Bug 743810

Summary: [RFE] corosync coroipcc: use malloc for path in service_connect
Product: Red Hat Enterprise Linux 6 Reporter: Jan Friesse <jfriesse>
Component: corosyncAssignee: Jan Friesse <jfriesse>
Status: CLOSED ERRATA QA Contact: Cluster QE <mspqa-list>
Severity: low Docs Contact:
Priority: low    
Version: 6.2CC: cluster-maint, jkortus, sdake
Target Milestone: rcKeywords: FutureFeature
Target Release: 6.3   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: corosync-1.4.1-5.el6 Doc Type: Enhancement
Doc Text:
Cause Each IPC connection is using 48K in stack. Consequence Multi threading applications with reduced stack size don't work correctly Fix Allocate private IPC data on heap Result Each IPC connection is using 48K in heap, so multi threading applications no longer need to justify it's stack size.
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-06-20 12:22:40 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On:    
Bug Blocks: 756082    
Attachments:
Description Flags
Proposed patch none

Description Jan Friesse 2011-10-06 07:21:07 UTC
Created attachment 526634 [details]
Proposed patch

Description of problem:

Coroipcc appropriately uses PATH_MAX sized variables for various data
structures handling files in the initialization of the client.  Due to
the use of 12 of these structures declared as stack variables, the
application stack balloons to over 12*4k. This is especially problematic
if threads are used by long running daemons to restart the connection
to corosync so as to be resilient in the face of system services
restarting (service corosync restart).

A simple alternative is to allocate temporary memory to avoid
requirements of large thread stacks.

Version-Release number of selected component (if applicable):
1.4.1

Additional info:
Patch exists in corosync 1.4.2

Unit test:

use massif tool, and testcpg where first operation is ctrl+d.

valgrind --tool=massif --stacks=yes   ./testcpg
ms_print massif.out.*

Before patch:
stacks(B) column is never higher then 20K

After patch:
stacks(B) column is higher then 30K

Comment 1 Jan Friesse 2011-10-06 07:23:10 UTC
In unit test, correct results are:

Before patch:
stacks(B) column is higher then 30K

After patch:
stacks(B) column is never higher then 20K

Comment 4 Jan Friesse 2012-03-07 07:20:01 UTC
    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
Cause
Each IPC connection is using 48K in stack.

Consequence
Multi threading applications with reduced stack size don't work correctly

Fix
Allocate private IPC data on heap

Result
Each IPC connection is using 48K in heap, so multi threading applications no longer need to justify it's stack size.

Comment 7 errata-xmlrpc 2012-06-20 12:22:40 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.

http://rhn.redhat.com/errata/RHBA-2012-0777.html