- 13 3月, 2017 7 次提交
-
-
由 Kevin Wolf 提交于
In some cases, bdrv_co_get_block_status() is called recursively for the whole backing chain. The automatically inserted bdrv_commit_top filter driver must not stop the recursion, so implement a callback that simply forwards the request to bs->backing. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Kevin Wolf 提交于
This fixes bdrv_co_get_block_status() for the bdrv_mirror_top block driver, which must fall through to bs->backing instead of bs->file. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Eric Blake 提交于
If bdrv_is_allocated() fails, we should react to that failure. For 2 of the 3 callers, reporting the error was easy. But in cluster_was_modified() and its lone caller get_cluster_count_for_direntry(), it's rather invasive to update the logic to pass the error back; so there, I went with merely documenting the issue by changing the return type to bool (in all likelihood, treating the cluster as modified will then trigger a read which will also fail, and eventually get to an error - but given the appalling number of abort() calls in this code, I'm not making it any worse). Signed-off-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Eric Blake 提交于
If bdrv_is_allocated() fails, we should immediately do the backup error action, rather than attempting backup_do_cow() (although that will likely fail too). Signed-off-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Eric Blake 提交于
The driver has failed to build since commit da34e65c, in qemu 2.6, due to a missing include of qapi/error.h for error_setg(). Since no one has complained in three releases, it is easier to remove the dead code than to keep it around, especially since it is not being built by default and therefore prone to bitrot. Signed-off-by: NEric Blake <eblake@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com> Reviewed-by: NFam Zheng <famz@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Fam Zheng 提交于
BlockLimits.max_transfer can be too high without this fix, guest will encounter I/O error or even get paused with werror=stop or rerror=stop. The cause is explained below. Linux has a separate limit, /sys/block/.../queue/max_segments, which in the worst case can be more restrictive than the BLKSECTGET which we already consider (note that they are two different things). So, the failure scenario before this patch is: 1) host device has max_sectors_kb = 4096 and max_segments = 64; 2) guest learns max_sectors_kb limit from QEMU, but doesn't know max_segments; 3) guest issues e.g. a 512KB request thinking it's okay, but actually it's not, because it will be passed through to host device as an SG_IO req that has niov > 64; 4) host kernel doesn't like the segmenting of the request, and returns -EINVAL; This patch checks the max_segments sysfs entry for the host device and calculates a "conservative" bytes limit using the page size, which is then merged into the existing max_transfer limit. Guest will discover this from the usual virtual block device interfaces. (In the case of scsi-generic, it will be done in the INQUIRY reply interception in device model.) The other possibility is to actually propagate it as a separate limit, but it's not better. On the one hand, there is a big complication: the limit is per-LUN in QEMU PoV (because we can attach LUNs from different host HBAs to the same virtio-scsi bus), but the channel to communicate it in a per-LUN manner is missing down the stack; on the other hand, two limits versus one doesn't change much about the valid size of I/O (because guest has no control over host segmenting). Also, the idea to fall back to bounce buffering in QEMU, upon -EINVAL, was explored. Unfortunately there is no neat way to ensure the bounce buffer is less segmented (in terms of DMA addr) than the guest buffer. Practically, this bug is not very common. It is only reported on a Emulex (lpfc), so it's okay to get it fixed in the easier way. Reviewed-by: NPaolo Bonzini <pbonzini@redhat.com> Signed-off-by: NFam Zheng <famz@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
Currently backup to nbd target is broken, as nbd doesn't have .bdrv_get_info realization. Signed-off-by: NVladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
- 07 3月, 2017 21 次提交
-
-
由 Fam Zheng 提交于
bdrv_set_backing_hd failure needn't be abort. Since we already have error parameter, use it. Signed-off-by: NFam Zheng <famz@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Fam Zheng 提交于
We have an errp and bdrv_root_attach_child can fail permission check, error_abort is not the best choice here. Signed-off-by: NFam Zheng <famz@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
QAPI type SocketAddressFlat differs from SocketAddress pointlessly: the discriminator value for variant InetSocketAddress is 'tcp' instead of 'inet'. Rename. The type is so far only used by the Gluster block drivers. Take care to keep 'tcp' working in things like -drive's file.server.0.type=tcp. The "gluster+tcp" URI scheme in pseudo-filenames stays the same. blockdev-add changes, but it has changed incompatibly since 2.8 already. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
As its documentation says, it's not specific to Gluster. Rename it, as I'm going to use it for something else. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
To reproduce, run $ valgrind qemu-system-x86_64 --nodefaults -S --drive driver=gluster,volume=testvol,path=/a/b/c,server.0.type=xxx Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NNiels de Vos <ndevos@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NNiels de Vos <ndevos@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
qemu_gluster_glfs_init() passes the names of QAPI enumeration type SocketTransport to glfs_set_volfile_server(). Works, because they were chosen to match. But the coupling is artificial. Use the appropriate literal strings instead. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NNiels de Vos <ndevos@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
This permits configuration with driver-specific options in addition to pseudo-filename parsed as URI. For instance, --drive driver=sheepdog,host=fido,vdi=dolly instead of --drive driver=sheepdog,file=sheepdog://fido/dolly It's also a first step towards supporting blockdev-add. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
sd_parse_uri() builds a string from host and port parts for inet_connect(). inet_connect() parses it into host, port and options. Whether this gets exactly the same host, port and no options for all inputs is not obvious. Cut out the string middleman and build a SocketAddress for socket_connect() instead. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NKevin Wolf <kwolf@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
Errors in the pseudo-filename are all reported with the same laconic "Can't parse filename" message. Add real error reporting, such as: $ qemu-system-x86_64 --drive driver=sheepdog,filename=sheepdog:/// qemu-system-x86_64: --drive driver=sheepdog,filename=sheepdog:///: missing file path in URI $ qemu-system-x86_64 --drive driver=sheepdog,filename=sheepgod:///vdi qemu-system-x86_64: --drive driver=sheepdog,filename=sheepgod:///vdi: URI scheme must be 'sheepdog', 'sheepdog+tcp', or 'sheepdog+unix' $ qemu-system-x86_64 --drive driver=sheepdog,filename=sheepdog+unix:///vdi?socke=sheepdog.sock qemu-system-x86_64: --drive driver=sheepdog,filename=sheepdog+unix:///vdi?socke=sheepdog.sock: unexpected query parameters The code to translate legacy syntax to URI fails to escape URI meta-characters. The new error messages are misleading then. Replace them by the old "Can't parse filename" message. "Internal error" would be more honest. Anyway, no worse than before. Also add a FIXME comment. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: NKevin Wolf <kwolf@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
sd_parse_uri() truncates long VDI names silently. Reject them instead. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
sd_parse_uri() and sd_snapshot_goto() screw up error checking after strtoul(), and truncate long tag names silently. Fix by replacing those parts by new sd_parse_snapid_or_tag(), which checks more carefully. sd_snapshot_delete() also parses snapshot IDs, but is currently too broken for me to touch. Mark TODO. Two calls of strtol() without error checking remain in parse_redundancy(). Mark them FIXME. More silent truncation of configuration strings remains elsewhere. Not marked. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
sd_snapshot_delete() should delete the snapshot whose ID matches @snapshot_id and whose name matches @name. But that's not what it does. If @snapshot_id is a valid ID, it deletes the snapshot with that ID, else it deletes the snapshot with that name. It doesn't use @name at all. Add suitable FIXME comments, so someone who actually knows Sheepdog can fix it. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
As a bdrv_create() method, sd_create() must set an error and return negative errno on failure. It prints the error instead of setting it when connect_to_sdog() fails. Fix that. While there, return the value of connect_to_sdog() like we do elsewhere, instead of -EIO. No functional change, as connect_to_sdog() returns no other error code. Many more suspicious uses of error_report() and error_report_err() remain in other functions. Left for another day. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: NKevin Wolf <kwolf@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
As a bdrv_snapshot_delete() method, sd_snapshot_delete() must set an error and return negative errno on failure. It sometimes returns -1, and sometimes neglects to set an error. It also prints error messages with error_report(). Fix all that. Moreover, its handling of an attempt to delete a nonexistent snapshot is wrong: it error_report()s and succeeds. Fix it to set an error and return -ENOENT instead. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Markus Armbruster 提交于
When qemu_opts_absorb_qdict() fails, sd_open() closes stdin, because sd->fd is still zero. Fortunately, qemu_opts_absorb_qdict() can't fail, because: 1. it only fails when qemu_opt_parse() fails, and 2. the only member of runtime_opts.desc[] is a QEMU_OPT_STRING, and 3. qemu_opt_parse() can't fail for QEMU_OPT_STRING. Defuse this ticking time bomb by jumping behind the file descriptor cleanup on error. Also do that for the error paths where sd->fd is still -1. The file descriptor cleanup happens to do nothing then, but let's not rely on that here. While there, rename label out to err, because it's on the error path, not the normal path out of the function. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Signed-off-by: NKevin Wolf <kwolf@redhat.com>
-
由 Kevin Wolf 提交于
When adding an Error parameter, bdrv_replace_in_backing_chain() would become nothing more than a wrapper around change_parent_backing_link(). So make the latter public, renamed as bdrv_replace_node(), and remove bdrv_replace_in_backing_chain(). Most of the callers just remove a node from the graph that they just inserted, so they can use &error_abort, but completion of a mirror job with 'replaces' set can actually fail. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NFam Zheng <famz@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Kevin Wolf 提交于
mirror_top_bs must be removed from the graph again when creating the dirty bitmap fails. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NFam Zheng <famz@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Kevin Wolf 提交于
mirror_top_bs takes write permissions on its backing file, which can make it impossible to attach that backing file node to another parent. However, this is exactly what needs to be done in order to remove mirror_top_bs from the backing chain. So give up the write permission first. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NFam Zheng <famz@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Kevin Wolf 提交于
The 'replaces' option of drive-mirror can be used to mirror a Quorum node to a new image and then let the target image replace one of the Quorum children. In order for this graph modification to succeed, the mirror job needs to lift its restrictions on the target node first before actually replacing the child. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NFam Zheng <famz@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com>
-
由 Kevin Wolf 提交于
Apparently some kind of mismerge happened in commit 8dfba279, which broke the error handling without any real reason by removing the assignment of the return value to ret in a blk_insert_bs() call. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NFam Zheng <famz@redhat.com> Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
-
- 05 3月, 2017 1 次提交
-
-
由 Markus Armbruster 提交于
The split between tests/test-qobject-input-visitor.c and tests/test-qobject-input-strict.c now makes less sense than ever. The next commit will take care of that. Signed-off-by: NMarkus Armbruster <armbru@redhat.com> Reviewed-by: NEric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-20-git-send-email-armbru@redhat.com>
-
- 03 3月, 2017 1 次提交
-
-
由 Paolo Bonzini 提交于
Reported by Coverity. Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
-
- 02 3月, 2017 1 次提交
-
-
由 Jeff Cody 提交于
This adds support for three additional options that may be specified by QAPI in blockdev-add: server: host, port auth method: either 'cephx' or 'none' The "server" and "auth-supported" QAPI parameters are arrays. To conform with the rados API, the array items are join as a single string with a ';' character as a delimiter when setting the configuration values. Reviewed-by: NMarkus Armbruster <armbru@redhat.com> Signed-off-by: NJeff Cody <jcody@redhat.com>
-
- 01 3月, 2017 9 次提交
-
-
由 Kevin Wolf 提交于
Aborting on error in bdrv_append() isn't correct. This patch fixes it and lets the callers handle failures. Test case 085 needs a reference output update. This is caused by the reversed order of bdrv_set_backing_hd() and change_parent_backing_link() in bdrv_append(): When the backing file of the new node is set, the parent nodes are still pointing to the old top, so the backing blocker is now initialised with the node name rather than the BlockBackend name. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Acked-by: NFam Zheng <famz@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com>
-
由 Kevin Wolf 提交于
Not all callers of bdrv_set_backing_hd() know for sure that attaching the backing file will be allowed by the permission system. Return the error from the function rather than aborting. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Acked-by: NFam Zheng <famz@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com>
-
由 Kevin Wolf 提交于
This adds an assertion that ensures that the necessary resize permission has been granted before bdrv_truncate() is called. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com> Acked-by: NFam Zheng <famz@redhat.com>
-
由 Kevin Wolf 提交于
This adds assertions that ensure that the necessary write permissions have been granted before someone attempts to write to a node. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Acked-by: NFam Zheng <famz@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com>
-
由 Kevin Wolf 提交于
This is where we want to check the permissions, so we need to have the BdrvChild around where they are stored. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Acked-by: NFam Zheng <famz@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com>
-
由 Kevin Wolf 提交于
The HMP command 'qemu-io' is a bit tricky because it wants to work on the original BlockBackend, but additional permissions could be required. The details are explained in a comment in the code, but in summary, just request whatever permissions the current qemu-io command needs. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com> Acked-by: NFam Zheng <famz@redhat.com>
-
由 Kevin Wolf 提交于
Management tools need to be able to know about every node in the graph and need a way to address them. Changing the graph structure was okay because libvirt doesn't really manage the node level yet, but future libvirt versions need to deal with both new and old version of qemu. This new option to blockdev-commit allows the client to set a node-name for the automatically inserted filter driver, and at the same time serves as a witness for a future libvirt that this version of qemu does automatically insert a filter driver. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Acked-by: NFam Zheng <famz@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com>
-
由 Kevin Wolf 提交于
Management tools need to be able to know about every node in the graph and need a way to address them. Changing the graph structure was okay because libvirt doesn't really manage the node level yet, but future libvirt versions need to deal with both new and old version of qemu. This new option to blockdev-mirror allows the client to set a node-name for the automatically inserted filter driver, and at the same time serves as a witness for a future libvirt that this version of qemu does automatically insert a filter driver. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com> Acked-by: NFam Zheng <famz@redhat.com>
-
由 Kevin Wolf 提交于
The correct permissions are relatively obvious here (and explained in code comments). For intermediate streaming, we need to reopen the top node read-write before creating the job now because the permissions system catches attempts to get the BLK_PERM_WRITE_UNCHANGED permission on a read-only node. Signed-off-by: NKevin Wolf <kwolf@redhat.com> Reviewed-by: NMax Reitz <mreitz@redhat.com> Acked-by: NFam Zheng <famz@redhat.com>
-