* pass coremask_arg to the working thread by value.
after 81249ab9d, get_val<>() returns a temporary variables instead of a
reference to the variant held by config. so calling string::c_str(), and
passing the returned `const char*` to a thread is not advisable. instead
we should pass a string by value. since spdk_env_init() will copy the
settings passed in by opt. it's safe even to destruct the coremask_arg
afterward.
* use ffsll() to find the LSB. it's not a bottleneck, but it's easier
and simpler, and probably improves the readability.
* refactor the NVMEManager::try_get() method: to define the variables
when they are used for the first time.
Signed-off-by: Kefu Chai <kchai@redhat.com>
* refs/pull/18274/head:
mds: fold mds_revoke_cap_timeout into mds_session_timeout
client: add new delegation testcases
client: add delegation support for cephfs
common: remove data_dir_option from common_preinit and global_pre_init
Reviewed-by: Gregory Farnum <gfarnum@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
This prevents the Inode from being deleted until after cap trimming is
finished. In particular, this prevents remove_all_caps from being called which
screws up the traversal of caps in trim_caps.
Fixes: http://tracker.ceph.com/issues/22157
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
rgw: fix a bug in rgw cache in delete_system_obj and get_system_obj.
Reviewed-by: Matt Benjamin <mbenjami@redhat.com>
Reviewed-by: Orit Wasserman <owasserm@redhat.com>
Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
Reviewed-by: Amit Kumar <amitkuma@redhat.com>
Improve the error message for 'ceph fs authorize' where the
given entity exists but has different auth caps.
Fixes: http://tracker.ceph.com/issues/21765
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
Revert CEPH_FEATURE_MSG_AUTH-related hunks of commit 7ba4d5324e
("msg/simple/Pipe: remove legacy feature case handle"), which
removed code that handled !CEPH_FEATURE_MSG_AUTH footers. It's needed
for kernels prior to 3.19, where cephx signatures aren't supported.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Revert CEPH_FEATURE_MSG_AUTH-related hunks of commit 83f7e392a1
("msg/async/AsyncConnection: remove legacy feature case handle"), which
removed code that handled !CEPH_FEATURE_MSG_AUTH footers. It's needed
for kernels prior to 3.19, where cephx signatures aren't supported.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
rgw: radosgw-admin zonegroup get and zone get return defaults when there is no realm
Reviewed-by: Orit Wasserman <owasserm@redhat.com>
Reviewed-by: Casey Bodley <cbodley@redhat.com>
Fixes the coverity issue:
CID 1405270 (#1 of 1): Division or modulo by zero (DIVIDE_BY_ZERO)
2. divide_by_zero: In function call ll_get_stripe_osd,
division by expression l.stripe_count which may be zero has
undefined behavior.
Signed-off-by: Amit Kumar <amitkuma@redhat.com>
rgw: keystone: bump up logging when error is received
Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
KernelDevice is tightly coupled with libaio. more work is needed to
decouple aio from it. but by guarding KernelDevice with HAVE_LIBAIO, we
can enable bluestore on platforms w/o libaio.
Signed-off-by: Kefu Chai <kchai@redhat.com>
Completed requests can also be in replay queue, they need to be processed
(treated as lookup request) in clientreplay stage. MDS will be stuck at
clientreplay stage if it delays processing requests of this type.
the bug was introduced by commit 0afbc033
(mds: don't rdlock locks in replica object while auth mds is recovering)
Fixes: http://tracker.ceph.com/issues/22163
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
Updated the COMMANDS section in RADOSGW-ADMIN doc page, using the latest commands from rgw_admin.cc::usage()
Signed-off-by: Jos Collin <jcollin@redhat.com>
