Re: [patch 08/11] Modify Ptrace routines to access breakpointregisters

[Alan Stern]

On Tue, 10 Mar 2009, Ingo Molnar wrote:

> 
> * prasad@xxxxxxxxxxxxxxxxxx <prasad@xxxxxxxxxxxxxxxxxx> wrote:
> 
> > -static unsigned long debugreg_addr_limit(struct task_struct *task)
> > -{
> > -#ifdef CONFIG_IA32_EMULATION
> > -	if (test_tsk_thread_flag(task, TIF_IA32))
> > -		return IA32_PAGE_OFFSET - 3;
> > -#endif
> > -	return TASK_SIZE_MAX - 7;
> > -}
> > -
> 
> I dont see where this security check has been carried over into 
> the generic code. The new code has:

Probably the IA32_EMULATION stuff was added after the hw-breakpoint 
patch was written.

>  +int arch_check_va_in_userspace(unsigned long va, struct task_struct *tsk)
>  +{
>  +       return (va < TASK_SIZE);
>  +}
> 
> but i think that misses the detail that it's not just the start 
> address of an x86 breakpoint that has to be considered, but also 
> the end addess of it.
> 
> For example a hardware breakpoint can be at 0xbfffffff with a 
> length of 4 bytes - thus overlapping into kernel-space by 3 
> bytes. It is important to not let that happen.

Quite correct.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/