We do some caching using a hash table indexed by the operations vector.
* Pass the vnode capabilities to the kernel module.
* In the kernel tailor the operation vectors for volumes and vnodes passed to
the VFS according to the respective capabilities. This way those vectors look
pretty much like those from the client FS. This saves unnecessary calls when
hooks are not implemented and should also fix compatibility problems in cases
where not implementing a hook and returning an error don't mean the same to
the VFS.
* Inlined some of the kernel module Volume class getters.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29572 a95241bf-73f2-0310-859d-f6bbb57e9c96
- Send() now also gets the message to send as parameter.
- Added methods to reserve and unreserve space in the buffer.
* RequestAllocator: Uses the port buffer reservation methods now. This allows
to let more than one RequestAllocator use a Port in a stack-like manner.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29565 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Change Port::Receive() semantics. It does no longer use the buffer associated
with the object. Instead it allocates heap memory for the caller. Since
there's only one caller who clones the message anyway, we save one copy this
way (though we have an additional syscall in userland) and don't overwrite
a request being prepared for sending.
* Changed RequestAllocator::ReadRequest() to also read the request from the
port.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29563 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Pass the request offset and length with the DoIORequest. This allows us to
already get the first vecs for do_iterative_fd_io() in userland, saving a
trip back and forth.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29546 a95241bf-73f2-0310-859d-f6bbb57e9c96
part.
* Implemented a good part of the kernel part for the support of
do_iterative_fd_io(). The forwarding of the callbacks are missing yet.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29540 a95241bf-73f2-0310-859d-f6bbb57e9c96
an on-stack structure for the request in a hash table when forwarding the
io() hook. Create a new object (HaikuKernelIORequest) instead and store it
in a hash table in HaikuKernelFileSystem.
* Some work on the userland side of do_iterative_fd_io().
* Added some missing requests in UserlandFSUtil::is_{kernel,userland}_request().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29505 a95241bf-73f2-0310-859d-f6bbb57e9c96
HashMap and HashSet classes to use the kernel utils OpenHashTable instead.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29429 a95241bf-73f2-0310-859d-f6bbb57e9c96
RecursiveLock in the kernel.
* Several adjustments according to UserlandFS header changes.
* Re-added reiserfs to image.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29410 a95241bf-73f2-0310-859d-f6bbb57e9c96
file cache are missing yet -- requires some refactoring in Volume.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29402 a95241bf-73f2-0310-859d-f6bbb57e9c96
/system/servers, now.
* Got rid of the dispatcher application. It was the entity the kernel module
talked to when mounting a yet unknown FS. It started the client FS server
applications. Now the kernel module starts them directly. They use
respectively named ports, so the kernel module can find them when already
running.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29377 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Split the FSCapabilities bit mask into three bit masks for the FS, the
volume, and the vnodes, since we have three independent FS interface
structures, now. The latter is not supported in the kernel add-on, yet.
* Server:
- Temporarily removed some things from the build (the BeOS interface
and the cache implementations).
- Some WIP in HaikuKernelVolume and [haiku_]kernel_emu.{h,cpp}.
- Added HaikuKernelNode, which wraps fs_vnode for the client FS.
The server is still quite a bit away from being buildable again.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29323 a95241bf-73f2-0310-859d-f6bbb57e9c96
kernel module compiles again, though newer features like VM file support,
asynchronous I/O, and FS layers aren't supported yet. The userland code is
still completely out of sync and won't even compile.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29318 a95241bf-73f2-0310-859d-f6bbb57e9c96
vfs_select.h, respectively moved most of it into the new kernel
private header wait_for_objects.h.
* Added new experimental API functions wait_for_objects[_etc](). They
work pretty much like poll(), but also for semaphores, ports, and
threads.
* Removed the "ref" parameter from notify_select_events() and the
select_sync_pool functions as well as from fd_ops::fd_[de]select(). It
is no longer needed. The FS interface select() hook still has it,
though -- the VFS will always pass 0.
* de]select_fd() take a select_info* instead of a select_sync* + ref
pair, now. Added respective functions for semaphores, ports, and
threads.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@22416 a95241bf-73f2-0310-859d-f6bbb57e9c96
rewind_query() -- are passed to the userland. get_vnode_name() has an
emulation in userland, in case the client FS doesn't implement it.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20522 a95241bf-73f2-0310-859d-f6bbb57e9c96
possibility to initialize the AutoLocker without locking the object even
if it is unlocked yet. Especially in loops Lock()/Unlock() come handy
when an otherwise constantly hold lock needs to be unlocked for a short
time.
I suppose we should move the kernel utils AutoLocker implementation
to headers/private/shared, and drop the less powerful ObjectLocker.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20405 a95241bf-73f2-0310-859d-f6bbb57e9c96
can be emulated -- and pass this info to the kernel add-on. Thus we can
avoid passing requests to the userland that can't be serviced anyway.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20331 a95241bf-73f2-0310-859d-f6bbb57e9c96
change.
* The new notification functions are used instead of send_notification()
and notify_listener() now. Mapped them in the BeOS kernel emulation
accordingly. RamFS node monitoring seems to work now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20298 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Made publish_vnode() available in userland. For old style FS add-ons
publish_vnode() is used when they request a new_vnode(). The semantics
of new_vnode() changed considerably in Haiku, but publish_vnode()
seems to do pretty much what the old new_vnode() did.
* The UserlandFS hosted RamFS begins to work under Haiku. It runs pretty
soon out of memory though (under vmware with 256 MB) and node
monitoring is broken ATM.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20264 a95241bf-73f2-0310-859d-f6bbb57e9c96
old interface is completely done in userland ATM.
It becomes more and more obvious that we probably need to provide
the kernel add-on with a bit more information about what the client FS
interface supports in the first place, so we can save unnecessary trips to
the userland. Opening/closing attributes for a FS using the old style
interface could be handled completely in the kernel add-on, for instance
(even if we lose a bit of accuracy wrt to open modes etc.).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20258 a95241bf-73f2-0310-859d-f6bbb57e9c96
need via library libuserlandfs_beos_kernel.so. Fine-tuned the legacy headers
so they can by used by the the kernel interface emulation code as well as by
the add-ons. This is actually a bit hacky, since we build everything in the
Haiku build environment and thus mix these old headers and Haiku's.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20254 a95241bf-73f2-0310-859d-f6bbb57e9c96
interface. The classes actually interfacing with the client FS add-on still
need to be adjusted.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20246 a95241bf-73f2-0310-859d-f6bbb57e9c96
FS interface. Adjusted old hooks, but didn't add the new ones yet. The
module builds now at least.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20235 a95241bf-73f2-0310-859d-f6bbb57e9c96
