Subversion Repositories HelenOS

Compare Revisions

Ignore whitespace Rev 4608 → Rev 4687

/tags/0.4.1/uspace/lib/libfs/libfs.c
0,0 → 1,490
/*
* Copyright (c) 2009 Jakub Jermar
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* - Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* - Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* - The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
 
/** @addtogroup libfs
* @{
*/
/**
* @file
* Glue code which is commonod to all FS implementations.
*/
 
#include "libfs.h"
#include "../../srv/vfs/vfs.h"
#include <errno.h>
#include <async.h>
#include <ipc/ipc.h>
#include <as.h>
#include <assert.h>
#include <dirent.h>
#include <mem.h>
#include <sys/stat.h>
 
/** Register file system server.
*
* This function abstracts away the tedious registration protocol from
* file system implementations and lets them to reuse this registration glue
* code.
*
* @param vfs_phone Open phone for communication with VFS.
* @param reg File system registration structure. It will be
* initialized by this function.
* @param info VFS info structure supplied by the file system
* implementation.
* @param conn Connection fibril for handling all calls originating in
* VFS.
*
* @return EOK on success or a non-zero error code on errror.
*/
int fs_register(int vfs_phone, fs_reg_t *reg, vfs_info_t *info,
async_client_conn_t conn)
{
/*
* Tell VFS that we are here and want to get registered.
* We use the async framework because VFS will answer the request
* out-of-order, when it knows that the operation succeeded or failed.
*/
ipc_call_t answer;
aid_t req = async_send_0(vfs_phone, VFS_IN_REGISTER, &answer);
 
/*
* Send our VFS info structure to VFS.
*/
int rc = ipc_data_write_start(vfs_phone, info, sizeof(*info));
if (rc != EOK) {
async_wait_for(req, NULL);
return rc;
}
 
/*
* Ask VFS for callback connection.
*/
ipc_connect_to_me(vfs_phone, 0, 0, 0, &reg->vfs_phonehash);
 
/*
* Allocate piece of address space for PLB.
*/
reg->plb_ro = as_get_mappable_page(PLB_SIZE);
if (!reg->plb_ro) {
async_wait_for(req, NULL);
return ENOMEM;
}
 
/*
* Request sharing the Path Lookup Buffer with VFS.
*/
rc = ipc_share_in_start_0_0(vfs_phone, reg->plb_ro, PLB_SIZE);
if (rc) {
async_wait_for(req, NULL);
return rc;
}
/*
* Pick up the answer for the request to the VFS_IN_REQUEST call.
*/
async_wait_for(req, NULL);
reg->fs_handle = (int) IPC_GET_ARG1(answer);
/*
* Create a connection fibril to handle the callback connection.
*/
async_new_connection(reg->vfs_phonehash, 0, NULL, conn);
/*
* Tell the async framework that other connections are to be handled by
* the same connection fibril as well.
*/
async_set_client_connection(conn);
 
return IPC_GET_RETVAL(answer);
}
 
void fs_node_initialize(fs_node_t *fn)
{
memset(fn, 0, sizeof(fs_node_t));
}
 
void libfs_mount(libfs_ops_t *ops, fs_handle_t fs_handle, ipc_callid_t rid,
ipc_call_t *request)
{
dev_handle_t mp_dev_handle = (dev_handle_t) IPC_GET_ARG1(*request);
fs_index_t mp_fs_index = (fs_index_t) IPC_GET_ARG2(*request);
fs_handle_t mr_fs_handle = (fs_handle_t) IPC_GET_ARG3(*request);
dev_handle_t mr_dev_handle = (dev_handle_t) IPC_GET_ARG4(*request);
int res;
ipcarg_t rc;
 
ipc_call_t call;
ipc_callid_t callid;
 
/* accept the phone */
callid = async_get_call(&call);
int mountee_phone = (int)IPC_GET_ARG1(call);
if ((IPC_GET_METHOD(call) != IPC_M_CONNECTION_CLONE) ||
mountee_phone < 0) {
ipc_answer_0(callid, EINVAL);
ipc_answer_0(rid, EINVAL);
return;
}
ipc_answer_0(callid, EOK); /* acknowledge the mountee_phone */
res = ipc_data_write_receive(&callid, NULL);
if (!res) {
ipc_hangup(mountee_phone);
ipc_answer_0(callid, EINVAL);
ipc_answer_0(rid, EINVAL);
return;
}
 
fs_node_t *fn = ops->node_get(mp_dev_handle, mp_fs_index);
if (!fn) {
ipc_hangup(mountee_phone);
ipc_answer_0(callid, ENOENT);
ipc_answer_0(rid, ENOENT);
return;
}
 
if (fn->mp_data.mp_active) {
ipc_hangup(mountee_phone);
ops->node_put(fn);
ipc_answer_0(callid, EBUSY);
ipc_answer_0(rid, EBUSY);
return;
}
 
rc = async_req_0_0(mountee_phone, IPC_M_CONNECT_ME);
if (rc != 0) {
ipc_hangup(mountee_phone);
ops->node_put(fn);
ipc_answer_0(callid, rc);
ipc_answer_0(rid, rc);
return;
}
ipc_call_t answer;
aid_t msg = async_send_1(mountee_phone, VFS_OUT_MOUNTED, mr_dev_handle,
&answer);
ipc_forward_fast(callid, mountee_phone, 0, 0, 0, IPC_FF_ROUTE_FROM_ME);
async_wait_for(msg, &rc);
if (rc == EOK) {
fn->mp_data.mp_active = true;
fn->mp_data.fs_handle = mr_fs_handle;
fn->mp_data.dev_handle = mr_dev_handle;
fn->mp_data.phone = mountee_phone;
}
/*
* Do not release the FS node so that it stays in memory.
*/
ipc_answer_3(rid, rc, IPC_GET_ARG1(answer), IPC_GET_ARG2(answer),
IPC_GET_ARG3(answer));
}
 
/** Lookup VFS triplet by name in the file system name space.
*
* The path passed in the PLB must be in the canonical file system path format
* as returned by the canonify() function.
*
* @param ops libfs operations structure with function pointers to
* file system implementation
* @param fs_handle File system handle of the file system where to perform
* the lookup.
* @param rid Request ID of the VFS_OUT_LOOKUP request.
* @param request VFS_OUT_LOOKUP request data itself.
*
*/
void libfs_lookup(libfs_ops_t *ops, fs_handle_t fs_handle, ipc_callid_t rid,
ipc_call_t *request)
{
unsigned first = IPC_GET_ARG1(*request);
unsigned last = IPC_GET_ARG2(*request);
unsigned next = first;
dev_handle_t dev_handle = IPC_GET_ARG3(*request);
int lflag = IPC_GET_ARG4(*request);
fs_index_t index = IPC_GET_ARG5(*request); /* when L_LINK specified */
char component[NAME_MAX + 1];
int len;
 
if (last < next)
last += PLB_SIZE;
 
fs_node_t *par = NULL;
fs_node_t *cur = ops->root_get(dev_handle);
fs_node_t *tmp = NULL;
 
if (cur->mp_data.mp_active) {
ipc_forward_slow(rid, cur->mp_data.phone, VFS_OUT_LOOKUP,
next, last, cur->mp_data.dev_handle, lflag, index,
IPC_FF_ROUTE_FROM_ME);
ops->node_put(cur);
return;
}
 
if (ops->plb_get_char(next) == '/')
next++; /* eat slash */
while (next <= last && ops->has_children(cur)) {
/* collect the component */
len = 0;
while ((next <= last) && (ops->plb_get_char(next) != '/')) {
if (len + 1 == NAME_MAX) {
/* component length overflow */
ipc_answer_0(rid, ENAMETOOLONG);
goto out;
}
component[len++] = ops->plb_get_char(next);
next++; /* process next character */
}
 
assert(len);
component[len] = '\0';
next++; /* eat slash */
 
/* match the component */
tmp = ops->match(cur, component);
if (tmp && tmp->mp_data.mp_active) {
if (next > last)
next = last = first;
else
next--;
ipc_forward_slow(rid, tmp->mp_data.phone,
VFS_OUT_LOOKUP, next, last, tmp->mp_data.dev_handle,
lflag, index, IPC_FF_ROUTE_FROM_ME);
ops->node_put(cur);
ops->node_put(tmp);
if (par)
ops->node_put(par);
return;
}
 
/* handle miss: match amongst siblings */
if (!tmp) {
if (next <= last) {
/* there are unprocessed components */
ipc_answer_0(rid, ENOENT);
goto out;
}
/* miss in the last component */
if (lflag & (L_CREATE | L_LINK)) {
/* request to create a new link */
if (!ops->is_directory(cur)) {
ipc_answer_0(rid, ENOTDIR);
goto out;
}
fs_node_t *fn;
if (lflag & L_CREATE)
fn = ops->create(dev_handle, lflag);
else
fn = ops->node_get(dev_handle,
index);
if (fn) {
int rc;
 
rc = ops->link(cur, fn, component);
if (rc != EOK) {
if (lflag & L_CREATE) {
(void)ops->destroy(fn);
}
ipc_answer_0(rid, rc);
} else {
ipc_answer_5(rid, EOK,
fs_handle, dev_handle,
ops->index_get(fn),
ops->size_get(fn),
ops->lnkcnt_get(fn));
ops->node_put(fn);
}
} else {
ipc_answer_0(rid, ENOSPC);
}
goto out;
}
ipc_answer_0(rid, ENOENT);
goto out;
}
 
if (par)
ops->node_put(par);
 
/* descend one level */
par = cur;
cur = tmp;
tmp = NULL;
}
 
/* handle miss: excessive components */
if (next <= last && !ops->has_children(cur)) {
if (lflag & (L_CREATE | L_LINK)) {
if (!ops->is_directory(cur)) {
ipc_answer_0(rid, ENOTDIR);
goto out;
}
 
/* collect next component */
len = 0;
while (next <= last) {
if (ops->plb_get_char(next) == '/') {
/* more than one component */
ipc_answer_0(rid, ENOENT);
goto out;
}
if (len + 1 == NAME_MAX) {
/* component length overflow */
ipc_answer_0(rid, ENAMETOOLONG);
goto out;
}
component[len++] = ops->plb_get_char(next);
next++; /* process next character */
}
assert(len);
component[len] = '\0';
fs_node_t *fn;
if (lflag & L_CREATE)
fn = ops->create(dev_handle, lflag);
else
fn = ops->node_get(dev_handle, index);
if (fn) {
int rc;
 
rc = ops->link(cur, fn, component);
if (rc != EOK) {
if (lflag & L_CREATE)
(void)ops->destroy(fn);
ipc_answer_0(rid, rc);
} else {
ipc_answer_5(rid, EOK,
fs_handle, dev_handle,
ops->index_get(fn),
ops->size_get(fn),
ops->lnkcnt_get(fn));
ops->node_put(fn);
}
} else {
ipc_answer_0(rid, ENOSPC);
}
goto out;
}
ipc_answer_0(rid, ENOENT);
goto out;
}
 
/* handle hit */
if (lflag & L_UNLINK) {
unsigned old_lnkcnt = ops->lnkcnt_get(cur);
int res = ops->unlink(par, cur, component);
ipc_answer_5(rid, (ipcarg_t)res, fs_handle, dev_handle,
ops->index_get(cur), ops->size_get(cur), old_lnkcnt);
goto out;
}
if (((lflag & (L_CREATE | L_EXCLUSIVE)) == (L_CREATE | L_EXCLUSIVE)) ||
(lflag & L_LINK)) {
ipc_answer_0(rid, EEXIST);
goto out;
}
if ((lflag & L_FILE) && (ops->is_directory(cur))) {
ipc_answer_0(rid, EISDIR);
goto out;
}
if ((lflag & L_DIRECTORY) && (ops->is_file(cur))) {
ipc_answer_0(rid, ENOTDIR);
goto out;
}
 
ipc_answer_5(rid, EOK, fs_handle, dev_handle, ops->index_get(cur),
ops->size_get(cur), ops->lnkcnt_get(cur));
 
out:
if (par)
ops->node_put(par);
if (cur)
ops->node_put(cur);
if (tmp)
ops->node_put(tmp);
}
 
void libfs_stat(libfs_ops_t *ops, fs_handle_t fs_handle, ipc_callid_t rid,
ipc_call_t *request)
{
dev_handle_t dev_handle = (dev_handle_t) IPC_GET_ARG1(*request);
fs_index_t index = (fs_index_t) IPC_GET_ARG2(*request);
fs_node_t *fn = ops->node_get(dev_handle, index);
 
ipc_callid_t callid;
size_t size;
if (!ipc_data_read_receive(&callid, &size) ||
size != sizeof(struct stat)) {
ipc_answer_0(callid, EINVAL);
ipc_answer_0(rid, EINVAL);
return;
}
 
struct stat stat;
memset(&stat, 0, sizeof(struct stat));
stat.fs_handle = fs_handle;
stat.dev_handle = dev_handle;
stat.index = index;
stat.lnkcnt = ops->lnkcnt_get(fn);
stat.is_file = ops->is_file(fn);
stat.size = ops->size_get(fn);
 
ipc_data_read_finalize(callid, &stat, sizeof(struct stat));
ipc_answer_0(rid, EOK);
}
 
/** Open VFS triplet.
*
* @param ops libfs operations structure with function pointers to
* file system implementation
* @param rid Request ID of the VFS_OUT_OPEN_NODE request.
* @param request VFS_OUT_OPEN_NODE request data itself.
*
*/
void libfs_open_node(libfs_ops_t *ops, fs_handle_t fs_handle, ipc_callid_t rid,
ipc_call_t *request)
{
dev_handle_t dev_handle = IPC_GET_ARG1(*request);
fs_index_t index = IPC_GET_ARG2(*request);
fs_node_t *node = ops->node_get(dev_handle, index);
if (node == NULL) {
ipc_answer_0(rid, ENOENT);
return;
}
ipc_answer_5(rid, EOK, fs_handle, dev_handle, index,
ops->size_get(node), ops->lnkcnt_get(node));
ops->node_put(node);
}
 
/** @}
*/
/tags/0.4.1/uspace/lib/libfs/libfs.h
0,0 → 1,94
/*
* Copyright (c) 2009 Jakub Jermar
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* - Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* - Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* - The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
 
/** @addtogroup libfs
* @{
*/
/**
* @file
*/
 
#ifndef LIBFS_LIBFS_H_
#define LIBFS_LIBFS_H_
 
#include <ipc/vfs.h>
#include <stdint.h>
#include <ipc/ipc.h>
#include <async.h>
#include <devmap.h>
 
typedef struct {
bool mp_active;
int phone;
fs_handle_t fs_handle;
dev_handle_t dev_handle;
} mp_data_t;
 
typedef struct {
mp_data_t mp_data; /**< Mount point info. */
void *data; /**< Data of the file system implementation. */
} fs_node_t;
 
typedef struct {
fs_node_t * (* match)(fs_node_t *, const char *);
fs_node_t * (* node_get)(dev_handle_t, fs_index_t);
void (* node_put)(fs_node_t *);
fs_node_t * (* create)(dev_handle_t, int);
int (* destroy)(fs_node_t *);
int (* link)(fs_node_t *, fs_node_t *, const char *);
int (* unlink)(fs_node_t *, fs_node_t *, const char *);
fs_index_t (* index_get)(fs_node_t *);
size_t (* size_get)(fs_node_t *);
unsigned (* lnkcnt_get)(fs_node_t *);
bool (* has_children)(fs_node_t *);
fs_node_t *(* root_get)(dev_handle_t);
char (* plb_get_char)(unsigned pos);
bool (* is_directory)(fs_node_t *);
bool (* is_file)(fs_node_t *);
} libfs_ops_t;
 
typedef struct {
int fs_handle; /**< File system handle. */
ipcarg_t vfs_phonehash; /**< Initial VFS phonehash. */
uint8_t *plb_ro; /**< Read-only PLB view. */
} fs_reg_t;
 
extern int fs_register(int, fs_reg_t *, vfs_info_t *, async_client_conn_t);
 
extern void fs_node_initialize(fs_node_t *);
 
extern void libfs_mount(libfs_ops_t *, fs_handle_t, ipc_callid_t, ipc_call_t *);
extern void libfs_lookup(libfs_ops_t *, fs_handle_t, ipc_callid_t, ipc_call_t *);
extern void libfs_stat(libfs_ops_t *, fs_handle_t, ipc_callid_t, ipc_call_t *);
extern void libfs_open_node(libfs_ops_t *, fs_handle_t, ipc_callid_t,
ipc_call_t *);
 
#endif
 
/** @}
*/
/tags/0.4.1/uspace/lib/libfs/Makefile
0,0 → 1,67
#
# Copyright (c) 2005 Martin Decky
# Copyright (c) 2007 Jakub Jermar
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
#
# - Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# - Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
# - The name of the author may not be used to endorse or promote products
# derived from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
# THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
 
## Common compiler flags
#
 
LIBC_PREFIX = ../libc
 
## Setup toolchain
#
 
include $(LIBC_PREFIX)/Makefile.toolchain
 
CFLAGS += -Iinclude
 
## Sources
#
 
SOURCES = \
libfs.c
 
OBJECTS := $(addsuffix .o,$(basename $(SOURCES)))
 
.PHONY: all clean depend
 
all: libfs.a
 
-include Makefile.depend
 
clean:
-rm -f libfs.a Makefile.depend
find . -name '*.o' -follow -exec rm \{\} \;
 
depend:
-makedepend -f - -- $(DEPEND_DEFS) $(CFLAGS) -- $(SOURCES) > Makefile.depend 2> /dev/null
 
libfs.a: depend $(OBJECTS)
$(AR) rc libfs.a $(OBJECTS)
 
%.o: %.c
$(CC) $(DEFS) $(CFLAGS) -c $< -o $@