| 13:38 | Ke> | apparently there are premade watchdog chips I could use for mcbin reset header, I assume open drain type watchdog should work here? |
| 13:39 | Ke> | also I assume that reset header ground is the same as the ATX power source ground, right? |
| 13:49 | Ke> | these patches are required for mcbin ATF btw. https://github.com/ARM-software/arm-trusted-firmware/pull/1214 |
| 14:39 | jnettlet[m]> | yes, along with kernel patches that aren't mainline y et |
| 15:08 | Ke> | jnettlet[m]: are you aware of anyone doing that on Marvell/SR side? |
| 15:09 | jnettlet[m]> | Ke: well this would have to happen on the Marvell side. I have patched my local builds, however that is only a fix for one of the vulnerabilities. |
| 15:10 | jnettlet[m]> | Really none of this is an issue if you are running single user machines. |
| 15:11 | jnettlet[m]> | where this vulnerability is really a problem is hosted VMs |
| 15:41 | Ke> | jnettlet[m]: any comment on the open drain hw watchdog? |
| 15:47 | jnettlet[m]> | Ke: haven't tested it. |
| 15:48 | Ke> | but you assume it should work on mcbin reset header? |
| 15:48 | Ke> | definitely the second pin is connected to ground at least |
| 15:59 | Ke> | also, I want the fixes everywhere, because I want layered protection |
| 16:05 | jnettlet[m]> | Ke: so you plan on using a gpio to keep the watchdog alive? |
| 16:06 | jnettlet[m]> | the way that the armada pins are designed I am not sure if that will work. |
| 16:12 | jnettlet[m]> | I think push pull is probably a better bet |