docs/radosgw: edit admin.rst "enable/disable user rate limit"

Edit "Enable/Disable User Rate Limit" in doc/radosgw/admin.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 1cd996b6fe2e3c923001a69772c7b7635d058f2b)

Merge pull request #55164 from zdover23/wip-doc-2024-01-13-backport-54530-to-quincy

quincy: ceph-menv:fix typo in README

ceph-menv:fix typo in README

Signed-off-by: yu.wang <wangyu@wyu.space>
(cherry picked from commit 4efd9317b83eca48c1a92724e7ebe304650fbc04)

Merge pull request #55098 from cbodley/wip-63840-quincy

quincy: qa: remove vstart runner from radosgw_admin task

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #54851 from mkogan1/wip-63752-quincy

quincy: rgw: d3n: fix valgrind reported leak related to libaio worker threads

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #54842 from cbodley/wip-63772-quincy

quincy: rgw/iam: admin/system users ignore iam policy parsing errors

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #54737 from yuvalif/wip-63702-quincy

quincy: rgw/kafka/amqp: fix race conditionn in async completion handlers

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #54675 from cbodley/wip-63627-quincy

quincy: rgw: object lock avoids 32-bit truncation of RetainUntilDate

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #54445 from tobias-urdin/service-token-quincy

quincy: rgw/auth: Add service token support for Keystone auth

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #54166 from cbodley/wip-58237-quincy

quincy: rgw: beast frontend checks for local_endpoint() errors

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #54165 from cbodley/wip-59132-quincy

quincy: rgw/s3: DeleteObjects response uses correct delete_marker flag

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #54162 from cbodley/wip-61353-quincy

quincy: rgw/s3: ListObjectsV2 returns correct object owners

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #53763 from cbodley/wip-63044-quincy

quincy: rgw/keystone: EC2Engine uses reject() for ERR_SIGNATURE_NO_MATCH

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #53473 from k0ste/wip-55700-quincy

quincy: radosgw-admin: don't crash on --placement-id without --storage-class

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #52795 from cbodley/wip-62310-quincy

quincy: rgw/crypt: apply rgw_crypt_default_encryption_key by default

Reviewed-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #52049 from trociny/wip-59612-quincy

quincy: rgw/sts: AssumeRole no longer writes to user metadata

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #51976 from trociny/wip-59065-quincy

quincy: rgw: fix multipart upload object leaks due to re-upload

Reviewed-by: Casey Bodley <cbodley@redhat.com>

Merge pull request #55158 from zdover23/wip-doc-2024-01-12-backport-55146-to-quincy

quincy: doc/radosgw: edit admin.rst "get user rate limit"

doc/radosgw: edit admin.rst "get user rate limit"

Edit "Get User Rate Limit" in doc/radosgw/admin.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 1c561e2beac67c85c40f2adf387c42aba8175ecb)

Merge pull request #54594 from idryomov/wip-53593-quincy

quincy: osd: don't require RWEXCL lock for stat+write ops.

Merge pull request #55151 from zdover23/wip-doc-2024-01-12-backport-55142-to-quincy

quincy: doc/radosgw - edit admin.rst "set user rate limit"

doc/radosgw - edit admin.rst "set user rate limit"

Edit "Set User Rate Limit" in doc/radosgw/admin.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 8660b4edc71113d5d2d51438c9fd6f388b35a214)

rgw: object lock uses 64-bit encoding for RetainUntilDate

the default encoding of ceph::real_time truncates seconds to uint32_t,
so stores the wrong timestamp for object lock enforcement

Fixes: https://tracker.ceph.com/issues/63537
Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 1fd4309fbbebf0d3e67aa2800d5fb3c7de19dcc7)

encoding: add round_trip_encode()/decode() for chrono types

the default encodings for chrono types were made to be
backward-compatible with utime_t, so truncated seconds to 32 bits

adds new functions that encode these chrono types using their underlying
representation, which for ceph::real_time and ceph::timespan is
'nanoseconds as uint64_t'

Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit b7aa16da4bbb444f6361e4f14be7ef64b0ea3bb1)

Merge pull request #55129 from zdover23/wip-doc-2024-01-10-backport-55106-to-quincy

quincy: doc/rgw: edit admin.rst - rate limit management

doc/rgw: edit admin.rst - rate limit management

Edit the "Rate Limit Management" section in doc/radosgw/admin.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 47fabe757cff630aaf23a85d268b7406057244e2)

Merge pull request #55092 from cbodley/wip-63965-quincy

quincy: make-dist: don't use --continue option for wget

Reviewed-by: Laura Flores <lflores@redhat.com>

Merge pull request #55085 from petrutlucian94/wip-63956-quincy

quincy: win32_deps_build.sh: change Boost URL

Reviewed-by: Guillaume Abrioux <gabrioux@ibm.com>

Merge pull request #55025 from idryomov/wip-63607-quincy

quincy: librados: make querying pools for selfmanaged snaps reliable

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge pull request #54123 from batrick/wip-63276-quincy

quincy: cmake: use or turn off liburing for rocksdb

Reviewed-by: Laura Flores <lflores@redhat.com>

qa: remove vstart runner from radosgw_admin task

Fixes: https://tracker.ceph.com/issues/63583
Signed-off-by: Ali Maredia <amaredia@redhat.com>
(cherry picked from commit a44fd94b9681441a3251d0988d1ddd159f3733f8)

Merge pull request #52461 from batrick/wip-62023-quincy

quincy: pybind/mgr: reopen database handle on blocklist

Reviewed-by: Laura Flores <lflores@redhat.com>

Merge pull request #53973 from Matan-B/wip-63181-quincy

quincy: osd/OSD: introduce reset_purged_snaps_last

Reviewed-by: Samuel Just <sjust@redhat.com>
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge pull request #53718 from leonid-s-usov/bp/msg-wait/quincy

quincy: msg/AsyncMessenger: re-evaluate the stop condition when woken up in 'wait()'

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>

Merge pull request #52458 from kamoltat/wip-ksirivad-backport-quincy-48209

quincy: osd/OSDMap: Check for uneven weights & != 2 buckets post stretch mode

Reviewed-by: Nitzan Mordechai <nmordech@redhat.com>

Merge pull request #54534 from pdvian/wip-63367-quincy

quincy: mgr: remove out&down osd from mgr daemons

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

Merge pull request #53886 from sseshasa/wip-63129-quincy

quincy: mon/ConfigMonitor: Show localized name in "config dump --format json" output

Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>

make-dist: don't use --continue option for wget

the boost jfrog mirror is broken and returns an HTML error page instead
of the archive. the file size of this page is 11534 bytes

when download_from() retries the download from download.ceph.com, the -c
option tells it to resume the download of the existing file. the
resulting boost_1_82_0.tar.bz2 ends up with the correct total file size
of 121325129 bytes, but the first 11534 bytes still correspond to the
HTML from jfrog. that causes the sha256sum mismatch

remove the -c option so that wget fetches the archive in its entirety

Signed-off-by: Casey Bodley <cbodley@redhat.com>
(cherry picked from commit 2ba7d6e59e8750867059c29a33c22bcb3e358b65)

Merge pull request #55083 from zdover23/wip-doc-2024-01-08-backport-55075-to-quincy

quincy: doc/radosgw: edit admin.rst - quota

win32_deps_build.sh: change Boost URL

The Boost download URL is broken again [1], we'll switch to a
working mirror.

[1] https://github.com/boostorg/boost/issues/842

Fixes: https://tracker.ceph.com/issues/63952
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
(cherry picked from commit 594d1e5e83e3d4c318ab1dd79660670a0852f7e6)

doc/radosgw: edit admin.rst - quota

Edit the "Quota Management" section (all of it) in
doc/radosgw/admin.rst.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 22fce3c7a55591f6755355751157bc7231776137)

Merge pull request #55072 from zdover23/wip-doc-2024-01-06-backport-55069-to-quincy

quincy: doc: update rgw admin api req params for get user info

Reviewed-by: Anthony D'Atri <anthony.datri@gmail.com>

Merge pull request #55074 from zdover23/wip-doc-2024-10-06-backport-55065-to-quincy

quincy: doc/radosrgw: edit admin.rst

doc/radosrgw: edit admin.rst

Edit "Add / Remove Admin Capabilities" in doc/radosgw/admin.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 1f0ff27898dfd06a17b18fadd52f68dfb73ccb94)

doc: update rgw admin api req params for get user info

Signed-off-by: Ali Maredia <amaredia@redhat.com>
(cherry picked from commit e6c58dcd614f71efa4de2f781b8425c480af285e)

Merge pull request #54950 from idryomov/wip-53897-quincy

quincy: librbd: don't report HOLE_UPDATED when diffing against a hole

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Laura Flores <lflores@redhat.com>

Merge pull request #54090 from mchangir/wip-63241-quincy

quincy: mgr/snap_schedule: make fs argument mandatory if more than one filesystem exists

qa: use correct imports to resolve fuse_mount and kernel_mount

Fixes: https://tracker.ceph.com/issues/62706
Signed-off-by: Milind Changire <mchangir@redhat.com>
(cherry picked from commit 994ae0d633843e61eae91f409ec3b0336db71dc3)

PendingReleaseNotes: note about mandatory fs argument

Signed-off-by: Milind Changire <mchangir@redhat.com>
(cherry picked from commit 71c6356f0a7551f3798dc0d5e58d9781736692d6)

doc/cephfs: add note about mandatory --fs argument to snap-schedule

Signed-off-by: Milind Changire <mchangir@redhat.com>
(cherry picked from commit 8c29772fbe60b7b7657fa53d84e24bf74c993dca)

qa: add test for mandatory fs argument to snap-schedule commands

Signed-off-by: Milind Changire <mchangir@redhat.com>
(cherry picked from commit 484336e1c3dccd42d594234614936ab1f4b0d35a)

mgr/snap-schedule: tweaks to keep mypy happy

Signed-off-by: Milind Changire <mchangir@redhat.com>
(cherry picked from commit c208ce9f3ced89755b1180178812ef82aa1a5346)

mgr/snap_schedule: validate fs before execution

Stop command execution if there are more than one filesystem and --fs
argument is missing in command-line, i.e. do not use the first fs in the
fsmap if there are more than one filesystem.
This is to ensure that user doesn't mistakenly run the command against
the first fs by missing to specify the desired fs.

Fixes: https://tracker.ceph.com/issues/62218
Signed-off-by: Milind Changire <mchangir@redhat.com>
(cherry picked from commit c612ce06300fa99f750ce99ef68dbd49ced16c34)

Merge pull request #54539 from aaSharma14/wip-63572-quincy

quincy: mgr/dashboard: Show the OSDs Out and Down panels as red whenever an OSD is in Out or Down state in Ceph Cluster grafana dashboard

Reviewed-by: Nizamudeen A <nia@redhat.com>

Merge pull request #55056 from zdover23/wip-doc-2024-01-04-backport-55048-to-quincy

quincy: doc/radosgw: edit "Add/Remove a Key"

doc/radosgw: edit "Add/Remove a Key"

Edit the section "Add/Remove a Key" in doc/radosgw/admin.rst. Each
operation (e.g. "Adding an S3 key pair for a user", "Removing an S3 key
pair for a user") now has its own subsection. This increased granularity
should make it easier in the future to link to each of these specific
operations, if needed.

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit f62e93cbe73cd8f624a6c99497051c1a0aaf3ab6)

Merge pull request #54405 from joscollin/wip-63475-quincy

quincy: client: call _getattr() for -ENODATA returned _getvxattr() calls

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #54317 from batrick/wip-63415-quincy

quincy: mon: fix mds metadata lost in one case.

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #54206 from neesingh-rh/wip-63263-quincy

quincy: mgr/volumes: fix `subvolume group rm` error message

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #54034 from vshankar/wip-63175-quincy

quincy: mds: adjust pre_segments_size for MDLog when trimming segments for st…

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #53917 from kotreshhr/wip-63163-quincy

quincy: pybind/mgr/volumes: log mutex locks to help debug deadlocks

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #53911 from kotreshhr/wip-62404-quincy-qa-fix

quincy: qa: Wait for purge to complete

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #53907 from vshankar/wip-62799-quincy

quincy: qa: move nfs (mgr/nfs) related tests to fs suite

Reviewed-by: Kotresh HR khiremat@redhat.com
Reviewed-by: Adam King adking@redhat.com

Merge pull request #53554 from batrick/wip-62898-quincy

quincy: qa: lengthen shutdown timeout for thrashed MDS

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #53551 from batrick/wip-62907-quincy

quincy: mds,qa: some balancer debug messages (<=5) not printed when debug_mds is >=5

Reviewed-by: Kotresh HR khiremat@redhat.com

Merge pull request #53549 from batrick/wip-62900-quincy

quincy: mds: log message when exiting due to asok command

Reviewed-by: Kotresh HR khiremat@redhat.com
Reviewed-by: Christopher Hoffman <choffman@redhat.com>

Merge pull request #53559 from batrick/wip-57777-quincy

quincy: doc: discuss the standard multi-tenant CephFS security model

Reviewed-by: Kotresh Hiremath Ravishankar <khiremat@redhat.com>

Merge pull request #55045 from zdover23/wip-doc-2024-01-02-backport-55038-to-quincy

quincy: doc/cephfs: Update disaster-recovery-experts.rst to mention Slack

doc/cephfs: Update disaster-recovery-experts.rst to mention Slack

Signed-off-by: Dhairya Parmar <dparmar@redhat.com>
(cherry picked from commit 6601e7f38c0b2c1326d40f3bb0599069e38421ff)

Merge pull request #55035 from zdover23/wip-doc-2024-01-01-backport-55030-to-quincy

quincy: doc/radosgw: edit "remove a subuser"

Merge pull request #55033 from zdover23/wip-doc-2024-01-01-backport-55031-to-quincy

quincy: doc/mgr: document lack of MSWin NFS 4.x support

doc/radosgw: edit "remove a subuser"

Edit the English language in the section "Remove a Subuser" in
doc/radosgw/admin.rst. This commit is made in response to Matt
Benjamin's request for improvement of this section
(https://github.com/ceph/ceph/pull/55028#discussion_r1438599833).

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 8fc3c488861887490f6ec6bac8f49770664b49a6)

doc/mgr: document lack of MSWin NFS 4.x support

Document that no version of Microsoft Windows supports mounting an NFS
v4.x export natively, as of the time of this commit.

This commit has been made in response to an anonymous note left on
https://pad.ceph.com/p/Report_Documentation_Bugs.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 03656c02a70d4ab719cbcf605517f8a2e9e81117)

Merge pull request #55028 from zdover23/wip-doc-2023-12-30-backport-55023-to-quincy

quincy: doc/radosgw: edit sections

doc/radosgw: edit sections

Edit the following sections in doc/radosgw/admin.rst:

* Remove a User
* Remove a Subuser

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 8631dc58f7144b32febadd9072c6691456ac8f28)

tools/rados: just return instead of exit(1) in "rados cppool" handler

Otherwise an occasional segfault occurs.  This instance was missed in
commit 2c149262888c ("tools/rados: always call rados.shutdown() before
exit()").

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit de6635582b7615eb54800a616711073a3c7b0dbd)

qa: make sure "rados cppool" requires --yes-i-really-mean-it for RBD

Safeguards in rados CLI tool isn't really the subject of this test, but
it fits nicely.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 2b135a2eb60e0dd5ab22d213d3d2435608c6a9d3)

librados: make querying pools for selfmanaged snaps reliable

If get_pool_is_selfmanaged_snaps_mode() is invoked on a fresh RADOS
client instance that still lacks an osdmap, it returns false, same as
for "this pool is not in selfmanaged snaps mode".  The same happens if
the pool in question doesn't exist since the signature doesn't allow to
return an error.

The motivation for this API was to prevent users from running "rados
cppool" on a pool with unmanaged snapshots and deleting the original
thinking that they have a full copy.  Unfortunately, it's exactly
"rados cppool" that fell into this trap, so no warning is printed and
--yes-i-really-mean-it flag isn't enforced.

Fixes: https://tracker.ceph.com/issues/63607
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 0999e63bfbbee46b8e19c3f05881eee64dba8b5e)

Conflicts:
PendingReleaseNotes [ moved to >=17.2.8 section ]

librados: clarify get_pool_is_selfmanaged_snaps_mode() semantics

Refer to the commit message of 8a9769a1d5fd ("librados: add
get_pool_is_selfmanaged_snaps_mode() function").

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 54edc7955b0593423d2c03ca308b5e5fcdc528a3)

Merge pull request #55021 from zdover23/wip-doc-2023-12-28-backport-55011-to-quincy

quincy: doc/radosgw: admin.rst - edit "Create a Subuser"

doc/radosgw: admin.rst - edit "Create a Subuser"

Edit the section "Create a Subuser" in doc/radosgw/admin.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 3c7a6c41df540f935b0b0b2c4fdc7375327a7662)

Merge pull request #55018 from zdover23/wip-doc-2023-12-28-backport-55016-to-quincy

quincy: doc/radosgw: admin.rst - edit sections

doc/radosgw: admin.rst - edit sections

Edit the following sections in doc/radosgw/admin.rst:

- Get User Info
- Modify User Info
- User Suspend/Enable (I split this into two sections)

Co-authored-by: Anthony D'Atri <anthony.datri@gmail.com>
Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 21ebd2e855a4dd2d2d1244bb3688db0c6613dc30)

test/pybind/rbd: don't ignore from_snapshot in check_diff()

Despite the test in test_diff_iterate() being correct, it started
failing:

    >       check_diff(self.image, 0, IMG_SIZE, 'snap1', [(0, 512, False)])
    ...
    a = [], b = [(0, 512, False)]
    ...
    >       assert a == b
    E       AssertionError

This is because check_diff() drops 'snap1' argument on the floor and
passes None to image.diff_iterate() instead.  This goes back to 2013,
see commit e88fe3cbbc8f ("rbd.py: add some missing functions").

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit f8ced6d1fe667864bee224104c23b62a2c1c0bfc)

test/librbd: redo TestMockObjectMapDiffRequest.*Delta tests

Existing *Delta tests cover:

- beginning of time -> HEAD, through intermediate snap
- snap -> snap, directly
- snap -> HEAD, directly

But coverage is too weak: none of the weird OBJECT_PENDING cases and
only a single diff-iterate vs deep-copy case is tested, for example.

Coverage is missing completely for:

- beginning of time -> HEAD, directly
- beginning of time -> snap, directly
- beginning of time -> snap, through intermediate snap
- snap -> snap, through intermediate snap
- snap -> HEAD, through intermediate snap

This adds the following tests:

- FromBeginningToHead
- FromBeginningToHeadIntermediateSnap (expands FullDelta)
- FromBeginningToSnap
- FromBeginningToSnapIntermediateSnap
- FromSnapToSnap (expands IntermediateDelta)
- FromSnapToSnapIntermediateSnap
- FromSnapToHead (expands EndDelta)
- FromSnapToHeadIntermediateSnap

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 7aff35c987b9f0e9f0d4198d032737da0ce12b36)

test/librbd: parametrize TestMockObjectMapDiffRequest tests

Exercise both diff-iterate and deep-copy modes of operation.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit f47f7f87d05da661786415dd8fbb0ac67c530e18)

librbd: OBJECT_PENDING should always be treated as dirty

OBJECT_PENDING is a transition state which normally isn't encountered
in (snapshot) object maps.  In case it's encountered, for example when
a snapshot is taken after losing power at the time a discard was being
handled, the object should be treated as dirty and produce a diff as
a result.

Assuming an object is marked OBJECT_PENDING, theoretically there are
four cases with respect to object's state in the next snapshot:

    1. OBJECT_NONEXISTENT
    2. OBJECT_EXISTS
    3. OBJECT_PENDING
    4. OBJECT_EXISTS_CLEAN

Prior to commit b81cd2460de7 ("librbd/object_map: diff state machine
should track object existence"), (3) was handled incorrectly (diff set
to DIFF_STATE_NONE instead of DIFF_STATE_UPDATED).

Post commit 399a45e11332 ("librbd/object_map: rbd diff between two
snapshots lists entire image content"), (4) is handled incorrectly
(diff set to DIFF_STATE_DATA instead of DIFF_STATE_DATA_UPDATED).

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 70c1991f225c137abf769111086b6c61abc0ece3)

test/librbd: expand DiffIterateTest.DiffIterateDiscard

Similar to DiffIterateTest.DiffIterateDeterministic, systematically
cover the most common cases involving full-object discards.  With this
in place, issue [1] can be reproduced by any of:

    (preparatory) before snap3 is taken
    (1) beginning of time -> HEAD
    (2) snap1 -> HEAD
    (5) beginning of time -> snap3
    (6) snap1 -> snap3

Sub-object discards aren't covered here because of further issues
[2][3].

[1] https://tracker.ceph.com/issues/53897
[2] https://tracker.ceph.com/issues/63770
[3] https://tracker.ceph.com/issues/63771

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit 6b45a2d920aad2edf17544807f7032e2b11efef3)

librbd: resurrect "exists" assert in simple_diff_cb()

This effectively reverts commit 3ccc3bb4bd35 ("librbd: diff_iterate
needs to handle holes in parent images") which just dropped the assert
instead of addressing the root cause of reported crashes.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit bcb107afcd2b0e34b1eefa4767691c5d8d109a82)

librbd: diff-iterate shouldn't ever report "new hole" against a hole

If an object doesn't exist in both start and end versions but there is
an intermediate snapshot which contains it (i.e. the object is written
to and captured at some point but then discarded prior to or in the end
version), diff-iterate reports "new hole" -- callback is invoked with
exists=false.  This occurs both on the slow list_snaps path and in
fast-diff mode.

Despite going all the way back to the introduction of diff-iterate in
commit 0296c7cdae91 ("librbd: implement diff_iterate"), this behavior
is wrong and contradicts diff-iterate API documentation added in commit
a69532e86450 ("librbd: document diff_iterate in header") in the same
series:

    If the source snapshot name is NULL, we interpret that as
    the beginning of time and return all allocated regions of the
    image.

It also triggered an assert added in commit c680531e070a ("librbd:
change diff_iterate interface to be more C-friendly") in the same
series.  Unfortunately, commit f1f6407221a0 ("test_librbd: add
diff_iterate test including discard"), also part of the same series,
added a test which expected the wrong behavior.  Very confusing!

A year later, a different manifestation of this bug was fixed in commit
9a1ab95176fe ("rbd: Fix rbd diff for non-existent objects"), but the
fix only covered the case where calc_snap_set_diff() goes past the end
snap ID while processing clones.  The case where it runs out of clones
to process before reaching the end snap ID remained mishandled.

A year after that, commit 3ccc3bb4bd35 ("librbd: diff_iterate needs to
handle holes in parent images") dropped the assert mentioned above and
this bug got enshrined in the newly introduced fast-diff mode.

Finally, a few years later, deep-copy actually started relying on this
bug in commit e5a21e904142 ("librbd: deep-copy image copy state machine
skips clean objects").  This necessitates bifurcation in DiffRequest
because deep-copy wants the "has this object been touched" semantics,
which is different from diff-iterate (and also potentially much more
expensive to produce!).

This commit brings a minimal update to TestMockObjectMapDiffRequest
tests and DiffIterateTest.DiffIterateDiscard.  Coverage is expanded in
the following commits.

Fixes: https://tracker.ceph.com/issues/53897
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
(cherry picked from commit be507aaed15fb7a193a90c5f88eda61b9be2af1b)

Merge pull request #54861 from idryomov/wip-63654-quincy

quincy: librbd: fix regressions in ObjectListSnapsRequest

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Ramana Raja <rraja@redhat.com>

Merge pull request #54819 from baergj/wip-63747-quincy

quincy: librbd: Append one journal event per image request

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>

Merge pull request #54770 from ajarr/wip-63716-quincy

quincy: qa/workunits/rbd/cli_generic.sh: narrow race window when checking that rbd_support module command fails after blocklisting the module's client

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>

Merge pull request #54292 from ajarr/wip-63386-quincy

quincy: qa/suites/rbd: add test to check rbd_support module recovery

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Ilya Dryomov <idryomov@redhat.com>

Merge pull request #54290 from ajarr/wip-63383-quincy

quincy: mgr/rbd_support: fix recursive locking on CreateSnapshotRequests lock

Reviewed-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Mykola Golub <mgolub@suse.com>

Merge pull request #54255 from pkalever/wip-63349-quincy

quincy: rbd-nbd: fix stuck with disable request

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Ilya Dryomov <idryomov@redhat.com>

Merge pull request #54054 from idryomov/wip-63028-quincy

quincy: pybind/rbd: don't produce info on errors in aio_mirror_image_get_info()

Reviewed-by: Mykola Golub <mgolub@suse.com>
Reviewed-by: Ramana Raja <rraja@redhat.com>

Merge pull request #55005 from zdover23/wip-doc-2023-12-26-backport-55003-to-quincy

quincy: doc/radosgw: admin.rst - edit "Create a User"

doc/radosgw: admin.rst - edit "Create a User"

Edit the section "Create a User" in doc/radosgw/admin.rst.

Signed-off-by: Zac Dover <zac.dover@proton.me>
(cherry picked from commit 46139c8fa9d4ae0e493bca17ba1199e3a67c39f8)

Merge pull request #54501 from zdover23/wip-doc-2023-11-15-backport-54499-to-quincy

quincy: doc/rgw/notifications: specify which event types are enabled by default

Reviewed-by: Matt Benjamin <mbenjamin@redhat.com>