[PATCH] powerpc/powernv/elog: Reduce elog message severity

OPAL interrupts kernel whenever it has new error log. Kernel calls
interrupt handler (elog_event()) to retrieve event. elog_event makes
OPAL API call (opal_get_elog_size()) to retrieve elog info.

In some case before kernel makes opal_get_elog_size() call, it gets interrupt
again. So second time when elog_event() calls opal_get_elog_size API OPAL
returns error. Its safe to ignore this error. Hence reduce the severity
of log message.

CC: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Signed-off-by: Vasant Hegde <redacted>
---
 arch/powerpc/platforms/powernv/opal-elog.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/powerpc/platforms/powernv/opal-elog.c b/arch/powerpc/platforms/powernv/opal-elog.c
index 62ef7ad995da..67f435bb1ec4 100644
--- a/arch/powerpc/platforms/powernv/opal-elog.c
+++ b/arch/powerpc/platforms/powernv/opal-elog.c

@@ -247,7 +247,7 @@ static irqreturn_t elog_event(int irq, void *data)
 
 	rc = opal_get_elog_size(&id, &size, &type);
 	if (rc != OPAL_SUCCESS) {
-		pr_err("ELOG: OPAL log info read failed\n");
+		pr_debug("ELOG: OPAL log info read failed\n");
 		return IRQ_HANDLED;
 	}
 

-- 
2.26.2

Vasant Hegde [off-list ref] writes:

OPAL interrupts kernel whenever it has new error log. Kernel calls
interrupt handler (elog_event()) to retrieve event. elog_event makes
OPAL API call (opal_get_elog_size()) to retrieve elog info.

In some case before kernel makes opal_get_elog_size() call, it gets interrupt
again. So second time when elog_event() calls opal_get_elog_size API OPAL
returns error.

Can you give more detail there? Do you have a stack trace?

We use IRQF_ONESHOT for elog_event(), which (I thought) meant it
shouldn't be called again until it has completed.

So I'm unclear how you're seeing the behaviour you describe.

cheers

On 10/9/20 10:34 AM, Michael Ellerman wrote:

Vasant Hegde [off-list ref] writes:

quoted

OPAL interrupts kernel whenever it has new error log. Kernel calls
interrupt handler (elog_event()) to retrieve event. elog_event makes
OPAL API call (opal_get_elog_size()) to retrieve elog info.

In some case before kernel makes opal_get_elog_size() call, it gets interrupt
again. So second time when elog_event() calls opal_get_elog_size API OPAL
returns error.

Can you give more detail there? Do you have a stack trace?

I observe below log in dmesg
  [  615.708709] ELOG: OPAL log info read failed

That means elog_event called twice for same error log event.
I don't have stack trace.

We use IRQF_ONESHOT for elog_event(), which (I thought) meant it
shouldn't be called again until it has completed.

Yes. I thought so.. But once in a while we get above message (mostly when we have multiple event bits are set).

-Vasant

So I'm unclear how you're seeing the behaviour you describe.

cheers

quoted

Its safe to ignore this error. Hence reduce the severity
of log message.

CC: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Signed-off-by: Vasant Hegde <redacted>
---
  arch/powerpc/platforms/powernv/opal-elog.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/powerpc/platforms/powernv/opal-elog.c b/arch/powerpc/platforms/powernv/opal-elog.c
index 62ef7ad995da..67f435bb1ec4 100644
--- a/arch/powerpc/platforms/powernv/opal-elog.c
+++ b/arch/powerpc/platforms/powernv/opal-elog.c

@@ -247,7 +247,7 @@ static irqreturn_t elog_event(int irq, void *data)
  
  	rc = opal_get_elog_size(&id, &size, &type);
  	if (rc != OPAL_SUCCESS) {
-		pr_err("ELOG: OPAL log info read failed\n");
+		pr_debug("ELOG: OPAL log info read failed\n");
  		return IRQ_HANDLED;
  	}
  

-- 
2.26.2