RE: PROBLEM: [x86] Running ptpd2 using an Intel 82572EI (e1000e) leads to a kernel oops (3.12.26)

-----Original Message-----
From: Nick Krause [mailto:xerofoify@gmail.com]
Sent: Wednesday, August 06, 2014 2:37 PM
To: Koehrer Mathias (ETAS/ESW5)
Cc: linux-kernel@vger.kernel.org; netdev@vger.kernel.org
Subject: Re: PROBLEM: [x86] Running ptpd2 using an Intel 82572EI (e1000e) leads
to a kernel oops (3.12.26)

On Wed, Aug 6, 2014 at 8:17 AM, Koehrer Mathias (ETAS/ESW5)
[off-list ref] wrote:

quoted

quoted

Mathias ,
After tracing this it seems to be either one of two things based on
the warn on messages in
__queue_work. Normally it would be because of this line.
WARN_ON_ONCE(!irqs_disabled());.
However it would also be the second warn on, If you want I can send a
simple patch with
a printk statement to see which one it is.
Regards Nick

That would be helpful. I can try to see what's happening.

Regards

Mathias

Sorry for the late reply but couldn't figure out how to print the flags.
Cheers Nick

Hi Nick,
thanks for the patch.
I applied it, however there was no change at all in the output.

My impression is, that it has to do with the network driver as everything works fine if I am using a different NIC.

When I apply the patch

Index: linux-3.12.26/drivers/net/ethernet/intel/e1000e/netdev.c
===================================================================

--- linux-3.12.26.orig/drivers/net/ethernet/intel/e1000e/netdev.c       2014-08-04 10:56:56.000000000 +0200
+++ linux-3.12.26/drivers/net/ethernet/intel/e1000e/netdev.c    2014-08-06 15:15:42.000000000 +0200

@@ -5549,7 +5549,8 @@
        count = e1000_tx_map(tx_ring, skb, first, adapter->tx_fifo_limit,
                             nr_frags);
        if (count) {
-               if (unlikely((skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP) &&
+               if ((adapter->flags & FLAG_HAS_HW_TIMESTAMP) &&
+                        unlikely((skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP) &&
                             !adapter->tx_hwtstamp_skb)) {
                        skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
                        tx_flags |= E1000_TX_FLAGS_HWTSTAMP;

to the e1000e driver it seems to work...
This makes sense as the work queue is only created if  "adapter->flags & FLAG_HAS_HW_TIMESTAMP"  is set.
With this NIC this is *not* the case.


Regards

Mathias

--