KVM: x86 emulator: make set_cr() callback return error if it fails
[linux-2.6.git] / fs / pipe.c
index f7dd21a..279eef9 100644 (file)
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -11,6 +11,7 @@
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/fs.h>
+#include <linux/log2.h>
 #include <linux/mount.h>
 #include <linux/pipe_fs_i.h>
 #include <linux/uio.h>
 #include <linux/pagemap.h>
 #include <linux/audit.h>
 #include <linux/syscalls.h>
+#include <linux/fcntl.h>
 
 #include <asm/uaccess.h>
 #include <asm/ioctls.h>
 
 /*
+ * The max size that a non-root user is allowed to grow the pipe. Can
+ * be set by root in /proc/sys/fs/pipe-max-size
+ */
+unsigned int pipe_max_size = 1048576;
+
+/*
+ * Minimum pipe size, as required by POSIX
+ */
+unsigned int pipe_min_size = PAGE_SIZE;
+
+/*
  * We use a start+len construction, which provides full use of the 
  * allocated memory.
  * -- Florian Coosmann (FGC)
@@ -68,8 +81,8 @@ void pipe_double_lock(struct pipe_inode_info *pipe1,
                pipe_lock_nested(pipe1, I_MUTEX_PARENT);
                pipe_lock_nested(pipe2, I_MUTEX_CHILD);
        } else {
-               pipe_lock_nested(pipe2, I_MUTEX_CHILD);
-               pipe_lock_nested(pipe1, I_MUTEX_PARENT);
+               pipe_lock_nested(pipe2, I_MUTEX_PARENT);
+               pipe_lock_nested(pipe1, I_MUTEX_CHILD);
        }
 }
 
@@ -222,6 +235,7 @@ void *generic_pipe_buf_map(struct pipe_inode_info *pipe,
 
        return kmap(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_map);
 
 /**
  * generic_pipe_buf_unmap - unmap a previously mapped pipe buffer
@@ -241,6 +255,7 @@ void generic_pipe_buf_unmap(struct pipe_inode_info *pipe,
        } else
                kunmap(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_unmap);
 
 /**
  * generic_pipe_buf_steal - attempt to take ownership of a &pipe_buffer
@@ -271,6 +286,7 @@ int generic_pipe_buf_steal(struct pipe_inode_info *pipe,
 
        return 1;
 }
+EXPORT_SYMBOL(generic_pipe_buf_steal);
 
 /**
  * generic_pipe_buf_get - get a reference to a &struct pipe_buffer
@@ -286,6 +302,7 @@ void generic_pipe_buf_get(struct pipe_inode_info *pipe, struct pipe_buffer *buf)
 {
        page_cache_get(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_get);
 
 /**
  * generic_pipe_buf_confirm - verify contents of the pipe buffer
@@ -301,6 +318,7 @@ int generic_pipe_buf_confirm(struct pipe_inode_info *info,
 {
        return 0;
 }
+EXPORT_SYMBOL(generic_pipe_buf_confirm);
 
 /**
  * generic_pipe_buf_release - put a reference to a &struct pipe_buffer
@@ -315,6 +333,7 @@ void generic_pipe_buf_release(struct pipe_inode_info *pipe,
 {
        page_cache_release(buf->page);
 }
+EXPORT_SYMBOL(generic_pipe_buf_release);
 
 static const struct pipe_buf_operations anon_pipe_buf_ops = {
        .can_merge = 1,
@@ -390,7 +409,7 @@ redo:
                        if (!buf->len) {
                                buf->ops = NULL;
                                ops->release(pipe, buf);
-                               curbuf = (curbuf + 1) & (PIPE_BUFFERS-1);
+                               curbuf = (curbuf + 1) & (pipe->buffers - 1);
                                pipe->curbuf = curbuf;
                                pipe->nrbufs = --bufs;
                                do_wakeup = 1;
@@ -472,7 +491,7 @@ pipe_write(struct kiocb *iocb, const struct iovec *_iov,
        chars = total_len & (PAGE_SIZE-1); /* size of the last buffer */
        if (pipe->nrbufs && chars != 0) {
                int lastbuf = (pipe->curbuf + pipe->nrbufs - 1) &
-                                                       (PIPE_BUFFERS-1);
+                                                       (pipe->buffers - 1);
                struct pipe_buffer *buf = pipe->bufs + lastbuf;
                const struct pipe_buf_operations *ops = buf->ops;
                int offset = buf->offset + buf->len;
@@ -518,8 +537,8 @@ redo1:
                        break;
                }
                bufs = pipe->nrbufs;
-               if (bufs < PIPE_BUFFERS) {
-                       int newbuf = (pipe->curbuf + bufs) & (PIPE_BUFFERS-1);
+               if (bufs < pipe->buffers) {
+                       int newbuf = (pipe->curbuf + bufs) & (pipe->buffers-1);
                        struct pipe_buffer *buf = pipe->bufs + newbuf;
                        struct page *page = pipe->tmp_page;
                        char *src;
@@ -580,7 +599,7 @@ redo2:
                        if (!total_len)
                                break;
                }
-               if (bufs < PIPE_BUFFERS)
+               if (bufs < pipe->buffers)
                        continue;
                if (filp->f_flags & O_NONBLOCK) {
                        if (!ret)
@@ -640,7 +659,7 @@ static long pipe_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
                        nrbufs = pipe->nrbufs;
                        while (--nrbufs >= 0) {
                                count += pipe->bufs[buf].len;
-                               buf = (buf+1) & (PIPE_BUFFERS-1);
+                               buf = (buf+1) & (pipe->buffers - 1);
                        }
                        mutex_unlock(&inode->i_mutex);
 
@@ -671,7 +690,7 @@ pipe_poll(struct file *filp, poll_table *wait)
        }
 
        if (filp->f_mode & FMODE_WRITE) {
-               mask |= (nrbufs < PIPE_BUFFERS) ? POLLOUT | POLLWRNORM : 0;
+               mask |= (nrbufs < pipe->buffers) ? POLLOUT | POLLWRNORM : 0;
                /*
                 * Most Unices do not set POLLERR for FIFOs but on Linux they
                 * behave exactly like pipes for poll().
@@ -777,36 +796,55 @@ pipe_rdwr_release(struct inode *inode, struct file *filp)
 static int
 pipe_read_open(struct inode *inode, struct file *filp)
 {
-       /* We could have perhaps used atomic_t, but this and friends
-          below are the only places.  So it doesn't seem worthwhile.  */
+       int ret = -ENOENT;
+
        mutex_lock(&inode->i_mutex);
-       inode->i_pipe->readers++;
+
+       if (inode->i_pipe) {
+               ret = 0;
+               inode->i_pipe->readers++;
+       }
+
        mutex_unlock(&inode->i_mutex);
 
-       return 0;
+       return ret;
 }
 
 static int
 pipe_write_open(struct inode *inode, struct file *filp)
 {
+       int ret = -ENOENT;
+
        mutex_lock(&inode->i_mutex);
-       inode->i_pipe->writers++;
+
+       if (inode->i_pipe) {
+               ret = 0;
+               inode->i_pipe->writers++;
+       }
+
        mutex_unlock(&inode->i_mutex);
 
-       return 0;
+       return ret;
 }
 
 static int
 pipe_rdwr_open(struct inode *inode, struct file *filp)
 {
+       int ret = -ENOENT;
+
        mutex_lock(&inode->i_mutex);
-       if (filp->f_mode & FMODE_READ)
-               inode->i_pipe->readers++;
-       if (filp->f_mode & FMODE_WRITE)
-               inode->i_pipe->writers++;
+
+       if (inode->i_pipe) {
+               ret = 0;
+               if (filp->f_mode & FMODE_READ)
+                       inode->i_pipe->readers++;
+               if (filp->f_mode & FMODE_WRITE)
+                       inode->i_pipe->writers++;
+       }
+
        mutex_unlock(&inode->i_mutex);
 
-       return 0;
+       return ret;
 }
 
 /*
@@ -858,25 +896,32 @@ struct pipe_inode_info * alloc_pipe_info(struct inode *inode)
 
        pipe = kzalloc(sizeof(struct pipe_inode_info), GFP_KERNEL);
        if (pipe) {
-               init_waitqueue_head(&pipe->wait);
-               pipe->r_counter = pipe->w_counter = 1;
-               pipe->inode = inode;
+               pipe->bufs = kzalloc(sizeof(struct pipe_buffer) * PIPE_DEF_BUFFERS, GFP_KERNEL);
+               if (pipe->bufs) {
+                       init_waitqueue_head(&pipe->wait);
+                       pipe->r_counter = pipe->w_counter = 1;
+                       pipe->inode = inode;
+                       pipe->buffers = PIPE_DEF_BUFFERS;
+                       return pipe;
+               }
+               kfree(pipe);
        }
 
-       return pipe;
+       return NULL;
 }
 
 void __free_pipe_info(struct pipe_inode_info *pipe)
 {
        int i;
 
-       for (i = 0; i < PIPE_BUFFERS; i++) {
+       for (i = 0; i < pipe->buffers; i++) {
                struct pipe_buffer *buf = pipe->bufs + i;
                if (buf->ops)
                        buf->ops->release(pipe, buf);
        }
        if (pipe->tmp_page)
                __free_page(pipe->tmp_page);
+       kfree(pipe->bufs);
        kfree(pipe);
 }
 
@@ -887,17 +932,6 @@ void free_pipe_info(struct inode *inode)
 }
 
 static struct vfsmount *pipe_mnt __read_mostly;
-static int pipefs_delete_dentry(struct dentry *dentry)
-{
-       /*
-        * At creation time, we pretended this dentry was hashed
-        * (by clearing DCACHE_UNHASHED bit in d_flags)
-        * At delete time, we restore the truth : not hashed.
-        * (so that dput() can proceed correctly)
-        */
-       dentry->d_flags |= DCACHE_UNHASHED;
-       return 0;
-}
 
 /*
  * pipefs_dname() is called from d_path().
@@ -909,7 +943,6 @@ static char *pipefs_dname(struct dentry *dentry, char *buffer, int buflen)
 }
 
 static const struct dentry_operations pipefs_dentry_operations = {
-       .d_delete       = pipefs_delete_dentry,
        .d_dname        = pipefs_dname,
 };
 
@@ -955,7 +988,7 @@ struct file *create_write_pipe(int flags)
        int err;
        struct inode *inode;
        struct file *f;
-       struct dentry *dentry;
+       struct path path;
        struct qstr name = { .name = "" };
 
        err = -ENFILE;
@@ -964,21 +997,16 @@ struct file *create_write_pipe(int flags)
                goto err;
 
        err = -ENOMEM;
-       dentry = d_alloc(pipe_mnt->mnt_sb->s_root, &name);
-       if (!dentry)
+       path.dentry = d_alloc(pipe_mnt->mnt_sb->s_root, &name);
+       if (!path.dentry)
                goto err_inode;
+       path.mnt = mntget(pipe_mnt);
 
-       dentry->d_op = &pipefs_dentry_operations;
-       /*
-        * We dont want to publish this dentry into global dentry hash table.
-        * We pretend dentry is already hashed, by unsetting DCACHE_UNHASHED
-        * This permits a working /proc/$pid/fd/XXX on pipes
-        */
-       dentry->d_flags &= ~DCACHE_UNHASHED;
-       d_instantiate(dentry, inode);
+       path.dentry->d_op = &pipefs_dentry_operations;
+       d_instantiate(path.dentry, inode);
 
        err = -ENFILE;
-       f = alloc_file(pipe_mnt, dentry, FMODE_WRITE, &write_pipefifo_fops);
+       f = alloc_file(&path, FMODE_WRITE, &write_pipefifo_fops);
        if (!f)
                goto err_dentry;
        f->f_mapping = inode->i_mapping;
@@ -990,7 +1018,7 @@ struct file *create_write_pipe(int flags)
 
  err_dentry:
        free_pipe_info(inode);
-       dput(dentry);
+       path_put(&path);
        return ERR_PTR(err);
 
  err_inode:
@@ -1009,20 +1037,14 @@ void free_write_pipe(struct file *f)
 
 struct file *create_read_pipe(struct file *wrf, int flags)
 {
-       struct file *f = get_empty_filp();
+       /* Grab pipe from the writer */
+       struct file *f = alloc_file(&wrf->f_path, FMODE_READ,
+                                   &read_pipefifo_fops);
        if (!f)
                return ERR_PTR(-ENFILE);
 
-       /* Grab pipe from the writer */
-       f->f_path = wrf->f_path;
        path_get(&wrf->f_path);
-       f->f_mapping = wrf->f_path.dentry->d_inode->i_mapping;
-
-       f->f_pos = 0;
        f->f_flags = O_RDONLY | (flags & O_NONBLOCK);
-       f->f_op = &read_pipefifo_fops;
-       f->f_mode = FMODE_READ;
-       f->f_version = 0;
 
        return f;
 }
@@ -1098,6 +1120,126 @@ SYSCALL_DEFINE1(pipe, int __user *, fildes)
 }
 
 /*
+ * Allocate a new array of pipe buffers and copy the info over. Returns the
+ * pipe size if successful, or return -ERROR on error.
+ */
+static long pipe_set_size(struct pipe_inode_info *pipe, unsigned long nr_pages)
+{
+       struct pipe_buffer *bufs;
+
+       /*
+        * We can shrink the pipe, if arg >= pipe->nrbufs. Since we don't
+        * expect a lot of shrink+grow operations, just free and allocate
+        * again like we would do for growing. If the pipe currently
+        * contains more buffers than arg, then return busy.
+        */
+       if (nr_pages < pipe->nrbufs)
+               return -EBUSY;
+
+       bufs = kcalloc(nr_pages, sizeof(struct pipe_buffer), GFP_KERNEL);
+       if (unlikely(!bufs))
+               return -ENOMEM;
+
+       /*
+        * The pipe array wraps around, so just start the new one at zero
+        * and adjust the indexes.
+        */
+       if (pipe->nrbufs) {
+               unsigned int tail;
+               unsigned int head;
+
+               tail = pipe->curbuf + pipe->nrbufs;
+               if (tail < pipe->buffers)
+                       tail = 0;
+               else
+                       tail &= (pipe->buffers - 1);
+
+               head = pipe->nrbufs - tail;
+               if (head)
+                       memcpy(bufs, pipe->bufs + pipe->curbuf, head * sizeof(struct pipe_buffer));
+               if (tail)
+                       memcpy(bufs + head, pipe->bufs, tail * sizeof(struct pipe_buffer));
+       }
+
+       pipe->curbuf = 0;
+       kfree(pipe->bufs);
+       pipe->bufs = bufs;
+       pipe->buffers = nr_pages;
+       return nr_pages * PAGE_SIZE;
+}
+
+/*
+ * Currently we rely on the pipe array holding a power-of-2 number
+ * of pages.
+ */
+static inline unsigned int round_pipe_size(unsigned int size)
+{
+       unsigned long nr_pages;
+
+       nr_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+       return roundup_pow_of_two(nr_pages) << PAGE_SHIFT;
+}
+
+/*
+ * This should work even if CONFIG_PROC_FS isn't set, as proc_dointvec_minmax
+ * will return an error.
+ */
+int pipe_proc_fn(struct ctl_table *table, int write, void __user *buf,
+                size_t *lenp, loff_t *ppos)
+{
+       int ret;
+
+       ret = proc_dointvec_minmax(table, write, buf, lenp, ppos);
+       if (ret < 0 || !write)
+               return ret;
+
+       pipe_max_size = round_pipe_size(pipe_max_size);
+       return ret;
+}
+
+long pipe_fcntl(struct file *file, unsigned int cmd, unsigned long arg)
+{
+       struct pipe_inode_info *pipe;
+       long ret;
+
+       pipe = file->f_path.dentry->d_inode->i_pipe;
+       if (!pipe)
+               return -EBADF;
+
+       mutex_lock(&pipe->inode->i_mutex);
+
+       switch (cmd) {
+       case F_SETPIPE_SZ: {
+               unsigned int size, nr_pages;
+
+               size = round_pipe_size(arg);
+               nr_pages = size >> PAGE_SHIFT;
+
+               ret = -EINVAL;
+               if (!nr_pages)
+                       goto out;
+
+               if (!capable(CAP_SYS_RESOURCE) && size > pipe_max_size) {
+                       ret = -EPERM;
+                       goto out;
+               }
+               ret = pipe_set_size(pipe, nr_pages);
+               break;
+               }
+       case F_GETPIPE_SZ:
+               ret = pipe->buffers * PAGE_SIZE;
+               break;
+       default:
+               ret = -EINVAL;
+               break;
+       }
+
+out:
+       mutex_unlock(&pipe->inode->i_mutex);
+       return ret;
+}
+
+/*
  * pipefs should _never_ be mounted by userland - too much of security hassle,
  * no real gain from having the whole whorehouse mounted. So we don't need
  * any operations on the root directory. However, we need a non-trivial