- Devel - libvirt List Archives

[libvirt] [PATCH 0/7] Update hotplug code to use device_add
by Daniel P. Berrange 27 Jan '10

27 Jan '10

I previously changed the command line syntax to use -device for starting guests. The key reason for this was to allow unique names and PCI addresses to be assigned to all devices. I forgot todo the same for hotplug :-( This series is not yet complete, but it goes some of the way towards fixing the hotplug code to use 'device_add' for creating devices. The main remaining problem is that I'm nt assigning aliases to devices, to they all get the name of '(null)'. Obviously I'll fix that before submitting again... Daniel

1 7

[libvirt] [PATCH] support XEN_SYSCTL_INTERFACE_VERSION
by Jim Fehlig 27 Jan '10

27 Jan '10

xen-unstable c/s 20762 bumped XEN_SYSCTL_INTERFACE_VERSION to 7. I don't see how the interface change affects libvirt, other than failing xenHypervisorInit() since version 7 is not tried. The attached patch accommodates the upcoming Xen 4.0 release by checking for XEN_SYSCTL_INTERFACE_VERSION 7. If found, it sets XEN_DOMCTL_INTERFACE_VERSION to 6, which is also new to Xen 4.0. Regards, Jim

2 2

[libvirt] [PATCH 0/1] Clean up udev init
by David Allan 27 Jan '10

27 Jan '10

Matthias' patch made me realize that I didn't write the udev init code as cleanly as I'd like, and the error handling was starting to get a little messy. The attached patch cleans it up. There should be no functional change. Dave

3 3

[libvirt] [PATCH] Add missing sata controller type to domain.rng
by Matthew Booth 27 Jan '10

27 Jan '10

* docs/schemas/domain.rng: Add sata controller type --- docs/schemas/domain.rng | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/docs/schemas/domain.rng b/docs/schemas/domain.rng index 49b57eb..827ff6f 100644 --- a/docs/schemas/domain.rng +++ b/docs/schemas/domain.rng @@ -529,6 +529,7 @@ <value>fdc</value> <value>ide</value> <value>scsi</value> + <value>sata</value> </choice> </attribute> </optional> -- 1.6.6

2 1

[libvirt] [PATCH] portability to non-glibc: don't use realpath(..., NULL)
by Jim Meyering 27 Jan '10

27 Jan '10

Using realpath like that is not portable, and providing a buffer instead of NULL is wasteful and harder to code properly. Instead, use gnulib's canonicalize_file_name, which does the same job portably: >From 4afea6b59e2be6c28b45ef59a6a2f896eed44dba Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 26 Jan 2010 17:13:45 +0100 Subject: [PATCH] portability to non-glibc: don't use realpath(..., NULL) it causes a NULL-dereference on some systems like Solaris 10. * src/node_device/node_device_linux_sysfs.c. Include <stdlib.h>. (get_sriov_function): Use canonicalize_filen_name, not realpath. * bootstrap (modules): Add canonicalize-lgpl. --- bootstrap | 1 + src/node_device/node_device_linux_sysfs.c | 6 ++++-- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/bootstrap b/bootstrap index aec5d05..cc3c6ef 100755 --- a/bootstrap +++ b/bootstrap @@ -68,6 +68,7 @@ modules=' areadlink base64 c-ctype +canonicalize-lgpl close connect getaddrinfo diff --git a/src/node_device/node_device_linux_sysfs.c b/src/node_device/node_device_linux_sysfs.c index 33e658d..c1fce5d 100644 --- a/src/node_device/node_device_linux_sysfs.c +++ b/src/node_device/node_device_linux_sysfs.c @@ -24,6 +24,7 @@ #include <fcntl.h> #include <sys/stat.h> +#include <stdlib.h> #include "node_device_driver.h" #include "node_device_hal.h" @@ -242,7 +243,8 @@ out: static int get_sriov_function(const char *device_link, struct pci_config_address **bdf) { - char *device_path = NULL, *config_address = NULL; + char *config_address = NULL; + char *device_path = NULL; char errbuf[64]; int ret = SRIOV_ERROR; @@ -259,7 +261,7 @@ static int get_sriov_function(const char *device_link, } - device_path = realpath(device_link, device_path); + device_path = canonicalize_file_name (device_link); if (device_path == NULL) { memset(errbuf, '\0', sizeof(errbuf)); VIR_ERROR("Failed to resolve device link '%s': '%s'", device_link, -- 1.7.0.rc0.140.gfbe7

2 2

[libvirt] [RFC] [Patch] Support for Linux macvtap device
by Stefan Berger 27 Jan '10

27 Jan '10

Hello! The attached patch provides support for the Linux macvtap device for Qemu by passing a file descriptor to Qemu command line similar to how it is done with a regular tap device. I have modified the network XML code to understand a definition as the following one here: <network> <name>vepanet</name> <uuid>4ebd5168-6321-4757-8397-f6e83484f402</uuid> <extbridge mode='vepa' dev='eth0'/> </network> This XML indicates the device that links to the external bridge, here eth0, and the mode this device is supposed to be set into, here 'vepa'. If 'extbridge' is found in the XML, all other XML elements that have been used so far, i.e., bridge, ip, dhcp, etc., are ignored. The above network description can then be referenced from the virtual machine definition using the typical interface description of type 'network': <interface type='network'> <source network='vepanet'/> <model type='virtio'/> </interface> The above network XML holds the necessary parameters to open a macvtap device. Using command line, one would issue the following command to achieve the same what now libvirt does internally, using a patched version of the 'ip' command: ip link add link eth0 <optional name of if> type macvtap mode vepa This then creates the network interface, i.e., macvtap12, along with entries in /sys/class/net/macvtap12/ where the content of the ifindex file returns the integer necessary to build the /dev/tap%d device to open the file descriptor of and pass to qemu via command line argument. Some details: In libvirt I am first searching for an unused interface name following the template macvtap%d unless the user has provided a name explicitly or the previous macvtap%d is now taken by another VM. Once that has succeeded, I follow the path through the filesystem to open the corresponding /dev/tap%d. Unlike the regular tap device, where the network interface disappears once the corresponding file descriptor is closed, the macvtap device needs explicit tear-down. So, when a VM terminates, I am deleting all macvtap type interface with the MAC address as the interface of the terminating VM. Some further background on this device was recently given here: https://www.redhat.com/archives/libvir-list/2010-January/msg00721.html The macvtap patch for Linux was posted here: http://lkml.org/lkml/2009/12/3/239 I'd be curious about comments on the code. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

2 3

[libvirt] [PATCH] qemu_monitor_json.c: avoid an unconditional leak
by Jim Meyering 27 Jan '10

27 Jan '10

Calling virJSONValueFree does not free the memory allocated via this: cmd = qemuMonitorJSONMakeCommand("drive_add", "s:pci_addr", dev, "s:opts", drivestr, NULL); so I added the VIR_FREE: >From 1d272e070b66fc16e2fcc23002879ee268a58a5f Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 27 Jan 2010 09:58:12 +0100 Subject: [PATCH] qemu_monitor_json.c: avoid an unconditional leak * src/qemu/qemu_monitor_json.c (qemuMonitorJSONAttachDrive): Don't leak the buffer behind a virJSONValuePtr. --- src/qemu/qemu_monitor_json.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 8e88c7e..0f970bb 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -1773,6 +1773,7 @@ int qemuMonitorJSONAttachDrive(qemuMonitorPtr mon, ret = -1; virJSONValueFree(cmd); + VIR_FREE(cmd); virJSONValueFree(reply); return ret; } -- 1.7.0.rc0.140.gfbe7

1 0

[libvirt] [PATCH] pci.c: correct erroneous expression
by Jim Meyering 27 Jan '10

27 Jan '10

An expression like this, can never be true, since the RHS is nonzero: ... !pciRead16(dev, PCI_CLASS_DEVICE) == PCI_CLASS_BRIDGE_PCI) This appears to be what was intended: pciRead16(dev, PCI_CLASS_DEVICE) != PCI_CLASS_BRIDGE_PCI) >From a86de65c1da4f2cc9408465f6bc9ded481def47d Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Tue, 26 Jan 2010 20:36:58 +0100 Subject: [PATCH] pci.c: correct erroneous expression * src/util/pci.c (pciDeviceDownstreamLacksACS): Fix a typo that rendered a subexpression always false. --- src/util/pci.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/util/pci.c b/src/util/pci.c index 09535bd..0274806 100644 --- a/src/util/pci.c +++ b/src/util/pci.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2009 Red Hat, Inc. + * Copyright (C) 2009-2010 Red Hat, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -1266,7 +1266,7 @@ pciDeviceDownstreamLacksACS(virConnectPtr conn, return -1; pos = dev->pcie_cap_pos; - if (!pos || !pciRead16(dev, PCI_CLASS_DEVICE) == PCI_CLASS_BRIDGE_PCI) + if (!pos || pciRead16(dev, PCI_CLASS_DEVICE) != PCI_CLASS_BRIDGE_PCI) return 0; flags = pciRead16(dev, pos + PCI_EXP_FLAGS); -- 1.7.0.rc0.140.gfbe7

2 2

[libvirt] [PATCH] Don't call disabled timer callbacks in event-test.c
by Matthias Bolte 26 Jan '10

26 Jan '10

This fixes a segfault in the remote driver that occurs for example when the event-test is run inside a domain-0 and libvirtd is also running. --- examples/domain-events/events-c/event-test.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/examples/domain-events/events-c/event-test.c b/examples/domain-events/events-c/event-test.c index b506687..b2eb1d5 100644 --- a/examples/domain-events/events-c/event-test.c +++ b/examples/domain-events/events-c/event-test.c @@ -335,9 +335,10 @@ int main(int argc, char **argv) sts = poll(&pfd, 1, TIMEOUT_MS); - /* We are assuming timeout of 0 here - so execute every time */ - if(t_cb && t_active) + /* if t_timeout < 0 then t_cb must not be called */ + if (t_cb && t_active && t_timeout >= 0) { t_cb(t_timeout,t_opaque); + } if (sts == 0) { /* DEBUG0("Poll timeout"); */ -- 1.6.3.3

2 2

[libvirt] [PATCH] esx: Output error details from libcurl
by Matthias Bolte 26 Jan '10

26 Jan '10

--- src/esx/esx_vi.c | 14 ++++++++------ src/esx/esx_vi.h | 1 + 2 files changed, 9 insertions(+), 6 deletions(-) diff --git a/src/esx/esx_vi.c b/src/esx/esx_vi.c index 0fe9535..bc94326 100644 --- a/src/esx/esx_vi.c +++ b/src/esx/esx_vi.c @@ -251,8 +251,8 @@ esxVI_CURL_Perform(esxVI_Context *ctx, const char *url) if (errorCode != CURLE_OK) { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, - "curl_easy_perform() returned an error: %s (%d)", - curl_easy_strerror(errorCode), errorCode); + "curl_easy_perform() returned an error: %s (%d) : %s", + curl_easy_strerror(errorCode), errorCode, ctx->curl_error); return -1; } @@ -262,8 +262,8 @@ esxVI_CURL_Perform(esxVI_Context *ctx, const char *url) if (errorCode != CURLE_OK) { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, "curl_easy_getinfo(CURLINFO_RESPONSE_CODE) returned an " - "error: %s (%d)", curl_easy_strerror(errorCode), - errorCode); + "error: %s (%d) : %s", curl_easy_strerror(errorCode), + errorCode, ctx->curl_error); return -1; } @@ -282,8 +282,8 @@ esxVI_CURL_Perform(esxVI_Context *ctx, const char *url) if (errorCode != CURLE_OK) { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, "curl_easy_getinfo(CURLINFO_REDIRECT_URL) returned " - "an error: %s (%d)", curl_easy_strerror(errorCode), - errorCode); + "an error: %s (%d) : %s", curl_easy_strerror(errorCode), + errorCode, ctx->curl_error); } else { ESX_VI_ERROR(VIR_ERR_INTERNAL_ERROR, "The server redirects from '%s' to '%s'", url, @@ -360,6 +360,8 @@ esxVI_Context_Connect(esxVI_Context *ctx, const char *url, esxVI_CURL_ReadString); curl_easy_setopt(ctx->curl_handle, CURLOPT_WRITEFUNCTION, esxVI_CURL_WriteBuffer); + curl_easy_setopt(ctx->curl_handle, CURLOPT_ERRORBUFFER, + ctx->curl_error); #if ESX_VI__CURL__ENABLE_DEBUG_OUTPUT curl_easy_setopt(ctx->curl_handle, CURLOPT_DEBUGFUNCTION, esxVI_CURL_Debug); curl_easy_setopt(ctx->curl_handle, CURLOPT_VERBOSE, 1); diff --git a/src/esx/esx_vi.h b/src/esx/esx_vi.h index 608dcbd..4e8b675 100644 --- a/src/esx/esx_vi.h +++ b/src/esx/esx_vi.h @@ -77,6 +77,7 @@ struct _esxVI_Context { char *ipAddress; CURL *curl_handle; struct curl_slist *curl_headers; + char curl_error[CURL_ERROR_SIZE]; virMutex curl_lock; char *username; char *password; -- 1.6.3.3

2 2

[libvirt] [PATCH] Add missing function parameter documentation
by Matthias Bolte 26 Jan '10

26 Jan '10

--- src/libvirt.c | 5 ++++- src/util/virterror.c | 1 + 2 files changed, 5 insertions(+), 1 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index d973907..32f9fc4 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -10939,7 +10939,10 @@ error: /** * virConnectCompareCPU: * @conn: virConnect connection - * @xml: XML describing the CPU to compare with host CPU + * @xmlDesc: XML describing the CPU to compare with host CPU + * @flags: currently unused, pass 0 + * + * Compares the given CPU description with the host CPU * * Returns comparison result according to enum virCPUCompareResult */ diff --git a/src/util/virterror.c b/src/util/virterror.c index e2128b9..d72b307 100644 --- a/src/util/virterror.c +++ b/src/util/virterror.c @@ -291,6 +291,7 @@ virGetLastError(void) /** * virSetError: + * @newerr: previously saved error object * * Set the current error from a previously saved error object * -- 1.6.3.3

2 2

[libvirt] Potential segfault in udev driver
by Matthias Bolte 26 Jan '10

26 Jan '10

udevDeviceMonitorStartup registers udevEventHandleCallback as event handle, but doesn't store the returned watch id to remove it later on. Also it's not clear to me whether the event handle should be register for the whole lifetime of the udev driver instance or just for the udevEnumerateDevices call. If for example the call to udevSetupSystemDev [1] fails udevDeviceMonitorShutdown is called to cleanup, but udevEventHandleCallback is still registered and may be called when driverState is NULL again, resulting in a segfault in udevEventHandleCallback. So to solve this the udevEventHandleCallback event handle must be removed at the appropriate place. Maximilian Wilhelm found this problem. [1] 22:42:20.960: error : udevSetupSystemDev:1460 : Failed to get udev device for syspath '/sys/devices/virtual/dmi/id' or '/sys/class/dmi/id' Matthias

3 7

[libvirt] [PATCH] Remove undefined symbols from libvirt_private.syms
by Matthias Bolte 26 Jan '10

26 Jan '10

--- src/libvirt_private.syms | 2 -- 1 files changed, 0 insertions(+), 2 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index e42c090..d56fb7d 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -177,12 +177,10 @@ virDomainObjListInit; virDomainObjListDeinit; virDomainObjRef; virDomainObjUnref; -virDomainDeviceAddressEqual; virDomainDevicePCIAddressEqual; virDomainDeviceAddressIsValid; virDomainDevicePCIAddressIsValid; virDomainDeviceInfoIsSet; -virDomainDeviceAddressClear; virDomainControllerTypeToString; virDomainControllerDefFree; virDomainDeviceAddressTypeToString; -- 1.6.3.3

2 2

[libvirt] [PATCH] udev: Set the state driver name
by Matthias Bolte 26 Jan '10

26 Jan '10

--- src/node_device/node_device_udev.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c index a625d76..dcd889d 100644 --- a/src/node_device/node_device_udev.c +++ b/src/node_device/node_device_udev.c @@ -1676,6 +1676,7 @@ static virDeviceMonitor udevDeviceMonitor = { }; static virStateDriver udevStateDriver = { + .name = "udev", .initialize = udevDeviceMonitorStartup, .cleanup = udevDeviceMonitorShutdown, .reload = udevDeviceMonitorReload, -- 1.6.3.3

2 2

[libvirt] [Request]Backport live migration flags from 0.7.3 to 0.6.3?
by edison 26 Jan '10

26 Jan '10

Hi, libvirt 0.7.3 adds the VIR_MIGRATE_PERSIST_DEST, VIR_MIGRATE_UNDEFINE_SOURCE flags in migration, which are very useful in our product. Unfortunately, we use libvirt 0.6.3.20, which is the default in RHEL 5.4. No such advanced migration feature in this version. I backported it for 0.6.3.20, in the attached patch. Do you think it's possible to be merged into 0.6.3.20?:)

2 1

[libvirt] [PATCH] usbFindBusByVendor: don't leak a DIR buffer and FD
by Jim Meyering 26 Jan '10

26 Jan '10

I know better than to say this is so simple/obvious that it doesn't need review, but I'm tempted... >From 02c7dfa830544bbafd62867fc70b3aba7cc07385 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 25 Jan 2010 16:36:07 +0100 Subject: [PATCH] usbFindBusByVendor: don't leak a DIR buffer and FD * src/util/hostusb.c (usbFindBusByVendor): Don't leak a DIR buffer and file descriptor. --- src/util/hostusb.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/util/hostusb.c b/src/util/hostusb.c index 8fbb486..9a37103 100644 --- a/src/util/hostusb.c +++ b/src/util/hostusb.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2009 Red Hat, Inc. + * Copyright (C) 2009-2010 Red Hat, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -151,6 +151,7 @@ static int usbFindBusByVendor(virConnectPtr conn, ret = 0; error: + closedir (dir); return ret; } -- 1.7.0.rc0.127.gab8271

5 9

[libvirt] [PATCH] Eliminate use of large buffer on stack in virFileMakePath.
by Laine Stump 26 Jan '10

26 Jan '10

virFileMakePath is a recursive function that was creates a buffer PATH_MAX bytes long for each recursion (one recursion for each element in the path). This changes it to have no buffers on the stack, and to allocate just one buffer total, no matter how many elements are in the path. Because the modified algorithm requires a char* to be passed in rather than const char *, it is now 2 functions - a toplevel API function that remains identical in function, and a 2nd helper function called for the recursions, which 1) doesn't allocate anything, and 2) takes a char* arg, so it can modify the contents. src/util/util.c: rewrite virFileMakePath --- src/util/util.c | 58 +++++++++++++++++++++++++++++++++++++++++++----------- 1 files changed, 46 insertions(+), 12 deletions(-) diff --git a/src/util/util.c b/src/util/util.c index 08d74a3..0ce5026 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -1443,34 +1443,68 @@ int virDirCreate(const char *path, mode_t mode, } #endif -int virFileMakePath(const char *path) -{ +static int virFileMakePathHelper(char *path) { struct stat st; - char parent[PATH_MAX]; - char *p; + char *p = NULL; int err; if (stat(path, &st) >= 0) return 0; - if (virStrcpyStatic(parent, path) == NULL) - return EINVAL; - - if (!(p = strrchr(parent, '/'))) + if ((p = strrchr(path, '/')) == NULL) return EINVAL; - if (p != parent) { + if (p != path) { *p = '\0'; - if ((err = virFileMakePath(parent))) + err = virFileMakePathHelper(path); + *p = '/'; + if (err != 0) return err; } - if (mkdir(path, 0777) < 0 && errno != EEXIST) + if (mkdir(path, 0777) < 0 && errno != EEXIST) { return errno; - + } return 0; } +int virFileMakePath(const char *path) +{ + struct stat st; + char *parent = NULL; + char *p; + int err = 0; + + if (stat(path, &st) >= 0) + goto cleanup; + + if ((parent = strdup(path)) == NULL) { + err = ENOMEM; + goto cleanup; + } + + if ((p = strrchr(parent, '/')) == NULL) { + err = EINVAL; + goto cleanup; + } + + if (p != parent) { + *p = '\0'; + if ((err = virFileMakePathHelper(parent)) != 0) { + goto cleanup; + } + } + + if (mkdir(path, 0777) < 0 && errno != EEXIST) { + err = errno; + goto cleanup; + } + +cleanup: + VIR_FREE(parent); + return err; +} + /* Build up a fully qualfiied path for a config file to be * associated with a persistent guest or network */ int virFileBuildPath(const char *dir, -- 1.6.6

2 2

[libvirt] [PATCH] Fix two instances of misspelled 'pseudo'
by Chris Lalancette 26 Jan '10

26 Jan '10

Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- docs/sitemap.html.in | 2 +- src/xen/xs_internal.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/sitemap.html.in b/docs/sitemap.html.in index 424897f..76d8faa 100644 --- a/docs/sitemap.html.in +++ b/docs/sitemap.html.in @@ -134,7 +134,7 @@ </li> <li> <a href="drvtest.html">Test</a> - <span>Psuedo-driver simulating APIs in memory for test suites</span> + <span>Pseudo-driver simulating APIs in memory for test suites</span> </li> <li> <a href="drvremote.html">Remote</a> diff --git a/src/xen/xs_internal.c b/src/xen/xs_internal.c index 8a64d4e..1ca3da0 100644 --- a/src/xen/xs_internal.c +++ b/src/xen/xs_internal.c @@ -853,7 +853,7 @@ int xenStoreDomainGetVNCPort(virConnectPtr conn, int domid) { * @conn: the hypervisor connection * @domid: id of the domain * - * Return the path to the psuedo TTY on which the guest domain's + * Return the path to the pseudo TTY on which the guest domain's * serial console is attached. * * Returns the path to the serial console. It is the callers -- 1.6.6

2 1

[libvirt] Live Migration with non-shared storage for kvm
by Kenneth Nagin 26 Jan '10

26 Jan '10

On January 5 I asked for comments on adding libvirt support for live migration with non-shared storage on kvm. I did not get any comments so I am repeating my request: Support for live migration between hosts that do not share storage was added to qemu-kvm release 0.12.1. It supports two flags: -b migration without shared storage with full disk copy -i migration without shared storage with incremental copy (same base image shared between source and destination). I suggest adding these flags to virDomainMigrate. If I'm not mistaken qemuMonitorTextMigrate is the function that actually invokes the kvm monitor. Thus, it would be necessary to pass the flags to qemuMonitorTextMigrate.. But qemuMonitorTextMigrate does not have a flag input parameter. I think the least disruptive way to support the new flags is use the existing "background" parameter to pass the flags. Of course this would require some changes to the upstream functions that are invoked for migration. What do you think? Kenneth Nagin

2 1

[libvirt] [PATCH 0/3] esx: Stop passing around virConnectPtr for error reporting
by Matthias Bolte 26 Jan '10

26 Jan '10

This is actually a single huge patch, but I had to split it in 3 parts, otherwise it's too big for the list. The parts won't compile on their own. The patch removes the virConnectPtr argument from internal functions of the ESX driver, no other functional change is included. Matthias

2 7

[libvirt] [PATCH] Fix a crash when restarting libvirtd.
by Chris Lalancette 25 Jan '10

25 Jan '10

If you shutdown libvirtd while a domain with PCI devices is running, then try to restart libvirtd, libvirtd will crash. This happens because qemuUpdateActivePciHostdevs() is calling pciDeviceListSteal() with a dev of 0x0 (NULL), and then trying to dereference it. This patch fixes it up so that qemuUpdateActivePciHostdevs() steals the devices after first Get()'ting them, avoiding the crash. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu/qemu_driver.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 55550ef..bbdbe33 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2147,6 +2147,7 @@ qemuUpdateActivePciHostdevs(struct qemud_driver *driver, virDomainDefPtr def) { pciDeviceList *pcidevs; + int i; int ret = -1; if (!def->nhostdevs) @@ -2155,8 +2156,9 @@ qemuUpdateActivePciHostdevs(struct qemud_driver *driver, if (!(pcidevs = qemuGetPciHostDeviceList(NULL, def))) return -1; - while (pciDeviceListCount(pcidevs) > 0) { - pciDevice *dev = pciDeviceListSteal(NULL, pcidevs, 0); + for (i = 0; i < pciDeviceListCount(pcidevs); i++) { + pciDevice *dev = pciDeviceListGet(pcidevs, i); + pciDeviceListSteal(NULL, pcidevs, dev); if (pciDeviceListAdd(NULL, driver->activePciHostdevs, dev) < 0) { -- 1.6.6

2 2

[libvirt] [PATCH] Revert "Fix libvirtd restart for domains with PCI passthrough devices"
by Chris Lalancette 25 Jan '10

25 Jan '10

This reverts commit cdc42d0a4865199a941d330dbb6ca1ef426323ae. As DanB pointed out, this patch is actually wrong. The real bug that was causing me to see this problem is a bug introduced in a RHEL-5 libvirt snapshot, and I'm going to fix the real bug there. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/conf/domain_conf.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 6d0e2dc..e548d1d 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -2876,7 +2876,7 @@ static int virDomainHostdevSubsysPciDefParseXML(virConnectPtr conn, const xmlNodePtr node, virDomainHostdevDefPtr def, - int flags ATTRIBUTE_UNUSED) { + int flags) { int ret = -1; xmlNodePtr cur; @@ -2890,7 +2890,8 @@ virDomainHostdevSubsysPciDefParseXML(virConnectPtr conn, if (virDomainDevicePCIAddressParseXML(conn, cur, addr) < 0) goto out; - } else if (xmlStrEqual(cur->name, BAD_CAST "state")) { + } else if ((flags & VIR_DOMAIN_XML_INTERNAL_STATUS) && + xmlStrEqual(cur->name, BAD_CAST "state")) { /* Legacy back-compat. Don't add any more attributes here */ char *devaddr = virXMLPropString(cur, "devaddr"); if (devaddr && -- 1.6.6

2 2

[libvirt] [PATCH] qemuMonitorTextAttachDrive: avoid two leaks
by Jim Meyering 25 Jan '10

25 Jan '10

Coverity reported the first leak: return-(-1)-would-leak-"safe_str". I spotted the 2nd. >From 6326ab7b52f3550f24015c27407c1d3c2a1fd64d Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 25 Jan 2010 20:19:01 +0100 Subject: [PATCH] qemuMonitorTextAttachDrive: avoid two leaks * src/qemu/qemu_monitor_text.c (qemuMonitorTextAttachDrive): Most other failures in this function would "goto cleanup", but one mistakenly returned directly, skipping the cleanup and resulting in a leak. In addition, iterating the "try_command" loop would clobber, and thus leak, the "cmd" allocated on the first iteration, so be careful to free it in addition to "reply" beforehand. --- src/qemu/qemu_monitor_text.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index ce5349b..380bcdc 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -1881,13 +1881,13 @@ try_command: if (qemudParseDriveAddReply(reply, driveAddr) < 0) { if (!tryOldSyntax && strstr(reply, "invalid char in expression")) { VIR_FREE(reply); + VIR_FREE(cmd); tryOldSyntax = 1; goto try_command; } qemudReportError (NULL, NULL, NULL, VIR_ERR_OPERATION_FAILED, _("adding %s disk failed: %s"), drivestr, reply); - VIR_FREE(reply); - return -1; + goto cleanup; } ret = 0; -- 1.7.0.rc0.127.gab8271

2 2

[libvirt] [PATCH] qemu: Search binaries in PATH instead of hardcoding /usr/bin
by Matthias Bolte 25 Jan '10

25 Jan '10

--- src/qemu/qemu_conf.c | 137 ++++++++++++++++++++++++++++++-------------------- 1 files changed, 82 insertions(+), 55 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 92a9348..6141ec6 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -379,20 +379,20 @@ static const struct qemu_feature_flags const arch_info_x86_64_flags [] = { /* The archicture tables for supported QEMU archs */ static const struct qemu_arch_info const arch_info_hvm[] = { - { "i686", 32, NULL, "/usr/bin/qemu", - "/usr/bin/qemu-system-x86_64", arch_info_i686_flags, 4 }, - { "x86_64", 64, NULL, "/usr/bin/qemu-system-x86_64", + { "i686", 32, NULL, "qemu", + "qemu-system-x86_64", arch_info_i686_flags, 4 }, + { "x86_64", 64, NULL, "qemu-system-x86_64", NULL, arch_info_x86_64_flags, 2 }, - { "arm", 32, NULL, "/usr/bin/qemu-system-arm", NULL, NULL, 0 }, - { "mips", 32, NULL, "/usr/bin/qemu-system-mips", NULL, NULL, 0 }, - { "mipsel", 32, NULL, "/usr/bin/qemu-system-mipsel", NULL, NULL, 0 }, - { "sparc", 32, NULL, "/usr/bin/qemu-system-sparc", NULL, NULL, 0 }, - { "ppc", 32, NULL, "/usr/bin/qemu-system-ppc", NULL, NULL, 0 }, + { "arm", 32, NULL, "qemu-system-arm", NULL, NULL, 0 }, + { "mips", 32, NULL, "qemu-system-mips", NULL, NULL, 0 }, + { "mipsel", 32, NULL, "qemu-system-mipsel", NULL, NULL, 0 }, + { "sparc", 32, NULL, "qemu-system-sparc", NULL, NULL, 0 }, + { "ppc", 32, NULL, "qemu-system-ppc", NULL, NULL, 0 }, }; static const struct qemu_arch_info const arch_info_xen[] = { - { "i686", 32, "xenner", "/usr/bin/xenner", NULL, arch_info_i686_flags, 4 }, - { "x86_64", 64, "xenner", "/usr/bin/xenner", NULL, arch_info_x86_64_flags, 2 }, + { "i686", 32, "xenner", "xenner", NULL, arch_info_i686_flags, 4 }, + { "x86_64", 64, "xenner", "xenner", NULL, arch_info_x86_64_flags, 2 }, }; @@ -768,8 +768,8 @@ qemudCapsInitGuest(virCapsPtr caps, int i; int haskvm = 0; int haskqemu = 0; - const char *kvmbin = NULL; - const char *binary = NULL; + char *kvmbin = NULL; + char *binary = NULL; time_t binary_mtime; virCapsGuestMachinePtr *machines = NULL; int nmachines = 0; @@ -779,10 +779,15 @@ qemudCapsInitGuest(virCapsPtr caps, /* Check for existance of base emulator, or alternate base * which can be used with magic cpu choice */ - if (access(info->binary, X_OK) == 0) - binary = info->binary; - else if (info->altbinary && access(info->altbinary, X_OK) == 0) - binary = info->altbinary; + binary = virFindFileInPath(info->binary); + + if (binary == NULL || access(binary, X_OK) != 0) { + VIR_FREE(binary); + binary = virFindFileInPath(info->altbinary); + + if (binary != NULL && access(binary, X_OK) != 0) + VIR_FREE(binary); + } /* Can use acceleration for KVM/KQEMU if * - host & guest arches match @@ -792,16 +797,30 @@ qemudCapsInitGuest(virCapsPtr caps, */ if (STREQ(info->arch, hostmachine) || (STREQ(hostmachine, "x86_64") && STREQ(info->arch, "i686"))) { - const char *const kvmbins[] = { "/usr/bin/qemu-kvm", /* Fedora */ - "/usr/bin/kvm" }; /* Upstream .spec */ + if (access("/dev/kvm", F_OK) == 0) { + const char *const kvmbins[] = { "qemu-kvm", /* Fedora */ + "kvm" }; /* Upstream .spec */ + + for (i = 0; i < ARRAY_CARDINALITY(kvmbins); ++i) { + kvmbin = virFindFileInPath(kvmbins[i]); + + if (kvmbin == NULL || access(kvmbin, X_OK) != 0) { + VIR_FREE(kvmbin); + continue; + } - for (i = 0; i < ARRAY_CARDINALITY(kvmbins); ++i) { - if (access(kvmbins[i], X_OK) == 0 && - access("/dev/kvm", F_OK) == 0) { haskvm = 1; - kvmbin = kvmbins[i]; - if (!binary) - binary = kvmbin; + + if (binary == NULL) { + binary = strdup(kvmbin); + + if (binary == NULL) { + virReportOOMError(NULL); + VIR_FREE(kvmbin); + return -1; + } + } + break; } } @@ -826,21 +845,18 @@ qemudCapsInitGuest(virCapsPtr caps, virCapsGuestMachinePtr machine; if (VIR_ALLOC(machine) < 0) { - virReportOOMError(NULL); - return -1; + goto no_memory; } if (!(machine->name = strdup(info->machine))) { - virReportOOMError(NULL); VIR_FREE(machine); - return -1; + goto no_memory; } if (VIR_ALLOC_N(machines, 1) < 0) { - virReportOOMError(NULL); VIR_FREE(machine->name); VIR_FREE(machine); - return -1; + goto no_memory; } machines[0] = machine; @@ -853,7 +869,7 @@ qemudCapsInitGuest(virCapsPtr caps, old_caps, &machines, &nmachines); if (probe && qemudProbeMachineTypes(binary, &machines, &nmachines) < 0) - return -1; + goto error; } /* We register kvm as the base emulator too, since we can @@ -865,21 +881,18 @@ qemudCapsInitGuest(virCapsPtr caps, binary, NULL, nmachines, - machines)) == NULL) { - for (i = 0; i < nmachines; i++) { - VIR_FREE(machines[i]->name); - VIR_FREE(machines[i]); - } - VIR_FREE(machines); - return -1; - } + machines)) == NULL) + goto error; + + machines = NULL; + nmachines = 0; guest->arch.defaultInfo.emulator_mtime = binary_mtime; if (qemudProbeCPUModels(binary, info->arch, &ncpus, NULL) == 0 && ncpus > 0 && !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0)) - return -1; + goto error; if (hvm) { if (virCapabilitiesAddGuestDomain(guest, @@ -888,7 +901,7 @@ qemudCapsInitGuest(virCapsPtr caps, NULL, 0, NULL) == NULL) - return -1; + goto error; if (haskqemu && virCapabilitiesAddGuestDomain(guest, @@ -897,7 +910,7 @@ qemudCapsInitGuest(virCapsPtr caps, NULL, 0, NULL) == NULL) - return -1; + goto error; if (haskvm) { virCapsGuestDomainPtr dom; @@ -911,9 +924,6 @@ qemudCapsInitGuest(virCapsPtr caps, binary_mtime = 0; } - machines = NULL; - nmachines = 0; - if (!STREQ(binary, kvmbin)) { int probe = 1; if (old_caps && binary_mtime) @@ -922,7 +932,7 @@ qemudCapsInitGuest(virCapsPtr caps, old_caps, &machines, &nmachines); if (probe && qemudProbeMachineTypes(kvmbin, &machines, &nmachines) < 0) - return -1; + goto error; } if ((dom = virCapabilitiesAddGuestDomain(guest, @@ -931,14 +941,12 @@ qemudCapsInitGuest(virCapsPtr caps, NULL, nmachines, machines)) == NULL) { - for (i = 0; i < nmachines; i++) { - VIR_FREE(machines[i]->name); - VIR_FREE(machines[i]); - } - VIR_FREE(machines); - return -1; + goto error; } + machines = NULL; + nmachines = 0; + dom->info.emulator_mtime = binary_mtime; } } else { @@ -948,7 +956,7 @@ qemudCapsInitGuest(virCapsPtr caps, NULL, 0, NULL) == NULL) - return -1; + goto error; } if (info->nflags) { @@ -957,11 +965,24 @@ qemudCapsInitGuest(virCapsPtr caps, info->flags[i].name, info->flags[i].default_on, info->flags[i].toggle) == NULL) - return -1; + goto error; } } + VIR_FREE(binary); + VIR_FREE(kvmbin); + return 0; + +no_memory: + virReportOOMError(NULL); + +error: + VIR_FREE(binary); + VIR_FREE(kvmbin); + virCapabilitiesFreeMachines(machines, nmachines); + + return -1; } @@ -1011,6 +1032,7 @@ virCapsPtr qemudCapsInit(virCapsPtr old_caps) { struct utsname utsname; virCapsPtr caps; int i; + char *xenner = NULL; /* Really, this never fails - look at the man-page. */ uname (&utsname); @@ -1051,7 +1073,9 @@ virCapsPtr qemudCapsInit(virCapsPtr old_caps) { goto no_memory; /* Then possibly the Xen paravirt guests (ie Xenner */ - if (access("/usr/bin/xenner", X_OK) == 0 && + xenner = virFindFileInPath("xenner"); + + if (xenner != NULL && access(xenner, X_OK) == 0 && access("/dev/kvm", F_OK) == 0) { for (i = 0 ; i < ARRAY_CARDINALITY(arch_info_xen) ; i++) /* Allow Xen 32-on-32, 32-on-64 and 64-on-64 */ @@ -1065,12 +1089,15 @@ virCapsPtr qemudCapsInit(virCapsPtr old_caps) { } } + VIR_FREE(xenner); + /* QEMU Requires an emulator in the XML */ virCapabilitiesSetEmulatorRequired(caps); return caps; no_memory: + VIR_FREE(xenner); virCapabilitiesFree(caps); return NULL; } -- 1.6.3.3

3 6

[libvirt] [PATCH] Look in /usr/libexec for the qemu-kvm binary.
by Chris Lalancette 25 Jan '10

25 Jan '10

On RHEL-5 the qemu-kvm binary is located in /usr/libexec. To reduce confusion for people trying to run upstream libvirt on RHEL-5 machines, make the qemu driver look in /usr/libexec for the qemu-kvm binary. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu/qemu_conf.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index c227fe1..2ba8366 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -797,7 +797,8 @@ qemudCapsInitGuest(virCapsPtr caps, */ if (STREQ(info->arch, hostmachine) || (STREQ(hostmachine, "x86_64") && STREQ(info->arch, "i686"))) { - const char *const kvmbins[] = { "/usr/bin/qemu-kvm", /* Fedora */ + const char *const kvmbins[] = { "/usr/libexec/qemu-kvm", /* RHEL */ + "/usr/bin/qemu-kvm", /* Fedora */ "/usr/bin/kvm" }; /* Upstream .spec */ for (i = 0; i < ARRAY_CARDINALITY(kvmbins); ++i) { -- 1.6.6

3 4