Re: [PATCH] x86/ftrace: Get rid of function_hook

2019-10-22 Thread Borislav Petkov
On Tue, Oct 22, 2019 at 01:38:42PM +0200, Jiri Slaby wrote: > On the top of Steven's comment: > Acked-by: Jiri Slaby > > Thanks for taking care of this while my tooth was causing me pain. Sure, np. Toothache is yuck. Ok, so after some back'n'forth yesterday on IRC, I think we agree on only rena

Re: [PATCH] x86/ftrace: Get rid of function_hook

2019-10-22 Thread Jiri Slaby
On 21. 10. 19, 16:10, Borislav Petkov wrote: > From: Borislav Petkov > > function_hook is used as a better name than the default __fentry__ > which is the profiling counter which gcc adds before every function's > prologue. Thus, it is not called from C and cannot have the same > semantics as a p

Re: [PATCH] x86/ftrace: Get rid of function_hook

2019-10-21 Thread Steven Rostedt
On Mon, 21 Oct 2019 16:10:38 +0200 Borislav Petkov wrote: > From: Borislav Petkov > > function_hook is used as a better name than the default __fentry__ > which is the profiling counter which gcc adds before every function's > prologue. Thus, it is not called from C and cannot have the same > s

[PATCH] x86/ftrace: Get rid of function_hook

2019-10-21 Thread Borislav Petkov
From: Borislav Petkov function_hook is used as a better name than the default __fentry__ which is the profiling counter which gcc adds before every function's prologue. Thus, it is not called from C and cannot have the same semantics as a pure C function - it saves/restores regs so that a C funct