Further looking into this issue I got the following info out of dtrace:
[i]crash.py is my scriptname to execute the python lines mentioned by K[/i]

  2  64709         e1000g_rxfree_func:entry   crash.py
  2  64710        e1000g_rxfree_func:return   crash.py
  2  64563     e1000g_intr_pciexpress:entry   sched
  2  64837    e1000g_check_acc_handle:entry   sched
  2  64838   e1000g_check_acc_handle:return   sched
  2  65035             e1000g_receive:entry   sched
  2  64835    e1000g_check_dma_handle:entry   sched
  2  64836   e1000g_check_dma_handle:return   sched
  2  64835    e1000g_check_dma_handle:entry   sched
  2  64836   e1000g_check_dma_handle:return   sched
  2  64837    e1000g_check_acc_handle:entry   sched
  2  64838   e1000g_check_acc_handle:return   sched
  2  65036            e1000g_receive:return   sched
  2  19233                 ip_input:receive
[b]  5  64613         e1000g_local_timer:entry   sched
  5  64615         e1000g_stall_check:entry   sched
  5  64616        e1000g_stall_check:return   sched
  5  64567          e1000g_link_check:entry   sched
  5  64541             e1000g_link_up:entry   sched
  5  65047       e1000_check_for_link:entry   sched
  5  64887 e1000_check_for_copper_link_generic:entry   sched
  5  64888 e1000_check_for_copper_link_generic:return   sched
  5  65048      e1000_check_for_link:return   sched
  5  64542            e1000g_link_up:return   sched
  5  64837    e1000g_check_acc_handle:entry   sched
  5  64838   e1000g_check_acc_handle:return   sched
  5  64568         e1000g_link_check:return   sched
  5  64799 e1000_igp_ttl_workaround_82547:entry   sched
  5  64800 e1000_igp_ttl_workaround_82547:return   sched
  5  64837    e1000g_check_acc_handle:entry   sched
  5  64838   e1000g_check_acc_handle:return   sched
  5  64617     restart_watchdog_timer:entry   sched
  5  64619         arm_watchdog_timer:entry   sched
  5  64620        arm_watchdog_timer:return   sched
  5  64618    restart_watchdog_timer:return   sched
  5  64614        e1000g_local_timer:return   sched
[/b]

So at some point the e1000g goes crazy and is not able to reset the device.
Searching in the bug database I found 
http://bugs.opensolaris.org/view_bug.do?bug_id=6597545
It is pretty similar to what we are experiencing except that our device is the 
pci8086,1096
As this bug is reported to be solved in b95 we tried with the e1000g driver in 
that release but we still experience the same issue.

S
 
 
This message posted from opensolaris.org
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to