On Thu, Mar 23, 2006 at 09:03:01AM +0100, Frank Behrens wrote: > Pyun YongHyeon <pyunyh_at_gmail.com> wrote on 20 Mar 2006 16:11: > > If you still see "watchdog timeout message" please let me know. > > Mar 23 03:17:15 <kern.crit> moon kernel: sk0: watchdog timeout > Mar 23 03:17:15 <kern.notice> moon kernel: sk0: link state changed to DOWN > Mar 23 03:24:00 <kern.notice> moon kernel: sk0: link state changed to UP > Mar 23 03:24:00 <kern.notice> moon kernel: sk0: link state changed to DOWN > Mar 23 03:24:00 <kern.crit> moon kernel: sk0: phy failed to come ready > Mar 23 03:24:00 <kern.notice> moon kernel: sk0: link state changed to UP > > The recovery at 03:24 was done by my "safety belt" in crontab: > (ifconfig sk0 | fgrep active >/dev/null) || (ifconfig sk0 down; ifconfig sk0 up; ifconfig sk0) > The driver used is from Mar 20. > Oh my... Ok, how about this one?(Replace if_sk.c with the file http://people.freebsd.org/~yongari/sk/if_sk.c) This one issues SK_TXBMU_TX_START command for each packet. Previously it used to issue single SK_TXBMU_TX_START command for all queued packets. Data sheet said nothing for issuing SK_TXBMU_TX_START command(i.e. one issue for each packet vs. one issue for all queued packets) and sending SK_TXBMU_TX_START for every packet would add an additional penalty as it needs a PCI write operation for each packet. Please let me know how it goes on your system. Thanks for reporting. -- Regards, Pyun YongHyeonReceived on Thu Mar 23 2006 - 10:04:56 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:54 UTC