<MrElendig2>
This message was brought to you by Private Internet Access
MrElendig2 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<anarsoul|2>
can someone enable +r mode on this channel, so only users with registered nick can write to this channel?
<anarsoul|2>
not sure if it's +r here on freenode though
BenG83 has quit [Quit: Leaving]
mzki has quit [Ping timeout: 265 seconds]
nuuuciano has quit [Ping timeout: 260 seconds]
nuuuciano has joined #linux-sunxi
mzki has joined #linux-sunxi
victhor has joined #linux-sunxi
<MoeIcenowy>
megi: what SoC?
<MoeIcenowy>
parent may have changed
<megi>
MoeIcenowy: A83T. I figured I was looking at the wrong bits. :) Frequency was correctly set to 50MHz. BSP contains differnt clock phase/delays settings for A83T. But I didn't obsere any effects when changing that.
<megi>
I just don't know why A83T SDMMC is so slow compared to H3.
<megi>
On H3 I get 22.6MiB/s on A83T I get 6MiB/s max
<megi>
:(
<megi>
Measured by dd iflag=direct ... bs=256k
<MoeIcenowy>
H6 SDMMC is also slow
<MoeIcenowy>
although not measured yet
lurchi_ is now known as lurchi__
nuuuciano has quit [Ping timeout: 240 seconds]
nuuuciano has joined #linux-sunxi
popolon has quit [Quit: WeeChat 1.9.1]
Pietre has quit [Ping timeout: 256 seconds]
anarsoul|2 has quit [Ping timeout: 240 seconds]
Pietre has joined #linux-sunxi
<MoeIcenowy>
mripard: btw it seems that the BSP now shows the version of MMC IP code
<MoeIcenowy>
in the format of vXpYx
<MoeIcenowy>
e.g. A33 has v4p10x, R40/H5 has v4p5x, H6/A63 has v4p6x
<MoeIcenowy>
but... why is v4p5x > v4p10x?
vagrantc has joined #linux-sunxi
cnxsoft has joined #linux-sunxi
<wens>
maybe they rolled back? :p
Awesomecase has joined #linux-sunxi
<Awesomecase>
or maybe this blog by freenode staff member Matthew 'mst' Trout https://MattSTrout.com/
<Awesomecase>
Hey, I thought you guys might be interested in this blog by freenode staff member Bryan 'kloeri' Ostergaard https://bryanostergaard.com/
<Tycale21>
This message was brought to you by Private Internet Access
Tycale21 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<MoeIcenowy>
interestingly there's a SMCV register in BSP
<MoeIcenowy>
which means SD/MMC Controller Version
<MoeIcenowy>
0x300
MillerBoss17 has joined #linux-sunxi
<MillerBoss17>
Hey, I thought you guys might be interested in this blog by freenode staff member Bryan 'kloeri' Ostergaard https://bryanostergaard.com/
<MillerBoss17>
or maybe this blog by freenode staff member Matthew 'mst' Trout https://MattSTrout.com/
<Guest20498>
This message was brought to you by Private Internet Access
Guest20498 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<jernej>
MoeIcenowy: Does that bus-width help with SD card speeds? how much?
selfbg has joined #linux-sunxi
<MoeIcenowy>
jernej: 4x
<jernej>
what is then final speed?
<MoeIcenowy>
still not satisfying
<MoeIcenowy>
5.6MB/s
<jernej>
well, step by step :)
TheSeven has quit [Ping timeout: 260 seconds]
[7] has joined #linux-sunxi
<jernej>
that should be already much better
<jernej>
and usable
netlynx has joined #linux-sunxi
netlynx has quit [Changing host]
netlynx has joined #linux-sunxi
<MoeIcenowy>
jernej: I asked tllim to ship you a Pine H64 model B
<jernej>
thanks
fire219 has joined #linux-sunxi
<KotCzarny>
rellla: ping
<KotCzarny>
jernej: MoeIcenowy: that sdcard slowness seems like a common bug somewhere
jemk has quit [Ping timeout: 256 seconds]
diego_r has quit [Ping timeout: 248 seconds]
kozy has quit [Quit: No Ping reply in 180 seconds.]
kozy has joined #linux-sunxi
jemk has joined #linux-sunxi
iamfrankenstein has joined #linux-sunxi
lkcl has quit [Ping timeout: 244 seconds]
lkcl has joined #linux-sunxi
muvlon_ has quit [Ping timeout: 256 seconds]
anarsoul|2 has joined #linux-sunxi
leviathan has joined #linux-sunxi
muvlon_ has joined #linux-sunxi
montjoie has joined #linux-sunxi
anarsoul|2 has quit [Ping timeout: 256 seconds]
philippefouquet has joined #linux-sunxi
vagrantc has joined #linux-sunxi
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-sunxi
<tllim>
@jernej and @icenowy, I will do that. However, I needs to revise model B to make sure that same as model A. It is pain on release two board that need to runs different OS build
tl_lim has joined #linux-sunxi
tllim has quit [Ping timeout: 260 seconds]
reinforce has joined #linux-sunxi
fl_0 has joined #linux-sunxi
jaganteki has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 240 seconds]
KellerFuchs21 has joined #linux-sunxi
KellerFuchs21 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<KellerFuchs21>
or maybe this blog by freenode staff member Matthew 'mst' Trout https://MattSTrout.com/
<KellerFuchs21>
Hey, I thought you guys might be interested in this blog by freenode staff member Bryan 'kloeri' Ostergaard https://bryanostergaard.com/
philippefouquet has quit [Remote host closed the connection]
ecrist121 has joined #linux-sunxi
ecrist121 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<messpl>
Hello, I have a question related to serial (UART). I am playing with it right now using A33-based board. From device tree I see that uart uses default linux driver, along with DMA configuration. I wrote simple C application that uses termios to access UART. The problem is I can see latnecy when I transmit somthing using UART. I send 128 bytes using write command but using logic analyser i see that data is sent in groups 1-3 bytes with small delays
<messpl>
(~0.5 ms). Should default driver use DMA configured in the device tree? With DMA I would expect to see at least complete transmission of DMA length.
<rellla>
bshah: ok, so MoeIcenowy, mripard and wens can not op/deop themselves?
<KotCzarny>
nope
<KotCzarny>
they can deop
<KotCzarny>
but not op themselves via chanserv
<rellla>
who is possible to make them op'able? founders?
<KotCzarny>
yup
<rellla>
ok, so then i will deop them again and libv, hno or mnemoc should make them op's if wanted. although i haven't heard from mnemoc and hno for a while...
<KotCzarny>
who are they ?
<rellla>
KotCzarny: part of the most active people in the past and part of the "founders" of linux-sunxi at all...
<rellla>
just read the oooold logs :p
<KotCzarny>
must be really long time ago, never seen them here
<rellla>
yeah, it is. omg.
<wens>
IIRC, linux-sunxi used to be hosted on mnemoc's server
<rellla>
KotCzarny: you may recognise the automatic pushes to linux-sunxi/linux-sunxi from amery. it's mnemoc.
<libv>
wens: i am paying for that server since a few years now as mnemoc was paying the full amount and i get to deduct this from my taxes
<libv>
i do not have the time or the interest to deal with it anymore, and the only party with a monetary interest in linux-sunxi is bootlin, but mripard has always treated the wiki very stepmotherly and you will not see any changes to the wiki by mripard outside of the overgrown mainline page which is more like a changelog than anything else
<libv>
so i am unsure how to proceed
<KotCzarny>
move everything to github wiki and be done?
<montjoie>
I have already proposed to do the sysadmin task for it
<libv>
montjoie: i would then also want to get rid of both the expense and the legal liability of this machine
<libv>
KotCzarny: how do you propose to do that move?
<montjoie>
do you know data size and access volumetry ?
<libv>
and do you really see the joke that is the github or similar wikis as equivalent to a proper mediawiki?
<KotCzarny>
linux-sunxi already has a github place
<MoeIcenowy>
try to archive legacy kernel about contents to archive.org?
<KotCzarny>
so it would be a matter of doing converting work
<KotCzarny>
assuming most of the functionality could be preserved
<libv>
also, what will you do when github declines?
<libv>
which it will
<montjoie>
libv: If volumetry can fit on my server
<KotCzarny>
would it?
<libv>
if you've been around long enough, you should know that it will
<libv>
sourceforge is a great example
<KotCzarny>
data is related to the project
<KotCzarny>
and afaik no copyrighted files (apart from specs) are there
<libv>
there used to be a time when there was little way around sourceforge
<libv>
heh, this is another thing
<libv>
some of the things hosted on linux-sunxi are grey, like the bsps with gpl violating code
<libv>
and the datasheets
<KotCzarny>
that could be moved to archive.org
<KotCzarny>
but the rest might be perfectly legal for github
<libv>
i do not have any info on data size or volume of traffic atm
<libv>
and i need to go get the sixt car and pack it up so that me and my family can go visit family in belgium, so i have little extra cycles today, and the next 1.5 weeks
<libv>
but, so far, the github wiki idea is stupid as it will kill the one thing that makes linux-sunxi stand out
<libv>
today still
<libv>
making it better than even the rpi.
<libv>
the device first wiki.
<libv>
github fucked up cloning a kernel for me btw, just 3 weeks ago
<MoeIcenowy>
maybe everyone can try to migrate dl.linux-sunxi.org to archive.org first
<MoeIcenowy>
as it's not active for a long time
AneoX_ has quit [Ping timeout: 264 seconds]
<libv>
anyway, think about it, send me some emails, i will be packing a car with babygear and driving it around, for the next 1.5 weeks
dddddd has joined #linux-sunxi
<libv>
but i warn you, if you manage to make the wiki even less accessible, you will have dropped linux-sunxi in the same dark ages that everyone else is in
<KotCzarny>
afair gh is also user editable with proper settings
<libv>
then there's the spam situation to deal with
<KotCzarny>
with additional bonus of allowing changes as a pulls
<MoeIcenowy>
libv: btw what's the current linux-sunxi.org server's hdd size?
<KotCzarny>
yeah, wiki on gh can be made public (ie. requiring valid gh account)
Wizzup has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<MoeIcenowy>
jernej: it works, thanks
afaerber has quit [Quit: Leaving]
<jernej>
glad that it works
afaerber has joined #linux-sunxi
<MoeIcenowy>
and I also met the strange breaking replug on the WS LCD
<MoeIcenowy>
maybe we need some anti-jitter code in hpd read code?
<MoeIcenowy>
when it's not working, the kernel log says `[ 326.428713] [drm:drm_helper_hpd_irq_event [drm_kms_helper]] [CONNECTOR:34:HDMI-A-1] status updated from disconnected to disconnected`
<jernej>
yeah, but implementation would be ugly in any case
<MoeIcenowy>
yes
<MoeIcenowy>
real world cannot be clean
<MoeIcenowy>
it's a law ;-)
<MoeIcenowy>
do you know about other DW HDMI platforms
<MoeIcenowy>
do they have the jitter problem?
<jernej>
you would need some kind of thread which would periodically check status and filter out jitter
<jernej>
same way as it is done in BSP
<MoeIcenowy>
maybe we can use some simpler way?
<jernej>
I'm not aware of any other platform with this issue
<MoeIcenowy>
similar to de-jittering on a button?
<MoeIcenowy>
(sample again
<jernej>
H6 probably doesn't have this issue
<MoeIcenowy>
seems so
<jernej>
do you see the issue often?
<MoeIcenowy>
very often
<MoeIcenowy>
it's the reason why replugging WS LCD doesn't work
<jernej>
then I guess this must be something else
<MoeIcenowy>
oh I mean very often when plugging
<MoeIcenowy>
(and deplugging
<jernej>
if interrupt is triggered, screen is detected...
<MoeIcenowy>
I plug it slowly
<MoeIcenowy>
so the jitter may be longer
<jernej>
which board?
reinforce has joined #linux-sunxi
<MoeIcenowy>
pine a64 lts
<MoeIcenowy>
jernej: sorry sometimes the reason why ws lcd doesn't work is not hpd
<jernej>
yeah, I think there is another reason
<MoeIcenowy>
but doing a delay-based de-jitter really fixes the hpd issue
<MoeIcenowy>
now I use a quite ugly scheme
<jernej>
can you show me your algo?
<MoeIcenowy>
read first - sleep 10ms - read second - if (first == second) return first - sleep 10ms - read third - if (first == third) return first else return second /* == third */
<jernej>
quick google check shows that some other platforms have also hdmi hpd jitter issue although they are not using DW HDMI (at least it seems so)
<jernej>
MoeIcenowy: but you are not doing that in interrupt context?
<MoeIcenowy>
I'm doing this in dw_hdmi_phy_read_hpd
<jernej>
I wonder if debouncing capacitor on hpd line would help
<MoeIcenowy>
which is called by drm fx
<jernej>
but I'm not sure what capacitance limit hdmi standard allows on hpd signal
<MoeIcenowy>
(indirectly
<MoeIcenowy>
jernej: BTW now both U-Boot and Linux failed to lit my father's HP ZR24w
<parco>
I wonder why the MIPI DSI port on the Banana Pi is laid out with the differential pairs separated by other signals
<anarsoul|2>
ask Banana Pi devs?
<parco>
Was hoping to find some here, they don't seem to have their own irc
<micken>
So again SMHC
<micken>
anarsoul|2: in u-boot.. if I write CMD3 (rel data address) to CMD I get 0x104 in raw intr, wich means command completed and timeout respons
<micken>
I tried to clock down
<micken>
but the result is still the same
<anarsoul|2>
micken: I'm not an expert in mmc driver
<micken>
understood
<micken>
quite fun to clock down mmc in u-boot, gets slower and slower and then stops working
f11f12 has joined #linux-sunxi
f11f12 has quit [Quit: f11f12]
DonkeyHotei has joined #linux-sunxi
<micken>
btw , I learned the hard way that the pmic register that sets voltage on sd/emmc/nand also affects lcd ( it is in pinebook schematics , just didn't notice)
<anarsoul|2>
:)
<anarsoul|2>
be careful, don't fry your LCD
<micken>
I tried to powercycle mmc , which made lcd go black =)
<micken>
don't worry I know the original value
<micken>
it was pointless anyway
<anarsoul|2>
micken: I think it's possible to set voltage higher than LCD can handle
<anarsoul|2>
so be careful
<micken>
yes I know
<micken>
I checked the value before I powered down ( and restored it after)
<micken>
but anyway it was a uneeded action, NetBSD doesn't play with powercycling
lurchi__ has quit [Remote host closed the connection]
chomwitt has joined #linux-sunxi
popolon has quit [Quit: WeeChat 1.9.1]
nuuuciano has quit [Read error: Connection reset by peer]