H96 Pro+ S912 Standby mode

Help & Support Hardware
101

I guess I forgot to say my problem still remains. The change from cec19 to cec3f had no effect. But thanks for the help!

If you could provide the u-boot for DDR3 CL11, I would be able to test it for you. :wink:

102

@bumerc

v2_2#kbi init
i2c_read: i2c transfer failed
Error reading the chip: -6

I just tried with the enabled kbi interface. But the I2C address does not match.
The kbi interface will read data from a “eeprom” what the H96Pro+ do not have.
Only the function “set_wol” is needed…

104

Hi,
first I took over the complete KBI solution without changing anything on the code. So you can directly see what works and what does not, without looking into the HW specifications. The redundant code can later be discarded.

105

Is there any HW documentation for H96 Pro+? If I read the VIM2 code/schematic correctly, there is some work to do to prepare the PHY for WOL and keep the appropriate power at the network interface by something looks like a PIC.
On H96 Pro + there is a PHY mentioned at kernel log, but seems not to be the same as for VIM2.
Thanks for any hint

106

Unfortunately, I have not seen any HW documentation for this device yet.

107

Sorry, I am wrong. The PHY seems to be the same:

eth0: PHY ID 001cc916 at 0 IRQ POLL (stmmac-0:00) active

108

Hi @atomizasser,

If I understand correctly you’re able to suspend and wake up your minix u9-h with your IR remote without any problems?
Is it possible to post the steps needed to do this? I would like my minix to suspend instead of power off but I’m confused about the things I need to do.
At the moment it just bugs out…

Thanks,

109

I followed the guide in https://github.com/CoreELEC/u-boot/wiki#upgrade-u-bootbin and used method 2 for reboot minix.
the minix enters in suspend mode, but after wake up wifi is not turned on.
@Portisch @bumerc is there any method to avoid wifi module turned off in suspend mode or to reconnect after wake up form suspend mode?
Thanks.

110

What defconfig did u use?
the one in the guide: make h96proplus_v2_2_defconfig ?

111

Well i used compiled version and YES i used h96proplus_v2_2 u -boot.

By the way, doing this i got wifi working after wake up from suspend state:

https://forum.libreelec.tv/thread/524-s905-builds-general-discussion/?postID=40599#post40599

1 Like
112

@derknut @bumerc
With the last commit the external RTL8211F is up and running in u-boot:

Now you need to set the wol mac end enable the chip when enter suspend. This is done like the set_wol in cmd_kbi:

#include <common.h>

#define MAC_LENGHT 6

static void set_wol(int enable)
{
char cmd[64];
int mode;

if ((enable&0x01) != 0) {

int mac_addr[MAC_LENGHT] = {0};

run_command(“phyreg w 0 0x1040”, 0);

run_command(“phyreg w 31 0xd40”, 0);
run_command(“phyreg w 22 0x20”, 0);
run_command(“phyreg w 31 0”, 0);

  char *s = getenv("mac");
  if ((s != NULL) && (strcmp(s, "00:00:00:00:00:00") != 0)) {
  	printf("set mac for wol = %s\n", s);
  	int i = 0;
  	for (i = 0; i < 6 && s[0] != '\0' && s[1] != '\0'; i++) {
  	mac_addr[i] = chartonum(s[0]) << 4 | chartonum(s[1]);
  	s +=3;
  	}
  } else {
  	printf("no mac for wol found\n");
  	return;
  }

run_command(“phyreg w 31 0xd8c”, 0);
sprintf(cmd, “phyreg w 16 0x%x%x”, mac_addr[1], mac_addr[0]);
run_command(cmd, 0);
sprintf(cmd, “phyreg w 17 0x%x%x”, mac_addr[3], mac_addr[2]);
run_command(cmd, 0);
sprintf(cmd, “phyreg w 18 0x%x%x”, mac_addr[5], mac_addr[4]);
run_command(cmd, 0);
run_command(“phyreg w 31 0”, 0);

run_command(“phyreg w 31 0xd8a”, 0);
run_command(“phyreg w 17 0x9fff”, 0);
run_command(“phyreg w 31 0”, 0);

run_command(“phyreg w 31 0xd8a”, 0);
run_command(“phyreg w 16 0x1000”, 0);
run_command(“phyreg w 31 0”, 0);

run_command(“phyreg w 31 0xd80”, 0);
run_command(“phyreg w 16 0x3000”, 0);
run_command(“phyreg w 17 0x0020”, 0);
run_command(“phyreg w 18 0x03c0”, 0);
run_command(“phyreg w 19 0x0000”, 0);
run_command(“phyreg w 20 0x0000”, 0);
run_command(“phyreg w 21 0x0000”, 0);
run_command(“phyreg w 22 0x0000”, 0);
run_command(“phyreg w 23 0x0000”, 0);
run_command(“phyreg w 31 0”, 0);

run_command(“phyreg w 31 0xd8a”, 0);
run_command(“phyreg w 19 0x1002”, 0);
run_command(“phyreg w 19 0x1002”, 0);

} else {
run_command(“phyreg w 31 0xd8a”, 0);
run_command(“phyreg w 16 0”, 0);
run_command(“phyreg w 17 0x7fff”, 0);
run_command(“phyreg w 31 0”, 0);
}
}

Then you have to check in detect_key function for the gpio: GPIOZ_15 or BOOT_8 (going from 1 to 0)
If 0 then wakeup by WOL. Manualy it is working fine - but I need help to implement the “set_wol” function…

113

That’s very good news. How did you test it, where you called set_wol()?
I would propose to write a new cmd module called “h96” or “wol” to setup the PHY. But I have no idea where to call it.
Thanks

114

set_wol needs to be called after entering the suspend mode. But this is not possible because it is a different task.
Also compiling does not work…

I started u-boot into console with the last commit.
Entered by hand line for line into the console to setup the RTL8211F.
This brings the PEMB pin high and the chip is waiting for a magic packet.
If a packet got received the hardware pin goes low.
Then the endless while loop detect_key in pwr_ctrl.c have to check the gpio and start the wakeup/resume.

The problem is also when the linux kernel starts the RTL8211f registers are getting modified so the wol would not work anymore. The set_wol have to be called again.
The only way I see right now is that the linux kernel setup the chip for wol and u-boot is checking the gpio for wol wakeup.

115

That’s sound promising and is the regular “Linux way” to activate WOL by “ethtool”. I will check the Linux part. Alternative it may be possible to add a quick and dirty hack to pwr_ctr.c to send MDIO commands to the PHY without a driver.

116

@derknut WOL is working now:


Use the commit “check for RTL8211F PMEB interrupt by Wake-on-Lan packet” only if you included first the linux kernel pull request! Otherwise u-boot will resume immediately from poweroff/suspend.

I will write a how to enable the WOL to the wiki:


1 Like
117

@atomizasser you can do the following as a quick fix for wifi on resume

  1. Go to \\Configfiles\ and make a folder sleep.d
  2. Then in that folder create a file named 01-wifi-resume.power with this content:
#!/bin/sh
case "$1" in
pre)
# <do something on suspend>
;;
post)
# <do something on resume>
connmanctl disable wifi
sleep 5
connmanctl enable wifi
;;
esac
118

Hi @Portisch,

I’ve installed your latest u-boot (just like atomizasser) on my minix u9-h.
However when I choose to suspend it doesn’t work (it’s the same as before).

I’ve uploaded the debug log from the moment i pressed the suspend. Can I provide anything else that may help to give you an idea why it doesn’t suspend? It would be great if you could take a look :slight_smile:
Or maybe there is something else blocking the suspend?

Thanks in advance,

kodi.log (8.8 KB)

edit: I’ve installed u-boot_q201_v2.2_green_PCB.tar.gz following the wiki on the github

119

This is the patch to compile CoreELEC with WOL on the wired eth interface enabled:linux-010_rtl8221f_wol_S912_gpioz_15.patch (4.8 KB)
Just copy it to

CoreELEC/packages/linux/patches

and compile the image with

PROJECT=Amlogic DEVICE=S912 ARCH=arm make

After update of the CoreELEC system update u-boot:
Compile and update u-boot with the last commit:

After flashing enable the WOL for the wired ethernet interface:

120

@freshprince only a dmesg log will help. Also a uart log of u-boot would be nice. Try to contact @atomizasser because I do not own such a box.

121

Hmm strange it keeps spamming a message in dmesg.
See attachment
dmesg.txt (17.6 KB)