On Mon, Aug 15, 2022 at 02:40:29PM +0200, Peter Zijlstra wrote: > On Mon, Aug 15, 2022 at 12:10:20PM +0200, Jiri Olsa wrote: > > On Mon, Aug 15, 2022 at 11:57:40AM +0200, Peter Zijlstra wrote: > > > On Thu, Aug 11, 2022 at 11:15:21AM +0200, Jiri Olsa wrote: > > > > Adding KPROBE_FLAG_ON_FUNC_ENTRY kprobe flag to indicate that > > > > attach address is on function entry. This is used in following > > > > changes in get_func_ip helper to return correct function address. > > > > > > IIRC (and I've not digested patch) the intent was to have func+0 mean > > > this. x86-IBT is not the only case where this applies, there are > > > multiple architectures where function entry is not +0. > > > > we can have kprobe created by user passing just the address > > > > in this case _kprobe_addr still computes the address's offset > > from the symbol but does not store it back to 'struct kprobe' > > Ah, this is an internal thing to record, in the struct kprobe if the > thing is on the function entry or not? yes, exactly.. it's flag saying the kprobe is on function entry jirka
