Bug 502844
| Summary: | CONFIG_ARPD needed by opennhrp | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Charles Lopes <tjarls> |
| Component: | kernel | Assignee: | John W. Linville <linville> |
| Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | medium | Docs Contact: | |
| Priority: | low | ||
| Version: | 11 | CC: | itamar, kernel-maint, maurizio.antillon, nhorman, quintela, timo.teras |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | 2.6.29.5-191.fc11 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2009-06-24 19:23:44 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
Neil, do you know any specific reason this is disabled in Fedora? from the kconfig.. "This code is experimental and also obsolete." To be fair, the code added by CONFIG_ARPD is quite small (around 40 lines). It is basically two functions that only get called when the parameter app_probes is not zero (.i.e. a user space daemon is present). Installations without the appropriate daemon will only get aa additional couple of extra "if" tests when an arp resolution is made. The experimental qualification here should not make us fear the introduction of buggy code into the kernel. As for the obsolete part, it is likely because the daemon arpd has hardly been needed ever since the kernel ARP implementation was changed to be tuneable by /proc. Still the interface has not been replaced by something else and it can be useful beyond its original application. NHRP and opennhrp provide a new use for it, and, in this particular application, there does not seem to be an alternate mechanism to get the necessary notifications from the network neighbour resolution. This bug appears to have been reported against 'rawhide' during the Fedora 11 development cycle. Changing version to '11'. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping For my $0.02...I have used ARPD in a former life and found it to be safe for that application. Review of the current code makes it look reasonably safe. So FWIW, I have no objection to enabling CONFIG_ARPD... The current help text for CONFIG_ARPD is bogus: http://marc.info/?l=linux-netdev&m=124477918912638&w=2 Patch to fix the text was sent and is currently pending review: http://marc.info/?l=linux-netdev&m=124478510017334&w=2 Additionally, there was one bug with the neighbor cache that was only triggerable with the Netlink ARP interface. So if you enable this option, you might also want to cherry-pick the following patch: http://git.kernel.org/?p=linux/kernel/git/davem/net-next-2.6.git;a=commitdiff;h=5ef12d98a19254ee5dc851bd83e214b43ec1f725 Enabled CONFIG_ARPD and added the Netlink ARP patch from comment 6 in rawhide kernels. Should be available in the next kernel _after_ kernel-2.6.30-1.fc12. Does this need to be for F-11? Or is F-12 good enough? Enabled in 2.6.29.5-179.rc1.fc11 and 2.6.30-6.fc12 Would it be inconvenient to also activate that on F-10, please? I am already using opennhrp on 7 systems running F-10. Otherwise, F-11 is already good enough, thanks. Next I intend to submit a package for opennhrp, unless Timo wants to do it himself. http://koji.fedoraproject.org/koji/buildinfo?buildID=106268 Hth...that is a 2.6.29 kernel -- I have no idea if/when F-10 will ever get an official update to 2.6.29... (In reply to comment #10) > Next I intend to submit a package for opennhrp, unless Timo wants to do it > himself. My experiences from RPM packaging is from five years ago. So please go ahead. If you want it on future opennhrp tarballs / git tree too, send it on the opennhrp list and I'm happy to commit it. (In reply to comment #11) > http://koji.fedoraproject.org/koji/buildinfo?buildID=106268 > > Hth...that is a 2.6.29 kernel -- I have no idea if/when F-10 will ever get an > official update to 2.6.29... That is good enough. Thanks. (In reply to comment #11) > http://koji.fedoraproject.org/koji/buildinfo?buildID=106268 > > Hth...that is a 2.6.29 kernel -- I have no idea if/when F-10 will ever get an > official update to 2.6.29... CONFIG_ARPD was not enabled. Fixed in 2.6.29.5-83 Doh! Thanks Chuck! kernel-2.6.29.5-191.fc11 has been submitted as an update for Fedora 11. http://admin.fedoraproject.org/updates/kernel-2.6.29.5-191.fc11 kernel-2.6.29.5-191.fc11 has been pushed to the Fedora 11 testing repository. If problems still persist, please make note of it in this bug report. If you want to test the update, you can install it with su -c 'yum --enablerepo=updates-testing update kernel'. You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F11/FEDORA-2009-6768 kernel-2.6.29.5-191.fc11 has been pushed to the Fedora 11 stable repository. If problems still persist, please make note of it in this bug report. |
The usage of opennhrp requires CONFIG_ARPD to be set. I looked through the kernel source code and that it looks like that flag does not change the kernel behaviour when no program uses the ARPD interface. While the help text for CONFIG_ARPD states: "If you say Y here, the kernel's internal ARP cache will never grow to more than 256 entries (the oldest entries are expired in a LIFO manner) and communication will be attempted with the user space ARP daemon arpd. Arpd then answers the address resolution request either from its own cache or by asking the net." the part about limiting the cache to 256 does not seem to be anymore valid. I do not expect that activating this flag has any detrimental side-effect.