Mr__Anderson has quit [Read error: Connection reset by peer]
Leepty has quit [Ping timeout: 240 seconds]
Mr__Anderson has joined #linux-sunxi
Mr__Anderson has quit [Ping timeout: 255 seconds]
Leepty has joined #linux-sunxi
<MoeIcenowy>
wens, mripard: how to force a clock to be under a certain parent?
qeed has joined #linux-sunxi
<MoeIcenowy>
I'm trying dual pipeline on A64 Pinebook
<MoeIcenowy>
I assigned the parents of tcon1 and hdmi clocks to pll-video1 in dt
<MoeIcenowy>
and the parents are really pll-video1 when boot (without HDMI cable inserted)
<MoeIcenowy>
however when I really insert the HDMI cable the parents become pll-video0 again and conflicted with the LCD clock
fkluknav has joined #linux-sunxi
gzamboni has joined #linux-sunxi
<MoeIcenowy>
should I add CLK_SET_RATE_NO_REPARENT to HDMI clocks and force the parents in CCU driver?
perr has quit [Remote host closed the connection]
perr has joined #linux-sunxi
<wens>
you probably need CLK_SET_RATE_NO_REPARENT in the clk driver, then do explicit clk_set_parent calls in your HDMI driver
<MoeIcenowy>
wens: do such kind of set_parent is ugly
<MoeIcenowy>
as the driver is of course also used on H3
<MoeIcenowy>
which HDMI pipeline is tcon0->hdmi
<mripard>
MoeIcenowy: forcing parents is just as ugly
<MoeIcenowy>
mripard: assign it in dt?
<mripard>
assign it anywhere
<mripard>
if pll-video0 is perfectly able to handle the rate, why would you use pll-video1 /
<mripard>
?
<wens>
probably because you have 2 pipelines active with incompatible dot clocks?
<MoeIcenowy>
wens: yes
<mripard>
wens: then pll-video0 is not able to handle that rate, and you shouldn't use it
<MoeIcenowy>
and tcon0 can only use pll-video0-2x or pll-mipi
<mripard>
but my point is that it should be dynamic, not static
<MoeIcenowy>
pll-mipi is also under pll-video0
<mripard>
if you can satisfy both pipelines with a single clock, there's no reason to use two clocks
enrico__ has joined #linux-sunxi
<MoeIcenowy>
mripard: nope it's impossible to satisfy them with a single clock.
<mripard>
MoeIcenowy: that's not "impossible".
<mripard>
if you have both pipelines running at the same resolution, it's definitely possible
<mripard>
and that's not an unlikely assumption
paulk-elm has joined #linux-sunxi
<MoeIcenowy>
mripard: but we shouldn't restrict this
<MoeIcenowy>
and in fact my debugging environment here have two pipelines both with definite timing -- one can be only 1024x600 (HDMI screen) and one only 1366x768 (eDP screen via RGB-eDP bridge)
<MoeIcenowy>
P.S. for pll-video0 and pll-video1 the parameters have no difference
Leepty has quit [Ping timeout: 268 seconds]
Mr__Anderson has joined #linux-sunxi
chomwitt has joined #linux-sunxi
Leepty has joined #linux-sunxi
Mr__Anderson has quit [Read error: Connection reset by peer]
popolon has joined #linux-sunxi
<MoeIcenowy>
I assigned hdmi and tcon1's parent to pll-video1 and set CLK_SET_RATE_NO_REPARENT to them
<MoeIcenowy>
HDMI has also CLK_SET_RATE_PARENT
<MoeIcenowy>
but the HDMI's CLK_SET_RATE_PARENT failed to work
<MoeIcenowy>
the pll-video1 is still 297MHz when setting HDMI to 128MHz
<MoeIcenowy>
(and this result in HDMI being 99MHz
<MoeIcenowy>
and then TCON1 got to 29.7MHz, which needs to be 32MHz
chomwitt1 has joined #linux-sunxi
chomwitt has quit [Ping timeout: 240 seconds]
yann-kaelig has joined #linux-sunxi
afaerber has joined #linux-sunxi
deskwizard has left #linux-sunxi ["faded away."]
Nacho___ has quit [Remote host closed the connection]
<KotCzarny>
(i have blanked the device, but wondering what smelly code in smelly 3.4 does)
<MoeIcenowy>
KotCzarny: BSP?
<MoeIcenowy>
I think BSP kernel uses many MBR backups, as they try to be compatible with NAND
<KotCzarny>
ho hum.
<MoeIcenowy>
and the wasted 19MiB space is also for "be compatible with NAND"
foxx_ has quit [Ping timeout: 258 seconds]
<oliv3r>
Hey guys, did anybody notice (potential) u-boot brackage with master? RTL8211X_PHY_FORCE_MASTER seems to have been stripped from the configs and because we no longer have SUNXI_GMAC we are now also not setting PHY_REALTEK anymore
paulk-elm has quit [Remote host closed the connection]
paulk-elm has joined #linux-sunxi
lurchi_ is now known as lurchi__
chomwitt1 has quit [Ping timeout: 240 seconds]
<MoeIcenowy>
dgp, tkaiser: I will try to adapt XR819 "new driver" from FriendlyARM to mainline
Mr__Anderson has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
paulk-elm has quit [Quit: Leaving]
paulk-elm has joined #linux-sunxi
LargePrime has quit [Ping timeout: 260 seconds]
cnxsoft has quit [Quit: cnxsoft]
<montjoie>
ouch an user answering to my first sun8i-emac driver serie...
<montjoie>
so lost!
<wens>
it happens
dave0x6d has joined #linux-sunxi
LargePrime has joined #linux-sunxi
yann has quit [Ping timeout: 240 seconds]
paulk-elm has quit [Remote host closed the connection]
paulk-elm has joined #linux-sunxi
jernej has joined #linux-sunxi
marble_visions has joined #linux-sunxi
<marble_visions>
hi all, I want to add mainline linux support for the nand flash chip (H27UBG8T2BTR-BC-Hynix) that comes on the olinuxino board
<marble_visions>
i've gotten it to the state where the driver is loading (kernel config + dtb are correct)
<marble_visions>
but the driver cannot find the ecc data
reinforce has quit [Quit: Leaving.]
<marble_visions>
from that I've gathered now I only need to add the struct for the chip in drivers/mtd/nand/nand_ids.c
<marble_visions>
is that correct?
<MoeIcenowy>
marble_vision: wait for 4.12
<marble_visions>
MoeIcenowy: oh, ok. is that refs/next at the moment so I can have a look?
<oliv3r>
is it wise (and how do I) create a 'hidden' symbol in a Kconfig? e.g. i want to 'select SOME_HIDDEN_OPTION' and then in the source use 'CONFIG_SOME_HIDDEN_OPTION'
afaerber has quit [Remote host closed the connection]
paulk-elm has quit [Read error: Connection reset by peer]
IgorPec has quit [Ping timeout: 255 seconds]
gaby has quit [Quit: leaving]
gaby has joined #linux-sunxi
<mic-e[m]>
hi everybody, about a month ago I asked for help with enabling the micrel KSZ9031 in u-boot 2017-03, in order to configure some of its extended registers.
<mic-e[m]>
the result was a patch that added #define CONFIG_PHY_MICREL and #define CONFIG_PHY_MICREL_KSZ9031 to include/configs/sunxi-common.h
<mic-e[m]>
now I have upgraded to u-boot 2017-05 and the patch has stopped working
<mic-e[m]>
I'm back to "mdio list" / "0 - Generic PHY <--> ethernet@01c50000" in u-boot's boot prompt
<oliv3r>
mic-e[m]: what board do you have?
<mic-e[m]>
oliv3r: it's a Olimex A20-SOM that's sitting on top of a custom PCB which has the micrel ksz9031 PHY.
<mic-e[m]>
(very similar to the Olimex LIME2)
<oliv3r>
mic-e[m]: ahh ok
<mic-e[m]>
I'm using armbian to build the image
<oliv3r>
mic-e[m]: well I belive there is a small bug introduced with the latest u-boots; i'm not sure if that is your issue however
<oliv3r>
did you create your own defconfig?
<mic-e[m]>
oliv3r: I don't think so, given that I don't even know what that is...
Ntemis has quit [Remote host closed the connection]
<mic-e[m]>
I'm mostly using armbian's LIME2 configs
<oliv3r>
well but you do not have a lime2 so it makes little sense to use that with the SOM
victhor has joined #linux-sunxi
<oliv3r>
espcially, since the lime2 uses a realtek PHY which forces master mode to boot
reinforce has joined #linux-sunxi
paulk-collins has joined #linux-sunxi
<mic-e[m]>
hm, how do I even figure out which defconfig armbian is using...
<oliv3r>
well if you are using the lime2 image, then it's using the lime2 defconfig
<oliv3r>
mind you, a defconfig is just that, a default config, you can take whatever you want as a defconfig, and then set options to end up where you need to be of course
<mic-e[m]>
ah yes, I see.
<mic-e[m]>
on armbian the defconfig is specified in config/boards/BOARDNAME.conf, and it is indeed set to the lime2 defconfig.
<mic-e[m]>
oliv3r: so what should I put in my custom defconfig? I don't see any references to the faulty RTL PHY in the u-boot 2017.05 LIME2 defconfig
victhor has quit [Ping timeout: 255 seconds]
<oliv3r>
it's implicitly set via include/configs/sunxi-common.h
perr has quit [Quit: Leaving]
<oliv3r>
since you are doing a custom board, here is how we do custom boards (based on a lime2 however in our case)
<paulk-collins>
run the defconfig kconfig target to get a defconfig from it
<mic-e[m]>
so the defconfig is basically a diff?
Gerwin_J has joined #linux-sunxi
fkluknav has quit [Ping timeout: 272 seconds]
Leepty has quit [Remote host closed the connection]
Gerwin_J has quit [Client Quit]
enrico__ has quit [Remote host closed the connection]
enrico__ has joined #linux-sunxi
<oliv3r>
mic-e[m]: that's the generated config :) so the first time, i 'loaded' a defconfig and generated the first version of that file; afterwards, i 'updated' that file
<oliv3r>
mic-e[m]: no, defconfig says 'for this board, you must have atleast these options enabled.
<oliv3r>
e.g. for the lime2, that's things like ARCH=sun7i; DRAM_FREQ=384Mhz; etc etc
matthias_bgg has quit [Quit: Leaving]
netlynx has joined #linux-sunxi
yann-kaelig has quit [Quit: Leaving]
IgorPec has joined #linux-sunxi
marble_visions has quit [Quit: Page closed]
enrico__ has quit [Quit: Bye]
diego__ has quit [Quit: Konversation terminated!]
Mr__Anderson has quit [Read error: Connection reset by peer]
|Jeroen| has joined #linux-sunxi
maz has quit [Quit: Leaving]
<montjoie>
qschulz: your patch seems working (at least no crash since 30 min of "watch sensors")
<montjoie>
still get "thermal thermal_zone0: failed to read out thermal zone (-110)"