[Mono-bugs] [Bug 504411] New: Mprotect fails with ENOMEM in GC_remap

[bugzilla_noreply at novell.com]

http://bugzilla.novell.com/show_bug.cgi?id=504411


           Summary: Mprotect fails with ENOMEM in GC_remap
    Classification: Mono
           Product: Mono: Runtime
           Version: 2.4.x
          Platform: PowerPC-64
        OS/Version: SLES 11
            Status: NEW
          Severity: Critical
          Priority: P5 - None
         Component: GC
        AssignedTo: lupus at novell.com
        ReportedBy: munroesj at us.ibm.com
         QAContact: mono-bugs at lists.ximian.com
          Found By: ---


Created an attachment (id=292550)
 --> (http://bugzilla.novell.com/attachment.cgi?id=292550)
Diag printf to stderr in op_dep.c (GC_remap)

User-Agent:       Mozilla/5.0 (X11; U; Linux ppc64; en-US; rv:1.8.1.19)
Gecko/20081213 SUSE/2.0.0.19-0.3 Firefox/2.0.0.19

On large workloads (OpenSIM) we are seeing frequent failures where GC_remap
executes a mprotect call that fails with errno 12 (ENOMEM). So far we has seen
this running mono-2.4 from the release tar on SLES11 PowerPC64. 

This may be associated with SLES11 using 64k pages (verus 4K in SLES10).

Reproducible: Sometimes

Steps to Reproduce:
1. So far only reproduced in the OpenSIM
2.
3.
Actual Results:  
GC_remap aborts after reporting:

Mprotect failed at 0x40002770000 (length 851968) with errno 12

Expected Results:  
Should see this. Lupus suggested in #monodev that libgc should never unmap heap
storage.

At first through this was ulimit issue, but the diag printf prove that ulimit
are all set to unlimited.

-- 
Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.