[PATCH 6/9] uprobes: flush cache after xol write

[Ananth N Mavinakayanahalli]

On Sat, Nov 03, 2012 at 05:33:01PM +0100, Oleg Nesterov wrote:
> On 10/29, Ananth N Mavinakayanahalli wrote:
> >
> > On Fri, Oct 26, 2012 at 06:39:51PM +0200, Oleg Nesterov wrote:
> > > >
> > > > > OTOH, I do not understand this stuff, everything is nop on x86. And
> > > > > when I look into Documentation/cachetlb.txt I am starting to think
> > > > > that may be this needs flush_icache_user_range instead?
> > > > >
> > > > > Rabin, Ananth could you clarify this?
> > > >
> > > > Yes. We need flush_icache_user_range(). Though for x86 its always been a
> > > > nop, one never knows if there is some Power4 or older machine out there
> > > > that is still being used. We are fine for Power5 and later.
> > >
> > > This is bad...
> > >
> > > flush_icache_user needs vma. perhaps just to check VM_EXEC...
> > >
> > > So let me repeat to be sure I really understand, do you confirm that
> > > _in general_ flush_dcache_page() is not enough?
> >
> > flush_dcache_page() on powerpc already checks for
> > CPU_FTR_COHERENT_ICACHE. So, yes, that is enough.
> 
> Thanks Ananth.
> 
> Still it is not clear to me if flush_dcache_page() would be always right
> if we add the new port.
> 
> OK. So I assume we need the fix and I am going to apply the patch below.
> 
> Ananth, Rabin, will you ack it (including the comment I affed) ?
> 
> Oleg.
> 
> ------------------------------------------------------------------------------
> [PATCH] uprobes: flush cache after xol write
> 
> From: Rabin Vincent <rabin at rab.in>
> 
> Flush the cache so that the instructions written to the XOL area are
> visible.
> 
> Signed-off-by: Rabin Vincent <rabin at rab.in>

Acked-by: Ananth N Mavinakayanahalli <ananth at in.ibm.com>

> 
> --- x/kernel/events/uprobes.c
> +++ x/kernel/events/uprobes.c
> @@ -1199,6 +1199,11 @@ static unsigned long xol_get_insn_slot(s
>  	vaddr = kmap_atomic(area->page);
>  	memcpy(vaddr + offset, uprobe->arch.insn, MAX_UINSN_BYTES);
>  	kunmap_atomic(vaddr);
> +	/*
> +	 * We probably need flush_icache_user_range() but it needs vma.
> +	 * This should work on supported architectures too.
> +	 */
> +	flush_dcache_page(area->page);
> 
>  	return current->utask->xol_vaddr;
>  }