* [PATCH] kprobes: bugfix: force unoptimize when disable kprobes.
@ 2015-01-05 12:32 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-05 12:32 UTC (permalink / raw)
To: masami.hiramatsu.pt, tixy, linux; +Cc: linux-kernel, linux-arm-kernel, lizefan
Original code failed to disarm the probed instruction after
echo 0 > /sys/kernel/debug/kprobes/enabled
if OPTPROBE is enabled.
This is caused by a piece of logically inconsistent code:
unoptimize_kprobe(p, false);
if (!kprobe_queued(p)) {
...
}
unoptimize_kprobe() with 'force' == false queues p onto
unoptimizing_list, so following kprobe_queued() check always fail unless
another core schedules optimizer and does the unoptimization very soon.
This logic causes arch_disarm_kprobe() failed to get execute, lefts a
breakpoint at the probed address, instead of restoring it.
This patch uses force unoptimize instead.
Signed-off-by: Wang Nan <wangnan0@huawei.com>
---
kernel/kprobes.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index b185464..9fbe0c3 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -869,7 +869,7 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
{
struct kprobe *_p;
- unoptimize_kprobe(p, false); /* Try to unoptimize */
+ unoptimize_kprobe(p, true); /* Try to unoptimize */
if (!kprobe_queued(p)) {
arch_disarm_kprobe(p);
--
1.8.4
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: force unoptimize when disable kprobes.
@ 2015-01-05 12:32 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-05 12:32 UTC (permalink / raw)
To: linux-arm-kernel
Original code failed to disarm the probed instruction after
echo 0 > /sys/kernel/debug/kprobes/enabled
if OPTPROBE is enabled.
This is caused by a piece of logically inconsistent code:
unoptimize_kprobe(p, false);
if (!kprobe_queued(p)) {
...
}
unoptimize_kprobe() with 'force' == false queues p onto
unoptimizing_list, so following kprobe_queued() check always fail unless
another core schedules optimizer and does the unoptimization very soon.
This logic causes arch_disarm_kprobe() failed to get execute, lefts a
breakpoint at the probed address, instead of restoring it.
This patch uses force unoptimize instead.
Signed-off-by: Wang Nan <wangnan0@huawei.com>
---
kernel/kprobes.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index b185464..9fbe0c3 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -869,7 +869,7 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
{
struct kprobe *_p;
- unoptimize_kprobe(p, false); /* Try to unoptimize */
+ unoptimize_kprobe(p, true); /* Try to unoptimize */
if (!kprobe_queued(p)) {
arch_disarm_kprobe(p);
--
1.8.4
^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: force unoptimize when disable kprobes.
2015-01-05 12:32 ` Wang Nan
@ 2015-01-12 11:42 ` Masami Hiramatsu
-1 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-12 11:42 UTC (permalink / raw)
To: Wang Nan; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
(2015/01/05 21:32), Wang Nan wrote:
> Original code failed to disarm the probed instruction after
>
> echo 0 > /sys/kernel/debug/kprobes/enabled
>
> if OPTPROBE is enabled.
>
> This is caused by a piece of logically inconsistent code:
>
> unoptimize_kprobe(p, false);
> if (!kprobe_queued(p)) {
> ...
> }
>
> unoptimize_kprobe() with 'force' == false queues p onto
> unoptimizing_list, so following kprobe_queued() check always fail unless
> another core schedules optimizer and does the unoptimization very soon.
> This logic causes arch_disarm_kprobe() failed to get execute, lefts a
> breakpoint at the probed address, instead of restoring it.
No, the root cause of this failure comes from the lack of checking
kprobes_all_disarmed in unoptimized_kprobe(). It should check the flag
and return soon if it is set.
So, I Nak this patch.
>
> This patch uses force unoptimize instead.
>
> Signed-off-by: Wang Nan <wangnan0@huawei.com>
> ---
> kernel/kprobes.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index b185464..9fbe0c3 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -869,7 +869,7 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
> {
> struct kprobe *_p;
>
> - unoptimize_kprobe(p, false); /* Try to unoptimize */
> + unoptimize_kprobe(p, true); /* Try to unoptimize */
>
> if (!kprobe_queued(p)) {
> arch_disarm_kprobe(p);
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: force unoptimize when disable kprobes.
@ 2015-01-12 11:42 ` Masami Hiramatsu
0 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-12 11:42 UTC (permalink / raw)
To: linux-arm-kernel
(2015/01/05 21:32), Wang Nan wrote:
> Original code failed to disarm the probed instruction after
>
> echo 0 > /sys/kernel/debug/kprobes/enabled
>
> if OPTPROBE is enabled.
>
> This is caused by a piece of logically inconsistent code:
>
> unoptimize_kprobe(p, false);
> if (!kprobe_queued(p)) {
> ...
> }
>
> unoptimize_kprobe() with 'force' == false queues p onto
> unoptimizing_list, so following kprobe_queued() check always fail unless
> another core schedules optimizer and does the unoptimization very soon.
> This logic causes arch_disarm_kprobe() failed to get execute, lefts a
> breakpoint at the probed address, instead of restoring it.
No, the root cause of this failure comes from the lack of checking
kprobes_all_disarmed in unoptimized_kprobe(). It should check the flag
and return soon if it is set.
So, I Nak this patch.
>
> This patch uses force unoptimize instead.
>
> Signed-off-by: Wang Nan <wangnan0@huawei.com>
> ---
> kernel/kprobes.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index b185464..9fbe0c3 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -869,7 +869,7 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
> {
> struct kprobe *_p;
>
> - unoptimize_kprobe(p, false); /* Try to unoptimize */
> + unoptimize_kprobe(p, true); /* Try to unoptimize */
>
> if (!kprobe_queued(p)) {
> arch_disarm_kprobe(p);
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-12 11:42 ` Masami Hiramatsu
@ 2015-01-12 12:09 ` Wang Nan
-1 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-12 12:09 UTC (permalink / raw)
To: masami.hiramatsu.pt; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
Original code failed to disarm the probed instruction after
echo 0 > /sys/kernel/debug/kprobes/enabled
if OPTPROBE is enabled.
This patch checks kprobes_all_disarmed in unoptimized_kprobe().
Signed-off-by: Wang Nan <wangnan0@huawei.com>
---
kernel/kprobes.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 9471710..f16936b 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
{
struct optimized_kprobe *op;
+ if (kprobes_all_disarmed)
+ return;
+
if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
return; /* This is not an optprobe nor optimized */
--
1.8.4
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-12 12:09 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-12 12:09 UTC (permalink / raw)
To: linux-arm-kernel
Original code failed to disarm the probed instruction after
echo 0 > /sys/kernel/debug/kprobes/enabled
if OPTPROBE is enabled.
This patch checks kprobes_all_disarmed in unoptimized_kprobe().
Signed-off-by: Wang Nan <wangnan0@huawei.com>
---
kernel/kprobes.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 9471710..f16936b 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
{
struct optimized_kprobe *op;
+ if (kprobes_all_disarmed)
+ return;
+
if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
return; /* This is not an optprobe nor optimized */
--
1.8.4
^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-12 12:09 ` Wang Nan
@ 2015-01-12 12:52 ` Masami Hiramatsu
-1 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-12 12:52 UTC (permalink / raw)
To: Wang Nan; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
(2015/01/12 21:09), Wang Nan wrote:
> Original code failed to disarm the probed instruction after
>
> echo 0 > /sys/kernel/debug/kprobes/enabled
>
> if OPTPROBE is enabled.
>
> This patch checks kprobes_all_disarmed in unoptimized_kprobe().
>
Looks good :)
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Thank you!
> Signed-off-by: Wang Nan <wangnan0@huawei.com>
> ---
> kernel/kprobes.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 9471710..f16936b 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
> {
> struct optimized_kprobe *op;
>
> + if (kprobes_all_disarmed)
> + return;
> +
> if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
> return; /* This is not an optprobe nor optimized */
>
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-12 12:52 ` Masami Hiramatsu
0 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-12 12:52 UTC (permalink / raw)
To: linux-arm-kernel
(2015/01/12 21:09), Wang Nan wrote:
> Original code failed to disarm the probed instruction after
>
> echo 0 > /sys/kernel/debug/kprobes/enabled
>
> if OPTPROBE is enabled.
>
> This patch checks kprobes_all_disarmed in unoptimized_kprobe().
>
Looks good :)
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Thank you!
> Signed-off-by: Wang Nan <wangnan0@huawei.com>
> ---
> kernel/kprobes.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 9471710..f16936b 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
> {
> struct optimized_kprobe *op;
>
> + if (kprobes_all_disarmed)
> + return;
> +
> if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
> return; /* This is not an optprobe nor optimized */
>
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-12 12:52 ` Masami Hiramatsu
@ 2015-01-19 3:04 ` Wang Nan
-1 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-19 3:04 UTC (permalink / raw)
To: Masami Hiramatsu; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
Hi Masami Hiramatsu,
I can't find this patch and '[PATCH] kprobes: bugfix: checks kprobes_all_disarmed
in unoptimized_kprobe().' in current mainline. How do these patches get there?
Should they be merged into Russell King's tree first?
Thank you!
On 2015/1/12 20:52, Masami Hiramatsu wrote:
> (2015/01/12 21:09), Wang Nan wrote:
>> Original code failed to disarm the probed instruction after
>>
>> echo 0 > /sys/kernel/debug/kprobes/enabled
>>
>> if OPTPROBE is enabled.
>>
>> This patch checks kprobes_all_disarmed in unoptimized_kprobe().
>>
>
> Looks good :)
>
> Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>
> Thank you!
>
>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>> ---
>> kernel/kprobes.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
>> index 9471710..f16936b 100644
>> --- a/kernel/kprobes.c
>> +++ b/kernel/kprobes.c
>> @@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
>> {
>> struct optimized_kprobe *op;
>>
>> + if (kprobes_all_disarmed)
>> + return;
>> +
>> if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
>> return; /* This is not an optprobe nor optimized */
>>
>>
>
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-19 3:04 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-19 3:04 UTC (permalink / raw)
To: linux-arm-kernel
Hi Masami Hiramatsu,
I can't find this patch and '[PATCH] kprobes: bugfix: checks kprobes_all_disarmed
in unoptimized_kprobe().' in current mainline. How do these patches get there?
Should they be merged into Russell King's tree first?
Thank you!
On 2015/1/12 20:52, Masami Hiramatsu wrote:
> (2015/01/12 21:09), Wang Nan wrote:
>> Original code failed to disarm the probed instruction after
>>
>> echo 0 > /sys/kernel/debug/kprobes/enabled
>>
>> if OPTPROBE is enabled.
>>
>> This patch checks kprobes_all_disarmed in unoptimized_kprobe().
>>
>
> Looks good :)
>
> Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>
> Thank you!
>
>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>> ---
>> kernel/kprobes.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
>> index 9471710..f16936b 100644
>> --- a/kernel/kprobes.c
>> +++ b/kernel/kprobes.c
>> @@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
>> {
>> struct optimized_kprobe *op;
>>
>> + if (kprobes_all_disarmed)
>> + return;
>> +
>> if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
>> return; /* This is not an optprobe nor optimized */
>>
>>
>
>
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-19 3:04 ` Wang Nan
@ 2015-01-19 9:05 ` Masami Hiramatsu
-1 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-19 9:05 UTC (permalink / raw)
To: Wang Nan; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
Hi Wang,
I've found a problem on this patch, since kprobes calls unoptioize_kprobe
with kprobes_all_disarmed=true when trying to disable all kprobes, this
cause a serious problem.
Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
Could you test it again?
Unless I could reproduce this bug, I'd like to keep this uncommitted.
Thank you,
(2015/01/19 12:04), Wang Nan wrote:
> Hi Masami Hiramatsu,
>
> I can't find this patch and '[PATCH] kprobes: bugfix: checks kprobes_all_disarmed
> in unoptimized_kprobe().' in current mainline. How do these patches get there?
> Should they be merged into Russell King's tree first?
>
> Thank you!
>
> On 2015/1/12 20:52, Masami Hiramatsu wrote:
>> (2015/01/12 21:09), Wang Nan wrote:
>>> Original code failed to disarm the probed instruction after
>>>
>>> echo 0 > /sys/kernel/debug/kprobes/enabled
>>>
>>> if OPTPROBE is enabled.
>>>
>>> This patch checks kprobes_all_disarmed in unoptimized_kprobe().
>>>
>>
>> Looks good :)
>>
>> Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>>
>> Thank you!
>>
>>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>>> ---
>>> kernel/kprobes.c | 3 +++
>>> 1 file changed, 3 insertions(+)
>>>
>>> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
>>> index 9471710..f16936b 100644
>>> --- a/kernel/kprobes.c
>>> +++ b/kernel/kprobes.c
>>> @@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
>>> {
>>> struct optimized_kprobe *op;
>>>
>>> + if (kprobes_all_disarmed)
>>> + return;
>>> +
>>> if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
>>> return; /* This is not an optprobe nor optimized */
>>>
>>>
>>
>>
>
>
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-19 9:05 ` Masami Hiramatsu
0 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-19 9:05 UTC (permalink / raw)
To: linux-arm-kernel
Hi Wang,
I've found a problem on this patch, since kprobes calls unoptioize_kprobe
with kprobes_all_disarmed=true when trying to disable all kprobes, this
cause a serious problem.
Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
Could you test it again?
Unless I could reproduce this bug, I'd like to keep this uncommitted.
Thank you,
(2015/01/19 12:04), Wang Nan wrote:
> Hi Masami Hiramatsu,
>
> I can't find this patch and '[PATCH] kprobes: bugfix: checks kprobes_all_disarmed
> in unoptimized_kprobe().' in current mainline. How do these patches get there?
> Should they be merged into Russell King's tree first?
>
> Thank you!
>
> On 2015/1/12 20:52, Masami Hiramatsu wrote:
>> (2015/01/12 21:09), Wang Nan wrote:
>>> Original code failed to disarm the probed instruction after
>>>
>>> echo 0 > /sys/kernel/debug/kprobes/enabled
>>>
>>> if OPTPROBE is enabled.
>>>
>>> This patch checks kprobes_all_disarmed in unoptimized_kprobe().
>>>
>>
>> Looks good :)
>>
>> Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
>>
>> Thank you!
>>
>>> Signed-off-by: Wang Nan <wangnan0@huawei.com>
>>> ---
>>> kernel/kprobes.c | 3 +++
>>> 1 file changed, 3 insertions(+)
>>>
>>> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
>>> index 9471710..f16936b 100644
>>> --- a/kernel/kprobes.c
>>> +++ b/kernel/kprobes.c
>>> @@ -630,6 +630,9 @@ static void unoptimize_kprobe(struct kprobe *p, bool force)
>>> {
>>> struct optimized_kprobe *op;
>>>
>>> + if (kprobes_all_disarmed)
>>> + return;
>>> +
>>> if (!kprobe_aggrprobe(p) || kprobe_disarmed(p))
>>> return; /* This is not an optprobe nor optimized */
>>>
>>>
>>
>>
>
>
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-19 9:05 ` Masami Hiramatsu
@ 2015-01-19 11:21 ` Wang Nan
-1 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-19 11:21 UTC (permalink / raw)
To: Masami Hiramatsu; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
On 2015/1/19 17:05, Masami Hiramatsu wrote:
> Hi Wang,
>
> I've found a problem on this patch, since kprobes calls unoptioize_kprobe
> with kprobes_all_disarmed=true when trying to disable all kprobes, this
> cause a serious problem.
>
> Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
> Could you test it again?
>
I tested it again based on 3.19-rc5 and found that the problem still exists.
My testing is based on QEMU.
First I tested my kprobeopt for ARM, then on x86_64. The test results are pasted
at the bottom of this mail. Commands after 'gdb attaches to QEMU' is my actions
on a gdb console attached to QEMU; commands after 'inside virtual machine' is
what I do in Linux run under QEMU.
** ARM result **
------ gdb attaches to QEMU -------
(gdb) x/3i sys_open
0xc013619c <SyS_open>: mov r12, sp <--- *original insn*
0xc01361a0 <SyS_open+4>: push {r11, r12, lr, pc}
0xc01361a4 <SyS_open+8>: sub r11, r12, #4
------ inside virtual machine -------
# echo 'p:myprobe sys_open' > /sys/kernel/debug/tracing/kprobe_events
# echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
------ gdb attaches to QEMU -------
cpu_v7_do_idle () at /home/w00229757/kernel-hydrogen/arch/arm/mm/proc-v7.S:74
74 ret lr
(gdb) x/3i sys_open
0xc013619c <SyS_open>: b 0xbf000000 <--- *optimized*
0xc01361a0 <SyS_open+4>: push {r11, r12, lr, pc}
0xc01361a4 <SyS_open+8>: sub r11, r12, #4
(gdb) c
------ inside virtual machine -------
# echo 0 > /sys/kernel/debug/kprobes/enabled
------ gdb attaches to QEMU -------
cpu_v7_do_idle () at /home/w00229757/kernel-hydrogen/arch/arm/mm/proc-v7.S:74
74 ret lr
(gdb) x/3i sys_open
0xc013619c <SyS_open>: ; <UNDEFINED> instruction: 0xe7f001f8 <--- *BREAKPOINT*
0xc01361a0 <SyS_open+4>: push {r11, r12, lr, pc}
0xc01361a4 <SyS_open+8>: sub r11, r12, #4
(gdb) c
** x86_64 result **
------ gdb attaches to QEMU -------
(gdb) x/10i sys_open
0xffffffff81184fe0 <SyS_open>: data32 data32 data32 xchg %ax,%ax
0xffffffff81184fe5 <SyS_open+5>: push %rbp
0xffffffff81184fe6 <SyS_open+6>: movzwl %dx,%ecx
0xffffffff81184fe9 <SyS_open+9>: mov %esi,%edx
0xffffffff81184feb <SyS_open+11>: mov %rsp,%rbp
0xffffffff81184fee <SyS_open+14>: mov %rdi,%rsi
0xffffffff81184ff1 <SyS_open+17>: or $0x80,%dh
0xffffffff81184ff4 <SyS_open+20>: mov $0xffffff9c,%edi
0xffffffff81184ff9 <SyS_open+25>: callq 0xffffffff81184da0 <do_sys_open>
0xffffffff81184ffe <SyS_open+30>: pop %rbp
(gdb) c
Continuing
------ inside virtual machine -------
# echo 'p:myprobe sys_open+20' > /sys/kernel/debug/tracing/kprobe_events
# echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
------ gdb attaches to QEMU -------
(gdb) x/10i sys_open
0xffffffff81184fe0 <SyS_open>: data32 data32 data32 xchg %ax,%ax
0xffffffff81184fe5 <SyS_open+5>: push %rbp
0xffffffff81184fe6 <SyS_open+6>: movzwl %dx,%ecx
0xffffffff81184fe9 <SyS_open+9>: mov %esi,%edx
0xffffffff81184feb <SyS_open+11>: mov %rsp,%rbp
0xffffffff81184fee <SyS_open+14>: mov %rdi,%rsi
0xffffffff81184ff1 <SyS_open+17>: or $0x80,%dh
0xffffffff81184ff4 <SyS_open+20>: jmpq 0xffffffffa0002000 <--- *optimized*
0xffffffff81184ff9 <SyS_open+25>: callq 0xffffffff81184da0 <do_sys_open>
0xffffffff81184ffe <SyS_open+30>: pop %rbp
(gdb) c
Continuing.
------ inside virtual machine -------
# echo 0 > /sys/kernel/debug/kprobes/enabled
------ gdb attaches to QEMU -------
(gdb) x/10i sys_open
0xffffffff81184fe0 <SyS_open>: data32 data32 data32 xchg %ax,%ax
0xffffffff81184fe5 <SyS_open+5>: push %rbp
0xffffffff81184fe6 <SyS_open+6>: movzwl %dx,%ecx
0xffffffff81184fe9 <SyS_open+9>: mov %esi,%edx
0xffffffff81184feb <SyS_open+11>: mov %rsp,%rbp
0xffffffff81184fee <SyS_open+14>: mov %rdi,%rsi
0xffffffff81184ff1 <SyS_open+17>: or $0x80,%dh
0xffffffff81184ff4 <SyS_open+20>: int3 <-- **BREAKPOINT**
0xffffffff81184ff5 <SyS_open+21>: pushfq
0xffffffff81184ff6 <SyS_open+22>: (bad)
(gdb)
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-19 11:21 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-19 11:21 UTC (permalink / raw)
To: linux-arm-kernel
On 2015/1/19 17:05, Masami Hiramatsu wrote:
> Hi Wang,
>
> I've found a problem on this patch, since kprobes calls unoptioize_kprobe
> with kprobes_all_disarmed=true when trying to disable all kprobes, this
> cause a serious problem.
>
> Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
> Could you test it again?
>
I tested it again based on 3.19-rc5 and found that the problem still exists.
My testing is based on QEMU.
First I tested my kprobeopt for ARM, then on x86_64. The test results are pasted
at the bottom of this mail. Commands after 'gdb attaches to QEMU' is my actions
on a gdb console attached to QEMU; commands after 'inside virtual machine' is
what I do in Linux run under QEMU.
** ARM result **
------ gdb attaches to QEMU -------
(gdb) x/3i sys_open
0xc013619c <SyS_open>: mov r12, sp <--- *original insn*
0xc01361a0 <SyS_open+4>: push {r11, r12, lr, pc}
0xc01361a4 <SyS_open+8>: sub r11, r12, #4
------ inside virtual machine -------
# echo 'p:myprobe sys_open' > /sys/kernel/debug/tracing/kprobe_events
# echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
------ gdb attaches to QEMU -------
cpu_v7_do_idle () at /home/w00229757/kernel-hydrogen/arch/arm/mm/proc-v7.S:74
74 ret lr
(gdb) x/3i sys_open
0xc013619c <SyS_open>: b 0xbf000000 <--- *optimized*
0xc01361a0 <SyS_open+4>: push {r11, r12, lr, pc}
0xc01361a4 <SyS_open+8>: sub r11, r12, #4
(gdb) c
------ inside virtual machine -------
# echo 0 > /sys/kernel/debug/kprobes/enabled
------ gdb attaches to QEMU -------
cpu_v7_do_idle () at /home/w00229757/kernel-hydrogen/arch/arm/mm/proc-v7.S:74
74 ret lr
(gdb) x/3i sys_open
0xc013619c <SyS_open>: ; <UNDEFINED> instruction: 0xe7f001f8 <--- *BREAKPOINT*
0xc01361a0 <SyS_open+4>: push {r11, r12, lr, pc}
0xc01361a4 <SyS_open+8>: sub r11, r12, #4
(gdb) c
** x86_64 result **
------ gdb attaches to QEMU -------
(gdb) x/10i sys_open
0xffffffff81184fe0 <SyS_open>: data32 data32 data32 xchg %ax,%ax
0xffffffff81184fe5 <SyS_open+5>: push %rbp
0xffffffff81184fe6 <SyS_open+6>: movzwl %dx,%ecx
0xffffffff81184fe9 <SyS_open+9>: mov %esi,%edx
0xffffffff81184feb <SyS_open+11>: mov %rsp,%rbp
0xffffffff81184fee <SyS_open+14>: mov %rdi,%rsi
0xffffffff81184ff1 <SyS_open+17>: or $0x80,%dh
0xffffffff81184ff4 <SyS_open+20>: mov $0xffffff9c,%edi
0xffffffff81184ff9 <SyS_open+25>: callq 0xffffffff81184da0 <do_sys_open>
0xffffffff81184ffe <SyS_open+30>: pop %rbp
(gdb) c
Continuing
------ inside virtual machine -------
# echo 'p:myprobe sys_open+20' > /sys/kernel/debug/tracing/kprobe_events
# echo 1 > /sys/kernel/debug/tracing/events/kprobes/myprobe/enable
------ gdb attaches to QEMU -------
(gdb) x/10i sys_open
0xffffffff81184fe0 <SyS_open>: data32 data32 data32 xchg %ax,%ax
0xffffffff81184fe5 <SyS_open+5>: push %rbp
0xffffffff81184fe6 <SyS_open+6>: movzwl %dx,%ecx
0xffffffff81184fe9 <SyS_open+9>: mov %esi,%edx
0xffffffff81184feb <SyS_open+11>: mov %rsp,%rbp
0xffffffff81184fee <SyS_open+14>: mov %rdi,%rsi
0xffffffff81184ff1 <SyS_open+17>: or $0x80,%dh
0xffffffff81184ff4 <SyS_open+20>: jmpq 0xffffffffa0002000 <--- *optimized*
0xffffffff81184ff9 <SyS_open+25>: callq 0xffffffff81184da0 <do_sys_open>
0xffffffff81184ffe <SyS_open+30>: pop %rbp
(gdb) c
Continuing.
------ inside virtual machine -------
# echo 0 > /sys/kernel/debug/kprobes/enabled
------ gdb attaches to QEMU -------
(gdb) x/10i sys_open
0xffffffff81184fe0 <SyS_open>: data32 data32 data32 xchg %ax,%ax
0xffffffff81184fe5 <SyS_open+5>: push %rbp
0xffffffff81184fe6 <SyS_open+6>: movzwl %dx,%ecx
0xffffffff81184fe9 <SyS_open+9>: mov %esi,%edx
0xffffffff81184feb <SyS_open+11>: mov %rsp,%rbp
0xffffffff81184fee <SyS_open+14>: mov %rdi,%rsi
0xffffffff81184ff1 <SyS_open+17>: or $0x80,%dh
0xffffffff81184ff4 <SyS_open+20>: int3 <-- **BREAKPOINT**
0xffffffff81184ff5 <SyS_open+21>: pushfq
0xffffffff81184ff6 <SyS_open+22>: (bad)
(gdb)
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-19 11:21 ` Wang Nan
@ 2015-01-19 12:45 ` Masami Hiramatsu
-1 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-19 12:45 UTC (permalink / raw)
To: Wang Nan; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
(2015/01/19 20:21), Wang Nan wrote:
> On 2015/1/19 17:05, Masami Hiramatsu wrote:
>> Hi Wang,
>>
>> I've found a problem on this patch, since kprobes calls unoptioize_kprobe
>> with kprobes_all_disarmed=true when trying to disable all kprobes, this
>> cause a serious problem.
>>
>> Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
>> Could you test it again?
>>
>
> I tested it again based on 3.19-rc5 and found that the problem still exists.
> My testing is based on QEMU.
>
> First I tested my kprobeopt for ARM, then on x86_64. The test results are pasted
> at the bottom of this mail. Commands after 'gdb attaches to QEMU' is my actions
> on a gdb console attached to QEMU; commands after 'inside virtual machine' is
> what I do in Linux run under QEMU.
Thank you for the reporting.
So, now I know what happened, the problem is "debugfs/kprobes/enabled doesn't work
correctly on optimized kprobes". Please make update the patch description.
I also reproduced the bug without gdb.
Here is the log.
----
[root@localhost ~]# cd /sys/kernel/debug/tracing/
[root@localhost tracing]# echo p do_fork+5 > kprobe_events # setup new event
[root@localhost tracing]# echo $$ > set_ftrace_pid # trace only this process
[root@localhost tracing]# echo 1 > events/kprobes/p_do_fork_5/enable # enable it
[root@localhost tracing]# cat trace # check the trace data
# tracer: nop
#
# entries-in-buffer/entries-written: 1/1 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-3883 [006] d... 279.799023: p_do_fork_5: (do_fork+0x5/0x360) # OK, now tracing
[root@localhost tracing]# cat ../kprobes/list
ffffffff810bc1c5 k do_fork+0x5 [OPTIMIZED] # and it is actually optimized
[root@localhost tracing]# echo 0 > ../kprobes/enabled # disable *ALL* kprobes
[root@localhost tracing]# echo > trace # clear events
[root@localhost tracing]# cat trace # this should show empty buffer
# tracer: nop
#
# entries-in-buffer/entries-written: 1/1 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360) # But still tracing!
[root@localhost tracing]# cat trace # Check again
# tracer: nop
#
# entries-in-buffer/entries-written: 2/2 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360)
bash-3883 [006] d... 345.592178: p_do_fork_5: (do_fork+0x5/0x360) # We are tracing!!
So, after global disabling kprobes, ALL kprobes event should be disabled, but not.
OK, I think your first patch is better than the second one, but not enough.
What we should do is use kprobes_all_disarmed for force option like below.
unoptimize_kprobe(p, kprobes_all_disarmed); /* Try to unoptimize */
We also would better to check the flag in unregistering path for skipping unneeded
disarming process when kprobes globally disarmed.
Thank you,
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-19 12:45 ` Masami Hiramatsu
0 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-19 12:45 UTC (permalink / raw)
To: linux-arm-kernel
(2015/01/19 20:21), Wang Nan wrote:
> On 2015/1/19 17:05, Masami Hiramatsu wrote:
>> Hi Wang,
>>
>> I've found a problem on this patch, since kprobes calls unoptioize_kprobe
>> with kprobes_all_disarmed=true when trying to disable all kprobes, this
>> cause a serious problem.
>>
>> Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
>> Could you test it again?
>>
>
> I tested it again based on 3.19-rc5 and found that the problem still exists.
> My testing is based on QEMU.
>
> First I tested my kprobeopt for ARM, then on x86_64. The test results are pasted
> at the bottom of this mail. Commands after 'gdb attaches to QEMU' is my actions
> on a gdb console attached to QEMU; commands after 'inside virtual machine' is
> what I do in Linux run under QEMU.
Thank you for the reporting.
So, now I know what happened, the problem is "debugfs/kprobes/enabled doesn't work
correctly on optimized kprobes". Please make update the patch description.
I also reproduced the bug without gdb.
Here is the log.
----
[root at localhost ~]# cd /sys/kernel/debug/tracing/
[root at localhost tracing]# echo p do_fork+5 > kprobe_events # setup new event
[root at localhost tracing]# echo $$ > set_ftrace_pid # trace only this process
[root at localhost tracing]# echo 1 > events/kprobes/p_do_fork_5/enable # enable it
[root at localhost tracing]# cat trace # check the trace data
# tracer: nop
#
# entries-in-buffer/entries-written: 1/1 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-3883 [006] d... 279.799023: p_do_fork_5: (do_fork+0x5/0x360) # OK, now tracing
[root at localhost tracing]# cat ../kprobes/list
ffffffff810bc1c5 k do_fork+0x5 [OPTIMIZED] # and it is actually optimized
[root at localhost tracing]# echo 0 > ../kprobes/enabled # disable *ALL* kprobes
[root at localhost tracing]# echo > trace # clear events
[root at localhost tracing]# cat trace # this should show empty buffer
# tracer: nop
#
# entries-in-buffer/entries-written: 1/1 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360) # But still tracing!
[root at localhost tracing]# cat trace # Check again
# tracer: nop
#
# entries-in-buffer/entries-written: 2/2 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360)
bash-3883 [006] d... 345.592178: p_do_fork_5: (do_fork+0x5/0x360) # We are tracing!!
So, after global disabling kprobes, ALL kprobes event should be disabled, but not.
OK, I think your first patch is better than the second one, but not enough.
What we should do is use kprobes_all_disarmed for force option like below.
unoptimize_kprobe(p, kprobes_all_disarmed); /* Try to unoptimize */
We also would better to check the flag in unregistering path for skipping unneeded
disarming process when kprobes globally disarmed.
Thank you,
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
2015-01-19 12:45 ` Masami Hiramatsu
@ 2015-01-19 12:59 ` Wang Nan
-1 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-19 12:59 UTC (permalink / raw)
To: Masami Hiramatsu; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
On 2015/1/19 20:45, Masami Hiramatsu wrote:
> (2015/01/19 20:21), Wang Nan wrote:
>> On 2015/1/19 17:05, Masami Hiramatsu wrote:
>>> Hi Wang,
>>>
>>> I've found a problem on this patch, since kprobes calls unoptioize_kprobe
>>> with kprobes_all_disarmed=true when trying to disable all kprobes, this
>>> cause a serious problem.
>>>
>>> Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
>>> Could you test it again?
>>>
>>
>> I tested it again based on 3.19-rc5 and found that the problem still exists.
>> My testing is based on QEMU.
>>
>> First I tested my kprobeopt for ARM, then on x86_64. The test results are pasted
>> at the bottom of this mail. Commands after 'gdb attaches to QEMU' is my actions
>> on a gdb console attached to QEMU; commands after 'inside virtual machine' is
>> what I do in Linux run under QEMU.
>
> Thank you for the reporting.
> So, now I know what happened, the problem is "debugfs/kprobes/enabled doesn't work
> correctly on optimized kprobes". Please make update the patch description.
>
> I also reproduced the bug without gdb.
> Here is the log.
>
> ----
> [root@localhost ~]# cd /sys/kernel/debug/tracing/
> [root@localhost tracing]# echo p do_fork+5 > kprobe_events # setup new event
> [root@localhost tracing]# echo $$ > set_ftrace_pid # trace only this process
> [root@localhost tracing]# echo 1 > events/kprobes/p_do_fork_5/enable # enable it
> [root@localhost tracing]# cat trace # check the trace data
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 1/1 #P:8
> #
> # _-----=> irqs-off
> # / _----=> need-resched
> # | / _---=> hardirq/softirq
> # || / _--=> preempt-depth
> # ||| / delay
> # TASK-PID CPU# |||| TIMESTAMP FUNCTION
> # | | | |||| | |
> bash-3883 [006] d... 279.799023: p_do_fork_5: (do_fork+0x5/0x360) # OK, now tracing
> [root@localhost tracing]# cat ../kprobes/list
> ffffffff810bc1c5 k do_fork+0x5 [OPTIMIZED] # and it is actually optimized
> [root@localhost tracing]# echo 0 > ../kprobes/enabled # disable *ALL* kprobes
> [root@localhost tracing]# echo > trace # clear events
> [root@localhost tracing]# cat trace # this should show empty buffer
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 1/1 #P:8
> #
> # _-----=> irqs-off
> # / _----=> need-resched
> # | / _---=> hardirq/softirq
> # || / _--=> preempt-depth
> # ||| / delay
> # TASK-PID CPU# |||| TIMESTAMP FUNCTION
> # | | | |||| | |
> bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360) # But still tracing!
> [root@localhost tracing]# cat trace # Check again
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 2/2 #P:8
> #
> # _-----=> irqs-off
> # / _----=> need-resched
> # | / _---=> hardirq/softirq
> # || / _--=> preempt-depth
> # ||| / delay
> # TASK-PID CPU# |||| TIMESTAMP FUNCTION
> # | | | |||| | |
> bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360)
> bash-3883 [006] d... 345.592178: p_do_fork_5: (do_fork+0x5/0x360) # We are tracing!!
>
> So, after global disabling kprobes, ALL kprobes event should be disabled, but not.
>
> OK, I think your first patch is better than the second one, but not enough.
> What we should do is use kprobes_all_disarmed for force option like below.
>
> unoptimize_kprobe(p, kprobes_all_disarmed); /* Try to unoptimize */
>
> We also would better to check the flag in unregistering path for skipping unneeded
> disarming process when kprobes globally disarmed.
>
> Thank you,
>
Thanks to your quick reply. I'll post an improved v1 patch tomorrow.
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe().
@ 2015-01-19 12:59 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-19 12:59 UTC (permalink / raw)
To: linux-arm-kernel
On 2015/1/19 20:45, Masami Hiramatsu wrote:
> (2015/01/19 20:21), Wang Nan wrote:
>> On 2015/1/19 17:05, Masami Hiramatsu wrote:
>>> Hi Wang,
>>>
>>> I've found a problem on this patch, since kprobes calls unoptioize_kprobe
>>> with kprobes_all_disarmed=true when trying to disable all kprobes, this
>>> cause a serious problem.
>>>
>>> Moreover, I couldn't reproduce your reported bug on my 3.19-rc4 kernel.
>>> Could you test it again?
>>>
>>
>> I tested it again based on 3.19-rc5 and found that the problem still exists.
>> My testing is based on QEMU.
>>
>> First I tested my kprobeopt for ARM, then on x86_64. The test results are pasted
>> at the bottom of this mail. Commands after 'gdb attaches to QEMU' is my actions
>> on a gdb console attached to QEMU; commands after 'inside virtual machine' is
>> what I do in Linux run under QEMU.
>
> Thank you for the reporting.
> So, now I know what happened, the problem is "debugfs/kprobes/enabled doesn't work
> correctly on optimized kprobes". Please make update the patch description.
>
> I also reproduced the bug without gdb.
> Here is the log.
>
> ----
> [root at localhost ~]# cd /sys/kernel/debug/tracing/
> [root at localhost tracing]# echo p do_fork+5 > kprobe_events # setup new event
> [root at localhost tracing]# echo $$ > set_ftrace_pid # trace only this process
> [root at localhost tracing]# echo 1 > events/kprobes/p_do_fork_5/enable # enable it
> [root at localhost tracing]# cat trace # check the trace data
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 1/1 #P:8
> #
> # _-----=> irqs-off
> # / _----=> need-resched
> # | / _---=> hardirq/softirq
> # || / _--=> preempt-depth
> # ||| / delay
> # TASK-PID CPU# |||| TIMESTAMP FUNCTION
> # | | | |||| | |
> bash-3883 [006] d... 279.799023: p_do_fork_5: (do_fork+0x5/0x360) # OK, now tracing
> [root at localhost tracing]# cat ../kprobes/list
> ffffffff810bc1c5 k do_fork+0x5 [OPTIMIZED] # and it is actually optimized
> [root at localhost tracing]# echo 0 > ../kprobes/enabled # disable *ALL* kprobes
> [root at localhost tracing]# echo > trace # clear events
> [root at localhost tracing]# cat trace # this should show empty buffer
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 1/1 #P:8
> #
> # _-----=> irqs-off
> # / _----=> need-resched
> # | / _---=> hardirq/softirq
> # || / _--=> preempt-depth
> # ||| / delay
> # TASK-PID CPU# |||| TIMESTAMP FUNCTION
> # | | | |||| | |
> bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360) # But still tracing!
> [root at localhost tracing]# cat trace # Check again
> # tracer: nop
> #
> # entries-in-buffer/entries-written: 2/2 #P:8
> #
> # _-----=> irqs-off
> # / _----=> need-resched
> # | / _---=> hardirq/softirq
> # || / _--=> preempt-depth
> # ||| / delay
> # TASK-PID CPU# |||| TIMESTAMP FUNCTION
> # | | | |||| | |
> bash-3883 [006] d... 337.770785: p_do_fork_5: (do_fork+0x5/0x360)
> bash-3883 [006] d... 345.592178: p_do_fork_5: (do_fork+0x5/0x360) # We are tracing!!
>
> So, after global disabling kprobes, ALL kprobes event should be disabled, but not.
>
> OK, I think your first patch is better than the second one, but not enough.
> What we should do is use kprobes_all_disarmed for force option like below.
>
> unoptimize_kprobe(p, kprobes_all_disarmed); /* Try to unoptimize */
>
> We also would better to check the flag in unregistering path for skipping unneeded
> disarming process when kprobes globally disarmed.
>
> Thank you,
>
Thanks to your quick reply. I'll post an improved v1 patch tomorrow.
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: makes kprobes/enabled works correctly for optimized kprobes.
2015-01-19 12:45 ` Masami Hiramatsu
@ 2015-01-20 2:51 ` Wang Nan
-1 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-20 2:51 UTC (permalink / raw)
To: masami.hiramatsu.pt; +Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
debugfs/kprobes/enabled doesn't work correctly on optimized kprobes.
Masami Hiramatsu has a test report on x86_64 platform:
https://lkml.org/lkml/2015/1/19/274
This patch forces it to unoptimize kprobe if kprobes_all_disarmed
is set. It also checks the flag in unregistering path for skipping
unneeded disarming process when kprobes globally disarmed.
Signed-off-by: Wang Nan <wangnan0@huawei.com>
---
kernel/kprobes.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 9471710..fb995ef 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -869,7 +869,8 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
{
struct kprobe *_p;
- unoptimize_kprobe(p, false); /* Try to unoptimize */
+ /* Try to unoptimize */
+ unoptimize_kprobe(p, kprobes_all_disarmed);
if (!kprobe_queued(p)) {
arch_disarm_kprobe(p);
@@ -1571,7 +1572,13 @@ static struct kprobe *__disable_kprobe(struct kprobe *p)
/* Try to disarm and disable this/parent probe */
if (p == orig_p || aggr_kprobe_disabled(orig_p)) {
- disarm_kprobe(orig_p, true);
+ /*
+ * If kprobes_all_disarmed is set, orig_p
+ * should have already been disarmed, so
+ * skip unneed disarming process.
+ */
+ if (!kprobes_all_disarmed)
+ disarm_kprobe(orig_p, true);
orig_p->flags |= KPROBE_FLAG_DISABLED;
}
}
--
1.8.4
^ permalink raw reply related [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: makes kprobes/enabled works correctly for optimized kprobes.
@ 2015-01-20 2:51 ` Wang Nan
0 siblings, 0 replies; 22+ messages in thread
From: Wang Nan @ 2015-01-20 2:51 UTC (permalink / raw)
To: linux-arm-kernel
debugfs/kprobes/enabled doesn't work correctly on optimized kprobes.
Masami Hiramatsu has a test report on x86_64 platform:
https://lkml.org/lkml/2015/1/19/274
This patch forces it to unoptimize kprobe if kprobes_all_disarmed
is set. It also checks the flag in unregistering path for skipping
unneeded disarming process when kprobes globally disarmed.
Signed-off-by: Wang Nan <wangnan0@huawei.com>
---
kernel/kprobes.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 9471710..fb995ef 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -869,7 +869,8 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
{
struct kprobe *_p;
- unoptimize_kprobe(p, false); /* Try to unoptimize */
+ /* Try to unoptimize */
+ unoptimize_kprobe(p, kprobes_all_disarmed);
if (!kprobe_queued(p)) {
arch_disarm_kprobe(p);
@@ -1571,7 +1572,13 @@ static struct kprobe *__disable_kprobe(struct kprobe *p)
/* Try to disarm and disable this/parent probe */
if (p == orig_p || aggr_kprobe_disabled(orig_p)) {
- disarm_kprobe(orig_p, true);
+ /*
+ * If kprobes_all_disarmed is set, orig_p
+ * should have already been disarmed, so
+ * skip unneed disarming process.
+ */
+ if (!kprobes_all_disarmed)
+ disarm_kprobe(orig_p, true);
orig_p->flags |= KPROBE_FLAG_DISABLED;
}
}
--
1.8.4
^ permalink raw reply related [flat|nested] 22+ messages in thread
* Re: [PATCH] kprobes: bugfix: makes kprobes/enabled works correctly for optimized kprobes.
2015-01-20 2:51 ` Wang Nan
@ 2015-01-20 7:12 ` Masami Hiramatsu
-1 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-20 7:12 UTC (permalink / raw)
To: Wang Nan, Ingo Molnar
Cc: tixy, linux, linux-kernel, linux-arm-kernel, lizefan
(2015/01/20 11:51), Wang Nan wrote:
> debugfs/kprobes/enabled doesn't work correctly on optimized kprobes.
> Masami Hiramatsu has a test report on x86_64 platform:
>
> https://lkml.org/lkml/2015/1/19/274
>
> This patch forces it to unoptimize kprobe if kprobes_all_disarmed
> is set. It also checks the flag in unregistering path for skipping
> unneeded disarming process when kprobes globally disarmed.
>
OK, here is the test result.
----
[root@localhost tracing]# echo p do_fork+5 > kprobe_events
[root@localhost tracing]# echo $$ > set_ftrace_pid
[root@localhost tracing]# echo 1 > events/kprobes/p_do_fork_5/enable
[root@localhost tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 1/1 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-11466 [004] d... 4548.420463: p_do_fork_5: (do_fork+0x5/0x360)
[root@localhost tracing]# cat ../kprobes/list
ffffffff810bc1c5 k do_fork+0x5 [OPTIMIZED]
[root@localhost tracing]# echo 0 > ../kprobes/enabled # kprobes globally disabled
[root@localhost tracing]# echo > trace # clear trace buffer
[root@localhost tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 0/0 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | | # No event traced!
[root@localhost tracing]# echo 0 > events/kprobes/p_do_fork_5/enable
[root@localhost tracing]# echo > kprobe_events
[root@localhost tracing]# cat ../kprobes/list
[root@localhost tracing]# # And we can safely remove the probe
-----
Now it looks good to me :)
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Ingo,
Could you pull this patch to your -tip tree?
Thank you,
> Signed-off-by: Wang Nan <wangnan0@huawei.com>
> ---
> kernel/kprobes.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 9471710..fb995ef 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -869,7 +869,8 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
> {
> struct kprobe *_p;
>
> - unoptimize_kprobe(p, false); /* Try to unoptimize */
> + /* Try to unoptimize */
> + unoptimize_kprobe(p, kprobes_all_disarmed);
>
> if (!kprobe_queued(p)) {
> arch_disarm_kprobe(p);
> @@ -1571,7 +1572,13 @@ static struct kprobe *__disable_kprobe(struct kprobe *p)
>
> /* Try to disarm and disable this/parent probe */
> if (p == orig_p || aggr_kprobe_disabled(orig_p)) {
> - disarm_kprobe(orig_p, true);
> + /*
> + * If kprobes_all_disarmed is set, orig_p
> + * should have already been disarmed, so
> + * skip unneed disarming process.
> + */
> + if (!kprobes_all_disarmed)
> + disarm_kprobe(orig_p, true);
> orig_p->flags |= KPROBE_FLAG_DISABLED;
> }
> }
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
* [PATCH] kprobes: bugfix: makes kprobes/enabled works correctly for optimized kprobes.
@ 2015-01-20 7:12 ` Masami Hiramatsu
0 siblings, 0 replies; 22+ messages in thread
From: Masami Hiramatsu @ 2015-01-20 7:12 UTC (permalink / raw)
To: linux-arm-kernel
(2015/01/20 11:51), Wang Nan wrote:
> debugfs/kprobes/enabled doesn't work correctly on optimized kprobes.
> Masami Hiramatsu has a test report on x86_64 platform:
>
> https://lkml.org/lkml/2015/1/19/274
>
> This patch forces it to unoptimize kprobe if kprobes_all_disarmed
> is set. It also checks the flag in unregistering path for skipping
> unneeded disarming process when kprobes globally disarmed.
>
OK, here is the test result.
----
[root at localhost tracing]# echo p do_fork+5 > kprobe_events
[root at localhost tracing]# echo $$ > set_ftrace_pid
[root at localhost tracing]# echo 1 > events/kprobes/p_do_fork_5/enable
[root at localhost tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 1/1 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | |
bash-11466 [004] d... 4548.420463: p_do_fork_5: (do_fork+0x5/0x360)
[root at localhost tracing]# cat ../kprobes/list
ffffffff810bc1c5 k do_fork+0x5 [OPTIMIZED]
[root at localhost tracing]# echo 0 > ../kprobes/enabled # kprobes globally disabled
[root at localhost tracing]# echo > trace # clear trace buffer
[root at localhost tracing]# cat trace
# tracer: nop
#
# entries-in-buffer/entries-written: 0/0 #P:8
#
# _-----=> irqs-off
# / _----=> need-resched
# | / _---=> hardirq/softirq
# || / _--=> preempt-depth
# ||| / delay
# TASK-PID CPU# |||| TIMESTAMP FUNCTION
# | | | |||| | | # No event traced!
[root at localhost tracing]# echo 0 > events/kprobes/p_do_fork_5/enable
[root at localhost tracing]# echo > kprobe_events
[root at localhost tracing]# cat ../kprobes/list
[root at localhost tracing]# # And we can safely remove the probe
-----
Now it looks good to me :)
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Ingo,
Could you pull this patch to your -tip tree?
Thank you,
> Signed-off-by: Wang Nan <wangnan0@huawei.com>
> ---
> kernel/kprobes.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 9471710..fb995ef 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -869,7 +869,8 @@ static void __disarm_kprobe(struct kprobe *p, bool reopt)
> {
> struct kprobe *_p;
>
> - unoptimize_kprobe(p, false); /* Try to unoptimize */
> + /* Try to unoptimize */
> + unoptimize_kprobe(p, kprobes_all_disarmed);
>
> if (!kprobe_queued(p)) {
> arch_disarm_kprobe(p);
> @@ -1571,7 +1572,13 @@ static struct kprobe *__disable_kprobe(struct kprobe *p)
>
> /* Try to disarm and disable this/parent probe */
> if (p == orig_p || aggr_kprobe_disabled(orig_p)) {
> - disarm_kprobe(orig_p, true);
> + /*
> + * If kprobes_all_disarmed is set, orig_p
> + * should have already been disarmed, so
> + * skip unneed disarming process.
> + */
> + if (!kprobes_all_disarmed)
> + disarm_kprobe(orig_p, true);
> orig_p->flags |= KPROBE_FLAG_DISABLED;
> }
> }
>
--
Masami HIRAMATSU
Software Platform Research Dept. Linux Technology Research Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt at hitachi.com
^ permalink raw reply [flat|nested] 22+ messages in thread
end of thread, other threads:[~2015-01-20 7:12 UTC | newest]
Thread overview: 22+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-01-05 12:32 [PATCH] kprobes: bugfix: force unoptimize when disable kprobes Wang Nan
2015-01-05 12:32 ` Wang Nan
2015-01-12 11:42 ` Masami Hiramatsu
2015-01-12 11:42 ` Masami Hiramatsu
2015-01-12 12:09 ` [PATCH] kprobes: bugfix: checks kprobes_all_disarmed in unoptimized_kprobe() Wang Nan
2015-01-12 12:09 ` Wang Nan
2015-01-12 12:52 ` Masami Hiramatsu
2015-01-12 12:52 ` Masami Hiramatsu
2015-01-19 3:04 ` Wang Nan
2015-01-19 3:04 ` Wang Nan
2015-01-19 9:05 ` Masami Hiramatsu
2015-01-19 9:05 ` Masami Hiramatsu
2015-01-19 11:21 ` Wang Nan
2015-01-19 11:21 ` Wang Nan
2015-01-19 12:45 ` Masami Hiramatsu
2015-01-19 12:45 ` Masami Hiramatsu
2015-01-19 12:59 ` Wang Nan
2015-01-19 12:59 ` Wang Nan
2015-01-20 2:51 ` [PATCH] kprobes: bugfix: makes kprobes/enabled works correctly for optimized kprobes Wang Nan
2015-01-20 2:51 ` Wang Nan
2015-01-20 7:12 ` Masami Hiramatsu
2015-01-20 7:12 ` Masami Hiramatsu
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.