UVM Fault with CARP on re(4)

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

UVM Fault with CARP on re(4)

>Synopsis:      UVM Fault with CARP on re(4)
>Category:      Kernel
        System      : OpenBSD 5.8
        Details     : OpenBSD 5.8 (GENERIC) #1066: Sun Aug 16 02:33:00 MDT 2015
                         [hidden email]:/usr/src/sys/arch/i386/compile/GENERIC

        Architecture: OpenBSD.i386
        Machine     : i386
        Two gateways OpenBsd 5.8 with CARP, TP-Link TG-3269 adapater (driver re(4))
        nonregular untimely crash -> launch ddb with message :

        uvm_fault(0xd0b68820, 0xd1aba000, 0, 1) -> e
        kernel: page fault trap, code=0
        Stopped at memcpy+0x13:    repe movsl      (%esi),%es:(%edi)

        trace ddb command give :

        memcpy(d5e8c200,2000,3b9aca00,2,1) at memcpy+0x13
        m_copym2(d5e8c200,0,3b9aca00,2,d0b26130) at m_copym2+0x2e
        carp_input(d18e8054,d5e8c200,0,d5e8c900,d5e8c900) at carp_input+0xe5
        if_input_process(d0b26120,f5bdcf68,d03bb0c0,0,0) at if_input_process+0xba
        taskq_thread(d1869000) at taskq_thread+0x4f

        it always ends up happening...

        I noted : http://permalink.gmane.org/gmane.os.openbsd.tech/47436
        or http://openbsd-archive.7691.n7.nabble.com/sis-4-induced-uvm-faults-on-net4801-td287109.html

        it seems that the problem is similar (sis)