| Message ID | 1361875228-15769-15-git-send-email-xiawenc@linux.vnet.ibm.com |
|---|---|
| State | New |
| Headers | show |
On 02/26/2013 03:40 AM, Wenchao Xia wrote: > This patch added the support of showing internal snapshots on an > image in the backing chain of a block device in hmp layer, by > calling a qmp function. > > Signed-off-by: Wenchao Xia <xiawenc@linux.vnet.ibm.com> > --- > hmp.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- > monitor.c | 6 ++-- > 2 files changed, 83 insertions(+), 4 deletions(-) > > diff --git a/hmp.c b/hmp.c > index 5e68b2f..ffa92ff 100644 > +static ImageInfo *find_image_info(const DeviceImageInfoList *device_image_list, > + const char *image_name) > +{ > + ImageInfo *image_info; > + > + > + /* search the chain */ > + while (device_image_list) { > + if (device_image_list->value->has_image) { > + image_info = device_image_list->value->image; > + if (!strcmp(image_info->filename, image_name)) { > + return image_info; > + } > + } Is this loop smart enough, or should it be borrowing from some of the additional complexity in bdrv_find_backing_image for properly chasing down relative names? > + if (image_info) { > + if (image_info->has_snapshots) { > + list = image_info->snapshots; > + monitor_printf(mon, "Device '%s', Image name '%s':\n", > + device_name, image_info->filename); > + monitor_dump_snapshotinfolist(mon, list); > + } else { > + monitor_printf(mon, "Device '%s' have no valid " s/have/has/ > + "internal snapshot.\n", > + device_name); > + } > + } else { > + monitor_printf(mon, "Device '%s' have no correspond image now.\n", s/have no correspond image now/has no corresponding image/
于 2013-2-27 0:13, Eric Blake 写道: > On 02/26/2013 03:40 AM, Wenchao Xia wrote: >> This patch added the support of showing internal snapshots on an >> image in the backing chain of a block device in hmp layer, by >> calling a qmp function. >> >> Signed-off-by: Wenchao Xia <xiawenc@linux.vnet.ibm.com> >> --- >> hmp.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- >> monitor.c | 6 ++-- >> 2 files changed, 83 insertions(+), 4 deletions(-) >> >> diff --git a/hmp.c b/hmp.c >> index 5e68b2f..ffa92ff 100644 > >> +static ImageInfo *find_image_info(const DeviceImageInfoList *device_image_list, >> + const char *image_name) >> +{ >> + ImageInfo *image_info; >> + > >> + >> + /* search the chain */ >> + while (device_image_list) { >> + if (device_image_list->value->has_image) { >> + image_info = device_image_list->value->image; >> + if (!strcmp(image_info->filename, image_name)) { >> + return image_info; >> + } >> + } > > Is this loop smart enough, or should it be borrowing from some of the > additional complexity in bdrv_find_backing_image for properly chasing > down relative names? > OK, it will be enhanced. > >> + if (image_info) { >> + if (image_info->has_snapshots) { >> + list = image_info->snapshots; >> + monitor_printf(mon, "Device '%s', Image name '%s':\n", >> + device_name, image_info->filename); >> + monitor_dump_snapshotinfolist(mon, list); >> + } else { >> + monitor_printf(mon, "Device '%s' have no valid " > > s/have/has/ > OK. >> + "internal snapshot.\n", >> + device_name); >> + } >> + } else { >> + monitor_printf(mon, "Device '%s' have no correspond image now.\n", > > s/have no correspond image now/has no corresponding image/ > OK.
于 2013-2-27 10:29, Wenchao Xia 写道: > 于 2013-2-27 0:13, Eric Blake 写道: >> On 02/26/2013 03:40 AM, Wenchao Xia wrote: >>> This patch added the support of showing internal snapshots on an >>> image in the backing chain of a block device in hmp layer, by >>> calling a qmp function. >>> >>> Signed-off-by: Wenchao Xia <xiawenc@linux.vnet.ibm.com> >>> --- >>> hmp.c | 81 >>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- >>> monitor.c | 6 ++-- >>> 2 files changed, 83 insertions(+), 4 deletions(-) >>> >>> diff --git a/hmp.c b/hmp.c >>> index 5e68b2f..ffa92ff 100644 >> >>> +static ImageInfo *find_image_info(const DeviceImageInfoList >>> *device_image_list, >>> + const char *image_name) >>> +{ >>> + ImageInfo *image_info; >>> + >> >>> + >>> + /* search the chain */ >>> + while (device_image_list) { >>> + if (device_image_list->value->has_image) { >>> + image_info = device_image_list->value->image; >>> + if (!strcmp(image_info->filename, image_name)) { >>> + return image_info; >>> + } >>> + } >> >> Is this loop smart enough, or should it be borrowing from some of the >> additional complexity in bdrv_find_backing_image for properly chasing >> down relative names? >> > OK, it will be enhanced. > To make things more straight, I guess a new hmp API is better: info image <device> -b which may show all image info on a device, including internal snapshot. In this case, hmp/qmp command are connected more closely and saves the trouble of find correct one in hmp level. If you agree I'll discard info snapshots <device> in next version. >> >>> + if (image_info) { >>> + if (image_info->has_snapshots) { >>> + list = image_info->snapshots; >>> + monitor_printf(mon, "Device '%s', Image name '%s':\n", >>> + device_name, image_info->filename); >>> + monitor_dump_snapshotinfolist(mon, list); >>> + } else { >>> + monitor_printf(mon, "Device '%s' have no valid " >> >> s/have/has/ >> > OK. > >>> + "internal snapshot.\n", >>> + device_name); >>> + } >>> + } else { >>> + monitor_printf(mon, "Device '%s' have no correspond >>> image now.\n", >> >> s/have no correspond image now/has no corresponding image/ >> > OK. > >
diff --git a/hmp.c b/hmp.c index 5e68b2f..ffa92ff 100644 --- a/hmp.c +++ b/hmp.c @@ -629,6 +629,73 @@ static void monitor_dump_snapshotinfolist(Monitor *mon, SnapshotInfoList *list) } } +/* Find the ImageInfo with correct image_name. If image_name = NULL, then + return the top one in the chain. */ +static ImageInfo *find_image_info(const DeviceImageInfoList *device_image_list, + const char *image_name) +{ + ImageInfo *image_info; + + /* return top one */ + if (!image_name) { + if (device_image_list->value->has_image) { + return device_image_list->value->image; + } else { + return NULL; + } + } + + /* search the chain */ + while (device_image_list) { + if (device_image_list->value->has_image) { + image_info = device_image_list->value->image; + if (!strcmp(image_info->filename, image_name)) { + return image_info; + } + } + device_image_list = device_image_list->next; + } + return NULL; +} + +static void hmp_info_snapshots_device(Monitor *mon, + const char *device_name, + const char *image_name) +{ + Error *err = NULL; + SnapshotInfoList *list; + + DeviceImageInfoList *device_image_list = qmp_query_images(true, + device_name, true, true, &err); + if (error_is_set(&err)) { + hmp_handle_error(mon, &err); + return; + } + if (!device_image_list) { + monitor_printf(mon, "Device '%s' can't get image info list.\n", + device_name); + return; + } + + ImageInfo *image_info = find_image_info(device_image_list, image_name); + if (image_info) { + if (image_info->has_snapshots) { + list = image_info->snapshots; + monitor_printf(mon, "Device '%s', Image name '%s':\n", + device_name, image_info->filename); + monitor_dump_snapshotinfolist(mon, list); + } else { + monitor_printf(mon, "Device '%s' have no valid " + "internal snapshot.\n", + device_name); + } + } else { + monitor_printf(mon, "Device '%s' have no correspond image now.\n", + device_name); + } + qapi_free_DeviceImageInfoList(device_image_list); +} + static void hmp_info_snapshots_vm(Monitor *mon) { Error *err = NULL; @@ -651,7 +718,19 @@ static void hmp_info_snapshots_vm(Monitor *mon) void hmp_info_snapshots(Monitor *mon, const QDict *qdict) { - hmp_info_snapshots_vm(mon); + const char *device_name = qdict_get_try_str(qdict, "device"); + const char *image_name = qdict_get_try_str(qdict, "image"); + + if (device_name) { + hmp_info_snapshots_device(mon, device_name, image_name); + } else { + if (image_name) { + monitor_printf(mon, "Parameter [image] is valid only when " + "device is specified.\n"); + return; + } + hmp_info_snapshots_vm(mon); + } } void hmp_quit(Monitor *mon, const QDict *qdict) diff --git a/monitor.c b/monitor.c index 5112375..1dca36f 100644 --- a/monitor.c +++ b/monitor.c @@ -2591,9 +2591,9 @@ static mon_cmd_t info_cmds[] = { }, { .name = "snapshots", - .args_type = "", - .params = "", - .help = "show the currently saved VM snapshots", + .args_type = "device:O?", + .params = "[device=str] [,image=str]", + .help = "show snapshots of whole vm or a single device", .mhandler.cmd = hmp_info_snapshots, }, {
This patch added the support of showing internal snapshots on an image in the backing chain of a block device in hmp layer, by calling a qmp function. Signed-off-by: Wenchao Xia <xiawenc@linux.vnet.ibm.com> --- hmp.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- monitor.c | 6 ++-- 2 files changed, 83 insertions(+), 4 deletions(-)