Re: [PATCH 1/2] perf/x86/intel/ds: Work around BTS leaking kernel addresses

From: Ingo Molnar
Date: Fri Aug 28 2015 - 01:30:45 EST



* Alexander Shishkin <alexander.shishkin@xxxxxxxxxxxxxxx> wrote:

> + for (at = base; at < top; at++) {
> + /*
> + * Note that right now *this* BTS code only works if
> + * attr::exclude_kernel is set, but let's keep this extra
> + * check here in case that changes.
> + */
> + if (event->attr.exclude_kernel &&
> + (at->from >= PAGE_OFFSET || at->to >= PAGE_OFFSET))
> + skip++;

Yeah, so that only works on 32-bit kernels, on 64-bit kernels the check for kernel
addresses is to see whether it's a negative address. PAGE_OFFSET points to above
any hypervisor's address, so even with your fix we could still leak hypervisor
addresses.

I.e. use the kernel_ip() primitive instead.

Thanks,

Ingo
--
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/