January 2008 - CIM - libvirt List Archives

[PATCH] Add a README.DMTF to the base_schema
by Dan Smith 15 Jan '08

15 Jan '08

# HG changeset patch # User Dan Smith <danms(a)us.ibm.com> # Date 1200334917 28800 # Node ID 4909cfcdc6288cd20d8c06a54b1fd458d8d298f3 # Parent 6cf619d174232d9c4f2cfab1543ed863443a92d2 Add a README.DMTF to the base_schema Signed-off-by: Dan Smith <danms(a)us.ibm.com> diff -r 6cf619d17423 -r 4909cfcdc628 base_schema/Makefile.am --- a/base_schema/Makefile.am Mon Jan 14 09:33:40 2008 -0800 +++ b/base_schema/Makefile.am Mon Jan 14 10:21:57 2008 -0800 @@ -6,4 +6,6 @@ dist_pkgdata_SCRIPTS = install_base_sche dist_pkgdata_SCRIPTS = install_base_schema.sh $(CIM_SCHEMA_ZIP): - wget http://www.dmtf.org/standards/cim/cim_schema_v$(CIM_SCHEMA_VER)/$(CIM_SCHEM… \ No newline at end of file + wget http://www.dmtf.org/standards/cim/cim_schema_v$(CIM_SCHEMA_VER)/$(CIM_SCHEM… + +EXTRA_DIST = README.DMTF diff -r 6cf619d17423 -r 4909cfcdc628 base_schema/README.DMTF --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/base_schema/README.DMTF Mon Jan 14 10:21:57 2008 -0800 @@ -0,0 +1,9 @@ +The DMTF CIM schema files included here are required for proper +initialization of the CIM server environment. They are permitted for +distribution per the guidelines at the top of the main MOF file: + +// DMTF is a not-for-profit association of industry members dedicated +// to promoting enterprise and systems management and interoperability. +// DMTF specifications and documents may be reproduced for uses +// consistent with this purpose by members and non-members, +// provided that correct attribution is given.

2 1

[PATCH] Bump version to 0.2
by Dan Smith 15 Jan '08

15 Jan '08

# HG changeset patch # User Dan Smith <danms(a)us.ibm.com> # Date 1200323991 28800 # Node ID 29c9bb8cb88428491764aa58f9bb214c3d7f5f03 # Parent db25d5f62c9d662cd262ea82efb6d3fde2806b62 Bump version to 0.2 Signed-off-by: Dan Smith <danms(a)us.ibm.com> diff -r db25d5f62c9d -r 29c9bb8cb884 configure.ac --- a/configure.ac Mon Jan 14 07:19:12 2008 -0800 +++ b/configure.ac Mon Jan 14 07:19:51 2008 -0800 @@ -1,6 +1,6 @@ # (C) Copyright IBM Corp. 2005 -AC_INIT(libvirt CMPI provider, 0.1, danms(a)us.ibm.com, libvirt-cim) +AC_INIT(libvirt CMPI provider, 0.2, danms(a)us.ibm.com, libvirt-cim) AC_CONFIG_SRCDIR([src/Virt_ComputerSystem.c])

2 2

[PATCH] Relax the KVM type detection logic
by Dan Smith 15 Jan '08

15 Jan '08

# HG changeset patch # User Dan Smith <danms(a)us.ibm.com> # Date 1200067225 28800 # Node ID a353b12b24858b16b3b6d77940178f5cd672bdce # Parent 4ae48640b6dfaf77c60300aa2f41a3be9395bf83 Relax the KVM type detection logic since it was too strict for no reason. Signed-off-by: Dan Smith <danms(a)us.ibm.com> diff -r 4ae48640b6df -r a353b12b2485 libxkutil/device_parsing.c --- a/libxkutil/device_parsing.c Fri Jan 11 07:59:47 2008 -0800 +++ b/libxkutil/device_parsing.c Fri Jan 11 08:00:25 2008 -0800 @@ -652,8 +652,7 @@ static int parse_os(struct domain *domin if ((STREQC(dominfo->os_info.fv.type, "hvm")) && (STREQC(dominfo->typestr, "xen"))) dominfo->type = DOMAIN_XENFV; - else if ((STREQC(dominfo->os_info.fv.type, "hvm")) && - (STREQC(dominfo->typestr, "kvm"))) + else if (STREQC(dominfo->typestr, "kvm")) dominfo->type = DOMAIN_KVM; else if (STREQC(dominfo->os_info.pv.type, "linux")) dominfo->type = DOMAIN_XENPV;

2 1

[PATCH] Make cleanup_dominfo() handle KVM domains
by Dan Smith 15 Jan '08

15 Jan '08

# HG changeset patch # User Dan Smith <danms(a)us.ibm.com> # Date 1200067187 28800 # Node ID 4ae48640b6dfaf77c60300aa2f41a3be9395bf83 # Parent e834b93342a431c9414fbb07c2aa34e28002744d Make cleanup_dominfo() handle KVM domains This avoids the following message in the CIMOM logs: device_parsing.c(818): Unknown domain type 2 Signed-off-by: Dan Smith <danms(a)us.ibm.com> diff -r e834b93342a4 -r 4ae48640b6df libxkutil/device_parsing.c --- a/libxkutil/device_parsing.c Fri Jan 11 07:40:28 2008 -0800 +++ b/libxkutil/device_parsing.c Fri Jan 11 07:59:47 2008 -0800 @@ -810,7 +810,8 @@ void cleanup_dominfo(struct domain **dom free(dom->os_info.pv.kernel); free(dom->os_info.pv.initrd); free(dom->os_info.pv.cmdline); - } else if (dom->type == DOMAIN_XENFV) { + } else if ((dom->type == DOMAIN_XENFV) || + (dom->type == DOMAIN_KVM)) { free(dom->os_info.fv.type); free(dom->os_info.fv.loader); free(dom->os_info.fv.boot);

2 1

[PATCH] CodingStyle updates based on experience gained in the past few months
by Dan Smith 14 Jan '08

14 Jan '08

# HG changeset patch # User Dan Smith <danms(a)us.ibm.com> # Date 1200085592 28800 # Node ID a541f98f909ab469ebd1631899045cf81faf4657 # Parent a42d6f035ed6d5d519e2efd5aa2fd45d5a7a95be CodingStyle updates based on experience gained in the past few months. Comments and adjustments welcomed. Signed-off-by: Dan Smith <danms(a)us.ibm.com> diff -r a42d6f035ed6 -r a541f98f909a doc/CodingStyle --- a/doc/CodingStyle Fri Jan 11 17:15:53 2008 +0800 +++ b/doc/CodingStyle Fri Jan 11 13:06:32 2008 -0800 @@ -1,6 +1,64 @@ + +The CodingStyle for libvirt-cim (and libcmpiutil) is mostly kernel +style, but with the following "clarifications": + - Eight-space indents -- No single-line if statements + +- No single-line if statements. This means none of this: + + if (foo) bar; + - Keywords and parens separated by one space (i.e. "if (" not "if(") -- 80-char width limit -- Identifier style: use_underbars_and_lowercase -- Macro style: ALL_UPPERCASE + +- 80-char width limit. Break long function calls by putting *every* + parameter of the call on its own line. There are a few exceptions + to the "every parameter" rule, such as for cu_statusf() calls. + Unless you've seen another example in the code, put each parameter + on a line. + +- Identifiers should be named with underbars_and_lowercase. The + libvirt style is to use CamelCase, but try to limit that to + operations directly related to libvirt and don't spread it to the + rest of the code. + +- Macros should be ALL_UPPERCASE + +- Braces around blocks should look like this: + + if (foo) { + . . . + } else if (bar) { + . . . + } else { + . . . + } + +- Absolutely all files should contain the following block at the end: + + /* + * Local Variables: + * mode: C + * c-set-style: "K&R" + * tab-width: 8 + * c-basic-offset: 8 + * indent-tabs-mode: nil + * End: + */ + + This ensures that emacs folks will have the style correct (they're + already halfway there by their choice in editor anyway). + +- Pointer and integer comparisons should be against something + concrete. The only time the following style should be used is if + foo is a bool: + + if (!foo) /* ONLY if foo is a bool */ + bar; + + For pointers and integers, do this: + + if (foo == NULL) /* if foo is a pointer */ + bar; + + if (foo == -1) /* if foo is an integer */ + bar;

2 1

Xen_ElementCapabilities return expected result from one side querying by non existing instance
by Guo Lian Yun 14 Jan '08

14 Jan '08

With a non existing instance, e.g. a wrong key name or value, Xen_VirtualSystemManagementCapabilities still return expected results below: wbemcli ain -ac Xen_ElementCapabilities http://root:password@localhost/root/virt:Xen_VirtualSystemManagementCapabil…" wrong" localhost:5988/root/virt:Xen_HostSystem.CreationClassName="Xen_HostSystem",Name="zeit" wbemcli ain -ac Xen_ElementCapabilities http://root:password@localhost/root/virt:Xen_VirtualSystemManagementCapabil…. Wrong="wrong" localhost:5988/root/virt:Xen_HostSystem.CreationClassName="Xen_HostSystem",Name="zeit" The same problem with Xen_VirtualSystemMigrationCapabilities. In conclusion, four association providers have the same problems, that is still return expected result with non existing instance from one or two sides querying. They are Xen_HostedResourcePool, Xen_ResourceAllocationFromPool, Xen_HostedService and Xen_ElementCapabilities. Please who can have a look into them? Thanks a lot! Best, Regards Daisy Guo Lian Yun E-mail: yunguol(a)cn.ibm.com IBM China Development Lab, Shanghai, China TEL: (86)-21-61008057

2 2

Xen_HostedService return expected queried result with non existing instance by two sides
by Guo Lian Yun 14 Jan '08

14 Jan '08

Hi, With a non existing instance, e.g. a wrong key name or value, Xen_HostedService still return expected results below: wbemcli ain -ac Xen_HostedService http://root:password@localhost/root/virt:Xen_ResourcePoolConfigurationServi…" wrong" localhost:5988/root/virt:Xen_HostSystem.CreationClassName="Xen_HostSystem",Name="zeit" wbemcli ain -ac Xen_HostedService http://root:password@localhost/root/virt:Xen_HostSystem.CreationClassName=" wrong",Wrong="zeit" localhost:5988/root/virt:Xen_ResourcePoolConfigurationService.CreationClassName="Xen_ResourcePoolConfigurationService",Name="RPCS",SystemCreationClassName="Xen_HostSystem",SystemName="zei" localhost:5988/root/virt:Xen_VirtualSystemManagementService.CreationClassName="Xen_VirtualSystemManagementService",Name="Management Service",SystemCreationClassName="Xen_HostSystem",SystemName="zeit" localhost:5988/root/virt:Xen_VirtualSystemMigrationService.CreationClassName="Xen_VirtualSystemMigrationService",Name="MigrationService",SystemCreationClassName="",SystemName="" Also, some keys value of Xen_VirtualSystemMigrationService are NULL. I think it's maybe provider's fault, the same issue of Xen_HostedResourcePool and Xen_ResourceAllocationFromPool, which have been reported before. Please who can have a look into them? Thanks a lot! Best, Regards Daisy Guo Lian Yun E-mail: yunguol(a)cn.ibm.com IBM China Development Lab, Shanghai, China TEL: (86)-21-61008057

2 2

[PATCH] Add code style phrase to inst_list.c
by Heidi Eckhart 12 Jan '08

12 Jan '08

# HG changeset patch # User Heidi Eckhart <heidieck(a)linux.vnet.ibm.com> # Date 1200052746 -3600 # Node ID eec4b4e334563efaf7e3f64e1e4d0770969dc925 # Parent 5278b5fa14a754ea8c6cfdd7870cdb7fc6c62fe7 Add code style phrase to inst_list.c Signed-off-by: Heidi Eckhart <heidieck(a)linux.vnet.ibm.com> diff -r 5278b5fa14a7 -r eec4b4e33456 inst_list.c --- a/inst_list.c Fri Jan 11 12:57:49 2008 +0100 +++ b/inst_list.c Fri Jan 11 12:59:06 2008 +0100 @@ -75,3 +75,12 @@ int inst_list_add(struct inst_list *list return 1; } +/* + * Local Variables: + * mode: C + * c-set-style: "K&R" + * tab-width: 8 + * c-basic-offset: 8 + * indent-tabs-mode: nil + * End: + */

3 2

[PATCH 0 of 3] .#2 Memory devices from XML and raw XML
by Dan Smith 12 Jan '08

12 Jan '08

Changes to the third patch only. Sorry the first round didn't have a commit log :)

3 7

[PATCH] Make inst_list_free() handle NULL pointers
by Heidi Eckhart 12 Jan '08

12 Jan '08

# HG changeset patch # User Heidi Eckhart <heidieck(a)linux.vnet.ibm.com> # Date 1200052669 -3600 # Node ID 5278b5fa14a754ea8c6cfdd7870cdb7fc6c62fe7 # Parent 393ba1344f5be2ea42b21adea0d377de440516dc Make inst_list_free() handle NULL pointers Signed-off-by: Heidi Eckhart <heidieck(a)linux.vnet.ibm.com> diff -r 393ba1344f5b -r 5278b5fa14a7 inst_list.c --- a/inst_list.c Fri Jan 11 10:33:08 2008 +0100 +++ b/inst_list.c Fri Jan 11 12:57:49 2008 +0100 @@ -50,6 +50,9 @@ void inst_list_init(struct inst_list *li void inst_list_free(struct inst_list *list) { + if (!list) + return; + free(list->list); inst_list_init(list); }

2 1