On Sat, Nov 30, 2013 at 08:19:32PM +0100, Pavel Machek wrote:
> so perhaps Nico Golde or Dan Carpenter can elaborate? I Cc-ed them
> now.
>
> Or is it some kind of super-secret issue and still under embargo for
> 10 days?
> Pavel
Nope, it's not secret. Here is the original report from Nico and
Fabian. Please give them credit when you fix the bug.
Reported-by: Nico Golde <nico@xxxxxxxxx>
Reported-by: Fabian Yamaguchi <fabs@xxxxxxxxx>
Felipe Contreras says we could just remove mmap() support and the driver
would still work, but no one has submitted a patch to do that. Really
this driver needs an actual maintainer who will respond to security bugs
and also start to move the driver out of staging.
regards,
dan carpenter
---- Bug Report ----
source file: drivers/staging/tidspbridge/rmgr/drv_interface.c
issue : mapping of physical memory without address range checks
259 static int bridge_mmap(struct file *filp, struct vm_area_struct *vma)
260 {
261 u32 status;
262
263 /* VM_IO | VM_DONTEXPAND | VM_DONTDUMP are set by remap_pfn_range() */
264 vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
265
266 dev_dbg(bridge, "%s: vm filp %p start %lx end %lx page_prot %ulx "
267 "flags %lx\n", __func__, filp,
268 vma->vm_start, vma->vm_end, vma->vm_page_prot,
269 vma->vm_flags);
270
271 status = remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
272 vma->vm_end - vma->vm_start,
273 vma->vm_page_prot);
274 if (status != 0)
275 status = -EAGAIN;
276
277 return status;
278 }
The function provides an interface to remap physical memory to user space, but
does not provide any checks to ensure that the memory is within the region that
should be accessible.
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html