Feedback from the initial pass at RFC for making a common pool object was that it got really confusing trying to manage the multitude of changes. See patch 3 of the series: http://www.redhat.com/archives/libvir-list/2017-February/msg00519.html So while I figure out what it'll take to make a better object framework, I figure I could start splitting out things a bit to make future work a bit easier to understand. Also rather than trying to do them all in one massive series, I can do shorter series contained to each of the driver conf's. Besides let's figure out whether all the thoughts and ideas are good. I would think moving to virnodedeviceobj wouldn't be an issue (other than needing to deal with backports). However, I'm sure there could be some "alternate opinions" over whether to make static functions not start with vir[DriverSubsystem] and rather go with driverSubsystem. Changing the API's to use the Obj is mainly a consistency thing. Let's see what falls out of this before I start queuing up other ones as it's a lot of rote work! John Ferlan (3): conf: Introduce virnodedeviceobj conf: Clean up nodedev code conf: Use consistent function name prefixes for virnodedeviceobj po/POTFILES.in | 1 + src/Makefile.am | 3 +- src/conf/node_device_conf.c | 981 +++++++++-------------------------- src/conf/node_device_conf.h | 50 -- src/conf/virnodedeviceobj.c | 558 ++++++++++++++++++++ src/conf/virnodedeviceobj.h | 78 +++ src/libvirt_private.syms | 23 +- src/node_device/node_device_driver.c | 24 +- src/node_device/node_device_driver.h | 2 +- src/node_device/node_device_hal.c | 10 +- src/node_device/node_device_udev.c | 12 +- src/test/test_driver.c | 29 +- 12 files changed, 940 insertions(+), 831 deletions(-) create mode 100644 src/conf/virnodedeviceobj.c create mode 100644 src/conf/virnodedeviceobj.h -- 2.9.3 -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list
