Messages in this thread | | | Subject | Re: [PATCH 0/9] [GIT PULL] tracing: updates for 2.6.37 | From | Steven Rostedt <> | Date | Fri, 15 Oct 2010 22:03:29 -0400 |
| |
On Wed, 2010-10-13 at 17:46 -0400, Jason Baron wrote: > On Fri, Oct 08, 2010 at 11:59:47PM -0400, Steven Rostedt wrote: > > On Fri, 2010-10-08 at 10:45 +0200, Ingo Molnar wrote: > > > hm, i'm getting: > > > > > > arch/x86/kernel/alternative.c: Assembler messages: > > > arch/x86/kernel/alternative.c:674: Error: symbol `ftrace_test_jmp' is already defined > > > arch/x86/kernel/alternative.c:678: Error: symbol `ftrace_test_p6nop' is already defined > > > arch/x86/kernel/alternative.c:680: Error: symbol `ftrace_test_nop5' is already defined > > > > > > Config attached. > > > > I'll see if I can get to this early next week. I'll be leaving to NYC on > > Monday for the End Users Summit. I'll try to look at this while there. > > > > Thanks! > > > > -- Steve > > > > just saw this...looks like this is broken for gcc < 4.5. Anyways, the > patch below fixes the compilation by making arch_init_ideal_nop5() > static (and moving its definition ahead of its usage). I'm not quite > sure why this makes the compiler happy...applies on top of this series. >
I can't reproduce this. And I don't like patches that have comments like: "not quite sure why this makes the compiler happy".
Is this still an issue? I tried out gcc 4.4.2, 4.4.4 and gcc 4.5.1 with the config given, and did not get the above errors.
-- Steve
| |