- Devel - libvirt List Archives

[libvirt] virt-dmesg and virt-uname
by Richard W.M. Jones 20 Jun '11

20 Jun '11

I just pushed the virt-dmesg package to Fedora Rawhide[1]. http://koji.fedoraproject.org/koji/buildinfo?buildID=248890 This tool uses libvirt to let you list the kernel messages from a Linux KVM guest without needing to log in or run any special agent in the guest. It's useful when a guest crashes or gets stuck. Usage is pretty simple: # virt-dmesg F14x64 | head <6>[ 0.000000] Initializing cgroup subsys cpuset <6>[ 0.000000] Initializing cgroup subsys cpu <5>[ 0.000000] Linux version 2.6.35.13-91.fc14.x86_64 (mockbuild(a)x86-15.phx2.fedoraproject.org) (gcc version 4.5.1 20100924 (Red Hat 4.5.1-4) (GCC) ) #1 SMP Tue May 3 13:23:06 UTC 2011 <6>[ 0.000000] Command line: ro root=/dev/mapper/vg_f13x64-lv_root rd_LVM_LV=vg_f13x64/lv_root rd_LVM_LV=vg_f13x64/lv_swap rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=uk rhgb quiet <6>[ 0.000000] BIOS-provided physical RAM map: <6>[ 0.000000] BIOS-e820: 0000000000000000 - 000000000009f400 (usable) <6>[ 0.000000] BIOS-e820: 000000000009f400 - 00000000000a0000 (reserved) <6>[ 0.000000] BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved) <6>[ 0.000000] BIOS-e820: 0000000000100000 - 000000003fffd000 (usable) <6>[ 0.000000] BIOS-e820: 000000003fffd000 - 0000000040000000 (reserved) In some cases virt-dmesg won't be able to find the symbol table, in which case you can fall back to dumping out the kernel and grepping it, eg: # virt-dmesg --dump-kernel F14x64 | strings | grep '^<[0-9]>' | less There's an accompanying tool called virt-uname which does what you would think: # virt-uname F14x64 Linux f14x64.home.annexia.org 2.6.35.13-91.fc14.x86_64 #1 SMP Tue May 3 13:23:06 UTC 2011 x86_64 (none) Please let me know if you find (Linux) guests for which virt-dmesg doesn't work properly (particularly the --dump-kernel option which ought to work for any Linux guest). Describe the guest as accurately as you can and provide the output of 'virt-dmesg -v Guest'. Rich. [1] Sadly because of some regressions in libvirt, it *won't* work in Fedora 15 or earlier unless you are able to upgrade to libvirt >= 0.9.1-3. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

1 1

[libvirt] [PATCH 1/3] virsh: clarify snapshot vs. save
by Eric Blake 20 Jun '11

20 Jun '11

* tools/virsh.c (info_snapshot_create, info_save): Clarify description. * tools/virsh.pod (save): Likewise. --- tools/virsh.c | 6 +++--- tools/virsh.pod | 8 +++++++- 2 files changed, 10 insertions(+), 4 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 1c0b1dc..6b5f3d9 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -1469,7 +1469,7 @@ cmdStart(vshControl *ctl, const vshCmd *cmd) */ static const vshCmdInfo info_save[] = { {"help", N_("save a domain state to a file")}, - {"desc", N_("Save a running domain.")}, + {"desc", N_("Save the RAM state of a running domain.")}, {NULL, NULL} }; @@ -10947,8 +10947,8 @@ cmdQuit(vshControl *ctl, const vshCmd *cmd ATTRIBUTE_UNUSED) * "snapshot-create" command */ static const vshCmdInfo info_snapshot_create[] = { - {"help", N_("Create a snapshot")}, - {"desc", N_("Snapshot create")}, + {"help", N_("Create a snapshot from XML")}, + {"desc", N_("Create a snapshot (disk and RAM) from XML")}, {NULL, NULL} }; diff --git a/tools/virsh.pod b/tools/virsh.pod index c6879cb..33f72ed 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -594,7 +594,8 @@ the state file was created. =item B<save> I<domain-id> I<state-file> -Saves a running domain to a state file so that it can be restored +Saves a running domain (RAM, but not disk state) to a state file so that +it can be restored later. Once saved, the domain will no longer be running on the system, thus the memory allocated for the domain will be free for other domains to use. B<virsh restore> restores from this state file. @@ -603,6 +604,11 @@ This is roughly equivalent to doing a hibernate on a running computer, with all the same limitations. Open network connections may be severed upon restore, as TCP timeouts may have expired. +Domain saved state files assume that disk images will be unchanged +between the creation and restore point. For a more complete system +restore point, where the disk state is saved alongside the memory +state, see the B<snapshot> family of commands. + =item B<schedinfo> optional I<--set> B<parameter=value> I<domain-id> I<--config> I<--live> I<--current> -- 1.7.1

3 10

[libvirt] [PATCH v2]: set and restore MAC address of a NIC when using PASSTHROUGH mode
by Gerhard Stenzel 20 Jun '11

20 Jun '11

This is a rework of the patch Dirk sent out last week. The attached patch addresses the problem that when a PASSTHROUGH mode DIRECT NIC connection is made the MAC address of the NIC is not automatically set and reset to the configured VM MAC and back again. The patch fixes this problem by setting and resetting the MAC while remembering the previous setting while the VM is running. This also works if libvirtd is restarted while the VM is running. The patch passes make syntax-check Signed-off-by: Dirk Herrendoerfer <d.herrendoerfer at herrendoerfer.name> Signed-off-by: Gerhard Stenzel <gerhard.stenzel(a)de.ibm.com> --- src/libvirt_macvtap.syms | 2 + src/qemu/qemu_command.c | 48 +++++++++++++++ src/qemu/qemu_process.c | 45 ++++++++++++++ src/util/macvtap.c | 149 ++++++++++++++++++++++++++++++++++++++++++++++ src/util/macvtap.h | 6 ++ 5 files changed, 250 insertions(+), 0 deletions(-) Index: libvirt/src/libvirt_macvtap.syms =================================================================== --- libvirt.orig/src/libvirt_macvtap.syms +++ libvirt/src/libvirt_macvtap.syms @@ -5,6 +5,8 @@ # macvtap.h delMacvtap; +getMacaddr; openMacvtapTap; +setMacaddr; vpAssociatePortProfileId; vpDisassociatePortProfileId; Index: libvirt/src/qemu/qemu_command.c =================================================================== --- libvirt.orig/src/qemu/qemu_command.c +++ libvirt/src/qemu/qemu_command.c @@ -125,6 +125,77 @@ qemuPhysIfaceConnect(virDomainDefPtr def net->model && STREQ(net->model, "virtio")) vnet_hdr = 1; + /** Note: When using PASSTHROUGH mode with MACVTAP devices the link + * device's MAC address must be set to the VMs MAC address. In + * order to not confuse the first switch or bridge in line this MAC + * address must be reset when the VM is shut down. + * This is especially important when using SRIOV capable cards that + * emulate their switch in firmware. + */ + if (net->data.direct.mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + unsigned char oldmac[6]; + rc = getMacaddr(&oldmac, net->data.direct.linkdev); + if (rc) { + virReportSystemError(rc, + _("Getting MAC address from '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x' failed."), + net->data.direct.linkdev, + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]); + } else { + char *oldmacpath = NULL; + char *oldmacname = NULL; + char *newmacname = NULL; + char *basedir = "/var/run/libvirt/network"; + + if (virAsprintf(&oldmacname,"%02x:%02x:%02x:%02x:%02x:%02x", + oldmac[0], oldmac[1], oldmac[2], + oldmac[3], oldmac[4], oldmac[5]) < 0) { + virReportOOMError(); + } + if (virAsprintf(&oldmacpath,"%s/%s", basedir, oldmacname) < 0) { + virReportOOMError(); + } + + if (virAsprintf(&newmacname, "%s/%s@%02x:%02x:%02x:%02x:%02x:%02x", + basedir, + net->data.direct.linkdev, + net->mac[0],net->mac[1],net->mac[2], + net->mac[3],net->mac[4],net->mac[5]) < 0) { + virReportOOMError(); + } + + rc = symlink (oldmacname, newmacname); + if (rc) { + virReportSystemError(errno, + _("MAC link file creation failed for %s."), + net->data.direct.linkdev); + } + rc = creat(oldmacpath, S_IWUSR); + if (rc) { + virReportSystemError(errno, + _("MAC link file creation failed for %s."), + oldmacpath); + } + rc = VIR_CLOSE(rc); + if (rc) { + virReportSystemError(errno, + _("MAC link file closing failed for %s."), + oldmacpath); + } + } + + rc = setMacaddr(net->mac, net->data.direct.linkdev); + if (rc) { + virReportSystemError(errno, + _("Setting MAC address on '%s' to " + "'%02x:%02x:%02x:%02x:%02x:%02x' failed."), + net->data.direct.linkdev, + net->mac[0], net->mac[1], net->mac[2], + net->mac[3], net->mac[4], net->mac[5]); + } + } + rc = openMacvtapTap(net->ifname, net->mac, net->data.direct.linkdev, net->data.direct.mode, vnet_hdr, def->uuid, &net->data.direct.virtPortProfile, &res_ifname, Index: libvirt/src/qemu/qemu_process.c =================================================================== --- libvirt.orig/src/qemu/qemu_process.c +++ libvirt/src/qemu/qemu_process.c @@ -2708,6 +2708,48 @@ void qemuProcessStop(struct qemud_driver if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { delMacvtap(net->ifname, net->mac, net->data.direct.linkdev, &net->data.direct.virtPortProfile); + + if (net->data.direct.mode == VIR_DOMAIN_NETDEV_MACVTAP_MODE_PASSTHRU) { + char *newmacname = NULL; + char *oldmacname = NULL; + char *oldmacpath = NULL; + char *basedir = "/var/run/libvirt/network"; + + if (virAsprintf(&newmacname, "%s/%s@%02x:%02x:%02x:%02x:%02x:%02x", + basedir, + net->data.direct.linkdev, + net->mac[0],net->mac[1],net->mac[2], + net->mac[3],net->mac[4],net->mac[5]) < 0) { + virReportOOMError(); + } + + ret = virFileResolveLink(newmacname, &oldmacpath); + if (ret < 0) { + virReportSystemError(errno, + _("Reading MAC address from '%s' failed !"), + newmacname); + } else { + unsigned char oldmac[6]; + + oldmacname = basename(oldmacpath); + + if (virParseMacAddr(oldmacname, &oldmac[0]) != 0) { + virReportSystemError(VIR_ERR_INTERNAL_ERROR, + _("Cannot parse MAC address from '%s'"), + oldmacname); + } + + VIR_WARN(("Resetting MAC address on '%s' " + "to '%02x:%02x:%02x:%02x:%02x:%02x'."), + net->data.direct.linkdev, + oldmac[0],oldmac[1],oldmac[2], + oldmac[3],oldmac[4],oldmac[5]); + /*reset mac and remove link file-ignore results*/ + ret = setMacaddr(oldmac, net->data.direct.linkdev); + ret = unlink(oldmacpath); + ret = unlink(newmacname); + } + } VIR_FREE(net->ifname); } } Index: libvirt/src/util/macvtap.c =================================================================== --- libvirt.orig/src/util/macvtap.c +++ libvirt/src/util/macvtap.c @@ -87,6 +87,7 @@ # define LLDPAD_PID_FILE "/var/run/lldpad.pid" +#define MACADDRSIZE 6 enum virVirtualPortOp { ASSOCIATE = 0x1, @@ -191,6 +192,149 @@ err_exit: # if WITH_MACVTAP +/** + * getMacaddr: + * Get the MAC address of a network device + * + * @macaddress: Pointer where the MAC address will be stored + * @srcdev: The interface name of the NIC to get the MAC from + * + * Returns zero in case of success, + * negative value otherwise with error reported. + * + */ +int +getMacaddr(const unsigned char *macaddress, const char *srcdev ) +{ + int sockfd; + int io; + struct ifreq ifr; + + strcpy(ifr.ifr_name, srcdev); + + sockfd = socket(AF_INET, SOCK_STREAM, 0); + if(sockfd < 0){ + return -1; + } + + io = ioctl(sockfd, SIOCGIFHWADDR, (char *)&ifr); + if(io < 0){ + return -1; + } + + memcpy(macaddress, ifr.ifr_ifru.ifru_hwaddr.sa_data, MACADDRSIZE); + + return 0; +} + +/** + * setMacaddr: + * Set the MAC address of a network device + * + * @macaddress: MAC address to assign to the NIC + * @srcdev: The interface name of the NIC + * + * Returns zero in case of success, + * negative value otherwise with error reported. + * + */ +int +setMacaddr(const unsigned char *macaddress, const char *srcdev ) +{ + int rc = 0; + struct nlmsghdr *resp; + struct nlmsgerr *err; + struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC }; + int ifindex; + unsigned char *recvbuf = NULL; + unsigned int recvbuflen; + struct nl_msg *nl_msg; + + if (ifaceGetIndex(true, srcdev, &ifindex) != 0) + return -1; + + nl_msg = nlmsg_alloc_simple(RTM_SETLINK, NLM_F_REQUEST); + + if (!nl_msg) { + virReportOOMError(); + return -1; + } + + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0) + goto buffer_too_small; + + if (nla_put_u32(nl_msg, IFLA_LINK, ifindex) < 0) + goto buffer_too_small; + + if (nla_put(nl_msg, IFLA_ADDRESS, MACADDRSIZE, macaddress) < 0) + goto buffer_too_small; + + if (srcdev && + nla_put(nl_msg, IFLA_IFNAME, strlen(srcdev)+1, srcdev) < 0) + goto buffer_too_small; + + if (nlComm(nl_msg, &recvbuf, &recvbuflen, 0) < 0) { + rc = -1; + goto err_exit; + } + + if (recvbuflen < NLMSG_LENGTH(0) || recvbuf == NULL) + goto malformed_resp; + + resp = (struct nlmsghdr *)recvbuf; + + switch (resp->nlmsg_type) { + case NLMSG_ERROR: + err = (struct nlmsgerr *)NLMSG_DATA(resp); + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err))) + goto malformed_resp; + + switch (err->error) { + + case 0: + break; + + case -EEXIST: + rc = -1; + break; + + default: + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("error setting device mac address")); + rc = -1; + } + break; + + case NLMSG_DONE: + break; + + default: + goto malformed_resp; + } + +err_exit: + nlmsg_free(nl_msg); + + VIR_FREE(recvbuf); + + return rc; + +malformed_resp: + nlmsg_free(nl_msg); + + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("malformed netlink response message")); + VIR_FREE(recvbuf); + return -1; + +buffer_too_small: + nlmsg_free(nl_msg); + + macvtapError(VIR_ERR_INTERNAL_ERROR, "%s", + _("allocated netlink buffer is too small")); + return -1; +} + static int link_add(const char *type, const unsigned char *macaddress, int macaddrsize, Index: libvirt/src/util/macvtap.h =================================================================== --- libvirt.orig/src/util/macvtap.h +++ libvirt/src/util/macvtap.h @@ -75,6 +75,10 @@ enum virVMOperationType { # include "internal.h" +int getMacaddr(const unsigned char *macaddress, const char *srcdev ); + +int setMacaddr(const unsigned char *macaddress, const char *srcdev ); + int openMacvtapTap(const char *ifname, const unsigned char *macaddress, const char *linkdev, =================================================================== Best regards, Gerhard Stenzel ------------------------------------------------------------------------------------- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Martin Jetter Geschaeftsfuehrung: Dirk Wittkopp Sitz der Gesellschaft: Boeblingen Registergericht: Amtsgericht Stuttgart, HRB 243294

3 3

[libvirt] [PATCH] virsh: Fix compile error due to const mismatch in cmdVcpupin
by Matthias Bolte 20 Jun '11

20 Jun '11

Broken by commit c4a8ca71b10eb adding the reset option. Instead of altering a const string move the while loop into an else clause. --- tools/virsh.c | 110 ++++++++++++++++++++++++++++----------------------------- 1 files changed, 54 insertions(+), 56 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index e04c9fc..0e9efad 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3000,7 +3000,7 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) int cpumaplen; int i, cpu, lastcpu, maxcpu; bool unuse = false; - char *cur; + const char *cur; int config = vshCommandOptBool(cmd, "config"); int live = vshCommandOptBool(cmd, "live"); int current = vshCommandOptBool(cmd, "current"); @@ -3061,72 +3061,70 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) cpumap = vshCalloc(ctl, 0, cpumaplen); /* Parse cpulist */ - cur = cpulist; - if (*cur == 0) { + if (*cpulist == '\0') { goto parse_error; - } else if (*cur == 'r') { + } else if (*cpulist == 'r') { for (cpu = 0; cpu < maxcpu; cpu++) VIR_USE_CPU(cpumap, cpu); - *cur = 0; - } - - while (*cur != 0) { - - /* the char '^' denotes exclusive */ - if (*cur == '^') { - cur++; - unuse = true; - } - - /* parse physical CPU number */ - if (!c_isdigit(*cur)) - goto parse_error; - cpu = virParseNumber(&cur); - if (cpu < 0) { - goto parse_error; - } - if (cpu >= maxcpu) { - vshError(ctl, _("Physical CPU %d doesn't exist."), cpu); - goto parse_error; - } - virSkipSpaces(&cur); - - if ((*cur == ',') || (*cur == 0)) { - if (unuse) { - VIR_UNUSE_CPU(cpumap, cpu); - } else { - VIR_USE_CPU(cpumap, cpu); + } else { + cur = cpulist; + while (*cur != '\0') { + /* the char '^' denotes exclusive */ + if (*cur == '^') { + cur++; + unuse = true; } - } else if (*cur == '-') { - /* the char '-' denotes range */ - if (unuse) { + + /* parse physical CPU number */ + if (!c_isdigit(*cur)) goto parse_error; - } - cur++; - virSkipSpaces(&cur); - /* parse the end of range */ - lastcpu = virParseNumber(&cur); - if (lastcpu < cpu) { + cpu = virParseNumber(&cur); + if (cpu < 0) { goto parse_error; } - if (lastcpu >= maxcpu) { - vshError(ctl, _("Physical CPU %d doesn't exist."), maxcpu); + if (cpu >= maxcpu) { + vshError(ctl, _("Physical CPU %d doesn't exist."), cpu); goto parse_error; } - for (i = cpu; i <= lastcpu; i++) { - VIR_USE_CPU(cpumap, i); - } virSkipSpaces(&cur); - } - if (*cur == ',') { - cur++; - virSkipSpaces(&cur); - unuse = false; - } else if (*cur == 0) { - break; - } else { - goto parse_error; + if ((*cur == ',') || (*cur == '\0')) { + if (unuse) { + VIR_UNUSE_CPU(cpumap, cpu); + } else { + VIR_USE_CPU(cpumap, cpu); + } + } else if (*cur == '-') { + /* the char '-' denotes range */ + if (unuse) { + goto parse_error; + } + cur++; + virSkipSpaces(&cur); + /* parse the end of range */ + lastcpu = virParseNumber(&cur); + if (lastcpu < cpu) { + goto parse_error; + } + if (lastcpu >= maxcpu) { + vshError(ctl, _("Physical CPU %d doesn't exist."), maxcpu); + goto parse_error; + } + for (i = cpu; i <= lastcpu; i++) { + VIR_USE_CPU(cpumap, i); + } + virSkipSpaces(&cur); + } + + if (*cur == ',') { + cur++; + virSkipSpaces(&cur); + unuse = false; + } else if (*cur == '\0') { + break; + } else { + goto parse_error; + } } } -- 1.7.0.4

1 2

[libvirt] [PATCH] Fix compile warnings in virsh vcpupin
by Daniel P. Berrange 20 Jun '11

20 Jun '11

The 'char *cur' variable was being assigned from a 'const char *' string, thus discarding constness. As well as causing a compile warning, it masked a piece of code which attempts to assign to the previously const string. * tools/virsh.c: Fix const-ness of 'cur' variable in vcpupin --- tools/virsh.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 3df0683..ed26b93 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -3000,7 +3000,7 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) int cpumaplen; int i, cpu, lastcpu, maxcpu; bool unuse = false; - char *cur; + const char *cur; int config = vshCommandOptBool(cmd, "config"); int live = vshCommandOptBool(cmd, "live"); int current = vshCommandOptBool(cmd, "current"); @@ -3067,7 +3067,7 @@ cmdVcpupin(vshControl *ctl, const vshCmd *cmd) } else if (*cur == 'r') { for (cpu = 0; cpu < maxcpu; cpu++) VIR_USE_CPU(cpumap, cpu); - *cur = 0; + cur = ""; } while (*cur != 0) { -- 1.7.4.4

2 1

[libvirt] nwfilter: limit VM traffic to specific MAC
by Shahar Havivi 20 Jun '11

20 Jun '11

Hi, I am trying to add custom filter to block VM traffic to other VMs by limiting the traffic only to the gateways MAC address. The filter XML: <filter name='rhev' chain='root'> <uuid>cd4e5890-ccc9-1b0f-303f-e7fe7123646d</uuid> <filterref filter='allow-dhcp'/> <rule action='drop' direction='out' priority='500'> <mac match='no' dstmacaddr='$MAC'/> </rule> </filter> The MAC is not the interface MAC address it's the gateways MAC that pass as a parameter (I use the gateway address hardcoded as well). The VM is getting DHCP ip but cannot get any traffic, I notice that when I edit (comment and uncomment) the drop rule, the filter is working fine, ie no traffic other then the gateway. 1. Am I doing something wrong? 1. What is the table name that libvirt use for ebtables? Shahar.

2 3

[libvirt] Availability of patchchecker
by Daniel Veillard 20 Jun '11

20 Jun '11

As some may have noticed, I'm often behind on patch review and it happens I notice patches way too late i.e. just before a release. For some time I though that having something automatic to remind me about unreviewed/unapplied patches would be a good way to fight this. I know that Eric, Matthias, Dan ... scans mostly everything but well are all getting quite busy and sometime things get forgotten and that's bad for the person who spent time on it and it's bad for the community. Asking to resend old patches tend to inflate the problem it's not a solution and humanly it's poor taste. So I spent most part of the week writing patchchecker, it's an automated tool, working out of the public mail archives and a git checkout, and was tested only with libvirt but should work also with other projects using MHonarc for mail archives and git (or could be made more flexible, yadda yadda ...) http://www.libvirt.org/git/?p=patchchecker.git git clone git://libvirt.org/patchchecker.git Check the README, there is an indexer to run regulary to extract from the mail archives, and a command to try to show intereting things, right now it reports only patches which didn't get any comments or acks. It detects patchset but doesn't fully use them yet. It's crude but it's automatic ! Next step is to so a bit of XML output and XSLT to export the output as a web page on libvirt.org , I will also need to add support for patch versioning i.e. being able to obsolete an older version with a newer mail (or patch set). I guess discussion and patches should be held and sent to this list, which already raises a limitation of this tool: it currently support only one git checkout, and if you send a patch for it it will complain if it doesn't get applied after being ACK'ed :-) BTW one of the results is that we should try to use the formal ACK keyword all the time (capital) when accepting and commiting a patch and I may occasionally send reply mail to threads with ACK or something like COMMITED if the algorithm of the checker get stuck on one patch. I really expect the amount of manual intervention to be limited to this I want the tool to help me, not give me/us more work :-) Hopefully it may prove useful to a few of us, and possibly be adapted to other projects (as usual I take patches :-) 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/

3 3

[libvirt] [PATCH] remote: Generate virDomainGetBlockPullInfo
by Matthias Bolte 20 Jun '11

20 Jun '11

It was already generatable but skipped. --- daemon/remote.c | 36 ------------------------------------ src/remote/remote_driver.c | 32 -------------------------------- src/remote/remote_protocol.x | 4 ++-- 3 files changed, 2 insertions(+), 70 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index 1914b25..8e0f47d 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1737,42 +1737,6 @@ cleanup: return rv; } -static int -remoteDispatchDomainGetBlockPullInfo(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client ATTRIBUTE_UNUSED, - virConnectPtr conn, - remote_message_header *hdr ATTRIBUTE_UNUSED, - remote_error * rerr, - remote_domain_get_block_pull_info_args *args, - remote_domain_get_block_pull_info_ret *ret) -{ - virDomainPtr dom = NULL; - virDomainBlockPullInfo tmp; - int rv = -1; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - if (!(dom = get_nonnull_domain(conn, args->dom))) - goto cleanup; - - if (virDomainGetBlockPullInfo(dom, args->path, &tmp, args->flags) < 0) - goto cleanup; - ret->cur = tmp.cur; - ret->end = tmp.end; - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - if (dom) - virDomainFree(dom); - return rv; -} - - /*-------------------------------------------------------------*/ static int diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index f4b43e0..5db1064 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -2671,38 +2671,6 @@ done: return rv; } -static int remoteDomainGetBlockPullInfo(virDomainPtr domain, - const char *path, - virDomainBlockPullInfoPtr info, - unsigned int flags) -{ - int rv = -1; - remote_domain_get_block_pull_info_args args; - remote_domain_get_block_pull_info_ret ret; - struct private_data *priv = domain->conn->privateData; - - remoteDriverLock(priv); - - make_nonnull_domain(&args.dom, domain); - args.path = (char *)path; - args.flags = flags; - - if (call(domain->conn, priv, 0, REMOTE_PROC_DOMAIN_GET_BLOCK_PULL_INFO, - (xdrproc_t)xdr_remote_domain_get_block_pull_info_args, - (char *)&args, - (xdrproc_t)xdr_remote_domain_get_block_pull_info_ret, - (char *)&ret) == -1) - goto done; - - info->cur = ret.cur; - info->end = ret.end; - rv = 0; - -done: - remoteDriverUnlock(priv); - return rv; -} - /*----------------------------------------------------------------------*/ static virDrvOpenStatus ATTRIBUTE_NONNULL (1) diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index 2b9784b..b136e41 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -1003,7 +1003,7 @@ struct remote_domain_get_block_pull_info_args { unsigned int flags; }; -struct remote_domain_get_block_pull_info_ret { +struct remote_domain_get_block_pull_info_ret { /* insert@2 */ unsigned hyper cur; unsigned hyper end; }; @@ -2409,7 +2409,7 @@ enum remote_procedure { REMOTE_PROC_DOMAIN_BLOCK_PULL_ALL = 230, /* autogen autogen */ REMOTE_PROC_DOMAIN_BLOCK_PULL_ABORT = 231, /* autogen autogen */ - REMOTE_PROC_DOMAIN_GET_BLOCK_PULL_INFO = 232, /* skipgen skipgen */ + REMOTE_PROC_DOMAIN_GET_BLOCK_PULL_INFO = 232, /* autogen autogen */ REMOTE_PROC_DOMAIN_EVENT_BLOCK_PULL = 233 /* skipgen skipgen */ /* -- 1.7.0.4

2 2

[libvirt] [PATCH] remote: Handle functions thar return optional strings
by Matthias Bolte 20 Jun '11

20 Jun '11

--- daemon/remote.c | 64 ------------------------------------------ daemon/remote_generator.pl | 28 ++++++++++++++++++ src/remote/remote_protocol.x | 2 +- 3 files changed, 29 insertions(+), 65 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index 8e0f47d..f2e193d 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1485,70 +1485,6 @@ cleanup: } static int -remoteDispatchDomainScreenshot(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client, - virConnectPtr conn, - remote_message_header *hdr, - remote_error *rerr, - remote_domain_screenshot_args *args, - remote_domain_screenshot_ret *ret) -{ - int rv = -1; - struct qemud_client_stream *stream = NULL; - virDomainPtr dom = NULL; - char *mime, **mime_p; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - ret->mime = NULL; - - if (!(dom = get_nonnull_domain (conn, args->dom))) - goto cleanup; - - if (!(stream = remoteCreateClientStream(conn, hdr))) - goto cleanup; - - if (!(mime = virDomainScreenshot(dom, stream->st, args->screen, args->flags))) - goto cleanup; - - if (remoteAddClientStream(client, stream, 1) < 0) { - virStreamAbort(stream->st); - goto cleanup; - } - - if (VIR_ALLOC(mime_p) < 0) { - virReportOOMError(); - goto cleanup; - } - - *mime_p = strdup(mime); - if (*mime_p == NULL) { - virReportOOMError(); - VIR_FREE(mime_p); - goto cleanup; - } - - ret->mime = mime_p; - - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - VIR_FREE(mime); - if (dom) - virDomainFree(dom); - if (stream && rv != 0) { - virStreamAbort(stream->st); - remoteFreeClientStream(client, stream); - } - return rv; -} - -static int remoteDispatchNodeGetCPUStats (struct qemud_server *server ATTRIBUTE_UNUSED, struct qemud_client *client ATTRIBUTE_UNUSED, virConnectPtr conn, diff --git a/daemon/remote_generator.pl b/daemon/remote_generator.pl index bcf5fd4..d9d9ded 100755 --- a/daemon/remote_generator.pl +++ b/daemon/remote_generator.pl @@ -351,6 +351,7 @@ elsif ($opt_b) { my @optionals_list = (); my @getters_list = (); my @args_list = (); + my @prepare_ret_list = (); my @ret_list = (); my @free_list = (); my @free_list_on_error = ("remoteDispatchError(rerr);"); @@ -551,6 +552,27 @@ elsif ($opt_b) { $single_ret_var = $1; $single_ret_by_ref = 0; $single_ret_check = " == NULL"; + } elsif ($ret_member =~ m/^remote_string (\S+);/) { + push(@vars_list, "char *$1 = NULL"); + push(@vars_list, "char **$1_p = NULL"); + push(@ret_list, "ret->$1 = $1_p;"); + push(@free_list, " VIR_FREE($1);"); + push(@free_list_on_error, "VIR_FREE($1_p);"); + push(@prepare_ret_list, + "if (VIR_ALLOC($1_p) < 0) {\n" . + " virReportOOMError();\n" . + " goto cleanup;\n" . + " }\n" . + " \n" . + " *$1_p = strdup($1);\n" . + " if (*$1_p == NULL) {\n" . + " virReportOOMError();\n" . + " goto cleanup;\n" . + " }\n"); + + $single_ret_var = $1; + $single_ret_by_ref = 0; + $single_ret_check = " == NULL"; } elsif ($ret_member =~ m/^remote_nonnull_(domain|network|storage_pool|storage_vol|interface|node_device|secret|nwfilter|domain_snapshot) (\S+);/) { my $type_name = name_to_ProcName($1); @@ -864,6 +886,12 @@ elsif ($opt_b) { print "\n"; } + if (@prepare_ret_list) { + print " "; + print join("\n ", @prepare_ret_list); + print "\n"; + } + if (@ret_list) { print " "; print join("\n ", @ret_list); diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index b136e41..0aa3dfb 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -2386,7 +2386,7 @@ enum remote_procedure { REMOTE_PROC_STORAGE_VOL_DOWNLOAD = 209, /* autogen autogen | readstream@1 */ REMOTE_PROC_DOMAIN_INJECT_NMI = 210, /* autogen autogen */ - REMOTE_PROC_DOMAIN_SCREENSHOT = 211, /* skipgen autogen | readstream@1 */ + REMOTE_PROC_DOMAIN_SCREENSHOT = 211, /* autogen autogen | readstream@1 */ REMOTE_PROC_DOMAIN_GET_STATE = 212, /* skipgen skipgen */ REMOTE_PROC_DOMAIN_MIGRATE_BEGIN3 = 213, /* skipgen skipgen */ REMOTE_PROC_DOMAIN_MIGRATE_PREPARE3 = 214, /* skipgen skipgen */ -- 1.7.0.4

2 2

[libvirt] [PATCH] remote generator: Allow to annotate arrays with typecasts
by Matthias Bolte 20 Jun '11

20 Jun '11

Removes special case code from the generator and handle additional methods. The generated version of remoteDispatchDomainPinVcpu(Flags) has no length check, but this check was useless anyway as it was applied to data that was already deserialized from its XDR form. --- daemon/remote.c | 82 -------------------------------------- daemon/remote_generator.pl | 90 ++++++++++++++++++++++++++++++------------ src/remote/remote_protocol.x | 17 +++++--- 3 files changed, 74 insertions(+), 115 deletions(-) diff --git a/daemon/remote.c b/daemon/remote.c index f2e193d..37fbed0 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1269,88 +1269,6 @@ cleanup: } static int -remoteDispatchDomainPinVcpu(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client ATTRIBUTE_UNUSED, - virConnectPtr conn, - remote_message_header *hdr ATTRIBUTE_UNUSED, - remote_error *rerr, - remote_domain_pin_vcpu_args *args, - void *ret ATTRIBUTE_UNUSED) -{ - virDomainPtr dom = NULL; - int rv = -1; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - if (!(dom = get_nonnull_domain(conn, args->dom))) - goto cleanup; - - if (args->cpumap.cpumap_len > REMOTE_CPUMAP_MAX) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("cpumap_len > REMOTE_CPUMAP_MAX")); - goto cleanup; - } - - if (virDomainPinVcpu(dom, args->vcpu, - (unsigned char *) args->cpumap.cpumap_val, - args->cpumap.cpumap_len) < 0) - goto cleanup; - - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - if (dom) - virDomainFree(dom); - return rv; -} - -static int -remoteDispatchDomainPinVcpuFlags(struct qemud_server *server ATTRIBUTE_UNUSED, - struct qemud_client *client ATTRIBUTE_UNUSED, - virConnectPtr conn, - remote_message_header *hdr ATTRIBUTE_UNUSED, - remote_error *rerr, - remote_domain_pin_vcpu_flags_args *args, - void *ret ATTRIBUTE_UNUSED) -{ - virDomainPtr dom = NULL; - int rv = -1; - - if (!conn) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open")); - goto cleanup; - } - - if (!(dom = get_nonnull_domain(conn, args->dom))) - goto cleanup; - - if (args->cpumap.cpumap_len > REMOTE_CPUMAP_MAX) { - virNetError(VIR_ERR_INTERNAL_ERROR, "%s", _("cpumap_len > REMOTE_CPUMAP_MAX")); - goto cleanup; - } - - if (virDomainPinVcpuFlags(dom, - args->vcpu, - (unsigned char *) args->cpumap.cpumap_val, - args->cpumap.cpumap_len, - args->flags) < 0) - goto cleanup; - - rv = 0; - -cleanup: - if (rv < 0) - remoteDispatchError(rerr); - if (dom) - virDomainFree(dom); - return rv; -} - -static int remoteDispatchDomainGetMemoryParameters(struct qemud_server *server ATTRIBUTE_UNUSED, struct qemud_client *client diff --git a/daemon/remote_generator.pl b/daemon/remote_generator.pl index d9d9ded..146c56b 100755 --- a/daemon/remote_generator.pl +++ b/daemon/remote_generator.pl @@ -403,20 +403,32 @@ elsif ($opt_b) { " virDomainSnapshotFree(snapshot);\n" . " if (dom)\n" . " virDomainFree(dom);"); - } elsif ($args_member =~ m/^(remote_string|remote_nonnull_string|remote_uuid|opaque) (\S+)<\S+>;/) { + } elsif ($args_member =~ m/^(?:remote_string|remote_uuid) (\S+)<\S+>;/) { if (! @args_list) { push(@args_list, "conn"); } - if ($call->{ProcName} eq "SecretSetValue") { - push(@args_list, "(const unsigned char *)args->$2.$2_val"); - } elsif ($call->{ProcName} eq "CPUBaseline") { - push(@args_list, "(const char **)args->$2.$2_val"); - } else { - push(@args_list, "args->$2.$2_val"); + push(@args_list, "args->$1.$1_val"); + push(@args_list, "args->$1.$1_len"); + } elsif ($args_member =~ m/^(?:opaque|remote_nonnull_string) (\S+)<\S+>;(.*)$/) { + if (! @args_list) { + push(@args_list, "conn"); + } + + my $cast = ""; + my $arg_name = $1; + my $annotation = $2; + + if ($annotation ne "") { + if ($annotation =~ m/\s*\/\*\s*(.*)\s*\*\//) { + $cast = $1; + } else { + die "malformed cast annotation for argument: $args_member"; + } } - push(@args_list, "args->$2.$2_len"); + push(@args_list, "${cast}args->$arg_name.${arg_name}_val"); + push(@args_list, "args->$arg_name.${arg_name}_len"); } elsif ($args_member =~ m/^(?:unsigned )?int (\S+)<\S+>;/) { if (! @args_list) { push(@args_list, "conn"); @@ -998,37 +1010,63 @@ elsif ($opt_k) { } elsif ($args_member =~ m/^remote_string (\S+);/) { push(@args_list, "const char *$1"); push(@setters_list, "args.$1 = $1 ? (char **)&$1 : NULL;"); - } elsif ($args_member =~ m/^remote_nonnull_string (\S+)<(\S+)>;/) { - push(@args_list, "const char **$1"); - push(@args_list, "unsigned int ${1}len"); - push(@setters_list, "args.$1.${1}_val = (char **)$1;"); - push(@setters_list, "args.$1.${1}_len = ${1}len;"); - push(@args_check_list, { name => "\"$1\"", arg => "${1}len", limit => $2 }); + } elsif ($args_member =~ m/^remote_nonnull_string (\S+)<(\S+)>;(.*)$/) { + my $type_name = "const char **"; + my $arg_name = $1; + my $limit = $2; + my $annotation = $3; + + if ($annotation ne "") { + if ($annotation =~ m/\s*\/\*\s*$(.*)$\s*\*\//) { + $type_name = $1; + } else { + die "malformed cast annotation for argument: $args_member"; + } + } + + push(@args_list, "$type_name$arg_name"); + push(@args_list, "unsigned int ${arg_name}len"); + push(@setters_list, "args.$arg_name.${arg_name}_val = (char **)$arg_name;"); + push(@setters_list, "args.$arg_name.${arg_name}_len = ${arg_name}len;"); + push(@args_check_list, { name => "\"$arg_name\"", arg => "${arg_name}len", limit => $2 }); } elsif ($args_member =~ m/^remote_nonnull_string (\S+);/) { push(@args_list, "const char *$1"); push(@setters_list, "args.$1 = (char *)$1;"); - } elsif ($args_member =~ m/^(remote_string|opaque) (\S+)<(\S+)>;/) { - my $type_name = $1; - my $arg_name = $2; - my $limit = $3; + } elsif ($args_member =~ m/^opaque (\S+)<(\S+)>;(.*)$/) { + my $type_name = "const char *"; + my $arg_name = $1; + my $limit = $2; + my $annotation = $3; + + if ($annotation ne "") { + if ($annotation =~ m/\s*\/\*\s*$(.*)$\s*\*\//) { + $type_name = $1; + } else { + die "malformed cast annotation for argument: $args_member"; + } + } + + push(@args_list, "$type_name$arg_name"); if ($call->{ProcName} eq "SecretSetValue") { - push(@args_list, "const unsigned char *$arg_name"); + # SPECIAL: virSecretSetValue uses size_t instead of int push(@args_list, "size_t ${arg_name}len"); - } elsif ($call->{ProcName} eq "DomainPinVcpu") { - push(@args_list, "unsigned char *$arg_name"); - push(@args_list, "int ${arg_name}len"); - } elsif ($call->{ProcName} eq "DomainPinVcpuFlags") { - push(@args_list, "unsigned char *$arg_name"); - push(@args_list, "int ${arg_name}len"); } else { - push(@args_list, "const char *$arg_name"); push(@args_list, "int ${arg_name}len"); } push(@setters_list, "args.$arg_name.${arg_name}_val = (char *)$arg_name;"); push(@setters_list, "args.$arg_name.${arg_name}_len = ${arg_name}len;"); push(@args_check_list, { name => "\"$arg_name\"", arg => "${arg_name}len", limit => $limit }); + } elsif ($args_member =~ m/^remote_string (\S+)<(\S+)>;/) { + my $arg_name = $1; + my $limit = $2; + + push(@args_list, "const char *$arg_name"); + push(@args_list, "int ${arg_name}len"); + push(@setters_list, "args.$arg_name.${arg_name}_val = (char *)$arg_name;"); + push(@setters_list, "args.$arg_name.${arg_name}_len = ${arg_name}len;"); + push(@args_check_list, { name => "\"$arg_name\"", arg => "${arg_name}len", limit => $limit }); } elsif ($args_member =~ m/^((?:unsigned )?int) (\S+)<(\S+)>;/) { my $type_name = $1; my $arg_name = $2; diff --git a/src/remote/remote_protocol.x b/src/remote/remote_protocol.x index 0aa3dfb..a77fd81 100644 --- a/src/remote/remote_protocol.x +++ b/src/remote/remote_protocol.x @@ -357,7 +357,10 @@ struct remote_node_get_memory_stats { * If the 'remote_CALL_ret' maps to a struct in the public API then it is * also filled via call-by-reference and must be annotated with a * insert@<offset> comment to indicate the offset in the parameter list of - * the function to be called. */ + * the function to be called. + * + * Dynamic opaque and remote_nonnull_string arrays can be annotated with an + * optional typecast */ struct remote_open_args { /* NB. "name" might be NULL although in practice you can't @@ -890,13 +893,13 @@ struct remote_domain_get_vcpus_flags_ret { struct remote_domain_pin_vcpu_args { remote_nonnull_domain dom; unsigned int vcpu; - opaque cpumap<REMOTE_CPUMAP_MAX>; + opaque cpumap<REMOTE_CPUMAP_MAX>; /* (unsigned char *) */ }; struct remote_domain_pin_vcpu_flags_args { remote_nonnull_domain dom; unsigned int vcpu; - opaque cpumap<REMOTE_CPUMAP_MAX>; + opaque cpumap<REMOTE_CPUMAP_MAX>; /* (unsigned char *) */ unsigned int flags; }; @@ -1701,7 +1704,7 @@ struct remote_secret_get_xml_desc_ret { struct remote_secret_set_value_args { remote_nonnull_secret secret; - opaque value<REMOTE_SECRET_VALUE_MAX>; + opaque value<REMOTE_SECRET_VALUE_MAX>; /* (const unsigned char *) */ unsigned int flags; }; @@ -1819,7 +1822,7 @@ struct remote_cpu_compare_ret { struct remote_cpu_baseline_args { - remote_nonnull_string xmlCPUs<REMOTE_CPU_BASELINE_MAX>; + remote_nonnull_string xmlCPUs<REMOTE_CPU_BASELINE_MAX>; /* (const char **) */ unsigned int flags; }; @@ -2182,7 +2185,7 @@ enum remote_procedure { REMOTE_PROC_DOMAIN_LOOKUP_BY_NAME = 23, /* autogen autogen */ REMOTE_PROC_DOMAIN_LOOKUP_BY_UUID = 24, /* autogen autogen */ REMOTE_PROC_NUM_OF_DEFINED_DOMAINS = 25, /* autogen autogen */ - REMOTE_PROC_DOMAIN_PIN_VCPU = 26, /* skipgen autogen */ + REMOTE_PROC_DOMAIN_PIN_VCPU = 26, /* autogen autogen */ REMOTE_PROC_DOMAIN_REBOOT = 27, /* autogen autogen */ REMOTE_PROC_DOMAIN_RESUME = 28, /* autogen autogen */ REMOTE_PROC_DOMAIN_SET_AUTOSTART = 29, /* autogen autogen */ @@ -2401,7 +2404,7 @@ enum remote_procedure { REMOTE_PROC_INTERFACE_CHANGE_ROLLBACK = 222, /* autogen autogen */ REMOTE_PROC_DOMAIN_GET_SCHEDULER_PARAMETERS_FLAGS = 223, /* skipgen autogen */ REMOTE_PROC_DOMAIN_EVENT_CONTROL_ERROR = 224, /* skipgen skipgen */ - REMOTE_PROC_DOMAIN_PIN_VCPU_FLAGS = 225, /* skipgen autogen */ + REMOTE_PROC_DOMAIN_PIN_VCPU_FLAGS = 225, /* autogen autogen */ REMOTE_PROC_DOMAIN_SEND_KEY = 226, /* autogen autogen */ REMOTE_PROC_NODE_GET_CPU_STATS = 227, /* skipgen skipgen */ REMOTE_PROC_NODE_GET_MEMORY_STATS = 228, /* skipgen skipgen */ -- 1.7.0.4

2 2