Bug 724937
| Summary: | hwloc-1.2-0.fc16 fails xmlbuffer self check on PPC, but passes on PPC64 | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Karsten Hopp <karsten> |
| Component: | hwloc | Assignee: | Jiri Hladky <hladky.jiri> |
| Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | rawhide | CC: | Brice.Goglin, hladky.jiri |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | powerpc | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | hwloc-1.3-1.fc16 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2011-11-25 02:16:01 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
Karsten Hopp
2011-07-22 11:16:50 UTC
Just tested hwloc-1.2.1, bug is still there, contacting hwloc developers ppc-koji build --scratch dist-f16 rpmbuild/SRPMS/hwloc-1.2.1-0.fc14.src.rpm Please see a complete build log at http://ppc.koji.fedoraproject.org/koji/getfile?taskID=285892&name=build.log Thanks Jirka PASS: glibc-sched exported to buffer 0x10568a30 length 1835 re-exported to buffer 0x1056d118 length 1834 ### First exported buffer is: <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE topology SYSTEM "hwloc.dtd"> <topology> <object type="Machine" os_level="-1" os_index="0" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003" local_memory="16091512832"> <page_type size="17179869184" count="0"/> <page_type size="65536" count="245537"/> <page_type size="16777216" count="0"/> <info name="Backend" value="Linux"/> <info name="OSName" value="Linux"/> <info name="OSRelease" value="2.6.32-131.6.1.el6.ppc64"/> <info name="OSVersion" value="#1 SMP Mon Jun 20 14:15:43 EDT 2011"/> <info name="HostName" value="ppc-comm01"/> <info name="Architecture" value="ppc"/> <object type="Socket" os_level="-1" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003"> <object type="Cache" os_level="-1" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003" cache_size="4194304" depth="2" cache_linesize="128"> <object type="Cache" os_level="-1" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003" cache_size="65536" depth="1" cache_linesize="128"> <object type="Core" os_level="-1" os_index="0" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003"> <object type="PU" os_level="-1" os_index="0" cpuset="0x00000001" complete_cpuset="0x00000001" online_cpuset="0x00000001" allowed_cpuset="0x00000001"/> <object type="PU" os_level="-1" os_index="1" cpuset="0x00000002" complete_cpuset="0x00000002" online_cpuset="0x00000002" allowed_cpuset="0x00000002"/> </object> </object> </object> </object> </object> </topology> ### End of first export buffer ### Second exported buffer is: <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE topology SYSTEM "hwloc.dtd"> <topology> <object type="Machine" os_level="-1" os_index="0" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003" local_memory="16091512832"> <page_type size="4294967295" count="0"/> <page_type size="65536" count="245537"/> <page_type size="16777216" count="0"/> <info name="Backend" value="Linux"/> <info name="OSName" value="Linux"/> <info name="OSRelease" value="2.6.32-131.6.1.el6.ppc64"/> <info name="OSVersion" value="#1 SMP Mon Jun 20 14:15:43 EDT 2011"/> <info name="HostName" value="ppc-comm01"/> <info name="Architecture" value="ppc"/> <object type="Socket" os_level="-1" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003"> <object type="Cache" os_level="-1" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003" cache_size="4194304" depth="2" cache_linesize="128"> <object type="Cache" os_level="-1" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003" cache_size="65536" depth="1" cache_linesize="128"> <object type="Core" os_level="-1" os_index="0" cpuset="0x00000003" complete_cpuset="0x00000003" online_cpuset="0x00000003" allowed_cpuset="0x00000003"> <object type="PU" os_level="-1" os_index="0" cpuset="0x00000001" complete_cpuset="0x00000001" online_cpuset="0x00000001" allowed_cpuset="0x00000001"/> <object type="PU" os_level="-1" os_index="1" cpuset="0x00000002" complete_cpuset="0x00000002" online_cpuset="0x00000002" allowed_cpuset="0x00000002"/> </object> </object> </object> </object> </object> </topology> ### End of second export buffer FAIL: xmlbuffer ======================================================== 1 of 26 tests failed Please report to http://www.open-mpi.org/community/help/ ======================================================== Looks like we cast the pagesizes to unsigned long during XML import+export. Please try this patch. It should work with your 16Go pages :)
Thanks!
Brice
Index: src/topology-xml.c
===================================================================
--- src/topology-xml.c (révision 3812)
+++ src/topology-xml.c (copie de travail)
@@ -280,9 +280,9 @@
const xmlChar *value = hwloc__xml_import_attr_value(attr);
if (value) {
if (!strcmp((char *) attr->name, "size"))
- size = strtoul((char *) value, NULL, 10);
+ size = strtoull((char *) value, NULL, 10);
else if (!strcmp((char *) attr->name, "count"))
- count = strtoul((char *) value, NULL, 10);
+ count = strtoull((char *) value, NULL, 10);
else
fprintf(stderr, "ignoring unknown pagetype attribute %s\n", (char *) attr->name);
}
Ho, you'll need this too, otherwise the lines would be missordered. I reproduced and fixes the problem on x86_32 so I assume it'll work for you too. Index: src/topology.c =================================================================== --- src/topology.c (révision 3828) +++ src/topology.c (copie de travail) @@ -889,7 +889,12 @@ const struct hwloc_obj_memory_page_type_s *a = _a; const struct hwloc_obj_memory_page_type_s *b = _b; /* consider 0 as larger so that 0-size page_type go to the end */ - return b->size ? (int)(a->size - b->size) : -1; + if (!b->size) + return -1; + /* don't cast a-b in int since those are ullongs */ + if (b->size == a->size) + return 0; + return a->size < b->size ? -1 : 1; } Hi Brice, I have tried to apply your patches https://bugzilla.redhat.com/show_bug.cgi?id=724937#c2 https://bugzilla.redhat.com/show_bug.cgi?id=724937#c3 to both hwloc-1.2 and hwloc-1.2.1 but it's failing: =================================================================== patching file src/topology.c Hunk #1 FAILED at 889. patching file src/topology-xml.c Hunk #1 FAILED at 280. =================================================================== Could you please provide a new complete patch using version hwloc-1.2.1 as base? http://www.open-mpi.org/software/hwloc/v1.2/downloads/hwloc-1.2.1.tar.bz2 Thanks a lot! Jirka The patch I backported to v1.2 is https://svn.open-mpi.org/trac/hwloc/changeset/3834 By the way, there's a 1.2.2rc1 online, and I will do the final 1.2.2 next week. Brice Hi Brice, thanks a lot for creating 1.2.2rc1. I have tested it and the issue is fixed:-) I will wait for 1.2.2 to submit a new rpm for Fedora. Thanks Jiri hwloc-1.2.2-0.fc16 has been submitted as an update for Fedora 16. https://admin.fedoraproject.org/updates/hwloc-1.2.2-0.fc16 hwloc-1.2.2-0.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/hwloc-1.2.2-0.fc15 Package hwloc-1.2.2-0.fc16: * should fix your issue, * was pushed to the Fedora 16 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing hwloc-1.2.2-0.fc16' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/hwloc-1.2.2-0.fc16 then log in and leave karma (feedback). hwloc-1.2.2-1.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/hwloc-1.2.2-1.fc15 hwloc-1.2.2-1.fc16 has been submitted as an update for Fedora 16. https://admin.fedoraproject.org/updates/hwloc-1.2.2-1.fc16 hwloc-1.3-0.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/hwloc-1.3-0.fc15 hwloc-1.3-1.fc16 has been submitted as an update for Fedora 16. https://admin.fedoraproject.org/updates/hwloc-1.3-1.fc16 hwloc-1.3-0.fc15 has been pushed to the Fedora 15 stable repository. If problems still persist, please make note of it in this bug report. hwloc-1.3-1.fc16 has been pushed to the Fedora 16 stable repository. If problems still persist, please make note of it in this bug report. |