[PATCH] arm64: Correct virt_addr_valid

[Catalin Marinas]

On Thu, Dec 12, 2013 at 10:09:05PM +0000, Laura Abbott wrote:
> On 12/12/2013 10:02 AM, Russell King - ARM Linux wrote:
> > On Thu, Dec 12, 2013 at 05:57:54PM +0000, Catalin Marinas wrote:
> >> On Wed, Dec 11, 2013 at 09:13:33PM +0000, Russell King - ARM Linux wrote:
> >>> There is actually a concern here, and that's if the v:p translation isn't
> >>> linear, could it return false results?
> >>>
> >>> According to my grep skills, we have one platform where this is true -
> >>> Realview:
> >>>
> >>>   * 256MB @ 0x00000000 -> PAGE_OFFSET
> >>>   * 512MB @ 0x20000000 -> PAGE_OFFSET + 0x10000000
> >>>   * 256MB @ 0x80000000 -> PAGE_OFFSET + 0x30000000
> >>>
> >>> The v:p translation is done via:
> >>>
> >>>           ((virt) >= PAGE_OFFSET2 ? (virt) - PAGE_OFFSET2 + 0x80000000 : \
> >>>            (virt) >= PAGE_OFFSET1 ? (virt) - PAGE_OFFSET1 + 0x20000000 : \
> >>>            (virt) - PAGE_OFFSET)
> >>>
> >>> Now the questions - what do values below PAGE_OFFSET give us?  Very
> >>> large numbers, which pfn_valid() should return false for.  What about
> >>> values > PAGE_OFFSET2 + 256MB?  The same.
> >>>
> >>> So this all _looks_ fine.  Wait a moment, what about highmem?  Let's say
> >>> that the last 256MB is only available as highmem, and let's go back to
> >>> Laura's patch:
> >>>
> >>> old:
> >>> #define	virt_addr_valid(kaddr)	(((void *)(kaddr) >= (void *)PAGE_OFFSET) && \
> >>> 				 ((void *)(kaddr) < (void *)high_memory))
> >>> new:
> >>> #define	virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
> >>>
> >>> The former _excludes_ highmem, but the latter _includes_ it.
> >>>
> >>> virt_addr_valid(v) should only ever return _true_ for the lowmem area,
> >>> never anywhere else - that's part of its point.  It's there to answer
> >>> the question "is this a valid virtual pointer which I can dereference".
> >>>
> >>> So... We actually need a combination of both of these tests.
> >>
> >> Just to avoid any confusion, on arm64 we don't have non-linear v:p
> >> translation as there is plenty of VA space to live with holes. So the
> >> original patch is fine.
> >
> > The point I make above actually has nothing to do with non-linear v:p
> > translations.

OK, I re-read it now.

> Yes, I believe the point was that if we call virt_addr_valid on a 
> not-direct-mapped address it should return false. We still need the 
> range check on arm64 systems as well to ensure this.

On arm64 we don't have highmem, so all RAM would be directly mapped (and
linear). Is there a case on a 64-bit architecture where pfn_valid() is
true but the memory not mapped? We don't unmap any memory which is
pfn_valid().

-- 
Catalin