WIll soft reboot on linux "reboot" will effect the signal nRESETIN_OUT

Hi,

I will forward this to the SW team.

From the HW point of view, nRESETIN_OUT is triggered by all reset sources, except JTAG TRSTz (Table 8-25 from the AM335X TRM Rev. L).

Yes according to TI AM335X ref manual it should, but software "reboot" is not affecting this signal.

Are you sure? What external components do you have connected to this line?

Do you have an external pullup resistor on this line? Any capacitors?

Yes this signal is pulled up by 10K register.  Only through hardware reset this signal is getting affecting.

If you copied the Beaglebone Black schematic please check whether you have capacitors on this line. The BBB has a 1uF capacitor that effectively filters out the short pulse that is generated from the processor on this pin.

Another thing I just checked is that the LAN8710 PHY requires a minimum 100us reset pulse. The Warm Reset Pulse width is defined by PRM_RSTTIME register bits [7:0] + max. 50 main oscillator cycles. Default value is 6 + 50 = 56 / 24MHz = 2.3us, which will not be long enough to reset the PHY. Please refer to sections 8.1.7.4.1.2 and 8.1.13.5.2 from the AM335X TRM Rev. L for details.

Yes, you can.

Sorry, I'm not a SW expert, so I couldn't help on this. You can try changing the value with the devmem2 command to see if there will be any effect. This register contents should not be affected by warm reset.

this should be modified in bootloader or linux kernel ?

I think i cannot achieve 100us because maximum value is 255+50/24MHZ=12.2us. please clarify on this.

Yes, this calculation is correct. In this case I cannot see what you can do, except move PHY reset to a GPIO.

But in beaglebone black also we are using the same nRESETIN_OUT signal to reset the phy.

I am facing the same issue on customized board as per this post
e2e.ti.com/.../335017

I have seen that voltage on PHYAD1 changes to 1 so phy address is getting changed and i am getting "PHY not found error" some time on power up. say once in 20-30 power cycles.

So i have seen that phy mask is getting changed

Failure log

6.723815] davinci_mdio davinci_mdio.0: davinci mdio revision 1.6
[ 6.730133] davinci_mdio davinci_mdio.0: detected phy mask 3f3f3f01
[ 6.744079] davinci_mdio.0: probed
[ 6.747589] davinci_mdio davinci_mdio.0: phy[1]: device 0:01, driver Micrel KSZ8895 or KSZ8864

[ 9.462432] PHY 0:00 not found
[ 9.512542] ksz8895_config_init: phydev->phy_id=0x221450, phydev->addr=0x4
[ 9.520294] net eth1: CPSW phy found : id is : 0x221450

Normal Log.

2.165649] davinci_mdio davinci_mdio.0: davinci mdio revision 1.6
[ 2.171997] davinci_mdio davinci_mdio.0: detected phy mask 3f3f3f00
[ 2.186523] davinci_mdio.0: probed
[ 2.190032] davinci_mdio davinci_mdio.0: phy[0]: device 0:00, driver SMSC LAN8710/LAN8720
[ 2.198455] davinci_mdio davinci_mdio.0: phy[1]: device 0:01, driver Micrel KSZ8895 or KSZ8864

9572] net eth0: CPSW phy found : id is : 0x7c0f1