Merge branch 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block

* 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block:
  [PATCH] splice: add ->splice_write support for /dev/null
  [PATCH] splice: rearrange moving to/from pipe helpers
  [PATCH] Add support for the sys_vmsplice syscall
  [PATCH] splice: fix offset problems
  [PATCH] splice: fix min() warning

1 files changed, 282 insertions, 73 deletions

diff --git a/fs/splice.c b/fs/splice.c
index 0559e7577a04..447ebc0a37f3 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -27,15 +27,22 @@
 #include <linux/buffer_head.h>
 #include <linux/module.h>
 #include <linux/syscalls.h>
+#include <linux/uio.h>
+
+struct partial_page {
+        unsigned int offset;
+        unsigned int len;
+};
 
 /*
- * Passed to the actors
+ * Passed to splice_to_pipe
  */
-struct splice_desc {
-        unsigned int len, total_len;    /* current and remaining length */
+struct splice_pipe_desc {
+        struct page **pages;            /* page map */
+        struct partial_page *partial;   /* pages[] may not be contig */
+        int nr_pages;                   /* number of pages in map */
         unsigned int flags;             /* splice flags */
-        struct file *file;              /* file to read/write */
-        loff_t pos;                     /* file position */
+        struct pipe_buf_operations *ops;/* ops associated with output pipe */
 };
 
 /*
@@ -128,6 +135,19 @@ static void page_cache_pipe_buf_unmap(struct pipe_inode_info *info,
         kunmap(buf->page);
 }
 
+static void *user_page_pipe_buf_map(struct file *file,
+                                    struct pipe_inode_info *pipe,
+                                    struct pipe_buffer *buf)
+{
+        return kmap(buf->page);
+}
+
+static void user_page_pipe_buf_unmap(struct pipe_inode_info *pipe,
+                                     struct pipe_buffer *buf)
+{
+        kunmap(buf->page);
+}
+
 static void page_cache_pipe_buf_get(struct pipe_inode_info *info,
                                     struct pipe_buffer *buf)
 {
@@ -143,19 +163,33 @@ static struct pipe_buf_operations page_cache_pipe_buf_ops = {
         .get = page_cache_pipe_buf_get,
 };
 
+static int user_page_pipe_buf_steal(struct pipe_inode_info *pipe,
+                                    struct pipe_buffer *buf)
+{
+        return 1;
+}
+
+static struct pipe_buf_operations user_page_pipe_buf_ops = {
+        .can_merge = 0,
+        .map = user_page_pipe_buf_map,
+        .unmap = user_page_pipe_buf_unmap,
+        .release = page_cache_pipe_buf_release,
+        .steal = user_page_pipe_buf_steal,
+        .get = page_cache_pipe_buf_get,
+};
+
 /*
  * Pipe output worker. This sets up our pipe format with the page cache
  * pipe buffer operations. Otherwise very similar to the regular pipe_writev().
  */
-static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages,
-                            int nr_pages, unsigned long len,
-                            unsigned int offset, unsigned int flags)
+static ssize_t splice_to_pipe(struct pipe_inode_info *pipe,
+                              struct splice_pipe_desc *spd)
 {
-        int ret, do_wakeup, i;
+        int ret, do_wakeup, page_nr;
 
         ret = 0;
         do_wakeup = 0;
-        i = 0;
+        page_nr = 0;
 
         if (pipe->inode)
                 mutex_lock(&pipe->inode->i_mutex);
@@ -171,27 +205,19 @@ static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages,
                 if (pipe->nrbufs < PIPE_BUFFERS) {
                         int newbuf = (pipe->curbuf + pipe->nrbufs) & (PIPE_BUFFERS - 1);
                         struct pipe_buffer *buf = pipe->bufs + newbuf;
-                        struct page *page = pages[i++];
-                        unsigned long this_len;
 
-                        this_len = PAGE_CACHE_SIZE - offset;
-                        if (this_len > len)
-                                this_len = len;
-
-                        buf->page = page;
-                        buf->offset = offset;
-                        buf->len = this_len;
-                        buf->ops = &page_cache_pipe_buf_ops;
+                        buf->page = spd->pages[page_nr];
+                        buf->offset = spd->partial[page_nr].offset;
+                        buf->len = spd->partial[page_nr].len;
+                        buf->ops = spd->ops;
                         pipe->nrbufs++;
+                        page_nr++;
+                        ret += buf->len;
+
                         if (pipe->inode)
                                 do_wakeup = 1;
 
-                        ret += this_len;
-                        len -= this_len;
-                        offset = 0;
-                        if (!--nr_pages)
-                                break;
-                        if (!len)
+                        if (!--spd->nr_pages)
                                 break;
                         if (pipe->nrbufs < PIPE_BUFFERS)
                                 continue;
@@ -199,7 +225,7 @@ static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages,
                         break;
                 }
 
-                if (flags & SPLICE_F_NONBLOCK) {
+                if (spd->flags & SPLICE_F_NONBLOCK) {
                         if (!ret)
                                 ret = -EAGAIN;
                         break;
@@ -234,8 +260,8 @@ static ssize_t move_to_pipe(struct pipe_inode_info *pipe, struct page **pages,
                 kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
         }
 
-        while (i < nr_pages)
-                page_cache_release(pages[i++]);
+        while (page_nr < spd->nr_pages)
+                page_cache_release(spd->pages[page_nr++]);
 
         return ret;
 }
@@ -246,17 +272,24 @@ __generic_file_splice_read(struct file *in, loff_t *ppos,
                            unsigned int flags)
 {
         struct address_space *mapping = in->f_mapping;
-        unsigned int loff, offset, nr_pages;
+        unsigned int loff, nr_pages;
         struct page *pages[PIPE_BUFFERS];
+        struct partial_page partial[PIPE_BUFFERS];
         struct page *page;
         pgoff_t index, end_index;
         loff_t isize;
-        size_t bytes;
-        int i, error;
+        size_t total_len;
+        int error;
+        struct splice_pipe_desc spd = {
+                .pages = pages,
+                .partial = partial,
+                .flags = flags,
+                .ops = &page_cache_pipe_buf_ops,
+        };
 
         index = *ppos >> PAGE_CACHE_SHIFT;
-        loff = offset = *ppos & ~PAGE_CACHE_MASK;
-        nr_pages = (len + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+        loff = *ppos & ~PAGE_CACHE_MASK;
+        nr_pages = (len + loff + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
 
         if (nr_pages > PIPE_BUFFERS)
                 nr_pages = PIPE_BUFFERS;
@@ -266,15 +299,15 @@ __generic_file_splice_read(struct file *in, loff_t *ppos,
          * read-ahead if this is a non-zero offset (we are likely doing small
          * chunk splice and the page is already there) for a single page.
          */
-        if (!offset || nr_pages > 1)
-                do_page_cache_readahead(mapping, in, index, nr_pages);
+        if (!loff || spd.nr_pages > 1)
+                do_page_cache_readahead(mapping, in, index, spd.nr_pages);
 
         /*
          * Now fill in the holes:
          */
         error = 0;
-        bytes = 0;
-        for (i = 0; i < nr_pages; i++, index++) {
+        total_len = 0;
+        for (spd.nr_pages = 0; spd.nr_pages < nr_pages; spd.nr_pages++, index++) {
                 unsigned int this_len;
 
                 if (!len)
@@ -283,7 +316,7 @@ __generic_file_splice_read(struct file *in, loff_t *ppos,
                 /*
                  * this_len is the max we'll use from this page
                  */
-                this_len = min(len, PAGE_CACHE_SIZE - loff);
+                this_len = min_t(unsigned long, len, PAGE_CACHE_SIZE - loff);
 find_page:
                 /*
                  * lookup the page for this index
@@ -367,26 +400,29 @@ readpage:
                          */
                         if (end_index == index) {
                                 loff = PAGE_CACHE_SIZE - (isize & ~PAGE_CACHE_MASK);
-                                if (bytes + loff > isize) {
+                                if (total_len + loff > isize) {
                                         page_cache_release(page);
                                         break;
                                 }
                                 /*
                                  * force quit after adding this page
                                  */
-                                nr_pages = i;
+                                nr_pages = spd.nr_pages;
                                 this_len = min(this_len, loff);
+                                loff = 0;
                         }
                 }
 fill_it:
-                pages[i] = page;
-                bytes += this_len;
+                pages[spd.nr_pages] = page;
+                partial[spd.nr_pages].offset = loff;
+                partial[spd.nr_pages].len = this_len;
                 len -= this_len;
+                total_len += this_len;
                 loff = 0;
         }
 
-        if (i)
-                return move_to_pipe(pipe, pages, i, bytes, offset, flags);
+        if (spd.nr_pages)
+                return splice_to_pipe(pipe, &spd);
 
         return error;
 }
@@ -439,14 +475,13 @@ EXPORT_SYMBOL(generic_file_splice_read);
 
 /*
  * Send 'sd->len' bytes to socket from 'sd->file' at position 'sd->pos'
- * using sendpage().
+ * using sendpage(). Return the number of bytes sent.
  */
 static int pipe_to_sendpage(struct pipe_inode_info *info,
                             struct pipe_buffer *buf, struct splice_desc *sd)
 {
         struct file *file = sd->file;
         loff_t pos = sd->pos;
-        unsigned int offset;
         ssize_t ret;
         void *ptr;
         int more;
@@ -461,16 +496,13 @@ static int pipe_to_sendpage(struct pipe_inode_info *info,
         if (IS_ERR(ptr))
                 return PTR_ERR(ptr);
 
-        offset = pos & ~PAGE_CACHE_MASK;
         more = (sd->flags & SPLICE_F_MORE) || sd->len < sd->total_len;
 
-        ret = file->f_op->sendpage(file, buf->page, offset, sd->len, &pos,more);
+        ret = file->f_op->sendpage(file, buf->page, buf->offset, sd->len,
+                                   &pos, more);
 
         buf->ops->unmap(info, buf);
-        if (ret == sd->len)
-                return 0;
-
-        return -EIO;
+        return ret;
 }
 
 /*
@@ -499,7 +531,7 @@ static int pipe_to_file(struct pipe_inode_info *info, struct pipe_buffer *buf,
         struct file *file = sd->file;
         struct address_space *mapping = file->f_mapping;
         gfp_t gfp_mask = mapping_gfp_mask(mapping);
-        unsigned int offset;
+        unsigned int offset, this_len;
         struct page *page;
         pgoff_t index;
         char *src;
@@ -515,6 +547,10 @@ static int pipe_to_file(struct pipe_inode_info *info, struct pipe_buffer *buf,
         index = sd->pos >> PAGE_CACHE_SHIFT;
         offset = sd->pos & ~PAGE_CACHE_MASK;
 
+        this_len = sd->len;
+        if (this_len + offset > PAGE_CACHE_SIZE)
+                this_len = PAGE_CACHE_SIZE - offset;
+
         /*
          * Reuse buf page, if SPLICE_F_MOVE is set.
          */
@@ -558,7 +594,7 @@ find_page:
                  * the full page.
                  */
                 if (!PageUptodate(page)) {
-                        if (sd->len < PAGE_CACHE_SIZE) {
+                        if (this_len < PAGE_CACHE_SIZE) {
                                 ret = mapping->a_ops->readpage(file, page);
                                 if (unlikely(ret))
                                         goto out;
@@ -582,7 +618,7 @@ find_page:
                 }
         }
 
-        ret = mapping->a_ops->prepare_write(file, page, 0, sd->len);
+        ret = mapping->a_ops->prepare_write(file, page, offset, offset+this_len);
         if (ret == AOP_TRUNCATED_PAGE) {
                 page_cache_release(page);
                 goto find_page;
@@ -592,18 +628,22 @@ find_page:
         if (!(buf->flags & PIPE_BUF_FLAG_STOLEN)) {
                 char *dst = kmap_atomic(page, KM_USER0);
 
-                memcpy(dst + offset, src + buf->offset, sd->len);
+                memcpy(dst + offset, src + buf->offset, this_len);
                 flush_dcache_page(page);
                 kunmap_atomic(dst, KM_USER0);
         }
 
-        ret = mapping->a_ops->commit_write(file, page, 0, sd->len);
+        ret = mapping->a_ops->commit_write(file, page, offset, offset+this_len);
         if (ret == AOP_TRUNCATED_PAGE) {
                 page_cache_release(page);
                 goto find_page;
         } else if (ret)
                 goto out;
 
+        /*
+         * Return the number of bytes written.
+         */
+        ret = this_len;
         mark_page_accessed(page);
         balance_dirty_pages_ratelimited(mapping);
 out:
@@ -616,17 +656,14 @@ out_nomem:
         return ret;
 }
 
-typedef int (splice_actor)(struct pipe_inode_info *, struct pipe_buffer *,
-                           struct splice_desc *);
-
 /*
  * Pipe input worker. Most of this logic works like a regular pipe, the
  * key here is the 'actor' worker passed in that actually moves the data
  * to the wanted destination. See pipe_to_file/pipe_to_sendpage above.
  */
-static ssize_t move_from_pipe(struct pipe_inode_info *pipe, struct file *out,
-                              loff_t *ppos, size_t len, unsigned int flags,
-                              splice_actor *actor)
+ssize_t splice_from_pipe(struct pipe_inode_info *pipe, struct file *out,
+                         loff_t *ppos, size_t len, unsigned int flags,
+                         splice_actor *actor)
 {
         int ret, do_wakeup, err;
         struct splice_desc sd;
@@ -652,16 +689,22 @@ static ssize_t move_from_pipe(struct pipe_inode_info *pipe, struct file *out,
                                 sd.len = sd.total_len;
 
                         err = actor(pipe, buf, &sd);
-                        if (err) {
+                        if (err <= 0) {
                                 if (!ret && err != -ENODATA)
                                         ret = err;
 
                                 break;
                         }
 
-                        ret += sd.len;
-                        buf->offset += sd.len;
-                        buf->len -= sd.len;
+                        ret += err;
+                        buf->offset += err;
+                        buf->len -= err;
+
+                        sd.len -= err;
+                        sd.pos += err;
+                        sd.total_len -= err;
+                        if (sd.len)
+                                continue;
 
                         if (!buf->len) {
                                 buf->ops = NULL;
@@ -672,8 +715,6 @@ static ssize_t move_from_pipe(struct pipe_inode_info *pipe, struct file *out,
                                         do_wakeup = 1;
                         }
 
-                        sd.pos += sd.len;
-                        sd.total_len -= sd.len;
                         if (!sd.total_len)
                                 break;
                 }
@@ -741,7 +782,7 @@ generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out,
         struct address_space *mapping = out->f_mapping;
         ssize_t ret;
 
-        ret = move_from_pipe(pipe, out, ppos, len, flags, pipe_to_file);
+        ret = splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_file);
         if (ret > 0) {
                 struct inode *inode = mapping->host;
 
@@ -783,7 +824,7 @@ EXPORT_SYMBOL(generic_file_splice_write);
 ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe, struct file *out,
                                 loff_t *ppos, size_t len, unsigned int flags)
 {
-        return move_from_pipe(pipe, out, ppos, len, flags, pipe_to_sendpage);
+        return splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_sendpage);
 }
 
 EXPORT_SYMBOL(generic_splice_sendpage);
@@ -870,7 +911,7 @@ long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
 
                 /*
                  * We don't have an immediate reader, but we'll read the stuff
-                 * out of the pipe right after the move_to_pipe(). So set
+                 * out of the pipe right after the splice_to_pipe(). So set
                  * PIPE_READERS appropriately.
                  */
                 pipe->readers = 1;
@@ -1010,6 +1051,174 @@ static long do_splice(struct file *in, loff_t __user *off_in,
         return -EINVAL;
 }
 
+/*
+ * Map an iov into an array of pages and offset/length tupples. With the
+ * partial_page structure, we can map several non-contiguous ranges into
+ * our ones pages[] map instead of splitting that operation into pieces.
+ * Could easily be exported as a generic helper for other users, in which
+ * case one would probably want to add a 'max_nr_pages' parameter as well.
+ */
+static int get_iovec_page_array(const struct iovec __user *iov,
+                                unsigned int nr_vecs, struct page **pages,
+                                struct partial_page *partial)
+{
+        int buffers = 0, error = 0;
+
+        /*
+         * It's ok to take the mmap_sem for reading, even
+         * across a "get_user()".
+         */
+        down_read(&current->mm->mmap_sem);
+
+        while (nr_vecs) {
+                unsigned long off, npages;
+                void __user *base;
+                size_t len;
+                int i;
+
+                /*
+                 * Get user address base and length for this iovec.
+                 */
+                error = get_user(base, &iov->iov_base);
+                if (unlikely(error))
+                        break;
+                error = get_user(len, &iov->iov_len);
+                if (unlikely(error))
+                        break;
+
+                /*
+                 * Sanity check this iovec. 0 read succeeds.
+                 */
+                if (unlikely(!len))
+                        break;
+                error = -EFAULT;
+                if (unlikely(!base))
+                        break;
+
+                /*
+                 * Get this base offset and number of pages, then map
+                 * in the user pages.
+                 */
+                off = (unsigned long) base & ~PAGE_MASK;
+                npages = (off + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
+                if (npages > PIPE_BUFFERS - buffers)
+                        npages = PIPE_BUFFERS - buffers;
+
+                error = get_user_pages(current, current->mm,
+                                       (unsigned long) base, npages, 0, 0,
+                                       &pages[buffers], NULL);
+
+                if (unlikely(error <= 0))
+                        break;
+
+                /*
+                 * Fill this contiguous range into the partial page map.
+                 */
+                for (i = 0; i < error; i++) {
+                        const int plen = min_t(size_t, len, PAGE_SIZE) - off;
+
+                        partial[buffers].offset = off;
+                        partial[buffers].len = plen;
+
+                        off = 0;
+                        len -= plen;
+                        buffers++;
+                }
+
+                /*
+                 * We didn't complete this iov, stop here since it probably
+                 * means we have to move some of this into a pipe to
+                 * be able to continue.
+                 */
+                if (len)
+                        break;
+
+                /*
+                 * Don't continue if we mapped fewer pages than we asked for,
+                 * or if we mapped the max number of pages that we have
+                 * room for.
+                 */
+                if (error < npages || buffers == PIPE_BUFFERS)
+                        break;
+
+                nr_vecs--;
+                iov++;
+        }
+
+        up_read(&current->mm->mmap_sem);
+
+        if (buffers)
+                return buffers;
+
+        return error;
+}
+
+/*
+ * vmsplice splices a user address range into a pipe. It can be thought of
+ * as splice-from-memory, where the regular splice is splice-from-file (or
+ * to file). In both cases the output is a pipe, naturally.
+ *
+ * Note that vmsplice only supports splicing _from_ user memory to a pipe,
+ * not the other way around. Splicing from user memory is a simple operation
+ * that can be supported without any funky alignment restrictions or nasty
+ * vm tricks. We simply map in the user memory and fill them into a pipe.
+ * The reverse isn't quite as easy, though. There are two possible solutions
+ * for that:
+ *
+ *      - memcpy() the data internally, at which point we might as well just
+ *        do a regular read() on the buffer anyway.
+ *      - Lots of nasty vm tricks, that are neither fast nor flexible (it
+ *        has restriction limitations on both ends of the pipe).
+ *
+ * Alas, it isn't here.
+ *
+ */
+static long do_vmsplice(struct file *file, const struct iovec __user *iov,
+                        unsigned long nr_segs, unsigned int flags)
+{
+        struct pipe_inode_info *pipe = file->f_dentry->d_inode->i_pipe;
+        struct page *pages[PIPE_BUFFERS];
+        struct partial_page partial[PIPE_BUFFERS];
+        struct splice_pipe_desc spd = {
+                .pages = pages,
+                .partial = partial,
+                .flags = flags,
+                .ops = &user_page_pipe_buf_ops,
+        };
+
+        if (unlikely(!pipe))
+                return -EBADF;
+        if (unlikely(nr_segs > UIO_MAXIOV))
+                return -EINVAL;
+        else if (unlikely(!nr_segs))
+                return 0;
+
+        spd.nr_pages = get_iovec_page_array(iov, nr_segs, pages, partial);
+        if (spd.nr_pages <= 0)
+                return spd.nr_pages;
+
+        return splice_to_pipe(pipe, &spd);
+}
+
+asmlinkage long sys_vmsplice(int fd, const struct iovec __user *iov,
+                             unsigned long nr_segs, unsigned int flags)
+{
+        struct file *file;
+        long error;
+        int fput;
+
+        error = -EBADF;
+        file = fget_light(fd, &fput);
+        if (file) {
+                if (file->f_mode & FMODE_WRITE)
+                        error = do_vmsplice(file, iov, nr_segs, flags);
+
+                fput_light(file, fput);
+        }
+
+        return error;
+}
+
 asmlinkage long sys_splice(int fd_in, loff_t __user *off_in,
                            int fd_out, loff_t __user *off_out,
                            size_t len, unsigned int flags)