frozeus
-
Posts
36 -
Joined
-
Last visited
Content Type
Forums
Store
Crowdfunding
Applications
Events
Raffles
Community Map
Posts posted by frozeus
-
-
I still think it's related to timing and power of the PHY. Not memory messup.
-
You need also to rmmod the MAC module (something like dwmac-sun8i) it's him who manage the phy-supply.
rmmod phy && rmmod mac
modprobe mac (should also modprobe the phy) -
-
@martinayotte I also don't use network manager. And I think it's more related to the timing here. Maybe the regulator as time to discharge because network manager is slower or maybe it makes the things more properly.
If you have an oscilloscope checking the phy reg should the easiest way to solve this.
Else maybe you can try
Ifdown eth0; sleep 2; reboot ?
I'm not familiar with the kernel and how the regulator are shutdown with network.
Is it possible to compile Ethernet phy and stack as a module and try to rmmod / modprobe and see if you trig the same issue ?
Clement
-
No issue on my Beelink GS1 on top of sunxi/for-next
-
19 minutes ago, martinayotte said:
Right !
It is just a pandora box been opened ...
We just discovered that : cold boot != warn boot
-
12 minutes ago, @lex said:
I don't see your u-boot patched , it is the only ATF that should be patched?
https://github.com/clementperon/u-boot/commits/beelink_gs1
"arm: sunxi: h6: fix reset using r_wdog"
-
Ethernet is still working at first cold boot right ?
So my thought is the PHY is still powered at boot and doesn't have a proper reset at probe in Linux.
So the reset patch didn't add the issue, the issue is here and we see it because we never rebooted before.
-
no the SUNXI_WDT_QUIRK is only set to 1 in sun50i_h6/platform.mk not for all platforms.
-
I have proposed a proper patch to ATF hope it will be accepted soon
https://github.com/ARM-software/arm-trusted-firmware/pull/1926
-
It has been merged so no need to add :
https://github.com/armbian/build/tree/master/patch/atf/atf-sunxi64
-
4 minutes ago, martinayotte said:
Oh ! On the board side ...
Next time, it is better to short the USB-TTL side, they are less expensive ...
Yes, at least the RX is still working.
BTW i have sent the patch for U-Boot if you could add your "Tested-by" tag for the reset patch.
Thanks
-
I made a short circuit on my board and the TX pin is no more working...
-
My UART is broken could you test this patch ?
https://github.com/clementperon/u-boot/commit/dc1abccb3628aefa6df3d44e027f33d8afc3fc86
Thanks
-
You just need to put it in build/patch/atf/atf-sunxi64/
-
Before trying any patch it would better to first understand, how it works on PineH64.
-
-
12 minutes ago, martinayotte said:
You are better looking at the schematic of any H6 (OPi3, OPiOnePlus, OPILite2) before arguing, PL9 is S-IR-RX, PH7 is really SPDIF-OUT ...
Oups indeed, My brain switch between IR and SPDIF.
The patch is from me and yes I tested it. Still not posted it because it's more a hack.
-
Changing to the R_WDOG instead of WDOG in ATF fix the issue.
-
SPIDF-OUT is PL9.
I think the legacy version doesn't use the watchdog or not the same to reboot the board.EDIT: I think it uses the ARISC co-processor
does someone with the h64 can test if the WDOG works ?
/* This should reset the system after 0.5s */ # devmem $((0x030090a0 + 0x0014)) 0x00000001 # devmem $((0x030090a0 + 0x0018)) 0x00000000 # devmem $((0x030090a0 + 0x0018)) 32 0x1 /* This should reset the system after 1s (as recommended in the User Manual) */ # devmem $((0x030090a0 + 0x0014)) 0x00000001 # devmem $((0x030090a0 + 0x0018)) 0x00000000 # devmem $((0x030090a0 + 0x0018)) 32 $((0x1 << 4)) # devmem $((0x030090a0 + 0x0018)) 0x00000010 # devmem $((0x030090a0 + 0x0018)) 32 $((0x1 << 4 | 0x1))
-
"IIRC there are 3 watchdogs on the H6. One is the trusted watchdog, which should be reserved for underlying firmware. The other two are the normal and R-block watchdog. Although they look the same, the "normal" one can't reset the system. That one is actually the one ATF currently uses. So reboot doesn't work on the H6.
ChenYu" - https://lkml.org/lkml/2019/4/5/425
If it's true it's strange that the reboot works on PineH64...
-
I"m quite sure the issue is comming from SCPI and ATF.
-
ssh to the board, then dmesg and copy/paste
-
Hhmmm doesn't fix the issue :'(
[ 3.497925] sunxi-wdt 30090a0.watchdog: Watchdog enabled (timeout=16 sec, nowayout=0)
still stuck
H6 Famous Reboot problem
in Allwinner sunxi
Posted
Great ! But still a workaround, will not work for people doing nfsroot.
@martinayotte is there a Dev branch of 5.1 ?
FYI I have test the IR receiver on H6 works fine :
https://github.com/clementperon/linux/commits/h6_ir