Re: [Libvir] PATCH: 2/16: Internal driver API
"Daniel P. Berrange" <berrange(a)redhat.com> wrote:
This defines the internal driver API for the storage APIs. The
pattern follows that used for the existing APIs. NB, both the
storage pool and storage volume objects are now top level objects.
Previous iterations of this code have the volume as a child of
the pool. This unneccessarily complicated the reference counting
and forced you to always have a pool available first.
include/libvirt/virterror.h | 4
src/driver.h | 150 ++++
src/hash.c | 317 +++++++++
src/internal.h | 68 ++
src/libvirt.c | 1403 ++++++++++++++++++++++++++++++++++++++++++++
src/virterror.c | 21
6 files changed, 1961 insertions(+), 5 deletions(-)
...
diff -r 42e6f49e4e69 src/internal.h
--- a/src/internal.h Thu Feb 07 12:33:13 2008 -0500
+++ b/src/internal.h Thu Feb 07 16:52:34 2008 -0500
...
@@ -187,6 +209,8 @@ struct _virConnect {
pthread_mutex_t lock;
virHashTablePtr domains; /* hash table for known domains */
virHashTablePtr networks; /* hash table for known domains */
+ virHashTablePtr storagePools;/* hash table for known storage pools */
+ virHashTablePtr storageVols;/* hash table for known storage vols */
virJobPtr jobs; /* list of active background jobs */
int refs; /* reference count */
};
Hi Dan,
I tried to apply all of these patches, and ended up with a bunch of
failed hunks. For example, the one above fails due to the fact that
there is no "jobs" member in checked-in sources, yet the patch
requires that it be there in the existing context.
I inserted the two new members manually.
Here's the list:
8 out of 15 hunks FAILED -- saving rejects to file src/virsh.c.rej
5 out of 15 hunks FAILED -- saving rejects to file python/generator.py.rej
29 out of 183 hunks FAILED -- saving rejects to file docs/libvirt-refs.xml.rej
7 out of 21 hunks FAILED -- saving rejects to file docs/libvirt-api.xml.rej
1 out of 7 hunks FAILED -- saving rejects to file src/internal.h.rej
I'm not too worried about the .xml or even .py conflicts, for now.
The failed hunks in virsh.c all involve VIR_JOB_CANCELLED, so I suspect
that the failures are all ignorable.
Bottom line: looked problematic, but no big deal