On Mon, Apr 12, 2021 at 01:05:09PM -0700, Kees Cook wrote:
> On Mon, Apr 12, 2021 at 10:00:16AM +0200, Peter Zijlstra wrote:
> > +struct vpr_data {
> > + int (*fn)(pte_t pte, unsigned long addr, void *data);
> > + void *data;
> > +};
>
> Eeerg. This is likely to become an attack target itself. Stored function
> pointer with stored (3rd) argument.
>
> This doesn't seem needed: only DRM uses it, and that's for error
> reporting. I'd rather plumb back errors in a way to not have to add
> another place in the kernel where we do func+arg stored calling.
Is this any better? It does have the stored pointer, but not a stored
argument, assuming you don't count returns as arguments I suppose.
The alternative is refactoring apply_to_page_range() :-/
---
struct vpr_data {
bool (*fn)(pte_t pte, unsigned long addr);
unsigned long addr;
};
static int vpr_fn(pte_t *pte, unsigned long addr, void *data)
{
struct vpr_data *vpr = data;
if (!vpr->fn(*pte, addr)) {
vpr->addr = addr;
return -EINVAL;
}
return 0;
}
/**
* verify_page_range() - Scan (and fill) a range of virtual memory and validate PTEs
* @mm: mm identifying the virtual memory map
* @addr: starting virtual address of the range
* @size: size of the range
* @fn: function that verifies the PTEs
*
* Scan a region of virtual memory, filling in page tables as necessary and
* calling a provided function on each leaf, providing a copy of the
* page-table-entry.
*
* Similar apply_to_page_range(), but does not provide direct access to the
* page-tables.
*
* NOTE! this function does not work correctly vs large pages.
*
* Return: the address that failed verification or 0 on success.
*/
unsigned long verify_page_range(struct mm_struct *mm,
unsigned long addr, unsigned long size,
bool (*fn)(pte_t pte, unsigned long addr))
{
struct vpr_data vpr = {
.fn = fn,
.addr = 0,
};
apply_to_page_range(mm, addr, size, vpr_fn, &vpr);
return vpr.addr;
}
EXPORT_SYMBOL_GPL(verify_page_range);
