rellla changed the topic of #linux-sunxi to: Allwinner/sunxi /development discussion - did you try looking at our wiki? - Don't ask to ask. Just ask and wait! - - Logs at - *only registered users can talk*
popolon has quit [Quit: WeeChat 2.4]
aalm has quit [Ping timeout: 258 seconds]
aalm has joined #linux-sunxi
altious has joined #linux-sunxi
lurchi__ is now known as lurchi_
lurchi_ is now known as lurchi__
chewitt has joined #linux-sunxi
victhor has quit [Ping timeout: 258 seconds]
cnxsoft has joined #linux-sunxi
OutBackDingo has quit [Ping timeout: 246 seconds]
OutBackDingo has joined #linux-sunxi
tkaiser_ has joined #linux-sunxi
tkaiser has quit [Ping timeout: 246 seconds]
chewitt has quit [Quit: Zzz..]
shfil has quit [Quit: Connection closed for inactivity]
chewitt has joined #linux-sunxi
dddddd has quit [Remote host closed the connection]
chewitt has quit [Quit: Adios!]
kaspter has quit [Ping timeout: 246 seconds]
kaspter has joined #linux-sunxi
kaspter has quit [Ping timeout: 258 seconds]
xcko has quit [Remote host closed the connection]
xcko has joined #linux-sunxi
kaspter has joined #linux-sunxi
lurchi_ has joined #linux-sunxi
lurchi__ has quit [Ping timeout: 246 seconds]
foxx_ has joined #linux-sunxi
camus has joined #linux-sunxi
kaspter has quit [Quit: kaspter]
camus has quit [Ping timeout: 246 seconds]
selfbg has joined #linux-sunxi
Perlovka_ has joined #linux-sunxi
Perlovka has quit [Ping timeout: 268 seconds]
Putti has joined #linux-sunxi
tllim has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
Putti has quit [Ping timeout: 258 seconds]
reinforce has joined #linux-sunxi
msimpson has joined #linux-sunxi
cmeerw has joined #linux-sunxi
red_devil1980 has quit [Quit: Leaving.]
msimpson_ has joined #linux-sunxi
msimpson has quit [Ping timeout: 268 seconds]
wwilly__ has quit [Quit: Leaving]
wwilly has joined #linux-sunxi
nuuuciano_ has quit [Ping timeout: 258 seconds]
<wens> MoeIcenowy: got some nokia dc cables and they were too small lol
<wens> MoeIcenowy: seems the q8 tablets might be using 2.5/0.7 mm plugs
foxx_ has quit [Ping timeout: 255 seconds]
vagrantc has quit [Quit: leaving]
foxx_ has joined #linux-sunxi
jerbob92 has quit [Remote host closed the connection]
kaspter has joined #linux-sunxi
kaspter has quit [Ping timeout: 268 seconds]
kaspter has joined #linux-sunxi
mzki has quit [Ping timeout: 255 seconds]
kaspter has quit [Ping timeout: 246 seconds]
kaspter has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
tllim has quit [Ping timeout: 245 seconds]
tllim has joined #linux-sunxi
tllim has quit [Client Quit]
tnovotny has joined #linux-sunxi
Gerwin_J has quit [Read error: Connection reset by peer]
Gerwin_J has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
dddddd has joined #linux-sunxi
Gerwin_J has quit [Read error: Connection reset by peer]
megi has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
mpmc has quit [Quit: ZNC 1.7.2+deb2~bpo9+1 -]
wwilly_ has joined #linux-sunxi
mpmc has joined #linux-sunxi
Gerwin_J has quit [Read error: Connection reset by peer]
wwilly has quit [Ping timeout: 252 seconds]
Gerwin_J has joined #linux-sunxi
Gerwin_J has quit [Read error: Connection reset by peer]
\\Mr_C\\ has joined #linux-sunxi
afaerber has joined #linux-sunxi
msimpson_ has quit [Ping timeout: 246 seconds]
msimpson_ has joined #linux-sunxi
tuxillo has quit [Remote host closed the connection]
tkaiser has joined #linux-sunxi
tkaiser_ has quit [Read error: Connection reset by peer]
tuxillo has joined #linux-sunxi
<dddddd> MoeIcenowy, Asara... u-boot for Olimex TERES-I doesn't work because (as jo0nas found) since f89d6133eef2 we need CONFIG_SPL_TEXT_BASE in the defconfig.
* jo0nas is preparing a patch
<jo0nas> seems beelink_gs1_defconfig is also affected, but I don't know that board so cannot provide patch
kaspter has quit [Ping timeout: 246 seconds]
Andy-D has joined #linux-sunxi
kaspter has joined #linux-sunxi
<jo0nas> why are my patches flagged as non-member posts requiring moderator approval?
victhor has joined #linux-sunxi
kaspter has quit [Ping timeout: 258 seconds]
msimpson_ has quit [Ping timeout: 244 seconds]
dirbaio3 has left #linux-sunxi [#linux-sunxi]
msimpson_ has joined #linux-sunxi
Javi-IMG has joined #linux-sunxi
msimpson_ has quit [Ping timeout: 246 seconds]
msimpson_ has joined #linux-sunxi
jbizcocho has joined #linux-sunxi
<jbizcocho> Hello?
<jbizcocho> Is there any A80 developers here?
Javi-IMG has quit [Quit: Page closed]
<jbizcocho> I'm from Imagination Technologies. We've prepared a new driver for the A80 (kernel version 3.4.39), this driver contains the latest OpenGLES, OpenCL and Vulkan drivers, so A80 developers can start using the Rogue GPU again
<KotCzarny> is this driver binary only? or can be ported to latest kernel?
<jbizcocho> Our intention is to keep supporting this platform with the latest drivers
<jbizcocho> but we've seen that the latest kernels are not well supported in this platform
<KotCzarny> 3.4.39 is ancient, no one uses it anymore
<mru> ancient doesn't quite describe it
<jbizcocho> Our ideal port would be for Kernel 4.4
<jbizcocho> but we couldn't even get it to boot properly
<mru> 4.4 is ancient, 3.4 is antediluvian
<libv> jbizcocho: partly because of the pvr, A31(s) and A80 have not been very popular
<karlp> 4.4 is still technically LTS at least... but pretty poor starting point yes.
<jbizcocho> We want to change that, and start supporting properly certaing platforms
<jbizcocho> we can port to any kernel you guys want
<KotCzarny> i think A31 includes PVR too, and is much more supported than A80
<libv> jbizcocho: img had the opportunity to do that half a decade ago
<libv> jbizcocho: i am the guy who started lima
<mru> 4.9, 4.14, and 4.19 are also lts
<libv> and i am also the guy who made sure that no-one wasted time on pvr
<karlp> we're moving forwards here libv, no need to be stuck int he past
<libv> karlp: then it's all yours
<jbizcocho> Well, I'm a Linux guy who wants to change things from inside the company
<libv> jbizcocho: do you really think i have not heard that before?
<jbizcocho> So the first step for this is to start supporting platforms with new drivers
<jbizcocho> I have here the latest DDK 1.11 ready to share it
<jbizcocho> It's done and I have green light
<libv> my "mole" inside the company left quite a while ago
<jbizcocho> I see, well now you don't even need a mole
<libv> but there were like yearly or half-yearly all hands where open sourcing was brought up, and then reasoned down
<jbizcocho> I am here to change this aspect from IMG
<jbizcocho> Ok, for the momment
<libv> jbizcocho: imho, you should first actually change img, then come back
<jbizcocho> Opensourcing anything is out of the question
<jbizcocho> but we can support the latest kernels
<libv> well, so much for moving to the future
<jbizcocho> with binaries
<libv> karlp: as said, it's all yours
<jbizcocho> My intention is to transform the A80 into an usable platform from the GPU standpoint
<jbizcocho> after this we'll start supporting also other platforms
<wens> jbizcocho: mainline (5.1) should work well by now
<jbizcocho> But the A80 is not supported in 5.1
<wens> where did you read that?
<jbizcocho> also, the latest kernel that we support is 4.15
<jbizcocho> there is a matrix in the sunxi website
<jbizcocho> the A80 kernel support is really limited
<jbizcocho> We would ideally want to support LTS kernels
<wens> my bad, I should've updated it :/
<wens> let me check which version started supporting display on a80
kaspter has joined #linux-sunxi
<mru> jbizcocho: what's stopping you?
<jbizcocho> because our iteration process for Q-A
<libv> jbizcocho: would it not make sense to do renesas instead?
<libv> you actually have customers there
<jbizcocho> what do you mean?
<wens> jbizcocho: RGB out on the A80 is supported starting with 4.17. With the Cubieboard 4 (CC-A80), which has an RGB-to-VGA converter, you can have display output
<libv> are the rcars no longer shipping rogues?
<jbizcocho> Yes
<libv> starting with allwinner is wrong from an img internal business pov
<libv> and it very much sounds like going with allwinner is a personal project and not a img management approved thing
<jbizcocho> From all the platforms that exists out there, we're going to select a few to support them
<jbizcocho> with latest drivers directly from imagination
<jbizcocho> and each customer will be asked individually
<jbizcocho> for the moment AllWinner is ok
<wens> I have to agree with libv. I don't think AW is even selling the A80 anymore
<jbizcocho> and we are talking with Mediatek
<jbizcocho> Renesas doesn't want us to tinker with their platforms
<jbizcocho> so we respect that
<jbizcocho> we cannot wildly start supporting platforms
<jbizcocho> It doesn't work like that
<libv> jbizcocho: you can also not wildly expect spare-timers to do imgs work for it
<wens> I thought you guys already released something for Mediatek, IIRC in some ChromiumOS tree
<libv> especially in light of how img has been viewed for the best part of the decade
<jbizcocho> Yes, for the Chromebook
<jbizcocho> the idea is to generate development platforms that we can support
<jbizcocho> for example with the new PVR-GDB GPU debugger that we are going to release
<jbizcocho> this is a GPU debugger that doesn't work in retail drivers
<libv> is this pdump, or an evolution of it?
<libv> i used to work for nokia and know a thing or two about the pvr sgx and its debugging infrastructure
<libv> which is why i made sure that no-one wastes time with pvr
<mru> and just how did you do that?
<libv> mru: the making sure that no-one worked on powervr?
<mru> yes
<mru> I doubt you have the authority to dictate what others do
<libv> when i did lima, people came to me
<libv> and i explained, time and time again, why pvr was a waste of time
<libv> robclark wanted to go pvr
<libv> erik faye lund wanted to go pvr
<libv> i talked then out of it
<mru> do you mean when you promised to do lima, then stalled forever? or was that someone else?
<libv> mru: i actually did a lot of it
* karlp grins
* wens sigh
<KotCzarny> yeah, would be fun seeing lima-ng without basing off lima
<KotCzarny> which is sad, because mali4xx and pvr (older ones) are old as fuck
<KotCzarny> yet, kept as national treasures
<libv> also, all of the arm gpu REing projects pretty much copy pasted the basic RE process and some code
<wens> KotCzarny: cheaper with regards to licensing and/or integration?
<KotCzarny> wens, yeah, no time spent == no money spent, which checks out, until competition chips are getting hacked opensource drivers
<libv> as for the pvr, people have taken my statements as gospel, and you see my statements parrotted all over the place in the years following the initial release of my work, and thus the time where i spent talking down imagination technologies
* mru thinks someone is a little full of himself
msimpson_ has quit [Ping timeout: 246 seconds]
<libv> heh
<libv> mru: feel free to spend half a day on google and find when people started talking about pvr being a moving target and not worth REing
<libv> this was my strategy from the start, march 2011, from before bionic had LD_PRELOAD
<libv> as at that time, i was still being paid by nokia, particularly, to provide reliable sequential pdump through debugfs
<libv> which broke the imagination simulator
<libv> and no-one could tell us what was wrong
<libv> and of course, not even nokia was allowed to have or run the simulator
<jbizcocho> sorry I'm not aware of any of this
<mru> so you convinced everybody not to work on pvr
<jbizcocho> I know that historically IMG has not been proactive supporting platforms
<libv> mru: the current dogma that pvr is not worth REing has its origins in the lima irc channel and my blog posts
<jbizcocho> But I still don't get where is all this negativity towards img coming from
<mru> then you promised to work on lima, and others waited so as not to duplicate effort, but you never delivered
<mru> that's two gpu platforms sabotaged
<mru> and somehow you're supposed to be the good guy here?
<karlp> jbizcocho: nothing you've done, don't worry.
fALSO has quit [Remote host closed the connection]
<libv> mru: and you are? and what did you do?
<mru> I didn't sabotage anyone
<mru> so that's a start
<libv> secondly, img never actively acted against REing
<libv> they just have a "technical solution" that is not worth REing
<jbizcocho> I don't even know what REing means
<KotCzarny> reverse engineering
<libv> which is what i explained to the world, and which people then took on
<jbizcocho> Listen guys, I want to change things around
<jbizcocho> really
<jbizcocho> Raytracing we'll come (hopefully soon)
<jbizcocho> and the new GPU 19 iteration
<jbizcocho> I want to prove managment that this is worth it
<jbizcocho> starting with the A80
<diego71> jbizcocho: what about TI cpu?
<karlp> jbizcocho: so just keep oin keeping on, but do try and start somewhere more modern than 4.4, certainly don't bother with 3.x..
<jbizcocho> and honestly libv I don't feel you're helping
<libv> jbizcocho: perhaps
<libv> jbizcocho: but i have heard this before
<jbizcocho> Well I don't mind that
<jbizcocho> I'm telling you that this will change
nuuuciano_ has joined #linux-sunxi
<jbizcocho> no opensoruce, but at least we will release new drivers
<jbizcocho> for latest kernels
<libv> and starry-eyed good intentions lead to a lot of disappointment
<libv> jbizcocho: which then is nothing the sunxi community should spend time on
<mru> libv: maybe you should spend a little less time telling others what to do
<libv> mru: unlike you, i actually have spent a lot of time on the topic of arm gpus
<mru> jbizcocho: I would urge you to make the kernel parts of the driver open source
<libv> they are.
<mru> keep the userspace part closed if you must
<libv> *facepalm*
<jbizcocho> The kernel is already open
<mru> ok, good
<jbizcocho> but we give it on demand
<mru> it wasn't always
<libv> mru: datapoints.
<libv> i have seen no mention, ever, of img not adhering to the GPL
<jbizcocho> Ok, so is ok if we target 4.14?
<mru> I never said they didn't
<mru> jbizcocho: why not 5.2?
<jbizcocho> is it LTS?
<mru> no, but it's the latest
<mru> 4.19 is latest lts
<jbizcocho> We need to support LTS, otherwise it might be too time consuming (at the beginning)
<mru> 5.4 or so will likely be the next one
<mru> and you won't be ready before that's released
kaspter has quit [Ping timeout: 246 seconds]
<jbizcocho> According to the kernel LTS list
fALSO has joined #linux-sunxi
<jbizcocho> supporting anything appart from 4.4 or 4.9 doesn't make much sense
<jbizcocho> because 4.19 EOL is 2020
<jbizcocho> whereas 4.9 is 2023
<mru> those are ancient
<jbizcocho> what is the benefit of getting the latest
<jbizcocho> is there any feature that is different?
<mru> why are they still making new versions?
<jbizcocho> have you benchmark them? is there any difference in performance?
<KotCzarny> yes
<jbizcocho> Can you point me to some results page?
<KotCzarny> jump from 4.9 to 4.14 was ~20% when i compared network stack on marvel 375 board
tllim has joined #linux-sunxi
<jbizcocho> Is there any guide for me to get 4.14 working on the A80?
<jo0nas> jbizcocho: one benefit of getting the latest is that it will stay maintained for longer
<jbizcocho> Right now our kernel modules are not integrated in mainline kernel
<jbizcocho> so we have to make extra effort per kernel revision
<jbizcocho> this whole project could help for us to integrate in our mainline kernel
<buZz> why not upstream your modules then? :P
<jbizcocho> I'll propose that once we get this first one up and running
<libv> jbizcocho: "this whole project" being your project
<libv> i would dread to see anyone waste his spare time on pvr, while there is so much other crap to do
<buZz> oh lol ; 13:45:42 < jbizcocho> Opensourcing anything is out of the question
<mru> libv: shut up, you're not helping
<buZz> imgtec ? :P
<libv> mru: careful.
<mru> or what?
* buZz spreads some love
<mru> you'll tell me what to do?
<jbizcocho> Let's be productive, any hint on how to get 4.14 running here on my board? Using 18.04 for example
<libv> mru: feel free to take your conversation private if you do not want others to butt in
<buZz> jbizcocho: why isnt 4.14 working as-is ? isnt A80 pretty ok supported? (beside video)
<mru> libv: feel free to start your own channel if want to dictate the conversation
<libv> mru: i actually did
<libv> in 2012
<libv> but i refrain from telling others to shut up
<jbizcocho> Common guys this is not useful
<jbizcocho> I am looking right now at
<jbizcocho> it looks not very well supported to me
<jbizcocho> Ok so I just need to grab 4.14 from kernel org right?
<jbizcocho> no special patcher or anything?
<jbizcocho> patches*
<libv> jbizcocho: wens stated that he probably did not keep things in the wiki too up to date, he is known to actually go document things if/when he finds time
<libv> (with respect to a80)
<jbizcocho> Sorry I didn't read that
<jbizcocho> Too much text
<KotCzarny> jbizcocho: that's the point of 'mainlining', nothing additional needed to run once soc is mainlined, might not be true for additional devices (touch, display etc)
chewitt has joined #linux-sunxi
<libv> KotCzarny: on the other hand, img has to support "commercial" customers, and telling everyone to only use mainline there is also a no-go
<jbizcocho> KotCzarny thanks for the note
msimpson has joined #linux-sunxi
<libv> img made its bed years ago, and gets to sleep in it, and i still maintain that no-one from the sunxi community should spend his/her time on this
<jbizcocho> Ok, so last question, as I said before I have here the drivers baked for 3.4. Is there any place in the sunxi community where we can upload them?
<jbizcocho> Or you guys prefer for us to host in IMG website
<libv> jbizcocho: as you can imagine, i am not too keen on hosting your binaries on our servers
<jbizcocho> Why not?
<jbizcocho> I still don't understand your hostility
<libv> because i do not want people to think that we officially ship this
<karlp> more usefully, I'm not sure anyone's interested in binaries for 3.4.
<jbizcocho> Don't know where is coming from
<jbizcocho> Is not a good thing for Sunxi to partner with us in this?
<libv> jbizcocho: img can provide tarballs on its own servers, and link to them from useful howtos in our wiki
<jbizcocho> Ok, that can work
<libv> jbizcocho: there is a difference between us hosting your binaries and us telling people how to use your binaries
<jbizcocho> Well, the A80 could be promoted to a developer platform for graphics
<jbizcocho> using Vulkan and so on
<libv> jbizcocho: also, it makes sense for you guys to keep things in git trees on github or so
<jbizcocho> Sure
<libv> and to have your kernel side and userspace side in the same repo as to avoid further abi inconsistencies
<jbizcocho> I agree
<libv> and to have this documented on our wiki
<jbizcocho> that sound great
<jbizcocho> Also, we will release a Wayland version
<jbizcocho> so it will be possible to run Weston on it
<jbizcocho> and perhaps Gnome 3
<KotCzarny> dont forget to license properly, so it wont be tied to some specific board or manufacturer
<jbizcocho> We'll the driver will just work on the A80
<KotCzarny> or distro maintainers will be forced to ignore them
<jbizcocho> Okay
<jbizcocho> just in case anyone wants to try out the 3.4 drivers
<jbizcocho> I'll provide ftp access
<jbizcocho> when I have 4.14 running, I'll come back
<jbizcocho> meanwhile we'll start putting together something more professional in our website
<mru> why 4.14?
<KotCzarny> mru: he said 4.15 is the limit
<jbizcocho> Is the latest driver that passed our Q&A
<libv> jbizcocho: as for the license, do not make the mistake arm did initially
<jbizcocho> which is... ?
<libv> do not use a click through eula that prohibits things
<libv> just state that the binary is provided without warranty and such
<jbizcocho> yes that's the idea
<libv> statements like "no REing or dissassembly" are just ludicrous
<jbizcocho> this will be EDs
<jbizcocho> for developers to have the latest features
<libv> both from a legal pov, and from the reality that no-one cares about pvr
<libv> (legal pov, at least in the eu)
<libv> so such a eula is just about annoying users
<jbizcocho> Okay
<wens> libv: which part doesn't stand in EU? click through EULA or no REing?
<libv> the reverse engineering clauses are just bs.
chewitt has quit [Quit: Adios!]
<libv> and the eula therefor is just annoying
<libv> i am not sure what the legal status of click-through eulas is
kaspter has joined #linux-sunxi
<jbizcocho> Guys
<jbizcocho> is there any place where I could upload the driver?
<jbizcocho> our public ftp is not working properly
<jbizcocho> this is for 3.4.39
<libv> jbizcocho: throw these things in a git tree, so you have unified version control between kernel and userspace
<libv> and then throw it up on github
<libv> kernel here being, the kernel part of the rogue codebase
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 -]
MoeIcenowy has joined #linux-sunxi
Gerwin_J has joined #linux-sunxi
<libv> jbizcocho: if needs be, use git-submodules to keep the gpled and binary parts physically separate
<libv> but please provide solid abi control, otherwise you will lose your users as fast as you get them
<buZz> jbizcocho: re; 3.4, i doubt anyone is going to run it
<buZz> but well, at least uploading it somewhere would be useful i guess
<libv> buZz: well... what we experienced at nokia, and one of the reasons why people did take up my statements and strategy towards img, is that it's a constant struggle with img to keep the data structures between kernel, userspace and microcode in conformance
<buZz> agreed, i've seen powervr work and not work on old kernels, but just from sideline
<buZz> from the #dragonbox-pyra channel
<buZz> and generally multiple TI OMAP devices
<libv> structures would shift depending on chip revision, build options, and even vendor specific changes (usually implemented by img)
<libv> and keeping kernel and userspace (which includes the microcode) in separate locations is a sure fire way to almost never have anything working
<buZz> ah, JonnyH is in that channel
<buZz> i -think- he works @ imgtec
<libv> so just providing binaries in a tarball is a very short term solution
<buZz> exactly
<buZz> probably 'totally fine' for corperate clients
<buZz> but nowhere -near- fine for 'linux users'
<libv> well, at nokia it was a constant struggle
<buZz> libv: did you see maemo-leste? :)
<buZz> does it warm your heart?
<libv> i did, but not really
<buZz> :)
<buZz> ah well, i like that ppl are reviving maemo
<libv> maemo has been dead since 20110211
<buZz> it did a lot right
<libv> and i am not sure it is the best use of valuable time
<KotCzarny> depends on the definition of 'dead'
<aalm> rotten?
<KotCzarny> libv: maemo-leste is targetting any open source linux with gpu stack working
<KotCzarny> not just specific models
<libv> buZz: i did however go and buy a new N9 2 years ago
<libv> just to own a pristine one
<buZz> :D
<buZz> KotCzarny: i think even x86 now?
<libv> my nokia "test" n9 was a reliable companion for way too long
<KotCzarny> target was as i've stated
<buZz> :)
<buZz> i stopped caring about 'phones' or any GSM carrying hw
<libv> and what maemo did great was a terminal app with a real keyboard (with arrows) in a transparent overlay
<buZz> but i have one n900 somewhere with broken cell modem
<buZz> and i do love PDAs
<libv> i used to use irssi and mutt from my n9 and i actually was relatively happy
<KotCzarny> i've even convinced fmg to get A33 based tablet to have something to work on
<libv> KotCzarny: :)
<KotCzarny> and i think wizzup made some lime-a20 images too
JohnDoe_71Rus has quit [Quit: KVIrc 5.0.0 Aria]
<libv> i actually have the flash adapter for the n9 and a caged up n950 on a shelf here
Gerwin_J has quit [Quit: Gerwin_J]
<buZz> :D
<libv> i am not sure whether i still have the info on how to use those, or if that only ever lived on the nokia internal wiki
DonkeyHotei has joined #linux-sunxi
<libv> as nokia was dying i was busy poking at lima, so it was not on my radar to retain that info
<buZz> just asked another former nokia employee, he has no backups of that to give
<libv> yeah, again, not the best way to spend time
<libv> while commendable, Wizzup his skills are imho better used elsewhere, but i very much understand the attraction of swimming against the stream ;)
<buZz> ah well, asking is free ;)
<buZz> i think wizzup just loves this task as-is
<buZz> and the hackerspace he is with still has a lot of n900 users :P
<KotCzarny> having modern open sourced linux mobile device is handy
<KotCzarny> even if hw specs are obsolete
<buZz> imho, n900 is pretty decent, specwise
<buZz> if you dont have to run a webbrowser
<buZz> if you dont have to run a modern webbrowser
<buZz> lol
<buZz> NCSA Mosaic probably -flies- on it
<KotCzarny> depends greatly on content
<KotCzarny> current web is broken (big images, lots of js)
<libv> wens: i have spent a few days ratholing on one of my petpeeves
<libv> i have replaced the register tables of ov5640 with code, and spent way too much time staring at the awkward datasheet
<libv> i'll post the results of just the churn in a few mins
kaspter has quit [Quit: kaspter]
veremitz has quit [Quit: ZNC -]
veremitz has joined #linux-sunxi
<libv> pawel, hrm, old synapses do not return his last name, was thrown out from suse at the same time 23 people were thrown out here in nue, has been at denx for a few years now
<libv> he is also big on keeping the n900 running with up to date kernel/uboot
<libv> meh, getting old.
<fALSO> nice n900
<libv> machek, of course
* Wizzup missed the conversation somehow, reading backlog
<libv> :)
<KotCzarny> macek?
<wens> layoffs?
<libv> way back when
<Wizzup> KotCzarny: probably
<libv> february 2009, at the height of the financial crisis panic
<Wizzup> well, it's fun, and the pine64 guys are making a phone that might ship with maemo
<Wizzup> (amongst other distros)
<libv> Wizzup: if you're having fun, then that's all that counts
<Wizzup> yeah, it's fun and a learning experience, and very much swimming against the stream :)
<Wizzup> there was a lt talk at fosdem about the project though, not sure if you saw it :)
<libv> Wizzup: link?
selfbg has quit [Remote host closed the connection]
<libv> ah, you talked
<libv> Wizzup: next year, place this talk at the hwe devroom
<libv> there you get at least half an hour
<Wizzup> I tried to get a larger talk in another room (think distro devroom or so) but it didn't get accepted :)
<libv> yeah, penta needs work so that a talk can be targetted at multiple devrooms
<libv> "penta needs work" is quite the understatement :)
<Wizzup> hehe
<libv> config PVR_DEBUG_PDUMP...
<libv> are you guys carrying my debugfs changes?
<libv> edm trace :))))
<Wizzup> I think this is the current tree for the n900
<Wizzup> couple pages down are various commits made by you
<Wizzup> there's also but we haven't done much with it in the last few months (this is not about foss driver, just unified kernel effort for closed driver)
<Wizzup> I think I see the pdumpfs stuff
<libv> nice
<libv> that just made my day :)))
<Wizzup> - v 5.0 or v 5.1 with pvr
<libv> this works on the sgx of n900?
<Wizzup> yes
<libv> where is the binary from? n9?
<Wizzup> I think this is from nemomobile for n900
<libv> we were shipping with full pdump enabled iirc, so we could collect the crashes and have them automatically hit our bugzilla
<Wizzup> so that's meego era I think?
<libv> one of the warstories i heard when i joined, was the battle to get img to actually work to fix their constant hwrecoveryresets
<libv> we did get them down to almost zero in the end
<libv> but no other player but nokia (at the time) would've been able to convince img that this was a fixable issue worth fixing
<Wizzup> did it cause lag spikes?
<libv> no idea, i just know that my and imres pvr commits hit the nemomobile tree back in the early 2010s
<libv> not sure, these things should just should not happen, despite claims of img that these were normal
<Wizzup> hehe
<KotCzarny> quality code, it seems
<Wizzup> btw, maemo leste doesn't ship with any closed userland other than pvr driver, all maemo related components are foss (some are REd)
<Wizzup> we're about to take a big leap forward with ofono support
<libv> it's no shame with pvr hw to ship the binary, you have been given no other option
<libv> and realistically, any attempt to RE the pvr would've run aground pretty quickly
<Wizzup> my thoughts exactly, next will be to make the droid4 one work with drm, that's probably a bit of work...
<paulk-leonov> heh the best thing to do about the powervr is pretend it doesn't exist :)
<paulk-leonov> which is close to possible on beefy platforms
<Wizzup> that'd be true, were it not that the main (non allwinner) phones that are well supported in mainline are omap + pvr
<mru> those are obsolete
<paulk-leonov> Wizzup, well, OMAP has a fast DSP, could be used for 2D acceleration
<paulk-leonov> mru, who cares?
<mru> anyone who wants a usable device
<paulk-leonov> they still work and are available in large quantities
<paulk-leonov> yeah, because of sucky userpace
<Wizzup> droid4 has 1GB of ram, it's a decent upgrade from n900 ;)
<paulk-leonov> not because of limited hardware
<paulk-leonov> userspace*
<paulk-leonov> I mean sure, if you design a system with the idea that showing an image must require a GPU, then you set a pretty-high entry bar
<Wizzup> (maemo does this ;))
<libv> well...
<paulk-leonov> it's unfair to say the devices are shit because of that
<libv> we did use the pvr for tiny blits at one point
<mru> my employer has a few thousand am3517 based devices in the field
<mru> they're becoming a PITA
<paulk-leonov> the devices have lots of abilities that are well under-used
<paulk-leonov> mru, I don't see any issue with a GL-less system and proper mainline support
<mru> they're just underpowered for what we want to do
<paulk-leonov> and we have enough of this crazy "throw away your device every 6 month" nonsense
<mru> and graphics has nothing to do with it
foxx_ has quit [Ping timeout: 246 seconds]
<paulk-leonov> mru, oh could be, but there are lots of cases where they can be very useful
<paulk-leonov> for phones, they definitely can do
<paulk-leonov> I mean the "it's outdated" argument alone is anti-ecological to a severe extent...
<mru> the am33xx devices are performing better
<paulk-leonov> the point of processors is that they can be reprogrammes instead of throwing away the hardware, so why are we doing it the other way round?
<paulk-leonov> reprogrammed*
Gerwin_J has joined #linux-sunxi
<mru> only within their capabilities
sunshavi has quit [Ping timeout: 245 seconds]
itdnhr_ has quit [Ping timeout: 245 seconds]
<paulk-leonov> yes, but that assumes that we are correctly exploiting their capabilities in the first place
<paulk-leonov> which is not really very much the case
<paulk-leonov> at least on multimedia aspects
<mru> yes, that's always bugged me
<mru> but still, a 600 MHz cpu and 256 MB RAM only goes so far
<paulk-leonov> :)
* paulk-leonov remembers using the Openmoko Neo Freerunner as a daily phone
<libv> part of why sunxi exists, and why the wiki is so important, is to avoid the throwaway culture
<libv> when a15 and then a7 happened people jumped on arm cpus to implement virt
<libv> then at best made a blog entry
<libv> and then moved on
<mru> we're using some allwinner chips as well
<libv> leaving no mid to long term useful information behind
<mru> A20 and H3
<fALSO> i wish i could find an working guide
<fALSO> on getting opengl es working
<mru> fALSO: on what chip?
<fALSO> H3 or H5
<fALSO> or H6
<fALSO> i have all of them
<mru> sorry, can't help
<mru> our H3 devices are headless
<fALSO> ;-)
<mru> I did get it "working" on A20
<Wizzup> mru: the droid4 is four times are powerful as the n900 though
<mru> except there's apparently no libgbm
<Wizzup> but yeah, pinephone might be fun.
<mru> so anything that needs that is broken
<fALSO> i wanted to run DOOM
<fALSO> doom 2
<fALSO> ;-)
<fALSO> gzdoom or zandronum
<fALSO> i managed to build it
<fALSO> then i got like 0.0000000000000000001 FPS
<fALSO> :-P
<mru> maybe I can sneak doom onto our devices
<fALSO> lololololol
<fALSO> dont get in troubles because of me
<mru> someone might react to the sudden jump in software update size though
<fALSO> LOL mru so much monitorized ?
<fALSO> even the update size
<mru> not closely
<mru> but I suspect someone might notice if it doubled
cnxsoft has quit [Quit: cnxsoft]
<fALSO> damn... i would never notice that
<mru> then again, maybe it wouldn't
<fALSO> =)
<mru> doom isn't all that huge
<fALSO> nah.
foxx_ has joined #linux-sunxi
<diego71> fALSO: doom work fine on 486, there should be no need of opengl...
<fALSO> well
<fALSO> 2d on on a orange pi
<fALSO> is very very slw
<fALSO> is very very slow
<fALSO> no aceleration
<fALSO> im not sure if you can compare
<fALSO> probably an old PCI S3 Virge was better
<fALSO> ;-P
<diego71> I run doom on omap3 600Mhz and it works fine
<fALSO> which port of doom ?
<mru> doom ran fine on a 486
<libv> was that the port that oliver mcfadden made, or did he make a port of quake?
<diego71> I think is zdoom + freedoom
<fALSO> i tried gzdoom in software mode, it was impossible to use
<KotCzarny> doom was 2d fps
<KotCzarny> well, 2.5D
<mru> 3d graphics, 2d map
<mru> or 2.5D map
<mru> wolfenstein was pure 2d map
JohnDoe_71Rus has joined #linux-sunxi
<fALSO> well, i know
<fALSO> but 2D on a orangepi at least
<fALSO> is very slow
<fALSO> just try to move an windows in xorg
<fALSO> you can see it being redrawn :-P
<diego71> fALSO: correction: prboom + freedoom
<fALSO> do you have the sources ?
Mandragor_ has quit [Ping timeout: 252 seconds]
<megi> fALSO: it's not that bad, depends on the xf86 driver
<DuClare> Also switch to 8-bit gfx if you're stuck with no dma
Mandragor has joined #linux-sunxi
<megi> I believe I also got a faster memcpy using NEON instrinsics
<megi> or some such, it's quite some time
<megi> load as much as I can into registers in one go and then store it back to mem
<fALSO> gotta try it again
<KotCzarny> 16:33 #linux-sunxi mru > but still, a 600 MHz cpu and 256 MB RAM only goes so far
<KotCzarny> lol
<fALSO> if im bothering you guys, or talking something off topic
<fALSO> tell me, that ill shut up
<KotCzarny> 600mhz cpu and 256m of ram can do a LOT
<buZz> i think its fine :P
<fALSO> man
<fALSO> i RAN doom on my ipod classic
tllim has quit [Quit: Leaving]
<buZz> i ran doom on a 386
<fALSO> its much much much worse than any orange pi
<buZz> worked
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 -]
<mru> KotCzarny: not as much as a 1 GHz quad-core and 1 GB RAM
<KotCzarny> sure, but a lot more than 120mhz, 16mb system
MoeIcenowy has joined #linux-sunxi
<fALSO> it was with some alternative firmware
<fALSO> rockbox
<buZz> :)
<fALSO> it is also prdoom
<fALSO> nah prBoom
<KotCzarny> which is where doom/quake originally ran
<fALSO> yap but doom for pc was probably filled with inline assembly
<KotCzarny> but they were programmed to squeeze every bit of cpu time out of the system
<fALSO> yap
<buZz> wasnt doom or quake written in C
<buZz> on a NeXT station
<mru> mostly C, bits of assembly for critical stuff
<buZz> one of em was , anyway
<buZz> carmack doesnt code ASM
<buZz> mcgee does
<buZz> who was only pulled in later
<mru> anyway, the customers don't care about running doom
<fALSO> yes
<fALSO> =)
<fALSO> only useless guys like falso
MoeIcenowy has quit [Client Quit]
MoeIcenowy has joined #linux-sunxi
<buZz> :)
<KotCzarny> unless you start selling diy consoles and your customers play retro
<buZz> its pretty nuts how alive that 'retro' stuff is nowadays
<buZz> just how many GB emulators does one person need
MoeIcenowy has quit [Client Quit]
MoeIcenowy has joined #linux-sunxi
popolon has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
MoeIcenowy has quit [Client Quit]
MoeIcenowy has joined #linux-sunxi
<plaes> jbizcocho: just make sure it runs with mainline kernel
<libv> Wizzup: i will bring up the fact that a handheld microphone is a bit noisy, 15min talks could have body mics
altious2 has joined #linux-sunxi
altious2 has quit [Remote host closed the connection]
Nyuutwo_ has joined #linux-sunxi
<jbizcocho> Hi plaes, I took mainline 4.14
Nyuutwo has quit [Ping timeout: 246 seconds]
aalm has quit [Ping timeout: 255 seconds]
aalm has joined #linux-sunxi
Putti has joined #linux-sunxi
victhor has quit [Ping timeout: 258 seconds]
vagrantc has joined #linux-sunxi
msimpson has quit [Ping timeout: 255 seconds]
clemens3 has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
jerbob92 has joined #linux-sunxi
<Wizzup> libv: *nod*
BenG83 has joined #linux-sunxi
reinforce has joined #linux-sunxi
BenG83 has quit [Remote host closed the connection]
BenG83 has joined #linux-sunxi
gnufan_home has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
netlynx has joined #linux-sunxi
Mandragor has quit [Ping timeout: 246 seconds]
Mandragor has joined #linux-sunxi
DonkeyHotei has quit [Ping timeout: 244 seconds]
clemens3 has quit [Quit: WeeChat 1.6]
sunshavi has joined #linux-sunxi
tkaiser has quit [Remote host closed the connection]
dev1990 has joined #linux-sunxi
jaganteki has joined #linux-sunxi
AneoX has quit [Ping timeout: 245 seconds]
AneoX has joined #linux-sunxi
merlin1991 has quit [Remote host closed the connection]
BenG83 has quit [Remote host closed the connection]
merlin1991 has joined #linux-sunxi
tnovotny has quit [Quit: Leaving]
lurchi_ is now known as lurchi__
<fALSO> Linux orangepipc 5.1.0-10240-g63863ee8e2f6 #27 SMP Tue May 14 12:02:12 WEST 2019 armv7l GNU/Linux
<fALSO> ;-D
<jernej> KotCzarny: I found a fix for a bug you mentioned some time ago, where 4K resolution doesn't work on H3, H5 and A64 if U-Boot video driver is disabled
<jernej> and yes, is clock issue
<KotCzarny> yay!
<KotCzarny> was it in kernel or uboot?
<jernej> I'll send fix probably today or tomorrow
<jernej> kernel
<jernej> hdmi phy clock driver read register before hdmi reset was released and naturally, it read 0
<jernej> but when reset was released, clock divider suddenly become 2
<KotCzarny> sucky kind of bugs to debug
BenG83 has joined #linux-sunxi
<jernej> nah, I learnt hard way to always check set clock rate dynamics
BenG83 has quit [Client Quit]
<[TheBug]> nice
jaganteki has quit [Ping timeout: 256 seconds]
afaerber has quit [Quit: Leaving]
foxx_ has quit [Ping timeout: 258 seconds]
nuuuciano_ has quit [Ping timeout: 246 seconds]
netlynx has quit [Quit: Ex-Chat]
nuuuciano has joined #linux-sunxi
AneoX has quit [Ping timeout: 268 seconds]
AneoX_ has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: KVIrc 5.0.0 Aria]
Putti has quit [Ping timeout: 245 seconds]
victhor has joined #linux-sunxi
altious has quit [Quit: Leaving]
altious has joined #linux-sunxi
shfil has joined #linux-sunxi
cmeerw has quit [Read error: Connection reset by peer]
tkaiser has joined #linux-sunxi
wwilly_ has quit [Ping timeout: 246 seconds]
wwilly_ has joined #linux-sunxi
altious2 has joined #linux-sunxi
altious has quit [Ping timeout: 244 seconds]
altious2 has quit [Client Quit]
Mangy_Dog has quit [Ping timeout: 246 seconds]
gnufan_home has quit [Quit: Leaving.]
gnufan_home has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
gnufan_home has quit [Quit: Leaving.]
reinforce has quit [Quit: Leaving.]
Mangy_Dog has quit [Remote host closed the connection]
lurchi__ is now known as lurchi_
lurchi_ is now known as lurchi__
AneoX has joined #linux-sunxi
lurchi__ has quit [Ping timeout: 244 seconds]
AneoX_ has quit [Ping timeout: 246 seconds]
lurchi__ has joined #linux-sunxi
pmpp_ has quit [Remote host closed the connection]
pmpp has joined #linux-sunxi