Re: [PATCH] x86: remove arbitrary instruction size limit in instruction decoder

[Peter Zijlstra <peterz@infradead.org>]

On Wed, Nov 12, 2014 at 02:53:52PM -0800, Dave Hansen wrote:
> We shouldn't simply error out when we get short copy_from_user*()
> results from userspace (like intel_pmu_pebs_fixup_ip() does
> currently).  It is perfectly valid to be executing an instruction
> within MAX_INSN_SIZE bytes of an unreadable page. We should be
> able to gracefully handle short reads in those cases.

> diff -puN arch/x86/kernel/cpu/perf_event_intel_ds.c~x86-insn-decoder-remove-arbitrary-limit arch/x86/kernel/cpu/perf_event_intel_ds.c
> --- a/arch/x86/kernel/cpu/perf_event_intel_ds.c~x86-insn-decoder-remove-arbitrary-limit	2014-11-12 12:45:52.954753152 -0800
> +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c	2014-11-12 12:45:52.970753874 -0800

> @@ -758,11 +759,12 @@ static int intel_pmu_pebs_fixup_ip(struc
>  		return 1;
>  	}
>  
> +	size = ip - to;
>  	if (!kernel_ip(ip)) {
> -		int size, bytes;
> +		int bytes;
>  		u8 *buf = this_cpu_read(insn_buffer);
>  
> -		size = ip - to; /* Must fit our buffer, see above */
> +		/* 'size' must fit our buffer, see above */
>  		bytes = copy_from_user_nmi(buf, (void __user *)to, size);
>  		if (bytes != 0)
>  			return 0;

Right, so we should better deal with short copies there. Should be
doable.