Vyatta system-level configuration templates/scripts (mirror of https://github.com/vyos/vyatta-cfg-system.git) https://git.amelek.net/vyos/vyatta-cfg-system.git/atom?h=master 2013-03-28T23:47:01+00:00 2013-03-28T23:47:01+00:00 Stephen Hemminger stephen@networkplumber.org 2013-03-28T23:47:01+00:00 urn:sha1:dcabba83be38ee6f79ab5a347f0cb3cce5a9eca8 Bug 8804 Previous change was passing a logfile to biosdevname, but parent directory was not necessarily created at that point! Log file is useless anyway, just redirect to dev null. 2013-02-06T23:40:49+00:00 James Davidson james.davidson@vyatta.com 2013-02-06T22:48:21+00:00 urn:sha1:80b635570ceb26bcbc6dda9e769961978bfe4b7e Bug 8676 biosdevname does not behave well on Xen so simply skip it when running in Xen. (cherry picked from commit dc53d09decae1cd8098ee92c7c1dab72dc80ae50) 2011-03-03T23:27:29+00:00 Stephen Hemminger stephen.hemminger@vyatta.com 2011-03-03T23:27:29+00:00 urn:sha1:ba5aeb6086e8c6aef86c18e2646b7c80e05f8e3d Bug 6818 The regex to extract wireless name was incorrect. Recode for clarity rather than adding more complex regex. 2011-02-01T18:52:23+00:00 Bob Gilligan gilligan@vyatta.com 2011-02-01T18:27:27+00:00 urn:sha1:a85d727c301007fb6fff667e6899bcbc00d7b2be It appears that biosdevname can be confused if interface renaming is being performed by other udev worker processes at the same time that it is running. Since vyatta_net_name does protect the body of its code with a lock, we can be sure that no other renaming will be started in other udev worker processes while this one is running in vyatta_net_name. But renaming being performed by other udev worker processes that were triggered by previous executions of vyatta_net_name may still be in progress. This change gives any interface renaming that was called for by previous invocations of vyatta_net_name a chance to complete before we call biosdevname. (cherry picked from commit 9dd41e8481e84bf80fcb0e9c3d7843c8a0b20059) 2011-01-24T23:08:04+00:00 Stephen Hemminger stephen.hemminger@vyatta.com 2011-01-24T21:44:51+00:00 urn:sha1:a1087973b84c87fde430de4f66b188c0473fd54f Bug 6718 On livecd config.boot does not exist yet, so no existing hw-id mappings. (cherry picked from commit 37bbcaef6083a843cb3dd9976854a30ab7cebe16) 2011-01-24T19:54:44+00:00 Stephen Hemminger stephen.hemminger@vyatta.com 2011-01-24T12:51:10+00:00 urn:sha1:39f9647acbc45c1f7868d2f40342b276679ea3b8 Put more messages in to help debug current problem and future ones (cherry picked from commit 8cc21571cb4efde31f3cf6b58ad8e3d80d4487cf) 2011-01-24T19:54:44+00:00 Stephen Hemminger stephen.hemminger@vyatta.com 2011-01-24T10:59:13+00:00 urn:sha1:4180f0bfb54ac3628e3f6f465bdc5284bd6d4d34 Log to file /dev/udev/net-name.log since that is available tmpfs (cherry picked from commit 15ce02d66348ed35308f5813991e171d74561248) 2011-01-24T19:54:44+00:00 Stephen Hemminger stephen.hemminger@vyatta.com 2011-01-24T07:03:46+00:00 urn:sha1:16db4272b78fb6d50fc92f6a24fbf12e2762c64f This reverts commit b1af934db367d589cf7c78b84e2181a1e72b2aaf. Loggging to /var/log does not work during boot. The root filesystem is read-only when udev is running this script. (cherry picked from commit 47301dbc029652f97847c9288d58c951db56d916) 2011-01-21T06:23:49+00:00 Bob Gilligan gilligan@vyatta.com 2011-01-21T06:23:49+00:00 urn:sha1:f90053272e601b3b87049a7f0a5d6c9f49e98940 This restores logging that we had in the previous release, which has proven useful in troubleshooting interface renaming problems. 2011-01-17T07:06:23+00:00 Bob Gilligan gilligan@vyatta.com 2011-01-17T06:58:22+00:00 urn:sha1:8619e721c4b62085a8fa83c4284b0d87d477d453 In the new version of biosdevname (version 0.3.4), the default policy generates interface names of the form "em1". That's not the naming convention that we use. The "all_ethN" policy does use the traditional ethernet naming convention. (cherry picked from commit f05a91c02e27bb19717ddf8e1ffc07ed7b1bbbc5)