Re: A question about PROT_NONE on ARM and ARM26

[Jamie Lokier]

Russell King wrote:
> Trust me, it does.  Unless you fully understand how the MMU and domains
> work on ARM, you've little chance of working it out from the code.

Thanks, that's fine.  I just wanted you to confirm PROT_NONE works
with set_fs(KERNEL_DS), as it's not apparent from your earlier
description.  I don't need to know _how_ it works - I can read manuals
too - although you description was interesting.

> > Instead of comparing the address against TI_ADDR_LIMIT, compare it
> > against the hard-coded userspace limit.
> 
> Wrong.  That means that if userspace passes an address above the hard
> coded limit, we _WILL_ bypass all protections and access that memory.

No - it does check against TI_ADDR_LIMIT in the case that the address
is above the hard-coded limit, so prevents that.

The optimisation is valid on all architectures, actually, including
current ARM where it saves a few instructions in the common path.

Here's the potential improvement to current 32-bit ARM.  It's
4 instructions instead of 8 and one less load, in the common case:

__get_user_4:
	cmp	r0, #TASK_SIZE-4
4:	ldrlet	r1, [r0]
	movle	r0, #0
	movle	pc, lr
	bic	r1, sp, #0x1f00
	bic	r1, r1, #0x00ff
	ldr	r1, [r1, #TI_ADDR_LIMIT]
	sub	r1, r1, #4
	cmp	r0, r1
14:	ldrlet	r1, [r0]
	movle	r0, #0
	movle	pc, lr
	b	__get_user_bad

Finally, I think I see a bug in current ARM.  Shouldn't this use
ldrlet instead of ldrlst?  Think about accesses to addresses
TASK_SIZE-4 and 0xfffffffc.

	ldr	r1, [r1, #TI_ADDR_LIMIT]
	sub	r1, r1, #4
	cmp	r0, r1
4:	ldrlst	r1, [r0]

Thanks,
-- Jamie
-
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/