Re: [PATCH] Powerpc/Watchpoint: Restore nvgprs while returning from exception

[Michael Ellerman <mpe@ellerman.id.au>]

Ravi Bangoria <ravi.bangoria@linux.ibm.com> writes:

> Powerpc hw triggers watchpoint before executing the instruction.
> To make trigger-after-execute behavior, kernel emulates the
> instruction. If the instruction is 'load something into non-
> volatile register', exception handler should restore emulated
> register state while returning back, otherwise there will be
> register state corruption. Ex, Adding a watchpoint on a list
> can corrput the list:
>
>   # cat /proc/kallsyms | grep kthread_create_list
>   c00000000121c8b8 d kthread_create_list
>
> Add watchpoint on kthread_create_list->next:
>
>   # perf record -e mem:0xc00000000121c8c0
>
> Run some workload such that new kthread gets invoked. Ex, I
> just logged out from console:
>
>   list_add corruption. next->prev should be prev (c000000001214e00), \
> 	but was c00000000121c8b8. (next=c00000000121c8b8).
>   WARNING: CPU: 59 PID: 309 at lib/list_debug.c:25 __list_add_valid+0xb4/0xc0
>   CPU: 59 PID: 309 Comm: kworker/59:0 Kdump: loaded Not tainted 5.1.0-rc7+ #69
>   ...
>   NIP __list_add_valid+0xb4/0xc0
>   LR __list_add_valid+0xb0/0xc0
>   Call Trace:
>   __list_add_valid+0xb0/0xc0 (unreliable)
>   __kthread_create_on_node+0xe0/0x260
>   kthread_create_on_node+0x34/0x50
>   create_worker+0xe8/0x260
>   worker_thread+0x444/0x560
>   kthread+0x160/0x1a0
>   ret_from_kernel_thread+0x5c/0x70

This all depends on what code the compiler generates for the list
access. Can you include a disassembly of the relevant code in your
kernel so we have an example of the bad case.

> diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
> index 9481a11..96de0d1 100644
> --- a/arch/powerpc/kernel/exceptions-64s.S
> +++ b/arch/powerpc/kernel/exceptions-64s.S
> @@ -1753,7 +1753,7 @@ handle_dabr_fault:
>  	ld      r5,_DSISR(r1)
>  	addi    r3,r1,STACK_FRAME_OVERHEAD
>  	bl      do_break
> -12:	b       ret_from_except_lite
> +12:	b       ret_from_except

This probably warrants a comment explaining why we can't use the (badly
named) "lite" version.

cheers