| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
legacy-option-headers provides the headers included by
`src/common/options/legacy_config_opts.h`. these headers are generated from
corresponding .yaml.in file. because these headers file are generated at
build time, we need to construct the file-level dependencies between the
targets using them and these headers, otherwise the compiler could fail
to compile the tree if any of the headers are not generated yet, when
compiling .cc file which (indirectly) includes it.
in order to address this, in this change, we
1. search for all .cc files which have `#include "common/config.h"` in it
2. and find out the targets building the .cc file, then
3. add `legacy-option-headers` to its linkage using CMake
this should partially address the above race condition we've been running
into on slow build hosts.
because we have not audited the .h files including `common/config.h`,
this change should be considered a partial fix.
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
|
| |\
| |
| |
| |
| | |
dokan: simple case insensitive emulation
Reviewed-by: Venky Shankar <vshankar@redhat.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
While CephFS is case sensitive, Windows software commonly assume
that the filesystem is case insensitive.
In order to unblock certain use cases, a short term solution is to
simply normalize paths when creating or accessing files or
directories.
This change adds two ceph-dokan parameters:
* --case-insensitive: if set, paths are normalized
* --force-lowercase: normalized paths are converted to lowercase
instead of uppercase
This trivial implementation has some limitations:
* the original case is not preserved
* we could later on use xattr to store the original name
* can't access existing files that have a different case
* handled at ceph-dokan level
The advantage is that it's simple, shouldn't impact performance
and doesn't require libcephfs or mds changes.
In the future, we may conider a more robust implementation
at the mds and/or libcephfs level.
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Windows CLI arguments use either ANSI (main()) or UTF-16 (wmain()).
Meanwhile, Ceph libraries expect UTF-8 and raise exceptions when
trying to use Unicode CLI arguments or log Unicode output:
rbd.exe create test_unicode_șțăâ --size=32M
terminate called after throwing an instance of 'std::runtime_error'
what(): invalid utf8
We'll use a Windows application manifest, setting the "activeCodePage"
property [1][2]. This enables the Windows UCRT UTF-8 mode so that
functions that receive char* arguments will expect UTF-8 instead of ANSI,
including main(). One exception is CreateProcess, which will need the
UTF-16 form (CreateProcessW).
Despite the locale being set to utf-8, we'll have to explicitly set
the console output to utf-8 using SetConsoleOutputCP(CP_UTF8).
In order to use the UTF-8 locale, we'll have to switch the mingw-llvm
runtime from msvcrt to ucrt.
This also fixes ceph-dokan crashes that currently occur when non-ANSI
paths are logged.
[1] https://learn.microsoft.com/en-us/windows/win32/sbscs/application-manifests#activecodepage
[2] https://learn.microsoft.com/en-us/windows/apps/design/globalizing/use-utf8-code-page
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
| |
We're using std::stoul to parse cli args, however we aren't
catching the exceptions.
This change will handle the exceptions and log the according
error message. For consistency, we'll use the std conversion
functions throughout ceph-dokan.
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The WNBD adapter may be reset in certain situations (e.g. driver
upgrade, MS WHQL tests, etc).
We're going to monitor the WNBD adapter using WMI[1] events, restarting
the rbd-wnbd disk mappings whenever necessary. Adapter monitoring can be
enabled by passing the --adapter-monitoring-enabled flag to the service.
This feature is optional for the following reasons:
* it's mainly used during development / driver certification
* we had to use a relatively small polling interval, which might imply
additional resource usage. WMI quotas also have to be considered.
While at it, we're updating two lambdas that are submitted to thread pools,
avoiding default reference capturing and explicitly specifying the variables
that get copied.
[1] https://learn.microsoft.com/en-us/windows/win32/wmisdk/wmi-start-page
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Windows APIs heavily use wchar. ceph-dokan and rbd-wnbd
have some duplicated helpers that convert wstrings to/from
utf8 strings.
To avoid duplication and allow reusing those helpers, we're moving
them to common/win32/wstring.h.
We're using the "win32" subfolder because it's unlikely that this
will ever be used on other platforms.
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |\
| |
| |
| |
| | |
dokan: Made file/dir access mode configurable
Reviewed-by: Venky Shankar <vshankar@redhat.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
By default, when mounting a filesystem using ceph-dokan, the file
and directory access modes were hardcoded as 755.
Now, both the file and directory access modes are configurable
using the --file-mode and --dir-mode optargs, accepting values
ranging from 001 to 777.
If no value is specified, the default value of 755 will be used.
Signed-off-by: Stefan Chivu <schivu@cloudbasesolutions.com>
|
| |/
|
|
|
|
|
| |
By mistake, src/dokan/dbg.cc is using the "rbd" logging subsystem.
We'll update it to use the "client" subsystem.
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
| |
Added ceph-dokan CLI optarg for configuring the value of
the maximum path length. By default, it will be set to 256 and
it will have a maximum value of 32767.
Signed-off-by: Stefan Chivu <schivu@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
| |
There was a typo in the function name cephfs_errno_to_ntsatus
defined in common/win32/errno.cc.
It has been replaced with cephfs_errno_to_ntstatus_map to avoid
conflicts with the existing cephfs_errno_to_ntstatus function.
Signed-off-by: Stefan Chivu <schivu@cloudbasesolutions.com>
|
| |
|
|
| |
Signed-off-by: Stefan Chivu <schivu@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
| |
Dokany v2 no longer allows specifying the worker thread count.
For this reason, we're deprecating this ceph-dokan parameter.
Dokany v2 does allow requesting single thread mode but that's probably
not worth exposing.
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
| |
Dokany v2 requires DokanInit() to be called before DokanMain().
Also, we need to call DokanShutdown() after unmounting the fs.
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
to fix the FTBFS due to following warning:
```
/home/jenkins-build/build/workspace/ceph-windows-pull-requests/ceph/build.deps/src/dokany/dokan/dokan.h:723:22: error: narrowing conversion of '-1' from 'int' to 'long unsigned int' [-Wnarrowing]
723 | #define DOKAN_ERROR -1
| ^
```
also, clean up the following warning:
```
/home/jenkins-build/build/workspace/ceph-windows-pull-requests/ceph/src/dokan/dbg.cc:142:62: warning: NULL used in arithmetic [-Wpointer-arith]
142 | o << "\n\tIsDirectory: " << (DokanFileInfo->IsDirectory != NULL);
|
```
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
to avoid the conflicting declaration of NTSTATUS from bcrypt.h and our
own typedef. as after switching to boost 1.79, we would have following compiling
failure:
In file included from ../src/dokan/options.cc:14:
../src/dokan/ceph_dokan.h:16:15: error: conflicting declaration 'typedef DWORD NTSTATUS'
16 | typedef DWORD NTSTATUS;
| ^~~~~~~~
In file included from ../build.deps/mingw/boost/include/boost/asio/impl/connect_pipe.ipp:29,
from ../build.deps/mingw/boost/include/boost/asio/connect_pipe.hpp:79,
from ../build.deps/mingw/boost/include/boost/asio.hpp:64,
from ../src/include/win32/winsock_wrapper.h:20,
from <command-line>:
/usr/share/mingw-w64/include/bcrypt.h:27:16: note: previous declaration as 'typedef LONG NTSTATUS'
27 | typedef LONG NTSTATUS,*PNTSTATUS;
| ^~~~~~~~
Signed-off-by: Kefu Chai <tchaikov@gmail.com>
|
| |
|
|
|
|
|
| |
A followup to PR #42820 that modified argv_to_vec() signature
(for style and performance).
Signed-off-by: Ronen Friedman <rfriedma@redhat.com>
|
| |
|
|
|
|
|
|
|
| |
* add "using" declarations in .cc files.
so we don't rely on "using namespace std" in one or more included
headers.
Signed-off-by: Kefu Chai <kchai@redhat.com>
|
| |
|
|
|
|
|
|
|
| |
This is a simple change that updates the logged mounted directory.
We're incorrectly using "ceph_getcwd" instead of the actual root
path.
Fixes: https://tracker.ceph.com/issues/49662
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
| |
"-m" is the short option for "--removable". We tried to keep the
syntax as close as possible to the old ceph-dokan project but this
flag is already used.
We'll drop the short option, keeping the long one ("--removable").
Fixes: https://tracker.ceph.com/issues/49662
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
At the moment, the Windows volume name and serial number are hardcoded.
This change makes the volume name configurable, defaulting to "Ceph"
or "Ceph - <fs_name>". This makes it easier to identify Ceph mounts.
At the same time, we're going to retrieve the Ceph filesystem identifier
instead of using the hardcoded value.
Fixes: https://tracker.ceph.com/issues/49662
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
At the moment, Windows CephFS mounts can only be removed by
terminating the daemon (e.g. sending CTRL-C) or through the
Windows mount manager if the "-o -m" parameters were passed
when the mapping was created.
This change adds the "ceph-dokan unmap" command, which takes
the mountpoint as input.
Fixes: https://tracker.ceph.com/issues/49662
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
|
|
In order to expose ceph filesystems to Windows hosts, we propose
including ceph-dokan[1][2] in the Ceph tree, while updating it to
work with the latest CephFS and Dokany APIs.
Dokany is a well maintained project (fork of the original Dokan
project), allowing filesystems to be implemented in userspace,
even providing a Fuse compatibility layer.
One reason for not using the FUSE compatibility layer is that it's
only covering the high level API while Ceph is using the low level
FUSE API, which among other things is inode centric.
Changes made by this patch compared to the upstream ceph-dokan:
* support latest stable Dokany API. The upstream version relies on
the legacy unmaintained Dokan API
* return proper error codes, converting standard errno.h values to
NTSTATUS
* minor changes to support latest cephfs API
* drop duplicated ceph code, no longer needed if we're to include it
in tree. This makes it much easier to maintain.
* drop redundant permission checks, leaving it up to libcephfs
* use ceph argparse helpers
* use ceph logging and daemon initialization
* fixed unicode handling
* switched to ceph coding style
* made ceph.conf param optional, using the default path if available
* enabled setting file timestamps
* append support
* configurable timeouts set once per mount
* ensure that the error code is always logged
* various cleanups (removed unused entry points, checks that have
been moved to dokany, simplified conditional statements,
unnecessary conversions in the hot path, etc).
[1] https://github.com/ketor/ceph-dokan
[2] https://github.com/ceph/ceph-dokan
Signed-off-by: Lucian Petrut <lpetrut@cloudbasesolutions.com>
|
