rellla changed the topic of #linux-sunxi to: Allwinner/sunxi /development discussion - did you try looking at our wiki? https://linux-sunxi.org - Don't ask to ask. Just ask and wait! - https://github.com/linux-sunxi/ - Logs at http://irclog.whitequark.org/linux-sunxi - *only registered users can talk*
<libv> hrm, code to adjust dram priority of cpu, csi0 and csi1
<libv> http://linux-sunxi.org/A10_DRAM_Controller_Register_Guide has SDR_HPCR "host port configuration register"
<libv> and 0x290 seems to be cpu, 0x2a0 csi0, 0x2bc csi1
_whitelogger has joined #linux-sunxi
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
Da_Coynul has joined #linux-sunxi
return0e has quit [Ping timeout: 246 seconds]
jo0nas has quit [Ping timeout: 248 seconds]
jo0nas has joined #linux-sunxi
tllim has quit [Read error: Connection reset by peer]
return0e has joined #linux-sunxi
craigo has quit [Ping timeout: 246 seconds]
embed-3d has quit [Ping timeout: 258 seconds]
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
s_frit has quit [Remote host closed the connection]
s_frit has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
vagrantc has joined #linux-sunxi
nexgen has quit [Quit: Leaving]
Gerwin_J has joined #linux-sunxi
vagrantc has quit [Quit: leaving]
<wens> IIRC, for A13/A20 the priority stuff is in the MBUS controller?
<wens> anarsoul: can we make it not do polling?
<anarsoul> wens: I think it's thermal zone stuff that does polling, thermal driver uses interrupts
<anarsoul> but thermal does it at unreasonable rate
<anarsoul> *thermal zone
<anarsoul> I'm testing it with lower modclk atm
<wens> I can't recall if it stores the current reading (from the most recent interrupt) or not
nexgen has joined #linux-sunxi
<anarsoul> thermal driver does not
<anarsoul> and I don't think that it's supposed to do that, it's job of thermal zone framework
<anarsoul> looks like pmu on A64 isn't working :\
<anarsoul> no interrupts from arm-pmu and perf reports nothing
<anarsoul> wens: have you ever seen pmu working on A64?
<anarsoul> and looks like jmcneill complained about the same issue ~year ago
TheSeven has quit [Ping timeout: 264 seconds]
[7] has joined #linux-sunxi
lurchi_ is now known as lurchi__
NeuroScr has quit [Quit: NeuroScr]
dddddd has quit [Remote host closed the connection]
NeuroScr has joined #linux-sunxi
<wens> anarsoul: no, don't use pmu
<wens> I *
florian has quit [Ping timeout: 245 seconds]
florian has joined #linux-sunxi
lurchi__ is now known as lurchi_
lurchi__ has joined #linux-sunxi
lurchi_ has quit [Ping timeout: 245 seconds]
<wens> anarsoul: maybe you could ask the person that added it?
NeuroScr has quit [Quit: NeuroScr]
aalm has quit [Quit: xyz 2.3]
Gerwin_J has quit [Ping timeout: 264 seconds]
<anarsoul> wens: I disabled pmu node and now perf works fine
<wens> but then it doesn't use the pmu, right?
<anarsoul> right
<anarsoul> PMU on A64 generates no interrupts
<anarsoul> (or interrupt numbers are wrong - but they're the same as in BSP kernel)
<KotCzarny> i wonder if they work on the r_* side
<KotCzarny> ie. with arisc
<anarsoul> KotCzarny: it's ARM-specific part
<anarsoul> it has no use on arisc side
<KotCzarny> but pmu isnt arm specific?
<anarsoul> performance monitoring unit
<anarsoul> not power management unit
<KotCzarny> ahm
<anarsoul> so it is ARM specific
<KotCzarny> maybe they needed irq line for something else and ran out of space ;)
<anarsoul> no idea
<anarsoul> anyway, we should disable it in dts, no pmu is better than broken pmu. I'll send the patch tomorrow
yann has quit [Ping timeout: 258 seconds]
<wens> anarsoul: please CC the original author, who claimed it worked on teres-I
jo0nas has quit [Ping timeout: 244 seconds]
jo0nas has joined #linux-sunxi
nexgen2 has joined #linux-sunxi
swiftgeek has quit [Ping timeout: 245 seconds]
nexgen has quit [Read error: Connection reset by peer]
[7] has quit [Ping timeout: 264 seconds]
JohnDoe_71Rus has joined #linux-sunxi
<KotCzarny> amd graphics back to arm ecosystem?
TheSeven has joined #linux-sunxi
<KotCzarny> those smartphone fans dont seem that crazy now ;)
airwind has joined #linux-sunxi
jo0nas has quit [Ping timeout: 272 seconds]
jo0nas has joined #linux-sunxi
<wens> IIRC Samsung only uses their own chips in mid or low end phones, or high end ones outside of the US
NeuroScr has joined #linux-sunxi
yann has joined #linux-sunxi
smaeul has quit [Ping timeout: 276 seconds]
<buZz> lol yes plz
<buZz> put a AMD gpu in my overheating andrerp phone
<buZz> so i can overheat while i overheat
<buZz> -awesome-
<KotCzarny> you need a fan in your life
<buZz> lol yeah
<KotCzarny> the fan that will always be with you
<buZz> andrerp phones with fans like 19" dell poweredge's
<KotCzarny> that will never let you down
<buZz> :)
<buZz> PFFFFFF 'i cant hear you over my fan, 1 moment'
<JohnDoe_71Rus> buZz: this model phone for Syberia :)
<KotCzarny> they have gsm/net in syberia?
megi has quit [Ping timeout: 246 seconds]
<JohnDoe_71Rus> https://tinyurl.com/y4tmf8rg for me work only G-cache site
qschulz has quit [Ping timeout: 246 seconds]
qschulz has joined #linux-sunxi
florian has quit [Ping timeout: 248 seconds]
AneoX has joined #linux-sunxi
swiftgeek has joined #linux-sunxi
megi has joined #linux-sunxi
<MoeIcenowy> wens: I think Samsung use Exynos in high-end phones in CDMA-less regions
ramtheconqueror has joined #linux-sunxi
ramtheconqueror has quit [Remote host closed the connection]
s_frit has quit [Remote host closed the connection]
ramtheconqueror has joined #linux-sunxi
s_frit has joined #linux-sunxi
suprothunderbolt has quit [Ping timeout: 268 seconds]
<wens> MoeIcenowy: oh, so that's the distinction
<wens> MoeIcenowy: people here are always cursing Samsung for using "inferior chipsets" lol
<MoeIcenowy> wens: Taiwan has CDMA carrier?
<MoeIcenowy> I just know the China ver is Qualcomm because China Telecom uses CDMA
<MoeIcenowy> (but at least Galaxy Note 5 choose to use Exynos even at China -- the corresponding Qualcomm 8-series product at that time is 810
<wens> MoeIcenowy: we have one (that is sort of fading away), but Samsung phones don't support it :)
<wens> hence all Samsung phones in Taiwan are Exynos based AFAIK
<wens> and people cry foul
<MoeIcenowy> oh why cry
<wens> because the price is the same when compared to the US, but you get Exynos instead of SnapDragon
<wens> though it's not as bad as the TSMC vs Samsung bit for iPhones :p
<wens> that was worse. IIRC it had a noticeable effect on performance and battery life
<wens> a friend, then working at TSMC, pre-ordered iPhones through the company, and got ones with Samsung chips
reinforce has joined #linux-sunxi
ramtheconqueror has quit [Remote host closed the connection]
florian has joined #linux-sunxi
<florian> hi all
<KotCzarny> regarding qualcomm vs exynos: https://blade.tencent.com/en/advisories/qualpwn/
<KotCzarny> ;)
<KotCzarny> We didn't test all the Qualcomm chips. We only tested on the Google Pixel2/Pixel3. Results of our tests indicate that unpatched phones running on Qualcomm Snapdragon 835,845 may be vulnerable.
<wens> *sigh*
diego_r has joined #linux-sunxi
rpirea has joined #linux-sunxi
rpirea has quit [Quit: rpirea]
rpirea has joined #linux-sunxi
<libv> wens: was that samsung vs qualcom chip thing not primarily about the modem?
<libv> at least, that was how it started i think
Andy-D has joined #linux-sunxi
<willmore> KotCzarny, hasn't there been AMD in ARM phone SoCs for years? Maybe I forgot how Ardreno came about....
<KotCzarny> willmore, but this might be a refresher
<libv> willmore: yes
<libv> when the credit crunch happened, amd needed cash bad
<willmore> Yep. :(
<willmore> Penny wise, pound foolis
<willmore> h
<libv> they shut down the suse partnership over radeonhd
<libv> which was not going well with amd ati internal struggles anyway
<willmore> "mistakes were made"
<libv> they sold off their fabs to the saudis
<willmore> I remember.
<libv> now globalfoundries
<wens> libv: I don't think they ever officially explained it , or they did and it fell on deaf ears
<libv> and they sold off their mobile bits, which was then shipping in... marvell and... something else
<willmore> MoeIcenowy, I didn't know China Telecom uses CDMA! I didn't think that junk was used outside of the Americas.
<libv> z340 comes to mind
<libv> for like 60m
<wens> willmore: China also has some CDMA variant not seen elsewhere
<willmore> wens, I guess that's not surprising. China tends to do some odd stuff, too. Can't let the US do all the stupid things.
<willmore> KotCzarny, Ardreno2, GPU buggaloo?
<libv> the credit crunch happened like more than a year after the whole of AMD (including that steaming pile of bankrupt shit, ATI) was worth less than what they paid for ATI 1.5ys earlier
<libv> and before the credit crunch the stockmarket was crazy
<libv> so amd was in a real bind
<libv> and their cpu plans were not great either, they were too busy trying to fix up ati i think
<libv> (and ended up being eaten alive from within)
<libv> we had partner meetings where bulldozer and such were pre-pre-announced
<libv> i was too busy with other things to notice how perhaps their ideas depended on ati too much
<libv> but for the radeonhd project it was clear how the revenue streams of cpus versus discrete graphics dominated politics
<libv> and by mid 2008, it was bad.
<libv> and then the credit crunch happened
<wens> I would've assumed discrete graphics made more money
<libv> well, amd cpu products were no longer competitive
<libv> and they were quickly losing the server market then
<libv> and linux == server market
<MoeIcenowy> wens: do you mean TD-SCDMA
<libv> this is why we got to free ati drivers and why we got to do radeonhd
<libv> amd bought ati, ati and fglrx were seriously bad marketing for their server business
<libv> so amd managers looked for a solution, and some amd managers were suse employees when SuSE enabled the amd hammer (as no-one else would)
<libv> and amd still spends a lot of money on the suse toolchain guys under now 16-18y old partner agreements
<libv> and as luck had it, the greatest display developer ever (;p) just had his job interview at suse when one of those managers called
<libv> so we took on that challenge and wrote a working driver
<libv> first massively hailed by amd management
<libv> but as time went by, ati got the upper hand, as amd was getting more and more in trouble
<libv> and then... total collapse
<wens> MoeIcenowy: yes
<libv> heh, looking at the dates of when bulldozer was shipped, that thing was massively delayed
<libv> amd was talking it up in early 2008 already
<libv> and it only was announced in Q3 2011, and shipped in Q1 2012
<KotCzarny> amd survived tho
<KotCzarny> and is now kicking around
<KotCzarny> maybe someone sacrificied their firstborn
<KotCzarny> *sacrificed
<libv> KotCzarny: it took them a decade
<libv> and they had to go back to their roots
<libv> i think it needed a serious downswing in ati relevance compared to nvidia first
reinforce has quit [Quit: Leaving.]
Da_Coynul has joined #linux-sunxi
<libv> for the political will to internally develop to finally go back to where they came from before they bought ati
<libv> and to let ati not influence the greater company
cnxsoft has quit [Quit: cnxsoft]
<KotCzarny> lets hope they dont run out of the firstborns, competition is nice
<libv> well, you saw what intel has been doing since sandy bridge
<libv> all we got was minor ipc and price increases since then
<KotCzarny> they got better with igpu
<libv> and peripherals, so usb3 and beyond, nvme
<wens> still waiting for ice lake to check out igpu improvements
<libv> but the actual cpu space was without competition and there was therefore no drive to push things forward
<libv> and the much touted arm servers... well
<libv> intel ran away screaming from the mobile market
<libv> (and killed off wayland with it)
<libv> and we still have not seen much in the way of big iron with arm cores
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
return0e has quit [Remote host closed the connection]
return0e has joined #linux-sunxi
jstefanop has joined #linux-sunxi
return0e has quit [Remote host closed the connection]
return0e has joined #linux-sunxi
return0e has quit [Remote host closed the connection]
Da_Coynul has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
_whitelogger has joined #linux-sunxi
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
Andy-D has quit [Ping timeout: 252 seconds]
craigo has joined #linux-sunxi
dddddd has joined #linux-sunxi
chewitt has quit [Quit: Adios!]
Andy-D has joined #linux-sunxi
<libv> meh, DEFE kills csi1 memory bandwidth
<libv> and of course, our driver does not properly disable it when we're done
<libv> i am going to have to use yuv444 on debe it seems and use the csc with 1,1,1
Andy-D has quit [Ping timeout: 252 seconds]
<libv> wens: about dram controller priority, i added some port info that i could find
<libv> do you have, in the back of your mind, some recollection of whether more info on that was ever known, or if someone went and figured out what value was for what subsystem?
jstefanop has quit [Remote host closed the connection]
<wens> libv: I only remember someone got the info for mbus, but not a10 though
<libv> ok
<libv> yeah, this seems to have been 2014
<libv> linux-sunxi, since august 2012 btw. 7ys
<wens> how come? # killed off wayland
<mru> just the usual libv hyperbole
<libv> wens: well, intel stopped paying their own developers to work on wayland in late 2013 early 2014
<mru> intel killed _their_ support for it
<libv> krh suddenly started working on mesa
<mru> not the entire project
<libv> then moved on to google
<libv> and one former nokia teammate first was moved to improving chromeos performance, then a lot of people in .br were made redundant
<libv> i do not know the backstory
<libv> but it was after chris wilson was forced to move mir patches out of the X driver for intel
<libv> and around the time when it became clear to intel that intel mobile was not going to fly
<libv> (i'd like to think that my hardhitting blog entry caused all that though ;p)
<mru> we know you think highly of yourself
<wens> mru: was your musb patch ever picked up?
<mru> which patch?
<mru> the dt one?
<mru> don't think so
<wens> hmm
<mru> nope, not in linux-next
<wens> probably needs a resend
<mru> _another_ resend
<wens> :/
<mru> done
gaston_ has joined #linux-sunxi
jstefanop has joined #linux-sunxi
jstefanop has quit [Ping timeout: 245 seconds]
nexgen2 has quit [Quit: Leaving]
afaerber has quit [Quit: Leaving]
kaspter has quit [Ping timeout: 245 seconds]
yann has quit [Ping timeout: 272 seconds]
<fALSO> megi, youre MEGOUS ?
<fALSO> megi, i've been trying to clone your linux kernel repo
<fALSO> megi, but after a few hours it ends saying that the repo is corrupt
<fALSO> the only way i can clone it is if i make a shallow copy
<mru> hours? slow connection?
<fALSO> but after that... if i git pull to update it
<fALSO> it tries to download everything again, and errors out
<fALSO> mru,
<fALSO> remote: Enumerating objects: 3887430
<fALSO> hours on this...
<mru> slow remote then
<KotCzarny> for a quickie try --depth=1
<KotCzarny> if you only need quick checkout
<fALSO> yes, that what i did, but with that, you cant update it
<mru> or add the remote to an existing clone
<mru> then you only need to get the difference
tuxillo has quit [Ping timeout: 246 seconds]
<KotCzarny> yup, but at least you can check what works etc
<fALSO> i cant clone it complety
<fALSO> im gonna wait until it fails so that i can show you guys the error ;-P
<fALSO> i had to reboot earlier so i dont have it
tl_lim has joined #linux-sunxi
tllim has joined #linux-sunxi
<megi> fALSO: clone it from github
<megi> fALSO: I'll keep that bug as a punishment for people trying to deplete my bandwidth quota on my VPS
<megi> :D
tllim has quit [Client Quit]
<megi> your download ended, sorry
<KotCzarny> :)
<libv> fALSO: i keep a linus tree somewhere and update that regularly
<libv> i clone from there
<libv> then remove the origin, and add the target origin
<libv> so only the relevant changes need to be retrieved over the net
<megi> yes please, that I'd recommend too, clone linus's tree and 'git add remote someone someurl ; git fetch --all'
<megi> s/add remote/remote add/
vagrantc has joined #linux-sunxi
airwind has quit [Quit: airwind]
reinforce has joined #linux-sunxi
kaspter has joined #linux-sunxi
<fALSO> megi, going to try it out
<fALSO> megi, do you have a mirror of your repo in github?
<megi> yes
<fALSO> damn
<fALSO> sorry for wasting your bw then
<megi> fALSO: no problem, I was mostly kidding ;)
<fALSO> megi, ahhh
<fALSO> got the error now too
<fALSO> fishy ~/opipc/linux-megous # git pull
<fALSO> fatal: protocol error: bad pack header
<fALSO> fatal: the remote end hung up unexpectedly
<megi> but the vps has low ram, and it's what is causing the slowdowns
<fALSO> gonna set the origin to github
<fALSO> thanks for the help
<fALSO> its working :)
jstefanop has joined #linux-sunxi
jstefanop has quit [Ping timeout: 248 seconds]
jonkerj has quit [Quit: brb]
jonkerj has joined #linux-sunxi
s_frit has quit [Remote host closed the connection]
s_frit has joined #linux-sunxi
tuxillo has joined #linux-sunxi
return0e has joined #linux-sunxi
s_frit has quit [Remote host closed the connection]
s_frit has joined #linux-sunxi
SopaXorzTaker has joined #linux-sunxi
tl_lim has quit [Ping timeout: 245 seconds]
netlynx has joined #linux-sunxi
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
nexgen has joined #linux-sunxi
kaspter has quit [Remote host closed the connection]
kaspter has joined #linux-sunxi
aballier has quit [Read error: Connection reset by peer]
aballier has joined #linux-sunxi
AneoX has quit [Quit: Textual IRC Client: www.textualapp.com]
gaston_ has quit [Ping timeout: 244 seconds]
rpirea has quit [Quit: rpirea]
lurchi__ has quit [Quit: Konversation terminated!]
lurchi__ has joined #linux-sunxi
aalm has joined #linux-sunxi
diego_r has quit [Ping timeout: 244 seconds]
gaston_ has joined #linux-sunxi
florian has quit [Quit: Leaving]
The_Loko has joined #linux-sunxi
tl_lim has joined #linux-sunxi
florian has joined #linux-sunxi
<ElBarto> anarsoul: I've just seen your patch for removing the pmu node on a64, are you a bit familiar with pmu stuff on arm64 ?
<anarsoul> ElBarto: not really, why?
<ElBarto> I don't understand how perf can show the cycle count as it's not defined in the json file (for some reason that I still don't understand)
<ElBarto> I have our pmu related stuff fixed on arm64 for FreeBSD but I'm trying to mimic how it's done on amd64
<ElBarto> and we have pmc stat which is basically the same as perf stat in the Linux world
<ElBarto> Ampere Emag seems to define CPU_CYCLES but it's not present in the cortex files
<anarsoul> ElBarto: sorry, I'm not familiar neither with hardware or with driver
<ElBarto> ok, thanks :)
<ElBarto> and btw I also didn't managed to make it worked on A64
<anarsoul> anyway looks like it doesn't work on A64, NetBSD dev reported that it's broken for him ~year ago
<ElBarto> it costed me a few hours/days
Nakaori has quit [Ping timeout: 246 seconds]
<ElBarto> was it Jared who reported that ?
<anarsoul> I think so
<ElBarto> I haven't tested on H5 yet
<anarsoul> could you reply to the thread that it didn't work for you as well?
<ElBarto> yeah I wanted to dig a bit deeper before doing that tonight
<ElBarto> the interrupts number are definitely wrong btw :)
Putti has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
<ElBarto> with the correct interrupts I don't kernel panic anymore but I'm not receiving any events
<ElBarto> I don't have any interrupts fired too ... weird that I kernel panic'ed before
vagrantc has quit [Quit: leaving]
JohnDoe_71Rus has quit [Quit: KVIrc KVIrc Aria 5.0.0, revision: 5.0.0+git-7422-2fe1a3bca, build type: debug, sources date: 20160102, built on: 2019-07-01 08:27:19 UTC 5.0.0+git-7422-2fe1a3bca http://www.kvirc.net/]
parco has joined #linux-sunxi
netlynx has quit [Quit: Ex-Chat]
<parco> hello, how can I read temperature of allwinner R16/A33 chip?
<megi> oh my, you need a supercomputer for dtc_check
<megi> eh, dtbs_check
Putti has quit [Ping timeout: 272 seconds]
SopaXorzTaker has quit [Remote host closed the connection]
<Mangy_Dog> parco i cant remember the name of the tool but i think its included with the armbian buids if not its apt getable...
<Mangy_Dog> but check that youtuber
<montjoie> megi: specify the yaml you want to check for faster check
<Mangy_Dog> the austrailian guy
<Mangy_Dog> who does loads of SMC reviews
<Mangy_Dog> sbc
<parco> Mangy_Dog: i'm just trying to "cat /sys/class/thermal/thermal_zone0/temp" and getting Invalid Argument
<Mangy_Dog> thay i have no idea about :p
<parco> the only Australian youtuber i'm familiar with is David jones
<Mangy_Dog> yeah not him
<megi> montjoie: thanks for the suggestion
<Mangy_Dog> let me see if i can find him
<Mangy_Dog> i think hes in my sub list
<Mangy_Dog> him
florian has quit [Ping timeout: 246 seconds]
<Mangy_Dog> anyway look at a couple of review vids he uses a tool to read board temps under loads
<Mangy_Dog> at the least it might show the port
<Mangy_Dog> to read off
<parco> thanks!
florian has joined #linux-sunxi
<parco> ok thats fixed, i had to enable the SUN4I_GPADC in the linux kernel
<parco> now I can properly read from /sys/class/thermal/thermal_zone0/temp
<anarsoul> ElBarto: can you test it on H5 or H6?
ed_peguillan has quit [Read error: Connection reset by peer]
gaston_ has quit [Quit: Konversation terminated!]
<ElBarto> anarsoul: I don't have any H6 boards but I plan to test on H5 tomorow
The_Loko has quit [Remote host closed the connection]
vagrantc has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
sunilmohan has quit [Read error: Connection reset by peer]
sunilmohan has joined #linux-sunxi
Da_Coynul has joined #linux-sunxi
Mangy_Dog has quit [Ping timeout: 245 seconds]
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
Da_Coynul has joined #linux-sunxi
florian has quit [Ping timeout: 246 seconds]
suprothunderbolt has joined #linux-sunxi
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
tl_lim has quit [Read error: Connection reset by peer]