Opened 11 years ago
Closed 11 years ago
#9953 closed bug (fixed)
ping gives crazy std deviation values
Reported by: | Pete | Owned by: | phoudoin |
---|---|---|---|
Priority: | normal | Milestone: | R1 |
Component: | Network & Internet | Version: | R1/alpha4.1 |
Keywords: | Cc: | ||
Blocked By: | Blocking: | ||
Platform: | All |
Description
I'm a bit surprised no one has called this out before, but when checking out my network performance with ping, I noticed that the std-deviation values it was printing were way out of whack -- larger than the actual round-trip times!
Took me a while to blame the Newton's square root calculation, but one of the comparisons there had the wrong sense!
A fix is attached.
Attachments (1)
Change History (6)
by , 11 years ago
Attachment: | 0001-Fix-to-std-dev-calculation-in-ping.patch added |
---|
comment:1 by , 11 years ago
patch: | 0 → 1 |
---|
follow-up: 3 comment:2 by , 11 years ago
Dunno where our ping.c comes from but the comparison was already corrected in 1999. I guess this validates your patch, which can thus be applied.
comment:3 by , 11 years ago
Replying to korli:
Dunno where our ping.c comes from but the comparison was already corrected in 1999.
Ahh -- I guess that's the derivation route anyway. I did various web searches, but couldn't find a ping that came close to matching. Many (earlier ones?) don't even calculate std deviation, and others used doubles and the library sqrt.
I guess this validates your patch, which can thus be applied.
Thanks.
comment:4 by , 11 years ago
Owner: | changed from | to
---|---|
Status: | new → assigned |
comment:5 by , 11 years ago
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
Patch applied in hrev46042. Thanks.
patch to ping.c to fix qsqrt