* [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
@ 2024-06-20 17:15 Mateusz Guzik
2024-06-20 18:23 ` Neeraj Upadhyay
2024-07-05 6:30 ` kernel test robot
0 siblings, 2 replies; 6+ messages in thread
From: Mateusz Guzik @ 2024-06-20 17:15 UTC (permalink / raw)
To: john.johansen
Cc: paul, jmorris, serge, linux-kernel, apparmor,
linux-security-module, Neeraj.Upadhyay, Mateusz Guzik
apparmor: try to avoid refing the label in apparmor_file_open
If the label is not stale (which is the common case), the fact that the
passed file object holds a reference can be leverged to avoid the
ref/unref cycle. Doing so reduces performance impact of apparmor on
parallel open() invocations.
When benchmarking on a 24-core vm using will-it-scale's open1_process
("Separate file open"), the results are (ops/s):
before: 6092196
after: 8309726 (+36%)
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
---
v2:
- reword the commit message
If you want any changes made to it can you just do them on your own
accord? :) Will be faster for both of us than another mail trip.
security/apparmor/include/cred.h | 20 ++++++++++++++++++++
security/apparmor/lsm.c | 5 +++--
2 files changed, 23 insertions(+), 2 deletions(-)
diff --git a/security/apparmor/include/cred.h b/security/apparmor/include/cred.h
index 58fdc72af664..7265d2f81dd5 100644
--- a/security/apparmor/include/cred.h
+++ b/security/apparmor/include/cred.h
@@ -63,6 +63,26 @@ static inline struct aa_label *aa_get_newest_cred_label(const struct cred *cred)
return aa_get_newest_label(aa_cred_raw_label(cred));
}
+static inline struct aa_label *aa_get_newest_cred_label_condref(const struct cred *cred,
+ bool *needput)
+{
+ struct aa_label *l = aa_cred_raw_label(cred);
+
+ if (unlikely(label_is_stale(l))) {
+ *needput = true;
+ return aa_get_newest_label(l);
+ }
+
+ *needput = false;
+ return l;
+}
+
+static inline void aa_put_label_condref(struct aa_label *l, bool needput)
+{
+ if (unlikely(needput))
+ aa_put_label(l);
+}
+
/**
* aa_current_raw_label - find the current tasks confining label
*
diff --git a/security/apparmor/lsm.c b/security/apparmor/lsm.c
index 2cea34657a47..4bf87eac4a56 100644
--- a/security/apparmor/lsm.c
+++ b/security/apparmor/lsm.c
@@ -461,6 +461,7 @@ static int apparmor_file_open(struct file *file)
struct aa_file_ctx *fctx = file_ctx(file);
struct aa_label *label;
int error = 0;
+ bool needput;
if (!path_mediated_fs(file->f_path.dentry))
return 0;
@@ -477,7 +478,7 @@ static int apparmor_file_open(struct file *file)
return 0;
}
- label = aa_get_newest_cred_label(file->f_cred);
+ label = aa_get_newest_cred_label_condref(file->f_cred, &needput);
if (!unconfined(label)) {
struct mnt_idmap *idmap = file_mnt_idmap(file);
struct inode *inode = file_inode(file);
@@ -494,7 +495,7 @@ static int apparmor_file_open(struct file *file)
/* todo cache full allowed permissions set and state */
fctx->allow = aa_map_file_to_perms(file);
}
- aa_put_label(label);
+ aa_put_label_condref(label, needput);
return error;
}
--
2.43.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
2024-06-20 17:15 [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open Mateusz Guzik
@ 2024-06-20 18:23 ` Neeraj Upadhyay
2024-06-20 18:30 ` Mateusz Guzik
2024-07-05 6:30 ` kernel test robot
1 sibling, 1 reply; 6+ messages in thread
From: Neeraj Upadhyay @ 2024-06-20 18:23 UTC (permalink / raw)
To: Mateusz Guzik, john.johansen
Cc: paul, jmorris, serge, linux-kernel, apparmor,
linux-security-module
On 6/20/2024 10:45 PM, Mateusz Guzik wrote:
> apparmor: try to avoid refing the label in apparmor_file_open
>
> If the label is not stale (which is the common case), the fact that the
> passed file object holds a reference can be leverged to avoid the
Minor: Typo 'leveraged'
> ref/unref cycle. Doing so reduces performance impact of apparmor on
> parallel open() invocations.
>
> When benchmarking on a 24-core vm using will-it-scale's open1_process
> ("Separate file open"), the results are (ops/s):
> before: 6092196
> after: 8309726 (+36%)
>
> Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
> ---
Trying to understand the changes done here. So, while the file cred can be updated
async to the task (referring to the comment from John here [1]), the file cred label
cannot change during apparmor_file_open() execution?
Reviewed-by: Neeraj upadhyay <Neeraj.Upadhyay@amd.com>
Thanks
Neeraj
[1] https://lore.kernel.org/lkml/9bfaeec2-535d-4401-8244-7560f660a065@canonical.com/
>
> v2:
> - reword the commit message
>
> If you want any changes made to it can you just do them on your own
> accord? :) Will be faster for both of us than another mail trip.
>
> security/apparmor/include/cred.h | 20 ++++++++++++++++++++
> security/apparmor/lsm.c | 5 +++--
> 2 files changed, 23 insertions(+), 2 deletions(-)
>
> diff --git a/security/apparmor/include/cred.h b/security/apparmor/include/cred.h
> index 58fdc72af664..7265d2f81dd5 100644
> --- a/security/apparmor/include/cred.h
> +++ b/security/apparmor/include/cred.h
> @@ -63,6 +63,26 @@ static inline struct aa_label *aa_get_newest_cred_label(const struct cred *cred)
> return aa_get_newest_label(aa_cred_raw_label(cred));
> }
>
> +static inline struct aa_label *aa_get_newest_cred_label_condref(const struct cred *cred,
> + bool *needput)
> +{
> + struct aa_label *l = aa_cred_raw_label(cred);
> +
> + if (unlikely(label_is_stale(l))) {
> + *needput = true;
> + return aa_get_newest_label(l);
> + }
> +
> + *needput = false;
> + return l;
> +}
> +
> +static inline void aa_put_label_condref(struct aa_label *l, bool needput)
> +{
> + if (unlikely(needput))
> + aa_put_label(l);
> +}
> +
> /**
> * aa_current_raw_label - find the current tasks confining label
> *
> diff --git a/security/apparmor/lsm.c b/security/apparmor/lsm.c
> index 2cea34657a47..4bf87eac4a56 100644
> --- a/security/apparmor/lsm.c
> +++ b/security/apparmor/lsm.c
> @@ -461,6 +461,7 @@ static int apparmor_file_open(struct file *file)
> struct aa_file_ctx *fctx = file_ctx(file);
> struct aa_label *label;
> int error = 0;
> + bool needput;
>
> if (!path_mediated_fs(file->f_path.dentry))
> return 0;
> @@ -477,7 +478,7 @@ static int apparmor_file_open(struct file *file)
> return 0;
> }
>
> - label = aa_get_newest_cred_label(file->f_cred);
> + label = aa_get_newest_cred_label_condref(file->f_cred, &needput);
> if (!unconfined(label)) {
> struct mnt_idmap *idmap = file_mnt_idmap(file);
> struct inode *inode = file_inode(file);
> @@ -494,7 +495,7 @@ static int apparmor_file_open(struct file *file)
> /* todo cache full allowed permissions set and state */
> fctx->allow = aa_map_file_to_perms(file);
> }
> - aa_put_label(label);
> + aa_put_label_condref(label, needput);
>
> return error;
> }
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
2024-06-20 18:23 ` Neeraj Upadhyay
@ 2024-06-20 18:30 ` Mateusz Guzik
2024-06-20 19:03 ` John Johansen
0 siblings, 1 reply; 6+ messages in thread
From: Mateusz Guzik @ 2024-06-20 18:30 UTC (permalink / raw)
To: Neeraj Upadhyay
Cc: john.johansen, paul, jmorris, serge, linux-kernel, apparmor,
linux-security-module
On Thu, Jun 20, 2024 at 8:23 PM Neeraj Upadhyay <Neeraj.Upadhyay@amd.com> wrote:
>
>
>
> On 6/20/2024 10:45 PM, Mateusz Guzik wrote:
> > apparmor: try to avoid refing the label in apparmor_file_open
> >
> > If the label is not stale (which is the common case), the fact that the
> > passed file object holds a reference can be leverged to avoid the
>
> Minor: Typo 'leveraged'
>
> > ref/unref cycle. Doing so reduces performance impact of apparmor on
> > parallel open() invocations.
> >
> > When benchmarking on a 24-core vm using will-it-scale's open1_process
> > ("Separate file open"), the results are (ops/s):
> > before: 6092196
> > after: 8309726 (+36%)
> >
> > Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
> > ---
>
>
> Trying to understand the changes done here. So, while the file cred can be updated
> async to the task (referring to the comment from John here [1]), the file cred label
> cannot change during apparmor_file_open() execution?
>
Refing a label retains racy vs refing it.
On stock code you can test the flag, determine it's not stale, grab
the ref and have it become stale immediately after. My patch avoids
the atomic dance for the common case, does not alter anything
correctness-wise AFAICS.
I am assuming the race is tolerated and checking here is only done to
make sure the new label is seen eventually.
Not having the race is possible with a bunch of trickery like seqc,
but so far does not look like this is necessary.
>
> Reviewed-by: Neeraj upadhyay <Neeraj.Upadhyay@amd.com>
>
>
> Thanks
> Neeraj
>
> [1] https://lore.kernel.org/lkml/9bfaeec2-535d-4401-8244-7560f660a065@canonical.com/
>
>
> >
> > v2:
> > - reword the commit message
> >
> > If you want any changes made to it can you just do them on your own
> > accord? :) Will be faster for both of us than another mail trip.
> >
> > security/apparmor/include/cred.h | 20 ++++++++++++++++++++
> > security/apparmor/lsm.c | 5 +++--
> > 2 files changed, 23 insertions(+), 2 deletions(-)
> >
> > diff --git a/security/apparmor/include/cred.h b/security/apparmor/include/cred.h
> > index 58fdc72af664..7265d2f81dd5 100644
> > --- a/security/apparmor/include/cred.h
> > +++ b/security/apparmor/include/cred.h
> > @@ -63,6 +63,26 @@ static inline struct aa_label *aa_get_newest_cred_label(const struct cred *cred)
> > return aa_get_newest_label(aa_cred_raw_label(cred));
> > }
> >
> > +static inline struct aa_label *aa_get_newest_cred_label_condref(const struct cred *cred,
> > + bool *needput)
> > +{
> > + struct aa_label *l = aa_cred_raw_label(cred);
> > +
> > + if (unlikely(label_is_stale(l))) {
> > + *needput = true;
> > + return aa_get_newest_label(l);
> > + }
> > +
> > + *needput = false;
> > + return l;
> > +}
> > +
> > +static inline void aa_put_label_condref(struct aa_label *l, bool needput)
> > +{
> > + if (unlikely(needput))
> > + aa_put_label(l);
> > +}
> > +
> > /**
> > * aa_current_raw_label - find the current tasks confining label
> > *
> > diff --git a/security/apparmor/lsm.c b/security/apparmor/lsm.c
> > index 2cea34657a47..4bf87eac4a56 100644
> > --- a/security/apparmor/lsm.c
> > +++ b/security/apparmor/lsm.c
> > @@ -461,6 +461,7 @@ static int apparmor_file_open(struct file *file)
> > struct aa_file_ctx *fctx = file_ctx(file);
> > struct aa_label *label;
> > int error = 0;
> > + bool needput;
> >
> > if (!path_mediated_fs(file->f_path.dentry))
> > return 0;
> > @@ -477,7 +478,7 @@ static int apparmor_file_open(struct file *file)
> > return 0;
> > }
> >
> > - label = aa_get_newest_cred_label(file->f_cred);
> > + label = aa_get_newest_cred_label_condref(file->f_cred, &needput);
> > if (!unconfined(label)) {
> > struct mnt_idmap *idmap = file_mnt_idmap(file);
> > struct inode *inode = file_inode(file);
> > @@ -494,7 +495,7 @@ static int apparmor_file_open(struct file *file)
> > /* todo cache full allowed permissions set and state */
> > fctx->allow = aa_map_file_to_perms(file);
> > }
> > - aa_put_label(label);
> > + aa_put_label_condref(label, needput);
> >
> > return error;
> > }
--
Mateusz Guzik <mjguzik gmail.com>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
2024-06-20 18:30 ` Mateusz Guzik
@ 2024-06-20 19:03 ` John Johansen
2024-06-21 4:12 ` Neeraj Upadhyay
0 siblings, 1 reply; 6+ messages in thread
From: John Johansen @ 2024-06-20 19:03 UTC (permalink / raw)
To: Mateusz Guzik, Neeraj Upadhyay
Cc: paul, jmorris, serge, linux-kernel, apparmor,
linux-security-module
On 6/20/24 11:30, Mateusz Guzik wrote:
> On Thu, Jun 20, 2024 at 8:23 PM Neeraj Upadhyay <Neeraj.Upadhyay@amd.com> wrote:
>>
>>
>>
>> On 6/20/2024 10:45 PM, Mateusz Guzik wrote:
>>> apparmor: try to avoid refing the label in apparmor_file_open
>>>
>>> If the label is not stale (which is the common case), the fact that the
>>> passed file object holds a reference can be leverged to avoid the
>>
>> Minor: Typo 'leveraged'
>>
>>> ref/unref cycle. Doing so reduces performance impact of apparmor on
>>> parallel open() invocations.
>>>
>>> When benchmarking on a 24-core vm using will-it-scale's open1_process
>>> ("Separate file open"), the results are (ops/s):
>>> before: 6092196
>>> after: 8309726 (+36%)
>>>
>>> Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
>>> ---
>>
>>
>> Trying to understand the changes done here. So, while the file cred can be updated
>> async to the task (referring to the comment from John here [1]), the file cred label
>> cannot change during apparmor_file_open() execution?
>>
>
> Refing a label retains racy vs refing it.
>
> On stock code you can test the flag, determine it's not stale, grab
> the ref and have it become stale immediately after. My patch avoids
> the atomic dance for the common case, does not alter anything
> correctness-wise AFAICS.
>
> I am assuming the race is tolerated and checking here is only done to
> make sure the new label is seen eventually.
>
> Not having the race is possible with a bunch of trickery like seqc,
> but so far does not look like this is necessary.
>
the race is some what tolerated because of the nature of what is being
done with the label. Basically labels go stale with policy updates, and
we do not guarantee atomic policy updates as the locking to do so would
cause a lot of performance issues.
Generally for mediation, the label is used in a read only fashion and
the state is taken at the start of hook entry and used the read side
value for the duration of the hook.
If the hook is to update the label it must take a lock and then get
any updated state before continuing to update the label.
This case in particular is unique in that apparmor doesn't update
the f_cred. The label can go stale but the file will continue to
reference the original label frpm the time the f_cred was set. So there
is no race that the f_creds reference might be put. This does however
mean that policy updates might result in the slow path, having to do
a refcount, getting triggered.
The afore mentioned replacement of unconfined and object delegation
work will change what apparmor is doing here and break this but like
I said that is a problem for those patches in the future.
>>
>> Reviewed-by: Neeraj upadhyay <Neeraj.Upadhyay@amd.com>
>>
>>
>> Thanks
>> Neeraj
>>
>> [1] https://lore.kernel.org/lkml/9bfaeec2-535d-4401-8244-7560f660a065@canonical.com/
>>
>>
>>>
>>> v2:
>>> - reword the commit message
>>>
>>> If you want any changes made to it can you just do them on your own
>>> accord? :) Will be faster for both of us than another mail trip.
>>>
>>> security/apparmor/include/cred.h | 20 ++++++++++++++++++++
>>> security/apparmor/lsm.c | 5 +++--
>>> 2 files changed, 23 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/security/apparmor/include/cred.h b/security/apparmor/include/cred.h
>>> index 58fdc72af664..7265d2f81dd5 100644
>>> --- a/security/apparmor/include/cred.h
>>> +++ b/security/apparmor/include/cred.h
>>> @@ -63,6 +63,26 @@ static inline struct aa_label *aa_get_newest_cred_label(const struct cred *cred)
>>> return aa_get_newest_label(aa_cred_raw_label(cred));
>>> }
>>>
>>> +static inline struct aa_label *aa_get_newest_cred_label_condref(const struct cred *cred,
>>> + bool *needput)
>>> +{
>>> + struct aa_label *l = aa_cred_raw_label(cred);
>>> +
>>> + if (unlikely(label_is_stale(l))) {
>>> + *needput = true;
>>> + return aa_get_newest_label(l);
>>> + }
>>> +
>>> + *needput = false;
>>> + return l;
>>> +}
>>> +
>>> +static inline void aa_put_label_condref(struct aa_label *l, bool needput)
>>> +{
>>> + if (unlikely(needput))
>>> + aa_put_label(l);
>>> +}
>>> +
>>> /**
>>> * aa_current_raw_label - find the current tasks confining label
>>> *
>>> diff --git a/security/apparmor/lsm.c b/security/apparmor/lsm.c
>>> index 2cea34657a47..4bf87eac4a56 100644
>>> --- a/security/apparmor/lsm.c
>>> +++ b/security/apparmor/lsm.c
>>> @@ -461,6 +461,7 @@ static int apparmor_file_open(struct file *file)
>>> struct aa_file_ctx *fctx = file_ctx(file);
>>> struct aa_label *label;
>>> int error = 0;
>>> + bool needput;
>>>
>>> if (!path_mediated_fs(file->f_path.dentry))
>>> return 0;
>>> @@ -477,7 +478,7 @@ static int apparmor_file_open(struct file *file)
>>> return 0;
>>> }
>>>
>>> - label = aa_get_newest_cred_label(file->f_cred);
>>> + label = aa_get_newest_cred_label_condref(file->f_cred, &needput);
>>> if (!unconfined(label)) {
>>> struct mnt_idmap *idmap = file_mnt_idmap(file);
>>> struct inode *inode = file_inode(file);
>>> @@ -494,7 +495,7 @@ static int apparmor_file_open(struct file *file)
>>> /* todo cache full allowed permissions set and state */
>>> fctx->allow = aa_map_file_to_perms(file);
>>> }
>>> - aa_put_label(label);
>>> + aa_put_label_condref(label, needput);
>>>
>>> return error;
>>> }
>
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
2024-06-20 19:03 ` John Johansen
@ 2024-06-21 4:12 ` Neeraj Upadhyay
0 siblings, 0 replies; 6+ messages in thread
From: Neeraj Upadhyay @ 2024-06-21 4:12 UTC (permalink / raw)
To: John Johansen, Mateusz Guzik
Cc: paul, jmorris, serge, linux-kernel, apparmor,
linux-security-module
On 6/21/2024 12:33 AM, John Johansen wrote:
> On 6/20/24 11:30, Mateusz Guzik wrote:
>> On Thu, Jun 20, 2024 at 8:23 PM Neeraj Upadhyay <Neeraj.Upadhyay@amd.com> wrote:
>>>
>>>
>>>
>>> On 6/20/2024 10:45 PM, Mateusz Guzik wrote:
>>>> apparmor: try to avoid refing the label in apparmor_file_open
>>>>
>>>> If the label is not stale (which is the common case), the fact that the
>>>> passed file object holds a reference can be leverged to avoid the
>>>
>>> Minor: Typo 'leveraged'
>>>
>>>> ref/unref cycle. Doing so reduces performance impact of apparmor on
>>>> parallel open() invocations.
>>>>
>>>> When benchmarking on a 24-core vm using will-it-scale's open1_process
>>>> ("Separate file open"), the results are (ops/s):
>>>> before: 6092196
>>>> after: 8309726 (+36%)
>>>>
>>>> Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
>>>> ---
>>>
>>>
>>> Trying to understand the changes done here. So, while the file cred can be updated
>>> async to the task (referring to the comment from John here [1]), the file cred label
>>> cannot change during apparmor_file_open() execution?
>>>
>>
>> Refing a label retains racy vs refing it.
>>
>> On stock code you can test the flag, determine it's not stale, grab
>> the ref and have it become stale immediately after. My patch avoids
>> the atomic dance for the common case, does not alter anything
>> correctness-wise AFAICS.
>>
>> I am assuming the race is tolerated and checking here is only done to
>> make sure the new label is seen eventually.
>>
>> Not having the race is possible with a bunch of trickery like seqc,
>> but so far does not look like this is necessary.
>>
>
> the race is some what tolerated because of the nature of what is being
> done with the label. Basically labels go stale with policy updates, and
> we do not guarantee atomic policy updates as the locking to do so would
> cause a lot of performance issues.
>
> Generally for mediation, the label is used in a read only fashion and
> the state is taken at the start of hook entry and used the read side
> value for the duration of the hook.
>
> If the hook is to update the label it must take a lock and then get
> any updated state before continuing to update the label.
>
> This case in particular is unique in that apparmor doesn't update
> the f_cred. The label can go stale but the file will continue to
> reference the original label frpm the time the f_cred was set. So there
> is no race that the f_creds reference might be put. This does however
> mean that policy updates might result in the slow path, having to do
> a refcount, getting triggered.
>
> The afore mentioned replacement of unconfined and object delegation
> work will change what apparmor is doing here and break this but like
> I said that is a problem for those patches in the future.
>
>
Thank you both for sharing the details!
Thanks
Neeraj
>>>
>>> Reviewed-by: Neeraj upadhyay <Neeraj.Upadhyay@amd.com>
>>>
>>>
>>> Thanks
>>> Neeraj
>>>
>>> [1] https://lore.kernel.org/lkml/9bfaeec2-535d-4401-8244-7560f660a065@canonical.com/
>>>
>>>
>>>>
>>>> v2:
>>>> - reword the commit message
>>>>
>>>> If you want any changes made to it can you just do them on your own
>>>> accord? :) Will be faster for both of us than another mail trip.
>>>>
>>>> security/apparmor/include/cred.h | 20 ++++++++++++++++++++
>>>> security/apparmor/lsm.c | 5 +++--
>>>> 2 files changed, 23 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/security/apparmor/include/cred.h b/security/apparmor/include/cred.h
>>>> index 58fdc72af664..7265d2f81dd5 100644
>>>> --- a/security/apparmor/include/cred.h
>>>> +++ b/security/apparmor/include/cred.h
>>>> @@ -63,6 +63,26 @@ static inline struct aa_label *aa_get_newest_cred_label(const struct cred *cred)
>>>> return aa_get_newest_label(aa_cred_raw_label(cred));
>>>> }
>>>>
>>>> +static inline struct aa_label *aa_get_newest_cred_label_condref(const struct cred *cred,
>>>> + bool *needput)
>>>> +{
>>>> + struct aa_label *l = aa_cred_raw_label(cred);
>>>> +
>>>> + if (unlikely(label_is_stale(l))) {
>>>> + *needput = true;
>>>> + return aa_get_newest_label(l);
>>>> + }
>>>> +
>>>> + *needput = false;
>>>> + return l;
>>>> +}
>>>> +
>>>> +static inline void aa_put_label_condref(struct aa_label *l, bool needput)
>>>> +{
>>>> + if (unlikely(needput))
>>>> + aa_put_label(l);
>>>> +}
>>>> +
>>>> /**
>>>> * aa_current_raw_label - find the current tasks confining label
>>>> *
>>>> diff --git a/security/apparmor/lsm.c b/security/apparmor/lsm.c
>>>> index 2cea34657a47..4bf87eac4a56 100644
>>>> --- a/security/apparmor/lsm.c
>>>> +++ b/security/apparmor/lsm.c
>>>> @@ -461,6 +461,7 @@ static int apparmor_file_open(struct file *file)
>>>> struct aa_file_ctx *fctx = file_ctx(file);
>>>> struct aa_label *label;
>>>> int error = 0;
>>>> + bool needput;
>>>>
>>>> if (!path_mediated_fs(file->f_path.dentry))
>>>> return 0;
>>>> @@ -477,7 +478,7 @@ static int apparmor_file_open(struct file *file)
>>>> return 0;
>>>> }
>>>>
>>>> - label = aa_get_newest_cred_label(file->f_cred);
>>>> + label = aa_get_newest_cred_label_condref(file->f_cred, &needput);
>>>> if (!unconfined(label)) {
>>>> struct mnt_idmap *idmap = file_mnt_idmap(file);
>>>> struct inode *inode = file_inode(file);
>>>> @@ -494,7 +495,7 @@ static int apparmor_file_open(struct file *file)
>>>> /* todo cache full allowed permissions set and state */
>>>> fctx->allow = aa_map_file_to_perms(file);
>>>> }
>>>> - aa_put_label(label);
>>>> + aa_put_label_condref(label, needput);
>>>>
>>>> return error;
>>>> }
>>
>>
>>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
2024-06-20 17:15 [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open Mateusz Guzik
2024-06-20 18:23 ` Neeraj Upadhyay
@ 2024-07-05 6:30 ` kernel test robot
1 sibling, 0 replies; 6+ messages in thread
From: kernel test robot @ 2024-07-05 6:30 UTC (permalink / raw)
To: Mateusz Guzik
Cc: oe-lkp, lkp, apparmor, ying.huang, feng.tang, fengwei.yin,
john.johansen, paul, jmorris, serge, linux-kernel,
linux-security-module, Neeraj.Upadhyay, Mateusz Guzik,
oliver.sang
Hello,
kernel test robot noticed a 93.1% improvement of will-it-scale.per_process_ops on:
commit: 22eddbc906d2fc507e4a01484e2985fdd6f3cb2f ("[PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open")
url: https://github.com/intel-lab-lkp/linux/commits/Mateusz-Guzik/apparmor-try-to-avoid-refing-the-label-in-apparmor_file_open/20240625-143723
base: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git 55027e689933ba2e64f3d245fb1ff185b3e7fc81
patch link: https://lore.kernel.org/all/20240620171528.167997-1-mjguzik@gmail.com/
patch subject: [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open
testcase: will-it-scale
test machine: 104 threads 2 sockets (Skylake) with 192G memory
parameters:
nr_task: 100%
mode: process
test: open2
cpufreq_governor: performance
In addition to that, the commit also has significant impact on the following tests:
+------------------+----------------------------------------------------------------+
| testcase: change | will-it-scale: will-it-scale.per_process_ops 90.7% improvement |
| test machine | 104 threads 2 sockets (Skylake) with 192G memory |
| test parameters | cpufreq_governor=performance |
| | mode=process |
| | nr_task=100% |
| | test=open1 |
+------------------+----------------------------------------------------------------+
Details are as below:
-------------------------------------------------------------------------------------------------->
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20240705/202407051440.a54b7d0f-oliver.sang@intel.com
=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
gcc-13/performance/x86_64-rhel-8.3/process/100%/debian-12-x86_64-20240206.cgz/lkp-skl-fpga01/open2/will-it-scale
commit:
55027e6899 ("Merge tag 'input-for-v6.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input")
22eddbc906 ("apparmor: try to avoid refing the label in apparmor_file_open")
55027e689933ba2e 22eddbc906d2fc507e4a01484e2
---------------- ---------------------------
%stddev %change %stddev
\ | \
4985 ± 4% -11.4% 4415 ± 5% perf-c2c.HITM.local
1788 -3.3% 1729 vmstat.system.cs
0.01 ± 2% -0.0 0.01 mpstat.cpu.all.soft%
19.61 +3.2 22.81 ± 3% mpstat.cpu.all.usr%
12729 ± 3% -8.1% 11703 proc-vmstat.nr_mapped
50690 ± 7% +52.4% 77253 ± 13% proc-vmstat.pgactivate
967492 +2.0% 986437 proc-vmstat.pgfault
4175664 +93.1% 8063854 ± 2% will-it-scale.104.processes
40150 +93.1% 77536 ± 2% will-it-scale.per_process_ops
4175664 +93.1% 8063854 ± 2% will-it-scale.workload
649.36 ± 12% +19.5% 776.11 ± 4% sched_debug.cfs_rq:/.util_avg.min
79.88 ± 16% -25.0% 59.92 ± 4% sched_debug.cfs_rq:/.util_avg.stddev
25.33 ± 8% -51.6% 12.26 ± 9% sched_debug.cpu.clock.stddev
2791 ± 4% -13.1% 2425 ± 8% sched_debug.cpu.curr->pid.min
2958 ± 5% +9.9% 3251 ± 4% sched_debug.cpu.nr_switches.stddev
4.716e+09 +86.3% 8.787e+09 ± 4% perf-stat.i.branch-instructions
1.09 -0.1 0.98 ± 6% perf-stat.i.branch-miss-rate%
51277236 ± 2% +60.0% 82061991 ± 4% perf-stat.i.branch-misses
1741 -2.8% 1692 perf-stat.i.context-switches
11.66 -47.6% 6.11 perf-stat.i.cpi
145.32 +11.1% 161.46 perf-stat.i.cpu-migrations
2.479e+10 +86.9% 4.634e+10 ± 4% perf-stat.i.instructions
0.09 ± 2% +89.9% 0.17 perf-stat.i.ipc
1.09 ± 2% -0.2 0.93 perf-stat.overall.branch-miss-rate%
11.67 -47.1% 6.17 ± 2% perf-stat.overall.cpi
0.09 +89.2% 0.16 ± 2% perf-stat.overall.ipc
1790350 -1.8% 1758176 perf-stat.overall.path-length
4.7e+09 +86.3% 8.758e+09 ± 4% perf-stat.ps.branch-instructions
51036303 ± 2% +60.2% 81758969 ± 4% perf-stat.ps.branch-misses
1734 -2.8% 1685 perf-stat.ps.context-switches
144.39 +11.4% 160.88 perf-stat.ps.cpu-migrations
2.471e+10 +87.0% 4.62e+10 ± 4% perf-stat.ps.instructions
7.476e+12 +89.7% 1.418e+13 ± 2% perf-stat.total.instructions
27.64 -22.1 5.54 ± 7% perf-profile.calltrace.cycles-pp.security_file_open.do_dentry_open.do_open.path_openat.do_filp_open
27.59 -22.1 5.50 ± 7% perf-profile.calltrace.cycles-pp.apparmor_file_open.security_file_open.do_dentry_open.do_open.path_openat
29.20 -21.6 7.60 ± 3% perf-profile.calltrace.cycles-pp.do_dentry_open.do_open.path_openat.do_filp_open.do_sys_openat2
34.98 -19.0 16.00 ± 4% perf-profile.calltrace.cycles-pp.do_open.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat
58.91 -10.1 48.79 perf-profile.calltrace.cycles-pp.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe.open64
58.99 -10.0 48.95 perf-profile.calltrace.cycles-pp.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe.open64
59.31 -9.9 49.38 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.open64
60.42 -9.3 51.07 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.open64
53.49 -8.3 45.16 perf-profile.calltrace.cycles-pp.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat.do_syscall_64
53.68 -8.2 45.50 perf-profile.calltrace.cycles-pp.do_filp_open.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
73.99 -8.1 65.90 perf-profile.calltrace.cycles-pp.open64
3.15 ± 4% -2.5 0.63 ± 16% perf-profile.calltrace.cycles-pp.putname.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
3.90 +0.1 4.03 ± 2% perf-profile.calltrace.cycles-pp.entry_SYSRETQ_unsafe_stack.__close
0.52 +0.4 0.94 ± 3% perf-profile.calltrace.cycles-pp.kmem_cache_alloc_noprof.alloc_empty_file.path_openat.do_filp_open.do_sys_openat2
0.65 +0.5 1.10 ± 2% perf-profile.calltrace.cycles-pp.strncpy_from_user.getname_flags.do_sys_openat2.__x64_sys_openat.do_syscall_64
3.34 ± 3% +0.6 3.90 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.open64
0.87 +0.6 1.51 ± 2% perf-profile.calltrace.cycles-pp.getname_flags.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
7.82 +0.8 8.60 ± 3% perf-profile.calltrace.cycles-pp.entry_SYSRETQ_unsafe_stack.open64
0.00 +0.8 0.78 ± 2% perf-profile.calltrace.cycles-pp.__check_object_size.strncpy_from_user.getname_flags.do_sys_openat2.__x64_sys_openat
0.00 +0.8 0.80 ± 3% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64.open64
0.00 +0.8 0.83 ± 3% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64.__close
1.48 +0.9 2.40 ± 3% perf-profile.calltrace.cycles-pp.link_path_walk.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat
2.19 ± 2% +1.3 3.54 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__close
3.28 ± 5% +2.7 5.95 ± 6% perf-profile.calltrace.cycles-pp.apparmor_current_getsecid_subj.security_current_getsecid_subj.ima_file_check.security_file_post_open.do_open
3.31 ± 5% +2.7 6.00 ± 6% perf-profile.calltrace.cycles-pp.security_current_getsecid_subj.ima_file_check.security_file_post_open.do_open.path_openat
3.62 ± 4% +2.8 6.40 ± 5% perf-profile.calltrace.cycles-pp.ima_file_check.security_file_post_open.do_open.path_openat.do_filp_open
3.66 ± 4% +2.8 6.44 ± 5% perf-profile.calltrace.cycles-pp.security_file_post_open.do_open.path_openat.do_filp_open.do_sys_openat2
12.01 +4.9 16.88 ± 2% perf-profile.calltrace.cycles-pp.apparmor_file_free_security.security_file_free.__fput.__x64_sys_close.do_syscall_64
12.04 +4.9 16.94 ± 2% perf-profile.calltrace.cycles-pp.security_file_free.__fput.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe
14.50 +5.8 20.27 perf-profile.calltrace.cycles-pp.__fput.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
17.92 +6.1 23.98 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__close
16.03 +6.2 22.19 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
15.50 +6.2 21.72 perf-profile.calltrace.cycles-pp.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
25.29 +8.2 33.52 perf-profile.calltrace.cycles-pp.__close
15.10 +8.9 23.97 perf-profile.calltrace.cycles-pp.init_file.alloc_empty_file.path_openat.do_filp_open.do_sys_openat2
14.08 +9.2 23.29 perf-profile.calltrace.cycles-pp.apparmor_file_alloc_security.security_file_alloc.init_file.alloc_empty_file.path_openat
14.36 +9.4 23.73 perf-profile.calltrace.cycles-pp.security_file_alloc.init_file.alloc_empty_file.path_openat.do_filp_open
15.80 +9.4 25.18 perf-profile.calltrace.cycles-pp.alloc_empty_file.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat
27.62 -22.1 5.50 ± 7% perf-profile.children.cycles-pp.apparmor_file_open
27.64 -22.1 5.55 ± 7% perf-profile.children.cycles-pp.security_file_open
29.21 -21.6 7.62 ± 3% perf-profile.children.cycles-pp.do_dentry_open
35.00 -19.0 16.04 ± 4% perf-profile.children.cycles-pp.do_open
58.96 -10.1 48.86 perf-profile.children.cycles-pp.do_sys_openat2
59.02 -10.0 48.97 perf-profile.children.cycles-pp.__x64_sys_openat
53.54 -8.3 45.22 perf-profile.children.cycles-pp.path_openat
53.70 -8.2 45.53 perf-profile.children.cycles-pp.do_filp_open
73.60 -7.6 66.03 perf-profile.children.cycles-pp.open64
75.93 -4.2 71.78 perf-profile.children.cycles-pp.do_syscall_64
79.03 -3.7 75.32 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
3.15 ± 4% -2.5 0.63 ± 16% perf-profile.children.cycles-pp.putname
0.80 ± 10% -0.6 0.16 ± 7% perf-profile.children.cycles-pp.lockref_get_not_dead
0.86 ± 9% -0.6 0.22 ±107% perf-profile.children.cycles-pp.__cmd_record
0.86 ± 9% -0.6 0.22 ±107% perf-profile.children.cycles-pp.cmd_record
0.86 ± 9% -0.6 0.22 ±108% perf-profile.children.cycles-pp.record__mmap_read_evlist
0.87 ± 9% -0.6 0.23 ±104% perf-profile.children.cycles-pp.main
0.87 ± 9% -0.6 0.23 ±104% perf-profile.children.cycles-pp.run_builtin
0.84 ± 9% -0.6 0.21 ±109% perf-profile.children.cycles-pp.perf_mmap__push
0.82 ± 8% -0.6 0.20 ±106% perf-profile.children.cycles-pp.write
0.81 ± 9% -0.6 0.20 ±110% perf-profile.children.cycles-pp.writen
0.80 ± 9% -0.6 0.20 ±110% perf-profile.children.cycles-pp.record__pushfn
0.48 ± 8% -0.4 0.12 ± 99% perf-profile.children.cycles-pp.ksys_write
0.42 ± 9% -0.3 0.11 ±109% perf-profile.children.cycles-pp.vfs_write
0.41 ± 9% -0.3 0.14 ± 6% perf-profile.children.cycles-pp.mnt_get_write_access
0.96 ± 5% -0.1 0.84 ± 6% perf-profile.children.cycles-pp.lockref_put_return
0.61 ± 3% -0.1 0.52 ± 4% perf-profile.children.cycles-pp.entry_SYSCALL_64_safe_stack
0.27 ± 6% -0.1 0.21 ± 10% perf-profile.children.cycles-pp.__fput_sync
0.29 ± 9% -0.1 0.24 ± 2% perf-profile.children.cycles-pp._raw_spin_lock
0.10 ± 6% -0.0 0.08 ± 7% perf-profile.children.cycles-pp.lockref_get
0.06 ± 9% +0.0 0.08 ± 6% perf-profile.children.cycles-pp.make_vfsuid
0.24 ± 6% +0.0 0.27 ± 2% perf-profile.children.cycles-pp.process_measurement
0.13 ± 5% +0.0 0.16 ± 4% perf-profile.children.cycles-pp.fd_install
0.08 ± 4% +0.0 0.11 ± 4% perf-profile.children.cycles-pp.ima_file_free
0.08 ± 5% +0.0 0.13 ± 2% perf-profile.children.cycles-pp.security_inode_permission
0.07 ± 6% +0.0 0.12 ± 7% perf-profile.children.cycles-pp.set_root
0.07 ± 10% +0.0 0.12 ± 6% perf-profile.children.cycles-pp.syscall_exit_to_user_mode_prepare
0.00 +0.1 0.05 ± 7% perf-profile.children.cycles-pp._find_next_zero_bit
0.00 +0.1 0.05 ± 7% perf-profile.children.cycles-pp.dnotify_flush
0.00 +0.1 0.05 ± 7% perf-profile.children.cycles-pp.get_unused_fd_flags
0.06 ± 11% +0.1 0.12 ± 6% perf-profile.children.cycles-pp.rcu_all_qs
0.06 +0.1 0.11 ± 6% perf-profile.children.cycles-pp.x64_sys_call
0.00 +0.1 0.06 ± 9% perf-profile.children.cycles-pp.locks_remove_file
0.00 +0.1 0.06 ± 6% perf-profile.children.cycles-pp.should_failslab
0.12 ± 4% +0.1 0.18 ± 2% perf-profile.children.cycles-pp.evm_file_release
0.10 ± 4% +0.1 0.16 ± 4% perf-profile.children.cycles-pp.nd_jump_root
0.00 +0.1 0.06 ± 9% perf-profile.children.cycles-pp.path_get
0.07 ± 6% +0.1 0.14 ± 9% perf-profile.children.cycles-pp.obj_cgroup_charge
0.02 ± 99% +0.1 0.09 ± 10% perf-profile.children.cycles-pp.file_close_fd_locked
0.00 +0.1 0.06 ± 7% perf-profile.children.cycles-pp.getname
0.00 +0.1 0.07 ± 11% perf-profile.children.cycles-pp.file_ra_state_init
0.00 +0.1 0.07 ± 5% perf-profile.children.cycles-pp.check_stack_object
0.29 ± 6% +0.1 0.36 ± 3% perf-profile.children.cycles-pp.alloc_fd
0.09 ± 7% +0.1 0.16 ± 4% perf-profile.children.cycles-pp.mntput_no_expire
0.16 ± 7% +0.1 0.23 ± 4% perf-profile.children.cycles-pp.path_init
0.00 +0.1 0.07 ± 8% perf-profile.children.cycles-pp.refill_obj_stock
0.20 ± 9% +0.1 0.27 ± 3% perf-profile.children.cycles-pp.file_close_fd
0.10 ± 10% +0.1 0.18 ± 5% perf-profile.children.cycles-pp.__lookup_mnt
0.09 ± 4% +0.1 0.17 ± 4% perf-profile.children.cycles-pp.mod_objcg_state
0.39 ± 4% +0.1 0.48 ± 2% perf-profile.children.cycles-pp.__fsnotify_parent
0.10 ± 4% +0.1 0.19 ± 3% perf-profile.children.cycles-pp.__check_heap_object
0.10 ± 3% +0.1 0.20 ± 3% perf-profile.children.cycles-pp.locks_remove_posix
0.22 ± 5% +0.1 0.32 ± 2% perf-profile.children.cycles-pp.syscall_exit_to_user_mode
0.11 ± 3% +0.1 0.21 ± 2% perf-profile.children.cycles-pp.memset_orig
0.08 ± 6% +0.1 0.18 ± 3% perf-profile.children.cycles-pp.build_open_flags
0.14 ± 17% +0.1 0.25 ± 9% perf-profile.children.cycles-pp.testcase
0.18 ± 2% +0.1 0.30 ± 3% perf-profile.children.cycles-pp.__virt_addr_valid
0.16 ± 2% +0.1 0.28 ± 2% perf-profile.children.cycles-pp.__cond_resched
0.40 ± 6% +0.1 0.54 ± 4% perf-profile.children.cycles-pp.may_open
0.30 +0.2 0.44 ± 2% perf-profile.children.cycles-pp.check_heap_object
0.29 +0.2 0.44 ± 2% perf-profile.children.cycles-pp.security_file_release
0.17 ± 2% +0.2 0.33 ± 4% perf-profile.children.cycles-pp.filp_flush
0.30 ± 5% +0.2 0.46 ± 5% perf-profile.children.cycles-pp.generic_permission
0.22 ± 5% +0.2 0.39 ± 6% perf-profile.children.cycles-pp.__memcg_slab_free_hook
0.35 ± 7% +0.2 0.52 ± 4% perf-profile.children.cycles-pp.__d_lookup_rcu
0.25 ± 6% +0.2 0.44 ± 5% perf-profile.children.cycles-pp.open_last_lookups
0.35 ± 5% +0.2 0.55 ± 4% perf-profile.children.cycles-pp.walk_component
0.26 ± 2% +0.2 0.48 ± 4% perf-profile.children.cycles-pp.__memcg_slab_post_alloc_hook
0.40 ± 6% +0.2 0.63 ± 4% perf-profile.children.cycles-pp.lookup_fast
0.32 ± 3% +0.2 0.55 ± 4% perf-profile.children.cycles-pp.step_into
0.64 ± 5% +0.2 0.88 ± 4% perf-profile.children.cycles-pp.inode_permission
0.50 +0.3 0.82 ± 2% perf-profile.children.cycles-pp.__check_object_size
12.65 +0.4 13.04 perf-profile.children.cycles-pp.entry_SYSRETQ_unsafe_stack
0.66 +0.5 1.12 ± 2% perf-profile.children.cycles-pp.strncpy_from_user
1.42 ± 2% +0.6 2.01 ± 3% perf-profile.children.cycles-pp.entry_SYSCALL_64
0.87 +0.6 1.52 ± 2% perf-profile.children.cycles-pp.getname_flags
0.93 ± 2% +0.7 1.64 ± 3% perf-profile.children.cycles-pp.kmem_cache_alloc_noprof
1.50 +0.9 2.42 ± 3% perf-profile.children.cycles-pp.link_path_walk
5.64 ± 2% +1.9 7.57 perf-profile.children.cycles-pp.syscall_return_via_sysret
3.29 ± 5% +2.7 5.96 ± 6% perf-profile.children.cycles-pp.apparmor_current_getsecid_subj
3.31 ± 5% +2.7 6.01 ± 6% perf-profile.children.cycles-pp.security_current_getsecid_subj
3.64 ± 4% +2.8 6.41 ± 5% perf-profile.children.cycles-pp.ima_file_check
3.66 ± 4% +2.8 6.44 ± 5% perf-profile.children.cycles-pp.security_file_post_open
12.02 +4.9 16.89 ± 2% perf-profile.children.cycles-pp.apparmor_file_free_security
12.06 +4.9 16.94 ± 2% perf-profile.children.cycles-pp.security_file_free
14.54 +5.8 20.33 perf-profile.children.cycles-pp.__fput
15.51 +6.2 21.74 perf-profile.children.cycles-pp.__x64_sys_close
25.38 +8.3 33.68 perf-profile.children.cycles-pp.__close
15.14 +8.9 24.00 perf-profile.children.cycles-pp.init_file
14.10 +9.2 23.31 perf-profile.children.cycles-pp.apparmor_file_alloc_security
14.37 +9.4 23.74 perf-profile.children.cycles-pp.security_file_alloc
15.82 +9.4 25.20 perf-profile.children.cycles-pp.alloc_empty_file
27.52 -22.0 5.48 ± 7% perf-profile.self.cycles-pp.apparmor_file_open
3.14 ± 4% -2.5 0.63 ± 16% perf-profile.self.cycles-pp.putname
0.80 ± 10% -0.6 0.16 ± 8% perf-profile.self.cycles-pp.lockref_get_not_dead
0.76 ± 12% -0.5 0.23 ± 30% perf-profile.self.cycles-pp.init_file
0.41 ± 8% -0.3 0.14 ± 7% perf-profile.self.cycles-pp.mnt_get_write_access
0.79 ± 5% -0.2 0.57 ± 4% perf-profile.self.cycles-pp.kmem_cache_free
0.94 ± 5% -0.1 0.81 ± 6% perf-profile.self.cycles-pp.lockref_put_return
0.54 ± 2% -0.1 0.44 ± 2% perf-profile.self.cycles-pp.do_syscall_64
0.26 ± 6% -0.1 0.20 ± 10% perf-profile.self.cycles-pp.__fput_sync
0.29 ± 9% -0.1 0.24 ± 2% perf-profile.self.cycles-pp._raw_spin_lock
0.10 ± 6% -0.0 0.08 ± 7% perf-profile.self.cycles-pp.lockref_get
0.06 ± 8% +0.0 0.07 ± 5% perf-profile.self.cycles-pp.path_init
0.05 ± 8% +0.0 0.07 perf-profile.self.cycles-pp.make_vfsuid
0.12 ± 5% +0.0 0.14 ± 3% perf-profile.self.cycles-pp.fd_install
0.24 ± 6% +0.0 0.27 perf-profile.self.cycles-pp.process_measurement
0.05 +0.0 0.08 ± 8% perf-profile.self.cycles-pp.dput
0.07 ± 5% +0.0 0.10 ± 3% perf-profile.self.cycles-pp.check_heap_object
0.08 ± 6% +0.0 0.11 ± 3% perf-profile.self.cycles-pp.ima_file_free
0.05 ± 8% +0.0 0.09 ± 7% perf-profile.self.cycles-pp.security_file_alloc
0.08 ± 5% +0.0 0.13 ± 2% perf-profile.self.cycles-pp.security_inode_permission
0.06 +0.0 0.10 ± 7% perf-profile.self.cycles-pp.__x64_sys_openat
0.09 ± 5% +0.0 0.14 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64_safe_stack
0.06 ± 7% +0.0 0.11 ± 5% perf-profile.self.cycles-pp.syscall_exit_to_user_mode_prepare
0.07 ± 5% +0.0 0.12 ± 9% perf-profile.self.cycles-pp.set_root
0.09 ± 4% +0.0 0.14 ± 4% perf-profile.self.cycles-pp.security_file_release
0.06 +0.0 0.11 ± 3% perf-profile.self.cycles-pp.x64_sys_call
0.12 ± 5% +0.1 0.17 ± 2% perf-profile.self.cycles-pp.syscall_exit_to_user_mode
0.00 +0.1 0.05 perf-profile.self.cycles-pp.locks_remove_file
0.00 +0.1 0.05 perf-profile.self.cycles-pp.security_current_getsecid_subj
0.00 +0.1 0.05 perf-profile.self.cycles-pp.security_file_free
0.05 ± 13% +0.1 0.10 ± 7% perf-profile.self.cycles-pp.rcu_all_qs
0.00 +0.1 0.05 ± 7% perf-profile.self.cycles-pp.dnotify_flush
0.00 +0.1 0.05 ± 7% perf-profile.self.cycles-pp.get_unused_fd_flags
0.05 ± 7% +0.1 0.10 ± 7% perf-profile.self.cycles-pp.lookup_fast
0.02 ± 99% +0.1 0.08 ± 4% perf-profile.self.cycles-pp.__x64_sys_close
0.08 ± 7% +0.1 0.13 ± 5% perf-profile.self.cycles-pp.ima_file_check
0.06 ± 8% +0.1 0.11 ± 5% perf-profile.self.cycles-pp.obj_cgroup_charge
0.03 ± 70% +0.1 0.09 ± 4% perf-profile.self.cycles-pp.getname_flags
0.09 ± 5% +0.1 0.15 ± 2% perf-profile.self.cycles-pp.do_sys_openat2
0.00 +0.1 0.06 ± 8% perf-profile.self.cycles-pp.alloc_empty_file
0.08 ± 5% +0.1 0.14 ± 5% perf-profile.self.cycles-pp.open_last_lookups
0.00 +0.1 0.06 perf-profile.self.cycles-pp.try_to_unlazy
0.10 ± 5% +0.1 0.16 ± 6% perf-profile.self.cycles-pp.__cond_resched
0.12 ± 4% +0.1 0.18 ± 2% perf-profile.self.cycles-pp.evm_file_release
0.32 ± 9% +0.1 0.38 ± 5% perf-profile.self.cycles-pp.inode_permission
0.09 ± 9% +0.1 0.16 ± 6% perf-profile.self.cycles-pp.mntput_no_expire
0.00 +0.1 0.06 ± 7% perf-profile.self.cycles-pp.check_stack_object
0.02 ±141% +0.1 0.08 ± 8% perf-profile.self.cycles-pp.file_close_fd_locked
0.02 ±141% +0.1 0.08 ± 5% perf-profile.self.cycles-pp.filp_flush
0.04 ± 44% +0.1 0.11 ± 8% perf-profile.self.cycles-pp.walk_component
0.00 +0.1 0.07 ± 8% perf-profile.self.cycles-pp.refill_obj_stock
0.07 +0.1 0.14 ± 2% perf-profile.self.cycles-pp.__check_object_size
0.10 ± 9% +0.1 0.17 ± 6% perf-profile.self.cycles-pp.__lookup_mnt
0.10 ± 10% +0.1 0.18 ± 9% perf-profile.self.cycles-pp.__memcg_slab_free_hook
0.09 ± 4% +0.1 0.17 ± 4% perf-profile.self.cycles-pp.mod_objcg_state
0.39 ± 4% +0.1 0.48 ± 2% perf-profile.self.cycles-pp.__fsnotify_parent
0.09 ± 4% +0.1 0.17 ± 6% perf-profile.self.cycles-pp.may_open
0.08 ± 5% +0.1 0.17 ± 3% perf-profile.self.cycles-pp.alloc_fd
0.10 ± 4% +0.1 0.19 ± 3% perf-profile.self.cycles-pp.__check_heap_object
0.12 ± 19% +0.1 0.21 ± 10% perf-profile.self.cycles-pp.testcase
0.11 +0.1 0.20 ± 3% perf-profile.self.cycles-pp.memset_orig
0.10 ± 3% +0.1 0.19 ± 3% perf-profile.self.cycles-pp.locks_remove_posix
0.07 ± 6% +0.1 0.18 ± 3% perf-profile.self.cycles-pp.build_open_flags
0.12 ± 3% +0.1 0.23 ± 5% perf-profile.self.cycles-pp.path_openat
0.17 ± 2% +0.1 0.28 ± 2% perf-profile.self.cycles-pp.__virt_addr_valid
0.15 ± 3% +0.1 0.27 ± 3% perf-profile.self.cycles-pp.__memcg_slab_post_alloc_hook
0.14 ± 5% +0.1 0.26 ± 2% perf-profile.self.cycles-pp.do_open
0.16 ± 2% +0.1 0.30 ± 4% perf-profile.self.cycles-pp.strncpy_from_user
0.16 ± 4% +0.1 0.30 ± 3% perf-profile.self.cycles-pp.do_filp_open
0.26 ± 6% +0.1 0.41 ± 5% perf-profile.self.cycles-pp.generic_permission
0.21 ± 4% +0.2 0.36 ± 7% perf-profile.self.cycles-pp.step_into
0.34 ± 8% +0.2 0.52 ± 4% perf-profile.self.cycles-pp.__d_lookup_rcu
0.46 ± 2% +0.3 0.77 ± 2% perf-profile.self.cycles-pp.link_path_walk
12.60 +0.3 12.93 perf-profile.self.cycles-pp.entry_SYSRETQ_unsafe_stack
0.50 ± 3% +0.3 0.83 ± 3% perf-profile.self.cycles-pp.kmem_cache_alloc_noprof
0.42 ± 3% +0.4 0.78 ± 7% perf-profile.self.cycles-pp.__close
3.12 +0.5 3.58 ± 2% perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe
1.29 ± 2% +0.5 1.77 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64
0.57 ± 8% +0.5 1.10 ± 4% perf-profile.self.cycles-pp.open64
0.83 ± 2% +0.6 1.40 ± 9% perf-profile.self.cycles-pp.do_dentry_open
1.16 ± 6% +0.7 1.85 ± 8% perf-profile.self.cycles-pp.__fput
5.64 ± 2% +1.9 7.56 perf-profile.self.cycles-pp.syscall_return_via_sysret
3.27 ± 5% +2.7 5.93 ± 6% perf-profile.self.cycles-pp.apparmor_current_getsecid_subj
11.98 +4.8 16.83 ± 2% perf-profile.self.cycles-pp.apparmor_file_free_security
14.04 +9.2 23.20 perf-profile.self.cycles-pp.apparmor_file_alloc_security
***************************************************************************************************
lkp-skl-fpga01: 104 threads 2 sockets (Skylake) with 192G memory
=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
gcc-13/performance/x86_64-rhel-8.3/process/100%/debian-12-x86_64-20240206.cgz/lkp-skl-fpga01/open1/will-it-scale
commit:
55027e6899 ("Merge tag 'input-for-v6.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input")
22eddbc906 ("apparmor: try to avoid refing the label in apparmor_file_open")
55027e689933ba2e 22eddbc906d2fc507e4a01484e2
---------------- ---------------------------
%stddev %change %stddev
\ | \
0.01 ± 3% -0.0 0.01 mpstat.cpu.all.soft%
19.35 ± 2% +4.0 23.35 ± 3% mpstat.cpu.all.usr%
21320 -1.0% 21111 proc-vmstat.nr_kernel_stack
12782 ± 2% -7.8% 11784 ± 2% proc-vmstat.nr_mapped
4188213 ± 2% +90.7% 7986330 ± 2% will-it-scale.104.processes
40271 ± 2% +90.7% 76791 ± 2% will-it-scale.per_process_ops
4188213 ± 2% +90.7% 7986330 ± 2% will-it-scale.workload
0.16 ± 10% +18.1% 0.19 ± 6% sched_debug.cfs_rq:/.h_nr_running.stddev
86.01 ± 6% -20.8% 68.12 ± 10% sched_debug.cfs_rq:/.util_avg.stddev
25.34 ± 6% -51.7% 12.24 ± 8% sched_debug.cpu.clock.stddev
0.15 ± 11% +22.0% 0.19 ± 7% sched_debug.cpu.nr_running.stddev
4.37e+09 ± 3% +86.1% 8.131e+09 ± 2% perf-stat.i.branch-instructions
1.22 ± 5% -0.1 1.07 ± 5% perf-stat.i.branch-miss-rate%
50836606 ± 2% +59.8% 81252843 ± 3% perf-stat.i.branch-misses
1732 -2.8% 1683 perf-stat.i.context-switches
12.25 ± 2% -46.3% 6.58 ± 2% perf-stat.i.cpi
144.30 +11.1% 160.37 perf-stat.i.cpu-migrations
2.28e+10 ± 3% +86.7% 4.256e+10 ± 2% perf-stat.i.instructions
0.09 ± 6% +74.9% 0.16 ± 4% perf-stat.i.ipc
0.92 ± 29% -54.4% 0.42 ± 39% perf-stat.overall.MPKI
1.16 -0.2 1.00 perf-stat.overall.branch-miss-rate%
12.45 ± 2% -46.4% 6.67 perf-stat.overall.cpi
0.08 ± 2% +86.7% 0.15 perf-stat.overall.ipc
1676348 -2.2% 1639350 perf-stat.overall.path-length
4.358e+09 ± 3% +86.0% 8.105e+09 ± 2% perf-stat.ps.branch-instructions
50625012 ± 2% +59.9% 80948460 ± 2% perf-stat.ps.branch-misses
1726 -2.9% 1677 perf-stat.ps.context-switches
143.46 +11.4% 159.75 perf-stat.ps.cpu-migrations
2.273e+10 ± 3% +86.6% 4.242e+10 ± 2% perf-stat.ps.instructions
7.021e+12 ± 2% +86.5% 1.309e+13 ± 2% perf-stat.total.instructions
27.80 -22.5 5.26 ± 13% perf-profile.calltrace.cycles-pp.security_file_open.do_dentry_open.do_open.path_openat.do_filp_open
27.75 -22.5 5.22 ± 13% perf-profile.calltrace.cycles-pp.apparmor_file_open.security_file_open.do_dentry_open.do_open.path_openat
29.25 -21.9 7.33 ± 7% perf-profile.calltrace.cycles-pp.do_dentry_open.do_open.path_openat.do_filp_open.do_sys_openat2
34.76 -19.3 15.41 ± 6% perf-profile.calltrace.cycles-pp.do_open.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat
58.25 -10.5 47.77 perf-profile.calltrace.cycles-pp.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe.open64
58.34 -10.4 47.93 perf-profile.calltrace.cycles-pp.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe.open64
58.68 -10.3 48.38 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.open64
60.03 -9.9 50.17 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.open64
52.68 -8.9 43.74 ± 2% perf-profile.calltrace.cycles-pp.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat.do_syscall_64
52.86 -8.8 44.08 ± 2% perf-profile.calltrace.cycles-pp.do_filp_open.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
73.80 -8.6 65.17 perf-profile.calltrace.cycles-pp.open64
3.29 ± 5% -2.3 1.02 ± 18% perf-profile.calltrace.cycles-pp.putname.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.30 ± 5% -0.3 0.97 ± 9% perf-profile.calltrace.cycles-pp.__legitimize_path.try_to_unlazy.complete_walk.do_open.path_openat
1.38 ± 4% -0.3 1.08 ± 8% perf-profile.calltrace.cycles-pp.try_to_unlazy.complete_walk.do_open.path_openat.do_filp_open
1.43 ± 4% -0.3 1.14 ± 8% perf-profile.calltrace.cycles-pp.complete_walk.do_open.path_openat.do_filp_open.do_sys_openat2
0.56 ± 8% +0.2 0.73 ± 6% perf-profile.calltrace.cycles-pp.__legitimize_mnt.__legitimize_path.try_to_unlazy.complete_walk.do_open
3.78 ± 3% +0.3 4.06 ± 2% perf-profile.calltrace.cycles-pp.entry_SYSRETQ_unsafe_stack.__close
0.55 ± 3% +0.4 0.94 ± 3% perf-profile.calltrace.cycles-pp.kmem_cache_alloc_noprof.alloc_empty_file.path_openat.do_filp_open.do_sys_openat2
0.63 ± 2% +0.5 1.09 perf-profile.calltrace.cycles-pp.strncpy_from_user.getname_flags.do_sys_openat2.__x64_sys_openat.do_syscall_64
0.85 ± 2% +0.6 1.49 perf-profile.calltrace.cycles-pp.getname_flags.do_sys_openat2.__x64_sys_openat.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.87 +0.7 1.53 perf-profile.calltrace.cycles-pp.link_path_walk.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat
3.15 ± 3% +0.7 3.89 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.open64
8.06 +0.7 8.81 ± 4% perf-profile.calltrace.cycles-pp.entry_SYSRETQ_unsafe_stack.open64
0.00 +0.8 0.76 ± 2% perf-profile.calltrace.cycles-pp.__check_object_size.strncpy_from_user.getname_flags.do_sys_openat2.__x64_sys_openat
0.00 +0.8 0.78 ± 2% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64.open64
0.00 +0.8 0.81 ± 2% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64.__close
2.30 ± 6% +1.5 3.80 ± 12% perf-profile.calltrace.cycles-pp.syscall_return_via_sysret.__close
3.17 ± 4% +2.5 5.66 ± 9% perf-profile.calltrace.cycles-pp.apparmor_current_getsecid_subj.security_current_getsecid_subj.ima_file_check.security_file_post_open.do_open
3.20 ± 4% +2.5 5.71 ± 9% perf-profile.calltrace.cycles-pp.security_current_getsecid_subj.ima_file_check.security_file_post_open.do_open.path_openat
3.52 ± 3% +2.6 6.11 ± 8% perf-profile.calltrace.cycles-pp.ima_file_check.security_file_post_open.do_open.path_openat.do_filp_open
3.55 ± 3% +2.6 6.16 ± 8% perf-profile.calltrace.cycles-pp.security_file_post_open.do_open.path_openat.do_filp_open.do_sys_openat2
12.05 +4.9 16.92 ± 4% perf-profile.calltrace.cycles-pp.apparmor_file_free_security.security_file_free.__fput.__x64_sys_close.do_syscall_64
12.09 +4.9 16.98 ± 4% perf-profile.calltrace.cycles-pp.security_file_free.__fput.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe
14.89 +5.7 20.57 ± 2% perf-profile.calltrace.cycles-pp.__fput.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
18.30 +6.0 24.31 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__close
16.42 +6.1 22.51 ± 2% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
15.87 +6.1 22.00 ± 2% perf-profile.calltrace.cycles-pp.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
25.66 +8.4 34.10 perf-profile.calltrace.cycles-pp.__close
15.06 +8.9 23.97 perf-profile.calltrace.cycles-pp.init_file.alloc_empty_file.path_openat.do_filp_open.do_sys_openat2
14.03 +9.2 23.25 perf-profile.calltrace.cycles-pp.apparmor_file_alloc_security.security_file_alloc.init_file.alloc_empty_file.path_openat
15.78 +9.4 25.16 perf-profile.calltrace.cycles-pp.alloc_empty_file.path_openat.do_filp_open.do_sys_openat2.__x64_sys_openat
14.30 +9.4 23.69 perf-profile.calltrace.cycles-pp.security_file_alloc.init_file.alloc_empty_file.path_openat.do_filp_open
27.78 -22.6 5.22 ± 13% perf-profile.children.cycles-pp.apparmor_file_open
27.80 -22.5 5.27 ± 13% perf-profile.children.cycles-pp.security_file_open
29.26 -21.9 7.35 ± 7% perf-profile.children.cycles-pp.do_dentry_open
34.78 -19.3 15.45 ± 6% perf-profile.children.cycles-pp.do_open
58.30 -10.5 47.84 perf-profile.children.cycles-pp.do_sys_openat2
58.36 -10.4 47.95 perf-profile.children.cycles-pp.__x64_sys_openat
52.72 -8.9 43.80 ± 2% perf-profile.children.cycles-pp.path_openat
52.89 -8.8 44.10 ± 2% perf-profile.children.cycles-pp.do_filp_open
73.30 -8.0 65.30 perf-profile.children.cycles-pp.open64
75.66 -4.5 71.18 perf-profile.children.cycles-pp.do_syscall_64
79.01 -4.2 74.84 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
3.29 ± 5% -2.3 1.02 ± 18% perf-profile.children.cycles-pp.putname
0.72 ± 5% -0.5 0.22 ± 22% perf-profile.children.cycles-pp.lockref_get_not_dead
0.81 ± 10% -0.4 0.37 ± 84% perf-profile.children.cycles-pp.perf_mmap__push
0.78 ± 10% -0.4 0.35 ± 85% perf-profile.children.cycles-pp.writen
0.77 ± 10% -0.4 0.35 ± 84% perf-profile.children.cycles-pp.record__pushfn
1.31 ± 5% -0.3 0.98 ± 9% perf-profile.children.cycles-pp.__legitimize_path
1.40 ± 4% -0.3 1.08 ± 8% perf-profile.children.cycles-pp.try_to_unlazy
1.44 ± 4% -0.3 1.15 ± 8% perf-profile.children.cycles-pp.complete_walk
0.46 ± 9% -0.3 0.20 ± 78% perf-profile.children.cycles-pp.ksys_write
0.38 ± 5% -0.2 0.16 ± 10% perf-profile.children.cycles-pp.mnt_get_write_access
0.42 ± 10% -0.2 0.19 ± 78% perf-profile.children.cycles-pp.vfs_write
0.33 ± 9% -0.2 0.12 ± 86% perf-profile.children.cycles-pp.shmem_file_write_iter
0.29 ± 10% -0.2 0.10 ± 81% perf-profile.children.cycles-pp.generic_perform_write
0.32 ± 11% -0.1 0.25 ± 3% perf-profile.children.cycles-pp._raw_spin_lock
0.24 +0.0 0.26 ± 5% perf-profile.children.cycles-pp.__hrtimer_run_queues
0.16 ± 8% +0.0 0.18 ± 4% perf-profile.children.cycles-pp.update_process_times
0.18 ± 8% +0.0 0.21 ± 4% perf-profile.children.cycles-pp.tick_nohz_handler
0.08 ± 4% +0.0 0.11 perf-profile.children.cycles-pp.ima_file_free
0.13 ± 2% +0.0 0.17 ± 5% perf-profile.children.cycles-pp.fd_install
0.31 ± 4% +0.0 0.35 ± 2% perf-profile.children.cycles-pp.alloc_fd
0.08 ± 6% +0.0 0.12 ± 4% perf-profile.children.cycles-pp.set_root
0.07 ± 9% +0.0 0.11 ± 6% perf-profile.children.cycles-pp.syscall_exit_to_user_mode_prepare
0.11 ± 6% +0.0 0.16 ± 3% perf-profile.children.cycles-pp.nd_jump_root
0.05 +0.0 0.09 ± 5% perf-profile.children.cycles-pp.security_inode_permission
0.06 ± 7% +0.0 0.11 ± 3% perf-profile.children.cycles-pp.rcu_all_qs
0.18 ± 5% +0.0 0.22 ± 4% perf-profile.children.cycles-pp.path_init
0.09 ± 9% +0.0 0.14 ± 8% perf-profile.children.cycles-pp.mntput_no_expire
0.06 +0.1 0.11 perf-profile.children.cycles-pp.x64_sys_call
0.00 +0.1 0.05 perf-profile.children.cycles-pp._find_next_zero_bit
0.00 +0.1 0.05 perf-profile.children.cycles-pp.dnotify_flush
0.00 +0.1 0.06 ± 9% perf-profile.children.cycles-pp.get_unused_fd_flags
0.00 +0.1 0.06 ± 9% perf-profile.children.cycles-pp.locks_remove_file
0.40 ± 4% +0.1 0.46 ± 7% perf-profile.children.cycles-pp.__fsnotify_parent
0.00 +0.1 0.06 ± 6% perf-profile.children.cycles-pp.check_stack_object
0.00 +0.1 0.06 ± 9% perf-profile.children.cycles-pp.path_get
0.00 +0.1 0.06 perf-profile.children.cycles-pp.should_failslab
0.07 ± 9% +0.1 0.13 ± 10% perf-profile.children.cycles-pp.obj_cgroup_charge
0.00 +0.1 0.06 ± 7% perf-profile.children.cycles-pp.make_vfsuid
0.00 +0.1 0.07 ± 7% perf-profile.children.cycles-pp.file_ra_state_init
0.00 +0.1 0.07 ± 7% perf-profile.children.cycles-pp.getname
0.02 ±141% +0.1 0.08 ± 5% perf-profile.children.cycles-pp.file_close_fd_locked
0.20 ± 9% +0.1 0.27 ± 5% perf-profile.children.cycles-pp.file_close_fd
0.09 ± 5% +0.1 0.16 ± 3% perf-profile.children.cycles-pp.mod_objcg_state
0.00 +0.1 0.07 ± 10% perf-profile.children.cycles-pp.refill_obj_stock
0.10 ± 3% +0.1 0.17 ± 4% perf-profile.children.cycles-pp.__lookup_mnt
0.10 ± 6% +0.1 0.19 ± 2% perf-profile.children.cycles-pp.__check_heap_object
0.15 ± 3% +0.1 0.24 ± 3% perf-profile.children.cycles-pp.walk_component
0.10 +0.1 0.20 ± 2% perf-profile.children.cycles-pp.locks_remove_posix
0.12 ± 3% +0.1 0.22 ± 5% perf-profile.children.cycles-pp.memset_orig
0.23 ± 4% +0.1 0.33 perf-profile.children.cycles-pp.__d_lookup_rcu
0.19 ± 3% +0.1 0.30 perf-profile.children.cycles-pp.__virt_addr_valid
0.24 ± 5% +0.1 0.34 ± 14% perf-profile.children.cycles-pp.syscall_exit_to_user_mode
0.16 ± 3% +0.1 0.28 ± 3% perf-profile.children.cycles-pp.__cond_resched
0.08 ± 6% +0.1 0.19 ± 4% perf-profile.children.cycles-pp.build_open_flags
0.11 ± 19% +0.1 0.24 ± 11% perf-profile.children.cycles-pp.testcase
0.26 ± 3% +0.1 0.40 perf-profile.children.cycles-pp.lookup_fast
0.17 ± 6% +0.2 0.32 ± 3% perf-profile.children.cycles-pp.generic_permission
0.22 ± 6% +0.2 0.38 ± 6% perf-profile.children.cycles-pp.__memcg_slab_free_hook
0.30 ± 2% +0.2 0.46 ± 2% perf-profile.children.cycles-pp.check_heap_object
0.17 ± 2% +0.2 0.33 perf-profile.children.cycles-pp.filp_flush
0.57 ± 8% +0.2 0.73 ± 6% perf-profile.children.cycles-pp.__legitimize_mnt
0.34 ± 2% +0.2 0.50 ± 2% perf-profile.children.cycles-pp.may_open
0.28 ± 4% +0.2 0.46 ± 3% perf-profile.children.cycles-pp.__memcg_slab_post_alloc_hook
0.28 ± 2% +0.2 0.46 ± 2% perf-profile.children.cycles-pp.step_into
0.30 ± 2% +0.2 0.50 ± 2% perf-profile.children.cycles-pp.open_last_lookups
0.42 +0.2 0.64 perf-profile.children.cycles-pp.inode_permission
0.48 ± 2% +0.3 0.80 ± 2% perf-profile.children.cycles-pp.__check_object_size
0.63 ± 2% +0.5 1.10 perf-profile.children.cycles-pp.strncpy_from_user
1.41 +0.6 2.01 ± 2% perf-profile.children.cycles-pp.entry_SYSCALL_64
0.85 ± 2% +0.6 1.49 perf-profile.children.cycles-pp.getname_flags
0.88 +0.7 1.55 perf-profile.children.cycles-pp.link_path_walk
0.95 ± 2% +0.7 1.62 ± 2% perf-profile.children.cycles-pp.kmem_cache_alloc_noprof
5.56 ± 2% +2.3 7.82 ± 5% perf-profile.children.cycles-pp.syscall_return_via_sysret
3.17 ± 4% +2.5 5.66 ± 9% perf-profile.children.cycles-pp.apparmor_current_getsecid_subj
3.20 ± 4% +2.5 5.72 ± 9% perf-profile.children.cycles-pp.security_current_getsecid_subj
3.53 ± 3% +2.6 6.12 ± 8% perf-profile.children.cycles-pp.ima_file_check
3.56 ± 3% +2.6 6.16 ± 8% perf-profile.children.cycles-pp.security_file_post_open
12.07 +4.9 16.94 ± 4% perf-profile.children.cycles-pp.apparmor_file_free_security
12.10 +4.9 16.98 ± 4% perf-profile.children.cycles-pp.security_file_free
14.93 +5.7 20.62 ± 2% perf-profile.children.cycles-pp.__fput
15.88 +6.1 22.02 ± 2% perf-profile.children.cycles-pp.__x64_sys_close
25.75 +8.5 34.26 perf-profile.children.cycles-pp.__close
15.08 +8.9 24.00 perf-profile.children.cycles-pp.init_file
14.05 +9.2 23.27 perf-profile.children.cycles-pp.apparmor_file_alloc_security
14.31 +9.4 23.70 perf-profile.children.cycles-pp.security_file_alloc
15.80 +9.4 25.19 perf-profile.children.cycles-pp.alloc_empty_file
27.68 -22.5 5.20 ± 13% perf-profile.self.cycles-pp.apparmor_file_open
3.28 ± 5% -2.3 1.02 ± 18% perf-profile.self.cycles-pp.putname
0.72 ± 5% -0.5 0.22 ± 21% perf-profile.self.cycles-pp.lockref_get_not_dead
0.76 ± 10% -0.5 0.29 ± 21% perf-profile.self.cycles-pp.init_file
0.38 ± 5% -0.2 0.16 ± 10% perf-profile.self.cycles-pp.mnt_get_write_access
0.78 ± 9% -0.2 0.61 ± 7% perf-profile.self.cycles-pp.kmem_cache_free
0.32 ± 11% -0.1 0.24 ± 4% perf-profile.self.cycles-pp._raw_spin_lock
0.04 ± 44% +0.0 0.07 ± 7% perf-profile.self.cycles-pp.try_to_unlazy
0.05 +0.0 0.08 ± 6% perf-profile.self.cycles-pp.dput
0.08 ± 4% +0.0 0.11 ± 4% perf-profile.self.cycles-pp.ima_file_free
0.10 ± 3% +0.0 0.13 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64_safe_stack
0.06 ± 6% +0.0 0.09 ± 5% perf-profile.self.cycles-pp.rcu_all_qs
0.07 ± 7% +0.0 0.10 ± 7% perf-profile.self.cycles-pp.syscall_exit_to_user_mode_prepare
0.05 ± 8% +0.0 0.09 ± 5% perf-profile.self.cycles-pp.security_file_alloc
0.07 ± 5% +0.0 0.11 ± 4% perf-profile.self.cycles-pp.set_root
0.05 +0.0 0.09 ± 4% perf-profile.self.cycles-pp.security_inode_permission
0.06 +0.0 0.10 ± 3% perf-profile.self.cycles-pp.__x64_sys_openat
0.10 ± 5% +0.0 0.14 ± 2% perf-profile.self.cycles-pp.security_file_release
0.11 ± 12% +0.0 0.15 ± 4% perf-profile.self.cycles-pp.do_sys_openat2
0.06 +0.0 0.10 ± 4% perf-profile.self.cycles-pp.check_heap_object
0.09 ± 9% +0.0 0.14 ± 7% perf-profile.self.cycles-pp.mntput_no_expire
0.06 ± 6% +0.0 0.10 ± 4% perf-profile.self.cycles-pp.x64_sys_call
0.00 +0.1 0.05 perf-profile.self.cycles-pp.dnotify_flush
0.00 +0.1 0.05 perf-profile.self.cycles-pp.file_close_fd
0.00 +0.1 0.05 perf-profile.self.cycles-pp.locks_remove_file
0.23 ± 3% +0.1 0.28 ± 3% perf-profile.self.cycles-pp.inode_permission
0.00 +0.1 0.05 ± 7% perf-profile.self.cycles-pp.complete_walk
0.00 +0.1 0.05 ± 7% perf-profile.self.cycles-pp.get_unused_fd_flags
0.05 ± 7% +0.1 0.10 ± 7% perf-profile.self.cycles-pp.obj_cgroup_charge
0.00 +0.1 0.05 ± 8% perf-profile.self.cycles-pp.check_stack_object
0.40 ± 4% +0.1 0.45 ± 7% perf-profile.self.cycles-pp.__fsnotify_parent
0.00 +0.1 0.06 ± 9% perf-profile.self.cycles-pp.alloc_empty_file
0.03 ± 70% +0.1 0.09 ± 4% perf-profile.self.cycles-pp.getname_flags
0.00 +0.1 0.06 ± 9% perf-profile.self.cycles-pp.security_current_getsecid_subj
0.08 +0.1 0.14 ± 6% perf-profile.self.cycles-pp.ima_file_check
0.02 ± 99% +0.1 0.08 ± 5% perf-profile.self.cycles-pp.__x64_sys_close
0.00 +0.1 0.06 perf-profile.self.cycles-pp.lookup_fast
0.00 +0.1 0.06 perf-profile.self.cycles-pp.make_vfsuid
0.10 ± 5% +0.1 0.16 ± 5% perf-profile.self.cycles-pp.__cond_resched
0.10 ± 9% +0.1 0.17 ± 7% perf-profile.self.cycles-pp.__memcg_slab_free_hook
0.13 ± 8% +0.1 0.19 ± 21% perf-profile.self.cycles-pp.syscall_exit_to_user_mode
0.06 ± 6% +0.1 0.12 ± 4% perf-profile.self.cycles-pp.__check_object_size
0.02 ±141% +0.1 0.08 ± 5% perf-profile.self.cycles-pp.filp_flush
0.09 ± 5% +0.1 0.16 ± 3% perf-profile.self.cycles-pp.may_open
0.09 ± 5% +0.1 0.16 ± 2% perf-profile.self.cycles-pp.mod_objcg_state
0.00 +0.1 0.07 ± 10% perf-profile.self.cycles-pp.refill_obj_stock
0.09 ± 6% +0.1 0.16 ± 3% perf-profile.self.cycles-pp.__lookup_mnt
0.08 ± 5% +0.1 0.16 ± 3% perf-profile.self.cycles-pp.alloc_fd
0.00 +0.1 0.08 ± 8% perf-profile.self.cycles-pp.file_close_fd_locked
0.10 ± 6% +0.1 0.18 ± 4% perf-profile.self.cycles-pp.__check_heap_object
0.09 ± 4% +0.1 0.18 ± 3% perf-profile.self.cycles-pp.open_last_lookups
0.10 ± 4% +0.1 0.19 perf-profile.self.cycles-pp.locks_remove_posix
0.17 ± 4% +0.1 0.26 ± 2% perf-profile.self.cycles-pp.__memcg_slab_post_alloc_hook
0.18 ± 5% +0.1 0.28 ± 3% perf-profile.self.cycles-pp.do_open
0.12 ± 4% +0.1 0.21 ± 5% perf-profile.self.cycles-pp.memset_orig
0.22 ± 3% +0.1 0.32 ± 2% perf-profile.self.cycles-pp.__d_lookup_rcu
0.18 ± 3% +0.1 0.28 perf-profile.self.cycles-pp.__virt_addr_valid
0.13 ± 4% +0.1 0.23 ± 6% perf-profile.self.cycles-pp.path_openat
0.08 ± 6% +0.1 0.18 ± 5% perf-profile.self.cycles-pp.build_open_flags
0.18 ± 2% +0.1 0.29 perf-profile.self.cycles-pp.step_into
0.08 ± 21% +0.1 0.20 ± 12% perf-profile.self.cycles-pp.testcase
0.16 ± 5% +0.1 0.29 ± 3% perf-profile.self.cycles-pp.do_filp_open
0.15 ± 7% +0.1 0.28 ± 2% perf-profile.self.cycles-pp.generic_permission
0.16 ± 4% +0.1 0.30 ± 3% perf-profile.self.cycles-pp.strncpy_from_user
0.56 ± 7% +0.2 0.72 ± 6% perf-profile.self.cycles-pp.__legitimize_mnt
0.30 +0.3 0.55 perf-profile.self.cycles-pp.link_path_walk
0.43 ± 3% +0.3 0.73 ± 12% perf-profile.self.cycles-pp.__close
0.49 +0.3 0.82 ± 3% perf-profile.self.cycles-pp.kmem_cache_alloc_noprof
3.36 ± 3% +0.3 3.70 perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe
1.23 ± 6% +0.5 1.70 ± 4% perf-profile.self.cycles-pp.__fput
1.28 +0.5 1.77 ± 3% perf-profile.self.cycles-pp.entry_SYSCALL_64
0.58 ± 6% +0.5 1.10 ± 3% perf-profile.self.cycles-pp.open64
0.78 ± 5% +0.6 1.40 ± 10% perf-profile.self.cycles-pp.do_dentry_open
5.55 ± 2% +2.3 7.82 ± 5% perf-profile.self.cycles-pp.syscall_return_via_sysret
3.16 ± 4% +2.5 5.63 ± 9% perf-profile.self.cycles-pp.apparmor_current_getsecid_subj
12.02 +4.8 16.87 ± 4% perf-profile.self.cycles-pp.apparmor_file_free_security
13.98 +9.2 23.16 perf-profile.self.cycles-pp.apparmor_file_alloc_security
Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2024-07-05 6:30 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-06-20 17:15 [PATCH v2] apparmor: try to avoid refing the label in apparmor_file_open Mateusz Guzik
2024-06-20 18:23 ` Neeraj Upadhyay
2024-06-20 18:30 ` Mateusz Guzik
2024-06-20 19:03 ` John Johansen
2024-06-21 4:12 ` Neeraj Upadhyay
2024-07-05 6:30 ` kernel test robot
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).