- Devel - libvirt List Archives

[libvirt] [PATCH] libxl: use serial device for console when targetType is serial
by Jim Fehlig 23 Jun '16

23 Jun '16

When domXML contains only <console type='pty'> and no corresponding <serial>, the console is "stolen" [1] and used as the first <serial> device. When this "stolen" console is accessed from the libxl driver (in libxlConsoleCallback and libxlDomainOpenConsole), check if the targetType is VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL, and use the "stolen" device in def->serials[0] instead. Prior to this change, creating a domain with input XML containing only a <console> device and subsequently attempting to access its console with 'virsh console' would fail error: internal error: character device <null> is not using a PTY [1] See comments associated with virDomainDefAddConsoleCompat() in $LIBVIRT-SRC/src/conf/domain_conf.c: Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/libxl/libxl_domain.c | 8 ++++++-- src/libxl/libxl_driver.c | 5 ++++- 2 files changed, 10 insertions(+), 3 deletions(-) diff --git a/src/libxl/libxl_domain.c b/src/libxl/libxl_domain.c index 221af87..6bcb4d9 100644 --- a/src/libxl/libxl_domain.c +++ b/src/libxl/libxl_domain.c @@ -964,13 +964,17 @@ libxlConsoleCallback(libxl_ctx *ctx, libxl_event *ev, void *for_callback) virObjectLock(vm); for (i = 0; i < vm->def->nconsoles; i++) { virDomainChrDefPtr chr = vm->def->consoles[i]; - if (chr && chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) { + if (i == 0 && + chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL) + chr = vm->def->serials[0]; + + if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) { libxl_console_type console_type; char *console = NULL; int ret; console_type = - (chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL ? + (chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_SERIAL ? LIBXL_CONSOLE_TYPE_SERIAL : LIBXL_CONSOLE_TYPE_PV); ret = libxl_console_get_tty(ctx, ev->domid, chr->target.port, console_type, diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index c8c1f07..3189f1c 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -4443,8 +4443,11 @@ libxlDomainOpenConsole(virDomainPtr dom, priv = vm->privateData; - if (vm->def->nconsoles) + if (vm->def->nconsoles) { chr = vm->def->consoles[0]; + if (chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL) + chr = vm->def->serials[0]; + } if (!chr) { virReportError(VIR_ERR_INTERNAL_ERROR, -- 2.1.4

3 4

[libvirt] [PATCH] conf: Remove dead console compat formatting
by Cole Robinson 23 Jun '16

23 Jun '16

This code was attempting to handle some implicit <console> XML formatting for manually assembled DomainDef, since previously the console<->serial compat copying was only done at XML parse time. Nowadays it's done via virDomainDefPostParse -> virDomainDefAddConsoleCompat, which all manual DomainDef builders already call, so we can drop this workaround. --- src/conf/domain_conf.c | 14 -------------- 1 file changed, 14 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 75ad03f..6ab7450 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -23463,20 +23463,6 @@ virDomainDefFormatInternal(virDomainDefPtr def, if (virDomainChrDefFormat(buf, &console, flags) < 0) goto error; } - /* The back-compat console device stub is added when parsing the domain XML - * and handled above, this is for formatting definitions created via other - * means. - */ - if (def->os.type == VIR_DOMAIN_OSTYPE_HVM && - def->nconsoles == 0 && - def->nserials > 0) { - virDomainChrDef console; - memcpy(&console, def->serials[n], sizeof(console)); - console.deviceType = VIR_DOMAIN_CHR_DEVICE_TYPE_CONSOLE; - console.targetType = VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL; - if (virDomainChrDefFormat(buf, &console, flags) < 0) - goto error; - } for (n = 0; n < def->nchannels; n++) if (virDomainChrDefFormat(buf, def->channels[n], flags) < 0) -- 2.7.4

2 1

[libvirt] [PATCH RFC] libxl: set serial source path for console type=serial
by Joao Martins 23 Jun '16

23 Jun '16

Guests use a <console /> (and sometimes <serial /> pair) to represent the console. On the callback that is called when console is brought up (NB: before domain starts), we fill the path of the console element with the appropriate "/dev/pts/X". For PV guests it all works fine, although for HVM guests it doesn't. Consequently we end up seeing erronous behaviour on HVM guests where console path is not displayed on the XML but still can be accessed with virDomainOpenConsole after booting guest. Consumers of this XML (e.g. Openstack) also won't be able to use the console path. Finally, the path set in consoles array won't persist across daemon reboots, thus rendering admins/users with no access to console with a message such as: "error: operation failed: PTY device is not yet assigned" This is because: for HVM guests, DefFormatInternal will ignore the console element and instead write it with the content of the serial element for target type = serial which isn't touched in our libxlConsoleCallback. To address that we introduce a new helper routine libxlConsoleSetSourcePath() that sets the source path and thus also handling this case. That is by setting the source path on with serial element akin to the one indexed by console "port". This way we keep similar behaviour for PV and HVM guests. Signed-off-by: Joao Martins <joao.m.martins(a)oracle.com> --- src/libxl/libxl_domain.c | 35 +++++++++++++++++++++++++++++++---- 1 file changed, 31 insertions(+), 4 deletions(-) diff --git a/src/libxl/libxl_domain.c b/src/libxl/libxl_domain.c index 221af87..4a46143 100644 --- a/src/libxl/libxl_domain.c +++ b/src/libxl/libxl_domain.c @@ -955,6 +955,35 @@ libxlNetworkPrepareDevices(virDomainDefPtr def) return 0; } +static int +libxlConsoleSetSourcePath(virDomainDefPtr def, virDomainChrDefPtr console, + char *path) +{ + int ret = -1; + size_t i; + + if (!path || path[0] == '\0') + return ret; + + if (VIR_STRDUP(console->source.data.file.path, path) < 0) + return ret; + + if (console->targetType != VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL) + return 0; + + for (i = 0; i < def->nserials; i++) { + virDomainChrDefPtr serial = def->serials[i]; + + if (serial->target.port == console->target.port && + VIR_STRDUP(serial->source.data.file.path, path) >= 0) { + ret = 0; + break; + } + } + + return ret; +} + static void libxlConsoleCallback(libxl_ctx *ctx, libxl_event *ev, void *for_callback) { @@ -977,10 +1006,8 @@ libxlConsoleCallback(libxl_ctx *ctx, libxl_event *ev, void *for_callback) &console); if (!ret) { VIR_FREE(chr->source.data.file.path); - if (console && console[0] != '\0') { - ignore_value(VIR_STRDUP(chr->source.data.file.path, - console)); - } + if (libxlConsoleSetSourcePath(vm->def, chr, console) < 0) + VIR_WARN("Failed to set console source path"); } VIR_FREE(console); } -- 2.1.4

3 7

[libvirt] [PATCH v3] storage: Fix several issues with transient pool cleanup
by Jovanka Gulicoska 23 Jun '16

23 Jun '16

There are several cases where we do not handle transient pool destroy and cleanup correctly. For example: https://bugzilla.redhat.com/show_bug.cgi?id=1227475 Move the pool cleanup logic to a new function storagePoolSetInactive and use it consistently. --- src/storage/storage_driver.c | 45 +++++++++++++++++++++++++------------------- 1 file changed, 26 insertions(+), 19 deletions(-) diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index e2d729f..930ae5a 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -77,6 +77,23 @@ static void storageDriverUnlock(void) } static void +storagePoolSetInactive(virStoragePoolObjPtr *poolptr) +{ + virStoragePoolObjPtr pool = *poolptr; + + pool->active = false; + + if (pool->configFile == NULL) { + virStoragePoolObjRemove(&driver->pools, pool); + *poolptr = NULL; + } else if (pool->newDef) { + virStoragePoolDefFree(pool->def); + pool->def = pool->newDef; + pool->newDef = NULL; + } +} + +static void storagePoolUpdateState(virStoragePoolObjPtr pool) { bool active; @@ -115,16 +132,19 @@ storagePoolUpdateState(virStoragePoolObjPtr pool) if (backend->refreshPool(NULL, pool) < 0) { if (backend->stopPool) backend->stopPool(NULL, pool); + active = false; virReportError(VIR_ERR_INTERNAL_ERROR, _("Failed to restart storage pool '%s': %s"), pool->def->name, virGetLastErrorMessage()); goto error; } + active = true; } - pool->active = active; ret = 0; error: + if (!active) + storagePoolSetInactive(&pool); if (ret < 0) { if (stateFile) unlink(stateFile); @@ -201,6 +221,8 @@ storageDriverAutostart(void) virReportError(VIR_ERR_INTERNAL_ERROR, _("Failed to autostart storage pool '%s': %s"), pool->def->name, virGetLastErrorMessage()); + storagePoolSetInactive(&pool); + continue; } else { pool->active = true; } @@ -737,8 +759,7 @@ storagePoolCreateXML(virConnectPtr conn, unlink(stateFile); if (backend->stopPool) backend->stopPool(conn, pool); - virStoragePoolObjRemove(&driver->pools, pool); - pool = NULL; + storagePoolSetInactive(&pool); goto cleanup; } @@ -1068,16 +1089,7 @@ storagePoolDestroy(virStoragePoolPtr obj) VIR_STORAGE_POOL_EVENT_STOPPED, 0); - pool->active = false; - - if (pool->configFile == NULL) { - virStoragePoolObjRemove(&driver->pools, pool); - pool = NULL; - } else if (pool->newDef) { - virStoragePoolDefFree(pool->def); - pool->def = pool->newDef; - pool->newDef = NULL; - } + storagePoolSetInactive(&pool); ret = 0; @@ -1197,12 +1209,7 @@ storagePoolRefresh(virStoragePoolPtr obj, pool->def->uuid, VIR_STORAGE_POOL_EVENT_STOPPED, 0); - pool->active = false; - - if (pool->configFile == NULL) { - virStoragePoolObjRemove(&driver->pools, pool); - pool = NULL; - } + storagePoolSetInactive(&pool); goto cleanup; } -- 2.5.5

2 1

[libvirt] [PATCH 0/4] Move ccwaddrs and pciaddrs to domainDef
by Tomasz Flendrich 23 Jun '16

23 Jun '16

This patch depends on another one, because it was created on top of the changes in the other one. Link to the dependency: https://www.redhat.com/archives/libvir-list/2016-June/msg01227.html This is the next patch of the ongoing process of abstracting device address generation. In this patch, addres sets are moved from QEMU's (ccw and pci) and bhyve's (pci only) private data to domainDef. In the future, it's likely that more hypervisors will add the feature of explicitly specifying the device addresses, so handling the addresses should be made more generic. Tomasz Flendrich (4): Move functions from qemu_domain_address to domain_addr Move ccwaddrs from QEMU's private data to domainDef Move virDomainPCIAddressSetFree from domain_addr to domain_conf Move pciaddrs from private data to domainDef src/bhyve/bhyve_device.c | 16 +- src/bhyve/bhyve_domain.c | 2 - src/bhyve/bhyve_domain.h | 3 - src/conf/domain_addr.c | 497 ++++++++++++++++++++++++++++++++++-- src/conf/domain_addr.h | 67 ++--- src/conf/domain_conf.c | 21 ++ src/conf/domain_conf.h | 51 ++++ src/libvirt_private.syms | 11 +- src/qemu/qemu_domain.c | 4 +- src/qemu/qemu_domain.h | 3 - src/qemu/qemu_domain_address.c | 561 +++-------------------------------------- src/qemu/qemu_domain_address.h | 3 +- src/qemu/qemu_hotplug.c | 23 +- src/qemu/qemu_process.c | 6 +- tests/qemuhotplugtest.c | 2 +- 15 files changed, 629 insertions(+), 641 deletions(-) -- 2.7.4 (Apple Git-66)

4 12

[libvirt] [PATCH 0/2] Abstracting device address allocation
by Tomasz Flendrich 23 Jun '16

23 Jun '16

More hypervisors may implement the possibility of specifying device addresses, so it makes sense to make QEMU-specific code more generic to be prepared for it. If this approach will be acceptable, all the other address sets will be moved from QEMU's private data to domainDef, the qemu_domain_address.c will shrink in size and the functionality will be moved to domain_addr.[hc]. Tomasz Flendrich (2): Move virtio *Free() functions from domain_addr to domain_conf Move vioserialaddrs from QEMU's private data to domainDef src/conf/domain_addr.c | 22 ---------------------- src/conf/domain_addr.h | 17 ----------------- src/conf/domain_conf.c | 23 +++++++++++++++++++++++ src/conf/domain_conf.h | 19 +++++++++++++++++++ src/libvirt_private.syms | 3 ++- src/qemu/qemu_domain.c | 1 - src/qemu/qemu_domain.h | 1 - src/qemu/qemu_domain_address.c | 20 ++++++++------------ src/qemu/qemu_hotplug.c | 9 +++++---- 9 files changed, 57 insertions(+), 58 deletions(-) -- 2.7.4 (Apple Git-66)

2 3

[libvirt] [PATCH] qemu: make monitor command API available during async jobs
by Nikolay Shirokovskiy 23 Jun '16

23 Jun '16

One can not issue monitor commands manually during async calls thru designated API while this could be useful for testing/debugging purposes. qemuDomainQemuMonitorCommand uses job of type QEMU_JOB_MODIFY and any async call disable parallel execution of this type of job. The only state that is changed is taint variable. AFAIU the only place we can mess is resetting taint flag in qemuProcessStop routine under some async job. But this can not happen thanx to both virDomainObjIsActive check in qemuDomainQemuMonitorCommand and resetting active status in qemuProcessStop before taint flag. Change job type to QEMU_JOB_QUERY and thus make the API call available for most of async jobs. Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy(a)virtuozzo.com> --- src/qemu/qemu_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 931efae..d6aa640 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -15514,7 +15514,7 @@ static int qemuDomainQemuMonitorCommand(virDomainPtr domain, const char *cmd, if (virDomainQemuMonitorCommandEnsureACL(domain->conn, vm->def) < 0) goto cleanup; - if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_MODIFY) < 0) + if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_QUERY) < 0) goto cleanup; if (!virDomainObjIsActive(vm)) { -- 1.8.3.1

2 3

[libvirt] [PATCH 00/11] Add auto convergence migration parameters
by Jiri Denemark 22 Jun '16

22 Jun '16

This series adds support for a couple of migration parameters which can be used to tune the auto-convergence algorithm in QEMU. Jiri Denemark (11): qemu: Make qemuMonitorSetMigrationCompression saner qemu: Decouple migration parameters from compression settings qemu: Rename qemuMonitorMigrationCompression qemu: Rework qemuMonitorJSONGetMigrationParams qemu: Rework qemuMonitorJSONSetMigrationParams test: Rework qemuMonitorJSONGetMigrationParams test Add auto convergence migration parameters qemu: Introduce qemuMigrationSetParams qemu: Add support for cpu throttling parameters qemu: Implement auto convergence migration parameters Report auto convergence throttle rate in migration stats include/libvirt/libvirt-domain.h | 27 ++++++ src/qemu/qemu_domain.c | 6 ++ src/qemu/qemu_driver.c | 13 ++- src/qemu/qemu_migration.c | 195 ++++++++++++++++++++++++++++++--------- src/qemu/qemu_migration.h | 17 +++- src/qemu/qemu_monitor.c | 29 ++++-- src/qemu/qemu_monitor.h | 37 +++++--- src/qemu/qemu_monitor_json.c | 74 +++++++-------- src/qemu/qemu_monitor_json.h | 8 +- tests/qemumonitorjsontest.c | 60 ++++++------ tools/virsh-domain.c | 35 +++++++ tools/virsh.pod | 16 ++-- 12 files changed, 364 insertions(+), 153 deletions(-) -- 2.9.0

2 22

[libvirt] [PATCH] syntax-check: rewrite prohibit-duplicate-header in perl
by Ján Tomko 22 Jun '16

22 Jun '16

Invoke the script only once instead of once for every file. --- build-aux/prohibit-duplicate-header.pl | 22 ++++++++++++++++++++++ cfg.mk | 29 +++++------------------------ 2 files changed, 27 insertions(+), 24 deletions(-) create mode 100644 build-aux/prohibit-duplicate-header.pl diff --git a/build-aux/prohibit-duplicate-header.pl b/build-aux/prohibit-duplicate-header.pl new file mode 100644 index 0000000..f9ab3f5 --- /dev/null +++ b/build-aux/prohibit-duplicate-header.pl @@ -0,0 +1,22 @@ +#!/usr/bin/perl + +use strict; + +my $file = " "; +my $ret = 0; +my %includes = ( ); + +while (<>) { + if (not $file eq $ARGV) { + %includes = ( ); + $file = $ARGV; + } + if (/^# *include *[<"]([^>"]*\.h)[">]/) { + $includes{$1}++; + if ($includes{$1} == 2) { + $ret = 1; + print STDERR "$1 included multiple times in $ARGV\n"; + } + } +} +exit $ret; diff --git a/cfg.mk b/cfg.mk index 297ca3a..69ab6fe 100644 --- a/cfg.mk +++ b/cfg.mk @@ -813,30 +813,6 @@ sc_prohibit_exit_in_tests: halt='use return, not exit(), in tests' \ $(_sc_search_regexp) -# Don't include duplicate header in the source (either *.c or *.h) -sc_prohibit_duplicate_header: - @fail=0; for i in $$($(VC_LIST_EXCEPT) | grep '\.[chx]$$'); do \ - awk '/# *include.*\.h/ { \ - match($$0, /[<"][^>"]*[">]/); \ - arr[substr($$0, RSTART + 1, RLENGTH - 2)]++; \ - } \ - END { \ - for (key in arr) { \ - if (arr[key] > 1) { \ - fail=1; \ - printf("%d %s are included\n", arr[key], key); \ - } \ - } \ - if (fail == 1) { \ - printf("duplicate header(s) in " FILENAME "\n"); \ - exit 1; \ - } \ - }' $$i || fail=1; \ - done; \ - if test $$fail -eq 1; then \ - { echo '$(ME): avoid duplicate headers' 1>&2; exit 1; } \ - fi; - # Don't include "libvirt/*.h" in "" form. sc_prohibit_include_public_headers_quote: @prohibit='# *include *"libvirt/.*\.h"' \ @@ -1100,6 +1076,11 @@ ifneq ($(_gl-Makefile),) syntax-check: $(top_srcdir)/HACKING spacing-check test-wrap-argv endif +# Don't include duplicate header in the source (either *.c or *.h) +prohibit_duplicate_header: + $(AM_V_GEN)files= $$($(VC_LIST_EXCEPT) | grep '\.[chx]$$') \ + $(PERL) -W $(top_srcdir)/build-aux/prohibit-duplicate-header.pl $$files + spacing-check: $(AM_V_GEN)files=`$(VC_LIST) | grep '\.c$$'`; \ $(PERL) $(top_srcdir)/build-aux/check-spacing.pl $$files || \ -- 2.7.3

2 2

[libvirt] [PATCH] conf: Simplify conditions in CPU parser/formatter
by Jiri Denemark 22 Jun '16

22 Jun '16

Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/conf/cpu_conf.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/src/conf/cpu_conf.c b/src/conf/cpu_conf.c index 7230b89..b71528e 100644 --- a/src/conf/cpu_conf.c +++ b/src/conf/cpu_conf.c @@ -330,8 +330,7 @@ virCPUDefParseXML(xmlNodePtr node, goto error; if (n > 0) { - if (!def->model && def->mode != VIR_CPU_MODE_HOST_MODEL && - def->mode != VIR_CPU_MODE_HOST_PASSTHROUGH) { + if (!def->model && def->mode == VIR_CPU_MODE_CUSTOM) { virReportError(VIR_ERR_XML_ERROR, "%s", _("Non-empty feature list specified without " "CPU model")); @@ -518,10 +517,7 @@ virCPUDefFormatBuf(virBufferPtr buf, (def->mode == VIR_CPU_MODE_HOST_MODEL || (def->mode == VIR_CPU_MODE_CUSTOM && def->model))); - if (!def->model && - def->mode != VIR_CPU_MODE_HOST_MODEL && - def->mode != VIR_CPU_MODE_HOST_PASSTHROUGH && - def->nfeatures) { + if (!def->model && def->mode == VIR_CPU_MODE_CUSTOM && def->nfeatures) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Non-empty feature list specified without CPU model")); return -1; -- 2.9.0

2 1