qemu-e2k/block/qapi.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

1030 lines
33 KiB
C
Raw Normal View History

/*
* Block layer qmp and info dump related functions
*
* Copyright (c) 2003-2008 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "qemu/osdep.h"
#include "qemu/cutils.h"
#include "block/qapi.h"
#include "block/block_int.h"
#include "block/dirty-bitmap.h"
#include "block/throttle-groups.h"
block: add event when disk usage exceeds threshold Managing applications, like oVirt (http://www.ovirt.org), make extensive use of thin-provisioned disk images. To let the guest run smoothly and be not unnecessarily paused, oVirt sets a disk usage threshold (so called 'high water mark') based on the occupation of the device, and automatically extends the image once the threshold is reached or exceeded. In order to detect the crossing of the threshold, oVirt has no choice but aggressively polling the QEMU monitor using the query-blockstats command. This lead to unnecessary system load, and is made even worse under scale: deployments with hundreds of VMs are no longer rare. To fix this, this patch adds: * A new monitor command `block-set-write-threshold', to set a mark for a given block device. * A new event `BLOCK_WRITE_THRESHOLD', to report if a block device usage exceeds the threshold. * A new `write_threshold' field into the `BlockDeviceInfo' structure, to report the configured threshold. This will allow the managing application to use smarter and more efficient monitoring, greatly reducing the need of polling. [Updated qemu-iotests 067 output to add the new 'write_threshold' property. --Stefan] [Changed g_assert_false() to !g_assert() to fix the build on older glib versions. --Kevin] Signed-off-by: Francesco Romani <fromani@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-id: 1421068273-692-1-git-send-email-fromani@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2015-01-12 14:11:13 +01:00
#include "block/write-threshold.h"
#include "qapi/error.h"
#include "qapi/qapi-commands-block-core.h"
#include "qapi/qobject-output-visitor.h"
#include "qapi/qapi-visit-block-core.h"
#include "qapi/qmp/qbool.h"
#include "qapi/qmp/qdict.h"
#include "qapi/qmp/qlist.h"
#include "qapi/qmp/qnum.h"
#include "qapi/qmp/qstring.h"
#include "qemu/qemu-print.h"
#include "sysemu/block-backend.h"
BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk,
BlockDriverState *bs,
bool flat,
Error **errp)
{
ImageInfo **p_image_info;
ImageInfo *backing_info;
block: Leave BDS.backing_{file,format} constant Parts of the block layer treat BDS.backing_file as if it were whatever the image header says (i.e., if it is a relative path, it is relative to the overlay), other parts treat it like a cache for bs->backing->bs->filename (relative paths are relative to the CWD). Considering bs->backing->bs->filename exists, let us make it mean the former. Among other things, this now allows the user to specify a base when using qemu-img to commit an image file in a directory that is not the CWD (assuming, everything uses relative filenames). Before this patch: $ ./qemu-img create -f qcow2 foo/bot.qcow2 1M $ ./qemu-img create -f qcow2 -b bot.qcow2 foo/mid.qcow2 $ ./qemu-img create -f qcow2 -b mid.qcow2 foo/top.qcow2 $ ./qemu-img commit -b mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b $PWD/foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find '[...]/foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' After this patch: $ ./qemu-img commit -b mid.qcow2 foo/top.qcow2 Image committed. $ ./qemu-img commit -b foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b $PWD/foo/mid.qcow2 foo/top.qcow2 Image committed. With this change, bdrv_find_backing_image() must look at whether the user has overridden a BDS's backing file. If so, it can no longer use bs->backing_file, but must instead compare the given filename against the backing node's filename directly. Note that this changes the QAPI output for a node's backing_file. We had very inconsistent output there (sometimes what the image header said, sometimes the actual filename of the backing image). This inconsistent output was effectively useless, so we have to decide one way or the other. Considering that bs->backing_file usually at runtime contained the path to the image relative to qemu's CWD (or absolute), this patch changes QAPI's backing_file to always report the bs->backing->bs->filename from now on. If you want to receive the image header information, you have to refer to full-backing-filename. This necessitates a change to iotest 228. The interesting information it really wanted is the image header, and it can get that now, but it has to use full-backing-filename instead of backing_file. Because of this patch's changes to bs->backing_file's behavior, we also need some reference output changes. Along with the changes to bs->backing_file, stop updating BDS.backing_format in bdrv_backing_attach() as well. This way, ImageInfo's backing-filename and backing-filename-format fields will represent what the image header says and nothing else. iotest 245 changes in behavior: With the backing node no longer overriding the parent node's backing_file string, you can now omit the @backing option when reopening a node with neither a default nor a current backing file even if it used to have a backing node at some point. 273 also changes: The base image is opened without a format layer, so ImageInfo.backing-filename-format used to report "file" for the base image's overlay after blockdev-snapshot. However, the image header never says "file" anywhere, so it now reports $IMGFMT. Signed-off-by: Max Reitz <mreitz@redhat.com>
2018-08-01 20:34:11 +02:00
BlockDriverState *bs0, *backing;
BlockDeviceInfo *info;
ERRP_GUARD();
if (!bs->drv) {
error_setg(errp, "Block device %s is ejected", bs->node_name);
return NULL;
}
bdrv_refresh_filename(bs);
info = g_malloc0(sizeof(*info));
info->file = g_strdup(bs->filename);
info->ro = bdrv_is_read_only(bs);
info->drv = g_strdup(bs->drv->format_name);
info->encrypted = bs->encrypted;
info->cache = g_new(BlockdevCacheInfo, 1);
*info->cache = (BlockdevCacheInfo) {
.writeback = blk ? blk_enable_write_cache(blk) : true,
.direct = !!(bs->open_flags & BDRV_O_NOCACHE),
.no_flush = !!(bs->open_flags & BDRV_O_NO_FLUSH),
};
if (bs->node_name[0]) {
info->node_name = g_strdup(bs->node_name);
}
block: Leave BDS.backing_{file,format} constant Parts of the block layer treat BDS.backing_file as if it were whatever the image header says (i.e., if it is a relative path, it is relative to the overlay), other parts treat it like a cache for bs->backing->bs->filename (relative paths are relative to the CWD). Considering bs->backing->bs->filename exists, let us make it mean the former. Among other things, this now allows the user to specify a base when using qemu-img to commit an image file in a directory that is not the CWD (assuming, everything uses relative filenames). Before this patch: $ ./qemu-img create -f qcow2 foo/bot.qcow2 1M $ ./qemu-img create -f qcow2 -b bot.qcow2 foo/mid.qcow2 $ ./qemu-img create -f qcow2 -b mid.qcow2 foo/top.qcow2 $ ./qemu-img commit -b mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b $PWD/foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find '[...]/foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' After this patch: $ ./qemu-img commit -b mid.qcow2 foo/top.qcow2 Image committed. $ ./qemu-img commit -b foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b $PWD/foo/mid.qcow2 foo/top.qcow2 Image committed. With this change, bdrv_find_backing_image() must look at whether the user has overridden a BDS's backing file. If so, it can no longer use bs->backing_file, but must instead compare the given filename against the backing node's filename directly. Note that this changes the QAPI output for a node's backing_file. We had very inconsistent output there (sometimes what the image header said, sometimes the actual filename of the backing image). This inconsistent output was effectively useless, so we have to decide one way or the other. Considering that bs->backing_file usually at runtime contained the path to the image relative to qemu's CWD (or absolute), this patch changes QAPI's backing_file to always report the bs->backing->bs->filename from now on. If you want to receive the image header information, you have to refer to full-backing-filename. This necessitates a change to iotest 228. The interesting information it really wanted is the image header, and it can get that now, but it has to use full-backing-filename instead of backing_file. Because of this patch's changes to bs->backing_file's behavior, we also need some reference output changes. Along with the changes to bs->backing_file, stop updating BDS.backing_format in bdrv_backing_attach() as well. This way, ImageInfo's backing-filename and backing-filename-format fields will represent what the image header says and nothing else. iotest 245 changes in behavior: With the backing node no longer overriding the parent node's backing_file string, you can now omit the @backing option when reopening a node with neither a default nor a current backing file even if it used to have a backing node at some point. 273 also changes: The base image is opened without a format layer, so ImageInfo.backing-filename-format used to report "file" for the base image's overlay after blockdev-snapshot. However, the image header never says "file" anywhere, so it now reports $IMGFMT. Signed-off-by: Max Reitz <mreitz@redhat.com>
2018-08-01 20:34:11 +02:00
backing = bdrv_cow_bs(bs);
if (backing) {
info->backing_file = g_strdup(backing->filename);
}
if (!QLIST_EMPTY(&bs->dirty_bitmaps)) {
info->has_dirty_bitmaps = true;
info->dirty_bitmaps = bdrv_query_dirty_bitmaps(bs);
}
info->detect_zeroes = bs->detect_zeroes;
if (blk && blk_get_public(blk)->throttle_group_member.throttle_state) {
ThrottleConfig cfg;
BlockBackendPublic *blkp = blk_get_public(blk);
throttle_group_get_config(&blkp->throttle_group_member, &cfg);
info->bps = cfg.buckets[THROTTLE_BPS_TOTAL].avg;
info->bps_rd = cfg.buckets[THROTTLE_BPS_READ].avg;
info->bps_wr = cfg.buckets[THROTTLE_BPS_WRITE].avg;
info->iops = cfg.buckets[THROTTLE_OPS_TOTAL].avg;
info->iops_rd = cfg.buckets[THROTTLE_OPS_READ].avg;
info->iops_wr = cfg.buckets[THROTTLE_OPS_WRITE].avg;
info->has_bps_max = cfg.buckets[THROTTLE_BPS_TOTAL].max;
info->bps_max = cfg.buckets[THROTTLE_BPS_TOTAL].max;
info->has_bps_rd_max = cfg.buckets[THROTTLE_BPS_READ].max;
info->bps_rd_max = cfg.buckets[THROTTLE_BPS_READ].max;
info->has_bps_wr_max = cfg.buckets[THROTTLE_BPS_WRITE].max;
info->bps_wr_max = cfg.buckets[THROTTLE_BPS_WRITE].max;
info->has_iops_max = cfg.buckets[THROTTLE_OPS_TOTAL].max;
info->iops_max = cfg.buckets[THROTTLE_OPS_TOTAL].max;
info->has_iops_rd_max = cfg.buckets[THROTTLE_OPS_READ].max;
info->iops_rd_max = cfg.buckets[THROTTLE_OPS_READ].max;
info->has_iops_wr_max = cfg.buckets[THROTTLE_OPS_WRITE].max;
info->iops_wr_max = cfg.buckets[THROTTLE_OPS_WRITE].max;
info->has_bps_max_length = info->has_bps_max;
info->bps_max_length =
cfg.buckets[THROTTLE_BPS_TOTAL].burst_length;
info->has_bps_rd_max_length = info->has_bps_rd_max;
info->bps_rd_max_length =
cfg.buckets[THROTTLE_BPS_READ].burst_length;
info->has_bps_wr_max_length = info->has_bps_wr_max;
info->bps_wr_max_length =
cfg.buckets[THROTTLE_BPS_WRITE].burst_length;
info->has_iops_max_length = info->has_iops_max;
info->iops_max_length =
cfg.buckets[THROTTLE_OPS_TOTAL].burst_length;
info->has_iops_rd_max_length = info->has_iops_rd_max;
info->iops_rd_max_length =
cfg.buckets[THROTTLE_OPS_READ].burst_length;
info->has_iops_wr_max_length = info->has_iops_wr_max;
info->iops_wr_max_length =
cfg.buckets[THROTTLE_OPS_WRITE].burst_length;
info->has_iops_size = cfg.op_size;
info->iops_size = cfg.op_size;
info->group =
g_strdup(throttle_group_get_name(&blkp->throttle_group_member));
}
block: add event when disk usage exceeds threshold Managing applications, like oVirt (http://www.ovirt.org), make extensive use of thin-provisioned disk images. To let the guest run smoothly and be not unnecessarily paused, oVirt sets a disk usage threshold (so called 'high water mark') based on the occupation of the device, and automatically extends the image once the threshold is reached or exceeded. In order to detect the crossing of the threshold, oVirt has no choice but aggressively polling the QEMU monitor using the query-blockstats command. This lead to unnecessary system load, and is made even worse under scale: deployments with hundreds of VMs are no longer rare. To fix this, this patch adds: * A new monitor command `block-set-write-threshold', to set a mark for a given block device. * A new event `BLOCK_WRITE_THRESHOLD', to report if a block device usage exceeds the threshold. * A new `write_threshold' field into the `BlockDeviceInfo' structure, to report the configured threshold. This will allow the managing application to use smarter and more efficient monitoring, greatly reducing the need of polling. [Updated qemu-iotests 067 output to add the new 'write_threshold' property. --Stefan] [Changed g_assert_false() to !g_assert() to fix the build on older glib versions. --Kevin] Signed-off-by: Francesco Romani <fromani@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-id: 1421068273-692-1-git-send-email-fromani@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2015-01-12 14:11:13 +01:00
info->write_threshold = bdrv_write_threshold_get(bs);
bs0 = bs;
p_image_info = &info->image;
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
info->backing_file_depth = 0;
/*
* Skip automatically inserted nodes that the user isn't aware of for
* query-block (blk != NULL), but not for query-named-block-nodes
*/
bdrv_query_image_info(bs0, p_image_info, flat, blk != NULL, errp);
if (*errp) {
qapi_free_BlockDeviceInfo(info);
return NULL;
}
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
backing_info = info->image->backing_image;
while (backing_info) {
info->backing_file_depth++;
backing_info = backing_info->backing_image;
}
return info;
}
/*
* Returns 0 on success, with *p_list either set to describe snapshot
* information, or NULL because there are no snapshots. Returns -errno on
* error, with *p_list untouched.
*/
int bdrv_query_snapshot_info_list(BlockDriverState *bs,
SnapshotInfoList **p_list,
Error **errp)
{
int i, sn_count;
QEMUSnapshotInfo *sn_tab = NULL;
SnapshotInfoList *head = NULL, **tail = &head;
SnapshotInfo *info;
sn_count = bdrv_snapshot_list(bs, &sn_tab);
if (sn_count < 0) {
const char *dev = bdrv_get_device_name(bs);
switch (sn_count) {
case -ENOMEDIUM:
error_setg(errp, "Device '%s' is not inserted", dev);
break;
case -ENOTSUP:
error_setg(errp,
"Device '%s' does not support internal snapshots",
dev);
break;
default:
error_setg_errno(errp, -sn_count,
"Can't list snapshots of device '%s'", dev);
break;
}
return sn_count;
}
for (i = 0; i < sn_count; i++) {
info = g_new0(SnapshotInfo, 1);
info->id = g_strdup(sn_tab[i].id_str);
info->name = g_strdup(sn_tab[i].name);
info->vm_state_size = sn_tab[i].vm_state_size;
info->date_sec = sn_tab[i].date_sec;
info->date_nsec = sn_tab[i].date_nsec;
info->vm_clock_sec = sn_tab[i].vm_clock_nsec / 1000000000;
info->vm_clock_nsec = sn_tab[i].vm_clock_nsec % 1000000000;
info->icount = sn_tab[i].icount;
info->has_icount = sn_tab[i].icount != -1ULL;
QAPI_LIST_APPEND(tail, info);
}
g_free(sn_tab);
*p_list = head;
return 0;
}
/**
* Helper function for other query info functions. Store information about @bs
* in @info, setting @errp on error.
*/
static void bdrv_do_query_node_info(BlockDriverState *bs,
BlockNodeInfo *info,
Error **errp)
{
int64_t size;
const char *backing_filename;
BlockDriverInfo bdi;
int ret;
Error *err = NULL;
aio_context_acquire(bdrv_get_aio_context(bs));
size = bdrv_getlength(bs);
if (size < 0) {
error_setg_errno(errp, -size, "Can't get image size '%s'",
bs->exact_filename);
goto out;
}
bdrv_refresh_filename(bs);
info->filename = g_strdup(bs->filename);
info->format = g_strdup(bdrv_get_format_name(bs));
info->virtual_size = size;
info->actual_size = bdrv_get_allocated_file_size(bs);
info->has_actual_size = info->actual_size >= 0;
if (bs->encrypted) {
info->encrypted = true;
info->has_encrypted = true;
}
if (bdrv_get_info(bs, &bdi) >= 0) {
if (bdi.cluster_size != 0) {
info->cluster_size = bdi.cluster_size;
info->has_cluster_size = true;
}
info->dirty_flag = bdi.is_dirty;
info->has_dirty_flag = true;
}
info->format_specific = bdrv_get_specific_info(bs, &err);
if (err) {
error_propagate(errp, err);
goto out;
}
backing_filename = bs->backing_file;
if (backing_filename[0] != '\0') {
char *backing_filename2;
block: Leave BDS.backing_{file,format} constant Parts of the block layer treat BDS.backing_file as if it were whatever the image header says (i.e., if it is a relative path, it is relative to the overlay), other parts treat it like a cache for bs->backing->bs->filename (relative paths are relative to the CWD). Considering bs->backing->bs->filename exists, let us make it mean the former. Among other things, this now allows the user to specify a base when using qemu-img to commit an image file in a directory that is not the CWD (assuming, everything uses relative filenames). Before this patch: $ ./qemu-img create -f qcow2 foo/bot.qcow2 1M $ ./qemu-img create -f qcow2 -b bot.qcow2 foo/mid.qcow2 $ ./qemu-img create -f qcow2 -b mid.qcow2 foo/top.qcow2 $ ./qemu-img commit -b mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b $PWD/foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find '[...]/foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' After this patch: $ ./qemu-img commit -b mid.qcow2 foo/top.qcow2 Image committed. $ ./qemu-img commit -b foo/mid.qcow2 foo/top.qcow2 qemu-img: Did not find 'foo/mid.qcow2' in the backing chain of 'foo/top.qcow2' $ ./qemu-img commit -b $PWD/foo/mid.qcow2 foo/top.qcow2 Image committed. With this change, bdrv_find_backing_image() must look at whether the user has overridden a BDS's backing file. If so, it can no longer use bs->backing_file, but must instead compare the given filename against the backing node's filename directly. Note that this changes the QAPI output for a node's backing_file. We had very inconsistent output there (sometimes what the image header said, sometimes the actual filename of the backing image). This inconsistent output was effectively useless, so we have to decide one way or the other. Considering that bs->backing_file usually at runtime contained the path to the image relative to qemu's CWD (or absolute), this patch changes QAPI's backing_file to always report the bs->backing->bs->filename from now on. If you want to receive the image header information, you have to refer to full-backing-filename. This necessitates a change to iotest 228. The interesting information it really wanted is the image header, and it can get that now, but it has to use full-backing-filename instead of backing_file. Because of this patch's changes to bs->backing_file's behavior, we also need some reference output changes. Along with the changes to bs->backing_file, stop updating BDS.backing_format in bdrv_backing_attach() as well. This way, ImageInfo's backing-filename and backing-filename-format fields will represent what the image header says and nothing else. iotest 245 changes in behavior: With the backing node no longer overriding the parent node's backing_file string, you can now omit the @backing option when reopening a node with neither a default nor a current backing file even if it used to have a backing node at some point. 273 also changes: The base image is opened without a format layer, so ImageInfo.backing-filename-format used to report "file" for the base image's overlay after blockdev-snapshot. However, the image header never says "file" anywhere, so it now reports $IMGFMT. Signed-off-by: Max Reitz <mreitz@redhat.com>
2018-08-01 20:34:11 +02:00
info->backing_filename = g_strdup(backing_filename);
backing_filename2 = bdrv_get_full_backing_filename(bs, NULL);
/* Always report the full_backing_filename if present, even if it's the
* same as backing_filename. That they are same is useful info. */
if (backing_filename2) {
info->full_backing_filename = g_strdup(backing_filename2);
}
if (bs->backing_format[0]) {
info->backing_filename_format = g_strdup(bs->backing_format);
}
g_free(backing_filename2);
}
ret = bdrv_query_snapshot_info_list(bs, &info->snapshots, &err);
switch (ret) {
case 0:
if (info->snapshots) {
info->has_snapshots = true;
}
break;
/* recoverable error */
case -ENOMEDIUM:
case -ENOTSUP:
error_free(err);
break;
default:
error_propagate(errp, err);
goto out;
}
out:
aio_context_release(bdrv_get_aio_context(bs));
}
/**
* bdrv_query_block_node_info:
* @bs: block node to examine
* @p_info: location to store node information
* @errp: location to store error information
*
* Store image information about @bs in @p_info.
*
* @p_info will be set only on success. On error, store error in @errp.
*/
void bdrv_query_block_node_info(BlockDriverState *bs,
BlockNodeInfo **p_info,
Error **errp)
{
BlockNodeInfo *info;
ERRP_GUARD();
info = g_new0(BlockNodeInfo, 1);
bdrv_do_query_node_info(bs, info, errp);
if (*errp) {
qapi_free_BlockNodeInfo(info);
return;
}
*p_info = info;
}
/**
* bdrv_query_image_info:
* @bs: block node to examine
* @p_info: location to store image information
* @flat: skip backing node information
* @skip_implicit_filters: skip implicit filters in the backing chain
* @errp: location to store error information
*
* Store image information in @p_info, potentially recursively covering the
* backing chain.
*
* If @flat is true, do not query backing image information, i.e.
* (*p_info)->has_backing_image will be set to false and
* (*p_info)->backing_image to NULL even when the image does in fact have a
* backing image.
*
* If @skip_implicit_filters is true, implicit filter nodes in the backing chain
* will be skipped when querying backing image information.
* (@skip_implicit_filters is ignored when @flat is true.)
*
* @p_info will be set only on success. On error, store error in @errp.
*/
void bdrv_query_image_info(BlockDriverState *bs,
ImageInfo **p_info,
bool flat,
bool skip_implicit_filters,
Error **errp)
{
ImageInfo *info;
ERRP_GUARD();
info = g_new0(ImageInfo, 1);
bdrv_do_query_node_info(bs, qapi_ImageInfo_base(info), errp);
if (*errp) {
goto fail;
}
if (!flat) {
BlockDriverState *backing;
/*
* Use any filtered child here (for backwards compatibility to when
* we always took bs->backing, which might be any filtered child).
*/
backing = bdrv_filter_or_cow_bs(bs);
if (skip_implicit_filters) {
backing = bdrv_skip_implicit_filters(backing);
}
if (backing) {
bdrv_query_image_info(backing, &info->backing_image, false,
skip_implicit_filters, errp);
if (*errp) {
goto fail;
}
}
}
*p_info = info;
return;
fail:
assert(*errp);
qapi_free_ImageInfo(info);
}
/**
* bdrv_query_block_graph_info:
* @bs: root node to start from
* @p_info: location to store image information
* @errp: location to store error information
*
* Store image information about the graph starting from @bs in @p_info.
*
* @p_info will be set only on success. On error, store error in @errp.
*/
void bdrv_query_block_graph_info(BlockDriverState *bs,
BlockGraphInfo **p_info,
Error **errp)
{
BlockGraphInfo *info;
BlockChildInfoList **children_list_tail;
BdrvChild *c;
ERRP_GUARD();
info = g_new0(BlockGraphInfo, 1);
bdrv_do_query_node_info(bs, qapi_BlockGraphInfo_base(info), errp);
if (*errp) {
goto fail;
}
children_list_tail = &info->children;
QLIST_FOREACH(c, &bs->children, next) {
BlockChildInfo *c_info;
c_info = g_new0(BlockChildInfo, 1);
QAPI_LIST_APPEND(children_list_tail, c_info);
c_info->name = g_strdup(c->name);
bdrv_query_block_graph_info(c->bs, &c_info->info, errp);
if (*errp) {
goto fail;
}
}
*p_info = info;
return;
fail:
assert(*errp != NULL);
qapi_free_BlockGraphInfo(info);
}
/* @p_info will be set only on success. */
static void bdrv_query_info(BlockBackend *blk, BlockInfo **p_info,
Error **errp)
{
BlockInfo *info = g_malloc0(sizeof(*info));
BlockDriverState *bs = blk_bs(blk);
char *qdev;
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
/* Skip automatically inserted nodes that the user isn't aware of */
bs = bdrv_skip_implicit_filters(bs);
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
info->device = g_strdup(blk_name(blk));
info->type = g_strdup("unknown");
info->locked = blk_dev_is_medium_locked(blk);
info->removable = blk_dev_has_removable_media(blk);
qdev = blk_get_attached_dev_id(blk);
if (qdev && *qdev) {
info->qdev = qdev;
} else {
g_free(qdev);
}
if (blk_dev_has_tray(blk)) {
info->has_tray_open = true;
info->tray_open = blk_dev_is_tray_open(blk);
}
if (blk_iostatus_is_enabled(blk)) {
info->has_io_status = true;
info->io_status = blk_iostatus(blk);
}
if (bs && bs->drv) {
info->inserted = bdrv_block_device_info(blk, bs, false, errp);
if (info->inserted == NULL) {
goto err;
}
}
*p_info = info;
return;
err:
qapi_free_BlockInfo(info);
}
static uint64List *uint64_list(uint64_t *list, int size)
{
int i;
uint64List *out_list = NULL;
uint64List **tail = &out_list;
for (i = 0; i < size; i++) {
QAPI_LIST_APPEND(tail, list[i]);
}
return out_list;
}
static BlockLatencyHistogramInfo *
bdrv_latency_histogram_stats(BlockLatencyHistogram *hist)
{
BlockLatencyHistogramInfo *info;
if (!hist->bins) {
return NULL;
}
info = g_new0(BlockLatencyHistogramInfo, 1);
info->boundaries = uint64_list(hist->boundaries, hist->nbins - 1);
info->bins = uint64_list(hist->bins, hist->nbins);
return info;
}
static void bdrv_query_blk_stats(BlockDeviceStats *ds, BlockBackend *blk)
{
BlockAcctStats *stats = blk_get_stats(blk);
BlockAcctTimedStats *ts = NULL;
BlockLatencyHistogram *hgram;
ds->rd_bytes = stats->nr_bytes[BLOCK_ACCT_READ];
ds->wr_bytes = stats->nr_bytes[BLOCK_ACCT_WRITE];
ds->unmap_bytes = stats->nr_bytes[BLOCK_ACCT_UNMAP];
ds->rd_operations = stats->nr_ops[BLOCK_ACCT_READ];
ds->wr_operations = stats->nr_ops[BLOCK_ACCT_WRITE];
ds->unmap_operations = stats->nr_ops[BLOCK_ACCT_UNMAP];
ds->failed_rd_operations = stats->failed_ops[BLOCK_ACCT_READ];
ds->failed_wr_operations = stats->failed_ops[BLOCK_ACCT_WRITE];
ds->failed_flush_operations = stats->failed_ops[BLOCK_ACCT_FLUSH];
ds->failed_unmap_operations = stats->failed_ops[BLOCK_ACCT_UNMAP];
ds->invalid_rd_operations = stats->invalid_ops[BLOCK_ACCT_READ];
ds->invalid_wr_operations = stats->invalid_ops[BLOCK_ACCT_WRITE];
ds->invalid_flush_operations =
stats->invalid_ops[BLOCK_ACCT_FLUSH];
ds->invalid_unmap_operations = stats->invalid_ops[BLOCK_ACCT_UNMAP];
ds->rd_merged = stats->merged[BLOCK_ACCT_READ];
ds->wr_merged = stats->merged[BLOCK_ACCT_WRITE];
ds->unmap_merged = stats->merged[BLOCK_ACCT_UNMAP];
ds->flush_operations = stats->nr_ops[BLOCK_ACCT_FLUSH];
ds->wr_total_time_ns = stats->total_time_ns[BLOCK_ACCT_WRITE];
ds->rd_total_time_ns = stats->total_time_ns[BLOCK_ACCT_READ];
ds->flush_total_time_ns = stats->total_time_ns[BLOCK_ACCT_FLUSH];
ds->unmap_total_time_ns = stats->total_time_ns[BLOCK_ACCT_UNMAP];
ds->has_idle_time_ns = stats->last_access_time_ns > 0;
if (ds->has_idle_time_ns) {
ds->idle_time_ns = block_acct_idle_time_ns(stats);
}
ds->account_invalid = stats->account_invalid;
ds->account_failed = stats->account_failed;
while ((ts = block_acct_interval_next(stats, ts))) {
BlockDeviceTimedStats *dev_stats = g_malloc0(sizeof(*dev_stats));
TimedAverage *rd = &ts->latency[BLOCK_ACCT_READ];
TimedAverage *wr = &ts->latency[BLOCK_ACCT_WRITE];
TimedAverage *fl = &ts->latency[BLOCK_ACCT_FLUSH];
dev_stats->interval_length = ts->interval_length;
dev_stats->min_rd_latency_ns = timed_average_min(rd);
dev_stats->max_rd_latency_ns = timed_average_max(rd);
dev_stats->avg_rd_latency_ns = timed_average_avg(rd);
dev_stats->min_wr_latency_ns = timed_average_min(wr);
dev_stats->max_wr_latency_ns = timed_average_max(wr);
dev_stats->avg_wr_latency_ns = timed_average_avg(wr);
dev_stats->min_flush_latency_ns = timed_average_min(fl);
dev_stats->max_flush_latency_ns = timed_average_max(fl);
dev_stats->avg_flush_latency_ns = timed_average_avg(fl);
dev_stats->avg_rd_queue_depth =
block_acct_queue_depth(ts, BLOCK_ACCT_READ);
dev_stats->avg_wr_queue_depth =
block_acct_queue_depth(ts, BLOCK_ACCT_WRITE);
QAPI_LIST_PREPEND(ds->timed_stats, dev_stats);
}
hgram = stats->latency_histogram;
ds->rd_latency_histogram
= bdrv_latency_histogram_stats(&hgram[BLOCK_ACCT_READ]);
ds->wr_latency_histogram
= bdrv_latency_histogram_stats(&hgram[BLOCK_ACCT_WRITE]);
ds->flush_latency_histogram
= bdrv_latency_histogram_stats(&hgram[BLOCK_ACCT_FLUSH]);
}
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
static BlockStats *bdrv_query_bds_stats(BlockDriverState *bs,
bool blk_level)
{
BdrvChild *parent_child;
BlockDriverState *filter_or_cow_bs;
BlockStats *s = NULL;
s = g_malloc0(sizeof(*s));
s->stats = g_malloc0(sizeof(*s->stats));
if (!bs) {
return s;
}
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
/* Skip automatically inserted nodes that the user isn't aware of in
* a BlockBackend-level command. Stay at the exact node for a node-level
* command. */
if (blk_level) {
bs = bdrv_skip_implicit_filters(bs);
block: Skip implicit nodes in query-block/blockstats Commits 0db832f and 6cdbceb introduced the automatic insertion of filter nodes above the top layer of mirror and commit block jobs. The assumption made there was that since libvirt doesn't do node-level management of the block layer yet, it shouldn't be affected by added nodes. This is true as far as commands issued by libvirt are concerned. It only uses BlockBackend names to address nodes, so any operations it performs still operate on the root of the tree as intended. However, the assumption breaks down when you consider query commands, which return data for the wrong node now. These commands also return information on some child nodes (bs->file and/or bs->backing), which libvirt does make use of, and which refer to the wrong nodes, too. One of the consequences is that oVirt gets wrong information about the image size and stops the VM in response as long as a mirror or commit job is running: https://bugzilla.redhat.com/show_bug.cgi?id=1470634 This patch fixes the problem by hiding the implicit nodes created automatically by the mirror and commit block jobs in the output of query-block and BlockBackend-based query-blockstats as long as the user doesn't indicate that they are aware of those nodes by providing a node name for them in the QMP command to start the block job. The node-based commands query-named-block-nodes and query-blockstats with query-nodes=true still show all nodes, including implicit ones. This ensures that users that are capable of node-level management can still access the full information; users that only know BlockBackends won't use these commands. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Tested-by: Eric Blake <eblake@redhat.com>
2017-07-18 17:24:05 +02:00
}
if (bdrv_get_node_name(bs)[0]) {
s->node_name = g_strdup(bdrv_get_node_name(bs));
}
s->stats->wr_highest_offset = stat64_get(&bs->wr_highest_offset);
s->driver_specific = bdrv_get_specific_stats(bs);
parent_child = bdrv_primary_child(bs);
if (!parent_child ||
!(parent_child->role & (BDRV_CHILD_DATA | BDRV_CHILD_FILTERED)))
{
BdrvChild *c;
/*
* Look for a unique data-storing child. We do not need to look for
* filtered children, as there would be only one and it would have been
* the primary child.
*/
parent_child = NULL;
QLIST_FOREACH(c, &bs->children, next) {
if (c->role & BDRV_CHILD_DATA) {
if (parent_child) {
/*
* There are multiple data-storing children and we cannot
* choose between them.
*/
parent_child = NULL;
break;
}
parent_child = c;
}
}
}
if (parent_child) {
s->parent = bdrv_query_bds_stats(parent_child->bs, blk_level);
}
filter_or_cow_bs = bdrv_filter_or_cow_bs(bs);
if (blk_level && filter_or_cow_bs) {
/*
* Put any filtered or COW child here (for backwards
* compatibility to when we put bs0->backing here, which might
* be either)
*/
s->backing = bdrv_query_bds_stats(filter_or_cow_bs, blk_level);
}
return s;
}
BlockInfoList *qmp_query_block(Error **errp)
{
BlockInfoList *head = NULL, **p_next = &head;
BlockBackend *blk;
Error *local_err = NULL;
for (blk = blk_all_next(NULL); blk; blk = blk_all_next(blk)) {
BlockInfoList *info;
if (!*blk_name(blk) && !blk_get_attached_dev(blk)) {
continue;
}
info = g_malloc0(sizeof(*info));
bdrv_query_info(blk, &info->value, &local_err);
if (local_err) {
error_propagate(errp, local_err);
g_free(info);
qapi_free_BlockInfoList(head);
return NULL;
}
*p_next = info;
p_next = &info->next;
}
return head;
}
BlockStatsList *qmp_query_blockstats(bool has_query_nodes,
bool query_nodes,
Error **errp)
{
BlockStatsList *head = NULL, **tail = &head;
BlockBackend *blk;
BlockDriverState *bs;
/* Just to be safe if query_nodes is not always initialized */
if (has_query_nodes && query_nodes) {
for (bs = bdrv_next_node(NULL); bs; bs = bdrv_next_node(bs)) {
AioContext *ctx = bdrv_get_aio_context(bs);
aio_context_acquire(ctx);
QAPI_LIST_APPEND(tail, bdrv_query_bds_stats(bs, false));
aio_context_release(ctx);
}
} else {
for (blk = blk_all_next(NULL); blk; blk = blk_all_next(blk)) {
AioContext *ctx = blk_get_aio_context(blk);
BlockStats *s;
char *qdev;
if (!*blk_name(blk) && !blk_get_attached_dev(blk)) {
continue;
}
aio_context_acquire(ctx);
s = bdrv_query_bds_stats(blk_bs(blk), true);
s->device = g_strdup(blk_name(blk));
qdev = blk_get_attached_dev_id(blk);
if (qdev && *qdev) {
s->qdev = qdev;
} else {
g_free(qdev);
}
bdrv_query_blk_stats(s->stats, blk);
aio_context_release(ctx);
QAPI_LIST_APPEND(tail, s);
}
}
return head;
}
void bdrv_snapshot_dump(QEMUSnapshotInfo *sn)
{
char clock_buf[128];
char icount_buf[128] = {0};
int64_t secs;
char *sizing = NULL;
if (!sn) {
qemu_printf("%-10s%-17s%8s%20s%13s%11s",
"ID", "TAG", "VM SIZE", "DATE", "VM CLOCK", "ICOUNT");
} else {
g_autoptr(GDateTime) date = g_date_time_new_from_unix_local(sn->date_sec);
g_autofree char *date_buf = g_date_time_format(date, "%Y-%m-%d %H:%M:%S");
secs = sn->vm_clock_nsec / 1000000000;
snprintf(clock_buf, sizeof(clock_buf),
"%02d:%02d:%02d.%03d",
(int)(secs / 3600),
(int)((secs / 60) % 60),
(int)(secs % 60),
(int)((sn->vm_clock_nsec / 1000000) % 1000));
sizing = size_to_str(sn->vm_state_size);
if (sn->icount != -1ULL) {
snprintf(icount_buf, sizeof(icount_buf),
"%"PRId64, sn->icount);
}
qemu_printf("%-9s %-16s %8s%20s%13s%11s",
sn->id_str, sn->name,
sizing,
date_buf,
clock_buf,
icount_buf);
}
g_free(sizing);
}
static void dump_qdict(int indentation, QDict *dict);
static void dump_qlist(int indentation, QList *list);
static void dump_qobject(int comp_indent, QObject *obj)
{
switch (qobject_type(obj)) {
case QTYPE_QNUM: {
QNum *value = qobject_to(QNum, obj);
char *tmp = qnum_to_string(value);
qemu_printf("%s", tmp);
g_free(tmp);
break;
}
case QTYPE_QSTRING: {
QString *value = qobject_to(QString, obj);
qemu_printf("%s", qstring_get_str(value));
break;
}
case QTYPE_QDICT: {
QDict *value = qobject_to(QDict, obj);
dump_qdict(comp_indent, value);
break;
}
case QTYPE_QLIST: {
QList *value = qobject_to(QList, obj);
dump_qlist(comp_indent, value);
break;
}
case QTYPE_QBOOL: {
QBool *value = qobject_to(QBool, obj);
qemu_printf("%s", qbool_get_bool(value) ? "true" : "false");
break;
}
default:
abort();
}
}
static void dump_qlist(int indentation, QList *list)
{
const QListEntry *entry;
int i = 0;
for (entry = qlist_first(list); entry; entry = qlist_next(entry), i++) {
QType type = qobject_type(entry->value);
bool composite = (type == QTYPE_QDICT || type == QTYPE_QLIST);
qemu_printf("%*s[%i]:%c", indentation * 4, "", i,
composite ? '\n' : ' ');
dump_qobject(indentation + 1, entry->value);
if (!composite) {
qemu_printf("\n");
}
}
}
static void dump_qdict(int indentation, QDict *dict)
{
const QDictEntry *entry;
for (entry = qdict_first(dict); entry; entry = qdict_next(dict, entry)) {
QType type = qobject_type(entry->value);
bool composite = (type == QTYPE_QDICT || type == QTYPE_QLIST);
char *key = g_malloc(strlen(entry->key) + 1);
int i;
/* replace dashes with spaces in key (variable) names */
for (i = 0; entry->key[i]; i++) {
key[i] = entry->key[i] == '-' ? ' ' : entry->key[i];
}
key[i] = 0;
qemu_printf("%*s%s:%c", indentation * 4, "", key,
composite ? '\n' : ' ');
dump_qobject(indentation + 1, entry->value);
if (!composite) {
qemu_printf("\n");
}
g_free(key);
}
}
/*
* Return whether dumping the given QObject with dump_qobject() would
* yield an empty dump, i.e. not print anything.
*/
static bool qobject_is_empty_dump(const QObject *obj)
{
switch (qobject_type(obj)) {
case QTYPE_QNUM:
case QTYPE_QSTRING:
case QTYPE_QBOOL:
return false;
case QTYPE_QDICT:
return qdict_size(qobject_to(QDict, obj)) == 0;
case QTYPE_QLIST:
return qlist_empty(qobject_to(QList, obj));
default:
abort();
}
}
/**
* Dumps the given ImageInfoSpecific object in a human-readable form,
* prepending an optional prefix if the dump is not empty.
*/
void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec,
const char *prefix,
int indentation)
{
QObject *obj, *data;
Visitor *v = qobject_output_visitor_new(&obj);
qapi: Add new visit_complete() function Making each output visitor provide its own output collection function was the only remaining reason for exposing visitor sub-types to the rest of the code base. Add a polymorphic visit_complete() function which is a no-op for input visitors, and which populates an opaque pointer for output visitors. For maximum type-safety, also add a parameter to the output visitor constructors with a type-correct version of the output pointer, and assert that the two uses match. This approach was considered superior to either passing the output parameter only during construction (action at a distance during visit_free() feels awkward) or only during visit_complete() (defeating type safety makes it easier to use incorrectly). Most callers were function-local, and therefore a mechanical conversion; the testsuite was a bit trickier, but the previous cleanup patch minimized the churn here. The visit_complete() function may be called at most once; doing so lets us use transfer semantics rather than duplication or ref-count semantics to get the just-built output back to the caller, even though it means our behavior is not idempotent. Generated code is simplified as follows for events: |@@ -26,7 +26,7 @@ void qapi_event_send_acpi_device_ost(ACP | QDict *qmp; | Error *err = NULL; | QMPEventFuncEmit emit; |- QmpOutputVisitor *qov; |+ QObject *obj; | Visitor *v; | q_obj_ACPI_DEVICE_OST_arg param = { | info |@@ -39,8 +39,7 @@ void qapi_event_send_acpi_device_ost(ACP | | qmp = qmp_event_build_dict("ACPI_DEVICE_OST"); | |- qov = qmp_output_visitor_new(); |- v = qmp_output_get_visitor(qov); |+ v = qmp_output_visitor_new(&obj); | | visit_start_struct(v, "ACPI_DEVICE_OST", NULL, 0, &err); | if (err) { |@@ -55,7 +54,8 @@ void qapi_event_send_acpi_device_ost(ACP | goto out; | } | |- qdict_put_obj(qmp, "data", qmp_output_get_qobject(qov)); |+ visit_complete(v, &obj); |+ qdict_put_obj(qmp, "data", obj); | emit(QAPI_EVENT_ACPI_DEVICE_OST, qmp, &err); and for commands: | { | Error *err = NULL; |- QmpOutputVisitor *qov = qmp_output_visitor_new(); | Visitor *v; | |- v = qmp_output_get_visitor(qov); |+ v = qmp_output_visitor_new(ret_out); | visit_type_AddfdInfo(v, "unused", &ret_in, &err); |- if (err) { |- goto out; |+ if (!err) { |+ visit_complete(v, ret_out); | } |- *ret_out = qmp_output_get_qobject(qov); |- |-out: | error_propagate(errp, err); Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1465490926-28625-13-git-send-email-eblake@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-06-09 18:48:43 +02:00
visit_type_ImageInfoSpecific(v, NULL, &info_spec, &error_abort);
visit_complete(v, &obj);
data = qdict_get(qobject_to(QDict, obj), "data");
if (!qobject_is_empty_dump(data)) {
if (prefix) {
qemu_printf("%*s%s", indentation * 4, "", prefix);
}
dump_qobject(indentation + 1, data);
}
qobject_unref(obj);
qapi: Add new visit_complete() function Making each output visitor provide its own output collection function was the only remaining reason for exposing visitor sub-types to the rest of the code base. Add a polymorphic visit_complete() function which is a no-op for input visitors, and which populates an opaque pointer for output visitors. For maximum type-safety, also add a parameter to the output visitor constructors with a type-correct version of the output pointer, and assert that the two uses match. This approach was considered superior to either passing the output parameter only during construction (action at a distance during visit_free() feels awkward) or only during visit_complete() (defeating type safety makes it easier to use incorrectly). Most callers were function-local, and therefore a mechanical conversion; the testsuite was a bit trickier, but the previous cleanup patch minimized the churn here. The visit_complete() function may be called at most once; doing so lets us use transfer semantics rather than duplication or ref-count semantics to get the just-built output back to the caller, even though it means our behavior is not idempotent. Generated code is simplified as follows for events: |@@ -26,7 +26,7 @@ void qapi_event_send_acpi_device_ost(ACP | QDict *qmp; | Error *err = NULL; | QMPEventFuncEmit emit; |- QmpOutputVisitor *qov; |+ QObject *obj; | Visitor *v; | q_obj_ACPI_DEVICE_OST_arg param = { | info |@@ -39,8 +39,7 @@ void qapi_event_send_acpi_device_ost(ACP | | qmp = qmp_event_build_dict("ACPI_DEVICE_OST"); | |- qov = qmp_output_visitor_new(); |- v = qmp_output_get_visitor(qov); |+ v = qmp_output_visitor_new(&obj); | | visit_start_struct(v, "ACPI_DEVICE_OST", NULL, 0, &err); | if (err) { |@@ -55,7 +54,8 @@ void qapi_event_send_acpi_device_ost(ACP | goto out; | } | |- qdict_put_obj(qmp, "data", qmp_output_get_qobject(qov)); |+ visit_complete(v, &obj); |+ qdict_put_obj(qmp, "data", obj); | emit(QAPI_EVENT_ACPI_DEVICE_OST, qmp, &err); and for commands: | { | Error *err = NULL; |- QmpOutputVisitor *qov = qmp_output_visitor_new(); | Visitor *v; | |- v = qmp_output_get_visitor(qov); |+ v = qmp_output_visitor_new(ret_out); | visit_type_AddfdInfo(v, "unused", &ret_in, &err); |- if (err) { |- goto out; |+ if (!err) { |+ visit_complete(v, ret_out); | } |- *ret_out = qmp_output_get_qobject(qov); |- |-out: | error_propagate(errp, err); Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1465490926-28625-13-git-send-email-eblake@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-06-09 18:48:43 +02:00
visit_free(v);
}
qemu-img: Change info key names for protocol nodes Currently, when querying a qcow2 image, qemu-img info reports something like this: image: test.qcow2 file format: qcow2 virtual size: 64 MiB (67108864 bytes) disk size: 196 KiB cluster_size: 65536 Format specific information: compat: 1.1 compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: false extended l2: false Child node '/file': image: test.qcow2 file format: file virtual size: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 Notably, the way the keys are named is specific for image files: The filename is shown under "image", the BDS driver under "file format", and the BDS length under "virtual size". This does not make much sense for nodes that are not actually supposed to be guest images, like the /file child node shown above. Give bdrv_node_info_dump() a @protocol parameter that gives a hint that the respective node is probably just used for data storage and does not necessarily present the data for a VM guest disk. This renames the keys so that with this patch, the output becomes: image: test.qcow2 [...] Child node '/file': filename: test.qcow2 protocol type: file file length: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 (Perhaps we should also rename "Format specific information", but I could not come up with anything better that will not become problematic if we guess wrong with the protocol "heuristic".) This change affects iotest 302, which has protocol node information in its reference output. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-13-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2022-06-20 18:27:04 +02:00
/**
* Print the given @info object in human-readable form. Every field is indented
* using the given @indentation (four spaces per indentation level).
*
* When using this to print a whole block graph, @protocol can be set to true to
* signify that the given information is associated with a protocol node, i.e.
* just data storage for an image, such that the data it presents is not really
* a full VM disk. If so, several fields change name: For example, "virtual
* size" is printed as "file length".
* (Consider a qcow2 image, which is represented by a qcow2 node and a file
* node. Printing a "virtual size" for the file node does not make sense,
* because without the qcow2 node, it is not really a guest disk, so it does not
* have a "virtual size". Therefore, we call it "file length" instead.)
*
* @protocol is ignored when @indentation is 0, because we take that to mean
* that the associated node is the root node in the queried block graph, and
* thus is always to be interpreted as a standalone guest disk.
*/
void bdrv_node_info_dump(BlockNodeInfo *info, int indentation, bool protocol)
{
char *size_buf, *dsize_buf;
g_autofree char *ind_s = g_strdup_printf("%*s", indentation * 4, "");
qemu-img: Change info key names for protocol nodes Currently, when querying a qcow2 image, qemu-img info reports something like this: image: test.qcow2 file format: qcow2 virtual size: 64 MiB (67108864 bytes) disk size: 196 KiB cluster_size: 65536 Format specific information: compat: 1.1 compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: false extended l2: false Child node '/file': image: test.qcow2 file format: file virtual size: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 Notably, the way the keys are named is specific for image files: The filename is shown under "image", the BDS driver under "file format", and the BDS length under "virtual size". This does not make much sense for nodes that are not actually supposed to be guest images, like the /file child node shown above. Give bdrv_node_info_dump() a @protocol parameter that gives a hint that the respective node is probably just used for data storage and does not necessarily present the data for a VM guest disk. This renames the keys so that with this patch, the output becomes: image: test.qcow2 [...] Child node '/file': filename: test.qcow2 protocol type: file file length: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 (Perhaps we should also rename "Format specific information", but I could not come up with anything better that will not become problematic if we guess wrong with the protocol "heuristic".) This change affects iotest 302, which has protocol node information in its reference output. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-13-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2022-06-20 18:27:04 +02:00
if (indentation == 0) {
/* Top level, consider this a normal image */
protocol = false;
}
if (!info->has_actual_size) {
dsize_buf = g_strdup("unavailable");
} else {
dsize_buf = size_to_str(info->actual_size);
}
size_buf = size_to_str(info->virtual_size);
qemu-img: Change info key names for protocol nodes Currently, when querying a qcow2 image, qemu-img info reports something like this: image: test.qcow2 file format: qcow2 virtual size: 64 MiB (67108864 bytes) disk size: 196 KiB cluster_size: 65536 Format specific information: compat: 1.1 compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: false extended l2: false Child node '/file': image: test.qcow2 file format: file virtual size: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 Notably, the way the keys are named is specific for image files: The filename is shown under "image", the BDS driver under "file format", and the BDS length under "virtual size". This does not make much sense for nodes that are not actually supposed to be guest images, like the /file child node shown above. Give bdrv_node_info_dump() a @protocol parameter that gives a hint that the respective node is probably just used for data storage and does not necessarily present the data for a VM guest disk. This renames the keys so that with this patch, the output becomes: image: test.qcow2 [...] Child node '/file': filename: test.qcow2 protocol type: file file length: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 (Perhaps we should also rename "Format specific information", but I could not come up with anything better that will not become problematic if we guess wrong with the protocol "heuristic".) This change affects iotest 302, which has protocol node information in its reference output. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-13-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2022-06-20 18:27:04 +02:00
qemu_printf("%s%s: %s\n"
"%s%s: %s\n"
"%s%s: %s (%" PRId64 " bytes)\n"
"%sdisk size: %s\n",
qemu-img: Change info key names for protocol nodes Currently, when querying a qcow2 image, qemu-img info reports something like this: image: test.qcow2 file format: qcow2 virtual size: 64 MiB (67108864 bytes) disk size: 196 KiB cluster_size: 65536 Format specific information: compat: 1.1 compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: false extended l2: false Child node '/file': image: test.qcow2 file format: file virtual size: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 Notably, the way the keys are named is specific for image files: The filename is shown under "image", the BDS driver under "file format", and the BDS length under "virtual size". This does not make much sense for nodes that are not actually supposed to be guest images, like the /file child node shown above. Give bdrv_node_info_dump() a @protocol parameter that gives a hint that the respective node is probably just used for data storage and does not necessarily present the data for a VM guest disk. This renames the keys so that with this patch, the output becomes: image: test.qcow2 [...] Child node '/file': filename: test.qcow2 protocol type: file file length: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 (Perhaps we should also rename "Format specific information", but I could not come up with anything better that will not become problematic if we guess wrong with the protocol "heuristic".) This change affects iotest 302, which has protocol node information in its reference output. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-13-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2022-06-20 18:27:04 +02:00
ind_s, protocol ? "filename" : "image", info->filename,
ind_s, protocol ? "protocol type" : "file format",
info->format,
ind_s, protocol ? "file length" : "virtual size",
size_buf, info->virtual_size,
ind_s, dsize_buf);
g_free(size_buf);
g_free(dsize_buf);
if (info->has_encrypted && info->encrypted) {
qemu_printf("%sencrypted: yes\n", ind_s);
}
if (info->has_cluster_size) {
qemu_printf("%scluster_size: %" PRId64 "\n",
ind_s, info->cluster_size);
}
if (info->has_dirty_flag && info->dirty_flag) {
qemu_printf("%scleanly shut down: no\n", ind_s);
}
if (info->backing_filename) {
qemu_printf("%sbacking file: %s", ind_s, info->backing_filename);
if (!info->full_backing_filename) {
qemu_printf(" (cannot determine actual path)");
} else if (strcmp(info->backing_filename,
info->full_backing_filename) != 0) {
qemu_printf(" (actual path: %s)", info->full_backing_filename);
}
qemu_printf("\n");
if (info->backing_filename_format) {
qemu_printf("%sbacking file format: %s\n",
ind_s, info->backing_filename_format);
}
}
if (info->has_snapshots) {
SnapshotInfoList *elem;
qemu_printf("%sSnapshot list:\n", ind_s);
qemu_printf("%s", ind_s);
bdrv_snapshot_dump(NULL);
qemu_printf("\n");
/* Ideally bdrv_snapshot_dump() would operate on SnapshotInfoList but
* we convert to the block layer's native QEMUSnapshotInfo for now.
*/
for (elem = info->snapshots; elem; elem = elem->next) {
QEMUSnapshotInfo sn = {
.vm_state_size = elem->value->vm_state_size,
.date_sec = elem->value->date_sec,
.date_nsec = elem->value->date_nsec,
.vm_clock_nsec = elem->value->vm_clock_sec * 1000000000ULL +
elem->value->vm_clock_nsec,
.icount = elem->value->has_icount ?
elem->value->icount : -1ULL,
};
pstrcpy(sn.id_str, sizeof(sn.id_str), elem->value->id);
pstrcpy(sn.name, sizeof(sn.name), elem->value->name);
qemu_printf("%s", ind_s);
bdrv_snapshot_dump(&sn);
qemu_printf("\n");
}
}
if (info->format_specific) {
bdrv_image_info_specific_dump(info->format_specific,
"Format specific information:\n",
indentation);
}
}