summaryrefslogtreecommitdiff
path: root/xpp/waitfor_xpds
AgeCommit message (Collapse)Author
2009-11-04Add optional "hotplug mode" to init scriptsTzafrir Cohen
The hotplug mode is only used in XPP_HOTPLUG_DAHDI=yes is set in /etc/dahdi/init.conf . If it is set and the system actually has Astribanks, most of the init script will actually be run from the context of a udev event. Merged revisions 6987,7014,7334-7335,7409,7476 via svnmerge from https://origsvn.digium.com/svn/dahdi/tools/trunk ........ r6987 | tzafrir | 2009-08-13 18:15:49 +0300 (Thu, 13 Aug 2009) | 9 lines xpp: Add astribank_is_starting: astribank_is_running is used to tell when we may have an Astribank that is initializing (and may be re-enumerating and thus not listed as a device). It uses a semaphore as we can always write to one and we can't always write to a file. ........ r7014 | tzafrir | 2009-08-16 19:05:11 +0300 (Sun, 16 Aug 2009) | 4 lines xpp: A man page for astribank_is_starting xpp rev: 7321 ........ r7334 | tzafrir | 2009-10-04 15:35:51 +0200 (Sun, 04 Oct 2009) | 7 lines dahdi.init: move module loading to a separate function Move the module loading parts of the dahdi init script to a seperate function - load_modules. To be used by a later commit. There is not functional change here (but the diff is confusing). ........ r7335 | tzafrir | 2009-10-04 15:53:01 +0200 (Sun, 04 Oct 2009) | 16 lines Add XPP_HOTPLUG_DAHDI mode: cfg initiated mostly from hotplug If XPP_HOTPLUG_DAHDI=yes is set in /etc/dahdi/init.conf (which is not the default), the normal run of the 'start' operation of the dahdi init script will only be responsible for loading modules. The rest will be done (if there are indeed Astribanks on the system) by a second call to the init script once all of them have connected. The astribank_hook has also been mostly rewritten. Most of the functionality of twinstar_hook moved to it. The current twinstar_hook is a simple example script. XPP_HOTPLUG_DAHDI mode assumes that all Astribanks on the system are listed in /etc/dahdi/xpp_order (which can be generated using, e.g. 'dahdi_genconf xpporder') ........ r7409 | tzafrir | 2009-10-21 16:30:32 +0200 (Wed, 21 Oct 2009) | 21 lines Fix XPP_HOTPLUG_DAHDI: logic; end of init.d script 'Hotplug mode' was introduced in r7335. * The logic in the script was broken. - Negative logic is not such a grand idea to start with. * Interactive invocation of init.d ends when expected and not sooner. This change makes waitfor_xpds wait longer. Rather than waiting for all the Astribanks to load, it will now wait until the initialization of dahdi from the Astribanks hook script is run. This allows running e.g.: /etc/init.d/dahdi start; /etc/init.d/asterisk start It also means that 'astribank_is_starting' is actually used as a semaphore and not only as stamp file. As before, those changes have no effect if hotplug mode is not explicitly enabled (setting 'XPP_HOTPLUG_DAHDI=yes' in init.conf). ........ r7476 | tzafrir | 2009-11-04 22:12:35 +0200 (Wed, 04 Nov 2009) | 6 lines xpp: make astribank_is_starting build without warnings. autoconfig.h is needed as semtimedop() requires _GNU_SOURCE. xpp rev: 7471 ........ git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/branches/2.2@7477 a0bf4364-ded3-4de4-8d8a-66a801d63aff
2009-11-04Fix bashism in waitfor_xpdsTzafrir Cohen
Merged revisions 6829,7021 via svnmerge from http://svn.digium.com/svn/dahdi/tools/trunk git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/branches/2.2@7468 a0bf4364-ded3-4de4-8d8a-66a801d63aff
2009-05-28xpp: if $dahdi_hardware is empty, don't use it in an error messageTzafrir Cohen
git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/trunk@6668 a0bf4364-ded3-4de4-8d8a-66a801d63aff
2009-05-25Properly fix usage of waitfor_xpds in non-standard settings.Tzafrir Cohen
From xpp rev. 7137. git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/trunk@6655 a0bf4364-ded3-4de4-8d8a-66a801d63aff
2009-05-19dahdi.init: moved test for astribanks into waitfor_xpdsTzafrir Cohen
If the system starts with an Astribank with no firmware, the firmware loading may take a while and the Astribank only becomes ready some time after the DAHDI init script starts. This means that the test at the beginning of xpp_startup() in dahdi.init fails to find an Astribank, and the init script simply continues ignoring this. This changeset moves the testing inside the script waitfor_xpds. The script now: 1. Tests for an Astribank using dahdi_hardware and exists immediately if none found. 2. A loop (of maximum 10 seconds) waits for the first Astribank to be detected by the driver. 3. Waiting for the initialization of all the Astribanks to complete (The original waitfor_xpds) The support for the old procfs iinterface has been removed, as it has never been used with DAHDI. git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/trunk@6646 a0bf4364-ded3-4de4-8d8a-66a801d63aff
2008-11-24Add some missing copyright statements in the xpp/ directory as well.Tzafrir Cohen
git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/trunk@5379 a0bf4364-ded3-4de4-8d8a-66a801d63aff
2008-10-16xpp: userspace support for sysfx migration.Tzafrir Cohen
The userspace side of dahdi-linux r5097. * Perl modules default to using xpp sysfs but will fallback (with warning) to procfs interface. * An additional /usr/share/dahdi/waitfor_xpds to replace the line in the init.d script. git-svn-id: http://svn.asterisk.org/svn/dahdi/tools/trunk@5099 a0bf4364-ded3-4de4-8d8a-66a801d63aff