LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH V2 0/1] PCI/AER: add pcie TLP header information in the tracepoint
@ 2018-05-08 19:56 Thomas Tai
  2018-05-08 19:57 ` [PATCH V2] " Thomas Tai
  0 siblings, 1 reply; 4+ messages in thread
From: Thomas Tai @ 2018-05-08 19:56 UTC (permalink / raw)
  To: bhelgaas, rostedt, thomas.tai; +Cc: linux-pci, linux-kernel, mr.nuke.me

Hi Bjorn and Steven,
Would you kindly review the patch for me? following are the
changes details. Please feel free to comment.

Change log since last version:
V2:
1. use u8 for boolean variable
2. check tlp_header_valid before accessing tlp_header
3. replace sizeof(unsigned int) with 4
4. change the variable name buf to tlp_header for better reading

Thank you,
Thomas

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [PATCH V2] PCI/AER: add pcie TLP header information in the tracepoint
  2018-05-08 19:56 [PATCH V2 0/1] PCI/AER: add pcie TLP header information in the tracepoint Thomas Tai
@ 2018-05-08 19:57 ` Thomas Tai
  2018-05-08 21:07   ` Steven Rostedt
  0 siblings, 1 reply; 4+ messages in thread
From: Thomas Tai @ 2018-05-08 19:57 UTC (permalink / raw)
  To: bhelgaas, rostedt, thomas.tai; +Cc: linux-pci, linux-kernel, mr.nuke.me

When a PCIe AER occurs, the TLP header information is
printed in the kernel message but it is missing from
the tracepoint. A userspace program can use this information
in the tracepoint to better analyze problems.

To enable the tracepoint:
echo 1 > /sys/kernel/debug/tracing/events/ras/aer_event/enable

Example tracepoint output:
cat /sys/kernel/debug/tracing/trace
aer_event: 0000:01:00.0
PCIe Bus Error: severity=Uncorrected, non-fatal, Completer Abort
TLP Header={0x0,0x1,0x2,0x3}

Signed-off-by: Thomas Tai <thomas.tai@oracle.com>
---
 drivers/pci/pcie/aer/aerdrv_errprint.c |  4 ++--
 include/ras/ras_event.h                | 22 ++++++++++++++++++----
 2 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/drivers/pci/pcie/aer/aerdrv_errprint.c b/drivers/pci/pcie/aer/aerdrv_errprint.c
index cfc89dd..fd49a8d 100644
--- a/drivers/pci/pcie/aer/aerdrv_errprint.c
+++ b/drivers/pci/pcie/aer/aerdrv_errprint.c
@@ -189,7 +189,7 @@ void aer_print_error(struct pci_dev *dev, struct aer_err_info *info)
 		pci_err(dev, "  Error of this Agent(%04x) is reported first\n", id);
 
 	trace_aer_event(dev_name(&dev->dev), (info->status & ~info->mask),
-			info->severity);
+			info->severity, info->tlp_header_valid, &info->tlp);
 }
 
 void aer_print_port_info(struct pci_dev *dev, struct aer_err_info *info)
@@ -249,6 +249,6 @@ void cper_print_aer(struct pci_dev *dev, int aer_severity,
 		__print_tlp_header(dev, &aer->header_log);
 
 	trace_aer_event(dev_name(&dev->dev), (status & ~mask),
-			aer_severity);
+			aer_severity, tlp_header_valid, &aer->header_log);
 }
 #endif
diff --git a/include/ras/ras_event.h b/include/ras/ras_event.h
index 9c68986..f73b168 100644
--- a/include/ras/ras_event.h
+++ b/include/ras/ras_event.h
@@ -298,30 +298,44 @@
 TRACE_EVENT(aer_event,
 	TP_PROTO(const char *dev_name,
 		 const u32 status,
-		 const u8 severity),
+		 const u8 severity,
+		 const u8 tlp_header_valid,
+		 struct aer_header_log_regs *tlp),
 
-	TP_ARGS(dev_name, status, severity),
+	TP_ARGS(dev_name, status, severity, tlp_header_valid, tlp),
 
 	TP_STRUCT__entry(
 		__string(	dev_name,	dev_name	)
 		__field(	u32,		status		)
 		__field(	u8,		severity	)
+		__field(	u8, 		tlp_header_valid)
+		__array(	u32, 		tlp_header, 4	)
 	),
 
 	TP_fast_assign(
 		__assign_str(dev_name, dev_name);
 		__entry->status		= status;
 		__entry->severity	= severity;
+		__entry->tlp_header_valid = tlp_header_valid;
+		if (__entry->tlp_header_valid) {
+			__entry->tlp_header[0] = tlp->dw0;
+			__entry->tlp_header[1] = tlp->dw1;
+			__entry->tlp_header[2] = tlp->dw2;
+			__entry->tlp_header[3] = tlp->dw3;
+		}
 	),
 
-	TP_printk("%s PCIe Bus Error: severity=%s, %s\n",
+	TP_printk("%s PCIe Bus Error: severity=%s, %s, TLP Header=%s\n",
 		__get_str(dev_name),
 		__entry->severity == AER_CORRECTABLE ? "Corrected" :
 			__entry->severity == AER_FATAL ?
 			"Fatal" : "Uncorrected, non-fatal",
 		__entry->severity == AER_CORRECTABLE ?
 		__print_flags(__entry->status, "|", aer_correctable_errors) :
-		__print_flags(__entry->status, "|", aer_uncorrectable_errors))
+		__print_flags(__entry->status, "|", aer_uncorrectable_errors),
+		__entry->tlp_header_valid ?
+			__print_array(__entry->tlp_header, 4, 4) :
+			"Not available")
 );
 
 /*
-- 
1.8.3.1

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH V2] PCI/AER: add pcie TLP header information in the tracepoint
  2018-05-08 19:57 ` [PATCH V2] " Thomas Tai
@ 2018-05-08 21:07   ` Steven Rostedt
  2018-05-08 22:24     ` Thomas Tai
  0 siblings, 1 reply; 4+ messages in thread
From: Steven Rostedt @ 2018-05-08 21:07 UTC (permalink / raw)
  To: Thomas Tai; +Cc: bhelgaas, linux-pci, linux-kernel, mr.nuke.me

On Tue,  8 May 2018 15:57:00 -0400
Thomas Tai <thomas.tai@oracle.com> wrote:

> When a PCIe AER occurs, the TLP header information is
> printed in the kernel message but it is missing from
> the tracepoint. A userspace program can use this information
> in the tracepoint to better analyze problems.
> 
> To enable the tracepoint:
> echo 1 > /sys/kernel/debug/tracing/events/ras/aer_event/enable
> 
> Example tracepoint output:
> cat /sys/kernel/debug/tracing/trace
> aer_event: 0000:01:00.0
> PCIe Bus Error: severity=Uncorrected, non-fatal, Completer Abort
> TLP Header={0x0,0x1,0x2,0x3}
> 
> Signed-off-by: Thomas Tai <thomas.tai@oracle.com>
> ---
>  drivers/pci/pcie/aer/aerdrv_errprint.c |  4 ++--
>  include/ras/ras_event.h                | 22 ++++++++++++++++++----
>  2 files changed, 20 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/pci/pcie/aer/aerdrv_errprint.c b/drivers/pci/pcie/aer/aerdrv_errprint.c
> index cfc89dd..fd49a8d 100644
> --- a/drivers/pci/pcie/aer/aerdrv_errprint.c
> +++ b/drivers/pci/pcie/aer/aerdrv_errprint.c
> @@ -189,7 +189,7 @@ void aer_print_error(struct pci_dev *dev, struct aer_err_info *info)
>  		pci_err(dev, "  Error of this Agent(%04x) is reported first\n", id);
>  
>  	trace_aer_event(dev_name(&dev->dev), (info->status & ~info->mask),
> -			info->severity);
> +			info->severity, info->tlp_header_valid, &info->tlp);
>  }
>  
>  void aer_print_port_info(struct pci_dev *dev, struct aer_err_info *info)
> @@ -249,6 +249,6 @@ void cper_print_aer(struct pci_dev *dev, int aer_severity,
>  		__print_tlp_header(dev, &aer->header_log);
>  
>  	trace_aer_event(dev_name(&dev->dev), (status & ~mask),
> -			aer_severity);
> +			aer_severity, tlp_header_valid, &aer->header_log);
>  }
>  #endif
> diff --git a/include/ras/ras_event.h b/include/ras/ras_event.h
> index 9c68986..f73b168 100644
> --- a/include/ras/ras_event.h
> +++ b/include/ras/ras_event.h
> @@ -298,30 +298,44 @@
>  TRACE_EVENT(aer_event,
>  	TP_PROTO(const char *dev_name,
>  		 const u32 status,
> -		 const u8 severity),
> +		 const u8 severity,
> +		 const u8 tlp_header_valid,
> +		 struct aer_header_log_regs *tlp),
>  
> -	TP_ARGS(dev_name, status, severity),
> +	TP_ARGS(dev_name, status, severity, tlp_header_valid, tlp),
>  
>  	TP_STRUCT__entry(
>  		__string(	dev_name,	dev_name	)
>  		__field(	u32,		status		)
>  		__field(	u8,		severity	)
> +		__field(	u8, 		tlp_header_valid)
> +		__array(	u32, 		tlp_header, 4	)
>  	),
>  
>  	TP_fast_assign(
>  		__assign_str(dev_name, dev_name);
>  		__entry->status		= status;
>  		__entry->severity	= severity;
> +		__entry->tlp_header_valid = tlp_header_valid;
> +		if (__entry->tlp_header_valid) {

Although the compiler will probably optimize it anyway, but you
probably should have it be:

		if (tlp_header_valid) {

because why add the extra dereference.

> +			__entry->tlp_header[0] = tlp->dw0;
> +			__entry->tlp_header[1] = tlp->dw1;
> +			__entry->tlp_header[2] = tlp->dw2;
> +			__entry->tlp_header[3] = tlp->dw3;
> +		}
>  	),
>  
> -	TP_printk("%s PCIe Bus Error: severity=%s, %s\n",
> +	TP_printk("%s PCIe Bus Error: severity=%s, %s, TLP Header=%s\n",
>  		__get_str(dev_name),
>  		__entry->severity == AER_CORRECTABLE ? "Corrected" :
>  			__entry->severity == AER_FATAL ?
>  			"Fatal" : "Uncorrected, non-fatal",
>  		__entry->severity == AER_CORRECTABLE ?
>  		__print_flags(__entry->status, "|", aer_correctable_errors) :
> -		__print_flags(__entry->status, "|", aer_uncorrectable_errors))
> +		__print_flags(__entry->status, "|", aer_uncorrectable_errors),
> +		__entry->tlp_header_valid ?
> +			__print_array(__entry->tlp_header, 4, 4) :
> +			"Not available")

The rest looks fine to me.

-- Steve

>  );
>  
>  /*

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH V2] PCI/AER: add pcie TLP header information in the tracepoint
  2018-05-08 21:07   ` Steven Rostedt
@ 2018-05-08 22:24     ` Thomas Tai
  0 siblings, 0 replies; 4+ messages in thread
From: Thomas Tai @ 2018-05-08 22:24 UTC (permalink / raw)
  To: Steven Rostedt; +Cc: bhelgaas, linux-pci, linux-kernel, mr.nuke.me



On 2018-05-08 05:07 PM, Steven Rostedt wrote:
> On Tue,  8 May 2018 15:57:00 -0400
> Thomas Tai <thomas.tai@oracle.com> wrote:
> 
>> When a PCIe AER occurs, the TLP header information is
>> printed in the kernel message but it is missing from
>> the tracepoint. A userspace program can use this information
>> in the tracepoint to better analyze problems.
>>
>> To enable the tracepoint:
>> echo 1 > /sys/kernel/debug/tracing/events/ras/aer_event/enable
>>
>> Example tracepoint output:
>> cat /sys/kernel/debug/tracing/trace
>> aer_event: 0000:01:00.0
>> PCIe Bus Error: severity=Uncorrected, non-fatal, Completer Abort
>> TLP Header={0x0,0x1,0x2,0x3}
>>
>> Signed-off-by: Thomas Tai <thomas.tai@oracle.com>
>> ---
>>   drivers/pci/pcie/aer/aerdrv_errprint.c |  4 ++--
>>   include/ras/ras_event.h                | 22 ++++++++++++++++++----
>>   2 files changed, 20 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/pci/pcie/aer/aerdrv_errprint.c b/drivers/pci/pcie/aer/aerdrv_errprint.c
>> index cfc89dd..fd49a8d 100644
>> --- a/drivers/pci/pcie/aer/aerdrv_errprint.c
>> +++ b/drivers/pci/pcie/aer/aerdrv_errprint.c
>> @@ -189,7 +189,7 @@ void aer_print_error(struct pci_dev *dev, struct aer_err_info *info)
>>   		pci_err(dev, "  Error of this Agent(%04x) is reported first\n", id);
>>   
>>   	trace_aer_event(dev_name(&dev->dev), (info->status & ~info->mask),
>> -			info->severity);
>> +			info->severity, info->tlp_header_valid, &info->tlp);
>>   }
>>   
>>   void aer_print_port_info(struct pci_dev *dev, struct aer_err_info *info)
>> @@ -249,6 +249,6 @@ void cper_print_aer(struct pci_dev *dev, int aer_severity,
>>   		__print_tlp_header(dev, &aer->header_log);
>>   
>>   	trace_aer_event(dev_name(&dev->dev), (status & ~mask),
>> -			aer_severity);
>> +			aer_severity, tlp_header_valid, &aer->header_log);
>>   }
>>   #endif
>> diff --git a/include/ras/ras_event.h b/include/ras/ras_event.h
>> index 9c68986..f73b168 100644
>> --- a/include/ras/ras_event.h
>> +++ b/include/ras/ras_event.h
>> @@ -298,30 +298,44 @@
>>   TRACE_EVENT(aer_event,
>>   	TP_PROTO(const char *dev_name,
>>   		 const u32 status,
>> -		 const u8 severity),
>> +		 const u8 severity,
>> +		 const u8 tlp_header_valid,
>> +		 struct aer_header_log_regs *tlp),
>>   
>> -	TP_ARGS(dev_name, status, severity),
>> +	TP_ARGS(dev_name, status, severity, tlp_header_valid, tlp),
>>   
>>   	TP_STRUCT__entry(
>>   		__string(	dev_name,	dev_name	)
>>   		__field(	u32,		status		)
>>   		__field(	u8,		severity	)
>> +		__field(	u8, 		tlp_header_valid)
>> +		__array(	u32, 		tlp_header, 4	)
>>   	),
>>   
>>   	TP_fast_assign(
>>   		__assign_str(dev_name, dev_name);
>>   		__entry->status		= status;
>>   		__entry->severity	= severity;
>> +		__entry->tlp_header_valid = tlp_header_valid;
>> +		if (__entry->tlp_header_valid) {
> 
> Although the compiler will probably optimize it anyway, but you
> probably should have it be:
> 
> 		if (tlp_header_valid) {
> 
> because why add the extra dereference.

Thanks Steven, good thinking. Will change and send v3.

Thomas



> 
>> +			__entry->tlp_header[0] = tlp->dw0;
>> +			__entry->tlp_header[1] = tlp->dw1;
>> +			__entry->tlp_header[2] = tlp->dw2;
>> +			__entry->tlp_header[3] = tlp->dw3;
>> +		}
>>   	),
>>   
>> -	TP_printk("%s PCIe Bus Error: severity=%s, %s\n",
>> +	TP_printk("%s PCIe Bus Error: severity=%s, %s, TLP Header=%s\n",
>>   		__get_str(dev_name),
>>   		__entry->severity == AER_CORRECTABLE ? "Corrected" :
>>   			__entry->severity == AER_FATAL ?
>>   			"Fatal" : "Uncorrected, non-fatal",
>>   		__entry->severity == AER_CORRECTABLE ?
>>   		__print_flags(__entry->status, "|", aer_correctable_errors) :
>> -		__print_flags(__entry->status, "|", aer_uncorrectable_errors))
>> +		__print_flags(__entry->status, "|", aer_uncorrectable_errors),
>> +		__entry->tlp_header_valid ?
>> +			__print_array(__entry->tlp_header, 4, 4) :
>> +			"Not available")
> 
> The rest looks fine to me.
> 
> -- Steve
> 
>>   );
>>   
>>   /*
> 

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2018-05-08 22:24 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-05-08 19:56 [PATCH V2 0/1] PCI/AER: add pcie TLP header information in the tracepoint Thomas Tai
2018-05-08 19:57 ` [PATCH V2] " Thomas Tai
2018-05-08 21:07   ` Steven Rostedt
2018-05-08 22:24     ` Thomas Tai

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).