narmstrong changed the topic of #linux-amlogic to: Amlogic mainline kernel development discussion - our wiki http://linux-meson.com/ - ml linux-amlogic@lists.infradead.org - Publicly Logged on https://irclog.whitequark.org/linux-amlogic
<brads> Xogium: been lacking time of late by im not far from testing multiple types of emmc with C2 on 5.0.2. I had no issues with 5.0-rc6 but others have reported issues (particulary with arch linux)
<Xogium> brads: ah, well good to know I guess since I'm on arch
<Xogium> my eMMC is the black one
<Xogium> so I think samsung
<brads> ok I have both orange and back emmc's to try. I will make an effort to try this later today
<Xogium> one of those that had problem and on which I had to reduce the frequency to 100000000 instead of 200000000
<brads> Did that resolve the problem?
<Xogium> yeah, back then it stopped corrupting the card
<Xogium> I always left it at that speed since
<brads> OK no worries, I will do some full tests after work tonight
<Xogium> not to mention that its also in upstream now iirc
vagrantc has quit [Ping timeout: 258 seconds]
vagrantc has joined #linux-amlogic
chewitt has joined #linux-amlogic
ldevulder_ has joined #linux-amlogic
ldevulder has quit [Ping timeout: 246 seconds]
sputnik__ has joined #linux-amlogic
sputnik_ has quit [Ping timeout: 245 seconds]
sputnik__ has quit [Ping timeout: 255 seconds]
vagrantc has quit [Ping timeout: 258 seconds]
vagrantc has joined #linux-amlogic
sputnik_ has joined #linux-amlogic
vagrantc has quit [Quit: leaving]
random_yanek has quit [Quit: random_yanek]
random_yanek has joined #linux-amlogic
random_yanek has quit [Max SendQ exceeded]
random_yanek has joined #linux-amlogic
jakogut has quit [Ping timeout: 246 seconds]
jakogut has joined #linux-amlogic
ldevulder_ is now known as ldevulder
<jbrunet> MMC needs some work. I'm currently discussing the topic with amlogic but there is a lot of tests to do and takes time. Please be patient
return0e_ has joined #linux-amlogic
<Xogium> jbrunet: sure thing :) I figured I'd ask before doing anything
return0e has quit [Ping timeout: 246 seconds]
nsaenz has joined #linux-amlogic
JulienMasson has joined #linux-amlogic
chewitt has quit [Quit: Zzz..]
JulienMasson has quit [Remote host closed the connection]
JulienMasson has joined #linux-amlogic
sputnik_ has quit [Ping timeout: 250 seconds]
afaerber has quit [Quit: Leaving]
afaerber has joined #linux-amlogic
nsaenz has quit [Quit: Leaving]
jakogut has quit [Remote host closed the connection]
<ldevulder> narmstrong, hi! I have some time to do some tests with vim2, I activated the SPI-NOR and "burn" u-boot.bin on it and I will want to boot from SPI. I know that the khadas u-boot version uses "kbi" command to do this. It's simply a 'i2c' command wrapper, so I activated MESON_I2C and tried to do the same command: 'i2c mw 0x18 x20 0 1' but this doesn't work
<ldevulder> have you been able to boot from SPI on this SoC?
<narmstrong> ldevulder: no, I haven't got time for thi...
<narmstrong> ldevulder: i'm glad you are trying ! Yep the kbi is just a wrapper
<narmstrong> but it could be porter to u-boot upstream
<narmstrong> *ported
<ldevulder> narmstrong, yes, easily I think, and I saw that khadas team does a version for rk3390 on a recent u-boot
<narmstrong> ldevulder: would be great if you can enable it, so we can do true EBBR/EFI on the vim2 aswell
<ldevulder> but as the i2c command fails I 'm not quite confident :) maybe some code in meson-i2c will need some change, I will try to investigate (but that could be like the USB issue on s912, still no idea...)
<ldevulder> narmstrong, I will try :)
<ldevulder> next challenge :-p
nsaenz has joined #linux-amlogic
chewitt has joined #linux-amlogic
afaerber has quit [Quit: Leaving]
<ndufresne> chewitt, btw, build worked, but failed at very last step, trying to mkdir /flash and mounting it, not sure if this was suppose to be run as root or something
<chewitt> hmm odd
<chewitt> what distro are you building on?
<chewitt> we never build as root
<chewitt> we build all our official images on Ubuntu LTS releases, either 16.04 or 18.04
<chewitt> other things should also work, but Ubuntu appears to be the n00b builders go-to distro so we use the same to minimise differences
<ndufresne> Fedora 29
<ndufresne> basically it ends on a permission denied
<ndufresne> chewitt, do you have a variable to make the makefile verbose ?
<ndufresne> I tried V=1, that wasn't it
<chewitt> this is the bit at the end where it creates the squashfs files and then the .img ?
<chewitt> "image: creating file LibreELEC-Khadas_VIM.arm-9.1.0.img..." etc.
<chewitt> can you paste the output .. pastebin or spam me on PM
<ndufresne> give me a minute, I was doing this at home
<chewitt> no rush
<ndufresne> chewitt, that's how it fails, https://paste.fedoraproject.org/paste/mw3WgJSBOhbqz7cl-Tl1fA
<chewitt> cp projects/Amlogic/devices/Khadas_VIM/bootloader/release projects/Amlogic/devices/Nexbox_A1/bootloader/release
<chewitt> then re-run the build command
<chewitt> I think I built the Nexbox image once .. I don't have the hardware
<chewitt> so I didn't test/check the config for a while
<chewitt> my bad .. as usual :)
<ndufresne> running ...
vagrantc has joined #linux-amlogic
<chewitt> all good?
<ndufresne> yep, just need to try it now ;-
<ndufresne> somewhere this week I suppose
<ndufresne> chewitt, were is the image exactly ?
<chewitt> target/
<chewitt> should have a .img or .img.gz file there
<chewitt> and maybe a .tar
<chewitt> .img = install
<ndufresne> good, got a cleanup that dir ;-P
<chewitt> .tar = update
<chewitt> technically you can update using .tar or .img or .img.gz files, but .tar is fastest
<chewitt> scp to /storage/.update on the box and reboot to update
<ndufresne> I voluntarily bricked the device so that it would pick the u-boot from my SDCard ;-P
<chewitt> not normally required
<chewitt> there's usually a reset button somewhere
<ndufresne> it's unrelated to LibreElec of course, the board came with q201 default configuration u-boot
<chewitt> hold that in, power on, wait for 3-5 secs and release the reset button
<ndufresne> which didn't have ethernet working
<ndufresne> ok, just like the android image
<ndufresne> anyway, for later, when I'm next to it
<ndufresne> chewitt, my goal is to check if a a properly patched kernel boots, and if by chance it works, also to give LibreELEC a try of course
<ndufresne> (well, on the last one, I could just test on PC, or on RPi)
<chewitt> mainline Amlogic still has some rough edges
<chewitt> but as long as you stick to 1080p media it's pretty decent
<ndufresne> noted
<chewitt> I've been using is exclusively at home since last November without any major issues
<chewitt> the occasional box freeze, but kids have learned where the reset buttons are :)
<ndufresne> nice
<chewitt> and I saw similar freezes on ye olde 3.14 images
<ndufresne> last time I was doing tests with GStreamer, I had specific patterns that would lead to freeze
<chewitt> the current ffmpeg v4l2 implementation needs to be reworked a little
<ndufresne> why so ?
<chewitt> we (royal) learned a bit from poking the stateless implementation for RK/Allwinner
<chewitt> and we need the same to move RPi towards vc4 .. which is planned for LE10
<chewitt> and the sooner we start some public testing of that stuff, the sooner we find bugs
<ndufresne> someone at RPi show me some work in progress V4L2 driver for the CODECs, prior to that there was DMABuf support for MMAL
<chewitt> we have a huge RPi userbase so it's something critical for us to get right
<chewitt> we have the basics working .. but lots of work to do
<ndufresne> the biggest tricky bit with v4l2 is to renegotiate the codec with a new resoltion
<ndufresne> and the seeking, where you keep flushing the codec to another IDR
<chewitt> seeking is one of the "needs improving" bits of the current Amlogic image
<ndufresne> yes I know, but it works a bit, in contrast to resolution changes
<chewitt> it works, but it's not robust enough, it's pretty simple to lock things up
<ndufresne> the second hardest part is to deal with stride and padding adaption between drivers
<ndufresne> that's were you start implementing G/S_ELECTION, so that you have a compose rectangle and an padded rectangle
<ndufresne> but then you app need to deal with figuring out (when possible) the display or color converter requirement
<ndufresne> you can always solve it the Android it, and just have a map of the requirement for your specific HW (GRAlloc), not something I can do in gst though
<chewitt> ^ all sounds greek to me :)
<chewitt> I focus on the packaging stuff .. you wouldn't want to run the code I write
sputnik_ has joined #linux-amlogic
<xdarklight> ccaione: when I try to edit the wiki (linux-meson.com) I'm being redirected to https://www.netsons.com/403/ after submitting my changes
<xdarklight> ccaione: is it possible to add https to the wiki? Chrome complains that logging into pages using http is not secure
jakogut has joined #linux-amlogic
afaerber has joined #linux-amlogic
phh has quit [Quit: No Ping reply in 180 seconds.]
phh has joined #linux-amlogic
sputnik_ has quit [Read error: Connection reset by peer]
sputnik_ has joined #linux-amlogic
sputnik_ has quit [Remote host closed the connection]
sputnik_ has joined #linux-amlogic
chewitt has quit [Ping timeout: 255 seconds]
chewitt has joined #linux-amlogic
vagrantc has quit [Quit: leaving]