Skip to contentSkip to conversation

2.6.14-rcX: strange timestamp on ping

DISCUSSION Joachim Bremer 2 participants 2 messages Sep 29, 2005

lore
Share & tools
lore
#0 lore raw

Hello,

since very early in the 2.6.14 process ping or traceroute gives very strange timestamps. eg
64 bytes from 192.168.0.1: icmp_seq=1 ttl=127 time=4294971590970 ms

Translating the time to hex gives something like 0x03e8 xxxx xxxx which make me wonder a little. This was introduced by the patch:

[NET]: Store skb->timestamp as offset to a base timestamp

Reduces skb size by 8 bytes on 64-bit.

---
commit a61bbcf28a8cb0ba56f8193d512f7222e711a294
tree 33ae1976ab3b08aac516debb2742d2c6696d5436
parent 25ed891019b84498c83903ecf53df7ce35e9cff6

I checked this on two machines. Both where equipped with Opteron processors in 64bit mode. Using gcc 3.4.3, 4.0.1 and 4.0.2 gives exactly the same reults. Reverting that patch restores the correct timestamping

Bye

Joachim

______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS! Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193

#1 lore raw
Quoted 7 lines · Joachim Bremer

Hello,

since very early in the 2.6.14 process ping or traceroute gives very strange timestamps. eg
64 bytes from 192.168.0.1: icmp_seq=1 ttl=127 time=4294971590970 ms

Similar on ia64, but I don't see the problem across the network (either to or from my 2.6.14-rc2 system). But I do see an odd time for localhost!

$ ping localhost
PING linux-t10 (127.0.0.1) 56(84) bytes of data.
64 bytes from linux-t10 (127.0.0.1): icmp_seq=0 ttl=64 time=4294967 ms
64 bytes from linux-t10 (127.0.0.1): icmp_seq=1 ttl=64 time=4294967 ms

-Tony