Subject is technically right, but entirely misleading. This patch is more a refactor of the code that fixes a coverity error rather than the other way around. On Fri, Sep 25, 2015 at 12:31:45 -0400, John Ferlan wrote:
Coverity flagged that if the input 'mask'parameter was NULL, then the strtok_r call would have a NULL first and third parameter, which would cause an issue.
Although the caller doesn't call getCpuBitMapfromString with a non-NULL mask value, Coverity doesn't check that. A fix could have been to check
So a false positive again?
the 'mask' in the function and keep Coverity happy; however, it was noted in a previous review that virBitmap* functions could be used instead.
So this patch modifies xenapiDomainGetVcpus to use virBitmap* functions and removes getCpuBitMapfromString in favor of those.
Signed-off-by: John Ferlan <jferlan@redhat.com> --- src/xenapi/xenapi_driver.c | 12 +++++++----- src/xenapi/xenapi_utils.c | 21 --------------------- src/xenapi/xenapi_utils.h | 2 -- 3 files changed, 7 insertions(+), 28 deletions(-)
This makes me like this patch though.
ACK if you fix the commit message. Peter