2022-11-02 09:02:40

by Zhen Lei

[permalink] [raw]
Subject: [PATCH v8 6/9] livepatch: Use kallsyms_on_each_match_symbol() to improve performance

Based on the test results of kallsyms_on_each_match_symbol() and
kallsyms_on_each_symbol(), the average performance can be improved by
more than 1500 times.

Signed-off-by: Zhen Lei <[email protected]>
---
kernel/livepatch/core.c | 20 +++++++++++++++++++-
1 file changed, 19 insertions(+), 1 deletion(-)

diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c
index 9ada0bc5247be5d..50bfc3481a4ee38 100644
--- a/kernel/livepatch/core.c
+++ b/kernel/livepatch/core.c
@@ -153,6 +153,24 @@ static int klp_find_callback(void *data, const char *name,
return 0;
}

+static int klp_match_callback(void *data, unsigned long addr)
+{
+ struct klp_find_arg *args = data;
+
+ args->addr = addr;
+ args->count++;
+
+ /*
+ * Finish the search when the symbol is found for the desired position
+ * or the position is not defined for a non-unique symbol.
+ */
+ if ((args->pos && (args->count == args->pos)) ||
+ (!args->pos && (args->count > 1)))
+ return 1;
+
+ return 0;
+}
+
static int klp_find_object_symbol(const char *objname, const char *name,
unsigned long sympos, unsigned long *addr)
{
@@ -167,7 +185,7 @@ static int klp_find_object_symbol(const char *objname, const char *name,
if (objname)
module_kallsyms_on_each_symbol(klp_find_callback, &args);
else
- kallsyms_on_each_symbol(klp_find_callback, &args);
+ kallsyms_on_each_match_symbol(klp_match_callback, name, &args);

/*
* Ensure an address was found. If sympos is 0, ensure symbol is unique;
--
2.25.1



2022-11-23 13:54:58

by Petr Mladek

[permalink] [raw]
Subject: Re: [PATCH v8 6/9] livepatch: Use kallsyms_on_each_match_symbol() to improve performance

Hi,

I am sorry for the late review. I have been snowed under another
tasks.

On Wed 2022-11-02 16:49:18, Zhen Lei wrote:
> Based on the test results of kallsyms_on_each_match_symbol() and
> kallsyms_on_each_symbol(), the average performance can be improved by
> more than 1500 times.

Sounds great.

> --- a/kernel/livepatch/core.c
> +++ b/kernel/livepatch/core.c
> @@ -153,6 +153,24 @@ static int klp_find_callback(void *data, const char *name,
> return 0;
> }
>
> +static int klp_match_callback(void *data, unsigned long addr)
> +{
> + struct klp_find_arg *args = data;
> +
> + args->addr = addr;
> + args->count++;
> +
> + /*
> + * Finish the search when the symbol is found for the desired position
> + * or the position is not defined for a non-unique symbol.
> + */
> + if ((args->pos && (args->count == args->pos)) ||
> + (!args->pos && (args->count > 1)))
> + return 1;
> +
> + return 0;

This duplicates most of the klp_find_callback(). Please, call this
new function in klp_find_callback() instead of the duplicated code.
I mean to do:

static int klp_find_callback(void *data, const char *name, unsigned long addr)
{
struct klp_find_arg *args = data;

if (strcmp(args->name, name))
return 0;

return klp_match_callback(data, addr);
}

Otherwise, it looks good.

Best Regards,
Petr

2022-11-24 03:14:43

by Zhen Lei

[permalink] [raw]
Subject: Re: [PATCH v8 6/9] livepatch: Use kallsyms_on_each_match_symbol() to improve performance



On 2022/11/23 21:28, Petr Mladek wrote:
> Hi,
>
> I am sorry for the late review. I have been snowed under another
> tasks.
>
> On Wed 2022-11-02 16:49:18, Zhen Lei wrote:
>> Based on the test results of kallsyms_on_each_match_symbol() and
>> kallsyms_on_each_symbol(), the average performance can be improved by
>> more than 1500 times.
>
> Sounds great.
>
>> --- a/kernel/livepatch/core.c
>> +++ b/kernel/livepatch/core.c
>> @@ -153,6 +153,24 @@ static int klp_find_callback(void *data, const char *name,
>> return 0;
>> }
>>
>> +static int klp_match_callback(void *data, unsigned long addr)
>> +{
>> + struct klp_find_arg *args = data;
>> +
>> + args->addr = addr;
>> + args->count++;
>> +
>> + /*
>> + * Finish the search when the symbol is found for the desired position
>> + * or the position is not defined for a non-unique symbol.
>> + */
>> + if ((args->pos && (args->count == args->pos)) ||
>> + (!args->pos && (args->count > 1)))
>> + return 1;
>> +
>> + return 0;
>
> This duplicates most of the klp_find_callback(). Please, call this
> new function in klp_find_callback() instead of the duplicated code.
> I mean to do:
>
> static int klp_find_callback(void *data, const char *name, unsigned long addr)
> {
> struct klp_find_arg *args = data;
>
> if (strcmp(args->name, name))
> return 0;
>
> return klp_match_callback(data, addr);
> }

Good idea. But these patches have been merged into linux-next, how about I post
a new cleanup patch after v6.2-rc1?

>
> Otherwise, it looks good.
>
> Best Regards,
> Petr
> .
>

--
Regards,
Zhen Lei

2022-11-24 09:13:47

by Petr Mladek

[permalink] [raw]
Subject: Re: [PATCH v8 6/9] livepatch: Use kallsyms_on_each_match_symbol() to improve performance

On Thu 2022-11-24 10:36:23, Leizhen (ThunderTown) wrote:
> On 2022/11/23 21:28, Petr Mladek wrote:
> > Hi,
> >
> > I am sorry for the late review. I have been snowed under another
> > tasks.
> >
> > On Wed 2022-11-02 16:49:18, Zhen Lei wrote:
> >> Based on the test results of kallsyms_on_each_match_symbol() and
> >> kallsyms_on_each_symbol(), the average performance can be improved by
> >> more than 1500 times.
> >
> > Sounds great.
> >
> >> --- a/kernel/livepatch/core.c
> >> +++ b/kernel/livepatch/core.c
> >> @@ -153,6 +153,24 @@ static int klp_find_callback(void *data, const char *name,
> >> return 0;
> >> }
> >>
> >> +static int klp_match_callback(void *data, unsigned long addr)
> >> +{
> >> + struct klp_find_arg *args = data;
> >> +
> >> + args->addr = addr;
> >> + args->count++;
> >> +
> >> + /*
> >> + * Finish the search when the symbol is found for the desired position
> >> + * or the position is not defined for a non-unique symbol.
> >> + */
> >> + if ((args->pos && (args->count == args->pos)) ||
> >> + (!args->pos && (args->count > 1)))
> >> + return 1;
> >> +
> >> + return 0;
> >
> > This duplicates most of the klp_find_callback(). Please, call this
> > new function in klp_find_callback() instead of the duplicated code.
> > I mean to do:
> >
> > static int klp_find_callback(void *data, const char *name, unsigned long addr)
> > {
> > struct klp_find_arg *args = data;
> >
> > if (strcmp(args->name, name))
> > return 0;
> >
> > return klp_match_callback(data, addr);
> > }
>
> Good idea. But these patches have been merged into linux-next, how about I post
> a new cleanup patch after v6.2-rc1?

I am fine with it.

Best Regards,
Petr

2022-12-06 22:42:16

by Luis Chamberlain

[permalink] [raw]
Subject: Re: [PATCH v8 6/9] livepatch: Use kallsyms_on_each_match_symbol() to improve performance

On Thu, Nov 24, 2022 at 10:36:23AM +0800, Leizhen (ThunderTown) wrote:
>
>
> On 2022/11/23 21:28, Petr Mladek wrote:
> > Hi,
> >
> > I am sorry for the late review. I have been snowed under another
> > tasks.
> >
> > On Wed 2022-11-02 16:49:18, Zhen Lei wrote:
> >> Based on the test results of kallsyms_on_each_match_symbol() and
> >> kallsyms_on_each_symbol(), the average performance can be improved by
> >> more than 1500 times.
> >
> > Sounds great.
> >
> >> --- a/kernel/livepatch/core.c
> >> +++ b/kernel/livepatch/core.c
> >> @@ -153,6 +153,24 @@ static int klp_find_callback(void *data, const char *name,
> >> return 0;
> >> }
> >>
> >> +static int klp_match_callback(void *data, unsigned long addr)
> >> +{
> >> + struct klp_find_arg *args = data;
> >> +
> >> + args->addr = addr;
> >> + args->count++;
> >> +
> >> + /*
> >> + * Finish the search when the symbol is found for the desired position
> >> + * or the position is not defined for a non-unique symbol.
> >> + */
> >> + if ((args->pos && (args->count == args->pos)) ||
> >> + (!args->pos && (args->count > 1)))
> >> + return 1;
> >> +
> >> + return 0;
> >
> > This duplicates most of the klp_find_callback(). Please, call this
> > new function in klp_find_callback() instead of the duplicated code.
> > I mean to do:
> >
> > static int klp_find_callback(void *data, const char *name, unsigned long addr)
> > {
> > struct klp_find_arg *args = data;
> >
> > if (strcmp(args->name, name))
> > return 0;
> >
> > return klp_match_callback(data, addr);
> > }
>
> Good idea. But these patches have been merged into linux-next, how about I post
> a new cleanup patch after v6.2-rc1?

You can send the cleanup now. The code doesn't change drastically, just
base it on modules-next.

Luis

2022-12-07 02:27:46

by Zhen Lei

[permalink] [raw]
Subject: Re: [PATCH v8 6/9] livepatch: Use kallsyms_on_each_match_symbol() to improve performance



On 2022/12/7 6:08, Luis Chamberlain wrote:
> On Thu, Nov 24, 2022 at 10:36:23AM +0800, Leizhen (ThunderTown) wrote:
>>
>>
>> On 2022/11/23 21:28, Petr Mladek wrote:
>>> Hi,
>>>
>>> I am sorry for the late review. I have been snowed under another
>>> tasks.
>>>
>>> On Wed 2022-11-02 16:49:18, Zhen Lei wrote:
>>>> Based on the test results of kallsyms_on_each_match_symbol() and
>>>> kallsyms_on_each_symbol(), the average performance can be improved by
>>>> more than 1500 times.
>>>
>>> Sounds great.
>>>
>>>> --- a/kernel/livepatch/core.c
>>>> +++ b/kernel/livepatch/core.c
>>>> @@ -153,6 +153,24 @@ static int klp_find_callback(void *data, const char *name,
>>>> return 0;
>>>> }
>>>>
>>>> +static int klp_match_callback(void *data, unsigned long addr)
>>>> +{
>>>> + struct klp_find_arg *args = data;
>>>> +
>>>> + args->addr = addr;
>>>> + args->count++;
>>>> +
>>>> + /*
>>>> + * Finish the search when the symbol is found for the desired position
>>>> + * or the position is not defined for a non-unique symbol.
>>>> + */
>>>> + if ((args->pos && (args->count == args->pos)) ||
>>>> + (!args->pos && (args->count > 1)))
>>>> + return 1;
>>>> +
>>>> + return 0;
>>>
>>> This duplicates most of the klp_find_callback(). Please, call this
>>> new function in klp_find_callback() instead of the duplicated code.
>>> I mean to do:
>>>
>>> static int klp_find_callback(void *data, const char *name, unsigned long addr)
>>> {
>>> struct klp_find_arg *args = data;
>>>
>>> if (strcmp(args->name, name))
>>> return 0;
>>>
>>> return klp_match_callback(data, addr);
>>> }
>>
>> Good idea. But these patches have been merged into linux-next, how about I post
>> a new cleanup patch after v6.2-rc1?
>
> You can send the cleanup now. The code doesn't change drastically, just
> base it on modules-next.

OK

>
> Luis
> .
>

--
Regards,
Zhen Lei