Dan Kenigsberg has posted comments on this change.
Change subject: vdsm: Handling topology for ppc64 ......................................................................
Patch Set 7: Code-Review-1
(3 comments)
.................................................... Commit Message Line 8: Line 9: This change adds the method to extract the topology from a specific Line 10: command for ppc64 machines. The current method for x86_64 hasn't Line 11: been changed. Line 12: Due libvirt shows wrong information about ppc64 topology, the lscpu Please include a URL to the relevant libvirt bug. Ovirt may need to circumvent this in the short term, but let's make it easy to revert this patch when the libvirt bug is solved. Line 13: command is being used to obtain it. Line 14: The topology tests were changed to be executed in a mock Line 15: architecture and new test cases were added to handle ppc64 methods. Line 16:
.................................................... File tests/Makefile.am Line 91: cpu_info.out \ Line 92: caps_libvirt_intel_E5649.out \ Line 93: caps_libvirt_amd_6274.out \ Line 94: caps_libvirt_intel_E31220.out \ Line 95: caps_lscpu_ppc64_1_4_4.out \ Makefiles love tabs. Line 96: caps_lscpu_ppc64_2_4_8.out \ Line 97: glob_1c60971a-8647-44ac-ae33-6520887f8843.out \ Line 98: glusterVolumeProfileInfo.xml \ Line 99: glusterVolumeProfileInfoNfs.xml \
.................................................... File tests/capsTests.py Line 61: testPath = os.path.realpath(__file__) Line 62: dirName = os.path.split(testPath)[0] Line 63: # PPC64 1 socket, 4 cores, 4 threads per core Line 64: path = os.path.join(dirName, "caps_lscpu_ppc64_1_4_4.out") Line 65: t = caps.CpuTopology(file(path).read()) file() is no longer cool. Use open(). Line 66: self.assertEqual(t.threads(), 16) Line 67: self.assertEqual(t.cores(), 4) Line 68: self.assertEqual(t.sockets(), 4) Line 69: # PPC64 2 sockets, 8 cores, 8 threads per core
vdsm-patches@lists.fedorahosted.org