25 Oct, 2010
2 commits
-
Replace the BKL with a mutex to protect the venus_comm structure which
binds the mountpoint with the character device and holds the upcall
queues.Signed-off-by: Yoshihisa Abe
Signed-off-by: Jan Harkes
Signed-off-by: Linus Torvalds -
Now that shared inode state is locked using the cii->c_lock, the BKL is
only used to protect the upcall queues used to communicate with the
userspace cache manager. The remaining state is all local and we can
push the lock further down into coda_upcall().Signed-off-by: Yoshihisa Abe
Signed-off-by: Jan Harkes
Signed-off-by: Linus Torvalds
08 Aug, 2010
1 commit
-
CODA should not be using defines in the global name space of
that nature, prefix them with CODA_.Signed-off-by: Jens Axboe
28 Apr, 2010
1 commit
-
Otherwise we must export backing-dev.h as well, which doesn't make
any sense.Signed-off-by: Jens Axboe
22 Apr, 2010
1 commit
-
This ensures that dirty data gets flushed properly.
Signed-off-by: Jens Axboe
03 Feb, 2009
2 commits
-
These are only for kernel internals as pointed by Arnd Bergmann:
struct kstatfs
struct venus_comm
coda_vcp()Signed-off-by: Jaswinder Singh Rajput
-
fix the following 'make headers_check' warning:
usr/include/linux/coda_psdev.h:90: extern's make no sense in userspace
Signed-off-by: Jaswinder Singh Rajput
22 Jul, 2007
1 commit
-
This is an variation on the patch sent by Christoph Hellwig which kills
file_count abuse by the Coda kernel module by moving the coda_flush
functionality into coda_release. However part of reason we were using the
coda_flush callback was to allow Coda to pass errors that occur during
writeback from the userspace cache manager back to close().As Al Viro explained on linux-fsdevel, it is impossible to guarantee that
such errors can in fact be returned back to the caller. There are many
cases where the last reference to a file is not released by the close
system call and it is also impossible to pick some close as a 'last-close'
and delay it until all other references have been destroyed.The CODA_STORE/CODA_RELEASE upcall combination is clearly a broken design,
and it is better to remove it completely.Signed-off-by: Jan Harkes
Cc: Christoph Hellwig
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
20 Jul, 2007
3 commits
-
Similar information can easily be obtained with strace -c.
Signed-off-by: Jan Harkes
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The sb_info structure only contains a single pointer to the character device,
there is no need for the added indirection.Signed-off-by: Jan Harkes
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We ignore signals for about 30 seconds to give userspace a chance to see the
upcall. As we did not block signals we ended up in a busy loop for the
remainder of the period when a signal is received.Signed-off-by: Jan Harkes
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
24 Sep, 2006
1 commit
-
Signed-off-by: Jeff Garzik
23 Jun, 2006
1 commit
-
Give the statfs superblock operation a dentry pointer rather than a superblock
pointer.This complements the get_sb() patch. That reduced the significance of
sb->s_root, allowing NFS to place a fake root there. However, NFS does
require a dentry to use as a target for the statfs operation. This permits
the root in the vfsmount to be used instead.linux/mount.h has been added where necessary to make allyesconfig build
successfully.Interest has also been expressed for use with the FUSE and XFS filesystems.
Signed-off-by: David Howells
Acked-by: Al Viro
Cc: Nathan Scott
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
17 Apr, 2005
1 commit
-
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.Let it rip!