pmp-p has quit [Remote host closed the connection]
pmp-p has joined #linux-sunxi
Nakaori has joined #linux-sunxi
Naka has quit [Ping timeout: 256 seconds]
cnxsoft has joined #linux-sunxi
ldevulder has joined #linux-sunxi
<mru>
gediz0x539: my understanding is that if you've obtained a "confidential" document without signing an nda or otherwise agreeing to such terms, you're breaking any law by distributing it
<mru>
unless it's a matter of national security
<libv>
mru: without those allwinner docs, there would be no #linux-sunxi, no linux-sunxi.org and no upstream code
<libv>
so while you technically are correct, the very fact that you are here is because of those things being published
<mru>
sorry, I mistyped... should say "you're NOT breaking any law"
<libv>
ah, ok
<libv>
it's a grey zone, we know it is a grey zone
<mru>
if you did sign an nda, it would be breach of contract
<libv>
but allwinners remaining business depends on us
<libv>
and if it was two board makers who actually adhered to the gpl which initially made it all happen, as no-one else in the android device world was really adhering to the gpl
<libv>
and then there's the u-boot and kernel blobs that allwinner forced upon its customers, with even nintendo shipping them
<mru>
depending on the document, it could be argued that distribution is unethical even if legal
<libv>
and allwinner was caught with ffmpeg code in their media stuff
<mru>
for hardware datasheets, that would be a rather difficult argument to make
<libv>
as said, it's grey, we know it's grey, but without it none of this here would exist
<mru>
it's a very, very light grey
<mru>
how do you think newspapers can publish leaked goverment papers without consequence?
<libv>
ok
netlynx has joined #linux-sunxi
netlynx has joined #linux-sunxi
<libv>
well, i find it very important to have anything document, as what i saw when i got into the mali, is that it was all forum posts and nothing solid, useable or long lasting
<mru>
if a "secret" stamp was sufficient, there'd be no need for multi-page NDAs
<libv>
and why i dread the board makers wanting to copy our wiki to do their marketing directly
<mru>
huh?
<libv>
as, as we have seen with C.H.I.P and nextthingco, vendors vanish
<mru>
those clowns
<libv>
and before this turned into an almost pure development board community, we only had chinese settopbox manufacturers and tablet makers
<libv>
nextthingco timed things badly, a .us company, doing a crowdfunding campaign... just when we had ramped up the noise surrounding the ffmpeg code
<libv>
they were a full week into their kickstarter before then acknowledged that, and they only then went the open source route
<libv>
the mele A1000 tv set top box with and allwinner a10, had the uart connector already soldered on
<libv>
tom cubie who then worked for allwinner, started selling it over ebay to get a useful allwinner development machine in many peoples hands
<libv>
this started linux-sunxi, especially when mnemoc worked with tom cubie to get his cubieboard produced through crowdsupply
<libv>
even mele seems dead
<libv>
nothing new there since 2016
<libv>
without the cubieboard, we most likely would not have seen the likes of orange pi, banana pi and everything even less popular and more shortlived
lurchi_ is now known as lurchi__
juri__ has joined #linux-sunxi
lurchi__ is now known as lurchi_
juri_ has quit [Ping timeout: 264 seconds]
juri_ has joined #linux-sunxi
apritzel has quit [Ping timeout: 240 seconds]
juri__ has quit [Ping timeout: 246 seconds]
\\Mr_C\\ has quit [Quit: (Read error: Connection reset by beer)]
gediz0x539 has quit [Ping timeout: 246 seconds]
gaston1980 has joined #linux-sunxi
Mangy_Dog has quit [Ping timeout: 272 seconds]
<rellla>
yeah, good old cubieboard :) my first device was a mele, which i got directly from Tom :p
<rellla>
omg, time flies ...
_whitelogger has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
lurchi_ is now known as lurchi__
apritzel has joined #linux-sunxi
apritzel has quit [Ping timeout: 265 seconds]
<libv>
rellla: 'twas the summer of 2012
<rellla>
oh
mripard has quit [Ping timeout: 246 seconds]
mripard has joined #linux-sunxi
Esmil has quit [Ping timeout: 256 seconds]
Esmil has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
<mru>
can someone enlighten me as to how to make mesa and lima play nice?
<mru>
"MESA-LOADER: failed to open sun4i-drm: /usr/lib/dri/sun4i-drm_dri.so"
<mru>
is it looking for the wrong thing, or did something not get built?
<tuxd3v>
ls -l /usr/lib/dri/sun4i-drm_dri.so
<mru>
the file doesn't exist
<mru>
there's a lima_dri.so though
<tuxd3v>
does you compiled it?
<mru>
I compiled mesa
<tuxd3v>
find / -name sun4i-drm_dri.so
<mru>
doesn't exist
<tuxd3v>
you installed it were?
<mru>
look, is there a magic configure flag I'm missing?
<mru>
or is something else wrong?
<mru>
there is no sun4i-drm_dri.so anywhere in the entire town
<tuxd3v>
I don't know your setup..
<mru>
you don't need to
<mru>
I'm asking what it should be, not what it is
sunshavi has quit [Remote host closed the connection]
sunshavi has joined #linux-sunxi
<montjoie>
cubieboard2 bring me here (and temperature)
gaston1980 has quit [Quit: Konversation terminated!]
<tuxd3v>
I have noticed that dwmac-sun8i changed
<tuxd3v>
I had some patchs for it
<tuxd3v>
to implement: regulator_phy,regulator_phy_io
<tuxd3v>
I will try to adapt my patchs, but I don't really know if they are needed anymore..
lurchi__ is now known as lurchi_
<jernej>
tuxd3v: those patches are for OPi3 ethernet and yes, you still need them
hexdump0815 has joined #linux-sunxi
<hexdump0815>
jernej: apritzel: not sure if it is useful for anything, but i added some ugly hacks to the h616 legacy u-boot so that it works more normal, i.e. loads a dtb defined in and uses an APPEND string defined in an extlinux.conf (before it only used in-u-boot hardcoded stuff)
<hexdump0815>
jernej: apritzel: with this its now possible to boot a mainline kernel from the legacy boot - not sure, but maybe this is useful for comparing registers or similar?
<jernej>
hexdump0815: U-Boot is more or less solved, what it might be interesting are register dumps from BSP kernel
<jernej>
but that can be already done from Android or OPi images
<hexdump0815>
jernej: ok - no problem if this is not needed - just wanted to mention it at least ... i tried to boot mainline kernel with this legacy u-boot to see if maybe usb works somehow this way due to maybe some hidden initialization in legacy u-boot, but its the same like booted with mainline u-boot
<hexdump0815>
jernej: all done on a x96q h313 tv box which is working well with usb and hdmi with legacy kernel and u-boot
<apritzel>
hexdump0815: did you try to enable the other USB ports, maybe it's not USB 0 and 1?
<tuxd3v>
jernej, many thanks
<hexdump0815>
apritzel: i tried it and it did not work, but maybe i did it the wrong way - what would be the proper way in config and dts to test that?
<apritzel>
hexdump0815: in the DT I mean: &ohci2 { status = "okay";};
<apritzel>
this is where U-Boot looks for USB ports, there is no separate config other than enabling USB in general
lurchi_ is now known as lurchi__
<apritzel>
shouldn't hurt to enable all of them, I did this on the OPi Zero2 and connected some USB bracket to the header pins (which carries USB2 and USB3), that worked nicely
lurchi__ is now known as lurchi_
<hexdump0815>
apritzel: success :) - with 2&3 enabled i now see usb devices on both usb ports in the linux kernel
<apritzel>
hexdump0815: nice! Are there just 2 USB ports on that box?
<hexdump0815>
apritzel: yes - it only has two
<apritzel>
so port 0 (OTG, usable for FEL) is not connected then?
<hexdump0815>
not really sure - when i played around with legacy i thought that one of them was the otg port as initially only one port worked until i switched the otg port to fixed host mode and then the other port worked as well
<hexdump0815>
both ports seem to work perfectly fine on mainline now - i had two usb sticks connected to them and can use the filesystems on both
<apritzel>
hexdump0815: can you see the MMIO addresses in the kernel log, where the kernel found your sticks?
<smaeul>
so the relevant buses are 1 (5101000.usb) and 4 (5311000.usb)
<hexdump0815>
but 1 was enable before and did not work before i enabled 4 ...
<hexdump0815>
5101000.usb would be 0 and 5311000.usb would be 3 in the dtb - right?
<smaeul>
bus numbers 1 and 4 here are just the order that Linux probed them, you have to match them to the DTB based on the MMIO address
<smaeul>
yes, that's ehci0 and ehci3
gaston1980 has joined #linux-sunxi
<asdf28>
the makers of kernel 3.4 are such gods
<hexdump0815>
smaeul: apritzel: if i only leave 0&3 enabled i again get usb errors - does that mean that maybe there is something mixed up in the dtb, so that some power (or other) realated stuff it wired to 1&2 although they are not used?
<apritzel>
I am starting to become afraid that the ports in the PHY are not really independent :-(
<apritzel>
because now that I think about it my USB also started to work when I tested the other ports
<jernej>
is this specific to H616 or also on other SoCs?
<apritzel>
jernej: I can't remember this to be a problem before, so it would be something new to the H616/313
hexdump0815 has quit [Ping timeout: 248 seconds]
<apritzel>
well, the H6 had this weird mapping, where the two USB 2.0 ports are 0 and 3