Re: [Qemu-devel] [PATCH] mmap: add check if requested memory area fits target address space

["Kirill A. Shutemov" <kirill@shutemov.name>]

[-- Attachment #1: Type: text/plain, Size: 1775 bytes --]

On Mon, Oct 27, 2008 at 08:37:39PM +0100, andrzej zaborowski wrote:
> 2008/10/27 Kirill A. Shutemov <kirill@shutemov.name>:
> > On Mon, Oct 27, 2008 at 02:08:52PM +0100, andrzej zaborowski wrote:
> >> On 17/10/2008, Kirill A. Shutemov <kirill@shutemov.name> wrote:
> >> > Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> >> >  ---
> >> >   linux-user/mmap.c |    5 +++++
> >> >   1 files changed, 5 insertions(+), 0 deletions(-)
> >> >
> >> >  diff --git a/linux-user/mmap.c b/linux-user/mmap.c
> >> >  index bc20f4b..9a2f355 100644
> >> >  --- a/linux-user/mmap.c
> >> >  +++ b/linux-user/mmap.c
> >> >  @@ -388,6 +388,11 @@ abi_long target_mmap(abi_ulong start, abi_ulong len, int prot,
> >> >          end = start + len;
> >> >          real_end = HOST_PAGE_ALIGN(end);
> >> >
> >> >  +        if ((unsigned long)start + len > (abi_ulong) -1) {
> >> >  +            errno = EINVAL;
> >> >  +            goto fail;
> >> >  +        }
> >>
> >> I'm being picky but this would prevent the last byte from being used?
> >> :p  (or the last page because len is aligned?)
> >
> > No, it returns error if start + len is more than 0xFFFFFFFF (32-bit
> > target).
> >
> >>
> >> I'm not sure unsigned long is the best choice.
> >
> > Why?
> 
> I may be misunderstanding but I think the range of valid addresses
> should depend on target word size, not host (even if the combination
> where it matters is not yet supported).

start + len can be more than 0xFFFFFFFF ((abi_ulong) -1) on 32-bit targets,
so we should use host's long.

> On a 32-bit host the condition is always false.

It's ok. It can be true, only on 64-bit host.

-- 
Regards,  Kirill A. Shutemov
 + Belarus, Minsk
 + ALT Linux Team, http://www.altlinux.com/

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]