v2: https://www.redhat.com/archives/libvir-list/2017-May/msg00999.html

Some patches from v2 were pushed (3, 4, 6, 7, 8, 9, 11, & 12), but a
few remained from that series and are the first 5 patches of this series.

What changed?
   -> Reworked the virNodeDeviceObjRemove patch (former patch 2, but new
     series patch 1). That affected the Test patch (former patch 1, but
     now patch 2). This patch removes the address of obj logic and moves
     the onus of the ObjFree to the caller (see patch for reason).
   -> Patch 3 is the former patch 5, with no essential change
   -> Patch 4 is the former patch 10, with no essential change
   -> Former patch 13 and 14, were altered to remove the offending
      address of pointer logic. The result is patch 5 which just
      essentially former patch 14 without the address of pointer logic.

   -> Patches 6-12 for this series are new, but follow along through the
      logic to make things private.

John Ferlan (12):
  nodedev: Alter virNodeDeviceObjRemove
  test: Adjust cleanup/error paths for nodedev test APIs
  nodedev: Use common naming for virnodedeviceobj
  nodedev: Use consistent names for driver variables
  nodedev: Introduce virNodeDeviceObjNew
  nodedev: Introduce virNodeDeviceObjListNew
  nodedev: Alter node device obj list function names
  nodedev: Dereference the obj/def in virNodeDeviceObjListFind* APIs
  nodedev: Introduce virNodeDeviceGetSCSIHostCaps
  nodedev: Introduce virNodeDeviceObjListFindSCSIHostByWWNs
  nodedev: Privatize _virNodeDeviceObj and _virNodeDeviceObjList
  nodedev: Convert virNodeDeviceObj to use virObjectLockable

 src/conf/node_device_conf.c               |  82 ++++++
 src/conf/node_device_conf.h               |  20 +-
 src/conf/virnodedeviceobj.c               | 420 ++++++++++++++++++------------
 src/conf/virnodedeviceobj.h               |  67 ++---
 src/libvirt_private.syms                  |  20 +-
 src/node_device/node_device_driver.c      | 159 +++++------
 src/node_device/node_device_hal.c         |  47 ++--
 src/node_device/node_device_linux_sysfs.c |  77 +-----
 src/node_device/node_device_udev.c        |  52 ++--
 src/test/test_driver.c                    | 136 +++++-----
 10 files changed, 570 insertions(+), 510 deletions(-)

-- 
2.9.4

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

ping?

Perhaps more involved - the first couple are from the previous series.
Once you get past the I hate it when someone renames things - it should
be a familiar sequence of creating a lockable object.

Tks,

John

On 06/03/2017 09:11 AM, John Ferlan wrote:
> v2: https://www.redhat.com/archives/libvir-list/2017-May/msg00999.html
> 
> Some patches from v2 were pushed (3, 4, 6, 7, 8, 9, 11, & 12), but a
> few remained from that series and are the first 5 patches of this series.
> 
> What changed?
>    -> Reworked the virNodeDeviceObjRemove patch (former patch 2, but new
>      series patch 1). That affected the Test patch (former patch 1, but
>      now patch 2). This patch removes the address of obj logic and moves
>      the onus of the ObjFree to the caller (see patch for reason).
>    -> Patch 3 is the former patch 5, with no essential change
>    -> Patch 4 is the former patch 10, with no essential change
>    -> Former patch 13 and 14, were altered to remove the offending
>       address of pointer logic. The result is patch 5 which just
>       essentially former patch 14 without the address of pointer logic.
> 
>    -> Patches 6-12 for this series are new, but follow along through the
>       logic to make things private.
> 
> John Ferlan (12):
>   nodedev: Alter virNodeDeviceObjRemove
>   test: Adjust cleanup/error paths for nodedev test APIs
>   nodedev: Use common naming for virnodedeviceobj
>   nodedev: Use consistent names for driver variables
>   nodedev: Introduce virNodeDeviceObjNew
>   nodedev: Introduce virNodeDeviceObjListNew
>   nodedev: Alter node device obj list function names
>   nodedev: Dereference the obj/def in virNodeDeviceObjListFind* APIs
>   nodedev: Introduce virNodeDeviceGetSCSIHostCaps
>   nodedev: Introduce virNodeDeviceObjListFindSCSIHostByWWNs
>   nodedev: Privatize _virNodeDeviceObj and _virNodeDeviceObjList
>   nodedev: Convert virNodeDeviceObj to use virObjectLockable
> 
>  src/conf/node_device_conf.c               |  82 ++++++
>  src/conf/node_device_conf.h               |  20 +-
>  src/conf/virnodedeviceobj.c               | 420 ++++++++++++++++++------------
>  src/conf/virnodedeviceobj.h               |  67 ++---
>  src/libvirt_private.syms                  |  20 +-
>  src/node_device/node_device_driver.c      | 159 +++++------
>  src/node_device/node_device_hal.c         |  47 ++--
>  src/node_device/node_device_linux_sysfs.c |  77 +-----
>  src/node_device/node_device_udev.c        |  52 ++--
>  src/test/test_driver.c                    | 136 +++++-----
>  10 files changed, 570 insertions(+), 510 deletions(-)
> 

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

On Wed, Jun 14, 2017 at 09:27:44PM -0400, John Ferlan wrote:
>
> ping?
>
> Perhaps more involved - the first couple are from the previous series.
> Once you get past the I hate it when someone renames things - it should
> be a familiar sequence of creating a lockable object.

Just a small note, as I haven't mentioned anything about that in my reviews so
far, we're in a freeze so I'd like you to wait with pushing the patches until
we're open again - if that information might have gotten lost in the ton of
emails that arrive to the list, so no offense was meant by writing this ;), it's
just I've seen a feature (if I remember correctly, definitely not a bugfix)
merged withing devel freeze, not that long time ago actually.

Erik

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

On 06/29/2017 10:47 AM, Erik Skultety wrote:
> On Wed, Jun 14, 2017 at 09:27:44PM -0400, John Ferlan wrote:
>>
>> ping?
>>
>> Perhaps more involved - the first couple are from the previous series.
>> Once you get past the I hate it when someone renames things - it should
>> be a familiar sequence of creating a lockable object.
> 
> Just a small note, as I haven't mentioned anything about that in my reviews so
> far, we're in a freeze so I'd like you to wait with pushing the patches until
> we're open again - if that information might have gotten lost in the ton of
> emails that arrive to the list, so no offense was meant by writing this ;), it's
> just I've seen a feature (if I remember correctly, definitely not a bugfix)
> merged withing devel freeze, not that long time ago actually.
> 
> Erik
> 

No worries - I'd wait for the next release anyway, much safer that way.
Still great to get a review now in order to be ready once the race to
commit is back on ;-)

John

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