<Nazcafan>
nand is some kind of low level memory, whereas flash devices (sdcards) are higher level, and thus are much easier to handle, is this correct?
<hno>
NAND is low-level flash chips. SD-Cards are NAND chips + RAM + CPU + firmware presenting the NAND as a normal block device.
<Nazcafan>
which means much less code in order to support it correctly?
<hno>
USB sticks is pretty much the same thing in different packaging.
<Nazcafan>
and it is troublesome to have proper nand support for sunxi, or is it?
<hno>
Supporting SD-Cards is many orders of magnitute simpler than NAND, and more importanly works exactly the same no matter what brand or size of sd card you have.
<hno>
With bare NAND chips a lot differ depending on what exact chip you have.
<Nazcafan>
what is libnand, and how does it relate or not to mtd? is libnand some sort of userspace thing?
<Nazcafan>
(warning: absolute newbie here)
<hno>
libnand is allwinners NAND driver. Found in their bootloader, u-boot, livesuit (the device side parts of it) and linux-kernel.
<hno>
most versions are distributed as binary blobs only.
<hno>
and 99.99% of all livesuit images you find are GPL violating from it.
<Nazcafan>
and I guess the blob is different for each device/soc?
<hno>
no
<Nazcafan>
ah
<hno>
but exists in several incompatible versions.
<Nazcafan>
incompatible?
<hno>
but yes it is partly SoC dependent.
<Nazcafan>
do you mean non-interchangeable with respect to the kernel they are shipped with?
<hno>
incompatible == can't access a flash chip formatted with another version and will most times silently reformat the flash erasing any content.
<Nazcafan>
aaah
<Nazcafan>
that's mean!
<hno>
we do have source of two versions, which both works across the SoC versions.
<Nazcafan>
so how does the sunxi kernel addresses the nand support problematic?
<Nazcafan>
(ah, this answers partly my last question, I guess)
<randomblame>
had to run for a bit - did anyone find the 4.4 sdk for a23?
<randomblame>
or is it not in the hands of the community yet
<Nazcafan>
do you use the source for the sunxi-kernel?
<hno>
I ignore it, and try to document the hardware so another driver can be written.
<hno>
the hardware manual have a very extensive documentation of the nand controller which basically says "there is a NAND controller".
<Nazcafan>
I have my last newbie question before I go to bed (don't want to be a zombie at work tomorrow)
<Nazcafan>
on PC platform, I can install several kernels on the same machine and select the one I want at boot time through grub
<hno>
yes, same here using u-boot, just that you don't have a fancy menu.
<Nazcafan>
I can live without the fancy menu
<Nazcafan>
why did the allwiner guys decide to go for a kernel module for the live suite? ain't usblib easier to use?
<hno>
don't know. The module can be replaced by a LD_PRELOAD wrapper using libusb.
<hno>
the module doesn't do anything useful
<hno>
Nazcafan, we do have full FEL boot mechanism implemented to boot whatever you like over USB-FEL. See http://linux-sunxi.org/FEL/USBBoot
<Nazcafan>
did you document the FEL protocol anywhere?
<Turl>
Nazcafan: there's always the open source fel tool
<Turl>
I don't know if there's any doc, maybe on the wiki
<geecko>
rz2k, is there anything to change kernel-side to support dma-buf ?
<rz2k>
r3p0 will not work with r3p2 because blobs check the version
<rz2k>
no
<geecko>
rz2k, it's r3p2-01rel0 blobs on r3p2-01rel3 kernel driver
<rz2k>
oh, then check api version on mali_utgard.h
<geecko>
rz2k, i did that. you think it would work just by changing api version?
<rz2k>
blobs ioctl api version definition from there
<geecko>
yeah i know
<rz2k>
should work
<geecko>
nice.
<geecko>
rz2k, i backported ION driver from the 3.4 kernel, switching to dma-buf on mali, upgrading the framebuffer driver, and switching hwcomposer.
<geecko>
lots of steps
<geecko>
lots of chances to fail also :/
<geecko>
it's for android of course.
<rz2k>
hmm
<rz2k>
I'm not sure about android, sorry
<geecko>
rz2k, np ^^
<rz2k>
never got deep into it
<geecko>
rz2k, i booted the kernel with api version changed and UMP disabled, i have no errors on logcat or dmesg, but there's no display, and that was totally expected
<rz2k>
go strace everything, I'm not sure how hwcomposer/surfaceflinger works with mali, so I cant help with that
<rz2k>
and the gralloc stuff
<geecko>
rz2k, it's pretty cool if the blobs run without modification on the kernel though :)
<rz2k>
ls
<rz2k>
whoops
leviathanch2 has joined #linux-sunxi
<geecko>
rm -rf /
<geecko>
wooooops !!
<rz2k>
lol
<binaryferret>
hehe
<geecko>
dd if=/dev/urandom of=/dev/sda
<geecko>
:P
<geecko>
don't try that :P
<Nyuutwo>
dd if=/dev/zero of=/dev/root - once I done that
<Nyuutwo>
...
<binaryferret>
hehehe
<rz2k>
geecko: with native linux I've been able to run different blobs version
<geecko>
rz2k, different rels ?
<rz2k>
r3p0 kernel and r3p2 blobs iirc
<geecko>
haha, wtf !
<rz2k>
but that was long ago
<geecko>
ok so it should definitely work
<geecko>
nice :D
<geecko>
rz2k, oh btw, with dma-buf blobs, UMP will not work at all, right ?
FreezingCold has joined #linux-sunxi
<rz2k>
not sure, mali has atleast two memory zones available, trace when dma-buf gets invoked
<libv>
codekipper: i think we have board pictures :)
<libv>
codekipper: the pictures we'd need would be like the hc860 page
<Turl>
the more the merrier, I'm not gonna say no to sunxi pictures, even if we have some already :)
<codekipper>
libv: so top and all sides with some form of connection?, you're in luck as I was moments away from butching the sides to add a serial jack connection.
<libv>
:)
<libv>
codekipper: yeah, and a side view for the infobox
<codekipper>
libv: Have you started with the Q8H yet?, I've got one on the way and I've been looking at the code. Just wondering if you've located the serial pins yet?
<libv>
i am pretty certain that those two obvious pads are serial
<libv>
just haven't changed the script.bin to test them yet
<libv>
they really are way too obvious, and in such a case, they have never been wrong
<codekipper>
libv: yep, they do stand out as contenders. I should also have a MK808C (with what also looks like obvious serial points) in the next few weeks.
<libv>
codekipper: get a proper irc client going if you want to do more sunxi stuff :)
<libv>
webchat is not the best or most reliable
HeHoPMaJIeH has quit [Remote host closed the connection]
rz2k has quit [Remote host closed the connection]
leviathanch2 has quit [Ping timeout: 260 seconds]
kz1 has quit [Read error: Connection reset by peer]
kz1 has joined #linux-sunxi
leviathanch2 has joined #linux-sunxi
<codekipper>
libv: will do once I've got around work's pr0xy
pwhalen has quit [Ping timeout: 260 seconds]
<libv>
:)
<libv>
hrm, what did i use to tunnel out of nokia network...
<Turl>
one of those proxy over DNS things? :)
<plaes>
libv: corkscrew
<libv>
exactly :)
* libv
was digging through ancient irclogs
<libv>
codekipper: codescrew is the ticket
<plaes>
I was in there too :P
<libv>
:)
<libv>
corkscrew even
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
pwhalen has joined #linux-sunxi
<codekipper>
libv: I'm using squid at the mo for other things...I will have a little play with that and corkscrew, if all fails then fat fingers will be reaching for an android client
<Tartarus>
hey hno, any plans to submit u-boot stuff upstream again soon?
<Turl>
Tartarus: there was a discussion about that on the ML these last days, be sure to read it
eagles0513875_ has quit [Ping timeout: 272 seconds]
egbert has quit [Read error: Operation timed out]
kivutar has quit [Quit: Ex-Chat]
kivutar has joined #linux-sunxi
eagles0513875_ has joined #linux-sunxi
randomblame has joined #linux-sunxi
pwhalen has quit [Ping timeout: 260 seconds]
bertrik has quit [Read error: Connection reset by peer]
paulk-collins has quit [Remote host closed the connection]
egbert has joined #linux-sunxi
* randomblame
is now known as poppycat
AreaScout has quit []
Black_Horseman has joined #linux-sunxi
Black_Horseman has quit [Changing host]
Black_Horseman has joined #linux-sunxi
kivutar_ has joined #linux-sunxi
kivutar has quit [Ping timeout: 265 seconds]
pwhalen has joined #linux-sunxi
pwhalen has joined #linux-sunxi
pwhalen has quit [Changing host]
<hno>
arete74, yes the gpio command can both set and read any gpio capable pin on the processor.
t3st3r has quit [Ping timeout: 240 seconds]
t3st3r has joined #linux-sunxi
Nazcafan has joined #linux-sunxi
<Nazcafan>
hi everyone
<hno>
oliv3r, regarding secure boot. It should work fully, and if you encrypt it all then there is no big issue with allowing other boot methods in terms of information security
<Turl>
hno: it's not secure if you can run arbitrary code
<hno>
also from A20 onwards there is some efuses to lock things down quite a bit, including secure bootmode where brom requires encrypted bootloader.
<hno>
possibly also exists for earlier generations, documentation quite incomplete in this area.
<hno>
A20 SID manual also mentions HDCP key btw, but a little unclear on where that is stored.
hawi has joined #linux-sunxi
Black_Horseman has quit [Read error: Connection reset by peer]
ganbold_ has quit [Remote host closed the connection]
Black_Horseman has joined #linux-sunxi
bsdfox has quit [Ping timeout: 264 seconds]
nove has quit [Quit: nove]
bfree has quit [Ping timeout: 248 seconds]
bsdfox has joined #linux-sunxi
randomblame_ has joined #linux-sunxi
woprr has joined #linux-sunxi
randomblame has quit [*.net *.split]
gzamboni has quit [*.net *.split]
Seppoz has quit [*.net *.split]
binaryferret has quit [*.net *.split]
Seppoz has joined #linux-sunxi
binaryferret has joined #linux-sunxi
gzamboni has joined #linux-sunxi
honx_ has quit [Read error: Operation timed out]
ykchavan has quit [Quit: Leaving]
honx has joined #linux-sunxi
bfree has joined #linux-sunxi
Xaros has joined #linux-sunxi
Xaros has quit [Remote host closed the connection]
Black_Horseman has quit [Ping timeout: 260 seconds]