Re: [PATCH] x86, 32-bit: Fix invalid stack address while in softirq

[wyang1 <Wei.Yang@windriver.com>]

On 09/06/2012 11:54 PM, Steven Rostedt wrote:
> On Thu, 2012-09-06 at 17:36 +0200, Robert Richter wrote:
>
>> I meant:
>>
>> unsigned long kernel_stack_pointer(struct pt_regs *regs)
>> {
>> 	unsigned long context = (unsigned long)regs&  ~(THREAD_SIZE - 1);
>> 	unsigned long sp = (unsigned long)&regs->sp;
>> 	struct thread_info *tinfo;
>>
>> 	if (context == (sp&  ~(THREAD_SIZE - 1)))
>> 		return sp;
>> 	
>> 	tinfo = (struct thread_info *)context;
>> 	if (tinfo->previous_esp)
>> 		return tinfo->previous_esp;
>>
>> 	return (unsigned long)regs;
>> }
>>
>> -Robert
>>
>>> Maybe this is even better.
>>>
>>>
Stevent & Robert,

Actually, we also can revert the following commit:

commit 7b6c6c77732ca1d2498eda7eabb64f9648896e96
Author: Masami Hiramatsu <mhiramat@redhat.com>
Date:   Mon May 11 17:03:00 2009 -0400

     x86, 32-bit: fix kernel_trap_sp()

     Use &regs->sp instead of regs for getting the top of stack in 
kernel mode.
     (on x86-64, regs->sp always points the top of stack)


I meant:

static inline unsigned long kernel_stack_pointer(struct pt_regs *regs)
{
     #ifdef CONFIG_X86_32
        return (unsigned long)regs;
     #else
        return regs->sp;
     #endif
}


What do you think of it?

Thanks
Wei

> Yeah, this is probably the safest.
>
> Thanks,
>
> -- Steve
>
>
>