On 21/04/17 19:24, Dave Martin wrote:

On Thu, Apr 20, 2017 at 07:17:13PM +0100, Kristina Martsenko wrote:

quoted

When handling a data abort from EL0, we currently zero the top byte of
the faulting address, as we assume the address is a TTBR0 address, which
may contain a non-zero address tag. However, the address may be a TTBR1
address, in which case we should not zero the top byte. This patch fixes
that. The effect is that the full TTBR1 address is passed to the task's
signal handler (or printed out in the kernel log).

When handling a data abort from EL1, we leave the faulting address
intact, as we assume it's either a TTBR1 address or a TTBR0 address with
tag 0x00. This is true as far as I'm aware, we don't seem to access a
tagged TTBR0 address anywhere in the kernel. Regardless, it's easy to
forget about address tags, and code added in the future may not always
remember to remove tags from addresses before accessing them. So add tag
handling to the EL1 data abort handler as well. This also makes it
consistent with the EL0 data abort handler.

Possibly it doesn't matter whether the tag bits are cleared for an EL0
fault on a TTBR1 address, since userspace can't have a valid pointer in
this range to (mis)match the fault address against ... or did I miss
something?

I don't think you've missed anything. But I don't see why userspace
can't match against an invalid (TTBR1) address, I think that would be a
valid thing to do (even if unlikely).

Factoring out the tag handling makes the intent of the code clearer
though, either way.

I assume this means you're fine with the patch as is.

Thanks,
Kristina