[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [PATCH 3/3] qemuxml2argvmock: Mock virNumaNodesetIsAvailable



On Wed, Nov 05, 2014 at 06:07:18PM +0100, Michal Privoznik wrote:
And yet again one fix of 90286418. The problem is, if libvirt is
built with NUMA disabled (--without-numad --without-numactl), the
testsuite doesn't count on that and some tests fail. This is
because the virNumaNodesetIsAvailable() function is taken from
another section of virnuma.c which contains functions stubs for
case NUMA is disabled at build time. And the function basically
yields an error. However, we want our testsuite to be
deterministic and hence we ought to mock the function.

Signed-off-by: Michal Privoznik <mprivozn redhat com>
---
tests/qemuxml2argvmock.c | 9 +++++++++
1 file changed, 9 insertions(+)

diff --git a/tests/qemuxml2argvmock.c b/tests/qemuxml2argvmock.c
index 7218747..7b8bdd6 100644
--- a/tests/qemuxml2argvmock.c
+++ b/tests/qemuxml2argvmock.c
@@ -39,3 +39,12 @@ virNumaGetMaxNode(void)

   return maxnodesNum;
}
+
+bool
+virNumaNodesetIsAvailable(virDomainNumatunePtr numatune)
+{
+    int maxGuestNode = virDomainNumatuneSpecifiedMaxNode(numatune);
+    int maxHostNode = virNumaGetMaxNode();
+
+    return maxGuestNode <= maxHostNode;
+}
--
2.0.4


NACK, we shouldn't need to mock this function.  I went over the commit
once more and I've realized it doesn't support non-contiguous nodesets
even if it could.  And since it will use only functions that have both
versions (better one with NUMA support and a stub without it) it can
be moved out of those #ifdefs.  I have the patch almost ready :)

I'll also remove the need for util to have connection to conf (it will
make more sense, hopefully.

Martin

Attachment: signature.asc
Description: Digital signature


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]