ChanServ changed the topic of #linux-rockchip to: Rockchip development discussion | Wiki at http://linux-rockchip.info | Logs at http://irclog.whitequark.org/linux-rockchip | ML at http://groups.google.com/group/linux-rockchip
lkcl has joined #linux-rockchip
nasuga has quit [Ping timeout: 265 seconds]
cnxsoft has joined #linux-rockchip
nasuga has joined #linux-rockchip
cnxsoft has quit [Ping timeout: 248 seconds]
cnxsoft has joined #linux-rockchip
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft1 has joined #linux-rockchip
cnxsoft1 is now known as cnxsoft
nasuga has quit [Quit: Leaving]
JohnDoe_71Rus has joined #linux-rockchip
paulk-collins has joined #linux-rockchip
lkcl_ has joined #linux-rockchip
scelestic has joined #linux-rockchip
_whitelogger has joined #linux-rockchip
lkcl_ has quit [Ping timeout: 240 seconds]
JohnDoe_71Rus has quit [Quit: KVIrc 4.9.2 Aria http://www.kvirc.net/]
JohnDoe_71Rus has joined #linux-rockchip
betheynyx_ has joined #linux-rockchip
betheynyx has quit [Quit: (Input/output error)]
JohnDoe1 has joined #linux-rockchip
JohnDoe_71Rus has quit [Ping timeout: 245 seconds]
<ayaka> the reason I offer the gstreamer rockchip plugin is that
<ayaka> the VA-API is not ready that time, and some customer urge me to offer the 4K support
<ayaka> but I think that implementation would exist for a long time, at lease for the encoder part
<ayaka> and even for the decoder part, I don't whether it would be ready in the end of the next year, it is ready a difficult work
<ayaka> already I begin to understand the H264 syntax but it is huge work for a single one
lkcl has quit [Read error: Connection timed out]
lkcl has joined #linux-rockchip
<ayaka> lkcl, hello, I am just back from holiday
lkcl_ has joined #linux-rockchip
lkcl_ has quit [Ping timeout: 248 seconds]
<lkcl> ayaka: oh - already? :)
<ayaka> lkcl, yes, after viewed the last place in the morning, I have taken a plane and back to home
<lkcl> ayaka: i am recovering from a cold, you don't want it.
<ayaka> so you could arrange your visit to HQ now
<lkcl> cool
<ayaka> I am sorry about that
<ayaka> even the British can't bear the temperature of the southern?
<ayaka> I though it is warm enough
<lkcl> no it's not that: i've been up extremely late, odd hours
<lkcl> just looking up the address
<lkcl> can you email me a recommended hotel, what i think i will do is come over with marie and lilyana for a couple of days, towards the end of our 30-day visa limit, then we will go to macau for 5 minutes and come back in to china
<ayaka> lkcl, there is a hotel where close to my home
<lkcl> also is there one of those really really really fast trains nearby?
<lkcl> oh cool. great.
<ayaka> the Fuzhou station, not the Fuzhou south station
<lkcl> is this the right area on the map? https://www.google.com/maps/place/Tong+Pan+Lu,+Gulou+Qu,+Fuzhou+Shi,+Fujian+Sheng,+China/@26.1053925,119.1785358,11.5z/data=!4m5!3m4!1s0x344056d3f9326c2d:0x4ecb795ba1a6e4f9!8m2!3d26.1056113!4d119.2782646
<ayaka> wait a second
<lkcl> woo we have a journey of about 800km, but on the fast train that's only 4 hours :)
<ayaka> yes, it is
<ayaka> but the bus from train station to the center of the city would be a hour
<ayaka> if you is coming at Weekend, it would be possible to pick you up
<lkcl> ok.
<lkcl> hmm, we have to be out of china by the 20th, which is a tuesday.
<lkcl> we'll manage by bus, with the right info. marie speaks mandarin.
<ayaka> then you have to take a bus to here, you know the mass of Chinese station
<lkcl> yeah we'll manage :)
<lkcl> if you email me the details (hotel address, train station name, bus routes - in english and chinese) marie can review it
<ayaka> ok I would
<lkcl> thxman
<ayaka> but do remember to feed back your plan, I have to told my supervisor and let him arrange someone to talk to you
<lkcl> great.
<lkcl> well let's arrange for say 14-16 to visit rockchip, i will return saturday 17 or perhaps sunday 18.
<lkcl> so we will travel all of tuesday. i will confirm that with marie.
cnxsoft has quit [Quit: cnxsoft]
<lkcl> what do you think? 2 days @ rk HQ? or 3?
<ayaka> I am not sure, I don't know the attitude of my supervisor
<ayaka> it is ok to talk with him even on the Saturday
lkcl_ has joined #linux-rockchip
<ayaka> him won't matter work over time
lkcl_ has quit [Ping timeout: 250 seconds]
scelestic has quit [Read error: Connection reset by peer]
<lkcl> ayaka: ok great, then could you ask him, maybe we make it only one day with him, that's fine. if you have time a tour would be nice, maybe also an opportunity to meet some of the hardware engineers, show them the schematics and PCB.
<lkcl> all "unofficial" like :)
<lkcl> my main thing is that this is an "ambassador" visit, to let rockchip know what i'm doing.
<lkcl> they don't really know that the plan is to start off small and then ramp up later to absolutely MASSIVE volume.
<lkcl> but that it has to be done in cooperation and collaboration with the software libre community.
<lkcl> also that i don't *actually* need their "support" - certainly not the kind that says "pay us $100k" which i heard about a couple of years ago
<lkcl> it would be *helpful* to have a recommended distributor of the RK3288, but strictly speaking i don't even need that.
<lkcl> :)
<ayaka> lkcl, the unofficial would not be possible
<ayaka> as I am only a fresh man
<ayaka> so without his agreement, you could try to make him believe that what you are doing would help in making rockchip more well-know
<ayaka> it is only up to you
<lkcl> ayaka: that it would.... but it *has* to be via the route of "software libre". i am *NOT* - i literally *CANNOT* - sell or distribute illegal copyright-violating software, and i'm not really happy to be distributing proprietary software either.
<ayaka> yes, I am sure about that
<lkcl> ayaka: so, that means that the 100k and above sales are critically dependent on *your* project being fully completed. i can wait for that to happen, but the question is: can *rockchip* afford to wait for that to happen?
<ayaka> no, they won't ever start up a project for it
<lkcl> ayaka: it's nearly there. i'm "tolerating" MALI right now. the bootloader's libre, u-boot's libre, the linux kernel is libre... the OSes are libre.... that just leaves the VPU.
<ayaka> all the source for the vpu is open now
scelestic has joined #linux-rockchip
<lkcl> ayaka: that's okay... i just won't place the order for 100k units until they do... or until the VPU project you're doing becomes libre by another route
<ayaka> both the v4l2 implementation and for mask produce one
<lkcl> ayaka: awesome. the release of the gstreamer and first decode example was approved?
<lkcl> mask produce? you mean mass-produced?
<ayaka> lkcl, the gstreamer rockchip plugin have been released
<ayaka> yes
<lkcl> ayaka: GREAT.
<lkcl> that's a big hairy deal.
<ayaka> but it is not the final version
<ayaka> at lease there are three place need to improved
<lkcl> that's okay - as long as there's something vaguely working, i can live with that
<ayaka> need to improve
lkcl_ has joined #linux-rockchip
<ayaka> it works well, but it is written in a short time
<lkcl> that means when i make the crowdfunding campaign, i can demo video playback.
<ayaka> so some implementation way or code style need to be change
<ayaka> yes, I have ready tried the 4K video playback for HEVC and AVC
<lkcl> the important thing is to be able to attract the interest of people who want something that they can do libreoffice, internet browsing, office work, a few games, and watching some videos.
<lkcl> oooo woooow, coool.
<ayaka> but the video playing in browser would not use the VPU
<lkcl> i really want to see, with the board i'm making, if it will do 4k HDMI. that would be awesome. i made an effort to make sure the tracks are equal length
<ayaka> I need to put a gstreamer plugin for browser
<lkcl> ayaka: there's a way to make firefox use gstreamer at least for the decode engine
<ayaka> yes, in my memory, there is a plug-in for it
<ayaka> but I don't have time to try
<lkcl> i had to set it up for the A20 video demo
<lkcl> i researched this a *lot*
<lkcl> if you want it to do accelerated *display*, unfortunately by default gstreamer hands off to OpenGL
<lkcl> *bletch* :)
<phh> ayaka: current recommended branch is still vpu_interface? rkximagesync should be usable? it is still using DRM for rendering, but asks X for window overlay, I'm understanding correctly?
<lkcl> so if you have MALI up and running you'll end up with really low CPU usage
<phh> ayaka: also, are rkximagesync/rkdrmsync required, or we can use any sink?
<phh> s/sync/sink/
<ayaka> phh, the rkximagesync is not what I did
<ayaka> I would suggest you to use the kmssink(rockchip fork) for most of case
<phh> ok, I'll stick with it for the moment
<ayaka> and the future development for the performance would be done there
<ayaka> lkcl, the kmssink won't use opengl as well, but without mali and gpu, it works more fast
<phh> ayaka: ATM I only care about being able to decode 1080p h265, so that's not a problem, but ok
<lkcl> the A20 was able to do 1280x720 @ 30fps with 140% CPU usage (dual core 1ghz), that's *with* using the libcedrus VPU decoder (which actually did a lot less) and drawing to a pixbuf. it was the on-screen rendering in the browser via pixbuf that took up the enormous amount of CPU cycles... but it *did* work
<ayaka> phh, even for 4K, it is enough
<ayaka> I use the tears of steel downloaded from youtube
<ayaka> it works fine
<lkcl> ayaka: i do need to track / replicate what you're doing. where can i get the required kernel / gstreaamer fork?
<ayaka> lkcl, could you wait a second, I am testing patches for the upstream
<lkcl> ayaka: ack :)
<lkcl> email me the links / branchnames when you have a moment, i'm back to sleep soon
<ayaka> ok
<ayaka> then the tomorrow
<phh> ayaka: sorry if that's a trivial question, but I'm still not too clear, there is no requirement of vpudec to rk*, I can vpudec to anything, it will work? rk{ximagesink,drmsink} are only for higher performances?
<ayaka> phh, yes but it won't support dma
<ayaka> that is one of three need to improve
<lkcl> ayaka: no problem.
<ayaka> let phh told you about that
<ayaka> it seems that he have tried that
<phh> not really, I'm adding an extra challenge, to have the driver on top of mainline kernel
<ayaka> phh, you could
<ayaka> if you meet a problem, please check the iommu part
lkcl_ has quit [Ping timeout: 258 seconds]
<ayaka> it is really a mess sucking me a week
<ayaka> I cherry-pick more that thirty patches for that
<ayaka> more then
<phh> I feel like my problem is around rockchip_pmu_idle_request
<phh> right, I see which patches yuou're speaking of ;)
<ayaka> lkcl, you also need the mpp
<ayaka> phh, it is only about the reset
<phh> yes well, I feel like my problems are reset problems ATM
<ayaka> the vpu in rk3288 need to put the power domain into idle not the off before the reset of clock
<ayaka> phh, don't try the power off, it would shutdown a core of cpu(I don't know why)
<phh> good to know -_-'
<ayaka> and if you would like to modify it to a really drm driver, I would be appreciate
<ayaka> those display buffer current from the drm driver of vop
<phh> erm, that sounds unlikely
<phh> mmm I'm not sure I understand what you mean
<phh> well, we'll discuss that later, first I want to get it to work :P
<ayaka> ok
scelestic has quit [Read error: Connection reset by peer]
scelestic has joined #linux-rockchip
mrueg has quit [Remote host closed the connection]
mrueg has joined #linux-rockchip
<rperier> mmind00: wip/rk3188 should work ? I followed what is explained on the barebox wiki, only the rockchip SPL is started (no output from uboot)
lkcl has quit [Ping timeout: 264 seconds]
lkcl_ has joined #linux-rockchip
lkcl has joined #linux-rockchip
lkcl_ has quit [Ping timeout: 246 seconds]
lkcl_ has joined #linux-rockchip
JohnDoe1 has quit [Quit: KVIrc 4.9.2 Aria http://www.kvirc.net/]
<mmind00> rperier: not sure what you did, but the barebox wiki suggests, getting the rockchip loader, extracting FlashData and then using mkbootimg with FlashData and the u-boot.bin ... aka your commandline should not contain any mention of the uboot-spl at all
mrutland has quit [Quit: bye...]
mrutland has joined #linux-rockchip
mrutland has quit [Quit: bye...]
mrutland has joined #linux-rockchip
ssvb has quit [Ping timeout: 265 seconds]
lkcl_ has quit [Ping timeout: 258 seconds]
lkcl_ has joined #linux-rockchip
paulk-collins has quit [Remote host closed the connection]
ssvb has joined #linux-rockchip
nighty-- has quit [Quit: Disappears in a puff of smoke]
lkcl_ has quit [Read error: Connection reset by peer]
nighty has quit [Quit: Disappears in a puff of smoke]