[PATCH] Documentation/x86: Update split lock documentation

[PATCH] Documentation/x86: Update split lock documentation

[Tony Luck]

commit b041b525dab9 ("x86/split_lock: Make life miserable for split
lockers") added a delay and serialization of cplit locks. Commit
727209376f49 ("x86/split_lock: Add sysctl to control the misery mode")
provided a sysctl to turn off the misery.

Update the split lock documentation to describe the current state of
the code.

Signed-off-by: Tony Luck <tony.luck@intel.com>
---
 Documentation/x86/buslock.rst | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/Documentation/x86/buslock.rst b/Documentation/x86/buslock.rst
index 7c051e714943..31ec0ef78086 100644
--- a/Documentation/x86/buslock.rst
+++ b/Documentation/x86/buslock.rst
@@ -53,8 +53,14 @@ parameter "split_lock_detect". Here is a summary of different options:
 |off	  	   |Do nothing			|Do nothing		|
 +------------------+----------------------------+-----------------------+
 |warn		   |Kernel OOPs			|Warn once per task and |
-|(default)	   |Warn once per task and	|and continues to run.  |
-|		   |disable future checking	|			|
+|(default)	   |Warn once per task, add a	|and continues to run.  |
+|		   |delay, add synchronization	|			|
+|		   |to prevent more than one	|			|
+|		   |core from executing a	|			|
+|		   |split lock in parallel.	|			|
+|		   |sysctl split_lock_mitigate	|			|
+|		   |can be used to avoid the	|			|
+|		   |delay and synchronization	|			|
 |		   |When both features are	|			|
 |		   |supported, warn in #AC	|			|
 +------------------+----------------------------+-----------------------+
-- 
2.39.2

Re: [PATCH] Documentation/x86: Update split lock documentation

[Fenghua Yu]

Hi, Tony,

On 3/15/23 15:57, Tony Luck wrote:
> commit b041b525dab9 ("x86/split_lock: Make life miserable for split
> lockers") added a delay and serialization of cplit locks. Commit
                                              s/cplit/split/

> 727209376f49 ("x86/split_lock: Add sysctl to control the misery mode")
> provided a sysctl to turn off the misery.
> 
> Update the split lock documentation to describe the current state of
> the code.
> 
> Signed-off-by: Tony Luck <tony.luck@intel.com>
> ---
>   Documentation/x86/buslock.rst | 10 ++++++++--
>   1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/x86/buslock.rst b/Documentation/x86/buslock.rst
> index 7c051e714943..31ec0ef78086 100644
> --- a/Documentation/x86/buslock.rst
> +++ b/Documentation/x86/buslock.rst
> @@ -53,8 +53,14 @@ parameter "split_lock_detect". Here is a summary of different options:
>   |off	  	   |Do nothing			|Do nothing		|
>   +------------------+----------------------------+-----------------------+
>   |warn		   |Kernel OOPs			|Warn once per task and |
> -|(default)	   |Warn once per task and	|and continues to run.  |
> -|		   |disable future checking	|			|
> +|(default)	   |Warn once per task, add a	|and continues to run.  |
> +|		   |delay, add synchronization	|			|
> +|		   |to prevent more than one	|			|
> +|		   |core from executing a	|			|
> +|		   |split lock in parallel.	|			|
> +|		   |sysctl split_lock_mitigate	|			|
> +|		   |can be used to avoid the	|			|
> +|		   |delay and synchronization	|			|
>   |		   |When both features are	|			|
>   |		   |supported, warn in #AC	|			|
>   +------------------+----------------------------+-----------------------+

Other than the typo, looks good to me.

Thanks.

-Fenghua

Re: [PATCH] Documentation/x86: Update split lock documentation

[Guilherme G. Piccoli]

On 15/03/2023 20:13, Fenghua Yu wrote:
> Hi, Tony,
> 
> On 3/15/23 15:57, Tony Luck wrote:
>> commit b041b525dab9 ("x86/split_lock: Make life miserable for split
>> lockers") added a delay and serialization of cplit locks. Commit
>                                               s/cplit/split/
> 
>> 727209376f49 ("x86/split_lock: Add sysctl to control the misery mode")
>> provided a sysctl to turn off the misery.
>>
>> Update the split lock documentation to describe the current state of
>> the code.
>>
>> Signed-off-by: Tony Luck <tony.luck@intel.com>
>> ---
>>   Documentation/x86/buslock.rst | 10 ++++++++--
>>   1 file changed, 8 insertions(+), 2 deletions(-)
>>
>> diff --git a/Documentation/x86/buslock.rst b/Documentation/x86/buslock.rst
>> index 7c051e714943..31ec0ef78086 100644
>> --- a/Documentation/x86/buslock.rst
>> +++ b/Documentation/x86/buslock.rst
>> @@ -53,8 +53,14 @@ parameter "split_lock_detect". Here is a summary of different options:
>>   |off	  	   |Do nothing			|Do nothing		|
>>   +------------------+----------------------------+-----------------------+
>>   |warn		   |Kernel OOPs			|Warn once per task and |
>> -|(default)	   |Warn once per task and	|and continues to run.  |
>> -|		   |disable future checking	|			|
>> +|(default)	   |Warn once per task, add a	|and continues to run.  |
>> +|		   |delay, add synchronization	|			|
>> +|		   |to prevent more than one	|			|
>> +|		   |core from executing a	|			|
>> +|		   |split lock in parallel.	|			|
>> +|		   |sysctl split_lock_mitigate	|			|
>> +|		   |can be used to avoid the	|			|
>> +|		   |delay and synchronization	|			|
>>   |		   |When both features are	|			|
>>   |		   |supported, warn in #AC	|			|
>>   +------------------+----------------------------+-----------------------+
> 
> Other than the typo, looks good to me.
> 
> Thanks.
> 
> -Fenghua

Thanks for fixing the documentation Tony! Also looks great for me.
Cheers,


Guilherme

Re: [PATCH] Documentation/x86: Update split lock documentation

[Jonathan Corbet]

Tony Luck <tony.luck@intel.com> writes:

> commit b041b525dab9 ("x86/split_lock: Make life miserable for split
> lockers") added a delay and serialization of cplit locks. Commit
> 727209376f49 ("x86/split_lock: Add sysctl to control the misery mode")
> provided a sysctl to turn off the misery.
>
> Update the split lock documentation to describe the current state of
> the code.
>
> Signed-off-by: Tony Luck <tony.luck@intel.com>
> ---
>  Documentation/x86/buslock.rst | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)

Applied (with the typo pointed out by Fenghua fixed), thanks.

jon