![[LWN Logo]](/images/lcorner.png) |
|
![[LWN.net]](/images/Included.png) |
From: Rusty Russell <rusty@rustcorp.com.au>
To: linux-kernel@vger.kernel.org
Subject: [PATCH] /proc/sys replacement code (in progress)
Date: Thu, 20 Dec 2001 17:15:40 +1100
After my last /proc patch (see "Replacement for current /proc of shit"
thread), it has become clear that my new "single value" proc code is
more appropriate in a new filesystem. I have also removed the
(complex) dynamic directory support.
The patch against 2.5.1 is included below: it doesn't yet have the
glue layer to support the old sysctl code, which I'll work on when I
come back from Christmas holidays (early January).
For full details:
http://www.kernel.org/pub/linux/kernel/people/rusty/
Comments *ON THE CODE* eagerly sought,
Rusty.
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/fs/sys/test.c working-2.4.16-uml-proc/fs/sys/test.c
--- linux-2.4.16-uml/fs/sys/test.c Thu Jan 1 10:00:00 1970
+++ working-2.4.16-uml-proc/fs/sys/test.c Thu Dec 20 15:34:27 2001
@@ -0,0 +1,54 @@
+/* Test stuff for new sys */
+#include <linux/sys_fs.h>
+#include <linux/spinlock.h>
+#include <asm/semaphore.h>
+
+static spinlock_t test_lock = SPIN_LOCK_UNLOCKED;
+static DECLARE_MUTEX(test_sem);
+
+unsigned int a, b, c;
+
+static int set_off(const char *dirname,
+ const char *filename,
+ const char *buffer,
+ unsigned int size,
+ void *arg)
+{
+ if (size < sizeof("REMOVE")-1)
+ return -1; /* FIXME: -EINVAL */
+ if (strncmp("REMOVE", buffer, sizeof("REMOVE")-1) == 0) {
+ /* Remove everything */
+ unsys("test", "a");
+ unsys("test", "b");
+ unsys("test", "c");
+ unsys("test", "off");
+ }
+ return 0;
+}
+
+static int get_off(const char *dirname,
+ const char *filename,
+ char *buffer,
+ unsigned int size,
+ void *arg)
+{
+ return snprintf(buffer, size, "Write REMOVE here to remove me\n");
+}
+
+static int __init init_test(void)
+{
+ if (sys("test", "a", a, uint, S_IWUSR|S_IRUGO) != 0)
+ printk("Failed: a");
+ if (sys_spinlock("test", "b", b, uint, &test_lock, S_IWUSR|S_IRUGO)
+ != 0)
+ printk("Failed: b");
+ if (sys_sem("test", "c", c, uint, &test_sem, S_IWUSR|S_IRUGO) != 0)
+ printk("Failed: c");
+ if (sys_callfn("test", "off", get_off, set_off, NULL, S_IWUSR|S_IRUGO)
+ != 0)
+ printk("Failed: off");
+
+ return 0;
+}
+
+module_init(init_test);
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/include/linux/sys_fs.h working-2.4.16-uml-proc/include/linux/sys_fs.h
--- linux-2.4.16-uml/include/linux/sys_fs.h Thu Jan 1 10:00:00 1970
+++ working-2.4.16-uml-proc/include/linux/sys_fs.h Thu Dec 20 13:58:21 2001
@@ -0,0 +1,127 @@
+#ifndef _LINUX_SYS_FS_H
+#define _LINUX_SYS_FS_H
+/* Kernel presentation interface. Just be grateful I didn't call it
+ systables... */
+/* (C) 2001 Rusty Russell. */
+
+#include <linux/config.h>
+#include <linux/spinlock.h>
+#include <linux/types.h>
+#include <linux/stat.h>
+#include <linux/typecheck.h>
+
+/* Commit the contents of this (NUL-terminated) buffer if possible.
+ -errno indicates error. */
+typedef int (sys_commitfn_t)(const char *dirname,
+ const char *filename,
+ const char *buffer,
+ unsigned int size,
+ void *arg);
+/* Fetch the contents into buffer: return size used (or needed), or
+ -errno. */
+typedef int (sys_fetchfn_t)(const char *dirname,
+ const char *filename,
+ char *buffer,
+ unsigned int size,
+ void *arg);
+
+/* Register a sys entry of the given type. */
+#define sys(dir, fname, var, type, perms) \
+ __sys(dir, fname, S_IFREG|(perms), \
+ __new_sys(&var, \
+ ((perms)&S_IRUGO) ? sys_fetch_##type : NULL, \
+ ((perms)&S_IWUGO) ? sys_commit_##type : NULL))
+
+/* Register a sys entry protected by a spinlock. */
+#define sys_spinlock(dir, fname, var, type, lock, p) \
+ __sys(dir, fname, S_IFREG|(p), \
+ __new_sys_lock(&var, lock, \
+ ((p)&S_IRUGO) ? sys_fetch_##type : NULL, \
+ ((p)&S_IWUGO) ? sys_commit_##type : NULL))
+
+/* Register a sys entry protected by a semaphore. */
+#define sys_sem(dir, fname, var, type, sem, p) \
+ __sys(dir, fname, S_IFREG|(p), \
+ __new_sys_sem(&var, sem, \
+ ((p)&S_IRUGO) ? sys_fetch_##type : NULL, \
+ ((p)&S_IWUGO) ? sys_commit_##type : NULL))
+
+/* Generic sys interface. */
+#define sys_callfn(dir, fname, fetch, commit, datap, perms) \
+ __sys(dir, fname, S_IFREG|(perms), __new_sys(datap, fetch, commit))
+
+/* These exist, believe me */
+struct semaphore;
+struct sys_data;
+
+#ifdef CONFIG_SYS_FS
+/* Low level functions */
+int __sys(const char *dirname, const char *fname, int mode,
+ struct sys_data *pdata);
+struct sys_data *__new_sys(void *arg, sys_fetchfn_t *, sys_commitfn_t *);
+struct sys_data *__new_sys_lock(void *arg, spinlock_t *lock,
+ sys_fetchfn_t *, sys_commitfn_t *);
+struct sys_data *__new_sys_sem(void *arg, struct semaphore *sem,
+ sys_fetchfn_t *, sys_commitfn_t *);
+
+/* Release a sys entry */
+void unsys(const char *dir, const char *fname);
+
+#else
+static inline void unsys(const char *dir, const char *fname)
+{
+}
+
+static inline int __sys(const char *dirname, const char *fname, int mode,
+ struct sys_data *pdata)
+{
+ return 0;
+}
+
+static inline struct sys_data *__new_sys(void *arg,
+ sys_fetchfn_t *fetch,
+ sys_commitfn_t *commit)
+{
+ return (struct sys_data *)-1;
+}
+static inline struct sys_data *__new_sys_lock(void *arg, spinlock_t *lock,
+ sys_fetchfn_t *fetch,
+ sys_commitfn_t *commit)
+{
+ return (struct sys_data *)-1;
+}
+static inline struct sys_data *__new_sys_sem(void *arg, struct semaphore *sem,
+ sys_fetchfn_t *fetch,
+ sys_commitfn_t *commit)
+{
+ return (struct sys_data *)-1;
+}
+#endif /*CONFIG_SYS_FS*/
+
+/* Helper parsing routines. You can write your own, too. */
+sys_fetchfn_t sys_fetch_short;
+sys_fetchfn_t sys_fetch_ushort;
+sys_fetchfn_t sys_fetch_int;
+sys_fetchfn_t sys_fetch_uint;
+sys_fetchfn_t sys_fetch_long;
+sys_fetchfn_t sys_fetch_ulong;
+sys_fetchfn_t sys_fetch_bool;
+
+sys_commitfn_t sys_commit_short;
+sys_commitfn_t sys_commit_ushort;
+sys_commitfn_t sys_commit_int;
+sys_commitfn_t sys_commit_uint;
+sys_commitfn_t sys_commit_long;
+sys_commitfn_t sys_commit_ulong;
+sys_commitfn_t sys_commit_bool;
+
+/* Internal use */
+struct sys_data
+{
+ /* User-defined argument for routines */
+ void *arg;
+
+ sys_commitfn_t *commit;
+ sys_fetchfn_t *fetch;
+};
+#endif /* _LINUX_SYS_FS_H */
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/fs/Config.in working-2.4.16-uml-proc/fs/Config.in
--- linux-2.4.16-uml/fs/Config.in Thu Dec 13 17:00:55 2001
+++ working-2.4.16-uml-proc/fs/Config.in Thu Dec 20 12:31:41 2001
@@ -61,6 +61,8 @@
bool '/proc file system support' CONFIG_PROC_FS
+bool '/proc/sys file system support (EXPERIMENTAL)' CONFIG_SYS_FS $CONFIG_EXPERIMENTAL
+
dep_bool '/dev file system support (EXPERIMENTAL)' CONFIG_DEVFS_FS $CONFIG_EXPERIMENTAL
dep_bool ' Automatically mount at boot' CONFIG_DEVFS_MOUNT $CONFIG_DEVFS_FS
dep_bool ' Debug devfs' CONFIG_DEVFS_DEBUG $CONFIG_DEVFS_FS
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/fs/Makefile working-2.4.16-uml-proc/fs/Makefile
--- linux-2.4.16-uml/fs/Makefile Thu Dec 13 17:00:55 2001
+++ working-2.4.16-uml-proc/fs/Makefile Thu Dec 20 12:31:41 2001
@@ -23,6 +23,7 @@
endif
subdir-$(CONFIG_PROC_FS) += proc
+subdir-$(CONFIG_SYS_FS) += sys
subdir-y += partitions
# Do not add any filesystems before this line
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/fs/sys/Makefile working-2.4.16-uml-proc/fs/sys/Makefile
--- linux-2.4.16-uml/fs/sys/Makefile Thu Jan 1 10:00:00 1970
+++ working-2.4.16-uml-proc/fs/sys/Makefile Thu Dec 20 12:31:41 2001
@@ -0,0 +1,9 @@
+# Makefile for the Linux sys filesystem routines.
+
+O_TARGET := sys.o
+
+export-objs := helper.o sys_core.o
+
+obj-y := sys_core.o helper.o test.o
+
+include $(TOPDIR)/Rules.make
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/fs/sys/helper.c working-2.4.16-uml-proc/fs/sys/helper.c
--- linux-2.4.16-uml/fs/sys/helper.c Thu Jan 1 10:00:00 1970
+++ working-2.4.16-uml-proc/fs/sys/helper.c Thu Dec 20 12:31:41 2001
@@ -0,0 +1,233 @@
+/* Copyright (C) 2001 Rusty Russell.
+
+ 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 the Free Software
+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+#include <linux/sys_fs.h>
+#include <linux/slab.h>
+#include <linux/errno.h>
+#include <linux/dcache.h>
+#include <linux/init.h>
+#include <linux/module.h>
+
+/* Wrapper for user's real sys functions */
+struct sdata_wrapper
+{
+ struct sys_data sdata;
+ sys_fetchfn_t *fetch;
+ sys_commitfn_t *commit;
+ void *lock;
+ void *userarg;
+};
+
+static struct sys_data *new_wrapper(sys_fetchfn_t *userfetch,
+ sys_commitfn_t *usercommit,
+ void *userarg,
+ sys_fetchfn_t *wrapfetch,
+ sys_commitfn_t *wrapcommit,
+ void *lock)
+{
+ struct sdata_wrapper *swrap;
+
+ swrap = kmalloc(sizeof(*swrap), GFP_KERNEL);
+ if (swrap) {
+ swrap->sdata.arg = swrap;
+ swrap->sdata.fetch = wrapfetch;
+ swrap->sdata.commit = wrapcommit;
+ swrap->fetch = userfetch;
+ swrap->commit = usercommit;
+ swrap->lock = lock;
+ swrap->userarg = userarg;
+ }
+ return &swrap->sdata;
+}
+
+static int lock_fetch(const char *dirname, const char *fname,
+ char *buffer, unsigned int size, void *arg)
+{
+ struct sdata_wrapper *swrap = arg;
+ int ret;
+
+ spin_lock_irq(swrap->lock);
+ ret = swrap->fetch(dirname, fname, swrap->userarg, size, buffer);
+ spin_unlock_irq(swrap->lock);
+
+ return ret;
+}
+
+static int lock_commit(const char *dirname, const char *fname,
+ const char *buffer, unsigned int size, void *arg)
+{
+ struct sdata_wrapper *swrap = arg;
+ int ret;
+
+ spin_lock_irq(swrap->lock);
+ ret = swrap->commit(dirname, fname, buffer, size, swrap->userarg);
+ spin_unlock_irq(swrap->lock);
+
+ return ret;
+}
+
+struct sys_data *__new_sys_lock(void *arg, spinlock_t *lock,
+ sys_fetchfn_t *fetch,
+ sys_commitfn_t *commit)
+{
+ return new_wrapper(fetch, commit, arg, lock_fetch, lock_commit, lock);
+}
+
+static int sem_fetch(const char *dirname, const char *fname,
+ char *buffer, unsigned int size, void *arg)
+{
+ struct sdata_wrapper *swrap = arg;
+ int ret;
+
+ if (down_interruptible(swrap->lock) != 0)
+ return -EINTR;
+ ret = swrap->fetch(dirname, fname, swrap->userarg, size, buffer);
+ up(swrap->lock);
+
+ return ret;
+}
+
+static int sem_commit(const char *dirname, const char *fname,
+ const char *buffer, unsigned int size, void *arg)
+{
+ struct sdata_wrapper *swrap = arg;
+ int ret;
+
+ if (down_interruptible(swrap->lock) != 0)
+ return -EINTR;
+ ret = swrap->commit(dirname, fname, buffer, size, swrap->userarg);
+ up(swrap->lock);
+
+ return ret;
+}
+
+struct sys_data *__new_sys_sem(void *arg, struct semaphore *sem,
+ sys_fetchfn_t *fetch,
+ sys_commitfn_t *commit)
+{
+ return new_wrapper(fetch, commit, arg, sem_fetch, sem_commit, sem);
+}
+
+int sys_fetch_short(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *shortp)
+{
+ return snprintf(outbuf, size, "%hi\n", *(short *)shortp);
+}
+
+int sys_commit_short(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *shortp)
+{
+ if (sscanf(inbuf, "%hi", (short *)shortp) != 1) return -EINVAL;
+ return 0;
+}
+
+int sys_fetch_ushort(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *ushortp)
+{
+ return snprintf(outbuf, size, "%hu\n", *(unsigned short *)ushortp);
+}
+
+int sys_commit_ushort(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *ushortp)
+{
+ if (sscanf(inbuf, "%hu", (unsigned short *)ushortp) != 1)
+ return -EINVAL;
+ return 0;
+}
+
+int sys_fetch_int(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *intp)
+{
+ return snprintf(outbuf, size, "%i\n", *(int *)intp);
+}
+
+int sys_commit_int(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *intp)
+{
+ if (sscanf(inbuf, "%i", (int *)intp) != 1) return -EINVAL;
+ return 0;
+}
+
+int sys_fetch_uint(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *uintp)
+{
+ return snprintf(outbuf, size, "%u\n", *(unsigned int *)uintp);
+}
+
+int sys_commit_uint(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *uintp)
+{
+ if (sscanf(inbuf, "%u", (unsigned int *)uintp) != 1) return -EINVAL;
+ return 0;
+}
+
+int sys_fetch_long(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *longp)
+{
+ return snprintf(outbuf, size, "%li\n", *(long *)longp);
+}
+
+int sys_commit_long(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *longp)
+{
+ if (sscanf(inbuf, "%li", (long *)longp) != 1) return -EINVAL;
+ return 0;
+}
+
+int sys_fetch_ulong(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *ulongp)
+{
+ return snprintf(outbuf, size, "%lu\n", *(long *)ulongp);
+}
+
+int sys_commit_ulong(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *ulongp)
+{
+ if (sscanf(inbuf, "%lu", (unsigned long *)ulongp) != 1) return -EINVAL;
+ return 0;
+}
+
+int sys_fetch_bool(const char *dir, const char *fname,
+ char *outbuf, unsigned int size, void *boolp)
+{
+ if (*(int *)boolp) return snprintf(outbuf, size, "y\n");
+ else return snprintf(outbuf, size, "n\n");
+}
+
+int sys_commit_bool(const char *dir, const char *fname,
+ const char *inbuf, unsigned int size, void *boolp)
+{
+ if (inbuf[0] == 'y' || inbuf[0] == 'Y')
+ *(int *)boolp = 1;
+ else if (inbuf[0] == 'n' || inbuf[0] == 'N')
+ *(int *)boolp = 0;
+ else return sys_commit_int(dir, fname, inbuf, size, boolp);
+ return 0;
+}
+
+EXPORT_SYMBOL_GPL(sys_fetch_short);
+EXPORT_SYMBOL_GPL(sys_commit_short);
+EXPORT_SYMBOL_GPL(sys_fetch_ushort);
+EXPORT_SYMBOL_GPL(sys_commit_ushort);
+EXPORT_SYMBOL_GPL(sys_fetch_int);
+EXPORT_SYMBOL_GPL(sys_commit_int);
+EXPORT_SYMBOL_GPL(sys_fetch_uint);
+EXPORT_SYMBOL_GPL(sys_commit_uint);
+EXPORT_SYMBOL_GPL(sys_fetch_long);
+EXPORT_SYMBOL_GPL(sys_commit_long);
+EXPORT_SYMBOL_GPL(sys_fetch_ulong);
+EXPORT_SYMBOL_GPL(sys_commit_ulong);
+EXPORT_SYMBOL_GPL(sys_fetch_bool);
+EXPORT_SYMBOL_GPL(sys_commit_bool);
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/fs/sys/sys_core.c working-2.4.16-uml-proc/fs/sys/sys_core.c
--- linux-2.4.16-uml/fs/sys/sys_core.c Thu Jan 1 10:00:00 1970
+++ working-2.4.16-uml-proc/fs/sys/sys_core.c Thu Dec 20 15:34:02 2001
@@ -0,0 +1,550 @@
+/* Copyright (C) 2001 Rusty Russell.
+
+ 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 the Free Software
+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
+#include <linux/fs.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/stat.h>
+#include <linux/slab.h>
+#include <linux/pagemap.h>
+#include <linux/sys_fs.h>
+#include <linux/module.h>
+#include <asm/semaphore.h>
+#include <asm/uaccess.h>
+
+/* Sys mount point */
+static struct vfsmount *sys_mnt;
+
+/* Serialize insert/delete and mounts */
+static DECLARE_MUTEX(sys_semaphore);
+
+#define SIMPLE_SYSFS_MAGIC 0x62121174
+
+/* Approximate upper ceiling for memory usage per fs */
+#define SYSFS_MAX_SIZE (1024 - sizeof(unsigned int))
+
+/* TBD */
+static struct dentry_operations sys_dentry_ops;
+static struct file_operations sys_dirops, sys_fileops;
+static struct inode_operations sys_inodeops;
+static struct super_operations sys_super_ops;
+
+struct sys_buffer
+{
+ unsigned int len;
+ /* One is for the nul terminator */
+ char buffer[1];
+};
+
+struct sys_dirinfo
+{
+ /* When it drops to zero, unlink it */
+ unsigned int usage;
+ /* Full nul-terminated path name */
+ char dirname[0];
+};
+
+struct sys_data *__new_sys(void *arg,
+ sys_fetchfn_t *fetch,
+ sys_commitfn_t *commit)
+{
+ struct sys_data *pdata;
+
+ pdata = kmalloc(sizeof(*pdata), GFP_KERNEL);
+ if (pdata) {
+ pdata->arg = arg;
+ pdata->fetch = fetch;
+ pdata->commit = commit;
+ }
+ return pdata;
+}
+
+/* Convenience routine to make an inode */
+static struct inode *
+new_sys_inode(struct super_block *sb, int mode,
+ struct inode_operations *op,
+ struct file_operations *fop)
+{
+ struct inode *inode = new_inode(sb);
+
+ if (!inode)
+ return NULL;
+
+ inode->i_mode = mode;
+ inode->i_uid = 0;
+ inode->i_gid = 0;
+ inode->i_blksize = PAGE_CACHE_SIZE;
+ inode->i_blocks = 0;
+ inode->i_rdev = NODEV;
+ inode->i_mapping->a_ops = NULL;
+ inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+ switch (mode & S_IFMT) {
+ case S_IFREG:
+ inode->i_fop = fop;
+ break;
+ case S_IFDIR:
+ inode->i_fop = fop;
+ inode->i_op = op;
+ break;
+ default:
+ BUG();
+ break;
+ }
+
+ return inode;
+}
+
+/* Make a new sys entry in this directory. */
+static int make_sys_entry(struct dentry *dir,
+ const char *fname,
+ int mode,
+ void *data,
+ struct inode_operations *op,
+ struct file_operations *fop)
+{
+ struct inode *inode;
+ struct dentry *dentry;
+ struct qstr qstr;
+
+ /* Create qstr for this entry */
+ qstr.name = fname;
+ qstr.len = strlen(fname);
+ qstr.hash = full_name_hash(qstr.name, qstr.len);
+
+ down(&sys_semaphore);
+ /* Does it already exist? */
+ dentry = d_lookup(dir, &qstr);
+ if (dentry) {
+ dput(dentry);
+ up(&sys_semaphore);
+ return -EEXIST;
+ }
+
+ /* Doesn't exist: create inode */
+ inode = new_sys_inode(dir->d_sb, mode, op, fop);
+ if (!inode) {
+ up(&sys_semaphore);
+ return -ENOMEM;
+ }
+
+ /* Create dentry */
+ dentry = d_alloc(dir, &qstr);
+ if (!dentry) {
+ iput(inode);
+ up(&sys_semaphore);
+ return -ENOMEM;
+ }
+ dentry->d_op = &sys_dentry_ops;
+ dentry->d_fsdata = data;
+ d_add(dentry, inode);
+
+ /* Pin the dentry here, so it doesn't get pruned */
+ dget(dentry);
+ up(&sys_semaphore);
+ return 0;
+}
+
+/* Create (static) sys directory if neccessary. */
+static struct dentry *get_sys_dir(const char *dirname)
+{
+ struct dentry *dentry;
+ struct qstr qstr;
+ const char *delim;
+ struct sys_dirinfo *dinfo;
+
+ /* FIXME: Definitely need a better way --RR */
+ dentry = dget(sys_mnt->mnt_sb->s_root);
+ delim = dirname;
+
+ for (;;) {
+ struct dentry *newdentry;
+
+ /* Ignore multiple slashes */
+ while (*delim == '/') delim++;
+ qstr.name = delim;
+ delim = strchr(qstr.name, '/');
+ if (!delim) delim = qstr.name + strlen(qstr.name);
+ qstr.len = delim-(char *)qstr.name;
+ qstr.hash = full_name_hash(qstr.name, qstr.len);
+
+ if (qstr.len == 0)
+ break;
+
+ /* If entry doesn't exist, create it */
+ newdentry = d_lookup(dentry, &qstr);
+ if (!newdentry) {
+ char fname[qstr.len+1];
+ int ret;
+
+ dinfo = kmalloc(sizeof(*dinfo) + strlen(dirname) + 1,
+ GFP_KERNEL);
+ if (!dinfo) {
+ dput(dentry);
+ return ERR_PTR(-ENOMEM);
+ }
+ strcpy(dinfo->dirname, dirname);
+ dinfo->usage = 0;
+
+ strncpy(fname, qstr.name, qstr.len);
+ fname[qstr.len] = '\0';
+ ret = make_sys_entry(dentry, fname, S_IFDIR|0555,
+ dinfo, &sys_inodeops,&sys_dirops);
+ if (ret < 0) {
+ dput(dentry);
+ return ERR_PTR(ret);
+ }
+ newdentry = d_lookup(dentry, &qstr);
+ if (!newdentry)
+ BUG();
+ }
+ dput(dentry);
+ dentry = newdentry;
+ dinfo = dentry->d_fsdata;
+ dinfo->usage++;
+ }
+ return dentry;
+}
+
+/* Actually add a static sys file */
+int __sys(const char *dirname,
+ const char *fname,
+ int mode,
+ struct sys_data *pdata)
+{
+ struct dentry *dir;
+ int ret;
+
+ if (!pdata)
+ return -ENOMEM;
+
+ dir = get_sys_dir(dirname);
+ if (IS_ERR(dir)) {
+ kfree(pdata);
+ return PTR_ERR(dir);
+ }
+
+ ret = make_sys_entry(dir, fname, mode, pdata,
+ &sys_inodeops, &sys_fileops);
+ dput(dir);
+ if (ret != 0)
+ kfree(pdata);
+ return ret;
+}
+
+/* See if /sys entry exists (entries registered in directory). */
+static struct dentry *sys_lookup(struct inode *dir, struct dentry *dentry)
+{
+ /* Since we place new static entries in the dcache, if we get
+ here, we know the entry does not exist. Create a negative
+ dentry, and return NULL */
+ d_add(dentry, NULL);
+ return NULL;
+}
+
+/* On open, we grab contents if we're readable... */
+static int sys_file_snapshot(struct inode *inode, struct file *filp)
+{
+ struct sys_buffer *buf;
+ struct sys_data *pdata;
+ struct sys_dirinfo *dirdata;
+ int used;
+
+ buf = kmalloc(sizeof(*buf) + SYSFS_MAX_SIZE, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+ buf->len = 0;
+ filp->private_data = buf;
+
+ if (!(filp->f_mode & FMODE_READ))
+ return 0;
+
+ /* Grab the snapshot */
+ pdata = filp->f_dentry->d_fsdata;
+ dirdata = filp->f_dentry->d_parent->d_fsdata;
+
+ used = pdata->fetch(dirdata->dirname,
+ filp->f_dentry->d_name.name,
+ buf->buffer,
+ SYSFS_MAX_SIZE,
+ pdata->arg);
+ if (used < 0) {
+ kfree(buf);
+ return used;
+ }
+ if (used > SYSFS_MAX_SIZE)
+ BUG();
+
+ /* Nul terminate */
+ buf->len = used;
+ return 0;
+}
+
+/* On close, we commit contents if we've been written to... */
+static int sys_file_commit(struct inode *inode, struct file *filp)
+{
+ int ret = 0;
+ struct sys_data *pdata;
+ struct sys_dirinfo *dinfo;
+ struct sys_buffer *buf;
+
+ if (!(filp->f_mode & FMODE_WRITE))
+ goto out;
+
+ pdata = filp->f_dentry->d_fsdata;
+ dinfo = filp->f_dentry->d_parent->d_fsdata;
+ buf = filp->private_data;
+ /* nul-terminate buffer for commit's convenience */
+ buf->buffer[buf->len] = '\0';
+ ret = pdata->commit(dinfo->dirname, filp->f_dentry->d_name.name,
+ buf->buffer, buf->len, pdata->arg);
+
+ out:
+ kfree(filp->private_data);
+ return ret;
+}
+
+/* Copy from buffer */
+static ssize_t sys_file_read(struct file *filp,
+ char *ubuf,
+ size_t size,
+ loff_t *off)
+{
+ struct sys_buffer *buf;
+
+ if (down_interruptible(&sys_semaphore) != 0)
+ return -EINTR;
+
+ buf = filp->private_data;
+ if (size + *off > buf->len)
+ size = buf->len - *off;
+
+ /* Copy from static buffer */
+ if (copy_to_user(ubuf, buf->buffer, size) != 0) {
+ up(&sys_semaphore);
+ return -EFAULT;
+ }
+ up(&sys_semaphore);
+
+ *off += size;
+ return (ssize_t)size;
+}
+
+/* Copy to buffer */
+static ssize_t sys_file_write(struct file *filp,
+ const char *ubuf,
+ size_t size,
+ loff_t *off)
+{
+ struct sys_buffer *buf;
+ struct sys_data *pdata;
+
+ if (down_interruptible(&sys_semaphore) != 0)
+ return -EINTR;
+
+ pdata = filp->f_dentry->d_fsdata;
+
+ buf = filp->private_data;
+ if (*off + size > SYSFS_MAX_SIZE) {
+ up(&sys_semaphore);
+ return -ENOSPC;
+ }
+
+ /* Do actual copy */
+ if (copy_from_user(buf->buffer + *off, ubuf, size) != 0) {
+ up(&sys_semaphore);
+ return -EFAULT;
+ }
+ up(&sys_semaphore);
+ buf->len += size;
+ *off += size;
+
+ return size;
+}
+
+static int sys_nofetch(const char *dirname, const char *fname,
+ char *outbuf, unsigned int len, void *arg)
+{
+ return -ENOENT;
+}
+
+static int sys_nocommit(const char *dirname, const char *fname,
+ const char *inbuf, unsigned int len, void *arg)
+{
+ return -ENOENT;
+}
+
+/* Release a sys entry */
+void unsys(const char *dir, const char *fname)
+{
+ struct dentry *dentry, *parent;
+ const char *delim;
+ struct qstr qstr;
+ struct sys_data *pdata;
+
+ down(&sys_semaphore);
+
+ /* FIXME: There's a better way, right? --RR */
+ dentry = dget(sys_mnt->mnt_sb->s_root);
+ delim = dir;
+ for (;;) {
+ /* Ignore multiple slashes */
+ while (*delim == '/') delim++;
+ qstr.name = delim;
+ delim = strchr(qstr.name, '/');
+ if (!delim) delim = qstr.name + strlen(qstr.name);
+ qstr.len = delim-(char *)qstr.name;
+ qstr.hash = full_name_hash(qstr.name, qstr.len);
+
+ if (qstr.len == 0)
+ break;
+
+ dentry = d_lookup(dentry, &qstr);
+ if (!dentry)
+ BUG();
+ dput(dentry->d_parent);
+ }
+
+ qstr.name = fname;
+ qstr.len = strlen(fname);
+ qstr.hash = full_name_hash(qstr.name, qstr.len);
+ dentry = d_lookup(dentry, &qstr);
+ if (!dentry)
+ BUG();
+ dput(dentry->d_parent);
+
+ /* We have the dentry: change the private area so it doesn't
+ enter the caller any more. */
+ pdata = dentry->d_fsdata;
+ pdata->commit = sys_nocommit;
+ pdata->fetch = sys_nofetch;
+
+ /* Save parent */
+ parent = dentry->d_parent;
+
+ /* First unpin, then delete */
+ dput(dentry);
+ d_delete(dentry);
+
+ /* Now clean up parents. */
+ for (dentry = parent; parent != sys_mnt->mnt_root; dentry = parent) {
+ struct sys_dirinfo *dinfo = dentry->d_fsdata;
+
+ /* Grab parent first... */
+ parent = dentry->d_parent;
+ if (--dinfo->usage == 0) {
+ dput(dentry);
+ d_delete(dentry);
+ }
+ }
+
+ up(&sys_semaphore);
+}
+
+/* FIXME: I have no idea what all this does: stolen from old /proc --RR */
+static int sys_statfs(struct super_block *sb, struct statfs *buf)
+{
+ buf->f_type = SIMPLE_SYSFS_MAGIC;
+ buf->f_bsize = PAGE_SIZE/sizeof(long);
+ buf->f_bfree = 0;
+ buf->f_bavail = 0;
+ buf->f_ffree = 0;
+ buf->f_namelen = NAME_MAX;
+ return 0;
+}
+
+static struct super_block *sys_read_super(struct super_block *s,
+ void *data,
+ int silent)
+{
+ struct inode * root_inode;
+
+ s->s_blocksize = 1024;
+ s->s_blocksize_bits = 10;
+ s->s_magic = SIMPLE_SYSFS_MAGIC;
+ s->s_op = &sys_super_ops;
+
+ root_inode = new_sys_inode(s, S_IFDIR|0555, &sys_inodeops,&sys_dirops);
+ if (!root_inode) return NULL;
+
+ /* Block concurrent mounts */
+ down(&sys_semaphore);
+
+ s->s_root = d_alloc_root(root_inode);
+ if (!s->s_root) {
+ iput(root_inode);
+ up(&sys_semaphore);
+ return NULL;
+ }
+ up(&sys_semaphore);
+ return s;
+}
+
+/* Free the private area when dentry is freed. */
+static void sys_release(struct dentry *dentry)
+{
+ kfree(dentry->d_fsdata);
+}
+
+static struct inode_operations sys_inodeops = {
+ lookup: sys_lookup,
+};
+
+static struct file_operations sys_dirops = {
+ read: generic_read_dir,
+ readdir: dcache_readdir,
+};
+
+static struct dentry_operations sys_dentry_ops = {
+ d_release: sys_release,
+};
+
+static struct super_operations sys_super_ops = {
+ statfs: sys_statfs,
+ put_inode: force_delete,
+};
+
+static struct file_operations sys_fileops = {
+ open: sys_file_snapshot,
+ release: sys_file_commit,
+ read: sys_file_read,
+ write: sys_file_write,
+};
+
+static DECLARE_FSTYPE(sys_fs_type, "sys", sys_read_super, FS_SINGLE);
+
+static struct {
+ struct sys_dirinfo info;
+ char name[2];
+} root_fsdata = { { 1 }, "/" };
+
+static int __init init_sys_fs(void)
+{
+ register_filesystem(&sys_fs_type);
+ sys_mnt = kern_mount(&sys_fs_type);
+ sys_mnt->mnt_root->d_fsdata = &root_fsdata;
+ return 0;
+}
+
+static void __exit exit_sys_fs(void)
+{
+ unregister_filesystem(&sys_fs_type);
+}
+
+module_init(init_sys_fs);
+module_exit(exit_sys_fs);
+
+EXPORT_SYMBOL(__sys);
+EXPORT_SYMBOL(__new_sys);
+EXPORT_SYMBOL(unsys);
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/include/linux/sysctl.h working-2.4.16-uml-proc/include/linux/sysctl.h
--- linux-2.4.16-uml/include/linux/sysctl.h Thu Dec 13 17:01:04 2001
+++ working-2.4.16-uml-proc/include/linux/sysctl.h Thu Dec 20 13:54:36 2001
@@ -1,3 +1,4 @@
+/* Obsolete: use sys_fs.h -- Rusty Russell. */
/*
* sysctl.h: General linux system control interface
*
diff -urN -I \$.*\$ --exclude TAGS -X /home/rusty/current-dontdiff --minimal linux-2.4.16-uml/include/linux/typecheck.h working-2.4.16-uml-proc/include/linux/typecheck.h
--- linux-2.4.16-uml/include/linux/typecheck.h Thu Jan 1 10:00:00 1970
+++ working-2.4.16-uml-proc/include/linux/typecheck.h Thu Dec 20 12:31:41 2001
@@ -0,0 +1,29 @@
+#ifndef _LINUX_TYPECHECK_H
+#define _LINUX_TYPECHECK_H
+/* Simple macros to do compile-time type checking. IIRC, Jakub
+ Jelinek came up with this idea for the 2.4 module init code. */
+/* (C) 2001 Rusty Russell */
+
+#define __CHECK_byte(name, p) /* Yeah, whatever */
+#define __CHECK_short(name, p) \
+ extern inline short *__CHECK_##name(void) { return(p); }
+#define __CHECK_ushort(name, p) \
+ extern inline unsigned short *__CHECK_##name(void) { return(p); }
+#define __CHECK_int(name, p) \
+ extern inline int *__CHECK_##name(void) { return(p); }
+#define __CHECK_uint(name, p) \
+ extern inline unsigned int *__CHECK_##name(void) { return(p); }
+#define __CHECK_long(name, p) \
+ extern inline long *__CHECK_##name(void) { return(p); }
+#define __CHECK_ulong(name, p) \
+ extern inline unsigned long *__CHECK_##name(void) { return(p); }
+#define __CHECK_charp(name, p) \
+ extern inline char **__CHECK_##name(void) { return(p); }
+#define __CHECK_bool(name, p) \
+ extern inline int *__CHECK_##name(void) { return(p); }
+#define __CHECK_invbool(name, p) \
+ extern inline int *__CHECK_##name(void) { return(p); }
+#define __CHECK_string(name, p) \
+ extern inline char *__CHECK_##name(void) { return(p); }
+
+#endif /* _LINUX_TYPECHECK_H */
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/