- Devel - libvirt List Archives

[libvirt] Release of libvirt-0.8.4
by Daniel Veillard 10 Sep '10

10 Sep '10

As planned the release is now available on the server at ftp://libvirt.org/libvirt/ This is mostly an enhancement and bug fix release, there is less patches than for usual monthly releases, I guess a number of people had a break in August :-) Features: - various improvements to UML driver (Soren Hansen) Documentation: - docs: fix lxc examples (Serge Hallyn) - docs: mention domain <clock> improvements (Eric Blake) - docs: fix an incorrect keyword in updated hooks page (Justin Clift) - docs: significant expansion of custom hook script information (Justin Clift) Portability - cygwin: build fix (Stefan Berger) - build: allow mingw VPATH build (Eric Blake) - Generate libvirt_qemu.def from libvirt_qemu.syms for MinGW builds (Matthias Bolte) - Only require XDR when building libvirtd or the remote driver (Matthias Bolte) - build-sys: fix build when daemon is disabled by not installing libvirtd.8 (Diego Elio Pettenò) Bug Fixes: - Fix block statistics with newer versions of Xen (Guido Günther) - Fix off-by-1 in QEMU boot arg array handling (Daniel P. Berrange) - bridge: Fix static-only DHCP configuration (Jiri Denemark) - qemu: qemuMonitorJSONEjectMedia(): Fix arguments' type (Luiz Capitulino) - qemu: qemuMonitorJSONMigrate(): Fix arguments' type (Luiz Capitulino) - test: Don't overwrite storage volume target path and key (Matthias Bolte) - build: Fix permissions of sysconfig files (Jiri Denemark) - virsh: remove driver check from attach-disk command (Ryan Harper) - phyp: Bad comparison when checking for existing domain name (Eduardo Otubo) - esx: Fix esxVI_BuildSelectSet's invalid argument check (Matthias Bolte) - spec: Fix undefined with_libnl (Jiri Denemark) - Fix handling of sparse NUMA topologies (Daniel P. Berrange) - qemu: Check for errors when converting PCI address to string (Jiri Denemark) - qemu: Fix JSON migrate_set_downtime command (Jiri Denemark) - uml: fix logic bug in checking reply length (Eric Blake) - xen: Fix scheduler setting problems (Daniel Veillard) - nodedev: Free the right pointers when getting WWNs fails (Jiri Denemark) - nodedev: Fix sysfs paths for vport operations (Dave Allan) - xen: Fix device count on detach (Daniel Berrange) - remote: Fix incorrect use of private data field (Daniel Berrange) - esx: Fix memory leak when looking up an non-existing domain by name (Matthias Bolte) - Fix up qemu domain save/managed save locking. (Chris Lalancette) - qemu: Release PCI slot when detaching disk and net devices (Jiri Denemark) - qemu: Re-reserve all PCI addresses on libvirtd restart (Jiri Denemark) - Close fd's of persistent tap devices (Soren Hansen) - Fix tunnelled migration with qemu running as qemu:qemu. (Chris Lalancette) - Fix valgrind complaints when using kill -SIGHUP on libvirtd (Stefan Berger) - bridge: Fix potential segfault when preparing dnsmasq arguments (Jiri Denemark) - Fix return value usage (Doug Goldstein) - nodeinfo: skip offline CPUs (Eric Blake) - Do not use boot=on on IDE device (Daniel Veillard) Improvements: - virsh: Option for overriding disk type in attach-disk (Jiri Denemark) - Explicitly pass uml_dir argument to user-mode-linux (Soren Hansen) - Fix dependancies for remote generated files (Daniel P. Berrange) - Ensure remote daemon unions are always non-zero length (Daniel P. Berrange) - Add nwfilter support to UML driver (Soren Hansen) - esx: Use SessionIsActive when available (Matthias Bolte) - esx: Fall back to path as key when QueryVirtualDiskUuid isn't available (Matthias Bolte) - mingw: match recent changes in spec file (Eric Blake) - Update of localization files (Daniel Veillard) - esx: Use the VirtualDisk UUID as storage volume key (Matthias Bolte) - esx: Add .vmdk storage volume creation (Matthias Bolte) - OpenVZ: add ethernet interface type support (Jean-Baptiste Rouault) - esx: Rework datastore path parsing and handling (Matthias Bolte) - esx: Fix generator for string return values (Matthias Bolte) - build: add some modules (Eric Blake) - Use global directory as UML's monitorDir for privileged connections (Soren Hansen) - maint: track moved file (Eric Blake) - Add tests for Xen's blktap2 implementation (Jim Fehlig) - Add blktap2 support to xend driver (Jim Fehlig) - esx: Map the .vmx annotation to the domain XML description (Matthias Bolte) - esx: Add read-only storage volume access (Matthias Bolte) - Move libvirt-guests init script and config to tools (Jiri Denemark) - Support virDomainAttachDevice and virDomainDetachDevice for disks in UML (Soren Hansen) - Log return value for virConnectGetCapabilities (Daniel P. Berrange) - Try harder to send RPC error message back to client (Daniel P. Berrange) - Add explicit warning messages when failing to serialize to XDR (Daniel P. Berrange) - Allow chardev of type 'file' for UML domains. (Soren Hansen) - esx: Improve object-by-type lookup performance (Matthias Bolte) - xen tests: Fix PV-VFB tests with RHEL-5 API (Jiri Denemark) - xen tests: Fix missing "type ioemu" with rhel5-api (Jiri Denemark) - nodeinfotest: Print libvirt error on failure (Jiri Denemark) - xenapi: support xenapi 5.6.0 headers (Eric Blake) - Add support for -enable-kqemu flag (Daniel P. Berrange) - esx: Use MD5 sum of mount path as storage pool UUID (Matthias Bolte) - esx: Make sure dumpxml outputs proper ID for active domains (Matthias Bolte) - Add actions to virDomainLifecycle enum (Jim Fehlig) - vbox: add location used in rpmfusion release (Eric Blake) - storage: add support for Vendor and Model in XML (Patrick Dignan) - qemu: Fix copy&paste error in warning message (Jiri Denemark) - nwfilter: extend nwfilter reload support (Stefan Berger) - esx: Explicitly disable unused floppy devices (Matthias Bolte) - phyp: Add rudimentary network driver (Eduardo Otubo) - Make umlConnectTapDevice ask brAddTap for a persistent tap device. (Soren Hansen) - Make sure all command line arguments get passed to UML (Soren Hansen) - nwfilter: Discard class D,E IP addresses when sniffing pkts (Stefan Berger) - nwfilter: serialize execution of scripts with ebtables cmds (Stefan Berger) - Move the tunnelled migration unix socket to /var/lib/libvirt/qemu (Chris Lalancette) - qemu: avoid calling the balloon info command if disabled (Daniel Veillard) - bridge: Add --dhcp-no-override option to dnsmasq (Jiri Denemark) - esx: Improve VMX file name parsing and formatting (Matthias Bolte) - esx: Split VMX code into a general and an ESX specific part (Matthias Bolte) - allow memballoon type of none to desactivate it (Daniel Veillard) - Add a detailed message when tap device add fails (Doug Goldstein) - qemu: Hack around asynchronous device_del (Jiri Denemark) - build-sys: only build the test programs during the check phase. (Diego Elio Pettenò) - Add "ubd" to the list of disk prefixes (Soren Hansen) - Extend virSocketParseAddr() to allow a NULL result pointer (Daniel Veillard) - bootloader_args is named bootargs in xen-xm (Philipp Hahn) - Fix return value usage (Doug Goldstein) - qemu: improve error if tun device is missing (Doug Goldstein) - build: rerun bootstrap if po/Makevars got nuked (Eric Blake) Cleanups: - Remove hack to get static binaries in DV environment (Daniel Veillard) - build: avoid uninitialized variable warning (Eric Blake) - vbox: factor a large function (Eric Blake) - lxc: avoid large stacks with veth creation (Eric Blake) - openvz: use virAsprintf to avoid large stacks (Eric Blake) - openvz: formatting cleanups (Eric Blake) - network: use virAsprintf when appropriate (Eric Blake) - Move hextobin as virHexToBin to util.c (Matthias Bolte) - Rename qemudShrinkDisks to virDomainDiskRemove and move to domain_conf.c (Soren Hansen) - xml2sexprtest: Remove graphics from unrelated tests (Jiri Denemark) - xenapi: Fix compile error in previous commit (Matthias Bolte) - qemu: Remove code duplication (Jiri Denemark) - vbox: factor a large function (Eric Blake) - xenapi: avoid sprintf (Eric Blake) - storage: avoid s[n]printf (Eric Blake) - maint: whitespace cleanups (Eric Blake) - build: delete dead comments (Eric Blake) - nwfilter: use consistent OOM reporting (Eric Blake) - build: fix compiler warning (Eric Blake) - build: fix AppArmor compilation (Jamie Strandboge) - Avoid unnecessary bootstrap runs in VPATH builds (Jiri Denemark) - maint: update an email address preference (Eric Blake) - phyp: refactor phypListDomainsGeneric to eliminate buffer overflow (Eduardo Otubo) Thanks everybody who helped for the release with patches, bug reports or documentation ! Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

1 0

[libvirt] Best place to put links to the libvirt App Dev Guide?
by Justin Clift 10 Sep '10

10 Sep '10

Hi guys, Where's the best place in the libvirt.org menu structure, to point people to the (developing) libvirt App Dev Guide? My initial thinking is: a) Have a link to the .pdf in the Downloads section + b) Maybe a new top title section like: Documentation -> "Dev Guide" With links to the html, html (one big page), and pdf files. Sound reasonable? Regards and best wishes, Justin Clift

5 8

[libvirt] [PATCH v3] buf: Fix possible infinite loop in EscapeString, VSnprintf
by Cole Robinson 10 Sep '10

10 Sep '10

The current code will go into an infinite loop if the printf generated string is >= 1000, AND exactly 1 character smaller than the amount of free space in the buffer. When this happens, we are dropped into the loop body, but nothing will actually change, because count == (buf->size - buf->use - 1), and virBufferGrow returns unchanged if count < (buf->size - buf->use) Fix this by removing the '- 1' bit from 'size'. The *nprintf functions handle the NULL byte for us anyways, so we shouldn't need to manually accommodate for it. Here's a bug where we are actually hitting this issue: https://bugzilla.redhat.com/show_bug.cgi?id=602772 v2: Eric's improvements: while -> if (), remove extra va_list variable, make sure we report buffer error if snprintf fails v3: Add tests/virbuftest which reproduces the infinite loop before this patch, works correctly after --- src/util/buf.c | 70 ++++++++++++++++++++++++++++--------------- tests/.gitignore | 1 + tests/Makefile.am | 7 ++++- tests/virbuftest.c | 84 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 137 insertions(+), 25 deletions(-) create mode 100644 tests/virbuftest.c diff --git a/src/util/buf.c b/src/util/buf.c index fc1217b..553e2a0 100644 --- a/src/util/buf.c +++ b/src/util/buf.c @@ -224,7 +224,7 @@ void virBufferVSprintf(const virBufferPtr buf, const char *format, ...) { int size, count, grow_size; - va_list locarg, argptr; + va_list argptr; if ((format == NULL) || (buf == NULL)) return; @@ -236,27 +236,38 @@ virBufferVSprintf(const virBufferPtr buf, const char *format, ...) virBufferGrow(buf, 100) < 0) return; - size = buf->size - buf->use - 1; va_start(argptr, format); - va_copy(locarg, argptr); - while (((count = vsnprintf(&buf->content[buf->use], size, format, - locarg)) < 0) || (count >= size - 1)) { + + size = buf->size - buf->use; + if ((count = vsnprintf(&buf->content[buf->use], + size, format, argptr)) < 0) { + buf->error = 1; + goto err; + } + + /* Grow buffer if necessary and retry */ + if (count >= size) { buf->content[buf->use] = 0; - va_end(locarg); + va_end(argptr); + va_start(argptr, format); - grow_size = (count > 1000) ? count : 1000; + grow_size = (count + 1 > 1000) ? count + 1 : 1000; if (virBufferGrow(buf, grow_size) < 0) { - va_end(argptr); - return; + goto err; } - size = buf->size - buf->use - 1; - va_copy(locarg, argptr); + size = buf->size - buf->use; + if ((count = vsnprintf(&buf->content[buf->use], + size, format, argptr)) < 0) { + buf->error = 1; + goto err; + } } - va_end(argptr); - va_end(locarg); buf->use += count; - buf->content[buf->use] = '\0'; + +err: + va_end(argptr); + return; } /** @@ -336,23 +347,34 @@ virBufferEscapeString(const virBufferPtr buf, const char *format, const char *st if ((buf->use >= buf->size) && virBufferGrow(buf, 100) < 0) { - VIR_FREE(escaped); - return; + goto err; + } + + size = buf->size - buf->use; + if ((count = snprintf(&buf->content[buf->use], size, + format, (char *)escaped)) < 0) { + buf->error = 1; + goto err; } - size = buf->size - buf->use - 1; - while (((count = snprintf(&buf->content[buf->use], size, format, - (char *)escaped)) < 0) || (count >= size - 1)) { + /* Grow buffer if necessary and retry */ + if (count >= size) { buf->content[buf->use] = 0; - grow_size = (count > 1000) ? count : 1000; + grow_size = (count + 1 > 1000) ? count + 1 : 1000; if (virBufferGrow(buf, grow_size) < 0) { - VIR_FREE(escaped); - return; + goto err; + } + size = buf->size - buf->use; + + if ((count = snprintf(&buf->content[buf->use], size, + format, (char *)escaped)) < 0) { + buf->error = 1; + goto err; } - size = buf->size - buf->use - 1; } buf->use += count; - buf->content[buf->use] = '\0'; + +err: VIR_FREE(escaped); } diff --git a/tests/.gitignore b/tests/.gitignore index 3f32939..e7c74c5 100644 --- a/tests/.gitignore +++ b/tests/.gitignore @@ -21,6 +21,7 @@ storagepoolxml2xmltest nodeinfotest statstest qparamtest +virbuftest seclabeltest eventtest *.exe diff --git a/tests/Makefile.am b/tests/Makefile.am index 1dc7f66..dd59034 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -73,7 +73,7 @@ EXTRA_DIST = \ qemuhelpdata check_PROGRAMS = virshtest conftest \ - nodeinfotest statstest qparamtest + nodeinfotest statstest qparamtest virbuftest if WITH_XEN check_PROGRAMS += xml2sexprtest sexpr2xmltest \ @@ -151,6 +151,7 @@ TESTS = virshtest \ nodeinfotest \ statstest \ qparamtest \ + virbuftest \ $(test_scripts) if WITH_XEN @@ -352,6 +353,10 @@ qparamtest_SOURCES = \ qparamtest.c testutils.h testutils.c qparamtest_LDADD = $(LDADDS) +virbuftest_SOURCES = \ + virbuftest.c testutils.h testutils.c +virbuftest_LDADD = $(LDADDS) + if WITH_LIBVIRTD eventtest_SOURCES = \ eventtest.c testutils.h testutils.c ../daemon/event.c diff --git a/tests/virbuftest.c b/tests/virbuftest.c new file mode 100644 index 0000000..1cc9410 --- /dev/null +++ b/tests/virbuftest.c @@ -0,0 +1,84 @@ +#include <config.h> + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#include "internal.h" +#include "util.h" +#include "testutils.h" +#include "buf.h" +#include "memory.h" + +#define TEST_ERROR(...) \ + do { \ + if (virTestGetDebug()) \ + fprintf(stderr, __VA_ARGS__); \ + } while (0) + +struct testInfo { + int doEscape; +}; + +static int testBufInfiniteLoop(const void *data ATTRIBUTE_UNUSED) +{ + virBuffer bufinit = VIR_BUFFER_INITIALIZER; + virBufferPtr buf = &bufinit; + char *addstr = NULL, *bufret = NULL; + int ret = -1; + const struct testInfo *info = data; + + virBufferAddChar(buf, 'a'); + if (buf->a != 1002 || buf->b != 1) { + TEST_ERROR("Buf was not expected size, size=%d use=%d\n", + buf->a, buf->b); + goto out; + } + + /* + * Infinite loop triggers if: + * (strlen + 1 > 1000) && (strlen == buf-size - buf-use - 1) + */ + if (virAsprintf(&addstr, "%*s", buf->a - buf->b - 1, "a") < 0) { + goto out; + } + + if (info->doEscape) + virBufferEscapeString(buf, "%s", addstr); + else + virBufferVSprintf(buf, "%s", addstr); + + ret = 0; +out: + bufret = virBufferContentAndReset(buf); + if (!bufret) { + TEST_ERROR("Buffer had error set"); + ret = -1; + } + + VIR_FREE(addstr); + VIR_FREE(bufret); + return ret; +} + +static int +mymain(int argc ATTRIBUTE_UNUSED, + char **argv ATTRIBUTE_UNUSED) +{ + int ret = 0; + + +# define DO_TEST(msg, cb, data) \ + do { \ + struct testInfo info = { data }; \ + if (virtTestRun("Buf: " msg, 1, cb, &info) < 0) \ + ret = -1; \ + } while (0) + + DO_TEST("EscapeString infinite loop", testBufInfiniteLoop, 1); + DO_TEST("VSprintf infinite loop", testBufInfiniteLoop, 0); + + return(ret==0 ? EXIT_SUCCESS : EXIT_FAILURE); +} + +VIRT_TEST_MAIN(mymain) -- 1.7.2.1

2 2

[libvirt] [PATCH] fix path to xen device statistics for newer kernels
by Guido Günther 10 Sep '10

10 Sep '10

Hi, it seems Xen changed the paths to the block device statistics in sysfs: from "/sys/devices/xen-backend/vbd-%d-%d/statistics/%s" to "/sys/bus/xen-backend/devices/vbd-%d-%d/statistics/%s": http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=596004 Attached patch fixes this. O.k. to apply? -- Guido

3 2

[libvirt] [PATCH] virsh: Option for overriding disk type in attach-disk
by Jiri Denemark 10 Sep '10

10 Sep '10

Unless --driver tap|file option was given to attach-disk, virsh would generate <disk type='block'> XML which might be fine for Xen but not for other hypervisors. This patch introduces a new option --sourcetype which can be used to explicitly set the type of disk source. The option accepts either "file" or "block" types. --- tools/virsh.c | 12 ++++++++++-- 1 files changed, 10 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 7bc33d8..57ea618 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -8116,6 +8116,7 @@ static const vshCmdOptDef opts_attach_disk[] = { {"type", VSH_OT_STRING, 0, N_("target device type")}, {"mode", VSH_OT_STRING, 0, N_("mode of device reading and writing")}, {"persistent", VSH_OT_BOOL, 0, N_("persist disk attachment")}, + {"sourcetype", VSH_OT_STRING, 0, N_("type of source (block|file)")}, {NULL, 0, 0, NULL} }; @@ -8127,6 +8128,7 @@ cmdAttachDisk(vshControl *ctl, const vshCmd *cmd) int isFile = 0, ret = FALSE; char *buf = NULL, *tmp = NULL; unsigned int flags; + char *stype; if (!vshConnectionUsability(ctl, ctl->conn)) goto cleanup; @@ -8144,10 +8146,16 @@ cmdAttachDisk(vshControl *ctl, const vshCmd *cmd) subdriver = vshCommandOptString(cmd, "subdriver", NULL); type = vshCommandOptString(cmd, "type", NULL); mode = vshCommandOptString(cmd, "mode", NULL); + stype = vshCommandOptString(cmd, "sourcetype", NULL); - if (driver) { - if (STREQ(driver, "file") || STREQ(driver, "tap")) + if (!stype) { + if (driver && (STREQ(driver, "file") || STREQ(driver, "tap"))) isFile = 1; + } else if (STREQ(stype, "file")) { + isFile = 1; + } else if (STRNEQ(stype, "block")) { + vshError(ctl, _("Unknown source type: '%s'"), stype); + goto cleanup; } if (mode) { -- 1.7.2.2

3 3

[libvirt] [PATCH] bridge: Fix static-only DHCP configuration
by Jiri Denemark 10 Sep '10

10 Sep '10

For static-only DHCP, i.e. with no <range> but at least one <host> element within <dhcp> element, we have to add "--dhcp-range IP,static" option to dnsmasq to actually enable the service. Without this option, dnsmasq will not respond to DHCP requests. --- src/network/bridge_driver.c | 26 +++++++++++++++++++++++--- 1 files changed, 23 insertions(+), 3 deletions(-) diff --git a/src/network/bridge_driver.c b/src/network/bridge_driver.c index d6d3068..01d2171 100644 --- a/src/network/bridge_driver.c +++ b/src/network/bridge_driver.c @@ -394,6 +394,16 @@ networkBuildDnsmasqArgv(virNetworkObjPtr network, int nbleases = 0; char *pidfileArg; char buf[1024]; + unsigned int ranges; + + /* + * For static-only DHCP, i.e. with no range but at least one host element, + * we have to add a special --dhcp-range option to enable the service in + * dnsmasq. + */ + ranges = network->def->nranges; + if (!ranges && network->def->nhosts) + ranges = 1; /* * NB, be careful about syntax for dnsmasq options in long format @@ -424,11 +434,11 @@ networkBuildDnsmasqArgv(virNetworkObjPtr network, /*2 + *//* --interface virbr0 */ 2 + /* --except-interface lo */ 2 + /* --listen-address 10.0.0.1 */ - (2 * network->def->nranges) + /* --dhcp-range 10.0.0.2,10.0.0.254 */ + (2 * ranges) + /* --dhcp-range 10.0.0.2,10.0.0.254 */ /* --dhcp-lease-max=xxx if needed */ (network->def->nranges ? 1 : 0) + /* --dhcp-no-override if needed */ - (network->def->nranges ? 1 : 0) + + (ranges ? 1 : 0) + /* --dhcp-hostsfile=/var/lib/dnsmasq/$NAME.hostsfile */ (network->def->nhosts > 0 ? 1 : 0) + /* --enable-tftp --tftp-root /srv/tftp */ @@ -496,12 +506,22 @@ networkBuildDnsmasqArgv(virNetworkObjPtr network, nbleases += network->def->ranges[r].size; } + if (!network->def->nranges && network->def->nhosts) { + snprintf(buf, sizeof(buf), "%s,static", + network->def->ipAddress); + + APPEND_ARG(*argv, i++, "--dhcp-range"); + APPEND_ARG(*argv, i++, buf); + } + if (network->def->nranges > 0) { snprintf(buf, sizeof(buf), "--dhcp-lease-max=%d", nbleases); APPEND_ARG(*argv, i++, buf); - APPEND_ARG(*argv, i++, "--dhcp-no-override"); } + if (ranges) + APPEND_ARG(*argv, i++, "--dhcp-no-override"); + if (network->def->nhosts > 0) { dnsmasqContext *dctx = dnsmasqContextNew(network->def->name, DNSMASQ_STATE_DIR); char *hostsfileArg; -- 1.7.2.2

2 2

[libvirt] kqemu
by Zvi Dubitzky 10 Sep '10

10 Sep '10

Hi I could not get an answer from the users list . Maybe here someone can give some advice . we have a Linux 64 bit system with qemu 0.9.1-10 ( uname -a = 2.6.26-2-amd64) installed . kvm is included and loaded to memory . Still when we create a VM we sometimes get the following error message: libvir: QEMU error : internal error unable to start guest: You do not have enough space in '/dev/shm' for the 499 MB of QEMU virtual RAM. To have more space available provided you have enough RAM and swap, do as root: mount -o remount,size=515m /dev/shm Or disable the accelerator module with -no-kqemu After we do the mount (as advised) we can create the VM . Does the error mean that kvm is not active ? It seems that kqemu is an old way to do the duties of kvm e.g supply memory to VMs via /dev/shm . Is there a way to avoid that error ? I guess the -no-kqemu should be to qemu but we do not have control over the invokation of qemu by libvirt. thanks Zvi Dubitzky Email:dubi@il.ibm.com

1 0

[libvirt] QEMU json related fixes
by Luiz Capitulino 10 Sep '10

10 Sep '10

Found these problems while running libvirt-tck with up to date libvirt and qemu.git. Details in the patches. Thanks.

2 3

[libvirt] [PATCH] Add nwfilter support to UML driver
by Soren Hansen 09 Sep '10

09 Sep '10

Extend user-mode-linux driver to support nwfilter. Signed-off-by: Soren Hansen <soren(a)linux2go.dk> --- src/uml/uml_conf.c | 16 +++++++++++++--- src/uml/uml_driver.c | 8 +++++++- 2 files changed, 20 insertions(+), 4 deletions(-) diff --git a/src/uml/uml_conf.c b/src/uml/uml_conf.c index 4906192..f2eaef5 100644 --- a/src/uml/uml_conf.c +++ b/src/uml/uml_conf.c @@ -46,6 +46,7 @@ #include "verify.h" #include "bridge.h" #include "logging.h" +#include "domain_nwfilter.h" #define VIR_FROM_THIS VIR_FROM_UML @@ -108,7 +109,8 @@ virCapsPtr umlCapsInit(void) { static int -umlConnectTapDevice(virDomainNetDefPtr net, +umlConnectTapDevice(virConnectPtr conn, + virDomainNetDefPtr net, const char *bridge) { brControl *brctl = NULL; @@ -164,6 +166,14 @@ umlConnectTapDevice(virDomainNetDefPtr net, goto error; } + if (net->filter) { + if (virDomainConfNWFilterInstantiate(conn, net)) { + if (template_ifname) + VIR_FREE(net->ifname); + goto error; + } + } + brShutdown(brctl); return 0; @@ -239,7 +249,7 @@ umlBuildCommandLineNet(virConnectPtr conn, goto error; } - if (umlConnectTapDevice(def, bridge) < 0) { + if (umlConnectTapDevice(conn, def, bridge) < 0) { VIR_FREE(bridge); goto error; } @@ -250,7 +260,7 @@ umlBuildCommandLineNet(virConnectPtr conn, } case VIR_DOMAIN_NET_TYPE_BRIDGE: - if (umlConnectTapDevice(def, def->data.bridge.brname) < 0) + if (umlConnectTapDevice(conn, def, def->data.bridge.brname) < 0) goto error; /* ethNNN=tuntap,tapname,macaddr,gateway */ diff --git a/src/uml/uml_driver.c b/src/uml/uml_driver.c index 0a5c829..40345d5 100644 --- a/src/uml/uml_driver.c +++ b/src/uml/uml_driver.c @@ -58,6 +58,7 @@ #include "domain_conf.h" #include "datatypes.h" #include "logging.h" +#include "domain_nwfilter.h" #define VIR_FROM_THIS VIR_FROM_UML @@ -876,6 +877,7 @@ static int umlStartVMDaemon(virConnectPtr conn, if (umlBuildCommandLine(conn, driver, vm, &keepfd, &argv, &progenv) < 0) { close(logfd); + virDomainConfVMNWFilterTeardown(vm); umlCleanupTapDevices(conn, vm); return -1; } @@ -928,8 +930,11 @@ static int umlStartVMDaemon(virConnectPtr conn, VIR_FREE(progenv[i]); VIR_FREE(progenv); - if (ret < 0) + if (ret < 0) { + virDomainConfVMNWFilterTeardown(vm); umlCleanupTapDevices(conn, vm); + } + /* NB we don't mark it running here - we do that async with inotify */ @@ -965,6 +970,7 @@ static void umlShutdownVMDaemon(virConnectPtr conn ATTRIBUTE_UNUSED, vm->def->id = -1; vm->state = VIR_DOMAIN_SHUTOFF; + virDomainConfVMNWFilterTeardown(vm); umlCleanupTapDevices(conn, vm); if (vm->newDef) { -- 1.7.1

4 4

[libvirt] [PATCH] test: Don't overwrite storage volume target path and key
by Matthias Bolte 09 Sep '10

09 Sep '10

Only generate target path and key when they are not defined in the XML config. --- src/test/test_driver.c | 20 ++++++++++++-------- 1 files changed, 12 insertions(+), 8 deletions(-) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 6c06cbc..9d22339 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -706,17 +706,21 @@ static int testOpenVolumesForPool(xmlDocPtr xml, goto error; } - if (virAsprintf(&def->target.path, "%s/%s", - pool->def->target.path, - def->name) == -1) { - virReportOOMError(); - goto error; + if (def->target.path == NULL) { + if (virAsprintf(&def->target.path, "%s/%s", + pool->def->target.path, + def->name) == -1) { + virReportOOMError(); + goto error; + } } - def->key = strdup(def->target.path); if (def->key == NULL) { - virReportOOMError(); - goto error; + def->key = strdup(def->target.path); + if (def->key == NULL) { + virReportOOMError(); + goto error; + } } pool->def->allocation += def->allocation; -- 1.7.0.4

2 2