e1000e: 82571 Tx Data Corruption during Tx hang recovery
authorTushar Dave <tushar.n.dave@intel.com>
Wed, 1 Aug 2012 02:11:15 +0000 (02:11 +0000)
committerPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Tue, 7 Aug 2012 07:41:36 +0000 (00:41 -0700)
commiteca90f550494171f54f8a700caee65ec16455a5b
tree76911fcfe9b98697ef72b6dc267ddc9eac4fb795
parentb7ec70be01a87f2c85df3ae11046e74f9b67e323
e1000e: 82571 Tx Data Corruption during Tx hang recovery

A bus trace shows that while executing e1000e_down, TCTL is cleared except
for the PSP bit.  This occurs while in the middle of fetching a TSO packet
since the Tx packet buffer is full at that point. Before the device is
reset, the e1000_watchdog_task starts to run from the middle (it was
apparently pre-empted earlier, although that is not in the trace) and sets
TCTL.EN.  At that point, 82571 transmits the corrupted packet, apparently
because TCTL.MULR was cleared in the middle of fetching a packet, which is
forbidden.

Driver should just clear TCTL.EN in e1000_reset_hw_82571 instead of
clearing the entire register, so as not to change any settings in the
middle of fetching a packet.

Signed-off-by: Tushar Dave <tushar.n.dave@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
drivers/net/ethernet/intel/e1000e/82571.c