Hi Alexandre,

Could you compile with `make debug`, so that dmesg gets plenty of
debugging output? That will let me see if this is related to
handshakes or to transport messages. An initial theory, based on the
stacktrace there involving fmonty and such, is that the 32-bit
implementation of curve25519-donna uses a lot of stack space, and if
you're on a kernel that puts the softirq stack and the process stack
in the same place, this could be an issue. Also, would you mind
sending your kernel's .config?

Jason
_______________________________________________
WireGuard mailing list
[email protected]
https://lists.zx2c4.com/mailman/listinfo/wireguard

Reply via email to