2008-11-18 16:30:26

by Masami Hiramatsu

Subject: [PATCH -mm] bugfix: pass aggr_kprobe to arch_remove_kprobe

Hi Andrew,

[email protected] wrote:
> The patch titled
> kprobes: support probing module __exit function
> has been added to the -mm tree. Its filename is
> kprobes-support-probing-module-__exit-function.patch

I found that one bug was still alive. Here is the bugfix.

Thank you,

Call arch_remove_kprobe() with aggr_kprobe instead of user specific kprobe,
because the user specific kprobe on the gone or reused aggr_kprobe may have
invalid arch_specific_insn.

Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
---
kernel/kprobes.c | 14 +++++++-------
1 file changed, 7 insertions(+), 7 deletions(-)

Index: 2.6-rc/kernel/kprobes.c
===================================================================
--- 2.6-rc.orig/kernel/kprobes.c
+++ 2.6-rc/kernel/kprobes.c
@@ -766,14 +766,14 @@ static void __kprobes __unregister_kprob
{
struct kprobe *old_p;

- if (list_empty(&p->list) || list_is_singular(&p->list)) {
- if (!list_empty(&p->list)) {
- /* "p" is the last child of an aggr_kprobe */
- old_p = list_entry(p->list.next, struct kprobe, list);
- list_del(&p->list);
- kfree(old_p);
- }
+ if (list_empty(&p->list))
arch_remove_kprobe(p);
+ else if (list_is_singular(&p->list)) {
+ /* "p" is the last child of an aggr_kprobe */
+ old_p = list_entry(p->list.next, struct kprobe, list);
+ list_del(&p->list);
+ arch_remove_kprobe(old_p);
+ kfree(old_p);
}
}


--
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: [email protected]

2008-11-19 04:32:39

by Ananth N Mavinakayanahalli

Subject: Re: [PATCH -mm] bugfix: pass aggr_kprobe to arch_remove_kprobe

On Tue, Nov 18, 2008 at 11:29:01AM -0500, Masami Hiramatsu wrote:
> Hi Andrew,
>
> [email protected] wrote:
> > The patch titled
> > kprobes: support probing module __exit function
> > has been added to the -mm tree. Its filename is
> > kprobes-support-probing-module-__exit-function.patch
>
> I found that one bug was still alive. Here is the bugfix.
>
> Thank you,
>
> Call arch_remove_kprobe() with aggr_kprobe instead of user specific kprobe,
> because the user specific kprobe on the gone or reused aggr_kprobe may have
> invalid arch_specific_insn.
>
> Signed-off-by: Masami Hiramatsu <[email protected]>
> Cc: Ananth N Mavinakayanahalli <[email protected]>

Acked-by: Ananth N Mavinakayanahalli <[email protected]>

> ---
> kernel/kprobes.c | 14 +++++++-------
> 1 file changed, 7 insertions(+), 7 deletions(-)
>
> Index: 2.6-rc/kernel/kprobes.c
> ===================================================================
> --- 2.6-rc.orig/kernel/kprobes.c
> +++ 2.6-rc/kernel/kprobes.c
> @@ -766,14 +766,14 @@ static void __kprobes __unregister_kprob
> {
> struct kprobe *old_p;
>
> - if (list_empty(&p->list) || list_is_singular(&p->list)) {
> - if (!list_empty(&p->list)) {
> - /* "p" is the last child of an aggr_kprobe */
> - old_p = list_entry(p->list.next, struct kprobe, list);
> - list_del(&p->list);
> - kfree(old_p);
> - }
> + if (list_empty(&p->list))
> arch_remove_kprobe(p);
> + else if (list_is_singular(&p->list)) {
> + /* "p" is the last child of an aggr_kprobe */
> + old_p = list_entry(p->list.next, struct kprobe, list);
> + list_del(&p->list);
> + arch_remove_kprobe(old_p);
> + kfree(old_p);
> }
> }
>
>
> --
> Masami Hiramatsu
>
> Software Engineer
> Hitachi Computer Products (America) Inc.
> Software Solutions Division
>
> e-mail: [email protected]
>