diff mbox series

[v3] powerpc/kernel/sysfs: Export ldbar spr to sysfs

Message ID 1510232445-10869-1-git-send-email-anju@linux.vnet.ibm.com (mailing list archive)
State Changes Requested
Headers show
Series [v3] powerpc/kernel/sysfs: Export ldbar spr to sysfs | expand

Commit Message

Anju T Sudhakar Nov. 9, 2017, 1 p.m. UTC
Add ldbar spr to sysfs. The spr holds thread level In-Memory Collection (IMC)                       
counter configuration data. Exposing this will help to understand the current
configuration of thread-level counters in the system.
Primarily, Bit 0 of ldbar says whether the counters are enabled or not.
And bit 1 indicates the mode (if 0-Accumulation Mode/if 1-Trace Mode). 

Signed-off-by: Anju T Sudhakar <anju@linux.vnet.ibm.com>
---
 arch/powerpc/kernel/sysfs.c | 8 ++++++++
 1 file changed, 8 insertions(+)

Comments

Michael Ellerman March 6, 2018, 11:05 a.m. UTC | #1
Anju T Sudhakar <anju@linux.vnet.ibm.com> writes:

> diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
> index 4437c70..caefb64 100644
> --- a/arch/powerpc/kernel/sysfs.c
> +++ b/arch/powerpc/kernel/sysfs.c
> @@ -757,6 +759,9 @@ static int register_cpu_online(unsigned int cpu)
>  			device_create_file(s, &pmc_attrs[i]);
>  
>  #ifdef CONFIG_PPC64
> +	if (cpu_has_feature(CPU_FTR_ARCH_300))
> +		device_create_file(s, &dev_attr_ldbar);

Is this register readable in supervisor state?

cheers
Anju T Sudhakar March 13, 2018, 5:43 a.m. UTC | #2
Hi,


On Tuesday 06 March 2018 04:35 PM, Michael Ellerman wrote:
> Anju T Sudhakar <anju@linux.vnet.ibm.com> writes:
>
>> diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
>> index 4437c70..caefb64 100644
>> --- a/arch/powerpc/kernel/sysfs.c
>> +++ b/arch/powerpc/kernel/sysfs.c
>> @@ -757,6 +759,9 @@ static int register_cpu_online(unsigned int cpu)
>>   			device_create_file(s, &pmc_attrs[i]);
>>   
>>   #ifdef CONFIG_PPC64
>> +	if (cpu_has_feature(CPU_FTR_ARCH_300))
>> +		device_create_file(s, &dev_attr_ldbar);
> Is this register readable in supervisor state?

This is a nice catch, thanks. :)
The guest kernel can not access the register, it is only readable in the 
hypervisor state.
I will resend the patch with a condition check so that this spr will not 
get registered for guest kernel.


Regards,
Anju
>
> cheers
>
diff mbox series

Patch

diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
index 4437c70..caefb64 100644
--- a/arch/powerpc/kernel/sysfs.c
+++ b/arch/powerpc/kernel/sysfs.c
@@ -485,6 +485,7 @@  SYSFS_PMCSETUP(mmcra, SPRN_MMCRA);
 SYSFS_SPRSETUP(purr, SPRN_PURR);
 SYSFS_SPRSETUP(spurr, SPRN_SPURR);
 SYSFS_SPRSETUP(pir, SPRN_PIR);
+SYSFS_SPRSETUP(ldbar, SPRN_LDBAR);
 
 /*
   Lets only enable read for phyp resources and
@@ -492,6 +493,7 @@  SYSFS_SPRSETUP(pir, SPRN_PIR);
   Lets be conservative and default to pseries.
 */
 static DEVICE_ATTR(mmcra, 0600, show_mmcra, store_mmcra);
+static DEVICE_ATTR(ldbar, 0400, show_ldbar, NULL);
 static DEVICE_ATTR(spurr, 0400, show_spurr, NULL);
 static DEVICE_ATTR(purr, 0400, show_purr, store_purr);
 static DEVICE_ATTR(pir, 0400, show_pir, NULL);
@@ -757,6 +759,9 @@  static int register_cpu_online(unsigned int cpu)
 			device_create_file(s, &pmc_attrs[i]);
 
 #ifdef CONFIG_PPC64
+	if (cpu_has_feature(CPU_FTR_ARCH_300))
+		device_create_file(s, &dev_attr_ldbar);
+
 	if (cpu_has_feature(CPU_FTR_MMCRA))
 		device_create_file(s, &dev_attr_mmcra);
 
@@ -842,6 +847,9 @@  static int unregister_cpu_online(unsigned int cpu)
 			device_remove_file(s, &pmc_attrs[i]);
 
 #ifdef CONFIG_PPC64
+	if (cpu_has_feature(CPU_FTR_ARCH_300))
+		device_remove_file(s, &dev_attr_ldbar);
+
 	if (cpu_has_feature(CPU_FTR_MMCRA))
 		device_remove_file(s, &dev_attr_mmcra);