Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi...
Linus Torvalds [Sun, 30 May 2010 16:16:14 +0000 (09:16 -0700)]
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
  mm: export generic_pipe_buf_*() to modules
  fuse: support splice() reading from fuse device
  fuse: allow splice to move pages
  mm: export remove_from_page_cache() to modules
  mm: export lru_cache_add_*() to modules
  fuse: support splice() writing to fuse device
  fuse: get page reference for readpages
  fuse: use get_user_pages_fast()
  fuse: remove unneeded variable

fs/fuse/dev.c
fs/fuse/file.c
fs/fuse/fuse_i.h
fs/pipe.c
include/linux/fuse.h
mm/filemap.c
mm/swap.c

index e53df5e..9424796 100644 (file)
@@ -16,6 +16,9 @@
 #include <linux/pagemap.h>
 #include <linux/file.h>
 #include <linux/slab.h>
+#include <linux/pipe_fs_i.h>
+#include <linux/swap.h>
+#include <linux/splice.h>
 
 MODULE_ALIAS_MISCDEV(FUSE_MINOR);
 MODULE_ALIAS("devname:fuse");
@@ -499,6 +502,9 @@ struct fuse_copy_state {
        int write;
        struct fuse_req *req;
        const struct iovec *iov;
+       struct pipe_buffer *pipebufs;
+       struct pipe_buffer *currbuf;
+       struct pipe_inode_info *pipe;
        unsigned long nr_segs;
        unsigned long seglen;
        unsigned long addr;
@@ -506,16 +512,16 @@ struct fuse_copy_state {
        void *mapaddr;
        void *buf;
        unsigned len;
+       unsigned move_pages:1;
 };
 
 static void fuse_copy_init(struct fuse_copy_state *cs, struct fuse_conn *fc,
-                          int write, struct fuse_req *req,
+                          int write,
                           const struct iovec *iov, unsigned long nr_segs)
 {
        memset(cs, 0, sizeof(*cs));
        cs->fc = fc;
        cs->write = write;
-       cs->req = req;
        cs->iov = iov;
        cs->nr_segs = nr_segs;
 }
@@ -523,7 +529,18 @@ static void fuse_copy_init(struct fuse_copy_state *cs, struct fuse_conn *fc,
 /* Unmap and put previous page of userspace buffer */
 static void fuse_copy_finish(struct fuse_copy_state *cs)
 {
-       if (cs->mapaddr) {
+       if (cs->currbuf) {
+               struct pipe_buffer *buf = cs->currbuf;
+
+               if (!cs->write) {
+                       buf->ops->unmap(cs->pipe, buf, cs->mapaddr);
+               } else {
+                       kunmap_atomic(cs->mapaddr, KM_USER0);
+                       buf->len = PAGE_SIZE - cs->len;
+               }
+               cs->currbuf = NULL;
+               cs->mapaddr = NULL;
+       } else if (cs->mapaddr) {
                kunmap_atomic(cs->mapaddr, KM_USER0);
                if (cs->write) {
                        flush_dcache_page(cs->pg);
@@ -545,26 +562,61 @@ static int fuse_copy_fill(struct fuse_copy_state *cs)
 
        unlock_request(cs->fc, cs->req);
        fuse_copy_finish(cs);
-       if (!cs->seglen) {
-               BUG_ON(!cs->nr_segs);
-               cs->seglen = cs->iov[0].iov_len;
-               cs->addr = (unsigned long) cs->iov[0].iov_base;
-               cs->iov++;
-               cs->nr_segs--;
+       if (cs->pipebufs) {
+               struct pipe_buffer *buf = cs->pipebufs;
+
+               if (!cs->write) {
+                       err = buf->ops->confirm(cs->pipe, buf);
+                       if (err)
+                               return err;
+
+                       BUG_ON(!cs->nr_segs);
+                       cs->currbuf = buf;
+                       cs->mapaddr = buf->ops->map(cs->pipe, buf, 1);
+                       cs->len = buf->len;
+                       cs->buf = cs->mapaddr + buf->offset;
+                       cs->pipebufs++;
+                       cs->nr_segs--;
+               } else {
+                       struct page *page;
+
+                       if (cs->nr_segs == cs->pipe->buffers)
+                               return -EIO;
+
+                       page = alloc_page(GFP_HIGHUSER);
+                       if (!page)
+                               return -ENOMEM;
+
+                       buf->page = page;
+                       buf->offset = 0;
+                       buf->len = 0;
+
+                       cs->currbuf = buf;
+                       cs->mapaddr = kmap_atomic(page, KM_USER0);
+                       cs->buf = cs->mapaddr;
+                       cs->len = PAGE_SIZE;
+                       cs->pipebufs++;
+                       cs->nr_segs++;
+               }
+       } else {
+               if (!cs->seglen) {
+                       BUG_ON(!cs->nr_segs);
+                       cs->seglen = cs->iov[0].iov_len;
+                       cs->addr = (unsigned long) cs->iov[0].iov_base;
+                       cs->iov++;
+                       cs->nr_segs--;
+               }
+               err = get_user_pages_fast(cs->addr, 1, cs->write, &cs->pg);
+               if (err < 0)
+                       return err;
+               BUG_ON(err != 1);
+               offset = cs->addr % PAGE_SIZE;
+               cs->mapaddr = kmap_atomic(cs->pg, KM_USER0);
+               cs->buf = cs->mapaddr + offset;
+               cs->len = min(PAGE_SIZE - offset, cs->seglen);
+               cs->seglen -= cs->len;
+               cs->addr += cs->len;
        }
-       down_read(&current->mm->mmap_sem);
-       err = get_user_pages(current, current->mm, cs->addr, 1, cs->write, 0,
-                            &cs->pg, NULL);
-       up_read(&current->mm->mmap_sem);
-       if (err < 0)
-               return err;
-       BUG_ON(err != 1);
-       offset = cs->addr % PAGE_SIZE;
-       cs->mapaddr = kmap_atomic(cs->pg, KM_USER0);
-       cs->buf = cs->mapaddr + offset;
-       cs->len = min(PAGE_SIZE - offset, cs->seglen);
-       cs->seglen -= cs->len;
-       cs->addr += cs->len;
 
        return lock_request(cs->fc, cs->req);
 }
@@ -586,23 +638,178 @@ static int fuse_copy_do(struct fuse_copy_state *cs, void **val, unsigned *size)
        return ncpy;
 }
 
+static int fuse_check_page(struct page *page)
+{
+       if (page_mapcount(page) ||
+           page->mapping != NULL ||
+           page_count(page) != 1 ||
+           (page->flags & PAGE_FLAGS_CHECK_AT_PREP &
+            ~(1 << PG_locked |
+              1 << PG_referenced |
+              1 << PG_uptodate |
+              1 << PG_lru |
+              1 << PG_active |
+              1 << PG_reclaim))) {
+               printk(KERN_WARNING "fuse: trying to steal weird page\n");
+               printk(KERN_WARNING "  page=%p index=%li flags=%08lx, count=%i, mapcount=%i, mapping=%p\n", page, page->index, page->flags, page_count(page), page_mapcount(page), page->mapping);
+               return 1;
+       }
+       return 0;
+}
+
+static int fuse_try_move_page(struct fuse_copy_state *cs, struct page **pagep)
+{
+       int err;
+       struct page *oldpage = *pagep;
+       struct page *newpage;
+       struct pipe_buffer *buf = cs->pipebufs;
+       struct address_space *mapping;
+       pgoff_t index;
+
+       unlock_request(cs->fc, cs->req);
+       fuse_copy_finish(cs);
+
+       err = buf->ops->confirm(cs->pipe, buf);
+       if (err)
+               return err;
+
+       BUG_ON(!cs->nr_segs);
+       cs->currbuf = buf;
+       cs->len = buf->len;
+       cs->pipebufs++;
+       cs->nr_segs--;
+
+       if (cs->len != PAGE_SIZE)
+               goto out_fallback;
+
+       if (buf->ops->steal(cs->pipe, buf) != 0)
+               goto out_fallback;
+
+       newpage = buf->page;
+
+       if (WARN_ON(!PageUptodate(newpage)))
+               return -EIO;
+
+       ClearPageMappedToDisk(newpage);
+
+       if (fuse_check_page(newpage) != 0)
+               goto out_fallback_unlock;
+
+       mapping = oldpage->mapping;
+       index = oldpage->index;
+
+       /*
+        * This is a new and locked page, it shouldn't be mapped or
+        * have any special flags on it
+        */
+       if (WARN_ON(page_mapped(oldpage)))
+               goto out_fallback_unlock;
+       if (WARN_ON(page_has_private(oldpage)))
+               goto out_fallback_unlock;
+       if (WARN_ON(PageDirty(oldpage) || PageWriteback(oldpage)))
+               goto out_fallback_unlock;
+       if (WARN_ON(PageMlocked(oldpage)))
+               goto out_fallback_unlock;
+
+       remove_from_page_cache(oldpage);
+       page_cache_release(oldpage);
+
+       err = add_to_page_cache_locked(newpage, mapping, index, GFP_KERNEL);
+       if (err) {
+               printk(KERN_WARNING "fuse_try_move_page: failed to add page");
+               goto out_fallback_unlock;
+       }
+       page_cache_get(newpage);
+
+       if (!(buf->flags & PIPE_BUF_FLAG_LRU))
+               lru_cache_add_file(newpage);
+
+       err = 0;
+       spin_lock(&cs->fc->lock);
+       if (cs->req->aborted)
+               err = -ENOENT;
+       else
+               *pagep = newpage;
+       spin_unlock(&cs->fc->lock);
+
+       if (err) {
+               unlock_page(newpage);
+               page_cache_release(newpage);
+               return err;
+       }
+
+       unlock_page(oldpage);
+       page_cache_release(oldpage);
+       cs->len = 0;
+
+       return 0;
+
+out_fallback_unlock:
+       unlock_page(newpage);
+out_fallback:
+       cs->mapaddr = buf->ops->map(cs->pipe, buf, 1);
+       cs->buf = cs->mapaddr + buf->offset;
+
+       err = lock_request(cs->fc, cs->req);
+       if (err)
+               return err;
+
+       return 1;
+}
+
+static int fuse_ref_page(struct fuse_copy_state *cs, struct page *page,
+                        unsigned offset, unsigned count)
+{
+       struct pipe_buffer *buf;
+
+       if (cs->nr_segs == cs->pipe->buffers)
+               return -EIO;
+
+       unlock_request(cs->fc, cs->req);
+       fuse_copy_finish(cs);
+
+       buf = cs->pipebufs;
+       page_cache_get(page);
+       buf->page = page;
+       buf->offset = offset;
+       buf->len = count;
+
+       cs->pipebufs++;
+       cs->nr_segs++;
+       cs->len = 0;
+
+       return 0;
+}
+
 /*
  * Copy a page in the request to/from the userspace buffer.  Must be
  * done atomically
  */
-static int fuse_copy_page(struct fuse_copy_state *cs, struct page *page,
+static int fuse_copy_page(struct fuse_copy_state *cs, struct page **pagep,
                          unsigned offset, unsigned count, int zeroing)
 {
+       int err;
+       struct page *page = *pagep;
+
        if (page && zeroing && count < PAGE_SIZE) {
                void *mapaddr = kmap_atomic(page, KM_USER1);
                memset(mapaddr, 0, PAGE_SIZE);
                kunmap_atomic(mapaddr, KM_USER1);
        }
        while (count) {
-               if (!cs->len) {
-                       int err = fuse_copy_fill(cs);
-                       if (err)
-                               return err;
+               if (cs->write && cs->pipebufs && page) {
+                       return fuse_ref_page(cs, page, offset, count);
+               } else if (!cs->len) {
+                       if (cs->move_pages && page &&
+                           offset == 0 && count == PAGE_SIZE) {
+                               err = fuse_try_move_page(cs, pagep);
+                               if (err <= 0)
+                                       return err;
+                       } else {
+                               err = fuse_copy_fill(cs);
+                               if (err)
+                                       return err;
+                       }
                }
                if (page) {
                        void *mapaddr = kmap_atomic(page, KM_USER1);
@@ -627,8 +834,10 @@ static int fuse_copy_pages(struct fuse_copy_state *cs, unsigned nbytes,
        unsigned count = min(nbytes, (unsigned) PAGE_SIZE - offset);
 
        for (i = 0; i < req->num_pages && (nbytes || zeroing); i++) {
-               struct page *page = req->pages[i];
-               int err = fuse_copy_page(cs, page, offset, count, zeroing);
+               int err;
+
+               err = fuse_copy_page(cs, &req->pages[i], offset, count,
+                                    zeroing);
                if (err)
                        return err;
 
@@ -705,11 +914,10 @@ __acquires(&fc->lock)
  *
  * Called with fc->lock held, releases it
  */
-static int fuse_read_interrupt(struct fuse_conn *fc, struct fuse_req *req,
-                              const struct iovec *iov, unsigned long nr_segs)
+static int fuse_read_interrupt(struct fuse_conn *fc, struct fuse_copy_state *cs,
+                              size_t nbytes, struct fuse_req *req)
 __releases(&fc->lock)
 {
-       struct fuse_copy_state cs;
        struct fuse_in_header ih;
        struct fuse_interrupt_in arg;
        unsigned reqsize = sizeof(ih) + sizeof(arg);
@@ -725,14 +933,13 @@ __releases(&fc->lock)
        arg.unique = req->in.h.unique;
 
        spin_unlock(&fc->lock);
-       if (iov_length(iov, nr_segs) < reqsize)
+       if (nbytes < reqsize)
                return -EINVAL;
 
-       fuse_copy_init(&cs, fc, 1, NULL, iov, nr_segs);
-       err = fuse_copy_one(&cs, &ih, sizeof(ih));
+       err = fuse_copy_one(cs, &ih, sizeof(ih));
        if (!err)
-               err = fuse_copy_one(&cs, &arg, sizeof(arg));
-       fuse_copy_finish(&cs);
+               err = fuse_copy_one(cs, &arg, sizeof(arg));
+       fuse_copy_finish(cs);
 
        return err ? err : reqsize;
 }
@@ -746,18 +953,13 @@ __releases(&fc->lock)
  * request_end().  Otherwise add it to the processing list, and set
  * the 'sent' flag.
  */
-static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
-                             unsigned long nr_segs, loff_t pos)
+static ssize_t fuse_dev_do_read(struct fuse_conn *fc, struct file *file,
+                               struct fuse_copy_state *cs, size_t nbytes)
 {
        int err;
        struct fuse_req *req;
        struct fuse_in *in;
-       struct fuse_copy_state cs;
        unsigned reqsize;
-       struct file *file = iocb->ki_filp;
-       struct fuse_conn *fc = fuse_get_conn(file);
-       if (!fc)
-               return -EPERM;
 
  restart:
        spin_lock(&fc->lock);
@@ -777,7 +979,7 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
        if (!list_empty(&fc->interrupts)) {
                req = list_entry(fc->interrupts.next, struct fuse_req,
                                 intr_entry);
-               return fuse_read_interrupt(fc, req, iov, nr_segs);
+               return fuse_read_interrupt(fc, cs, nbytes, req);
        }
 
        req = list_entry(fc->pending.next, struct fuse_req, list);
@@ -787,7 +989,7 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
        in = &req->in;
        reqsize = in->h.len;
        /* If request is too large, reply with an error and restart the read */
-       if (iov_length(iov, nr_segs) < reqsize) {
+       if (nbytes < reqsize) {
                req->out.h.error = -EIO;
                /* SETXATTR is special, since it may contain too large data */
                if (in->h.opcode == FUSE_SETXATTR)
@@ -796,12 +998,12 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
                goto restart;
        }
        spin_unlock(&fc->lock);
-       fuse_copy_init(&cs, fc, 1, req, iov, nr_segs);
-       err = fuse_copy_one(&cs, &in->h, sizeof(in->h));
+       cs->req = req;
+       err = fuse_copy_one(cs, &in->h, sizeof(in->h));
        if (!err)
-               err = fuse_copy_args(&cs, in->numargs, in->argpages,
+               err = fuse_copy_args(cs, in->numargs, in->argpages,
                                     (struct fuse_arg *) in->args, 0);
-       fuse_copy_finish(&cs);
+       fuse_copy_finish(cs);
        spin_lock(&fc->lock);
        req->locked = 0;
        if (req->aborted) {
@@ -829,6 +1031,110 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
        return err;
 }
 
+static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
+                             unsigned long nr_segs, loff_t pos)
+{
+       struct fuse_copy_state cs;
+       struct file *file = iocb->ki_filp;
+       struct fuse_conn *fc = fuse_get_conn(file);
+       if (!fc)
+               return -EPERM;
+
+       fuse_copy_init(&cs, fc, 1, iov, nr_segs);
+
+       return fuse_dev_do_read(fc, file, &cs, iov_length(iov, nr_segs));
+}
+
+static int fuse_dev_pipe_buf_steal(struct pipe_inode_info *pipe,
+                                  struct pipe_buffer *buf)
+{
+       return 1;
+}
+
+static const struct pipe_buf_operations fuse_dev_pipe_buf_ops = {
+       .can_merge = 0,
+       .map = generic_pipe_buf_map,
+       .unmap = generic_pipe_buf_unmap,
+       .confirm = generic_pipe_buf_confirm,
+       .release = generic_pipe_buf_release,
+       .steal = fuse_dev_pipe_buf_steal,
+       .get = generic_pipe_buf_get,
+};
+
+static ssize_t fuse_dev_splice_read(struct file *in, loff_t *ppos,
+                                   struct pipe_inode_info *pipe,
+                                   size_t len, unsigned int flags)
+{
+       int ret;
+       int page_nr = 0;
+       int do_wakeup = 0;
+       struct pipe_buffer *bufs;
+       struct fuse_copy_state cs;
+       struct fuse_conn *fc = fuse_get_conn(in);
+       if (!fc)
+               return -EPERM;
+
+       bufs = kmalloc(pipe->buffers * sizeof (struct pipe_buffer), GFP_KERNEL);
+       if (!bufs)
+               return -ENOMEM;
+
+       fuse_copy_init(&cs, fc, 1, NULL, 0);
+       cs.pipebufs = bufs;
+       cs.pipe = pipe;
+       ret = fuse_dev_do_read(fc, in, &cs, len);
+       if (ret < 0)
+               goto out;
+
+       ret = 0;
+       pipe_lock(pipe);
+
+       if (!pipe->readers) {
+               send_sig(SIGPIPE, current, 0);
+               if (!ret)
+                       ret = -EPIPE;
+               goto out_unlock;
+       }
+
+       if (pipe->nrbufs + cs.nr_segs > pipe->buffers) {
+               ret = -EIO;
+               goto out_unlock;
+       }
+
+       while (page_nr < cs.nr_segs) {
+               int newbuf = (pipe->curbuf + pipe->nrbufs) & (pipe->buffers - 1);
+               struct pipe_buffer *buf = pipe->bufs + newbuf;
+
+               buf->page = bufs[page_nr].page;
+               buf->offset = bufs[page_nr].offset;
+               buf->len = bufs[page_nr].len;
+               buf->ops = &fuse_dev_pipe_buf_ops;
+
+               pipe->nrbufs++;
+               page_nr++;
+               ret += buf->len;
+
+               if (pipe->inode)
+                       do_wakeup = 1;
+       }
+
+out_unlock:
+       pipe_unlock(pipe);
+
+       if (do_wakeup) {
+               smp_mb();
+               if (waitqueue_active(&pipe->wait))
+                       wake_up_interruptible(&pipe->wait);
+               kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
+       }
+
+out:
+       for (; page_nr < cs.nr_segs; page_nr++)
+               page_cache_release(bufs[page_nr].page);
+
+       kfree(bufs);
+       return ret;
+}
+
 static int fuse_notify_poll(struct fuse_conn *fc, unsigned int size,
                            struct fuse_copy_state *cs)
 {
@@ -988,23 +1294,17 @@ static int copy_out_args(struct fuse_copy_state *cs, struct fuse_out *out,
  * it from the list and copy the rest of the buffer to the request.
  * The request is finished by calling request_end()
  */
-static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
-                              unsigned long nr_segs, loff_t pos)
+static ssize_t fuse_dev_do_write(struct fuse_conn *fc,
+                                struct fuse_copy_state *cs, size_t nbytes)
 {
        int err;
-       size_t nbytes = iov_length(iov, nr_segs);
        struct fuse_req *req;
        struct fuse_out_header oh;
-       struct fuse_copy_state cs;
-       struct fuse_conn *fc = fuse_get_conn(iocb->ki_filp);
-       if (!fc)
-               return -EPERM;
 
-       fuse_copy_init(&cs, fc, 0, NULL, iov, nr_segs);
        if (nbytes < sizeof(struct fuse_out_header))
                return -EINVAL;
 
-       err = fuse_copy_one(&cs, &oh, sizeof(oh));
+       err = fuse_copy_one(cs, &oh, sizeof(oh));
        if (err)
                goto err_finish;
 
@@ -1017,7 +1317,7 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
         * and error contains notification code.
         */
        if (!oh.unique) {
-               err = fuse_notify(fc, oh.error, nbytes - sizeof(oh), &cs);
+               err = fuse_notify(fc, oh.error, nbytes - sizeof(oh), cs);
                return err ? err : nbytes;
        }
 
@@ -1036,7 +1336,7 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 
        if (req->aborted) {
                spin_unlock(&fc->lock);
-               fuse_copy_finish(&cs);
+               fuse_copy_finish(cs);
                spin_lock(&fc->lock);
                request_end(fc, req);
                return -ENOENT;
@@ -1053,7 +1353,7 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
                        queue_interrupt(fc, req);
 
                spin_unlock(&fc->lock);
-               fuse_copy_finish(&cs);
+               fuse_copy_finish(cs);
                return nbytes;
        }
 
@@ -1061,11 +1361,13 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
        list_move(&req->list, &fc->io);
        req->out.h = oh;
        req->locked = 1;
-       cs.req = req;
+       cs->req = req;
+       if (!req->out.page_replace)
+               cs->move_pages = 0;
        spin_unlock(&fc->lock);
 
-       err = copy_out_args(&cs, &req->out, nbytes);
-       fuse_copy_finish(&cs);
+       err = copy_out_args(cs, &req->out, nbytes);
+       fuse_copy_finish(cs);
 
        spin_lock(&fc->lock);
        req->locked = 0;
@@ -1081,10 +1383,101 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
  err_unlock:
        spin_unlock(&fc->lock);
  err_finish:
-       fuse_copy_finish(&cs);
+       fuse_copy_finish(cs);
        return err;
 }
 
+static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
+                             unsigned long nr_segs, loff_t pos)
+{
+       struct fuse_copy_state cs;
+       struct fuse_conn *fc = fuse_get_conn(iocb->ki_filp);
+       if (!fc)
+               return -EPERM;
+
+       fuse_copy_init(&cs, fc, 0, iov, nr_segs);
+
+       return fuse_dev_do_write(fc, &cs, iov_length(iov, nr_segs));
+}
+
+static ssize_t fuse_dev_splice_write(struct pipe_inode_info *pipe,
+                                    struct file *out, loff_t *ppos,
+                                    size_t len, unsigned int flags)
+{
+       unsigned nbuf;
+       unsigned idx;
+       struct pipe_buffer *bufs;
+       struct fuse_copy_state cs;
+       struct fuse_conn *fc;
+       size_t rem;
+       ssize_t ret;
+
+       fc = fuse_get_conn(out);
+       if (!fc)
+               return -EPERM;
+
+       bufs = kmalloc(pipe->buffers * sizeof (struct pipe_buffer), GFP_KERNEL);
+       if (!bufs)
+               return -ENOMEM;
+
+       pipe_lock(pipe);
+       nbuf = 0;
+       rem = 0;
+       for (idx = 0; idx < pipe->nrbufs && rem < len; idx++)
+               rem += pipe->bufs[(pipe->curbuf + idx) & (pipe->buffers - 1)].len;
+
+       ret = -EINVAL;
+       if (rem < len) {
+               pipe_unlock(pipe);
+               goto out;
+       }
+
+       rem = len;
+       while (rem) {
+               struct pipe_buffer *ibuf;
+               struct pipe_buffer *obuf;
+
+               BUG_ON(nbuf >= pipe->buffers);
+               BUG_ON(!pipe->nrbufs);
+               ibuf = &pipe->bufs[pipe->curbuf];
+               obuf = &bufs[nbuf];
+
+               if (rem >= ibuf->len) {
+                       *obuf = *ibuf;
+                       ibuf->ops = NULL;
+                       pipe->curbuf = (pipe->curbuf + 1) & (pipe->buffers - 1);
+                       pipe->nrbufs--;
+               } else {
+                       ibuf->ops->get(pipe, ibuf);
+                       *obuf = *ibuf;
+                       obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
+                       obuf->len = rem;
+                       ibuf->offset += obuf->len;
+                       ibuf->len -= obuf->len;
+               }
+               nbuf++;
+               rem -= obuf->len;
+       }
+       pipe_unlock(pipe);
+
+       fuse_copy_init(&cs, fc, 0, NULL, nbuf);
+       cs.pipebufs = bufs;
+       cs.pipe = pipe;
+
+       if (flags & SPLICE_F_MOVE)
+               cs.move_pages = 1;
+
+       ret = fuse_dev_do_write(fc, &cs, len);
+
+       for (idx = 0; idx < nbuf; idx++) {
+               struct pipe_buffer *buf = &bufs[idx];
+               buf->ops->release(pipe, buf);
+       }
+out:
+       kfree(bufs);
+       return ret;
+}
+
 static unsigned fuse_dev_poll(struct file *file, poll_table *wait)
 {
        unsigned mask = POLLOUT | POLLWRNORM;
@@ -1226,8 +1619,10 @@ const struct file_operations fuse_dev_operations = {
        .llseek         = no_llseek,
        .read           = do_sync_read,
        .aio_read       = fuse_dev_read,
+       .splice_read    = fuse_dev_splice_read,
        .write          = do_sync_write,
        .aio_write      = fuse_dev_write,
+       .splice_write   = fuse_dev_splice_write,
        .poll           = fuse_dev_poll,
        .release        = fuse_dev_release,
        .fasync         = fuse_dev_fasync,
index b5fd6f9..ada0ade 100644 (file)
@@ -516,17 +516,26 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
        int i;
        size_t count = req->misc.read.in.size;
        size_t num_read = req->out.args[0].size;
-       struct inode *inode = req->pages[0]->mapping->host;
+       struct address_space *mapping = NULL;
 
-       /*
-        * Short read means EOF.  If file size is larger, truncate it
-        */
-       if (!req->out.h.error && num_read < count) {
-               loff_t pos = page_offset(req->pages[0]) + num_read;
-               fuse_read_update_size(inode, pos, req->misc.read.attr_ver);
-       }
+       for (i = 0; mapping == NULL && i < req->num_pages; i++)
+               mapping = req->pages[i]->mapping;
 
-       fuse_invalidate_attr(inode); /* atime changed */
+       if (mapping) {
+               struct inode *inode = mapping->host;
+
+               /*
+                * Short read means EOF. If file size is larger, truncate it
+                */
+               if (!req->out.h.error && num_read < count) {
+                       loff_t pos;
+
+                       pos = page_offset(req->pages[0]) + num_read;
+                       fuse_read_update_size(inode, pos,
+                                             req->misc.read.attr_ver);
+               }
+               fuse_invalidate_attr(inode); /* atime changed */
+       }
 
        for (i = 0; i < req->num_pages; i++) {
                struct page *page = req->pages[i];
@@ -535,6 +544,7 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
                else
                        SetPageError(page);
                unlock_page(page);
+               page_cache_release(page);
        }
        if (req->ff)
                fuse_file_put(req->ff);
@@ -549,6 +559,7 @@ static void fuse_send_readpages(struct fuse_req *req, struct file *file)
 
        req->out.argpages = 1;
        req->out.page_zeroing = 1;
+       req->out.page_replace = 1;
        fuse_read_fill(req, file, pos, count, FUSE_READ);
        req->misc.read.attr_ver = fuse_get_attr_version(fc);
        if (fc->async_read) {
@@ -588,6 +599,7 @@ static int fuse_readpages_fill(void *_data, struct page *page)
                        return PTR_ERR(req);
                }
        }
+       page_cache_get(page);
        req->pages[req->num_pages] = page;
        req->num_pages++;
        return 0;
@@ -993,10 +1005,7 @@ static int fuse_get_user_pages(struct fuse_req *req, const char __user *buf,
        nbytes = min_t(size_t, nbytes, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
        npages = (nbytes + offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
        npages = clamp(npages, 1, FUSE_MAX_PAGES_PER_REQ);
-       down_read(&current->mm->mmap_sem);
-       npages = get_user_pages(current, current->mm, user_addr, npages, !write,
-                               0, req->pages, NULL);
-       up_read(&current->mm->mmap_sem);
+       npages = get_user_pages_fast(user_addr, npages, !write, req->pages);
        if (npages < 0)
                return npages;
 
@@ -1579,9 +1588,9 @@ static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
        while (iov_iter_count(&ii)) {
                struct page *page = pages[page_idx++];
                size_t todo = min_t(size_t, PAGE_SIZE, iov_iter_count(&ii));
-               void *kaddr, *map;
+               void *kaddr;
 
-               kaddr = map = kmap(page);
+               kaddr = kmap(page);
 
                while (todo) {
                        char __user *uaddr = ii.iov->iov_base + ii.iov_offset;
index 2c0d14a..8f309f0 100644 (file)
@@ -177,6 +177,9 @@ struct fuse_out {
        /** Zero partially or not copied pages */
        unsigned page_zeroing:1;
 
+       /** Pages may be replaced with new ones */
+       unsigned page_replace:1;
+
        /** Number or arguments */
        unsigned numargs;
 
index 60da077..db6eaab 100644 (file)
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -230,6 +230,7 @@ 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
@@ -249,6 +250,7 @@ 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
@@ -279,6 +281,7 @@ 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
@@ -294,6 +297,7 @@ 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
@@ -309,6 +313,7 @@ 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
@@ -323,6 +328,7 @@ 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,
index 3e2925a..88e0eb5 100644 (file)
@@ -34,6 +34,9 @@
  * 7.13
  *  - make max number of background requests and congestion threshold
  *    tunables
+ *
+ * 7.14
+ *  - add splice support to fuse device
  */
 
 #ifndef _LINUX_FUSE_H
@@ -65,7 +68,7 @@
 #define FUSE_KERNEL_VERSION 7
 
 /** Minor version number of this interface */
-#define FUSE_KERNEL_MINOR_VERSION 13
+#define FUSE_KERNEL_MINOR_VERSION 14
 
 /** The node ID of the root inode */
 #define FUSE_ROOT_ID 1
index 45a2d18..20e5642 100644 (file)
@@ -151,6 +151,7 @@ void remove_from_page_cache(struct page *page)
        spin_unlock_irq(&mapping->tree_lock);
        mem_cgroup_uncharge_cache_page(page);
 }
+EXPORT_SYMBOL(remove_from_page_cache);
 
 static int sync_page(void *word)
 {
index 7cd60bf..3ce7bc3 100644 (file)
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -224,6 +224,7 @@ void __lru_cache_add(struct page *page, enum lru_list lru)
                ____pagevec_lru_add(pvec, lru);
        put_cpu_var(lru_add_pvecs);
 }
+EXPORT_SYMBOL(__lru_cache_add);
 
 /**
  * lru_cache_add_lru - add a page to a page list