- 04 Jan, 2011 1 commit
-
-
Jesper Juhl authored
Hi, In fs/nfs/proc.c::nfs_proc_symlink() we will leak memory if either nfs_alloc_fhandle() or nfs_alloc_fattr() returns NULL but the other one doesn't. This patch ensures memory allocated by one when the other fails is always released (this is safe since nfs_free_fattr() and nfs_free_fhandle() both call kfree which deals gracefully with NULL pointers). Signed-off-by:
Jesper Juhl <jj@chaosbits.net> Signed-off-by:
Trond Myklebust <Trond.Myklebust@netapp.com>
-
- 21 Dec, 2010 6 commits
-
-
Trond Myklebust authored
Signed-off-by: -
Joe Perches authored
Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Stanislav Kinsbursky authored
Update: added check for zero value as it was before (note: can't simply check mountd_port for positive value because it's typeof unsigned short) Default value for mount server port is set to NFS_UNSPEC_PORT (-1) and will not be changed during parsing mount options for mound data version 6. This default value will be showed for mountport in /proc/mounts always since current default check is for zero value. This small mistake leads to big problem, because during umount.nfs execution from old user-space utils (at least nfs-utils 1.0.9) this value will be used as the server port to connect to. This request will be rejected (since port is 65535) and thus nfs mount point can't be unmounted. Note from Chuck Lever (chuck.lever@oracle.com): this is only possible if /etc/mtab is a link to /proc/mounts. Not all systems have this configuration. Signed-off-by:
Stanislav Kinsbursky <skinsbursky@parallels.com> Signed-off-by:
-
J. Bruce Fields authored
Note that cl_lease_time is in jiffies. This can cause a very long wait in the NFS4ERR_CLID_INUSE case. Signed-off-by:
J. Bruce Fields <bfields@redhat.com> Signed-off-by:
-
Jesper Juhl authored
Take advantage of kmem_cache_zalloc() in nfs_page_alloc(). Save a call to memset() and a few bytes. Before: [jj@dragon linux-2.6]$ size fs/nfs/pagelist.o text data bss dec hex filename 1765 0 8 1773 6ed fs/nfs/pagelist.o After: [jj@dragon linux-2.6]$ size fs/nfs/pagelist.o text data bss dec hex filename 1749 0 8 1757 6dd fs/nfs/pagelist.o Signed-off-by: -
Tobias Klauser authored
IS_ERR() already implies unlikely(), so it can be omitted here. Signed-off-by:
Tobias Klauser <tklauser@distanz.ch> Signed-off-by:
-
- 16 Dec, 2010 33 commits
-
-
Chuck Lever authored
Clean up. The contents of the src_sap field is not used in nlm_alloc_host(). Signed-off-by:
Chuck Lever <chuck.lever@oracle.com> Signed-off-by:
-
Chuck Lever authored
Clean up. Remove the now unused helper nlm_lookup_host(). Signed-off-by:
-
Chuck Lever authored
Clean up. Signed-off-by:
-
Chuck Lever authored
Clean up. nlm_hosts now contains only server-side entries. Rename it to match convention of client side cache. Signed-off-by:
-
Chuck Lever authored
Clean up. Change nlmsvc_lookup_host() to be purpose-built for server-side nlm_host management. This replaces the generic nlm_lookup_host() helper function, just like on the client side. The lookup logic is specialized for server host lookups. The server side cache also gets its own specialized equivalent of the nlm_release_host() function. Signed-off-by:
-
Chuck Lever authored
NFS clients don't need the garbage collection processing that is performed on nlm_host structures. The client picks up an nlm_host at mount time and holds a reference to it until the file system is unmounted. Servers, on the other hand, don't have a precise way to tell when an nlm_host is no longer being used, so zero refcount nlm_host entries are left to expire in the cache after a time. Basically there's nothing holding a reference to an nlm_host between individual server-side NLM requests, but we can't afford the expense of recreating them for every new NLM request from a client. The nlm_host cache adds some lifetime hysteresis to entries in the cache so the next time a particular nlm_host is needed, it's likely to be discovered by a lookup rather than created from whole cloth. With the new implementation, client nlm_host cache items are no longer garbage collected, and are destroyed directly by a new release function specialized for client entries, nlmclnt_release_host(). They are cached in their own data structure, and have their own lookup logic, simplified and specialized for client nlm_host entries. However, the client nlm_host cache still shares reboot recovery logic with the server nlm_host cache. The NSM "peer rebooted" downcall for clients and servers still come through the same RPC call. This is a legacy formal API that would be difficult to alter, and besides, the user space NSM implementation can't tell the difference between peers that are clients or servers. For this reason, the client cache continues to share the nlm_host_mutex (and reboot recovery logic) with the server cache. Signed-off-by:
-
Chuck Lever authored
The nlm_release_call() function is invoked from both the server and the client side. We're about to introduce a distinct server- and client-side nlm_release_host(), so nlm_release_call() must first be split into a client-side and a server-side version. Signed-off-by:
-
Chuck Lever authored
Refactor the tail of nlm_gc_hosts() into nlm_destroy_host() so that this logic can be used separately from garbage collection. Rename it _locked() to document that it must be called with the hosts cache mutex held. Signed-off-by:
-
Chuck Lever authored
Refactor nlm_host allocation and initialization into a separate function. This will be the common piece of server and client nlm_host lookup logic after the nlm_host cache is split. Small change: use kmalloc() instead of kzalloc(), as we're overwriting almost all fields in the new nlm_host struct with non-zero values immediately after it is allocated. An added benefit is we now have an explicit reference to each field name where it is initialized (for all you cscope fans out there). Signed-off-by:
-
J. Bruce Fields authored
Minor reorganization; no change in behavior. This will save some duplicated code after we split the client and server host caches. Signed-off-by:
J. Bruce Fields <bfields@citi.umich.edu> [ cel: Forward-ported to 2.6.37 ] Signed-off-by:
-
J. Bruce Fields authored
We've got a lot of loops like this, and I find them a little easier to read with the macros. More such loops are coming. Signed-off-by:
-
Chuck Lever authored
Now that all client-side XDR decoder routines use xdr_streams, there should be no need to support the legacy calling sequence [rpc_rqst *, __be32 *, RPC res *] anywhere. We can construct an xdr_stream in the generic RPC code, instead of in each decoder function. This is a refactoring change. It should not cause different behavior. Signed-off-by:
-
Chuck Lever authored
Now that all client-side XDR encoder routines use xdr_streams, there should be no need to support the legacy calling sequence [rpc_rqst *, __be32 *, RPC arg *] anywhere. We can construct an xdr_stream in the generic RPC code, instead of in each encoder function. Also, all the client-side encoder functions return 0 now, making a return value superfluous. Take this opportunity to convert them to return void instead. This is a refactoring change. It should not cause different behavior. Signed-off-by:
-
Chuck Lever authored
Clean up. Just fixed a panic where the nrprocs field in a different upper layer client was set by hand incorrectly. Use the compiler-generated method used by the other upper layer protocols. Signed-off-by:
-
Chuck Lever authored
Clean up. The trend in the other XDR encoder functions is to BUG() when encoding problems occur, since a problem here is always due to a local coding error. Then, instead of a status, zero is unconditionally returned. Update the rpcbind XDR encoders to behave this way. To finish the update, use the new-style be32_to_cpup() and cpu_to_be32() macros, and compute the buffer sizes using raw integers instead of sizeof(). This matches the conventions used in other XDR functions. Signed-off-by:
-
Chuck Lever authored
Clean up. The UMNT request has a NULL response. There's no need to set up a mountres structure for it. Signed-off-by:
-
Chuck Lever authored
Clean up. The trend in the other XDR encoder functions is to BUG() when encoding problems occur, since a problem here is always due to a local coding error. Then, instead of a status, zero is unconditionally returned. Update the mount client XDR encoders to behave this way. To finish the update, use the new-style be32_to_cpup() and cpu_to_be32() macros, and compute the buffer sizes using raw integers instead of sizeof(). This matches the conventions used in other XDR functions. Signed-off-by:
-
Chuck Lever authored
Clean up. The trend in the other XDR encoder functions is to BUG() when encoding problems occur, since a problem here is always due to a local coding error. Then, instead of a status, zero is unconditionally returned. Update the NSM XDR encoders to behave this way. To finish the update, use the new-style be32_to_cpup() and cpu_to_be32() macros, and compute the buffer sizes using raw integers instead of sizeof(). This matches the conventions used in other XDR functions Signed-off-by:
-
Chuck Lever authored
Clean up. .../linux/nfs-2.6/fs/nfs/nfs4xdr.c: In function ‘decode_getdeviceinfo’: .../linux/nfs-2.6/fs/nfs/nfs4xdr.c:5008: warning: comparison between signed and unsigned integer expressions Signed-off-by:
-
Chuck Lever authored
Clean up. The pointer returned by ->decode_dirent() is no longer used as a pointer. The only call site (xdr_decode() in fs/nfs/dir.c) simply extracts the errno value encoded in the pointer. Replace the returned pointer with a standard integer errno return value. Also, pass the "server" argument as part of the nfs_entry instead of as a separate parameter. It's faster to derive "server" in nfs_readdir_xdr_to_array() since we already have the directory's inode handy. "server" ought to be invariant for a set of entries in the same directory, right? The legacy versions of decode_dirent() don't use "server" anyway, so it's wasted work for them to derive and pass "server" for each entry. Signed-off-by:
-
Chuck Lever authored
When computing the length of the header, be sure to include the four octets consumed by "count". Signed-off-by:
-
Chuck Lever authored
Clean up. nlmdbg_cookie2a() is used only in svclock.c. Signed-off-by:
-
Chuck Lever authored
Clean up. When I was making other changes in this area, checkscript.pl complained about the use of leading blanks in the PROC macros in the xdr files. Signed-off-by:
-
Chuck Lever authored
Clean up. Remove old-style NFSv4 XDR macros in favor of the style now used in fs/nfs/nfs4xdr.c. These were forgotten during the recent nfs4xdr.c rewrite. Additional whitespace cleanup adds to the size of this patch. Signed-off-by:
-
Chuck Lever authored
Clean up. Remove old-style NFSv4 XDR macros in favor of the style now used in fs/nfs/nfs4xdr.c. These were forgotten during the recent nfs4xdr.c rewrite. Signed-off-by:
-
Chuck Lever authored
We'd like to prevent local buffer overflows caused by malicious or broken servers. New xdr_stream style decoders can do that. For efficiency, we also want to be able to pass xdr_streams from call_encode() to all XDR encoding functions, rather than building an xdr_stream in every XDR encoding function in the kernel. Same idea as the NLM v3 XDR overhaul. Signed-off-by:
-
Chuck Lever authored
Clean up. Move the timestamp decoder to match the placement and naming conventions of the other helpers. Fold xdr_decode_fattr() into decode_fattr3(), which is now it's only user. Fold xdr_decode_wcc_attr() into decode_wcc_attr(), which is now it's only user. Signed-off-by:
-
Chuck Lever authored
Clean up. Remove unused legacy result decoder functions, and any now unused decoder helper functions. Signed-off-by:
-
Chuck Lever authored
The naming scheme of the new decoder functions, which follows the NFSv4 XDR decoder functions, is slightly different than the scheme used for the old functions. Rename the functions as a separate step to keep the patches clean. Signed-off-by:
-
Chuck Lever authored
We'd like to prevent local buffer overflows caused by malicious or broken servers. New xdr_stream style decoders can do that. For efficiency, we also eventually want to be able to pass xdr_streams from call_decode() to all XDR decoding functions, rather than building an xdr_stream in every XDR decoding function in the kernel. Static helper functions are left without the "inline" directive. This allows the compiler to choose automatically how to optimize these for size or speed. Signed-off-by:
-
Chuck Lever authored
Clean up. Move the timestamp and the sattr encoder to match the placement convention of the other helpers, update their coding style, and refresh their documenting comments. Signed-off-by:
-
Chuck Lever authored
Clean up. Remove unused legacy argument encoder functions, and any now unused encoder helper functions. Signed-off-by:
-
Chuck Lever authored
The naming scheme of the new encoder functions, which follows the NFSv4 XDR encoder functions, is slightly different than the scheme used for the old functions. Rename the functions as a separate step to keep the patches clean. Signed-off-by:
-
