Messages in this thread | | | Subject | Re: [PATCH v2 net-next 1/2] lib: introduce strncpy_from_unsafe() | From | Alexei Starovoitov <> | Date | Fri, 28 Aug 2015 15:08:35 -0700 |
| |
On 8/28/15 2:48 PM, Steven Rostedt wrote: >> * On success, returns the length of the string (not including the trailing >> >+ * NUL). > I think it includes the NUL.
oops. yes. that was a copy paste from strncpy_from_user comment. trace_kprobe usage wants NUL to be counted, so I intended to have it counted, but that brings the question what should be the semantics. Should it be similar to strncpy_from_user (not counting NUL) or similar to strlen_user (counts NUL) ? imo counting NUL makes a little bit more sense, since when a user says strncpy_from_unsafe(..., ..., 32) and it returns 32 as the whole buffer was filled, it looks cleaner. Thoughts?
| |