After this pile of changes I am back down to zero issues.
Patch 1 - is a repeat offender - this is the 3rd try, but I think
I have a better solution here. Details are in the commit message.
Patch 2-3 - are repeats from the last series as patches 6 and 9.
For the virtime.c code - I just removed the offending math and
deadcode leaving the entrails in the commit message and code.
For the virstoragefile.c code - I agree with eblake's logic and
just removed the entire check as dead.
Patch 4 - seems to be a real issue. Initially I thought it wasn't,
but after finding the original commit it seems it was a copypaste
type error. I initially thought perhaps if hard_limit was set as
the unlimited value that the design was don't change soft_limit and
swap_hard_limit
Patch 5 - false positive
Patch 6 - false positive mostly, but just add the proper checks like
other code
Patch 7 - the BAD_SIZEOF was being triggered on the math to the PROBE
macro. By generating local variables - the issue went away
Patch 8 - initially I wasn't so sure on this - that !!sock_path is just
one of those constructs I find "odd" to read in code. Anyway, as it turns
out this is a false positive as there is no way sock_path could be null
here - so I just added the sa_assert() and that satisfies Coverity
John Ferlan (8):
remote_driver: Resolve Coverity RESOURCE_LEAK
virtime: Resolve Coverity DEADCODE
virstoragefile: Resolve Coverity DEADCODE
domain_conf: Resolve Coverity COPY_PASTE_ERROR
virsh: Resolve Coverity DEADCODE
Resolve Coverity CHECKED_RETURN
qemu: Resolve Coverity BAD_SIZEOF
daemon: Resolve Coverity FORWARD_NULL
daemon/libvirtd.c | 1 +
src/conf/domain_conf.c | 4 ++--
src/qemu/qemu_monitor.c | 18 +++++++----------
src/remote/remote_driver.c | 49 ++++++++++++++++++++++++++--------------------
src/util/virstoragefile.c | 2 --
src/util/virtime.c | 14 ++++---------
tests/objecteventtest.c | 3 ++-
tools/virsh-domain.c | 1 +
tools/virt-login-shell.c | 2 +-
9 files changed, 46 insertions(+), 48 deletions(-)
--
1.9.3