Mailing List Archive: Quagga: Bugs

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value

Index | Next | Previous | View Threaded

bugzilla-daemon at allevil

Jun 6, 2008, 7:30 AM

Post #1 of 6 (4339 views)
Permalink

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug
report.

http://bugzilla.quagga.net/show_bug.cgi?id=459

cgeorge [at] paradise changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Ever Confirmed| |1

------- Additional Comments From cgeorge [at] paradise 2008-06-06 15:08 -------
If time() returns a negative value then OSPFD doesn't work correctly, and no
OSPF routes are created.

It is possible for time() to return a negative value on some systems when the
hw clock has been set to a time before 1970, because time_t is signed. It is
true on our system, running linux 2.2.4, and is probably true on other systems
as well.

We spotted the problem in the function ospf_spf_calculate_schedule in the file
ospf_spf.c. If time is negative a very large delay is set for the timer which
means, in effect, that the timer never matures and routes are never generated.

Possibly the solution is to use the value of ospf->spf_delay for delay, if time
() returns a negative value.

Perhaps another solution would just be to put out a warning (maybe error)
message when time() returns a negative value. The user can then set their hw
clock appropriatly.

This may seem like a trivial problem, but as new users, it took us about four
working days to find this problem, and my manager was screaming at us to dump
quagga and try something else!

Great software by the way.

Hope this helps.

Best Regards
Cliff George - Paradise Datacom

------- Additional Comments From cgeorge [at] paradise 2008-06-06 15:30 -------
Should have started this off as a NEW bug not an CONFIRMED one, sorry!

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
Quagga-bugs mailing list
Quagga-bugs [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-bugs

bugzilla-daemon at allevil

Jun 10, 2008, 8:18 AM

Post #2 of 6 (4104 views)
Permalink

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value [In reply to]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug
report.

http://bugzilla.quagga.net/show_bug.cgi?id=459

------- Additional Comments From paul [at] dishone 2008-06-10 16:18 -------
Sorry to hear that. A simple search through old bugs for 'time' would have lead
you to bugs 120 and bugs 134. Bug 134 indicates there are fixes for this problem
in 0.99.

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
Quagga-bugs mailing list
Quagga-bugs [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-bugs

bugzilla-daemon at allevil

Jun 10, 2008, 8:19 AM

Post #3 of 6 (4095 views)
Permalink

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value [In reply to]

bugzilla-daemon at allevil

Jun 10, 2008, 8:56 AM

Post #4 of 6 (4106 views)
Permalink

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value [In reply to]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug
report.

http://bugzilla.quagga.net/show_bug.cgi?id=459

------- Additional Comments From cgeorge [at] paradise 2008-06-10 16:56 -------
You may be correct, but I am not totally convinced.

Bug 134 deals with changes in time, while OSPF is running.

Our problem was that when the system started the hw system clock had a value
before 1970 in it and so time() returned a negative value, then OSPF didn't
work, at least not in 0.98.6. No changes to the system clock were made while
OSPF was running.

It could be that the changes that were made for bug 134 also fix this bug, but
that also may not be true.

According to bug 134, fixes have only been applied to 0.99, but only 0.98.6 is
described as stable, implying to us that 0.99 is not yet stable. We don't want
to use something that may not work correctly.

Sorry, not meaning to be controversial, just thorough

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
Quagga-bugs mailing list
Quagga-bugs [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-bugs

bugzilla-daemon at allevil

Jun 11, 2008, 2:27 PM

Post #5 of 6 (4090 views)
Permalink

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value [In reply to]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug
report.

http://bugzilla.quagga.net/show_bug.cgi?id=459

paul [at] dishone changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|DUPLICATE |

------- Additional Comments From paul [at] dishone 2008-06-11 22:27 -------
Hi Clifford,

time() returning a negative is an error. I'm not sure what Quagga is supposed to
do in such a case. That is indeed different to bug #134, my apologies -
re-opening bug. It would be useful to re-test this against 0.99 and observe the
behaviour there. Perhaps we can fix things - though, if time repeatedly returns
a useless value there may not be much that a timer-driven application will be
able to do to work-around this...

0.98 is indeed marked stable, and 0.99 marked unstable. That said, we're
reasonably conversative about those designations, and 0.99 has been in a
stabilisation mode for about a year or so now, and we hope to make it the stable
series soon. 0.98 is stable in the sense that it's a known quantity - not that
it has fewer bugs... For any fresh development, I'd urge 0.99 be evaluated
first, at this stage. ;)

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
Quagga-bugs mailing list
Quagga-bugs [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-bugs

bugzilla-daemon at allevil

Jun 12, 2008, 1:31 AM

Post #6 of 6 (4101 views)
Permalink

[Bug 459] OSPF doesn't work correctly if the function time() returns a negative value [In reply to]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug
report.

http://bugzilla.quagga.net/show_bug.cgi?id=459

------- Additional Comments From cgeorge [at] paradise 2008-06-12 09:31 -------
Hi Paul,

Thanks for your comments.

When I get a bit of spare time I will try out version 0.99 and see what happens
when time() comes back as negative.

Regards
Cliff

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
Quagga-bugs mailing list
Quagga-bugs [at] lists
http://lists.quagga.net/mailman/listinfo/quagga-bugs

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads