Re: [PATCH v2 net-next 4/5] bpf/verifier: remove varlen_map_value_access flag

2017-08-23 Thread Daniel Borkmann
On 08/23/2017 04:10 PM, Edward Cree wrote: The optimisation it does is broken when the 'new' register value has a variable offset and the 'old' was constant. I broke it with my pointer types unification (see Fixes tag below), before which the 'new' value would have type

Re: [PATCH v2 net-next 4/5] bpf/verifier: remove varlen_map_value_access flag

2017-08-23 Thread Alexei Starovoitov
On 8/23/17 7:10 AM, Edward Cree wrote: The optimisation it does is broken when the 'new' register value has a variable offset and the 'old' was constant. I broke it with my pointer types unification (see Fixes tag below), before which the 'new' value would have type PTR_TO_MAP_VALUE_ADJ and

[PATCH v2 net-next 4/5] bpf/verifier: remove varlen_map_value_access flag

2017-08-23 Thread Edward Cree
The optimisation it does is broken when the 'new' register value has a variable offset and the 'old' was constant. I broke it with my pointer types unification (see Fixes tag below), before which the 'new' value would have type PTR_TO_MAP_VALUE_ADJ and would thus not compare equal; other