[PATCH 0/4] Documentation: x86: resctrl_ui.txt fixes and clarification

[PATCH 0/4] Documentation: x86: resctrl_ui.txt fixes and clarification

[James Morse]

While updating resctrl_ui.rst for arm64 support these things stuck
out as being unclear or no longer true.

(I've shortened the CC list to just RDT+Documentation)


Thanks,

James Morse (4):
  Documentation: x86: Contiguous cbm isn't all X86
  Documentation: x86: Remove cdpl2 unspported statement and fix
    capitalisation
  Documentation: x86: Clarify MBA takes MB as referring to mba_sc
  Documentation: x86: fix some typos

 Documentation/x86/resctrl_ui.rst | 30 ++++++++++++++++++------------
 1 file changed, 18 insertions(+), 12 deletions(-)

-- 
2.20.1

[PATCH 1/4] Documentation: x86: Contiguous cbm isn't all X86

[James Morse]

Since commit 4d05bf71f157 ("x86/resctrl: Introduce AMD QOS feature")
resctrl has supported non-contiguous cache bit masks. The interface
for this is currently try-it-and-see.

Update the documentation to say Intel CPUs have this requirement,
instead of X86.

Cc: Babu Moger <Babu.Moger@amd.com>
Signed-off-by: James Morse <james.morse@arm.com>
---
 Documentation/x86/resctrl_ui.rst | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
index 225cfd4daaee..066f94e53418 100644
--- a/Documentation/x86/resctrl_ui.rst
+++ b/Documentation/x86/resctrl_ui.rst
@@ -342,7 +342,7 @@ For cache resources we describe the portion of the cache that is available
 for allocation using a bitmask. The maximum value of the mask is defined
 by each cpu model (and may be different for different cache levels). It
 is found using CPUID, but is also provided in the "info" directory of
-the resctrl file system in "info/{resource}/cbm_mask". X86 hardware
+the resctrl file system in "info/{resource}/cbm_mask". Intel hardware
 requires that these masks have all the '1' bits in a contiguous block. So
 0x3, 0x6 and 0xC are legal 4-bit masks with two bits set, but 0x5, 0x9
 and 0xA are not.  On a system with a 20-bit mask each bit represents 5%
-- 
2.20.1

Re: [PATCH 1/4] Documentation: x86: Contiguous cbm isn't all X86

[Fenghua Yu]

On Fri, Jun 07, 2019 at 04:14:06PM +0100, James Morse wrote:
> Since commit 4d05bf71f157 ("x86/resctrl: Introduce AMD QOS feature")
> resctrl has supported non-contiguous cache bit masks. The interface
> for this is currently try-it-and-see.
> 
> Update the documentation to say Intel CPUs have this requirement,
> instead of X86.
> 
> Cc: Babu Moger <Babu.Moger@amd.com>
> Signed-off-by: James Morse <james.morse@arm.com>
> ---
>  Documentation/x86/resctrl_ui.rst | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
> index 225cfd4daaee..066f94e53418 100644
> --- a/Documentation/x86/resctrl_ui.rst
> +++ b/Documentation/x86/resctrl_ui.rst
> @@ -342,7 +342,7 @@ For cache resources we describe the portion of the cache that is available
>  for allocation using a bitmask. The maximum value of the mask is defined
>  by each cpu model (and may be different for different cache levels). It
>  is found using CPUID, but is also provided in the "info" directory of
> -the resctrl file system in "info/{resource}/cbm_mask". X86 hardware
> +the resctrl file system in "info/{resource}/cbm_mask". Intel hardware
>  requires that these masks have all the '1' bits in a contiguous block. So
>  0x3, 0x6 and 0xC are legal 4-bit masks with two bits set, but 0x5, 0x9
>  and 0xA are not.  On a system with a 20-bit mask each bit represents 5%
> -- 
> 2.20.1
>
 
Acked-by: Fenghua Yu <fenghua.yu@intel.com>

Thanks.

-Fenghua Yu

[PATCH 2/4] Documentation: x86: Remove cdpl2 unspported statement and fix capitalisation

[James Morse]

"L2 cache does not support code and data prioritization". This isn't
true, elsewhere the document says it can be enabled with the cdpl2
mount option.

While we're here, these sample strings have lower-case code/data,
which isn't how the kernel exports them.

Signed-off-by: James Morse <james.morse@arm.com>
---
 Documentation/x86/resctrl_ui.rst | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
index 066f94e53418..638cd987937d 100644
--- a/Documentation/x86/resctrl_ui.rst
+++ b/Documentation/x86/resctrl_ui.rst
@@ -418,16 +418,22 @@ L3 schemata file details (CDP enabled via mount option to resctrl)
 When CDP is enabled L3 control is split into two separate resources
 so you can specify independent masks for code and data like this::
 
-	L3data:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
-	L3code:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
+	L3DATA:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
+	L3CODE:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
 
 L2 schemata file details
 ------------------------
-L2 cache does not support code and data prioritization, so the
-schemata format is always::
+CDP is supported at L2 using the 'cdpl2' mount option. The schemata
+format is either::
 
 	L2:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
 
+or
+
+	L2DATA:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
+	L2CODE:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
+
+
 Memory bandwidth Allocation (default mode)
 ------------------------------------------
 
-- 
2.20.1

Re: [PATCH 2/4] Documentation: x86: Remove cdpl2 unspported statement and fix capitalisation

[Fenghua Yu]

On Fri, Jun 07, 2019 at 04:14:07PM +0100, James Morse wrote:
> "L2 cache does not support code and data prioritization". This isn't
> true, elsewhere the document says it can be enabled with the cdpl2
> mount option.
> 
> While we're here, these sample strings have lower-case code/data,
> which isn't how the kernel exports them.
> 
> Signed-off-by: James Morse <james.morse@arm.com>
> ---
>  Documentation/x86/resctrl_ui.rst | 14 ++++++++++----
>  1 file changed, 10 insertions(+), 4 deletions(-)
> 
> diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
> index 066f94e53418..638cd987937d 100644
> --- a/Documentation/x86/resctrl_ui.rst
> +++ b/Documentation/x86/resctrl_ui.rst
> @@ -418,16 +418,22 @@ L3 schemata file details (CDP enabled via mount option to resctrl)
>  When CDP is enabled L3 control is split into two separate resources
>  so you can specify independent masks for code and data like this::
>  
> -	L3data:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
> -	L3code:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
> +	L3DATA:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
> +	L3CODE:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
>  
>  L2 schemata file details
>  ------------------------
> -L2 cache does not support code and data prioritization, so the
> -schemata format is always::
> +CDP is supported at L2 using the 'cdpl2' mount option. The schemata
> +format is either::
>  
>  	L2:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
>  
> +or
> +
> +	L2DATA:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
> +	L2CODE:<cache_id0>=<cbm>;<cache_id1>=<cbm>;...
> +
> +
>  Memory bandwidth Allocation (default mode)
>  ------------------------------------------
>  
> -- 
> 2.20.1
> 

Acked-by: Fenghua Yu <fenghua.yu@intel.com>

Thanks.

-Fenghua

[PATCH 3/4] Documentation: x86: Clarify MBA takes MB as referring to mba_sc

[James Morse]

"If the MBA is specified in MB then user can enter the max b/w in MB"
is a tautology. How can the user know if the schemata takes a percentage
or a MB/s value?

This is referring to whether the software controller is interpreting
the schemata's value. Make this clear.

Signed-off-by: James Morse <james.morse@arm.com>
---
 Documentation/x86/resctrl_ui.rst | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
index 638cd987937d..866b66aa289b 100644
--- a/Documentation/x86/resctrl_ui.rst
+++ b/Documentation/x86/resctrl_ui.rst
@@ -677,8 +677,8 @@ allocations can overlap or not. The allocations specifies the maximum
 b/w that the group may be able to use and the system admin can configure
 the b/w accordingly.
 
-If the MBA is specified in MB(megabytes) then user can enter the max b/w in MB
-rather than the percentage values.
+If resctrl is using the software controller (mba_sc) then user can enter the
+max b/w in MB rather than the percentage values.
 ::
 
   # echo "L3:0=3;1=c\nMB:0=1024;1=500" > /sys/fs/resctrl/p0/schemata
-- 
2.20.1

Re: [PATCH 3/4] Documentation: x86: Clarify MBA takes MB as referring to mba_sc

[Fenghua Yu]

On Fri, Jun 07, 2019 at 04:14:08PM +0100, James Morse wrote:
> "If the MBA is specified in MB then user can enter the max b/w in MB"
> is a tautology. How can the user know if the schemata takes a percentage
> or a MB/s value?
> 
> This is referring to whether the software controller is interpreting
> the schemata's value. Make this clear.
> 
> Signed-off-by: James Morse <james.morse@arm.com>
> ---
>  Documentation/x86/resctrl_ui.rst | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
> index 638cd987937d..866b66aa289b 100644
> --- a/Documentation/x86/resctrl_ui.rst
> +++ b/Documentation/x86/resctrl_ui.rst
> @@ -677,8 +677,8 @@ allocations can overlap or not. The allocations specifies the maximum
>  b/w that the group may be able to use and the system admin can configure
>  the b/w accordingly.
>  
> -If the MBA is specified in MB(megabytes) then user can enter the max b/w in MB
> -rather than the percentage values.
> +If resctrl is using the software controller (mba_sc) then user can enter the
> +max b/w in MB rather than the percentage values.
>  ::
>  
>    # echo "L3:0=3;1=c\nMB:0=1024;1=500" > /sys/fs/resctrl/p0/schemata
> -- 
> 2.20.1
> 

Acked-by: Fenghua Yu <fenghua.yu@intel.com>

Thanks.

-Fenghua

[PATCH 4/4] Documentation: x86: fix some typos

[James Morse]

These are all obvious typos.

Signed-off-by: James Morse <james.morse@arm.com>
---
 Documentation/x86/resctrl_ui.rst | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
index 866b66aa289b..5368cedfb530 100644
--- a/Documentation/x86/resctrl_ui.rst
+++ b/Documentation/x86/resctrl_ui.rst
@@ -40,7 +40,7 @@ mount options are:
 	Enable the MBA Software Controller(mba_sc) to specify MBA
 	bandwidth in MBps
 
-L2 and L3 CDP are controlled seperately.
+L2 and L3 CDP are controlled separately.
 
 RDT features are orthogonal. A particular system may support only
 monitoring, only control, or both monitoring and control.  Cache
@@ -118,7 +118,7 @@ related to allocation:
 			      Corresponding region is pseudo-locked. No
 			      sharing allowed.
 
-Memory bandwitdh(MB) subdirectory contains the following files
+Memory bandwidth(MB) subdirectory contains the following files
 with respect to allocation:
 
 "min_bandwidth":
@@ -209,7 +209,7 @@ All groups contain the following files:
 	CPUs to/from this group. As with the tasks file a hierarchy is
 	maintained where MON groups may only include CPUs owned by the
 	parent CTRL_MON group.
-	When the resouce group is in pseudo-locked mode this file will
+	When the resource group is in pseudo-locked mode this file will
 	only be readable, reflecting the CPUs associated with the
 	pseudo-locked region.
 
@@ -380,7 +380,7 @@ where L2 external  is 10GBps (hence aggregate L2 external bandwidth is
 240GBps) and L3 external bandwidth is 100GBps. Now a workload with '20
 threads, having 50% bandwidth, each consuming 5GBps' consumes the max L3
 bandwidth of 100GBps although the percentage value specified is only 50%
-<< 100%. Hence increasing the bandwidth percentage will not yeild any
+<< 100%. Hence increasing the bandwidth percentage will not yield any
 more bandwidth. This is because although the L2 external bandwidth still
 has capacity, the L3 external bandwidth is fully used. Also note that
 this would be dependent on number of cores the benchmark is run on.
@@ -398,7 +398,7 @@ In order to mitigate this and make the interface more user friendly,
 resctrl added support for specifying the bandwidth in MBps as well.  The
 kernel underneath would use a software feedback mechanism or a "Software
 Controller(mba_sc)" which reads the actual bandwidth using MBM counters
-and adjust the memowy bandwidth percentages to ensure::
+and adjust the memory bandwidth percentages to ensure::
 
 	"actual bandwidth < user specified bandwidth".
 
-- 
2.20.1

Re: [PATCH 4/4] Documentation: x86: fix some typos

[Fenghua Yu]

On Fri, Jun 07, 2019 at 04:14:09PM +0100, James Morse wrote:
> These are all obvious typos.
> 
> Signed-off-by: James Morse <james.morse@arm.com>
> ---
>  Documentation/x86/resctrl_ui.rst | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/Documentation/x86/resctrl_ui.rst b/Documentation/x86/resctrl_ui.rst
> index 866b66aa289b..5368cedfb530 100644
> --- a/Documentation/x86/resctrl_ui.rst
> +++ b/Documentation/x86/resctrl_ui.rst
> @@ -40,7 +40,7 @@ mount options are:
>  	Enable the MBA Software Controller(mba_sc) to specify MBA
>  	bandwidth in MBps
>  
> -L2 and L3 CDP are controlled seperately.
> +L2 and L3 CDP are controlled separately.
>  
>  RDT features are orthogonal. A particular system may support only
>  monitoring, only control, or both monitoring and control.  Cache
> @@ -118,7 +118,7 @@ related to allocation:
>  			      Corresponding region is pseudo-locked. No
>  			      sharing allowed.
>  
> -Memory bandwitdh(MB) subdirectory contains the following files
> +Memory bandwidth(MB) subdirectory contains the following files
>  with respect to allocation:
>  
>  "min_bandwidth":
> @@ -209,7 +209,7 @@ All groups contain the following files:
>  	CPUs to/from this group. As with the tasks file a hierarchy is
>  	maintained where MON groups may only include CPUs owned by the
>  	parent CTRL_MON group.
> -	When the resouce group is in pseudo-locked mode this file will
> +	When the resource group is in pseudo-locked mode this file will
>  	only be readable, reflecting the CPUs associated with the
>  	pseudo-locked region.
>  
> @@ -380,7 +380,7 @@ where L2 external  is 10GBps (hence aggregate L2 external bandwidth is
>  240GBps) and L3 external bandwidth is 100GBps. Now a workload with '20
>  threads, having 50% bandwidth, each consuming 5GBps' consumes the max L3
>  bandwidth of 100GBps although the percentage value specified is only 50%
> -<< 100%. Hence increasing the bandwidth percentage will not yeild any
> +<< 100%. Hence increasing the bandwidth percentage will not yield any
>  more bandwidth. This is because although the L2 external bandwidth still
>  has capacity, the L3 external bandwidth is fully used. Also note that
>  this would be dependent on number of cores the benchmark is run on.
> @@ -398,7 +398,7 @@ In order to mitigate this and make the interface more user friendly,
>  resctrl added support for specifying the bandwidth in MBps as well.  The
>  kernel underneath would use a software feedback mechanism or a "Software
>  Controller(mba_sc)" which reads the actual bandwidth using MBM counters
> -and adjust the memowy bandwidth percentages to ensure::
> +and adjust the memory bandwidth percentages to ensure::
>  
>  	"actual bandwidth < user specified bandwidth".
>  
> -- 
> 2.20.1
> 

Acked-by: Fenghua Yu <fenghua.yu@intel.com>

Thanks.

-Fenghua

Re: [PATCH 0/4] Documentation: x86: resctrl_ui.txt fixes and clarification

[Jonathan Corbet]

On Fri,  7 Jun 2019 16:14:05 +0100
James Morse <james.morse@arm.com> wrote:

> While updating resctrl_ui.rst for arm64 support these things stuck
> out as being unclear or no longer true.
> 
> (I've shortened the CC list to just RDT+Documentation)
> 
> 
> Thanks,
> 
> James Morse (4):
>   Documentation: x86: Contiguous cbm isn't all X86
>   Documentation: x86: Remove cdpl2 unspported statement and fix
>     capitalisation
>   Documentation: x86: Clarify MBA takes MB as referring to mba_sc
>   Documentation: x86: fix some typos
> 
>  Documentation/x86/resctrl_ui.rst | 30 ++++++++++++++++++------------
>  1 file changed, 18 insertions(+), 12 deletions(-)

I've applied this set, sorry for the delay.

Thanks,

jon