freebsd stable - Feb 2008 - if_re losinbg ability to communicate, even with latest patches

yesterday I had my first experience of an if_re interface simply stopping
responding. I ma running 7.0-RELEASE on the machine, but with
the lates files taken from here:

	http://people.freebsd.org/~yongari/re/7.0R/if_re.c
	http://people.freebsd.org/~yongari/re/7.0R/if_rlreg.h

Luckily I had worried about tis and have set the machine up to reboot
if it suspects that it';s netowkring is down, so twelve hours later it
did so and I can login again. Is there anythign useful I can get out
of the logs to try and work out what the problem is though ? A I see
lots of messages in the log like this:

"error sending response: not enough free resources"

but I am assuming that this is due to the ether stall, not the cause of
it. Would enabling polling on the interface help ? (i.e. is this possiblky
a lost interrupt issue?).

All advice welcome...

-pete.

On Thu, Feb 28, 2008 at 02:34:01PM +0000, Pete French wrote:
 > yesterday I had my first experience of an if_re interface simply stopping
 > responding. I ma running 7.0-RELEASE on the machine, but with
 > the lates files taken from here:
 > 
 > 	http://people.freebsd.org/~yongari/re/7.0R/if_re.c
 > 	http://people.freebsd.org/~yongari/re/7.0R/if_rlreg.h
 > 
 > Luckily I had worried about tis and have set the machine up to reboot
 > if it suspects that it';s netowkring is down, so twelve hours later it
 > did so and I can login again. Is there anythign useful I can get out
 > of the logs to try and work out what the problem is though ? A I see
 > lots of messages in the log like this:
 > 
 > "error sending response: not enough free resources"
 > 
 > but I am assuming that this is due to the ether stall, not the cause of
 > it. Would enabling polling on the interface help ? (i.e. is this possiblky
 > a lost interrupt issue?).
 > 

Any other console messages printed from re(4)?

Show me both verbosed boot messages related with re(4) and
'vmstat -i' output.

 > All advice welcome...
 > 
 > -pete.

-- 
Regards,
Pyun YongHyeon

On Thu, Feb 28, 2008, Pete French wrote:
> yesterday I had my first experience of an if_re interface simply stopping
> responding. I ma running 7.0-RELEASE on the machine, but with
> the lates files taken from here:
> 
> 	http://people.freebsd.org/~yongari/re/7.0R/if_re.c
> 	http://people.freebsd.org/~yongari/re/7.0R/if_rlreg.h
Just wondering whether you disabled any hardware offloadings. I'm
running a RELENG_7/amd64 as of 2008-Feb-22 with the re(4) driver from
the URLs above. As a precaution, I disabled all kinds of hardware
offloadings.

| # grep ^ifconfig_re.= /etc/rc.conf 
| ifconfig_re0="inet x.x.x.x netmask 255.255.255.224 media auto -rxcsum
-txcsum -tso -lro -vlanhwtag"

| # ifconfig re0
| re0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu
1500
|         options=88<VLAN_MTU,VLAN_HWCSUM>
|         ether 00:1d:92:x:x:x
|         inet6 fe80::21d:92ff:fexx:xxxx%re0 prefixlen 64 scopeid 0x1 
|         inet x.x.x.x     netmask 0xffffffe0 broadcast 78.46.41.63
|         inet x.x.x.x     netmask 0xfffffff8 broadcast 78.47.124.7
|         inet x.x.x.x     netmask 0xfffffff8 broadcast 78.47.124.7
|         inet x.x.x.x     netmask 0xfffffff8 broadcast 78.47.124.7
|         inet x.x.x.x     netmask 0xfffffff8 broadcast 78.47.124.7
|         inet x.x.x.x     netmask 0xfffffff8 broadcast 78.47.124.7
|         inet x.x.x.x     netmask 0xfffffff8 broadcast 78.47.124.7
|         media: Ethernet autoselect (100baseTX <full-duplex>)
|         status: active

| # pciconf -lv
[...]
| re0@pci0:2:0:0: class=0x020000 card=0x368c1462 chip=0x816810ec rev=0x01
hdr=0x00
|     vendor     = 'Realtek Semiconductor'
|     device     = 'RTL8168/8111 PCI-E Gigabit Ethernet NIC'
|     class      = network
|     subclass   = ethernet
[...]

After almost one week uptime, the machine is still stable. Before it
started lossing packets after very few hours uptime. I'm wondering if
this is related to the options I disabled, so I'm interested in your
settings.

-cs