I thought stacks were separated by redzone pages, but I'd have to double
check.
I recommend rebooting with kmem_flags = 0xf in your /etc/system.
-- Garrett
Andrew Gallatin wrote:
> I just got a very strange panic when running a torture
> test on my GLDv3 driver:
>
> in.rshd:
> #pf Page fault
> Bad kernel fault at addr=0x0
> pid=16617, pc=0xfffffffff84f87a7, sp=0xffffff00083ec5f8, eflags=0x10246
> cr0: 8005003b<pg,wp,ne,et,ts,mp,pe> cr4: 6f8<xmme,fxsr,pge,mce,pae,pse,de>
> cr2: 0
> cr3: 3493000
> cr8: c
>
> rdi: ffffff01d72a1a10 rsi: ffffff00083ec600 rdx: ffffff01d44eb120
> rcx: 3 r8: ffffff01cfde8500 r9: 85b
> rax: 0 rbx: ffffff01cecc9c60 rbp: 0
> r10: 300004c57 r11: ffffff01e6d9c000 r12: ffffff01cecc9c60
> r13: ffffff01d72a1a10 r14: ffffff01d2ecc080 r15: ffffff01d0935bb8
> fsb: 0 gsb: ffffff01ceaa6ac0 ds: 4b
> es: 4b fs: 0 gs: 1c3
> trp: e err: 0 rip: fffffffff84f87a7
> cs: 30 rfl: 10246 rsp: ffffff00083ec5f8
> ss: 38
>
> ffffff00083ec3e0 unix:die+c8 ()
> ffffff00083ec4f0 unix:trap+13b9 ()
> ffffff00083ec500 unix:cmntrap+e9 ()
>
>
> According to mdb, there is no stack:
>
> > $C
>
> I'm assuming the stack got corrupted somehow, but the current thread
> seems well within its stack:
>
> > ffffff01d44eb120::print struct _kthread t_stkbase
> t_stkbase = 0xffffff00083e8000
>
>
> Is it possible some other stack smashed into this thread's stack, and
> trashed it? Will Solaris panic if a thread exceeds its kernel stack
> space, or will it just corrupt whatever is below it? How do I debug
> something like this?
>
> Thanks,
>
> Drew
> _______________________________________________
> networking-discuss mailing list
> [email protected]
>
_______________________________________________
networking-discuss mailing list
[email protected]
