Bug 493345

Summary: Buffer overrun on device name printing 'sar -p -d 1' on 64 bit machines
Product: Red Hat Enterprise Linux 4 Reporter: Issue Tracker <tao>
Component: sysstatAssignee: Ivana Varekova <varekova>
Status: CLOSED ERRATA QA Contact: BaseOS QE <qe-baseos-auto>
Severity: high Docs Contact:
Priority: high    
Version: 4.7CC: benl, ddumas, mpoole, rvokal, syeghiay, tao
Target Milestone: rc   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-02-16 14:06:39 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Issue Tracker 2009-04-01 14:11:29 UTC
Escalated to Bugzilla from IssueTracker

Comment 1 Issue Tracker 2009-04-01 14:11:31 UTC
Description of problem:

Customer can see inconsistent output with the "sar -p -d 1" command.

The output of this command is different on a RHEL 4 32bit system and on RHEL 4 64 bit system.

On a RHEL 4 32bit machine the DEV column of the output shows correct device names (ram0,hda1 etc)
but on a RHEL 4 64 bit machine it only shows numbers (0,1 etc ).

Customer wants to know why there is a difference in the output and if its a bug.


How reproducible:

Steps to Reproduce:

I have reproduced this issue on a RHEL 4 system running the current kernel and updated sysstat package.

Kernel : 2.6.9-78.0.17.EL
sysstat : 5.0.5-19.el4

On a RHEL 4 32 bit system:

#[root@dhcp6-146 ~]# sar -p -d 1
Linux 2.6.9-78.0.17.EL (dhcp6-146.pnq.redhat.com)       03/27/2009

01:14:51 PM    DEV              tps    rd_sec/s  wr_sec/s
01:14:52 PM    ram0             0.00      0.00      0.00
01:14:52 PM    ram1             0.00      0.00      0.00
01:14:52 PM    ram2             0.00      0.00      0.00
...
01:14:52 PM    hda              0.00      0.00      0.00
01:14:52 PM    hda1             0.00      0.00      0.00
01:14:52 PM    hda2             0.00      0.00      0.00
...


On a RHEL 4 64 bit system:

[root@dhcp7-180 ~]# sar -p -d 1
Linux 2.6.9-78.0.17.EL (dhcp7-180.pnq.redhat.com)       03/27/2009

05:34:25 PM    DEV              tps    rd_sec/s  wr_sec/s
05:34:26 PM    0                0.00      0.00      0.00
05:34:26 PM    1                0.00      0.00      0.00
05:34:26 PM    2                0.00      0.00      0.00
...
05:34:26 PM    a                1.02     16.33      0.00
05:34:26 PM    a1               0.00      0.00      0.00
05:34:26 PM    a2               1.02     16.33      0.00
...

Please find attached complete outputs.

Actual results:

Output of the "sar -p -d 1 " shows only numbers on the 64 bit system 
whereas shows correct device names on 32 bit system


Expected results:

Outputs on both 32 bit and 64 bit RHEL 4 systems should be identical.
i.e the DEV column should show correct device names (ram0,hda1 etc)


Additional info:

This behavior is only visible on a RHEL 4 system 64 bit system.
RHEL 5 on 32 and 64 bit machines show correct output.

I checked "man sar" and found the following comment for the "-p" option:

<snip>
Pretty-print  device  names.  By default names are printed as devn-m where n and m are the major and minor numbers for the
device.  Use of this option displays the names of the devices as they (should) appear in /dev.   Name  mappings  are  con-
trolled by /etc/sysconfig/sysstat.ioconf.
</snip>

But the file /etc/sysconfig/sysstat.ioconf is identical on both 32 & 64 bit systems.
This event sent from IssueTracker by mpoole  [Support Engineering Group]
 issue 280142

Comment 2 Martin Poole 2009-04-01 14:55:23 UTC
After some debugging have tracked it ioc_conv in common.c overrunning the "out" buffer due to incorrectly defined length.

Fix was included upstream

2005/11/28: Version 6.0.2 - Sebastien Godard (sysstat <at> wanadoo.fr)

        * Off-by-one error in ioc_conv which was corrupting device names
          on 64-bit architecture [Debian bug #334305].


Just needs to change definition of "out", was

  static char out[16];

should be

  static char out[17];


Since any buffer overrun has possible unintended consequences other than the symptoms already noted I am marking this as a blocker.

Comment 9 errata-xmlrpc 2011-02-16 14:06:39 UTC
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHBA-2011-0224.html