<NiteHawk>
i'm still struggling a bit with 3.19.2 - but i was able to boot that customized 3.19.0 just fine
<ijc>
oliv3r: morning
<oliv3r>
i'm running 4.0-rc5 without problems :)
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
_massi has joined #linux-sunxi
<oliv3r>
mripard_: wow, going over v5 of the patchset now, this guy really invested some good time in it :) Happy to see :)
<oliv3r>
mripard_: thanks for pointing me into that direction, the ssd130x driver seen very little love in a logn time, except for the small return patch a few weeks ago
<oliv3r>
i'll go over it too, test it and submit the patch to add ssd1309 support
gianMOD has joined #linux-sunxi
gianMOD has quit [Remote host closed the connection]
iamfrankenstein has joined #linux-sunxi
<wigyori>
morning
FreezingCold has joined #linux-sunxi
prz has joined #linux-sunxi
bonbons has joined #linux-sunxi
gzamboni has quit [Ping timeout: 252 seconds]
gianMOD has joined #linux-sunxi
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-sunxi
gzamboni has joined #linux-sunxi
ssvb has quit [Ping timeout: 255 seconds]
awe00 has joined #linux-sunxi
orly_owl has quit [Ping timeout: 255 seconds]
hipboi has quit [Read error: Connection reset by peer]
hipboi has joined #linux-sunxi
iamfrankenstein has quit [Quit: iamfrankenstein]
<oliv3r>
ijc: hey ian, do you know if in u-boot, when I add one of those EXTRA_OPTIONS to a defconfig, can I still skip the probing via a parameter? Or is it all or nothing? I don't mind having a bigger binary that supports usb, ahci, but i don't want to probe for devices by default (takes to much time)
Akagi201 has joined #linux-sunxi
<ijc>
oliv3r: I'm afraid I don't understand the question. What devices are we talking about?
diego_r has quit [Ping timeout: 250 seconds]
Renard has joined #linux-sunxi
ricardocrudo has quit [Ping timeout: 256 seconds]
Akagi201_ has quit [Ping timeout: 250 seconds]
ricardocrudo has joined #linux-sunxi
premoboss has joined #linux-sunxi
<oliv3r>
ijc: usb and ahci; they take quite a while to probe for devices
<oliv3r>
ijc: so idaelly, i want to disable only the probing bit. If a user for example wants to boot from usb or what not, they can always call ahci scan; or usb scan (i suspect that's what it is, since mmc scan exists and can be used to scan for mmc stuff)
<oliv3r>
but by default, alyways boot from mmc/nand (wahtever it is compiled to do) making booting super fast
<ijc>
oliv3r: $boot_targets (or whatever it is called) is considered a user serviceable part, so you can set it to which ever set of devices you want to try in whatever order. I think we only actually probe if we come to try to actually boot from a device
<ijc>
$boot_targets was the right name
leviathancn has joined #linux-sunxi
lucaswang has joined #linux-sunxi
iamfrankenstein has joined #linux-sunxi
naobsd has joined #linux-sunxi
leviathancn has quit [Ping timeout: 272 seconds]
leviathancn2 has joined #linux-sunxi
lucaswang has quit [Ping timeout: 245 seconds]
orly_owl has joined #linux-sunxi
vpeter has quit [Ping timeout: 252 seconds]
hipboi has quit [Read error: Connection timed out]
hipboi has joined #linux-sunxi
<oliv3r>
ijc: ok thanks; i'll read into that and see if it stops the slow probing of devices
<oliv3r>
ijc: ah, okay if i understand correctly, I add boot_targets to my boot.scr and that should probe + boot in that order; well worth a test for sure
ssvb has joined #linux-sunxi
cnxsoft has quit [Ping timeout: 264 seconds]
ricardocrudo has quit [Ping timeout: 252 seconds]
<oliv3r>
ijc: doesn't seem to work; looks like i have to disable the device entirly (or hack u-boot ;)
<oliv3r>
ijc: basically, as soon as the CONFIG_CMD_USB is enabled to comile the commands (which you need) it automatically adds the scan to the def. environment
<oliv3r>
unless i can override the def. env. maybe
Andy-D has joined #linux-sunxi
Andy-D has quit [Ping timeout: 246 seconds]
diego_r has joined #linux-sunxi
<ijc>
oliv3r: I meant for you to set boot_targets in your env, I suspect boot.scr is too late, it'll have already probed each thing.
<ijc>
oliv3r: Does it probe usb even without trying to boot from usb? I think that isn't expected...
leviathancn2 has quit [Ping timeout: 252 seconds]
<oliv3r>
ijc it always probes ahci and usb; i don't know what the default boot order is, i'll find out where i best can set boot_targets before boot.scr (not awesome with the boot env yet)
<ijc>
oliv3r: Hrm, I thought Hans had tried to fix that, or maybe I'm think of something else?
<ijc>
Possibly it probes usb for potential keyboard if any sort of GFX support is on?
<oliv3r>
ijc: i haven't kept up to date with u-boot development, the only thing I know is how to generate he boot.scr and have modified it since :p
<ijc>
Worth having a look through "printenv" and seeing if there is a simple thing which could be nopped out
<oliv3r>
ijc: that could explain the usb probeing, but what about ahci probing?
<oliv3r>
ok sec
<oliv3r>
ah obviously i have no env, *** Warning - bad CRC, using default environment
<oliv3r>
which is no supprise as i have not set anything up :p
cnxsoft has quit [Remote host closed the connection]
<libv>
oh, right, readelf -s shows some high level functions.
cubeast has quit [Quit: Leaving]
<jemk>
and .text hasn't changed (except some jumps), so you can simply take the old symtab and substract the offset between old and new .text
<libv>
i just found the code to H264BuildDefList
<plaes>
:)
<plaes>
another nail in the coffin
<jemk>
and .text hasn't changed (except some jumps), so you can simply take the old symtab and substract the offset between old and new .text
<jemk>
...
<jemk>
i hate my irc client
<libv>
jemk: yeah, by the time you had written that, i had already found that code
<libv>
jemk: how did you notice that the .text section did not change?
<libv>
what tool or objdump/readelf arguments did you use?
<jemk>
libv: I guessed because of same size and bindiffed it
<jemk>
some words int the first part changed, these are jumps in disassembly
<jemk>
but the rest is identical
<libv>
with me, there is a difference between the offsets of the two ./_fini and the two _init/.init
<libv>
oh, .plt
Akagi201 has quit [Remote host closed the connection]
f15h has joined #linux-sunxi
<libv>
hrm
Akagi201 has joined #linux-sunxi
<libv>
how about me growing a brain and then running strip myself on the old binary]
reinforce has joined #linux-sunxi
<plaes>
:D
<ssvb>
libv: sigh, allwinner is acting as a silly kid, who is trying to get off the hook after getting caught doing something bad
<ssvb>
this looks really ridiculous
<ssvb>
I would be more in favor of applying a kind of "don't ask, don't tell" policy here, but allwinner is trying very hard to dig its own grave :(
<libv>
ssvb: it is completely ridiculous and it is becoming very very hard to claim that allwinner simply does not understand things
<iamfrankenstein>
Hello everyone, I have a question, I have a Q88 tablet that doesn't want to be reflashed trough USB (current image is broken). Is it possible to use JTAG to pull and push an image? I have a Q88 that does work so is it possible to dump the internal memory? Can i use OpenOCD for this? (i read something about address limit beyound 2gb) ( i don't expect step by step instructions, just want to know if its in theory possible).
<libv>
ssvb: as i said earlier, i gave them some public crap in august to get their attention, and then tried to work with them on the bigger issue(s) in the back
<libv>
iamfrankenstein: new device howto
<iamfrankenstein>
ow Q88 is a Allwinner A13 tablet
premoboss has quit [Ping timeout: 256 seconds]
<libv>
ssvb: their kernel with binaries stripped was the big clue for me that they did not want to be helpful
<libv>
allwinner has been piling things up since then.
<libv>
anyone who still wishes to defend allwinner today is intellectually dishonest, or, a liar.
kaspter has quit [Ping timeout: 246 seconds]
<plaes>
libv: or paid by allwinner ;)
<libv>
plaes: both are not mutually exclusive :p
<libv>
i would change the or to and
<ssvb>
libv: there are certain things, which can't be easily undone and I would be kind of satisfied to see changes for the better at least with the new stuff from now on
<libv>
ssvb: definitely
<libv>
ssvb: but allwinner had been given a route for improvement by many people many times over
<libv>
instead, they just dig themselves in deeper
<ssvb>
admittedly, they are doing all this crazy stuff after being cornered
<ssvb>
and learning a wrong lesson
<libv>
or are just utterly unwilling to learn
prz has joined #linux-sunxi
cubeast has joined #linux-sunxi
<plaes>
well, acquaintance of mine said "copyright in china means right to copy.."
gianMOD has quit []
gianMOD has joined #linux-sunxi
Renard has quit [Ping timeout: 246 seconds]
<mripard_>
oliv3r: not willing to understand how something works is usually not a very good way to fix an issue.
Renard has joined #linux-sunxi
<oliv3r>
mripard_: wow! not so negagive :( i am very willing amd i will for sure, just not enough hours in a day more pressi g matters first :(
konradoo77 has joined #linux-sunxi
Akagi201_ has joined #linux-sunxi
dev1990 has joined #linux-sunxi
Akagi201 has quit [Ping timeout: 255 seconds]
Netlynx has joined #linux-sunxi
Netlynx has joined #linux-sunxi
imcsk8 is now known as imcsk8|afk
<mripard_>
oliv3r: there's absolutely no need for a ioctl to call update_display
<mripard_>
mmap or not
<mripard_>
the page mapped is setup to generate page faults whenever someone accesses it
<mripard_>
and this is how the kernel knows that you wrote something to your framebuffer
<mripard_>
but I already told you that yesterday.
prz has quit [Ping timeout: 264 seconds]
gianMOD_ has joined #linux-sunxi
ricardocrudo has quit [Ping timeout: 250 seconds]
gianMOD has quit [Ping timeout: 244 seconds]
ricardocrudo has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 264 seconds]
<wens>
[6~[6~[6~[6~[6~[6~/sb c
konradoo77 has joined #linux-sunxi
<wens>
oops
domidumont has quit [Read error: Connection reset by peer]
khuey|away is now known as khuey
f15h has quit [Remote host closed the connection]
domidumont has joined #linux-sunxi
gianMOD_ has quit [Remote host closed the connection]
awe00 has quit [Ping timeout: 265 seconds]
leviathancn has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 255 seconds]
dev1990 has quit [Quit: Konversation terminated!]
ricardocrudo has quit [Remote host closed the connection]
afaerber has quit [Quit: Verlassend]
konradoo77 has joined #linux-sunxi
cooper__ has joined #linux-sunxi
<cooper__>
Just a quick quesJust a quick question about http://linux-sunxi.org/DDR_Calibration - do I really need *that* u-boot, or will mainline u-boot work just as well?
<ssvb>
Cooper__: you can use any u-boot, but the mainline u-boot is preferable
<cooper__>
Excellent.
diego71 has quit [Read error: Connection reset by peer]
* ssvb
still needs to update the wiki
_massi has quit [Remote host closed the connection]
afaerber has joined #linux-sunxi
<cooper__>
In board/sunxi I have both dram_sun4i_auto.c and dram_sun5i_auto.c - should I edit both the same way simultaneously, or only one of them?
diego71 has joined #linux-sunxi
<cooper__>
There's also a value printed for 'dcdc3_vol' which I can't find anywhere.
<ssvb>
you don't need to edit the dram_sun4i_auto.c and dram_sun5i_auto.c files, these settings should go to defconfig files
<ssvb>
only the dcdc3 voltage should be patched in the code
<ssvb>
Cooper__: as usual, ping me if you have additional questions
<cooper__>
Will do. Thanks.
<mripard_>
ssvb: we were discussing about kernelci.org yesterday (the automated boot farm) with wens and tyler-baker about running your cpufreq test tool, would that make sense to you?
<ssvb>
mripard_: the voltage tolerance seems to be device specific and I'm not sure if it makes sense to run it more than once on that farm
<ssvb>
other than catching the cases of setting obviously incorrect voltages via some regressions in the kernel
montjoie[home] has quit [Quit: leaving]
<cooper__>
ssvb: About that zq value, how do you go from a zq of "0x3B" to a u-boot defconfig value of "17688576"?
<ssvb>
mripard_: I also still need to rewrite this test program in C (for easier deployment and use), because I have already isolated the critical libjpeg-turbo part, which tends to fail
<cooper__>
ssvb: And more to the point, when atsampson's table says "dram_zq = 0x29 (0x1991800)", what value should I place in the setting?
<ssvb>
Cooper__: that's a good question, I'll write some notes about it
<tyler-baker>
ssvb, if you rewrite it in C that would be excellent :)
montjoie[home] has joined #linux-sunxi
<mripard_>
ssvb: it's all about catching regressions ;)
<mripard_>
and if you actually have some ideas of meaningful tests to run on that farm, I'm all ears
skoperst has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 246 seconds]
<ssvb>
mripard_: ok, I'll look into it
JohnDoe_71Rus has joined #linux-sunxi
gjm has joined #linux-sunxi
leviathanch has quit [Quit: No Ping reply in 180 seconds.]
leviathanch has joined #linux-sunxi
imcsk8|afk is now known as imcsk8
<cooper__>
ssvb: While you write notes about that zq setting, could you, for now, provide me with the value to get that 0x29 setting?
skoperst has quit [Ping timeout: 255 seconds]
<ssvb>
Cooper__: you can use either 0x29 directly or 0x1991800
<Skylark>
gak. A20 clock outputs are open drain - why.
<Skylark>
(is there a software way around that?)
gianMOD has joined #linux-sunxi
<wens>
they are?
Lannig has joined #linux-sunxi
<Lannig>
Hi folks
<Lannig>
I'm coming here from the wiki which is a tremendous source of information on A23 (which is what I'm into right now)
paulk-collins has joined #linux-sunxi
<Lannig>
I have a question that might be borderline to the topic since it relates to Android firmwares, but somehow I suspect that this subject might be relevant to Linux implementations as well
<Lannig>
May I fire?
<Skylark>
wens: yup! i learned the very hard way. now i have to graft some 0402 2k resistors on the pcb
<Skylark>
wens: the cubietruck, which is the only other user of this that i know of, has a pull on that pin - i didn't notice it, since it was on a different page
diego_r has quit [Ping timeout: 244 seconds]
<oliv3r>
mripard_: you did tell me that, after i tried it and experimented. reason i went ioctl route was because the mmapping was broken and not working duh :p
<wens>
Skylark: hmm, datasheet doesn't say
<Skylark>
you're totally right on that one :-(
<Skylark>
most CPUs i've implemented before had push-pull clock outputs
<oliv3r>
mripard_: also quite onviously i cant change the dispkay to a grayscale o e or anything. im stuck with the hardware that we have :(
<wens>
Skylark: i don't know, i would've thought all the PIO pins are tri-phase push-pull
<mripard_>
oliv3r: so, if mmap doesn't work on a driver, that must be because the memory management code is broken?
<mripard_>
Lannig: shoot
<oliv3r>
mripard_: not at all, but i had no idea what was going on and it was a 4 line test :p
<Lannig>
OK I'll shoot: until recently all sunxi firmwares had the h/w config stored into a script.bin/script0.bin in the bootloader partition
<Lannig>
This file is gone from recent KK firmwares. Doing string search on Phoenixsuit images show that it's *still* there, but where?
<oliv3r>
it was a quick brainfart is all inspired by the update dis-lay in the write function is all. no evil intentions ;)
<Skylark>
wens: when you switch it to the special function mode, it seems to only drive 0, never 1 - if i enable the internal pullup it does a bit for me, but not high enough so it needs some external pull too
<oliv3r>
btw why did you refer to directfb? i didnt mention the framework i am using (efl/evas)
<oliv3r>
anyhow thomas's stuff works great so thats cool
Patukas has joined #linux-sunxi
<Patukas>
hi
<mripard_>
oliv3r: because that's what was talking about in the mail you replied to
<oliv3r>
and i'll just patch evas's fb handling code, its not that much to do
<oliv3r>
mripard_: ah right, i was thinking all fb users actually
<oliv3r>
but its cool, not a huge issue for me
<oliv3r>
anyway, bed time! bye bye
<mripard_>
Lannig: no idea, sorry :/
dev1990 has joined #linux-sunxi
<Lannig>
Ok thanks for reading me anyway. I'll keep searching... :-/
<mripard_>
maybe wens will know
<Patukas>
hi, i have a cubietruck with i2s playback works, but i need the capture too. Anyone knows a modified drivers with i2s capture enable?
<Lannig>
I've spend an insane amount of time trying to reverse-engineer a boot SD created by Phoenixcard but I can't relate this to the NAND partition structure yet