Re: KASSERT(ci->ci_ilevel < IPL_HIGH) failed when running a linux binary

[Arnaud Lacombe <lacombar%gmail.com@localhost>]

Hi,

On Fri, Apr 16, 2010 at 9:26 PM, Mindaugas Rasiukevicius
<rmind%netbsd.org@localhost> wrote:
> Arnaud Lacombe <lacombar%gmail.com@localhost> wrote:
>> On a -current kernel from CVS of April 2nd, my machine is crashing on
>> following assertion:
>>
>> KASSERT(ci->ci_ilevel < IPL_HIGH) in arch/x86/x86/pmap.c
>>
>> It is fairly easy to reproduce: after a fresh reboot, while running
>> grep in background, starting a linux binary in a loop is enough
>> trigger the issue. <...>
>> dmesg(8) of an older previous kernel attached.
>
> Well, dmesg output is quite useless in such case.  Can you get a backtrace?
> That would be a good start.
>
D'oh... The printf() call in netbsd_elf_signature() at
sys/kern/kern_exec.c:882, trigger a fault. The exact trace is

netbsd_elf_signature() ->
  strlen() ->
    trap(): trap number 6 ->
      uvm_fault_internal() ->
        pmap_map_ptes() ...

The message printed during previous successful executions is:

/path/to/bin/gcc: bad tag 1: [14 4, 16 4, GNU PaX]

 - Arnaud