Bug 736006

Summary: Defects revealed by the Coverity scan
Product: Red Hat Enterprise Linux 6 Reporter: Michal Luscon <mluscon>
Component: papiAssignee: William Cohen <wcohen>
Status: CLOSED ERRATA QA Contact: qe-baseos-tools
Severity: medium Docs Contact:
Priority: medium    
Version: 6.2CC: kdudka, mnowak, pmuller, praiskup
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Fixed In Version: papi-4.1.0-3.el6 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-12-06 18:56:09 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---
Bug Depends On:    
Bug Blocks: 743047    

Description Michal Luscon 2011-09-06 11:54:58 UTC
Description of problem:

/src/papi_pfm_events.c:820 - If return value from the function load_preset_table is NULL, NULL variable notemap will be dereferenced in the function free_notes (line #835).

/src/ctests/do_loops.c:304 - Declaring variable "i" without initializer.
/src/ctests/do_loops.c:309 - Using uninitialized value "i" if argc<=1 . 

/src/ctests/test_utils.c:813 - Comparing "hw_info" to null implies that "hw_info" might be null.
/src/ctests/test_utils.c:820 - Passing null variable "hw_info->model_string" to function "strcmp", which dereferences it.

Version-Release number of selected component (if applicable):

Additional info:
These defects were not present in the previous version of package.

Comment 2 William Cohen 2011-09-06 20:48:21 UTC
The array is zero'ed out before use in papi_pfm_events.c. Seems that static analysis tool is not following the man page that is a no-op. According to "man 3 free":

free() frees the memory space pointed to by ptr, which must  have  been
returned by a previous call to malloc(), calloc() or realloc().  Other-
wise, or if free(ptr) has already been called before, undefined  behav-
ior occurs.  If ptr is NULL, no operation is performed.

The remaining two things flagged by the static analysis tool are in the test routines and do not end up in the binary rpms. Patches have been sent to the upstream papi developers to fix /src/ctests/do_loops.c and /src/ctests/test_utils.c

However, it seems of limited utility to fix this particular set of issues in the RHEL rpm.

Comment 3 Kamil Dudka 2011-09-06 21:40:59 UTC
(In reply to comment #2)
> The array is zero'ed out before use in papi_pfm_events.c. Seems that static
> analysis tool is not following the man page that is a no-op. According to "man
> 3 free":

The tool does not care about man pages.  It checks the code.  Unless I am mistaken, the code is wrong.  If load_preset_table() returns non-zero, you pass the NULL pointer to free_notes(), which is _not_ no-op.

Please explain how exactly is the situation above handled by the code.

Comment 4 William Cohen 2011-09-08 17:02:55 UTC
You are correct. Have patches to address the problems found in the scans. The patch /src/ctests/do_loops.c and /src/ctests/test_utils.c has been accepted upstream. The upstream has signficantly changed for papi_pfm_events.c, so that patch won't be applicable to papi upstream.

Comment 6 William Cohen 2011-09-09 19:57:08 UTC
Built papi-4.1.3-3 with patches to fix the problems identified by coverity in this bz.

Comment 11 errata-xmlrpc 2011-12-06 18:56:09 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.