21s! [swapper/3:0]
<anarsoul> hm
<anarsoul> do you have a backtrace?
<vagrantc> and then the machine is barely responsive or unresponsive for a while ... *sometimes* leaping into the future
<anarsoul> oh
<anarsoul> do you have a64 timer fix?
<vagrantc> i have the a64 timer workaround ... the one-liner in .dts ?
<vagrantc> or at least, i thought i did...
<anarsoul> that's not only dts
<anarsoul> you also need to patch the driver
chewitt has quit [Quit: Zzz..]
<vagrantc> anarsoul: ah, then i only have the workaround that smaeul pushed
<anarsoul> that's the workaround I'm talking about
<anarsoul> you need "arm64: arch_timer: Workaround for Allwinner A64 timer instability" and "arm64: dts: allwinner: a64: Enable A64 timer workaround"
<anarsoul> 2 patches
<anarsoul> I believe that's not enough
<vagrantc> if there's a correlary driver patch ...
<anarsoul> don't forget to enable SUN50I_ERRATUM_UNKNOWN1 in kernel config :)
<vagrantc> surely forgot that too
<vagrantc> ERRATUM_UNKNOWN1 ... so descriptive
<anarsoul> there's no erratum number
<anarsoul> since allwinner never admitted that there's a bug in cpu
<vagrantc> right...
<vagrantc> oh, that is apparently already in linux 5.0
<anarsoul> is it?
<vagrantc> so maybe i just need to enable the config
<anarsoul> only if it was backported
<vagrantc> $ git describe c950ca8c35eeb32224a63adc47e12f9e226da241
<vagrantc> v5.0-rc2-8-gc950ca8c35ee
<anarsoul> nah
<anarsoul> try "git describe --contains"
<anarsoul> ;)
lurchi_ has quit [Quit: Konversation terminated!]
<anarsoul> $ git describe --contains c950ca8c35eeb32224a63adc47e12f9e226da241
<anarsoul> v5.1-rc1~170^2^2~16
lurchi_ has joined #linux-sunxi
<vagrantc> ah, ok.
<vagrantc> wonder if i can manage to backport it to 4.19...
<anarsoul> it's pretty straightforward
<anarsoul> personally I've never seen this issue on my pinebooks even without patches
<anarsoul> I wonder if it affects only some A64?
<vagrantc> lucky you :)
<vagrantc> i've seen it on my early generation pine64+ as well as the pinebook
<vagrantc> thanks for pointing out my insufficient "too good to be true" patching :)
<anarsoul> np
<anarsoul> vagrantc: btw, are you publishing your debian image somewhere?
<vagrantc> anarsoul: :)
<anarsoul> and where I can find the image on I can find only arm64 ISO...
<vagrantc> no image, really ... just kernel+u-boot+arm-trusted-firmware ... there are debian-installer images, but they rely on the u-boot EFI/device-tree
<vagrantc> which, u-boot's EFI support, while coming along, isn't quite there yet
<vagrantc> but you also need a separate u-boot image
<vagrantc> kind of sub-optimal
<vagrantc> the u-boot image goes on microSD, and you can put the mini.iso on a USB stick or something
<anarsoul> I see
<vagrantc> it works, but obviously would be better with a single image on microSD
<anarsoul> I wish pinebook had spi flash for u-boot :(
<vagrantc> i do appreciate that it loads u-boot off of removeable media and prefers it by default
<anarsoul> yeah, that's nice feature
<vagrantc> when doing bootloader/arm-trusted-firmware testing ... it can sometimes go badly and nice to have an easy way to revert
MangyDog has quit [Ping timeout: 255 seconds]
jstefano_ has quit [Remote host closed the connection]
jstefanop has joined #linux-sunxi
jstefanop has quit [Ping timeout: 246 seconds]
Mangy_Dog has joined #linux-sunxi