client/Client: added client_mount_uid and gid as parameters to getattr call

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

mds/MDSAuthCap: fix creation ownership check

Check uid too.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: whitespace

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: only verify create git when not AUTH_UID_ANY

Signed-off-by: Sage Weil <sage@redhat.com>

mds: send cap flush ack even when access check failed.

Signed-off-by: Yan, Zheng <zyan@redhat.com>

client: force setattr to MDS when caller's {uid,gid} are not the specified ones

Signed-off-by: Yan, Zheng <zyan@redhat.com>

client: allow specifying default caller_{uid,gid} of MClientRequest

Signed-off-by: Yan, Zheng <zyan@redhat.com>

client: add options to specify caller_{uid,gid} of MClientCaps

Signed-off-by: Yan, Zheng <zyan@redhat.com>

messages: add caller_{uid,gid} to cap msgs

Signed-off-by: Yan, Zheng <zyan@redhat.com>

test/libcephfs/access: expand example test a bit

Signed-off-by: Sage Weil <sage@redhat.com>

mds/Server: fix check_access

Pass through correct mask.  Clean up formatting.

Signed-off-by: Sage Weil <sage@redhat.com>

client: properly set caller_{uid,gid} of readdir request

save uid/gid of ll_opendir caller in dir_result_t, use the saved
uid/gid for readdir request.

Signed-off-by: Yan, Zheng <zyan@redhat.com>

ceph_test_libcephfs: skeleton for access tests

Signed-off-by: Sage Weil <sage@redhat.com>

mds: fix Server::check_access

Signed-off-by: Yan, Zheng <zyan@redhat.com>

doc/cephfs: path-based restriction

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add check_access in _do_cap_update

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

move _check_access to SessionMap

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add _check_access for async cap updates

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add stray_prior_path for is_stray

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add stray_prior_path to store path before rename

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

mds/MDSAuthCaps: add test cases for is_capable

mds/Server: add chown and chgrp check access to setattr

mds/Server: add create access check for openc

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

Server: add create access check for mknod and symlink

test_auth_caps: add mkdir check with mode 557

test_auth_caps: resolve bug with other bits test case

test_auth_caps: remove grp mount

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

MDSAuthCaps: validate create access

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

add stray_prior_path to store path before unlink

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add open check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add link check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add rename check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add snaps(ls,mk,rm,rename) check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add setlayout, setdirlayout, setxattr check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add readdir check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

mds/Server: clean up check_access a bit

Signed-off-by: Sage Weil <sage@redhat.com>

MDSAuthCaps: add logic for group bits check

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

test_auth_caps: add test for user bits

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

mds/MDSAuthCaps: add permissions for user bits

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

Makefile: include ceph-fuse in base target

Signed-off-by: Sage Weil <sage@redhat.com>

client: behave if we can't getattr parents of mount point

If our cap locks us into a subdirectory (e.g., /foo), we will fail to
getattr on its parents (e.g., /).  This is expected.  Tolerate this case,
and warn that quotas may misbehave in that case (if they are set on one of
those parents).

Signed-off-by: Sage Weil <sage@redhat.com>

test_auth_caps: Move trap and cleanup to the top

test_auth_caps: Added test logic for world bits

MDSAuthCaps: add world bits check logic

Signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>

client: pass uid, gid to lookup

The MDS cares about who is performing lookup, too!

Signed-off-by: Sage Weil <sage@redhat.com>

add unlink, rmdir check_access test

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add unlink, rmdir check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add symlink test

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

add symlink check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

Add mknod check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

qa/workunits/fs/test_path_caps: tolerate existing directories

Signed-off-by: Sage Weil <sage@redhat.com>

functional test of mds cap path restriction

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

mds/Server.cc: drop leading / from path in is_capable check

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

unittest_mds_authcap: test lists of allow grants

Signed-off-by: Sage Weil <sage@redhat.com>

mds: mkdir check_access

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

mds/MDSAuthCaps: fix parse error message

Signed-off-by: Sage Weil <sage@redhat.com>

mds: whitespace

Signed-off-by: Sage Weil <sage@redhat.com>

mds: log to cluster log if mds cap parse fails

It's an admin error; tell an admin.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: debug is_capable

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCap: fix debug prefix

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCap: drop leading / in paths

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: fix allow_all

Empty path is '', not '/'.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: add cct for debug context

signed-off-by: Nishtha Rai <nishtha3rai@gmail.com>
Signed-off-by: Sage Weil <sage@redhat.com>

mds: calculate path in check_access()

Signed-off-by: Jashan Kamboj <jashank42@gmail.com>

mds/MDSAuthCaps: normalize path, drop useless constant.

Use an empty string for no path--this is more efficient.  This generalizes
to losing any leading '/' character.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: pass down inode uid.gid and mode

We will need this to evaluate the unix permissions.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/Server: add a few access checks

(incomplete coverage!)

Signed-off-by: Sage Weil <sage@redhat.com>

mds/Server: add check_access() hook

For now, we are doing a very basic permission check.  No callers yet.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCap: fix uid and gid types

int64_t for uid, so we can do a negative "none" value.
gid_t (uint32_t) for gid.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: cosmetic

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: use bitmask for is_capable()

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCap: fix path match

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: move allows() into MDSCapSpec

Also, fix this so that 'any' implies that we can read or write.

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: parse optional gid list

Signed-off-by: Sage Weil <sage@redhat.com>

mds/MDSAuthCaps: whitespace

Signed-off-by: Sage Weil <sage@redhat.com>

Merge pull request #5951 from guangyy/doc-pool

doc: update doc for with new pool settings

Reviewed-by: Loic Dachary <ldachary@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #5853 from branch-predictor/bp-rados-fix-noverify

common/obj_bencher.cc: fix verification crashing when there's no objects

Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #6009 from theanalyst/scripts/orig-issue

scripts: release_notes can track original issue

Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #5211 from 0003ydlom/smc3

MOSDOp/do_op/handle_op: Simple Messenger optimization

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Greg Farnum <gfarnum@redhat.com>
Tested-by: Kefu Chai <kchai@redhat.com>

Merge pull request #6127 from wido/mirror-docs-fixes

docs: Fix styling of newly added mirror docs

Merge pull request #6109 from liewegas/wip-13214

init-rbdmap: fix CMDPARAMS

Reviewed-by: Nathan Cutler <ncutler@suse.com>

init-rbdmap: fix CMDPARAMS

Fixes: #13214
Reported-by: Wyllys Ingersoll <wyllys.ingersoll@keepertech.com>
Signed-off-by: Sage Weil <sage@redhat.com>

docs: Fix styling of newly added mirror docs

cleanup: make the pool setting GET decription point to SET description

Signed-off-by: Guang Yang yguang@yahoo-inc.com

doc: update doc for with new pool settings

Signed-off-by: Guang Yang yguang@yahoo-inc.com

Merge remote-tracking branch 'gh/infernalis'

Merge pull request #6101 from liewegas/wip-fs

os/fs: fix do loop

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #6103 from SUSE/wip-re-redrop-fdupes

ceph.spec.in: re-re-drop fdupes

Reviewed-by: Boris Ranto <branto@redhat.com>

test/test_rados_tool.sh: implement regression test for bench verify crash

Truncate all objects after format test, then check whether rados bench rand/seq
still works, or just crashes.

Signed-off-by: Piotr Dałek <piotr.dalek@ts.fujitsu.com>

common/obj_bencher.cc: fix verification crashing when there's no objects

When specified pool is empty or doesn't contain valid bench objects, rados
can crash because it attempts to memcmp through NULL pointer. Added check
for valid object size (which also fixes the case when objects are there, but
they're truncated), and made object_size being size_t instead of int.
In seq_read_bench, there's a missing lock.Lock() which causes rados to trigger
assert when object reading fails. Moved lock.Lock() before completion_ret()
call.

Signed-off-by: Piotr Dałek <piotr.dalek@ts.fujitsu.com>

ceph.spec.in: re-re-drop fdupes

The %fdupes call was dropped by 53072b9019caf72e0313b2804ea174237ed7da33
Later, it got accidentally reinstated by ceb93e8e69e125c9358f63b0099e7509b6624bcf

Signed-off-by: Nathan Cutler <ncutler@suse.com>

Merge pull request #6025 from ukernel/wip-13167

journaler: detect unexpected holes in journal objects

Reviewed-by: Greg Farnum <gfarnum@redhat.com>

Merge pull request #5849 from ukernel/wip-32bits-ceph-fuse

Wip 32bits ceph fuse

Reviewed-by: Greg Farnum <gfarnum@redhat.com>

Merge pull request #4791 from ceph/wip-11783

cephfs: detect completed cap flushes

Reviewed-by: Greg Farnum <gfarnum@redhat.com>

Merge pull request #6070 from jcsp/wip-reformat-subtreemaptest

Fix reformatting journals containing SUBTREEMAP_TEST events

Merge pull request #6082 from jcsp/wip-client-quota-roots

Fix quota enforcement on subdir mounts

os/fs: fix aio submit method

continue in a do while(false) will always eval false and break
out.  To repeat, we need while (true) and an explicit break.

Reported-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
Signed-off-by: Sage Weil <sage@redhat.com>

Merge pull request #6078 from H3C/wip-13208

Objecter: maybe access wild pointer(op) in _op_submit_with_budget.

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #6064 from H3C/wip-rbd-bugfix1

Objecter: misc fix for segmentation fault.

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #6084 from liewegas/wip-revert-scrub

Revert "osd: new pool settings: scrub intervals"

Reviewed-by: Mykola Golub <mgolub@mirantis.com>