- 22 Jun, 2005 40 commits
-
-
Trond Myklebust authored
If the lock blocks, the server may send us a GRANTED message that races with the reply to our LOCK request. Make sure that we catch the GRANTED by queueing up our request on the nlm_blocked list before we send off the first LOCK rpc call. Signed-off-by:Trond Myklebust <Trond.Myklebust@netapp.com>
-
Trond Myklebust authored
Signed-off-by: -
Trond Myklebust authored
Signed-off-by: -
Trond Myklebust authored
Signed-off-by: -
Trond Myklebust authored
Basically copies the VFS's method for tracking writebacks and applies it to the struct nfs_page. Signed-off-by: -
Trond Myklebust authored
Use stable writes if we can see that we are only going to put a single write on the wire. Signed-off-by: -
Trond Myklebust authored
Even if the file is open for writes. Signed-off-by: -
Trond Myklebust authored
Unless we're doing O_APPEND writes, we really don't care about revalidating the file length. Just make sure that we catch any page cache invalidations. Signed-off-by: -
Trond Myklebust authored
Instead of looking at whether or not the file is open for writes before we accept to update the length using the server value, we should rather be looking at whether or not we are currently caching any writes. Failure to do so means in particular that we're not updating the file length correctly after obtaining a POSIX or BSD lock. Signed-off-by: -
Trond Myklebust authored
If we do not hold a valid stateid that is open for writes, there is little point in doing an extra open of the file, as the RFC does not appear to mandate this... Make setattr use the correct stateid if we're holding mandatory byte range locks. Signed-off-by: -
Trond Myklebust authored
Signed-off-by: -
Trond Myklebust authored
Signed-off-by: -
Olivier Galibert authored
NFSv3 currently returns the unsigned 64-bit cookie directly to userspace. The following patch causes the kernel to generate loff_t offsets for the benefit of userland. The current server-generated READDIR cookie is cached in the nfs_open_context instead of in filp->f_pos, so we still end up work correctly under directory insertions/deletion. Signed-off-by:
Olivier Galibert <galibert@pobox.com> Signed-off-by:
-
Chuck Lever authored
Make the socket transport kick the event queue to start socket connects immediately. This should improve responsiveness of applications that are sensitive to slow mount operations (like automounters). We are now also careful to cancel the connect worker before destroying the xprt. This eliminates a race where xprt_destroy can finish before the connect worker is even allowed to run. Test-plan: Destructive testing (unplugging the network temporarily). Connectathon with UDP and TCP. Hard-code impossibly small connect timeout. Version: Fri, 29 Apr 2005 15:32:01 -0400 Signed-off-by:
Chuck Lever <cel@netapp.com> Signed-off-by:
-
Chuck Lever authored
When the network layer reports a connection close, the RPC task waiting to reconnect should be notified so it can retry immediately instead of waiting for the normal connection establishment timeout. This reverts a change made in 2.6.6 as part of adding client support for RPC over TCP socket idle timeouts. Test-plan: Destructive testing with NFS over TCP mounts. Version: Fri, 29 Apr 2005 15:31:46 -0400 Signed-off-by:
-
Trond Myklebust authored
Cancel autodisconnect requests inside xprt_transmit() in order to avoid races. Use more efficient del_singleshot_timer_sync() Signed-off-by: -
Trond Myklebust authored
For internal purposes, the rpc_clnt_sigmask() call is replaced by a call to rpc_task_sigmask(), which ensures that the current task sigmask respects both the client cl_intr flag and the per-task NOINTR flag. Problem noted by Jiaying Zhang. Signed-off-by: -
Trond Myklebust authored
The changeset "trond.myklebust@fys.uio.no|ChangeSet|20050322152404|16979" (RPC: Ensure XDR iovec length is initialized correctly in call_header) causes the NFSv4 callback code to BUG() due to an incorrectly initialized scratch buffer. Signed-off-by: -
Reuben Farrelly authored
From: Reuben Farrelly <reuben-lkml@reub.net> With gcc-4.0: fs/nfs/nfs4proc.c:2976: error: static declaration of 'nfs4_file_inode_operations' follows non-static declaration fs/nfs/nfs4_fs.h:179: error: previous declaration of 'nfs4_file_inode_operations' was here Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
-
Andrew Morton authored
Older gcc's don't like this. fs/nfs/nfs4proc.c:2194: field `data' has incomplete type Signed-off-by:
-
Adrian Bunk authored
The Coverity checker noticed that such a simplification was possible. Signed-off-by:
Adrian Bunk <bunk@stusta.de> Signed-off-by:
-
Andreas Gruenbacher authored
* Pointer arithmetic bug: p is in word units. This fixes a memory corruption with big acls. * Initialize pg_class to prevent a NULL pointer access. Signed-off-by:
Andreas Gruenbacher <agruen@suse.de> Signed-off-by:
-
Trond Myklebust authored
Initialize the inode cache values correctly. Clean up __nfs3_forget_cached_acls() Signed-off-by: -
Andreas Gruenbacher authored
Attach acls to inodes in the icache to avoid unnecessary GETACL RPC round-trips. As long as the client doesn't retrieve any acls itself, only the default acls of exiting directories and the default and access acls of new directories will end up in the cache, which preserves some memory compared to always caching the access and default acl of all files. Signed-off-by:
Olaf Kirch <okir@suse.de> Signed-off-by:
-
Andreas Gruenbacher authored
NFSv3 has no concept of a umask on the server side: The client applies the umask locally, and sends the effective permissions to the server. This behavior is wrong when files are created in a directory that has a default ACL. In this case, the umask is supposed to be ignored, and only the default ACL determines the file's effective permissions. Usually its the server's task to conditionally apply the umask. But since the server knows nothing about the umask, we have to do it on the client side. This patch tries to fetch the parent directory's default ACL before creating a new file, computes the appropriate create mode to send to the server, and finally sets the new file's access and default acl appropriately. Many thanks to Buck Huppmann <buchk@pobox.com> for sending the initial version of this patch, as well as for arguing why we need this change. Signed-off-by:
-
Andreas Gruenbacher authored
This adds acl support fo nfs clients via the NFSACL protocol extension, by implementing the getxattr, listxattr, setxattr, and removexattr iops for the system.posix_acl_access and system.posix_acl_default attributes. This patch implements a dumb version that uses no caching (and thus adds some overhead). (Another patch in this patchset adds caching as well.) Signed-off-by:
-
Andreas Gruenbacher authored
This adds functions for encoding and decoding POSIX ACLs for the NFSACL protocol extension, and the GETACL and SETACL RPCs. The implementation is compatible with NFSACL in Solaris. Signed-off-by:
-
Andreas Gruenbacher authored
The NFS and NFSACL programs run on the same RPC transport. This patch adds support for this by converting svc_program into a chained list of programs (server-side). Signed-off-by:
-
Andreas Gruenbacher authored
Add the missing NFS3ERR_NOTSUPP error code (defined in NFSv3) to the system-to-protocol-error table in nfsd. The nfsacl extension uses this error code. Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Trond Myklebust authored
Signed-off-by: -
Olaf Kirch authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
Andreas Gruenbacher authored
Signed-off-by:
-
J. Bruce Fields authored
Currently we return -ENOMEM for every single failure to create a new auth. This is actually accurate for auth_null and auth_unix, but for auth_gss it's a bit confusing. Allow rpcauth_create (and the ->create methods) to return errors. With this patch, the user may sometimes see an EINVAL instead. Whee. Signed-off-by:
J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by:
-
J. Bruce Fields authored
We shouldn't be silently falling back from krb5p to krb5i. Signed-off-by:
-
J. Bruce Fields authored
Add nfs4_acl field to the nfs_inode, and use it to cache acls. Only cache acls of size up to a page. Also prepare for up to a page of acl data even when the user doesn't pass in a buffer, as when they want to get the acl length to decide what size buffer to allocate. Signed-off-by:
-
J. Bruce Fields authored
Client-side write support for NFSv4 ACLs. Signed-off-by:
-
J. Bruce Fields authored
Client-side support for NFSv4 acls: xdr encoding and decoding routines for writing acls Signed-off-by:
-
J. Bruce Fields authored
Client-side support for NFSv4 ACLs. Exports the raw xdr code via the system.nfs4_acl extended attribute. It is up to userspace to decode the acl (and to provide correctly xdr'd acls on setxattr), and to convert to/from POSIX ACLs if desired. This patch provides only the read support. Signed-off-by:
-
