Re: [PATCH net-next] bpf: add support for %s specifier to bpf_trace_printk()

[Alexei Starovoitov <ast@plumgrid.com>]

On 8/27/15 3:34 PM, David Miller wrote:
> From: Alexei Starovoitov <ast@plumgrid.com>
> Date: Wed, 26 Aug 2015 23:26:59 -0700
>
>> +/* similar to strncpy_from_user() but with extra checks */
>> +static void probe_read_string(char *buf, int size, long unsafe_ptr)
>> +{
>> +	char dst[4];
>> +	int i = 0;
>> +
>> +	size--;
>> +	for (;;) {
>> +		if (probe_kernel_read(dst, (void *) unsafe_ptr, 4))
>> +			break;
>
> I don't think this does the right thing when the string is not a multiple
> of 3 and ends at the last byte of a page that ends a valid region of
> kernel memory.
>
> Seeing this kind of error makes me skeptical to the overall value of
> optimizing this :-/

I've considered the case when first two bytes are valid, but the
other two are in a different page. In such case the probe_read_string()
will trim the string and won't be printing these two valid bytes.
I think that's very rare, so I'm picking higher performance for common
case. The strings over > 64 bytes also will be trimmed to 64.
It's a debugging facility, so I felt that's ok.
Fair or you still think it should be per byte copy?