Re: Testing PROT_NONE and other protections, and a surprise

[William Lee Irwin III]

William Lee Irwin III wrote:
>> It would be a bug if copy_to_user()/copy_from_user() failed to return
>> errors on attempted copies to/from areas with PROT_NONE protection.
>> I recommend writing a testcase and submitting it to LTP. I'll follow up
>> with an additional suggestion.

On Thu, Jul 01, 2004 at 04:26:06AM +0100, Jamie Lokier wrote:
> I've just written a thorough test.  The attached program tries every
> combination of PROT_* flags, and tells you what protection you really get.
> I don't know how tests get into LTP; perhaps I can leave that to you?
> When running it on i386, I got a *huge* surprise (to me).  A
> PROT_WRITE-only page can sometimes fault on read or exec.  This is the
> output:

This is unsurprising. The permissions can't be represented in pagetables,
but can opportunistically be enforced when exceptions are taken for other
reasons (e.g. TLB invalidations related to page replacement).


-- wli
-
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/