[PATCH] v9fs: transport modules
authorEric Van Hensbergen <ericvh@gmail.com>
Fri, 9 Sep 2005 20:04:22 +0000 (13:04 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Fri, 9 Sep 2005 20:57:57 +0000 (13:57 -0700)
This part of the patch contains transport routines.

Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Signed-off-by: Latchesar Ionkov <lucho@ionkov.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
fs/9p/mux.c [new file with mode: 0644]
fs/9p/mux.h [new file with mode: 0644]
fs/9p/trans_fd.c [new file with mode: 0644]
fs/9p/trans_sock.c [new file with mode: 0644]
fs/9p/transport.h [new file with mode: 0644]
fs/9p/v9fs.c

diff --git a/fs/9p/mux.c b/fs/9p/mux.c
new file mode 100644 (file)
index 0000000..8ebc1af
--- /dev/null
@@ -0,0 +1,440 @@
+/*
+ * linux/fs/9p/mux.c
+ *
+ * Protocol Multiplexer
+ *
+ *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
+ *  Copyright (C) 2004 by Latchesar Ionkov <lucho@ionkov.net>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/kthread.h>
+#include <linux/idr.h>
+
+#include "debug.h"
+#include "v9fs.h"
+#include "9p.h"
+#include "transport.h"
+#include "conv.h"
+#include "mux.h"
+
+/**
+ * dprintcond - print condition of session info
+ * @v9ses: session info structure
+ * @req: RPC request structure
+ *
+ */
+
+static inline int
+dprintcond(struct v9fs_session_info *v9ses, struct v9fs_rpcreq *req)
+{
+       dprintk(DEBUG_MUX, "condition: %d, %p\n", v9ses->transport->status,
+               req->rcall);
+       return 0;
+}
+
+/**
+ * xread - force read of a certain number of bytes
+ * @v9ses: session info structure
+ * @ptr: pointer to buffer
+ * @sz: number of bytes to read
+ *
+ * Chuck Cranor CS-533 project1
+ */
+
+static int xread(struct v9fs_session_info *v9ses, void *ptr, unsigned long sz)
+{
+       int rd = 0;
+       int ret = 0;
+       while (rd < sz) {
+               ret = v9ses->transport->read(v9ses->transport, ptr, sz - rd);
+               if (ret <= 0) {
+                       dprintk(DEBUG_ERROR, "xread errno %d\n", ret);
+                       return ret;
+               }
+               rd += ret;
+               ptr += ret;
+       }
+       return (rd);
+}
+
+/**
+ * read_message - read a full 9P2000 fcall packet
+ * @v9ses: session info structure
+ * @rcall: fcall structure to read into
+ * @rcalllen: size of fcall buffer
+ *
+ */
+
+static int
+read_message(struct v9fs_session_info *v9ses,
+            struct v9fs_fcall *rcall, int rcalllen)
+{
+       unsigned char buf[4];
+       void *data;
+       int size = 0;
+       int res = 0;
+
+       res = xread(v9ses, buf, sizeof(buf));
+       if (res < 0) {
+               dprintk(DEBUG_ERROR,
+                       "Reading of count field failed returned: %d\n", res);
+               return res;
+       }
+
+       if (res < 4) {
+               dprintk(DEBUG_ERROR,
+                       "Reading of count field failed returned: %d\n", res);
+               return -EIO;
+       }
+
+       size = buf[0] | (buf[1] << 8) | (buf[2] << 16) | (buf[3] << 24);
+       dprintk(DEBUG_MUX, "got a packet count: %d\n", size);
+
+       /* adjust for the four bytes of size */
+       size -= 4;
+
+       if (size > v9ses->maxdata) {
+               dprintk(DEBUG_ERROR, "packet too big: %d\n", size);
+               return -E2BIG;
+       }
+
+       data = kmalloc(size, GFP_KERNEL);
+       if (!data) {
+               eprintk(KERN_WARNING, "out of memory\n");
+               return -ENOMEM;
+       }
+
+       res = xread(v9ses, data, size);
+       if (res < size) {
+               dprintk(DEBUG_ERROR, "Reading of fcall failed returned: %d\n",
+                       res);
+               kfree(data);
+               return res;
+       }
+
+       /* we now have an in-memory string that is the reply.
+        * deserialize it. There is very little to go wrong at this point
+        * save for v9fs_alloc errors.
+        */
+       res = v9fs_deserialize_fcall(v9ses, size, data, v9ses->maxdata,
+                                    rcall, rcalllen);
+
+       kfree(data);
+
+       if (res < 0)
+               return res;
+
+       return 0;
+}
+
+/**
+ * v9fs_recv - receive an RPC response for a particular tag
+ * @v9ses: session info structure
+ * @req: RPC request structure
+ *
+ */
+
+static int v9fs_recv(struct v9fs_session_info *v9ses, struct v9fs_rpcreq *req)
+{
+       int ret = 0;
+
+       dprintk(DEBUG_MUX, "waiting for response: %d\n", req->tcall->tag);
+       ret = wait_event_interruptible(v9ses->read_wait,
+                      ((v9ses->transport->status != Connected) ||
+                       (req->rcall != 0) || dprintcond(v9ses, req)));
+
+       dprintk(DEBUG_MUX, "got it: rcall %p\n", req->rcall);
+       if (v9ses->transport->status == Disconnected)
+               return -ECONNRESET;
+
+       if (ret == 0) {
+               spin_lock(&v9ses->muxlock);
+               list_del(&req->next);
+               spin_unlock(&v9ses->muxlock);
+       }
+
+       return ret;
+}
+
+/**
+ * v9fs_send - send a 9P request
+ * @v9ses: session info structure
+ * @req: RPC request to send
+ *
+ */
+
+static int v9fs_send(struct v9fs_session_info *v9ses, struct v9fs_rpcreq *req)
+{
+       int ret = -1;
+       void *data = NULL;
+       struct v9fs_fcall *tcall = req->tcall;
+
+       data = kmalloc(v9ses->maxdata + V9FS_IOHDRSZ, GFP_KERNEL);
+       if (!data)
+               return -ENOMEM;
+
+       tcall->size = 0;        /* enforce size recalculation */
+       ret =
+           v9fs_serialize_fcall(v9ses, tcall, data,
+                                v9ses->maxdata + V9FS_IOHDRSZ);
+       if (ret < 0)
+               goto free_data;
+
+       spin_lock(&v9ses->muxlock);
+       list_add(&req->next, &v9ses->mux_fcalls);
+       spin_unlock(&v9ses->muxlock);
+
+       dprintk(DEBUG_MUX, "sending message: tag %d size %d\n", tcall->tag,
+               tcall->size);
+       ret = v9ses->transport->write(v9ses->transport, data, tcall->size);
+
+       if (ret != tcall->size) {
+               spin_lock(&v9ses->muxlock);
+               list_del(&req->next);
+               kfree(req->rcall);
+
+               spin_unlock(&v9ses->muxlock);
+               if (ret >= 0)
+                       ret = -EREMOTEIO;
+       } else
+               ret = 0;
+
+      free_data:
+       kfree(data);
+       return ret;
+}
+
+/**
+ * v9fs_mux_rpc - send a request, receive a response
+ * @v9ses: session info structure
+ * @tcall: fcall to send
+ * @rcall: buffer to place response into
+ *
+ */
+
+long
+v9fs_mux_rpc(struct v9fs_session_info *v9ses, struct v9fs_fcall *tcall,
+            struct v9fs_fcall **rcall)
+{
+       int tid = -1;
+       struct v9fs_fcall *fcall = NULL;
+       struct v9fs_rpcreq req;
+       int ret = -1;
+
+       if (!v9ses)
+               return -EINVAL;
+
+       if (rcall)
+               *rcall = NULL;
+
+       if (tcall->id != TVERSION) {
+               tid = v9fs_get_idpool(&v9ses->tidpool);
+               if (tid < 0)
+                       return -ENOMEM;
+       }
+
+       tcall->tag = tid;
+
+       req.tcall = tcall;
+       req.rcall = NULL;
+
+       ret = v9fs_send(v9ses, &req);
+
+       if (ret < 0) {
+               if (tcall->id != TVERSION)
+                       v9fs_put_idpool(tid, &v9ses->tidpool);
+               dprintk(DEBUG_MUX, "error %d\n", ret);
+               return ret;
+       }
+
+       ret = v9fs_recv(v9ses, &req);
+
+       fcall = req.rcall;
+
+       dprintk(DEBUG_MUX, "received: tag=%x, ret=%d\n", tcall->tag, ret);
+       if (ret == -ERESTARTSYS) {
+               if (v9ses->transport->status != Disconnected
+                   && tcall->id != TFLUSH) {
+                       unsigned long flags;
+
+                       dprintk(DEBUG_MUX, "flushing the tag: %d\n",
+                               tcall->tag);
+                       clear_thread_flag(TIF_SIGPENDING);
+                       v9fs_t_flush(v9ses, tcall->tag);
+                       spin_lock_irqsave(&current->sighand->siglock, flags);
+                       recalc_sigpending();
+                       spin_unlock_irqrestore(&current->sighand->siglock,
+                                              flags);
+                       dprintk(DEBUG_MUX, "flushing done\n");
+               }
+
+               goto release_req;
+       } else if (ret < 0)
+               goto release_req;
+
+       if (!fcall)
+               ret = -EIO;
+       else {
+               if (fcall->id == RERROR) {
+                       ret = v9fs_errstr2errno(fcall->params.rerror.error);
+                       if (ret == 0) { /* string match failed */
+                               if (fcall->params.rerror.errno)
+                                       ret = -(fcall->params.rerror.errno);
+                               else
+                                       ret = -ESERVERFAULT;
+                       }
+               } else if (fcall->id != tcall->id + 1) {
+                       dprintk(DEBUG_ERROR,
+                               "fcall mismatch: expected %d, got %d\n",
+                               tcall->id + 1, fcall->id);
+                       ret = -EIO;
+               }
+       }
+
+      release_req:
+       if (tcall->id != TVERSION)
+               v9fs_put_idpool(tid, &v9ses->tidpool);
+       if (rcall)
+               *rcall = fcall;
+       else
+               kfree(fcall);
+
+       return ret;
+}
+
+/**
+ * v9fs_recvproc - kproc to handle demultiplexing responses
+ * @data: session info structure
+ *
+ */
+
+static int v9fs_recvproc(void *data)
+{
+       struct v9fs_session_info *v9ses = (struct v9fs_session_info *)data;
+       struct v9fs_fcall *rcall = NULL;
+       struct v9fs_rpcreq *rptr;
+       struct v9fs_rpcreq *req;
+       struct v9fs_rpcreq *rreq;
+       int err = 0;
+
+       allow_signal(SIGKILL);
+       set_current_state(TASK_INTERRUPTIBLE);
+       complete(&v9ses->proccmpl);
+       while (!kthread_should_stop() && err >= 0) {
+               req = rptr = rreq = NULL;
+
+               rcall = kmalloc(v9ses->maxdata + V9FS_IOHDRSZ, GFP_KERNEL);
+               if (!rcall) {
+                       eprintk(KERN_ERR, "no memory for buffers\n");
+                       break;
+               }
+
+               err = read_message(v9ses, rcall, v9ses->maxdata + V9FS_IOHDRSZ);
+               if (err < 0) {
+                       kfree(rcall);
+                       break;
+               }
+               spin_lock(&v9ses->muxlock);
+               list_for_each_entry_safe(rreq, rptr, &v9ses->mux_fcalls, next) {
+                       if (rreq->tcall->tag == rcall->tag) {
+                               req = rreq;
+                               req->rcall = rcall;
+                               break;
+                       }
+               }
+
+               if (req && (req->tcall->id == TFLUSH)) {
+                       struct v9fs_rpcreq *treq = NULL;
+                       list_for_each_entry_safe(treq, rptr, &v9ses->mux_fcalls, next) {
+                               if (treq->tcall->tag ==
+                                   req->tcall->params.tflush.oldtag) {
+                                       list_del(&rptr->next);
+                                       kfree(treq->rcall);
+                                       break;
+                               }
+                       }
+               }
+
+               spin_unlock(&v9ses->muxlock);
+
+               if (!req) {
+                       dprintk(DEBUG_ERROR,
+                               "unexpected response: id %d tag %d\n",
+                               rcall->id, rcall->tag);
+
+                       kfree(rcall);
+               }
+
+               wake_up_all(&v9ses->read_wait);
+               set_current_state(TASK_INTERRUPTIBLE);
+       }
+
+       /* Inform all pending processes about the failure */
+       wake_up_all(&v9ses->read_wait);
+
+       if (signal_pending(current))
+               complete(&v9ses->proccmpl);
+
+       dprintk(DEBUG_MUX, "recvproc: end\n");
+       v9ses->recvproc = NULL;
+
+       return err >= 0;
+}
+
+/**
+ * v9fs_mux_init - initialize multiplexer (spawn kproc)
+ * @v9ses: session info structure
+ * @dev_name: mount device information (to create unique kproc)
+ *
+ */
+
+int v9fs_mux_init(struct v9fs_session_info *v9ses, const char *dev_name)
+{
+       char procname[60];
+
+       strncpy(procname, dev_name, sizeof(procname));
+       procname[sizeof(procname) - 1] = 0;
+
+       init_waitqueue_head(&v9ses->read_wait);
+       init_completion(&v9ses->fcread);
+       init_completion(&v9ses->proccmpl);
+       spin_lock_init(&v9ses->muxlock);
+       INIT_LIST_HEAD(&v9ses->mux_fcalls);
+       v9ses->recvproc = NULL;
+       v9ses->curfcall = NULL;
+
+       v9ses->recvproc = kthread_create(v9fs_recvproc, v9ses,
+                                        "v9fs_recvproc %s", procname);
+
+       if (IS_ERR(v9ses->recvproc)) {
+               eprintk(KERN_ERR, "cannot create receiving thread\n");
+               v9fs_session_close(v9ses);
+               return -ECONNABORTED;
+       }
+
+       wake_up_process(v9ses->recvproc);
+       wait_for_completion(&v9ses->proccmpl);
+
+       return 0;
+}
diff --git a/fs/9p/mux.h b/fs/9p/mux.h
new file mode 100644 (file)
index 0000000..d7d8fa1
--- /dev/null
@@ -0,0 +1,39 @@
+/*
+ * linux/fs/9p/mux.h
+ *
+ * Multiplexer Definitions
+ *
+ *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+
+/* structure to manage each RPC transaction */
+
+struct v9fs_rpcreq {
+       struct v9fs_fcall *tcall;
+       struct v9fs_fcall *rcall;
+
+       /* XXX - could we put scatter/gather buffers here? */
+
+       struct list_head next;
+};
+
+int v9fs_mux_init(struct v9fs_session_info *v9ses, const char *dev_name);
+long v9fs_mux_rpc(struct v9fs_session_info *v9ses,
+                 struct v9fs_fcall *tcall, struct v9fs_fcall **rcall);
diff --git a/fs/9p/trans_fd.c b/fs/9p/trans_fd.c
new file mode 100644 (file)
index 0000000..63b58ce
--- /dev/null
@@ -0,0 +1,172 @@
+/*
+ * linux/fs/9p/trans_fd.c
+ *
+ * File Descriptor Transport Layer
+ *
+ *  Copyright (C) 2005 by Eric Van Hensbergen <ericvh@gmail.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/net.h>
+#include <linux/ipv6.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/un.h>
+#include <asm/uaccess.h>
+#include <linux/inet.h>
+#include <linux/idr.h>
+#include <linux/file.h>
+
+#include "debug.h"
+#include "v9fs.h"
+#include "transport.h"
+
+struct v9fs_trans_fd {
+       struct file *in_file;
+       struct file *out_file;
+};
+
+/**
+ * v9fs_fd_recv - receive from a socket
+ * @v9ses: session information
+ * @v: buffer to receive data into
+ * @len: size of receive buffer
+ *
+ */
+
+static int v9fs_fd_recv(struct v9fs_transport *trans, void *v, int len)
+{
+       struct v9fs_trans_fd *ts = trans ? trans->priv : NULL;
+
+       if (!trans || trans->status != Connected || !ts)
+               return -EIO;
+
+       return kernel_read(ts->in_file, ts->in_file->f_pos, v, len);
+}
+
+/**
+ * v9fs_fd_send - send to a socket
+ * @v9ses: session information
+ * @v: buffer to send data from
+ * @len: size of send buffer
+ *
+ */
+
+static int v9fs_fd_send(struct v9fs_transport *trans, void *v, int len)
+{
+       struct v9fs_trans_fd *ts = trans ? trans->priv : NULL;
+       mm_segment_t oldfs = get_fs();
+       int ret = 0;
+
+       if (!trans || trans->status != Connected || !ts)
+               return -EIO;
+
+       set_fs(get_ds());
+       /* The cast to a user pointer is valid due to the set_fs() */
+       ret = vfs_write(ts->out_file, (void __user *)v, len, &ts->out_file->f_pos);
+       set_fs(oldfs);
+
+       return ret;
+}
+
+/**
+ * v9fs_fd_init - initialize file descriptor transport
+ * @v9ses: session information
+ * @addr: address of server to mount
+ * @data: mount options
+ *
+ */
+
+static int
+v9fs_fd_init(struct v9fs_session_info *v9ses, const char *addr, char *data)
+{
+       struct v9fs_trans_fd *ts = NULL;
+       struct v9fs_transport *trans = v9ses->transport;
+
+       if((v9ses->wfdno == ~0) || (v9ses->rfdno == ~0)) {
+               printk(KERN_ERR "v9fs: Insufficient options for proto=fd\n");
+               return -ENOPROTOOPT;
+       }
+
+       sema_init(&trans->writelock, 1);
+       sema_init(&trans->readlock, 1);
+
+       ts = kmalloc(sizeof(struct v9fs_trans_fd), GFP_KERNEL);
+
+       if (!ts)
+               return -ENOMEM;
+
+       ts->in_file = fget( v9ses->rfdno );
+       ts->out_file = fget( v9ses->wfdno );
+
+       if (!ts->in_file || !ts->out_file) {
+               if (ts->in_file)
+                       fput(ts->in_file);
+
+               if (ts->out_file)
+                       fput(ts->out_file);
+
+               kfree(ts);
+               return -EIO;
+       }
+
+       trans->priv = ts;
+       trans->status = Connected;
+
+       return 0;
+}
+
+
+/**
+ * v9fs_fd_close - shutdown file descriptor
+ * @trans: private socket structure
+ *
+ */
+
+static void v9fs_fd_close(struct v9fs_transport *trans)
+{
+       struct v9fs_trans_fd *ts;
+
+       if (!trans)
+               return;
+
+       trans->status = Disconnected;
+       ts = trans->priv;
+
+       if (!ts)
+               return;
+
+       if (ts->in_file)
+               fput(ts->in_file);
+
+       if (ts->out_file)
+               fput(ts->out_file);
+
+       kfree(ts);
+}
+
+struct v9fs_transport v9fs_trans_fd = {
+       .init = v9fs_fd_init,
+       .write = v9fs_fd_send,
+       .read = v9fs_fd_recv,
+       .close = v9fs_fd_close,
+};
+
diff --git a/fs/9p/trans_sock.c b/fs/9p/trans_sock.c
new file mode 100644 (file)
index 0000000..081d1c8
--- /dev/null
@@ -0,0 +1,282 @@
+/*
+ * linux/fs/9p/trans_socket.c
+ *
+ * Socket Transport Layer
+ *
+ *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
+ *  Copyright (C) 1997-2002 by Ron Minnich <rminnich@sarnoff.com>
+ *  Copyright (C) 1995, 1996 by Olaf Kirch <okir@monad.swb.de>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/net.h>
+#include <linux/ipv6.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/un.h>
+#include <asm/uaccess.h>
+#include <linux/inet.h>
+#include <linux/idr.h>
+
+#include "debug.h"
+#include "v9fs.h"
+#include "transport.h"
+
+#define V9FS_PORT 564
+
+struct v9fs_trans_sock {
+       struct socket *s;
+};
+
+/**
+ * v9fs_sock_recv - receive from a socket
+ * @v9ses: session information
+ * @v: buffer to receive data into
+ * @len: size of receive buffer
+ *
+ */
+
+static int v9fs_sock_recv(struct v9fs_transport *trans, void *v, int len)
+{
+       struct msghdr msg;
+       struct kvec iov;
+       int result;
+       mm_segment_t oldfs;
+       struct v9fs_trans_sock *ts = trans ? trans->priv : NULL;
+
+       if (trans->status == Disconnected)
+               return -EREMOTEIO;
+
+       result = -EINVAL;
+
+       oldfs = get_fs();
+       set_fs(get_ds());
+
+       iov.iov_base = v;
+       iov.iov_len = len;
+       msg.msg_name = NULL;
+       msg.msg_namelen = 0;
+       msg.msg_iovlen = 1;
+       msg.msg_control = NULL;
+       msg.msg_controllen = 0;
+       msg.msg_namelen = 0;
+       msg.msg_flags = MSG_NOSIGNAL;
+
+       result = kernel_recvmsg(ts->s, &msg, &iov, 1, len, 0);
+
+       dprintk(DEBUG_TRANS, "socket state %d\n", ts->s->state);
+       set_fs(oldfs);
+
+       if (result <= 0) {
+               if (result != -ERESTARTSYS)
+                       trans->status = Disconnected;
+       }
+
+       return result;
+}
+
+/**
+ * v9fs_sock_send - send to a socket
+ * @v9ses: session information
+ * @v: buffer to send data from
+ * @len: size of send buffer
+ *
+ */
+
+static int v9fs_sock_send(struct v9fs_transport *trans, void *v, int len)
+{
+       struct kvec iov;
+       struct msghdr msg;
+       int result = -1;
+       mm_segment_t oldfs;
+       struct v9fs_trans_sock *ts = trans ? trans->priv : NULL;
+
+       dprintk(DEBUG_TRANS, "Sending packet size %d (%x)\n", len, len);
+       dump_data(v, len);
+
+       down(&trans->writelock);
+
+       oldfs = get_fs();
+       set_fs(get_ds());
+       iov.iov_base = v;
+       iov.iov_len = len;
+       msg.msg_name = NULL;
+       msg.msg_namelen = 0;
+       msg.msg_iovlen = 1;
+       msg.msg_control = NULL;
+       msg.msg_controllen = 0;
+       msg.msg_namelen = 0;
+       msg.msg_flags = MSG_NOSIGNAL;
+       result = kernel_sendmsg(ts->s, &msg, &iov, 1, len);
+       set_fs(oldfs);
+
+       if (result < 0) {
+               if (result != -ERESTARTSYS)
+                       trans->status = Disconnected;
+       }
+
+       up(&trans->writelock);
+       return result;
+}
+
+/**
+ * v9fs_tcp_init - initialize TCP socket
+ * @v9ses: session information
+ * @addr: address of server to mount
+ * @data: mount options
+ *
+ */
+
+static int
+v9fs_tcp_init(struct v9fs_session_info *v9ses, const char *addr, char *data)
+{
+       struct socket *csocket = NULL;
+       struct sockaddr_in sin_server;
+       int rc = 0;
+       struct v9fs_trans_sock *ts = NULL;
+       struct v9fs_transport *trans = v9ses->transport;
+
+       sema_init(&trans->writelock, 1);
+       sema_init(&trans->readlock, 1);
+
+       ts = kmalloc(sizeof(struct v9fs_trans_sock), GFP_KERNEL);
+
+       if (!ts)
+               return -ENOMEM;
+
+       trans->priv = ts;
+       ts->s = NULL;
+
+       if (!addr)
+               return -EINVAL;
+
+       dprintk(DEBUG_TRANS, "Connecting to %s\n", addr);
+
+       sin_server.sin_family = AF_INET;
+       sin_server.sin_addr.s_addr = in_aton(addr);
+       sin_server.sin_port = htons(v9ses->port);
+       sock_create_kern(PF_INET, SOCK_STREAM, IPPROTO_TCP, &csocket);
+       rc = csocket->ops->connect(csocket,
+                                  (struct sockaddr *)&sin_server,
+                                  sizeof(struct sockaddr_in), 0);
+       if (rc < 0) {
+               eprintk(KERN_ERR,
+                       "v9fs_trans_tcp: problem connecting socket to %s\n",
+                       addr);
+               return rc;
+       }
+       csocket->sk->sk_allocation = GFP_NOIO;
+       ts->s = csocket;
+       trans->status = Connected;
+
+       return 0;
+}
+
+/**
+ * v9fs_unix_init - initialize UNIX domain socket
+ * @v9ses: session information
+ * @dev_name: path to named pipe
+ * @data: mount options
+ *
+ */
+
+static int
+v9fs_unix_init(struct v9fs_session_info *v9ses, const char *dev_name,
+              char *data)
+{
+       int rc;
+       struct socket *csocket;
+       struct sockaddr_un sun_server;
+       struct v9fs_transport *trans;
+       struct v9fs_trans_sock *ts;
+
+       rc = 0;
+       csocket = NULL;
+       trans = v9ses->transport;
+
+       if (strlen(dev_name) > UNIX_PATH_MAX) {
+               eprintk(KERN_ERR, "v9fs_trans_unix: address too long: %s\n",
+                       dev_name);
+               return -ENOMEM;
+       }
+
+       ts = kmalloc(sizeof(struct v9fs_trans_sock), GFP_KERNEL);
+       if (!ts)
+               return -ENOMEM;
+
+       trans->priv = ts;
+       ts->s = NULL;
+
+       sema_init(&trans->writelock, 1);
+       sema_init(&trans->readlock, 1);
+
+       sun_server.sun_family = PF_UNIX;
+       strcpy(sun_server.sun_path, dev_name);
+       sock_create_kern(PF_UNIX, SOCK_STREAM, 0, &csocket);
+       rc = csocket->ops->connect(csocket, (struct sockaddr *)&sun_server,
+               sizeof(struct sockaddr_un) - 1, 0);     /* -1 *is* important */
+       if (rc < 0) {
+               eprintk(KERN_ERR,
+                       "v9fs_trans_unix: problem connecting socket: %s: %d\n",
+                       dev_name, rc);
+               return rc;
+       }
+       csocket->sk->sk_allocation = GFP_NOIO;
+       ts->s = csocket;
+       trans->status = Connected;
+
+       return 0;
+}
+
+/**
+ * v9fs_sock_close - shutdown socket
+ * @trans: private socket structure
+ *
+ */
+
+static void v9fs_sock_close(struct v9fs_transport *trans)
+{
+       struct v9fs_trans_sock *ts = trans ? trans->priv : NULL;
+
+       if ((ts) && (ts->s)) {
+               dprintk(DEBUG_TRANS, "closing the socket %p\n", ts->s);
+               sock_release(ts->s);
+               ts->s = NULL;
+               trans->status = Disconnected;
+               dprintk(DEBUG_TRANS, "socket closed\n");
+       }
+
+       kfree(ts);
+}
+
+struct v9fs_transport v9fs_trans_tcp = {
+       .init = v9fs_tcp_init,
+       .write = v9fs_sock_send,
+       .read = v9fs_sock_recv,
+       .close = v9fs_sock_close,
+};
+
+struct v9fs_transport v9fs_trans_unix = {
+       .init = v9fs_unix_init,
+       .write = v9fs_sock_send,
+       .read = v9fs_sock_recv,
+       .close = v9fs_sock_close,
+};
diff --git a/fs/9p/transport.h b/fs/9p/transport.h
new file mode 100644 (file)
index 0000000..9e9cd41
--- /dev/null
@@ -0,0 +1,46 @@
+/*
+ * linux/fs/9p/transport.h
+ *
+ * Transport Definition
+ *
+ *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+
+enum v9fs_transport_status {
+       Connected,
+       Disconnected,
+       Hung,
+};
+
+struct v9fs_transport {
+       enum v9fs_transport_status status;
+       struct semaphore writelock;
+       struct semaphore readlock;
+       void *priv;
+
+       int (*init) (struct v9fs_session_info *, const char *, char *);
+       int (*write) (struct v9fs_transport *, void *, int);
+       int (*read) (struct v9fs_transport *, void *, int);
+       void (*close) (struct v9fs_transport *);
+};
+
+extern struct v9fs_transport v9fs_trans_tcp;
+extern struct v9fs_transport v9fs_trans_unix;
+extern struct v9fs_transport v9fs_trans_fd;
index 14d663ebfcbcc6946666d299c2a78327998ce0b2..a573b751dd9a6a8088fce4aa73456649bb9080dc 100644 (file)
@@ -296,11 +296,6 @@ v9fs_session_init(struct v9fs_session_info *v9ses,
        case PROTO_FD:
                trans_proto = &v9fs_trans_fd;
                *v9ses->remotename = 0;
-               if((v9ses->wfdno == ~0) || (v9ses->rfdno == ~0)) {
-                       printk(KERN_ERR "v9fs: Insufficient options for proto=fd\n");
-                       retval = -ENOPROTOOPT;
-                       goto SessCleanUp;
-               }
                break;
        default:
                printk(KERN_ERR "v9fs: Bad mount protocol %d\n", v9ses->proto);