Messages in this thread | | | Date | Wed, 29 Aug 2012 17:08:51 +0900 | From | Masami Hiramatsu <> | Subject | Re: BUG: Kprobe smoke test: 2 out of 6 tests failed |
| |
Hi Steven and Fengguang,
(2012/08/25 10:00), Steven Rostedt wrote: > On Fri, 2012-08-24 at 09:41 -0400, Steven Rostedt wrote: > >> On Fri, 2012-08-24 at 15:15 +0800, Fengguang Wu wrote: >>> Hi Steven, >>> >>> The following test fails are mostly due to this commit, or one of the >>> last 4 commits in >>> >>> tree: git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git tip/perf/core >>> head: d57c5d51a30152f3175d2344cb6395f08bf8ee0c >>> commit: d57c5d51a30152f3175d2344cb6395f08bf8ee0c [100/100] ftrace/x86: Add support for -mfentry to x86_64 >>> >>> [ 9.084881] Kprobe smoke test failed: register_jprobe returned -22 >>> [ 9.086786] Kprobe smoke test failed: register_jprobes returned -22 > > Masami, > > Seems that when we use fentry, we break jprobes. I thought the patches > that we added would just move the call to the next op, not fail totally?
Ah! yes, it would be kprobe-side problem.
Since ftrace_regs_call allows handlers to set regs->ip, it is not necessarily to move jprobe onto the next instruction (even it will be the simplest one). Instead, I think we can allow it on ftrace entries. Of course, ftrace side change is also needed, e.g. sorting handlers for moving such ip-modifying handler to the last.
Thank you,
-- Masami HIRAMATSU Software Platform Research Dept. Linux Technology Center Hitachi, Ltd., Yokohama Research Laboratory E-mail: masami.hiramatsu.pt@hitachi.com
| |