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
pionen has joined #linux-amlogic
dsd_ has quit [Quit: Lost terminal]
pionen has quit [Quit: leaving]
vagrantc has quit [Quit: leaving]
cthugha has joined #linux-amlogic
dsd has joined #linux-amlogic
cthugha has quit [Read error: Connection timed out]
ldevulder has quit [Ping timeout: 276 seconds]
cthugha has joined #linux-amlogic
dsd has quit [Quit: Lost terminal]
chewitt has quit [Quit: Adios!]
Barada has joined #linux-amlogic
Elpaulo_m has joined #linux-amlogic
commavir has quit [Ping timeout: 246 seconds]
commavir has joined #linux-amlogic
Elpaulo_m2 has joined #linux-amlogic
cthugha is now known as ldevulder
Elpaulo_m has quit [Ping timeout: 240 seconds]
Elpaulo_m has joined #linux-amlogic
Elpaulo_m2 has quit [Ping timeout: 240 seconds]
brujah has quit [Ping timeout: 240 seconds]
yann has quit [Ping timeout: 240 seconds]
brujah has joined #linux-amlogic
cyrozap has quit [Ping timeout: 256 seconds]
cyrozap has joined #linux-amlogic
yann has joined #linux-amlogic
yann has quit [Ping timeout: 264 seconds]
yann has joined #linux-amlogic
brujah has quit [Ping timeout: 248 seconds]
yann has quit [Ping timeout: 268 seconds]
brujah has joined #linux-amlogic
yann has joined #linux-amlogic
a5m has joined #linux-amlogic
yann has quit [Ping timeout: 264 seconds]
yann has joined #linux-amlogic
afaerber has quit [Quit: Leaving]
afaerber has joined #linux-amlogic
Kamikaze_84 has joined #linux-amlogic
dsd_ has joined #linux-amlogic
<Kamikaze_84> Hi, posted the other day about getting usb1-port1: disabled by hub (EMI?), re-enabling... messages on my nexbox a95x running linux-next with latest dts patches applied so I could run USB host mode. Just wondering if you think this is probably hardware related or if it could be driver related? Also apologies if I disconnect soon as my device is low on battery, will check logged messages later for any reply.
brujah has quit [Ping timeout: 276 seconds]
Kamikaze_84 has quit [Ping timeout: 260 seconds]
brujah has joined #linux-amlogic
Barada has quit [Quit: Barada]
Barada has joined #linux-amlogic
Elpaulo_m2 has joined #linux-amlogic
vagrantc has joined #linux-amlogic
Elpaulo_m has quit [Ping timeout: 268 seconds]
Elpaulo_m has joined #linux-amlogic
Barada has quit [Quit: Barada]
Elpaulo_m2 has quit [Ping timeout: 276 seconds]
Elpaulo_m2 has joined #linux-amlogic
Elpaulo_m has quit [Ping timeout: 264 seconds]
Elpaulo_m has joined #linux-amlogic
Elpaulo_m2 has quit [Ping timeout: 252 seconds]
Elpaulo_m2 has joined #linux-amlogic
Elpaulo_m has quit [Ping timeout: 240 seconds]
Elpaulo_m has joined #linux-amlogic
Elpaulo_m2 has quit [Ping timeout: 240 seconds]
Elpaulo_m2 has joined #linux-amlogic
Elpaulo_m has quit [Ping timeout: 276 seconds]
yann has quit [Ping timeout: 252 seconds]
brujah has quit [Ping timeout: 256 seconds]
yann has joined #linux-amlogic
brujah has joined #linux-amlogic
a5m has quit [Remote host closed the connection]
yann has quit [Ping timeout: 264 seconds]
commavir has quit [Ping timeout: 276 seconds]
commavir has joined #linux-amlogic
yann has joined #linux-amlogic
yann has quit [Ping timeout: 240 seconds]
<Ely> ndufresne: Have you been able to give more time to the dts64 ?
<ndufresne> not in the last 3 days :-(
<Ely> ah no problelm
<Ely> I actually just received my Potato so I'm gonna be able to get onto it as well
<Ely> did you push your progress somewhere ?
<ndufresne> no, what I started writing is a small C script that include the ucode headers and generates files from it
<ndufresne> it's a bit of a one time thing, considering passed 3.14 they don't use that anymore
<Ely> alright
<ndufresne> my goal right now is a bit small, due to being busy and all over the place, but it's to have a linux-firmware branch, with amlogic/, and iterate that until we think it's right
<ndufresne> though, about the GXL, it's the GXTVBB firmwares that you need
<ndufresne> the rest is pretty much the same, except that _single variant can be used pretty much everywhere
<ndufresne> I wonder if the core is just faster/clock higher
<ndufresne> Ely, you'll probably be able to fix the DT much faster then me, as I'm still learning ;-P
* ndufresne wish I had couple of weeks of free-time
<Ely> hehe
jakogut has quit [Quit: jakogut]
<Ely> So for S905X / GXL, I need the GXTVBB firmware in 4.9 ?
yann has joined #linux-amlogic
<Ely> actually, where I can find those at all ?
brujah has quit [Ping timeout: 264 seconds]
<Ely> oh and if I remember correctly, you said the vh264_mc in my repo is the same as the one in 3.14 headers ?
brujah has joined #linux-amlogic
yann has quit [Ping timeout: 256 seconds]
brujah has quit [Ping timeout: 264 seconds]
jakogut has joined #linux-amlogic
yann has joined #linux-amlogic
brujah has joined #linux-amlogic
trem has joined #linux-amlogic
Elpaulo_m has joined #linux-amlogic
<Ely> narmstrong: can I just dd my-image-libretech-cc.wic to the sdcard ?
Elpaulo_m2 has quit [Ping timeout: 256 seconds]
<narmstrong> @Ely: no switch to my master branch I dropped the wic and added a .sdcard file you can directly did
<narmstrong> *dd
<narmstrong> Like meta-odroid but better :-D
<Ely> ohhh your default branch is rocko I didn't notice
yann has quit [Ping timeout: 264 seconds]
<Ely> hehe
<Ely> with mainline uboot, me likey
<narmstrong> Yep did it yesterday :-)
<narmstrong> I hope it will build...
<Ely> It failed but it's on my end, I didn't whitelist u-boot-1 to meta-clang
<Ely> u-boot*
<Ely> It's building the rootfs so that's a good sign :D
<Ely> | dd: failed to open '/mnt/poky/build/tmp/deploy/images/libretech-cc/u-boot.bin.sd.bin': No such file or directory
<Ely> ERROR: Task (/mnt/poky/meta/recipes-extended/images/my-image.bb:do_image_sdcard) failed with exit code '1'
<Ely> Looks like I'm missing something
<Ely> is this a piece of firmware I should get somewhere ?
<narmstrong> It should be installed by u-boot !
<Ely> There's a u-boot.bin and u-boot-libretech-cc.bin
<Ely> maybe a naming issue ?
<Ely> both point to the same file
<Ely> u-boot-libretech-cc-2018.01-r0.bin
<narmstrong> It’s not this one, this one is the u-boot only, the .sd.bin one has the fip installed
<Ely> ah
<narmstrong> I can’t look right now
<Ely> no worries I'll figure it out
* narmstrong trying to do a fresh rebuild
<Ely> for some reason the u-boot recipe in meta/ was called.. Even though your layer has higher priority and they have the same version
<narmstrong> ah you are on poky master ?
<Ely> yeh
<Ely> bleeding the edge
<narmstrong> i should rename the u-boot bb to avoid conflicts
<Ely> Still though there's no reason the one in meta/ was called
<Ely> can't really figure that one out
<narmstrong> for sumo it should be a bbappend
<Ely> Alright
<Ely> Now it's amlogic-fip-libretech-cc not working :<
* Ely should maybe checkout rocko
<narmstrong> yeah the recipe is not very clean, cleanall it
<Ely> I see it's fetching some very recent toolchains
<Ely> :D
<narmstrong> yep, I could avoid that since I prebuilt these for LibreELEC
<narmstrong> but I was lazy and reused the amlogic u-boot recipe !
dickon has left #linux-amlogic [#linux-amlogic]
<narmstrong> I will switch master to sumo since it will be released in the next days
<Ely> I was missing zlib1g:i386 on my system, fortunately it was the only error
<Ely> narmstrong: cool!
<Ely> Alright got my image, thanks a lot
<Ely> and it boots first time no problem
<Ely> woo!
<narmstrong> yeah !
<Ely> if I want to update the kernel/dtb, I just update files in /boot ?
<narmstrong> yep
<Ely> fantastic
<narmstrong> and modules in /lib ;-)
<Ely> yes that one is more standard :P
<Ely> on the C1 I had to update the kernel/dtb by mounting the vfat partition first
<narmstrong> yep, the .sdcard image is a single ext4 partition, all in one thanks to mainline u-boot supporting ext4 and extlinux.conf !
<narmstrong> another argument to ditch the old buggy amlogic u-boot...
<Ely> There has to be a firmware/u-boot at a magic offset somewhere before the partition though, no ?
<narmstrong> yep in the first MBR sector, and 444bytes at offset 0
<narmstrong> this I why I can't use wic...
<Ely> :D
<lp0-on-fire> I just wish whoever started to port environment to kconfig in mainline u-boot would finish it
<lp0-on-fire> :D
<lp0-on-fire> because right now I'm stuck having my env literally nowhere
<narmstrong> people are never satisfied !
<lp0-on-fire> well, it's more like I just have no env storage, bit complicated to update embedded stuff
<lp0-on-fire> can't store the env in mmc device because if you enable that option the build fails
<Ely> also I pushed to https://github.com/Elyotna/linux yersteday & today: the H.264 decoder is now really stable, and I was able to map the vb2 mplane buffers to the canvases directly, thus dropping the big-ass 64MiB alloc.
<Ely> I'm probably gonna start modularizing the driver now and adding mpeg 1/2/4
<narmstrong> cool !
<Ely> .. and getting it to work on S905X which I hope won't be too much of a pita
<narmstrong> yeah could be great to try it on libreleec/kodi with software rendering since they already use ffmpeg !
<narmstrong> we may gain enough to have smoother decoding on 1080p
vagrantc has quit [Quit: leaving]
<Ely> Problem remains the tiling though. It'll be faster for sure but the picture will look terrible :D
<narmstrong> yeah, sure !
<narmstrong> so we will know the HW is actually decoding
<Ely> :)
sputnik_ has quit [Remote host closed the connection]
<narmstrong> did you try gstreamer aswell +
<narmstrong> ?
<Ely> no I haven't, I'm only testing with ffmpeg -c:v h264_v4l2m2m right now
<Ely> I dump the raw "nv21" frames and play them later on my PC
sputnik_ has joined #linux-amlogic
<narmstrong> ok, my colleague @ldts will be happy to know that !
<Ely> ldts?
<narmstrong> he thinks nobody uses the ffmpeg v4l2m2m
<narmstrong> Jorge Ramirez-Ortiz, he pushed the m2m support
<Ely> oh It's someone in the chat :D
<Ely> Cool, awesome work
<narmstrong> yep^
<Ely> been really helpful for me on this project
* ndufresne hope to have time to keep improving it, though ffmpeg needs DMABuf support
<ndufresne> Ely, btw, gstreamer have larger support for that, but might be a tad more picky with the driver
<ndufresne> (supports more features, that's why)
<Ely> ndufresne: Would you know of a simple gst-launch pipeline to test a V4L2 M2M driver ?
<Ely> and say output the frames in a file
<Ely> I am very, very bad at gstreamer (almost never used it)
<ndufresne> sure, first step is to make sure you driver is detected as such, so gst-inspect-1.0 video4linux2
<ndufresne> you can touch /dev/videoN to force it to probe again
<ndufresne> when it should show up element like v4l2h264dec (I guess that's all you have implemented)
<Ely> okay
<ndufresne> and a simple pipeline would be gst-launch-1.0 filesrc location=media ! parsebin ! v4l2h264dec ! kmssink
<ndufresne> well, kmssink might not work, we don't have any YUV overlay yet
<ndufresne> so maybe gst-launch-1.0 filesrc location=media ! parsebin ! v4l2h264dec ! videoconvert ! kmssink
* Ely saves notes in a file
<ndufresne> you need 1.12+, obviously, the most recent gst is better, since I keep working on that ;-P
<ndufresne> The Venus driver was written using both gst and ffmpeg, since they trigger different corner case
<ndufresne> (and the ffmpeg backend was written in parallel)
<Ely> yocto has 1.12.4, hopefully this should be enough
<ndufresne> should do
<Ely> I'll try as soon as I get the basics on S905X
<ndufresne> oh, and you'll see GST_DEBUG="v4l2*:7" to make any sense (or to ask me to make any sense) or errors
<Ely> Thanks
<Ely> OOT, but regarding the firmwares, where can I find the ones for gxtvbb btw ?
<Ely> even the ones in raw files or anything. I'm just a bit lost about whether the ones in 3.14 will do or if I should look elsewhere
<Ely> Like where did you find all the _single and _mc firmwares ?
steamport|sleep is now known as steamport
<ndufresne> sorry, yes, let me give you pointers
<ndufresne> the base directory is drivers/amlogic/amports/arch/ucode/
<Ely> in 3.14 ?
<ndufresne> yes, aml-3.14, as pulled by buildroot
<ndufresne> Ely, what I verified is that the one you have put in your branch, is that same as the one defined as vh264_mc
<ndufresne> drivers/amlogic/amports/arch/ucode/h264/vh264_mc.c
<Ely> great
<ndufresne> drivers/amlogic/amports/arch/ucode/h264/h264c_linux.h
<ndufresne> And for the S905x, it seems that the one defines as gxtvbb_vh264_mc is used
<ndufresne> (it has multiple part, header, data, mmco, list, slice (whatever that means)
<ndufresne> (So the one from gxtvbb_h264header_linux.h)
yann has joined #linux-amlogic
<Ely> awesome thanks for the investigation :)
<ndufresne> though not used in the code, vh264_mc.c flags that gxtvbb firwmare as being for GXTVBB/GXL/GXM/TXL/TXLX
<ndufresne> but I would not really trust that, because then they also define gxm_vh264_mc for GXM and TXL/TXLX
afaerber has quit [Quit: Leaving]
<Ely> haha
<ndufresne> there is a if/else if/... in drivers/amlogic/amports/vh264.c that I think we can trust, line 2623
yann has quit [Remote host closed the connection]
<Ely> yup
<Ely> MESON_CPU_MAJOR_ID_GXTVBB
<Ely> #define MESON_CPU_MAJOR_ID_GXTVBB 0x20
<Ely> #define MESON_CPU_MAJOR_ID_GXL 0x21
<Ely> #define MESON_CPU_MAJOR_ID_GXM 0x22
trem has quit [Quit: Leaving]
dsd_ has quit [Quit: Lost terminal]
<Ely> Oh boy.. You guys also have been hit by the wonders of "amlogic, codec_io" ?