All part of the effort I have to have a common object model. This series
is node device test, driver, and virnodedevobj related.

There's also a couple of bug fixes at the beginning of the series from
things I have found during this effort.

There's still a few more patches in local branches to make the
virNodeDeviceObjListPtr private as well, but those have some merge
needs with other patches currently on list elsewhere, so I'll hold
onto them for now.

John Ferlan (13):
  test: Adjust cleanup/error paths for nodedev test APIs
  nodedev: Fix locking in virNodeDeviceObjRemove
  nodedev: Need to check for vport capable scsi_host for vHBA searches
  nodedev: Move node_device_linux_sysfs from node_driver to conf
  nodedev: Use common naming for virnodedeviceobj
  nodedev: Cleanup driver code and prototypes
  nodedev: Alter param to nodeDeviceUpdateDriverName
  nodedev: Alter param to nodeDeviceUpdateCaps
  nodedev: Create helper for finding by name in driver
  nodedev: Use consistent names for driver variables
  nodedev: Introduce virNodeDeviceObjGetDef
  nodedev: Remove privateData from virNodeDeviceObj
  nodedev: Introduce virNodeDeviceObjNew

 src/Makefile.am                                    |   8 +-
 src/conf/node_device_conf.h                        |   2 -
 .../node_device_linux_sysfs.c                      |  24 +-
 .../node_device_linux_sysfs.h                      |   9 +-
 src/conf/virnodedeviceobj.c                        | 170 ++++++-----
 src/conf/virnodedeviceobj.h                        |   2 +
 src/libvirt_private.syms                           |   6 +
 src/node_device/node_device_driver.c               | 235 +++++++--------
 src/node_device/node_device_driver.h               |  93 ++++--
 src/node_device/node_device_hal.c                  |  60 ++--
 src/node_device/node_device_udev.c                 | 323 ++++++++++++---------
 src/test/test_driver.c                             | 113 ++++---
 12 files changed, 576 insertions(+), 469 deletions(-)
 rename src/{node_device => conf}/node_device_linux_sysfs.c (89%)
 rename src/{node_device => conf}/node_device_linux_sysfs.h (82%)

-- 
2.9.3

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

John Ferlan <jferlan@redhat.com> [2017-05-19, 09:08AM -0400]:
>All part of the effort I have to have a common object model. This series
>is node device test, driver, and virnodedevobj related.
>
>There's also a couple of bug fixes at the beginning of the series from
>things I have found during this effort.
>
>There's still a few more patches in local branches to make the
>virNodeDeviceObjListPtr private as well, but those have some merge
>needs with other patches currently on list elsewhere, so I'll hold
>onto them for now.
>

Am I missing something or where does the privatization actually happen?

In general, a lot of the cleanups you are performing are beneficial for
the readability. My biggest concern is that this will generate a lot of
work when backporting patches.
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On 05/23/2017 04:26 AM, Bjoern Walk wrote:
> John Ferlan <jferlan@redhat.com> [2017-05-19, 09:08AM -0400]:
>> All part of the effort I have to have a common object model. This series
>> is node device test, driver, and virnodedevobj related.
>>
>> There's also a couple of bug fixes at the beginning of the series from
>> things I have found during this effort.
>>
>> There's still a few more patches in local branches to make the
>> virNodeDeviceObjListPtr private as well, but those have some merge
>> needs with other patches currently on list elsewhere, so I'll hold
>> onto them for now.
>>
> 
> Am I missing something or where does the privatization actually happen?

Yeah about 5 or so more patches - bad subject for the cover letter then.
The "issue" is that the "next" patch in the series deals with
virNodeDeviceObjList and that required a test_driver change. Because the
patches in my local branch build upon others and changes in other
patches have conflicts with a test_driver.c change to _testDriver...

> 
> In general, a lot of the cleanups you are performing are beneficial for
> the readability. My biggest concern is that this will generate a lot of
> work when backporting patches.

Think I already addressed this in patch 4 response.

John

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list