sunrpc: teach the SUNRPC layer how to speak SMB
Subsystems (4)
cover sunrpc: teach the SUNRPC layer how to speak SMB
This patchset is still preliminary and is just an RFC...
First, some background. When I was at Connectathon this year, Trond mentioned an interesting idea to me. He said (paraphrasing):
"Why doesn't CIFS just use the RPC layer for transport? It's very efficient at shoveling bits out onto the wire. You'd just need to abstract out the XDR/RPC specific bits."
The idea has been floating around in the back of my head until recently, when I decided to give it a go. This patchset represents a first, rough draft at a patchset to do this. There's also a proof of concept module that demonstrates that it works as expected.
The patchset is still very rough. Reconnect behavior is still very "RPC-ish", for instance. There are doubtless other changes that will need to be made before I had anything merge-worthy.
At this point, I'm just interested in feedback on the general approach. Should I continue to pursue this or is it a non-starter for some reason?
The next step here, obviously is to start building a fs on top of it. I'd particularly be interested in using this as a foundation of a new smb2fs.
I've also got this patchset up on my public git tree:
http://git.samba.org/?p=jlayton/cifs.git;a=summary
Here are some questions I anticipate on this, and their answers:
------------------------------------------------------------------------
Q: Are you insane? Why would you attempt to do this?
A: Maybe...but in the last couple of years, I've spent a substantial amount of time working on the CIFS code. Much of that time has been spent fixing bugs. Many of those bugs exist in the low-level transport code which has been hacked on, kludged around and hand tweaked to where it is today. Unfortunately that's made it a very hard to work on mess. This drives away potential developers.
CIFS in particular is also designed around synchronous ops, which seriously limits throughput. Retrofitting it for asynchronous operation will be adding even more kludges. The sunrpc code is already fundamentally asynchronous.
------------------------------------------------------------------------
Q: Okay, what's the benefit of hooking it up to sunrpc rather than building a new transport layer (or fixing the transport in the other two filesystems)?
A: Using sunrpc allows us to share a lot of the rpc scheduler code with sunrpc. At a high level, NFS/RPC and SMB aren't really very different. Sure, they have different formats, and one is big endian on the wire and the other isn't...still there are significant similarities.
We also get access to the great upcall mechanisms that sunrpc has, and the possibility to share code like the gssapi upcalls. The sunrpc layer has a credential and authentication management framework that we can build on to make a truly multiuser CIFS/SMB filesystem.
I've heard it claimed before that Linux's sunrpc layer is over-engineered, but in this case that works in our favor...
------------------------------------------------------------------------
Q: can we hook up cifs or smbfs to use this as a transport?
A: Not trivially. CIFS in particular is not designed with each call having discrete encode and decode functions. They're sort of mashed together. smbfs might be possible...I'm a little less familiar with it, but it does have a transport layer that more closely resembles the sunrpc one. Still though, it'd take significant work to make that happen. I'm not opposed to the idea however.
In the end though, I think we'll probably need to design something new to sit on top of this. We will probably be able to borrow code and concepts from the other filesystems however.
------------------------------------------------------------------------
Q: could we use this as a transport layer for a smb2fs ?
A: Yes, I think so. This particular prototype is build around SMB1, but SMB2 could be supported with only minor modifications. One of the reasons for sending this patchset now before I've built a filesystem on top of it is because I know that SMB2 work is in progress. I'd like to see it based around a more asynchronous transport model, or at least built with cleaner layering so that we can eventually bolt on a different transport layer if we so choose.
Jeff Layton (9):
sunrpc: abstract out encoding function at rpc_clnt level
sunrpc: move check for too small reply size into rpc_verify_header
sunrpc: abstract our call decoding routine
sunrpc: move rpc_xdr_buf_init to clnt.h
sunrpc: make call_bind non-static
sunrpc: add new SMB transport class for sunrpc
sunrpc: add encoding and decoding routines for SMB
sunrpc: add Kconfig option for CONFIG_SUNRPC_SMB
smbtest: simple module for testing SMB/RPC code
fs/Makefile | 2 + fs/lockd/host.c | 4 + fs/lockd/mon.c | 4 + fs/nfs/client.c | 4 + fs/nfs/mount_clnt.c | 4 + fs/nfsd/nfs4callback.c | 4 + fs/smbtest/Makefile | 1 + fs/smbtest/smbtest.c | 204 +++++ include/linux/sunrpc/clnt.h | 24 +- include/linux/sunrpc/smb.h | 42 + include/linux/sunrpc/xprtsmb.h | 59 ++ net/sunrpc/Kconfig | 11 + net/sunrpc/Makefile | 1 + net/sunrpc/clnt.c | 98 ++- net/sunrpc/rpcb_clnt.c | 8 + net/sunrpc/smb.c | 120 +++ net/sunrpc/sunrpc_syms.c | 3 + net/sunrpc/xprtsmb.c | 1723 ++++++++++++++++++++++++++++++++++++++++ 18 files changed, 2272 insertions(+), 44 deletions(-) create mode 100644 fs/smbtest/Makefile create mode 100644 fs/smbtest/smbtest.c create mode 100644 include/linux/sunrpc/smb.h create mode 100644 include/linux/sunrpc/xprtsmb.h create mode 100644 net/sunrpc/smb.c create mode 100644 net/sunrpc/xprtsmb.c
RFC PATCH 1/9 sunrpc: abstract out encoding function at rpc_clnt level
In order to add the ability to do an SMB call with the sunrpc layer, we need to abstract out the call encoding. Add a function pointer that hangs off of the rpc_clnt to do the encoding. For now, all of the existing callers will simply set it to rpc_xdr_encode.
--- fs/lockd/host.c | 3 +++ fs/lockd/mon.c | 3 +++ fs/nfs/client.c | 3 +++ fs/nfs/mount_clnt.c | 3 +++ fs/nfsd/nfs4callback.c | 3 +++ include/linux/sunrpc/clnt.h | 7 +++++++ net/sunrpc/clnt.c | 14 ++++++++++---- net/sunrpc/rpcb_clnt.c | 6 ++++++ 8 files changed, 38 insertions(+), 4 deletions(-)
fs/lockd/host.c +3 -0
@@ -362,6 +362,9 @@ nlm_bind_host(struct nlm_host *host) .program = &nlm_program, .version = host->h_version, .authflavor = RPC_AUTH_UNIX, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_AUTOBIND), };
fs/lockd/mon.c +3 -0
@@ -75,6 +75,9 @@ static struct rpc_clnt *nsm_create(void) .program = &nsm_program, .version = NSM_VERSION, .authflavor = RPC_AUTH_NULL, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, .flags = RPC_CLNT_CREATE_NOPING, };
fs/nfs/client.c +3 -0
@@ -606,6 +606,9 @@ static int nfs_create_rpc_client(struct nfs_client *clp, .servername = clp->cl_hostname, .program = &nfs_program, .version = clp->rpc_ops->version, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, .authflavor = flavor, };
fs/nfs/mount_clnt.c +3 -0
@@ -159,6 +159,9 @@ int nfs_mount(struct nfs_mount_request *info) .servername = info->hostname, .program = &mnt_program, .version = info->version, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, .authflavor = RPC_AUTH_UNIX, };struct rpc_clnt *mnt_clnt;
fs/nfsd/nfs4callback.c +3 -0
@@ -494,6 +494,9 @@ int setup_callback_client(struct nfs4_client *clp) .authflavor = clp->cl_flavor, .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_QUIET), .client_name = clp->cl_principal, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, };struct rpc_clnt *client;
include/linux/sunrpc/clnt.h +7 -0
@@ -63,6 +63,9 @@ struct rpc_clnt {struct rpc_program * cl_program;char cl_inline_name[32 ];char *cl_principal;/* target to authenticate to */ +void (*cl_encode) (struct rpc_task *task); +size_t cl_callhdr_sz;/* in quadwords */ +size_t cl_replhdr_sz;/* in quadwords */ };/*
@@ -115,6 +118,9 @@ struct rpc_create_args {unsigned long flags;char *client_name;struct svc_xprt *bc_xprt;/* NFSv4.1 backchannel */ +size_t callhdr_size;/* in quadwords */ +size_t replhdr_size;/* in quadwords */ +void (*encode) (struct rpc_task *task); };/* Values for "flags" field */
@@ -150,6 +156,7 @@ struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred, int flags);void rpc_restart_call_prepare (struct rpc_task *);void rpc_restart_call (struct rpc_task *); +void rpc_xdr_encode (struct rpc_task *);void rpc_setbufsize (struct rpc_clnt *,unsigned int ,unsigned int );size_t rpc_max_payload (struct rpc_clnt *);void rpc_force_rebind (struct rpc_clnt *);
net/sunrpc/clnt.c +10 -4
@@ -200,6 +200,10 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args, stru clnt->cl_vers = version->number; clnt->cl_stats = program->stats; clnt->cl_metrics =rpc_alloc_iostats (clnt); + clnt->cl_encode = args->encode; + clnt->cl_callhdr_sz = args->callhdr_size; + clnt->cl_replhdr_sz = args->replhdr_size; + err = -ENOMEM ;if (clnt->cl_metrics ==NULL ) goto out_no_stats;
@@ -905,6 +909,7 @@ call_allocate(struct rpc_task *task)unsigned int slack = task->tk_msg.rpc_cred->cr_auth->au_cslack;struct rpc_rqst *req = task->tk_rqstp;struct rpc_xprt *xprt = task->tk_xprt; +struct rpc_clnt *clnt = task->tk_client;struct rpc_procinfo *proc = task->tk_msg.rpc_proc;dprint_status (task);
@@ -926,9 +931,9 @@ call_allocate(struct rpc_task *task)* and reply headers, and convert both values * to byte sizes. */ - req->rq_callsize =RPC_CALLHDRSIZE + (slack <<1 ) + proc->p_arglen; + req->rq_callsize = clnt->cl_callhdr_sz + (slack <<1 ) + proc->p_arglen; req->rq_callsize <<=2 ; - req->rq_rcvsize =RPC_REPHDRSIZE + slack + proc->p_replen; + req->rq_rcvsize = clnt->cl_replhdr_sz + slack + proc->p_replen; req->rq_rcvsize <<=2 ; req->rq_buffer = xprt->ops->buf_alloc(task,
@@ -975,7 +980,7 @@ rpc_xdr_buf_init(struct xdr_buf *buf, void *start, size_t len)/* * 3. Encode arguments of an RPC call */ -static void +void rpc_xdr_encode (struct rpc_task *task) {struct rpc_rqst *req = task->tk_rqstp;
@@ -1005,6 +1010,7 @@ rpc_xdr_encode(struct rpc_task *task) task->tk_status =rpcauth_wrap_req (task, encode, req, p, task->tk_msg.rpc_argp); } +EXPORT_SYMBOL_GPL (rpc_xdr_encode);/* * 4. Get the server port number if not yet set
@@ -1148,7 +1154,7 @@ call_transmit(struct rpc_task *task)/* Encode here so that rpcsec_gss can use correct sequence number. */ if (rpc_task_need_encode (task)) {BUG_ON (task->tk_rqstp->rq_bytes_sent !=0 ); -rpc_xdr_encode (task); + task->tk_client->cl_encode (task);/* Did the encode result in an error condition? */ if (task->tk_status !=0 ) {/* Was the error nonfatal? */
net/sunrpc/rpcb_clnt.c +6 -0
@@ -174,6 +174,9 @@ static struct rpc_clnt *rpcb_create_local(struct sockaddr *addr, .program = &rpcb_program, .version = version, .authflavor = RPC_AUTH_UNIX, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, .flags = RPC_CLNT_CREATE_NOPING, };
@@ -191,6 +194,9 @@ static struct rpc_clnt *rpcb_create(char *hostname, struct sockaddr *srvaddr, .program = &rpcb_program, .version = version, .authflavor = RPC_AUTH_UNIX, + .encode = rpc_xdr_encode, + .callhdr_size = RPC_CALLHDRSIZE, + .replhdr_size = RPC_REPHDRSIZE, .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_NONPRIVPORT), };
Signature · 1 line
1.6.0.6
RFC PATCH 2/9 sunrpc: move check for too small reply size into rpc_verify_header
In reply to Jeff Layton...this should introduce no behavioral changes.
--- net/sunrpc/clnt.c | 24 ++++++++++++------------ 1 files changed, 12 insertions(+), 12 deletions(-)
net/sunrpc/clnt.c +12 -12
@@ -1411,18 +1411,6 @@ call_decode(struct rpc_task *task)WARN_ON (memcmp (&req->rq_rcv_buf, &req->rq_private_buf,sizeof (req->rq_rcv_buf)) !=0 ); -if (req->rq_rcv_buf.len <12 ) { -if (!RPC_IS_SOFT (task)) { - task->tk_action = call_bind; - clnt->cl_stats->rpcretrans++; - goto out_retry; - } -dprintk ("RPC: %s: too small RPC reply size (%d bytes)\n" , - clnt->cl_protname, task->tk_status); - task->tk_action = call_timeout; - goto out_retry; - } - p =rpc_verify_header (task);if (IS_ERR (p)) {if (p ==ERR_PTR (-EAGAIN ))
@@ -1518,6 +1506,18 @@ rpc_verify_header(struct rpc_task *task)u32 n;int error = -EACCES ; +if (task->tk_rqstp->rq_rcv_buf.len <12 ) { +if (!RPC_IS_SOFT (task)) { + task->tk_action = call_bind; + task->tk_client->cl_stats->rpcretrans++; + goto out_retry; + } +dprintk ("RPC: %s: too small RPC reply size (%d bytes)\n" , + task->tk_client->cl_protname, task->tk_status); + task->tk_action = call_timeout; + goto out_retry; + } +if ((task->tk_rqstp->rq_rcv_buf.len &3 ) !=0 ) {/* RFC-1014 says that the representation of XDR data must be a * multiple of four bytes
Signature · 1 line
1.6.0.6
RFC PATCH 3/9 sunrpc: abstract our call decoding routine
In reply to Jeff Layton...all of the existing rpc users will use the new rpc_xdr_decode function.
--- fs/lockd/host.c | 1 + fs/lockd/mon.c | 1 + fs/nfs/client.c | 1 + fs/nfs/mount_clnt.c | 1 + fs/nfsd/nfs4callback.c | 1 + include/linux/sunrpc/clnt.h | 3 ++ net/sunrpc/clnt.c | 46 +++++++++++++++++++++++++++++------------- net/sunrpc/rpcb_clnt.c | 2 + 8 files changed, 42 insertions(+), 14 deletions(-)
fs/lockd/host.c +1 -0
@@ -363,6 +363,7 @@ nlm_bind_host(struct nlm_host *host) .version = host->h_version, .authflavor = RPC_AUTH_UNIX, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, .flags = (RPC_CLNT_CREATE_NOPING |
fs/lockd/mon.c +1 -0
@@ -76,6 +76,7 @@ static struct rpc_clnt *nsm_create(void) .version = NSM_VERSION, .authflavor = RPC_AUTH_NULL, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, .flags = RPC_CLNT_CREATE_NOPING,
fs/nfs/client.c +1 -0
@@ -607,6 +607,7 @@ static int nfs_create_rpc_client(struct nfs_client *clp, .program = &nfs_program, .version = clp->rpc_ops->version, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, .authflavor = flavor,
fs/nfs/mount_clnt.c +1 -0
@@ -160,6 +160,7 @@ int nfs_mount(struct nfs_mount_request *info) .program = &mnt_program, .version = info->version, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, .authflavor = RPC_AUTH_UNIX,
fs/nfsd/nfs4callback.c +1 -0
@@ -495,6 +495,7 @@ int setup_callback_client(struct nfs4_client *clp) .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_QUIET), .client_name = clp->cl_principal, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, };
include/linux/sunrpc/clnt.h +3 -0
@@ -64,6 +64,7 @@ struct rpc_clnt {char cl_inline_name[32 ];char *cl_principal;/* target to authenticate to */ void (*cl_encode) (struct rpc_task *task); +int (*cl_decode) (struct rpc_task *task);size_t cl_callhdr_sz;/* in quadwords */ size_t cl_replhdr_sz;/* in quadwords */ };
@@ -121,6 +122,7 @@ struct rpc_create_args {size_t callhdr_size;/* in quadwords */ size_t replhdr_size;/* in quadwords */ void (*encode) (struct rpc_task *task); +int (*decode) (struct rpc_task *task); };/* Values for "flags" field */
@@ -157,6 +159,7 @@ struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred,void rpc_restart_call_prepare (struct rpc_task *);void rpc_restart_call (struct rpc_task *);void rpc_xdr_encode (struct rpc_task *); +int rpc_xdr_decode (struct rpc_task *);void rpc_setbufsize (struct rpc_clnt *,unsigned int ,unsigned int );size_t rpc_max_payload (struct rpc_clnt *);void rpc_force_rebind (struct rpc_clnt *);
net/sunrpc/clnt.c +32 -14
74 lines not on this page · raw message
net/sunrpc/rpcb_clnt.c +2 -0
@@ -175,6 +175,7 @@ static struct rpc_clnt *rpcb_create_local(struct sockaddr *addr, .version = version, .authflavor = RPC_AUTH_UNIX, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, .flags = RPC_CLNT_CREATE_NOPING,
@@ -195,6 +196,7 @@ static struct rpc_clnt *rpcb_create(char *hostname, struct sockaddr *srvaddr, .version = version, .authflavor = RPC_AUTH_UNIX, .encode = rpc_xdr_encode, + .decode = rpc_xdr_decode, .callhdr_size = RPC_CALLHDRSIZE, .replhdr_size = RPC_REPHDRSIZE, .flags = (RPC_CLNT_CREATE_NOPING |
Signature · 1 line
1.6.0.6
RFC PATCH 4/9 sunrpc: move rpc_xdr_buf_init to clnt.h
In reply to Jeff Layton...so that it can be used elsewhere.
--- include/linux/sunrpc/clnt.h | 13 ++++++++++++- net/sunrpc/clnt.c | 12 ------------ 2 files changed, 12 insertions(+), 13 deletions(-)
include/linux/sunrpc/clnt.h +12 -1
@@ -165,7 +165,6 @@ size_t rpc_max_payload(struct rpc_clnt *);void rpc_force_rebind (struct rpc_clnt *);size_t rpc_peeraddr (struct rpc_clnt *,struct sockaddr *,size_t );const char *rpc_peeraddr2str (struct rpc_clnt *,enum rpc_display_format_t ); -size_t rpc_ntop (const struct sockaddr *,char *,const size_t );size_t rpc_pton (const char *,const size_t ,struct sockaddr *,const size_t );
@@ -197,6 +196,18 @@ static inline void rpc_set_port(struct sockaddr *sap, } } +static inline void +rpc_xdr_buf_init (struct xdr_buf *buf,void *start,size_t len) +{ + buf->head[0 ].iov_base = start; + buf->head[0 ].iov_len = len; + buf->tail[0 ].iov_len =0 ; + buf->page_len =0 ; + buf->flags =0 ; + buf->len =0 ; + buf->buflen = len; +} +#define IPV6_SCOPE_DELIMITER '%'#define IPV6_SCOPE_ID_LEN sizeof("%nnnnnnnnnn")
net/sunrpc/clnt.c +0 -12
@@ -966,18 +966,6 @@ rpc_task_force_reencode(struct rpc_task *task) task->tk_rqstp->rq_bytes_sent =0 ; } -static inline void -rpc_xdr_buf_init (struct xdr_buf *buf,void *start,size_t len) -{ - buf->head[0 ].iov_base = start; - buf->head[0 ].iov_len = len; - buf->tail[0 ].iov_len =0 ; - buf->page_len =0 ; - buf->flags =0 ; - buf->len =0 ; - buf->buflen = len; -} -/* * 3. Encode arguments of an RPC call */
Signature · 1 line
1.6.0.6
RFC PATCH 5/9 sunrpc: make call_bind non-static
In reply to Jeff Layton...is there a better way to do this?
--- include/linux/sunrpc/clnt.h | 1 + net/sunrpc/clnt.c | 4 ++-- 2 files changed, 3 insertions(+), 2 deletions(-)
include/linux/sunrpc/clnt.h +1 -0
@@ -171,6 +171,7 @@ size_t rpc_pton(const char *, const size_t,char *rpc_sockaddr2uaddr (const struct sockaddr *);size_t rpc_uaddr2sockaddr (const char *,const size_t ,struct sockaddr *,const size_t ); +void call_bind (struct rpc_task *task);static inline unsigned short rpc_get_port (const struct sockaddr *sap) {
net/sunrpc/clnt.c +2 -2
@@ -63,7 +63,6 @@ static void call_reserve(struct rpc_task *task);static void call_reserveresult (struct rpc_task *task);static void call_allocate (struct rpc_task *task);static void call_decode (struct rpc_task *task); -static void call_bind (struct rpc_task *task);static void call_bind_status (struct rpc_task *task);static void call_transmit (struct rpc_task *task);#if defined(CONFIG_NFS_V4_1 )
@@ -1004,7 +1003,7 @@ EXPORT_SYMBOL_GPL(rpc_xdr_encode);/* * 4. Get the server port number if not yet set */ -static void +void call_bind (struct rpc_task *task) {struct rpc_xprt *xprt = task->tk_xprt;
@@ -1018,6 +1017,7 @@ call_bind(struct rpc_task *task) xprt->ops->rpcbind (task); } } +EXPORT_SYMBOL_GPL (call_bind);/* * 4a. Sort out bind result
Signature · 1 line
1.6.0.6
RFC PATCH 6/9 sunrpc: add new SMB transport class for sunrpc
In reply to Jeff LaytonAdd a new transport class for SMB traffic that lives in xprtsmb.c and a corresponding header file for making use of it.
With this, it's possible to use the sunrpc layer to send and receive SMB traffic.
--- include/linux/sunrpc/xprtsmb.h | 59 ++ net/sunrpc/xprtsmb.c | 1720 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 1779 insertions(+), 0 deletions(-) create mode 100644 include/linux/sunrpc/xprtsmb.h create mode 100644 net/sunrpc/xprtsmb.c
include/linux/sunrpc/xprtsmb.h new file +59 -0
@@ -0,0 +1,59 @@ +/* +* net/sunrpc/xprtsmb.h -- SMB transport for sunrpc +* +* Copyright (c) 2009 Red Hat, Inc. +* Author(s): Jeff Layton ([email protected]) +* +* This library is free software; you can redistribute it and/or modify +* it under the terms of the GNU Lesser General Public License as published +* by the Free Software Foundation; either version 2.1 of the License, or +* (at your option) any later version. +* +* This library is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See +* the GNU Lesser General Public License for more details. +* +* You should have received a copy of the GNU Lesser General Public License +* along with this library; if not, write to the Free Software +* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA +*/ + +/* +* xprtsmb.h +* +* This file contains the public interfaces for the SMB transport for +* the sunrpc layer. +*/ + +#ifndef _LINUX_SUNRPC_XPRTSMB_H +#define _LINUX_SUNRPC_XPRTSMB_H + +/* +* RPC transport identifier for SMB +*/ +#define XPRT_TRANSPORT_SMB 1024 + +int init_smb_xprt (void ); +void cleanup_smb_xprt (void ); + +/* standard SMB header */ +struct smb_header { +__u8 protocol[4 ]; +__u8 command; +__le32 status; +__u8 flags; +__le16 flags2; +__le16 pidhigh; +__u8 signature[8 ]; +__u8 pad[2 ]; +__le16 tid; +__le16 pid; +__le16 uid; +__le16 mid; +} __attribute__((packed)); + +/* SMB Header Flags of interest */ +#define SMBFLG_RESPONSE 0x80/* response from server */ + +#endif /* _LINUX_SUNRPC_XPRTSMB_H */
net/sunrpc/xprtsmb.c new file +1720 -0
1,720 lines not on this page · raw message
Signature · 1 line
1.6.0.6
RFC PATCH 7/9 sunrpc: add encoding and decoding routines for SMB
In reply to Jeff LaytonAdd a new smb.c file to hold encode and decode routines for SMB packets.
--- include/linux/sunrpc/smb.h | 42 ++++++++++++++ include/linux/sunrpc/xprtsmb.h | 6 +- net/sunrpc/smb.c | 120 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 165 insertions(+), 3 deletions(-) create mode 100644 include/linux/sunrpc/smb.h create mode 100644 net/sunrpc/smb.c
include/linux/sunrpc/smb.h new file +42 -0
@@ -0,0 +1,42 @@ +/* +* net/sunrpc/smb.h -- SMB transport for sunrpc +* +* Copyright (c) 2009 Red Hat, Inc. +* Author(s): Jeff Layton ([email protected]) +* +* This library is free software; you can redistribute it and/or modify +* it under the terms of the GNU Lesser General Public License as published +* by the Free Software Foundation; either version 2.1 of the License, or +* (at your option) any later version. +* +* This library is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See +* the GNU Lesser General Public License for more details. +* +* You should have received a copy of the GNU Lesser General Public License +* along with this library; if not, write to the Free Software +* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA +*/ + +/* +* xprtsmb.h +* +* This file contains the public interfaces for the SMB transport for +* the sunrpc layer. +*/ + +#ifndef _LINUX_SUNRPC_SMB_H +#define _LINUX_SUNRPC_SMB_H + +/* +* This is the generic SMB function. rqstp is either a rpc_rqst (client +* side) or svc_rqst pointer (server side). +* Encode functions always assume there's enough room in the buffer. +*/ +typedef int (*ksmbproc_t )(void *rqstp,__le32 *data,void *obj); + +void smb_encode (struct rpc_task *task); +int smb_decode (struct rpc_task *task); + +#endif /* _LINUX_SUNRPC_SMB_H */
include/linux/sunrpc/xprtsmb.h +3 -3
@@ -34,9 +34,6 @@*/ #define XPRT_TRANSPORT_SMB 1024 -int init_smb_xprt (void ); -void cleanup_smb_xprt (void ); -/* standard SMB header */ struct smb_header {__u8 protocol[4 ];
@@ -56,4 +53,7 @@ struct smb_header {/* SMB Header Flags of interest */ #define SMBFLG_RESPONSE 0x80/* response from server */ +int init_smb_xprt (void ); +void cleanup_smb_xprt (void ); +#endif /* _LINUX_SUNRPC_XPRTSMB_H */
net/sunrpc/smb.c new file +120 -0
120 lines not on this page · raw message
Signature · 1 line
1.6.0.6
RFC PATCH 8/9 sunrpc: add Kconfig option for CONFIG_SUNRPC_SMB
In reply to Jeff Layton...and the necessary Makefile bits so to compile support into the sunrpc module.
--- net/sunrpc/Kconfig | 11 +++++++++++ net/sunrpc/Makefile | 1 + net/sunrpc/sunrpc_syms.c | 3 +++ net/sunrpc/xprtsmb.c | 7 +++++-- 4 files changed, 20 insertions(+), 2 deletions(-)
net/sunrpc/Kconfig +11 -0
@@ -54,3 +54,14 @@ config RPCSEC_GSS_SPKM3 available from http://linux-nfs.org/. If unsure, say N. + +config SUNRPC_SMB + bool "SMB/CIFS Transport for the SUNRPC Layer (EXPERIMENTAL)" + depends on SUNRPC && EXPERIMENTAL + help + This option adds the ability for the kernels SUNRPC Layer to + send and receive Sever Message Block (SMB) traffic. This + protocol is widely used by Microsoft Windows servers for + filesharing. + + If unsure, say N.
net/sunrpc/Makefile +1 -0
@@ -16,3 +16,4 @@ sunrpc-y := clnt.o xprt.o socklib.o xprtsock.o sched.o \ sunrpc-$(CONFIG_NFS_V4_1) += backchannel_rqst.o bc_svc.o sunrpc-$(CONFIG_PROC_FS) += stats.o sunrpc-$(CONFIG_SYSCTL) += sysctl.o +sunrpc-$(CONFIG_SUNRPC_SMB) += xprtsmb.o smb.o
net/sunrpc/sunrpc_syms.c +3 -0
@@ -21,6 +21,7 @@#include <linux/workqueue.h> #include <linux/sunrpc/rpc_pipe_fs.h> #include <linux/sunrpc/xprtsock.h> +#include <linux/sunrpc/xprtsmb.h> extern struct cache_detail ip_map_cache, unix_gid_cache;
@@ -45,6 +46,7 @@ init_sunrpc(void)cache_register (&unix_gid_cache);svc_init_xprt_sock ();/* svc sock transport */ init_socket_xprt ();/* clnt sock transport */ +init_smb_xprt ();rpcauth_init_module (); out:return err;
@@ -54,6 +56,7 @@ static void __exitcleanup_sunrpc (void) {rpcauth_remove_module (); +cleanup_smb_xprt ();cleanup_socket_xprt ();svc_cleanup_xprt_sock ();unregister_rpc_pipefs ();
net/sunrpc/xprtsmb.c +5 -2
@@ -1691,7 +1691,8 @@ static struct xprt_class xs_smb_transport = {* init_socket_xprt - set up xprtsock's sysctls, register with RPC client * */ -int init_smb_xprt (void ) +int +init_smb_xprt (void ) {#ifdef RPC_DEBUG if (!sunrpc_table_header)
@@ -1707,7 +1708,8 @@ int init_smb_xprt(void)* cleanup_socket_xprt - remove xprtsock's sysctls, unregister * */ -void cleanup_smb_xprt (void ) +void +cleanup_smb_xprt (void ) {#ifdef RPC_DEBUG if (sunrpc_table_header) {
@@ -1718,3 +1720,4 @@ void cleanup_smb_xprt(void)xprt_unregister_transport (&xs_smb_transport); } +
Signature · 1 line
1.6.0.6
RFC PATCH 9/9 smbtest: simple module for testing SMB/RPC code
In reply to Jeff LaytonProof of concept module. This just sends a NEGOTIATE_PROTOCOL request to a server and verifies the response.
--- fs/Makefile | 2 + fs/smbtest/Makefile | 1 + fs/smbtest/smbtest.c | 204 ++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 207 insertions(+), 0 deletions(-) create mode 100644 fs/smbtest/Makefile create mode 100644 fs/smbtest/smbtest.c
fs/Makefile +2 -0
@@ -124,3 +124,5 @@ obj-$(CONFIG_OCFS2_FS) += ocfs2/ obj-$(CONFIG_BTRFS_FS) += btrfs/ obj-$(CONFIG_GFS2_FS) += gfs2/ obj-$(CONFIG_EXOFS_FS) += exofs/ + +obj-m += smbtest/
fs/smbtest/Makefile new file +1 -0
@@ -0,0 +1 @@ +obj-m += smbtest.o
fs/smbtest/smbtest.c new file +204 -0
204 lines not on this page · raw message
Signature · 1 line
1.6.0.6
Quoted 125 lines · Jeff Layton
This patchset is still preliminary and is just an RFC...
First, some background. When I was at Connectathon this year, Trond mentioned an interesting idea to me. He said (paraphrasing):
"Why doesn't CIFS just use the RPC layer for transport? It's very efficient at shoveling bits out onto the wire. You'd just need to abstract out the XDR/RPC specific bits."
The idea has been floating around in the back of my head until recently, when I decided to give it a go. This patchset represents a first, rough draft at a patchset to do this. There's also a proof of concept module that demonstrates that it works as expected.
The patchset is still very rough. Reconnect behavior is still very "RPC-ish", for instance. There are doubtless other changes that will need to be made before I had anything merge-worthy.
At this point, I'm just interested in feedback on the general approach. Should I continue to pursue this or is it a non-starter for some reason?
The next step here, obviously is to start building a fs on top of it. I'd particularly be interested in using this as a foundation of a new smb2fs.
I've also got this patchset up on my public git tree:
http://git.samba.org/?p=jlayton/cifs.git;a=summary
Here are some questions I anticipate on this, and their answers:
------------------------------------------------------------------------
Q: Are you insane? Why would you attempt to do this?A: Maybe...but in the last couple of years, I've spent a substantial amount of time working on the CIFS code. Much of that time has been spent fixing bugs. Many of those bugs exist in the low-level transport code which has been hacked on, kludged around and hand tweaked to where it is today. Unfortunately that's made it a very hard to work on mess. This drives away potential developers.
CIFS in particular is also designed around synchronous ops, which seriously limits throughput. Retrofitting it for asynchronous operation will be adding even more kludges. The sunrpc code is already fundamentally asynchronous.
------------------------------------------------------------------------
Q: Okay, what's the benefit of hooking it up to sunrpc rather than building a new transport layer (or fixing the transport in the other two filesystems)?A: Using sunrpc allows us to share a lot of the rpc scheduler code with sunrpc. At a high level, NFS/RPC and SMB aren't really very different. Sure, they have different formats, and one is big endian on the wire and the other isn't...still there are significant similarities.
We also get access to the great upcall mechanisms that sunrpc has, and the possibility to share code like the gssapi upcalls. The sunrpc layer has a credential and authentication management framework that we can build on to make a truly multiuser CIFS/SMB filesystem.
I've heard it claimed before that Linux's sunrpc layer is over-engineered, but in this case that works in our favor...
------------------------------------------------------------------------
Q: can we hook up cifs or smbfs to use this as a transport?A: Not trivially. CIFS in particular is not designed with each call having discrete encode and decode functions. They're sort of mashed together. smbfs might be possible...I'm a little less familiar with it, but it does have a transport layer that more closely resembles the sunrpc one. Still though, it'd take significant work to make that happen. I'm not opposed to the idea however.
In the end though, I think we'll probably need to design something new to sit on top of this. We will probably be able to borrow code and concepts from the other filesystems however.
------------------------------------------------------------------------
Q: could we use this as a transport layer for a smb2fs ?A: Yes, I think so. This particular prototype is build around SMB1, but SMB2 could be supported with only minor modifications. One of the reasons for sending this patchset now before I've built a filesystem on top of it is because I know that SMB2 work is in progress. I'd like to see it based around a more asynchronous transport model, or at least built with cleaner layering so that we can eventually bolt on a different transport layer if we so choose.
Jeff Layton (9):
sunrpc: abstract out encoding function atrpc_clntlevel
sunrpc: move check for too small reply size intorpc_verify_header
sunrpc: abstract our call decoding routine
sunrpc: moverpc_xdr_buf_inittoclnt.h
sunrpc: makecall_bindnon-static
sunrpc: add new SMB transport class for sunrpc
sunrpc: add encoding and decoding routines for SMB
sunrpc: addKconfigoption forCONFIG_SUNRPC_SMB
smbtest: simple module for testing SMB/RPC codefs/Makefile | 2 + fs/lockd/host.c | 4 + fs/lockd/mon.c | 4 + fs/nfs/client.c | 4 + fs/nfs/mount_clnt.c | 4 + fs/nfsd/nfs4callback.c | 4 + fs/smbtest/Makefile | 1 + fs/smbtest/smbtest.c | 204 +++++ include/linux/sunrpc/clnt.h | 24 +- include/linux/sunrpc/smb.h | 42 + include/linux/sunrpc/xprtsmb.h | 59 ++ net/sunrpc/Kconfig | 11 + net/sunrpc/Makefile | 1 + net/sunrpc/clnt.c | 98 ++- net/sunrpc/rpcb_clnt.c | 8 + net/sunrpc/smb.c | 120 +++ net/sunrpc/sunrpc_syms.c | 3 + net/sunrpc/xprtsmb.c | 1723 ++++++++++++++++++++++++++++++++++++++++ 18 files changed, 2272 insertions(+), 44 deletions(-) create mode 100644 fs/smbtest/Makefile create mode 100644 fs/smbtest/smbtest.c create mode 100644 include/linux/sunrpc/smb.h create mode 100644 include/linux/sunrpc/xprtsmb.h create mode 100644 net/sunrpc/smb.c create mode 100644 net/sunrpc/xprtsmb.c
Quoted 5 lines
_______________________________________________
linux-cifs-client mailing list
[email protected]
https://lists.samba.org/mailman/listinfo/linux-cifs-client
Jeff,
So servers need to speak SUNRPC as well right? Are_threre/how_many servers are out there that speak CIFS/SMB over SUNRPC or SMB2 over SUNRPC?
_______________________________________________
linux-cifs-client mailing list
[email protected]
https://lists.samba.org/mailman/listinfo/linux-cifs-client
A message that is not in the archive (sent off-list or not yet received): 4a4634330909271921o125c8303p64974d850007ed8b-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org
Quoted 134 lines · On Sun, 27 Sep 2009 21:21:40 -0500 Shirish Pargaonkar…
Quoted 125 lines · Jeff Layton
This patchset is still preliminary and is just an RFC...
First, some background. When I was at Connectathon this year, Trond mentioned an interesting idea to me. He said (paraphrasing):
"Why doesn't CIFS just use the RPC layer for transport? It's very efficient at shoveling bits out onto the wire. You'd just need to abstract out the XDR/RPC specific bits."
The idea has been floating around in the back of my head until recently, when I decided to give it a go. This patchset represents a first, rough draft at a patchset to do this. There's also a proof of concept module that demonstrates that it works as expected.
The patchset is still very rough. Reconnect behavior is still very "RPC-ish", for instance. There are doubtless other changes that will need to be made before I had anything merge-worthy.
At this point, I'm just interested in feedback on the general approach. Should I continue to pursue this or is it a non-starter for some reason?
The next step here, obviously is to start building a fs on top of it. I'd particularly be interested in using this as a foundation of a new smb2fs.
I've also got this patchset up on my public git tree:
http://git.samba.org/?p=jlayton/cifs.git;a=summary
Here are some questions I anticipate on this, and their answers:
------------------------------------------------------------------------
Q: Are you insane? Why would you attempt to do this?A: Maybe...but in the last couple of years, I've spent a substantial amount of time working on the CIFS code. Much of that time has been spent fixing bugs. Many of those bugs exist in the low-level transport code which has been hacked on, kludged around and hand tweaked to where it is today. Unfortunately that's made it a very hard to work on mess. This drives away potential developers.
CIFS in particular is also designed around synchronous ops, which seriously limits throughput. Retrofitting it for asynchronous operation will be adding even more kludges. The sunrpc code is already fundamentally asynchronous.
------------------------------------------------------------------------
Q: Okay, what's the benefit of hooking it up to sunrpc rather than building a new transport layer (or fixing the transport in the other two filesystems)?A: Using sunrpc allows us to share a lot of the rpc scheduler code with sunrpc. At a high level, NFS/RPC and SMB aren't really very different. Sure, they have different formats, and one is big endian on the wire and the other isn't...still there are significant similarities.
We also get access to the great upcall mechanisms that sunrpc has, and the possibility to share code like the gssapi upcalls. The sunrpc layer has a credential and authentication management framework that we can build on to make a truly multiuser CIFS/SMB filesystem.
I've heard it claimed before that Linux's sunrpc layer is over-engineered, but in this case that works in our favor...
------------------------------------------------------------------------
Q: can we hook up cifs or smbfs to use this as a transport?A: Not trivially. CIFS in particular is not designed with each call having discrete encode and decode functions. They're sort of mashed together. smbfs might be possible...I'm a little less familiar with it, but it does have a transport layer that more closely resembles the sunrpc one. Still though, it'd take significant work to make that happen. I'm not opposed to the idea however.
In the end though, I think we'll probably need to design something new to sit on top of this. We will probably be able to borrow code and concepts from the other filesystems however.
------------------------------------------------------------------------
Q: could we use this as a transport layer for a smb2fs ?A: Yes, I think so. This particular prototype is build around SMB1, but SMB2 could be supported with only minor modifications. One of the reasons for sending this patchset now before I've built a filesystem on top of it is because I know that SMB2 work is in progress. I'd like to see it based around a more asynchronous transport model, or at least built with cleaner layering so that we can eventually bolt on a different transport layer if we so choose.
Jeff Layton (9):
sunrpc: abstract out encoding function atrpc_clntlevel
sunrpc: move check for too small reply size intorpc_verify_header
sunrpc: abstract our call decoding routine
sunrpc: moverpc_xdr_buf_inittoclnt.h
sunrpc: makecall_bindnon-static
sunrpc: add new SMB transport class for sunrpc
sunrpc: add encoding and decoding routines for SMB
sunrpc: addKconfigoption forCONFIG_SUNRPC_SMB
smbtest: simple module for testing SMB/RPC codefs/Makefile | 2 + fs/lockd/host.c | 4 + fs/lockd/mon.c | 4 + fs/nfs/client.c | 4 + fs/nfs/mount_clnt.c | 4 + fs/nfsd/nfs4callback.c | 4 + fs/smbtest/Makefile | 1 + fs/smbtest/smbtest.c | 204 +++++ include/linux/sunrpc/clnt.h | 24 +- include/linux/sunrpc/smb.h | 42 + include/linux/sunrpc/xprtsmb.h | 59 ++ net/sunrpc/Kconfig | 11 + net/sunrpc/Makefile | 1 + net/sunrpc/clnt.c | 98 ++- net/sunrpc/rpcb_clnt.c | 8 + net/sunrpc/smb.c | 120 +++ net/sunrpc/sunrpc_syms.c | 3 + net/sunrpc/xprtsmb.c | 1723 ++++++++++++++++++++++++++++++++++++++++ 18 files changed, 2272 insertions(+), 44 deletions(-) create mode 100644 fs/smbtest/Makefile create mode 100644 fs/smbtest/smbtest.c create mode 100644 include/linux/sunrpc/smb.h create mode 100644 include/linux/sunrpc/xprtsmb.h create mode 100644 net/sunrpc/smb.c create mode 100644 net/sunrpc/xprtsmb.cQuoted 5 lines
_______________________________________________
linux-cifs-client mailing list
[email protected]
https://lists.samba.org/mailman/listinfo/linux-cifs-client
Quoted 5 lines
Jeff,
So servers need to speak SUNRPC as well right? Are_threre/how_many servers are out there that speak CIFS/SMB over SUNRPC or SMB2 over SUNRPC?
No...sorry if that wasn't clear. The on the wire protocol is still the same as cifs or smbfs talks. You could use this to talk to windows servers for instance. The idea is to use the existing code in the Linux kernel's sunrpc engine and teach it how to speak a new "dialect" on the wire...the SMB or SMB2 protocol.
For example, to test this, I sent a NegotiateProtocol request to a samba server over its normal port 445. No ONC-RPC required.
Signature · 1 line
Jeff Layton <[email protected]>