diff options
author | Tom Zanussi <zanussi@us.ibm.com> | 2006-01-08 01:02:23 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@g5.osdl.org> | 2006-01-08 20:13:49 -0800 |
commit | 6625b861f8f0e429902b8671b3e70792cd99074e (patch) | |
tree | 0a141b0c15d8ed012d971553acde9f7029adcfe6 /fs/relayfs/buffers.c | |
parent | b33291c0bcecfa44baa905964eec4b8815dcbcdf (diff) | |
download | op-kernel-dev-6625b861f8f0e429902b8671b3e70792cd99074e.zip op-kernel-dev-6625b861f8f0e429902b8671b3e70792cd99074e.tar.gz |
[PATCH] relayfs: decouple buffer creation from inode creation
The patch series implementa or fixes 3 things that were specifically requested
or suggested by relayfs users:
- support for non-relay files (patches 1-6)
Currently, the relayfs API only supports the creation of directories
(relayfs_create_dir()) and relay files (relay_open()). These patches adds
support for non-relay files (relayfs_create_file()). This is so relayfs
applications can create 'control files' in relayfs itself rather than in /proc
or via a netlink channel, as is currently done in the relay-app examples.
Basically what this amounts to is exporting relayfs_create_file() with an
additional file_ops param that clients can use to supply file operations for
their own special-purpose files in relayfs.
- make exported relay file ops useful (patches 7-8)
The relayfs relay_file_operations have always been exported, the intent being
to make it possible to create relay files in other filesystems such as
debugfs. The problem, though, is that currently the file operations are too
tightly coupled to relayfs to actually be used for this purpose. This patch
fixes that by adding a couple of callback functions that allow a client to
hook into relay_open()/close() and supply the files that will be used to
represent the channel buffers; the default implementation if no callbacks are
defined is to create the files in relayfs.
- add an option to create global relay buffer (patches 9-10) The file creation
callback also supplies an optional param, is_global, that can be used by
clients to create a single global relayfs buffer instead of the default
per-cpu buffers. This was suggested as being useful for certain debugging
applications where it's more convenient to be able to get all the data from a
single channel without having to go to the bother of dealing with per-cpu
files.
- cleanup, some renaming and Documentation updates (patches 11-12)
There were several comments that the use of netlink in the example code was
non-intuitive and in fact the whole relay-app business was needlessly
confusing. Based on that feedback, the example code has been completely
converted over to relayfs control files as supported by this patch, and have
also been made completely self-contained.
The converted examples along with a couple of new examples that demonstrate
using exported relay files can be found in relay-apps tarball:
http://prdownloads.sourceforge.net/relayfs/relay-apps-0.9.tar.gz?download
This patch:
Separate buffer create/destroy from inode create/destroy. We want to be able
to associate other data and not just relay buffers with inodes. Buffer
create/destroy is moved out of inode.c and into relayfs core code.
Signed-off-by: Tom Zanussi <zanussi@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'fs/relayfs/buffers.c')
-rw-r--r-- | fs/relayfs/buffers.c | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/fs/relayfs/buffers.c b/fs/relayfs/buffers.c index 84e21ff..667b529 100644 --- a/fs/relayfs/buffers.c +++ b/fs/relayfs/buffers.c @@ -186,4 +186,5 @@ void relay_remove_buf(struct kref *kref) { struct rchan_buf *buf = container_of(kref, struct rchan_buf, kref); relayfs_remove(buf->dentry); + relay_destroy_buf(buf); } |