bfs_file_rename() and remove the bfs_file_delete() from bfs_file_rename().
After calling bfs_file_rename() it was too late to set "tnode->removed"
as the inode already disappeared.
- Use (parent_vnode, name, name_len) as key.
- Change smbfs_nget() to return a referenced but unlocked vnode and
adapt smbfs_setroot(), smbfs_create(), smbfs_mkdir() and smbfs_lookup().
(derived from clock_subr.h). Keep clock_subr.h with the kernel structures
and functions to reduce diffs, and have clock.h only include standalone
constants and macros.
- Take the vnode attributes from the NAME attribute.
- Remove now unused ntfs_times() and ntfs_filesize().
- Treat nodes without an unnamed DATA attribute as zero-sized regular files.
- Fix ntfs_loadntnode() to always read full cluster to prevent buffer
cache inconsistency.
- Change ntfs_vgetex() and ntfs_fget() to take a const attrname and
always pass a possibly empty string.
- Change printf to dprintf.
- Get rid of dead fields i_next, i_prev of struct ntnode.
There are a few situations where we must take care of the cache if direct
I/O was enabled:
- if we do direct I/O for write but not for read, then any write must
invalidate the cache so that a reader gets the written data and not
the not-updated cache.
- if we used a vnode without direct I/O and it is enabled for writing,
we must flush the cache before compeling the open operation, so that
the cachec write are not lost.
And at inactive time, we wipe direct I/O flags so that a new open without
direct I/O does not inherit direct I/O.
When puffs_cookie2vnode() misses an entry, vcache_get()
creates a new node (puffs_vfsop_loadvnode being called to
initialize the PUFFS part), then it discovers it is VNON,
and tries to vrele() it. vrele() calls VOP_INACTIVE(),
which led us in puffs_vnop_inactive() where we sent a
request to the filesystem for a node that already had been
reclaimed.
The fix is to check for VNON nodes in puffs_vnop_inactive()
and to return without doing anyting. This is suboptimal, but
a better workaround would probably need to modify vcache API,
with an impact on other filesystems. Let us keep it simple.
fail without taking down the kernel. It seems this is not the case anymore,
hence we can stop dropping errors in puffs_vnop_strategy()
Approved by pooka@
- Calling getnewvnode() with "mp == NULL" is wrong. Stop attaching a
vnode to system file nodes and change nilfs_bread() to translate
the block address and then uyse the device for the read.
- Move the vnode initialisation to nilfs_get_node() and use
nilfs_get_node_raw() to initialise the nilfs node only.
- Same for nilfs_reclaim() versus nilfs_dispose_node().
- Change nilfs_get_node() to return an unlocked vnode instead of
a nilfs node with locked vnode. Adapt nilfs_lookup() and nilfs_root().
- Don't treat unsupported node types (blk, chr ...) as regular,
return ENXIO instead.
- Fix nilfs_getattr() to mask the mode with ALLPERMS.
- Destroy sync_cv before free.
1) do it only if the file is open for writing, otherwise we send write
requests to the FS on a file that has never been open.
2) do it inside existing if (vap->va_size != VNOVAL) block
In the PUFFSDEBUG case, wrap do { ... } while (/*CONSTCOND*/0)
around the definitions. In the non-PUFFSDEBUG case, define them
as ((void)0) instead of as empty.
