diff options
Diffstat (limited to 'fs/fuse/file.c')
| -rw-r--r-- | fs/fuse/file.c | 1713 |
1 files changed, 1382 insertions, 331 deletions
diff --git a/fs/fuse/file.c b/fs/fuse/file.c index c8224587123..40ac2628ddc 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -13,6 +13,10 @@ #include <linux/kernel.h> #include <linux/sched.h> #include <linux/module.h> +#include <linux/compat.h> +#include <linux/swap.h> +#include <linux/aio.h> +#include <linux/falloc.h> static const struct file_operations fuse_direct_io_file_operations; @@ -23,7 +27,7 @@ static int fuse_send_open(struct fuse_conn *fc, u64 nodeid, struct file *file, struct fuse_req *req; int err; - req = fuse_get_req(fc); + req = fuse_get_req_nopages(fc); if (IS_ERR(req)) return PTR_ERR(req); @@ -55,7 +59,7 @@ struct fuse_file *fuse_file_alloc(struct fuse_conn *fc) return NULL; ff->fc = fc; - ff->reserved_req = fuse_request_alloc(); + ff->reserved_req = fuse_request_alloc(0); if (unlikely(!ff->reserved_req)) { kfree(ff); return NULL; @@ -85,18 +89,62 @@ struct fuse_file *fuse_file_get(struct fuse_file *ff) return ff; } +static void fuse_release_async(struct work_struct *work) +{ + struct fuse_req *req; + struct fuse_conn *fc; + struct path path; + + req = container_of(work, struct fuse_req, misc.release.work); + path = req->misc.release.path; + fc = get_fuse_conn(path.dentry->d_inode); + + fuse_put_request(fc, req); + path_put(&path); +} + static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req) { - path_put(&req->misc.release.path); + if (fc->destroy_req) { + /* + * If this is a fuseblk mount, then it's possible that + * releasing the path will result in releasing the + * super block and sending the DESTROY request. If + * the server is single threaded, this would hang. + * For this reason do the path_put() in a separate + * thread. + */ + atomic_inc(&req->count); + INIT_WORK(&req->misc.release.work, fuse_release_async); + schedule_work(&req->misc.release.work); + } else { + path_put(&req->misc.release.path); + } } -static void fuse_file_put(struct fuse_file *ff) +static void fuse_file_put(struct fuse_file *ff, bool sync) { if (atomic_dec_and_test(&ff->count)) { struct fuse_req *req = ff->reserved_req; - req->end = fuse_release_end; - fuse_request_send_background(ff->fc, req); + if (ff->fc->no_open) { + /* + * Drop the release request when client does not + * implement 'open' + */ + req->background = 0; + path_put(&req->misc.release.path); + fuse_put_request(ff->fc, req); + } else if (sync) { + req->background = 0; + fuse_request_send(ff->fc, req); + path_put(&req->misc.release.path); + fuse_put_request(ff->fc, req); + } else { + req->end = fuse_release_end; + req->background = 1; + fuse_request_send_background(ff->fc, req); + } kfree(ff); } } @@ -104,36 +152,62 @@ static void fuse_file_put(struct fuse_file *ff) int fuse_do_open(struct fuse_conn *fc, u64 nodeid, struct file *file, bool isdir) { - struct fuse_open_out outarg; struct fuse_file *ff; - int err; int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN; ff = fuse_file_alloc(fc); if (!ff) return -ENOMEM; - err = fuse_send_open(fc, nodeid, file, opcode, &outarg); - if (err) { - fuse_file_free(ff); - return err; + ff->fh = 0; + ff->open_flags = FOPEN_KEEP_CACHE; /* Default for no-open */ + if (!fc->no_open || isdir) { + struct fuse_open_out outarg; + int err; + + err = fuse_send_open(fc, nodeid, file, opcode, &outarg); + if (!err) { + ff->fh = outarg.fh; + ff->open_flags = outarg.open_flags; + + } else if (err != -ENOSYS || isdir) { + fuse_file_free(ff); + return err; + } else { + fc->no_open = 1; + } } if (isdir) - outarg.open_flags &= ~FOPEN_DIRECT_IO; + ff->open_flags &= ~FOPEN_DIRECT_IO; - ff->fh = outarg.fh; ff->nodeid = nodeid; - ff->open_flags = outarg.open_flags; file->private_data = fuse_file_get(ff); return 0; } EXPORT_SYMBOL_GPL(fuse_do_open); +static void fuse_link_write_file(struct file *file) +{ + struct inode *inode = file_inode(file); + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_inode *fi = get_fuse_inode(inode); + struct fuse_file *ff = file->private_data; + /* + * file may be written through mmap, so chain it onto the + * inodes's write_file list + */ + spin_lock(&fc->lock); + if (list_empty(&ff->write_entry)) + list_add(&ff->write_entry, &fi->write_files); + spin_unlock(&fc->lock); +} + void fuse_finish_open(struct inode *inode, struct file *file) { struct fuse_file *ff = file->private_data; + struct fuse_conn *fc = get_fuse_conn(inode); if (ff->open_flags & FOPEN_DIRECT_IO) file->f_op = &fuse_direct_io_file_operations; @@ -141,28 +215,45 @@ void fuse_finish_open(struct inode *inode, struct file *file) invalidate_inode_pages2(inode->i_mapping); if (ff->open_flags & FOPEN_NONSEEKABLE) nonseekable_open(inode, file); + if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) { + struct fuse_inode *fi = get_fuse_inode(inode); + + spin_lock(&fc->lock); + fi->attr_version = ++fc->attr_version; + i_size_write(inode, 0); + spin_unlock(&fc->lock); + fuse_invalidate_attr(inode); + if (fc->writeback_cache) + file_update_time(file); + } + if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache) + fuse_link_write_file(file); } int fuse_open_common(struct inode *inode, struct file *file, bool isdir) { struct fuse_conn *fc = get_fuse_conn(inode); int err; - - /* VFS checks this, but only _after_ ->open() */ - if (file->f_flags & O_DIRECT) - return -EINVAL; + bool lock_inode = (file->f_flags & O_TRUNC) && + fc->atomic_o_trunc && + fc->writeback_cache; err = generic_file_open(inode, file); if (err) return err; + if (lock_inode) + mutex_lock(&inode->i_mutex); + err = fuse_do_open(fc, get_node_id(inode), file, isdir); - if (err) - return err; - fuse_finish_open(inode, file); + if (!err) + fuse_finish_open(inode, file); + + if (lock_inode) + mutex_unlock(&inode->i_mutex); - return 0; + return err; } static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode) @@ -177,7 +268,7 @@ static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode) rb_erase(&ff->polled_node, &fc->polled_files); spin_unlock(&fc->lock); - wake_up_interruptible_sync(&ff->poll_wait); + wake_up_interruptible_all(&ff->poll_wait); inarg->fh = ff->fh; inarg->flags = flags; @@ -200,6 +291,12 @@ void fuse_release_common(struct file *file, int opcode) req = ff->reserved_req; fuse_prepare_release(ff, file->f_flags, opcode); + if (ff->flock) { + struct fuse_release_in *inarg = &req->misc.release.in; + inarg->release_flags |= FUSE_RELEASE_FLOCK_UNLOCK; + inarg->lock_owner = fuse_lock_owner_id(ff->fc, + (fl_owner_t) file); + } /* Hold vfsmount and dentry until release is finished */ path_get(&file->f_path); req->misc.release.path = file->f_path; @@ -208,8 +305,12 @@ void fuse_release_common(struct file *file, int opcode) * Normally this will send the RELEASE request, however if * some asynchronous READ or WRITE requests are outstanding, * the sending will be delayed. + * + * Make the release synchronous if this is a fuseblk mount, + * synchronous RELEASE is allowed (and desirable) in this case + * because the server can be trusted not to screw up. */ - fuse_file_put(ff); + fuse_file_put(ff, ff->fc->destroy_req != NULL); } static int fuse_open(struct inode *inode, struct file *file) @@ -219,6 +320,12 @@ static int fuse_open(struct inode *inode, struct file *file) static int fuse_release(struct inode *inode, struct file *file) { + struct fuse_conn *fc = get_fuse_conn(inode); + + /* see fuse_vma_close() for !writeback_cache case */ + if (fc->writeback_cache) + write_inode_now(inode, 1); + fuse_release_common(file, FUSE_RELEASE); /* return value is ignored by VFS */ @@ -230,6 +337,7 @@ void fuse_sync_release(struct fuse_file *ff, int flags) WARN_ON(atomic_read(&ff->count) > 1); fuse_prepare_release(ff, flags, FUSE_RELEASE); ff->reserved_req->force = 1; + ff->reserved_req->background = 0; fuse_request_send(ff->fc, ff->reserved_req); fuse_put_request(ff->fc, ff->reserved_req); kfree(ff); @@ -259,12 +367,13 @@ u64 fuse_lock_owner_id(struct fuse_conn *fc, fl_owner_t id) } /* - * Check if page is under writeback + * Check if any page in a range is under writeback * * This is currently done by walking the list of writepage requests * for the inode, which can be pretty inefficient. */ -static bool fuse_page_is_writeback(struct inode *inode, pgoff_t index) +static bool fuse_range_is_writeback(struct inode *inode, pgoff_t idx_from, + pgoff_t idx_to) { struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_inode *fi = get_fuse_inode(inode); @@ -277,7 +386,8 @@ static bool fuse_page_is_writeback(struct inode *inode, pgoff_t index) BUG_ON(req->inode != inode); curr_index = req->misc.write.in.offset >> PAGE_CACHE_SHIFT; - if (curr_index == index) { + if (idx_from < curr_index + req->num_pages && + curr_index <= idx_to) { found = true; break; } @@ -287,6 +397,11 @@ static bool fuse_page_is_writeback(struct inode *inode, pgoff_t index) return found; } +static inline bool fuse_page_is_writeback(struct inode *inode, pgoff_t index) +{ + return fuse_range_is_writeback(inode, index, index); +} + /* * Wait for page writeback to be completed. * @@ -301,9 +416,24 @@ static int fuse_wait_on_page_writeback(struct inode *inode, pgoff_t index) return 0; } +/* + * Wait for all pending writepages on the inode to finish. + * + * This is currently done by blocking further writes with FUSE_NOWRITE + * and waiting for all sent writes to complete. + * + * This must be called under i_mutex, otherwise the FUSE_NOWRITE usage + * could conflict with truncation. + */ +static void fuse_sync_writes(struct inode *inode) +{ + fuse_set_nowrite(inode); + fuse_release_nowrite(inode); +} + static int fuse_flush(struct file *file, fl_owner_t id) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_file *ff = file->private_data; struct fuse_req *req; @@ -316,7 +446,15 @@ static int fuse_flush(struct file *file, fl_owner_t id) if (fc->no_flush) return 0; - req = fuse_get_req_nofail(fc, file); + err = write_inode_now(inode, 1); + if (err) + return err; + + mutex_lock(&inode->i_mutex); + fuse_sync_writes(inode); + mutex_unlock(&inode->i_mutex); + + req = fuse_get_req_nofail_nopages(fc, file); memset(&inarg, 0, sizeof(inarg)); inarg.fh = ff->fh; inarg.lock_owner = fuse_lock_owner_id(fc, id); @@ -336,22 +474,8 @@ static int fuse_flush(struct file *file, fl_owner_t id) return err; } -/* - * Wait for all pending writepages on the inode to finish. - * - * This is currently done by blocking further writes with FUSE_NOWRITE - * and waiting for all sent writes to complete. - * - * This must be called under i_mutex, otherwise the FUSE_NOWRITE usage - * could conflict with truncation. - */ -static void fuse_sync_writes(struct inode *inode) -{ - fuse_set_nowrite(inode); - fuse_release_nowrite(inode); -} - -int fuse_fsync_common(struct file *file, int datasync, int isdir) +int fuse_fsync_common(struct file *file, loff_t start, loff_t end, + int datasync, int isdir) { struct inode *inode = file->f_mapping->host; struct fuse_conn *fc = get_fuse_conn(inode); @@ -363,23 +487,30 @@ int fuse_fsync_common(struct file *file, int datasync, int isdir) if (is_bad_inode(inode)) return -EIO; - if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir)) - return 0; + mutex_lock(&inode->i_mutex); /* * Start writeback against all dirty pages of the inode, then * wait for all outstanding writes, before sending the FSYNC * request. */ - err = write_inode_now(inode, 0); + err = filemap_write_and_wait_range(inode->i_mapping, start, end); if (err) - return err; + goto out; fuse_sync_writes(inode); + err = sync_inode_metadata(inode, 1); + if (err) + goto out; - req = fuse_get_req(fc); - if (IS_ERR(req)) - return PTR_ERR(req); + if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir)) + goto out; + + req = fuse_get_req_nopages(fc); + if (IS_ERR(req)) { + err = PTR_ERR(req); + goto out; + } memset(&inarg, 0, sizeof(inarg)); inarg.fh = ff->fh; @@ -399,12 +530,15 @@ int fuse_fsync_common(struct file *file, int datasync, int isdir) fc->no_fsync = 1; err = 0; } +out: + mutex_unlock(&inode->i_mutex); return err; } -static int fuse_fsync(struct file *file, int datasync) +static int fuse_fsync(struct file *file, loff_t start, loff_t end, + int datasync) { - return fuse_fsync_common(file, datasync, 0); + return fuse_fsync_common(file, start, end, datasync, 0); } void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos, @@ -427,9 +561,114 @@ void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos, req->out.args[0].size = count; } -static size_t fuse_send_read(struct fuse_req *req, struct file *file, +static void fuse_release_user_pages(struct fuse_req *req, int write) +{ + unsigned i; + + for (i = 0; i < req->num_pages; i++) { + struct page *page = req->pages[i]; + if (write) + set_page_dirty_lock(page); + put_page(page); + } +} + +/** + * In case of short read, the caller sets 'pos' to the position of + * actual end of fuse request in IO request. Otherwise, if bytes_requested + * == bytes_transferred or rw == WRITE, the caller sets 'pos' to -1. + * + * An example: + * User requested DIO read of 64K. It was splitted into two 32K fuse requests, + * both submitted asynchronously. The first of them was ACKed by userspace as + * fully completed (req->out.args[0].size == 32K) resulting in pos == -1. The + * second request was ACKed as short, e.g. only 1K was read, resulting in + * pos == 33K. + * + * Thus, when all fuse requests are completed, the minimal non-negative 'pos' + * will be equal to the length of the longest contiguous fragment of + * transferred data starting from the beginning of IO request. + */ +static void fuse_aio_complete(struct fuse_io_priv *io, int err, ssize_t pos) +{ + int left; + + spin_lock(&io->lock); + if (err) + io->err = io->err ? : err; + else if (pos >= 0 && (io->bytes < 0 || pos < io->bytes)) + io->bytes = pos; + + left = --io->reqs; + spin_unlock(&io->lock); + + if (!left) { + long res; + + if (io->err) + res = io->err; + else if (io->bytes >= 0 && io->write) + res = -EIO; + else { + res = io->bytes < 0 ? io->size : io->bytes; + + if (!is_sync_kiocb(io->iocb)) { + struct inode *inode = file_inode(io->iocb->ki_filp); + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_inode *fi = get_fuse_inode(inode); + + spin_lock(&fc->lock); + fi->attr_version = ++fc->attr_version; + spin_unlock(&fc->lock); + } + } + + aio_complete(io->iocb, res, 0); + kfree(io); + } +} + +static void fuse_aio_complete_req(struct fuse_conn *fc, struct fuse_req *req) +{ + struct fuse_io_priv *io = req->io; + ssize_t pos = -1; + + fuse_release_user_pages(req, !io->write); + + if (io->write) { + if (req->misc.write.in.size != req->misc.write.out.size) + pos = req->misc.write.in.offset - io->offset + + req->misc.write.out.size; + } else { + if (req->misc.read.in.size != req->out.args[0].size) + pos = req->misc.read.in.offset - io->offset + + req->out.args[0].size; + } + + fuse_aio_complete(io, req->out.h.error, pos); +} + +static size_t fuse_async_req_send(struct fuse_conn *fc, struct fuse_req *req, + size_t num_bytes, struct fuse_io_priv *io) +{ + spin_lock(&io->lock); + io->size += num_bytes; + io->reqs++; + spin_unlock(&io->lock); + + req->io = io; + req->end = fuse_aio_complete_req; + + __fuse_get_request(req); + fuse_request_send_background(fc, req); + + return num_bytes; +} + +static size_t fuse_send_read(struct fuse_req *req, struct fuse_io_priv *io, loff_t pos, size_t count, fl_owner_t owner) { + struct file *file = io->file; struct fuse_file *ff = file->private_data; struct fuse_conn *fc = ff->fc; @@ -440,6 +679,10 @@ static size_t fuse_send_read(struct fuse_req *req, struct file *file, inarg->read_flags |= FUSE_READ_LOCKOWNER; inarg->lock_owner = fuse_lock_owner_id(fc, owner); } + + if (io->async) + return fuse_async_req_send(fc, req, count, io); + fuse_request_send(fc, req); return req->out.args[0].size; } @@ -451,15 +694,43 @@ static void fuse_read_update_size(struct inode *inode, loff_t size, struct fuse_inode *fi = get_fuse_inode(inode); spin_lock(&fc->lock); - if (attr_ver == fi->attr_version && size < inode->i_size) { + if (attr_ver == fi->attr_version && size < inode->i_size && + !test_bit(FUSE_I_SIZE_UNSTABLE, &fi->state)) { fi->attr_version = ++fc->attr_version; i_size_write(inode, size); } spin_unlock(&fc->lock); } -static int fuse_readpage(struct file *file, struct page *page) +static void fuse_short_read(struct fuse_req *req, struct inode *inode, + u64 attr_ver) { + size_t num_read = req->out.args[0].size; + struct fuse_conn *fc = get_fuse_conn(inode); + + if (fc->writeback_cache) { + /* + * A hole in a file. Some data after the hole are in page cache, + * but have not reached the client fs yet. So, the hole is not + * present there. + */ + int i; + int start_idx = num_read >> PAGE_CACHE_SHIFT; + size_t off = num_read & (PAGE_CACHE_SIZE - 1); + + for (i = start_idx; i < req->num_pages; i++) { + zero_user_segment(req->pages[i], off, PAGE_CACHE_SIZE); + off = 0; + } + } else { + loff_t pos = page_offset(req->pages[0]) + num_read; + fuse_read_update_size(inode, pos, attr_ver); + } +} + +static int fuse_do_readpage(struct file *file, struct page *page) +{ + struct fuse_io_priv io = { .async = 0, .file = file }; struct inode *inode = page->mapping->host; struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_req *req; @@ -469,21 +740,16 @@ static int fuse_readpage(struct file *file, struct page *page) u64 attr_ver; int err; - err = -EIO; - if (is_bad_inode(inode)) - goto out; - /* - * Page writeback can extend beyond the liftime of the + * Page writeback can extend beyond the lifetime of the * page-cache page, so make sure we read a properly synced * page. */ fuse_wait_on_page_writeback(inode, page->index); - req = fuse_get_req(fc); - err = PTR_ERR(req); + req = fuse_get_req(fc, 1); if (IS_ERR(req)) - goto out; + return PTR_ERR(req); attr_ver = fuse_get_attr_version(fc); @@ -491,21 +757,36 @@ static int fuse_readpage(struct file *file, struct page *page) req->out.argpages = 1; req->num_pages = 1; req->pages[0] = page; - num_read = fuse_send_read(req, file, pos, count, NULL); + req->page_descs[0].length = count; + num_read = fuse_send_read(req, &io, pos, count, NULL); err = req->out.h.error; - fuse_put_request(fc, req); if (!err) { /* * Short read means EOF. If file size is larger, truncate it */ if (num_read < count) - fuse_read_update_size(inode, pos + num_read, attr_ver); + fuse_short_read(req, inode, attr_ver); SetPageUptodate(page); } - fuse_invalidate_attr(inode); /* atime changed */ + fuse_put_request(fc, req); + + return err; +} + +static int fuse_readpage(struct file *file, struct page *page) +{ + struct inode *inode = page->mapping->host; + int err; + + err = -EIO; + if (is_bad_inode(inode)) + goto out; + + err = fuse_do_readpage(file, page); + fuse_invalidate_atime(inode); out: unlock_page(page); return err; @@ -527,14 +808,10 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req) /* * Short read means EOF. If file size is larger, truncate it */ - if (!req->out.h.error && num_read < count) { - loff_t pos; + if (!req->out.h.error && num_read < count) + fuse_short_read(req, inode, req->misc.read.attr_ver); - 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 */ + fuse_invalidate_atime(inode); } for (i = 0; i < req->num_pages; i++) { @@ -547,7 +824,7 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req) page_cache_release(page); } if (req->ff) - fuse_file_put(req->ff); + fuse_file_put(req->ff, false); } static void fuse_send_readpages(struct fuse_req *req, struct file *file) @@ -577,6 +854,7 @@ struct fuse_fill_data { struct fuse_req *req; struct file *file; struct inode *inode; + unsigned nr_pages; }; static int fuse_readpages_fill(void *_data, struct page *page) @@ -592,16 +870,31 @@ static int fuse_readpages_fill(void *_data, struct page *page) (req->num_pages == FUSE_MAX_PAGES_PER_REQ || (req->num_pages + 1) * PAGE_CACHE_SIZE > fc->max_read || req->pages[req->num_pages - 1]->index + 1 != page->index)) { + int nr_alloc = min_t(unsigned, data->nr_pages, + FUSE_MAX_PAGES_PER_REQ); fuse_send_readpages(req, data->file); - data->req = req = fuse_get_req(fc); + if (fc->async_read) + req = fuse_get_req_for_background(fc, nr_alloc); + else + req = fuse_get_req(fc, nr_alloc); + + data->req = req; if (IS_ERR(req)) { unlock_page(page); return PTR_ERR(req); } } + + if (WARN_ON(req->num_pages >= req->max_pages)) { + fuse_put_request(fc, req); + return -EIO; + } + page_cache_get(page); req->pages[req->num_pages] = page; + req->page_descs[req->num_pages].length = PAGE_SIZE; req->num_pages++; + data->nr_pages--; return 0; } @@ -612,6 +905,7 @@ static int fuse_readpages(struct file *file, struct address_space *mapping, struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_fill_data data; int err; + int nr_alloc = min_t(unsigned, nr_pages, FUSE_MAX_PAGES_PER_REQ); err = -EIO; if (is_bad_inode(inode)) @@ -619,7 +913,11 @@ static int fuse_readpages(struct file *file, struct address_space *mapping, data.file = file; data.inode = inode; - data.req = fuse_get_req(fc); + if (fc->async_read) + data.req = fuse_get_req_for_background(fc, nr_alloc); + else + data.req = fuse_get_req(fc, nr_alloc); + data.nr_pages = nr_pages; err = PTR_ERR(data.req); if (IS_ERR(data.req)) goto out; @@ -635,23 +933,25 @@ out: return err; } -static ssize_t fuse_file_aio_read(struct kiocb *iocb, const struct iovec *iov, - unsigned long nr_segs, loff_t pos) +static ssize_t fuse_file_read_iter(struct kiocb *iocb, struct iov_iter *to) { struct inode *inode = iocb->ki_filp->f_mapping->host; + struct fuse_conn *fc = get_fuse_conn(inode); - if (pos + iov_length(iov, nr_segs) > i_size_read(inode)) { + /* + * In auto invalidate mode, always update attributes on read. + * Otherwise, only update if we attempt to read past EOF (to ensure + * i_size is up to date). + */ + if (fc->auto_inval_data || + (iocb->ki_pos + iov_iter_count(to) > i_size_read(inode))) { int err; - /* - * If trying to read past EOF, make sure the i_size - * attribute is up-to-date. - */ err = fuse_update_attributes(inode, NULL, iocb->ki_filp, NULL); if (err) return err; } - return generic_file_aio_read(iocb, iov, nr_segs, pos); + return generic_file_read_iter(iocb, to); } static void fuse_write_fill(struct fuse_req *req, struct fuse_file *ff, @@ -677,9 +977,10 @@ static void fuse_write_fill(struct fuse_req *req, struct fuse_file *ff, req->out.args[0].value = outarg; } -static size_t fuse_send_write(struct fuse_req *req, struct file *file, +static size_t fuse_send_write(struct fuse_req *req, struct fuse_io_priv *io, loff_t pos, size_t count, fl_owner_t owner) { + struct file *file = io->file; struct fuse_file *ff = file->private_data; struct fuse_conn *fc = ff->fc; struct fuse_write_in *inarg = &req->misc.write.in; @@ -690,88 +991,29 @@ static size_t fuse_send_write(struct fuse_req *req, struct file *file, inarg->write_flags |= FUSE_WRITE_LOCKOWNER; inarg->lock_owner = fuse_lock_owner_id(fc, owner); } - fuse_request_send(fc, req); - return req->misc.write.out.size; -} -static int fuse_write_begin(struct file *file, struct address_space *mapping, - loff_t pos, unsigned len, unsigned flags, - struct page **pagep, void **fsdata) -{ - pgoff_t index = pos >> PAGE_CACHE_SHIFT; + if (io->async) + return fuse_async_req_send(fc, req, count, io); - *pagep = grab_cache_page_write_begin(mapping, index, flags); - if (!*pagep) - return -ENOMEM; - return 0; + fuse_request_send(fc, req); + return req->misc.write.out.size; } -void fuse_write_update_size(struct inode *inode, loff_t pos) +bool fuse_write_update_size(struct inode *inode, loff_t pos) { struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_inode *fi = get_fuse_inode(inode); + bool ret = false; spin_lock(&fc->lock); fi->attr_version = ++fc->attr_version; - if (pos > inode->i_size) + if (pos > inode->i_size) { i_size_write(inode, pos); - spin_unlock(&fc->lock); -} - -static int fuse_buffered_write(struct file *file, struct inode *inode, - loff_t pos, unsigned count, struct page *page) -{ - int err; - size_t nres; - struct fuse_conn *fc = get_fuse_conn(inode); - unsigned offset = pos & (PAGE_CACHE_SIZE - 1); - struct fuse_req *req; - - if (is_bad_inode(inode)) - return -EIO; - - /* - * Make sure writepages on the same page are not mixed up with - * plain writes. - */ - fuse_wait_on_page_writeback(inode, page->index); - - req = fuse_get_req(fc); - if (IS_ERR(req)) - return PTR_ERR(req); - - req->in.argpages = 1; - req->num_pages = 1; - req->pages[0] = page; - req->page_offset = offset; - nres = fuse_send_write(req, file, pos, count, NULL); - err = req->out.h.error; - fuse_put_request(fc, req); - if (!err && !nres) - err = -EIO; - if (!err) { - pos += nres; - fuse_write_update_size(inode, pos); - if (count == PAGE_CACHE_SIZE) - SetPageUptodate(page); + ret = true; } - fuse_invalidate_attr(inode); - return err ? err : nres; -} - -static int fuse_write_end(struct file *file, struct address_space *mapping, - loff_t pos, unsigned len, unsigned copied, - struct page *page, void *fsdata) -{ - struct inode *inode = mapping->host; - int res = 0; - - if (copied) - res = fuse_buffered_write(file, inode, pos, copied, page); + spin_unlock(&fc->lock); - unlock_page(page); - page_cache_release(page); - return res; + return ret; } static size_t fuse_send_write_pages(struct fuse_req *req, struct file *file, @@ -781,13 +1023,14 @@ static size_t fuse_send_write_pages(struct fuse_req *req, struct file *file, size_t res; unsigned offset; unsigned i; + struct fuse_io_priv io = { .async = 0, .file = file }; for (i = 0; i < req->num_pages; i++) fuse_wait_on_page_writeback(inode, req->pages[i]->index); - res = fuse_send_write(req, file, pos, count, NULL); + res = fuse_send_write(req, &io, pos, count, NULL); - offset = req->page_offset; + offset = req->page_descs[0].offset; count = res; for (i = 0; i < req->num_pages; i++) { struct page *page = req->pages[i]; @@ -818,7 +1061,7 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req, int err; req->in.argpages = 1; - req->page_offset = offset; + req->page_descs[0].offset = offset; do { size_t tmp; @@ -842,9 +1085,7 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req, if (mapping_writably_mapped(mapping)) flush_dcache_page(page); - pagefault_disable(); tmp = iov_iter_copy_from_user_atomic(page, ii, offset, bytes); - pagefault_enable(); flush_dcache_page(page); if (!tmp) { @@ -856,6 +1097,7 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req, err = 0; req->pages[req->num_pages] = page; + req->page_descs[req->num_pages].length = tmp; req->num_pages++; iov_iter_advance(ii, tmp); @@ -868,28 +1110,41 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req, if (!fc->big_writes) break; } while (iov_iter_count(ii) && count < fc->max_write && - req->num_pages < FUSE_MAX_PAGES_PER_REQ && offset == 0); + req->num_pages < req->max_pages && offset == 0); return count > 0 ? count : err; } +static inline unsigned fuse_wr_pages(loff_t pos, size_t len) +{ + return min_t(unsigned, + ((pos + len - 1) >> PAGE_CACHE_SHIFT) - + (pos >> PAGE_CACHE_SHIFT) + 1, + FUSE_MAX_PAGES_PER_REQ); +} + static ssize_t fuse_perform_write(struct file *file, struct address_space *mapping, struct iov_iter *ii, loff_t pos) { struct inode *inode = mapping->host; struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_inode *fi = get_fuse_inode(inode); int err = 0; ssize_t res = 0; if (is_bad_inode(inode)) return -EIO; + if (inode->i_size < pos + iov_iter_count(ii)) + set_bit(FUSE_I_SIZE_UNSTABLE, &fi->state); + do { struct fuse_req *req; ssize_t count; + unsigned nr_pages = fuse_wr_pages(pos, iov_iter_count(ii)); - req = fuse_get_req(fc); + req = fuse_get_req(fc, nr_pages); if (IS_ERR(req)) { err = PTR_ERR(req); break; @@ -919,30 +1174,34 @@ static ssize_t fuse_perform_write(struct file *file, if (res > 0) fuse_write_update_size(inode, pos); + clear_bit(FUSE_I_SIZE_UNSTABLE, &fi->state); fuse_invalidate_attr(inode); return res > 0 ? res : err; } -static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov, - unsigned long nr_segs, loff_t pos) +static ssize_t fuse_file_write_iter(struct kiocb *iocb, struct iov_iter *from) { struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; - size_t count = 0; + size_t count = iov_iter_count(from); ssize_t written = 0; + ssize_t written_buffered = 0; struct inode *inode = mapping->host; ssize_t err; - struct iov_iter i; + loff_t endbyte = 0; + loff_t pos = iocb->ki_pos; - WARN_ON(iocb->ki_pos != pos); + if (get_fuse_conn(inode)->writeback_cache) { + /* Update size (EOF optimization) and mode (SUID clearing) */ + err = fuse_update_attributes(mapping->host, NULL, file, NULL); + if (err) + return err; - err = generic_segment_checks(iov, &nr_segs, &count, VERIFY_READ); - if (err) - return err; + return generic_file_write_iter(iocb, from); + } mutex_lock(&inode->i_mutex); - vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); /* We can write back this queue in page reclaim */ current->backing_dev_info = mapping->backing_dev_info; @@ -954,17 +1213,45 @@ static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov, if (count == 0) goto out; + iov_iter_truncate(from, count); err = file_remove_suid(file); if (err) goto out; - file_update_time(file); + err = file_update_time(file); + if (err) + goto out; + + if (file->f_flags & O_DIRECT) { + written = generic_file_direct_write(iocb, from, pos); + if (written < 0 || !iov_iter_count(from)) + goto out; + + pos += written; + + written_buffered = fuse_perform_write(file, mapping, from, pos); + if (written_buffered < 0) { + err = written_buffered; + goto out; + } + endbyte = pos + written_buffered - 1; + + err = filemap_write_and_wait_range(file->f_mapping, pos, + endbyte); + if (err) + goto out; - iov_iter_init(&i, iov, nr_segs, count, 0); - written = fuse_perform_write(file, mapping, &i, pos); - if (written >= 0) - iocb->ki_pos = pos + written; + invalidate_mapping_pages(file->f_mapping, + pos >> PAGE_CACHE_SHIFT, + endbyte >> PAGE_CACHE_SHIFT); + written += written_buffered; + iocb->ki_pos = pos + written_buffered; + } else { + written = fuse_perform_write(file, mapping, from, pos); + if (written >= 0) + iocb->ki_pos = pos + written; + } out: current->backing_dev_info = NULL; mutex_unlock(&inode->i_mutex); @@ -972,87 +1259,135 @@ out: return written ? written : err; } -static void fuse_release_user_pages(struct fuse_req *req, int write) +static inline void fuse_page_descs_length_init(struct fuse_req *req, + unsigned index, unsigned nr_pages) { - unsigned i; + int i; - for (i = 0; i < req->num_pages; i++) { - struct page *page = req->pages[i]; - if (write) - set_page_dirty_lock(page); - put_page(page); - } + for (i = index; i < index + nr_pages; i++) + req->page_descs[i].length = PAGE_SIZE - + req->page_descs[i].offset; +} + +static inline unsigned long fuse_get_user_addr(const struct iov_iter *ii) +{ + return (unsigned long)ii->iov->iov_base + ii->iov_offset; +} + +static inline size_t fuse_get_frag_size(const struct iov_iter *ii, + size_t max_size) +{ + return min(iov_iter_single_seg_count(ii), max_size); } -static int fuse_get_user_pages(struct fuse_req *req, const char __user *buf, +static int fuse_get_user_pages(struct fuse_req *req, struct iov_iter *ii, size_t *nbytesp, int write) { - size_t nbytes = *nbytesp; - unsigned long user_addr = (unsigned long) buf; - unsigned offset = user_addr & ~PAGE_MASK; - int npages; + size_t nbytes = 0; /* # bytes already packed in req */ /* Special case for kernel I/O: can copy directly into the buffer */ - if (segment_eq(get_fs(), KERNEL_DS)) { + if (ii->type & ITER_KVEC) { + unsigned long user_addr = fuse_get_user_addr(ii); + size_t frag_size = fuse_get_frag_size(ii, *nbytesp); + if (write) req->in.args[1].value = (void *) user_addr; else req->out.args[0].value = (void *) user_addr; + iov_iter_advance(ii, frag_size); + *nbytesp = frag_size; return 0; } - 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); - npages = get_user_pages_fast(user_addr, npages, !write, req->pages); - if (npages < 0) - return npages; + while (nbytes < *nbytesp && req->num_pages < req->max_pages) { + unsigned npages; + size_t start; + unsigned n = req->max_pages - req->num_pages; + ssize_t ret = iov_iter_get_pages(ii, + &req->pages[req->num_pages], + n * PAGE_SIZE, &start); + if (ret < 0) + return ret; + + iov_iter_advance(ii, ret); + nbytes += ret; + + ret += start; + npages = (ret + PAGE_SIZE - 1) / PAGE_SIZE; - req->num_pages = npages; - req->page_offset = offset; + req->page_descs[req->num_pages].offset = start; + fuse_page_descs_length_init(req, req->num_pages, npages); + + req->num_pages += npages; + req->page_descs[req->num_pages - 1].length -= + (PAGE_SIZE - ret) & (PAGE_SIZE - 1); + } if (write) req->in.argpages = 1; else req->out.argpages = 1; - nbytes = (req->num_pages << PAGE_SHIFT) - req->page_offset; - *nbytesp = min(*nbytesp, nbytes); + *nbytesp = nbytes; return 0; } -ssize_t fuse_direct_io(struct file *file, const char __user *buf, - size_t count, loff_t *ppos, int write) +static inline int fuse_iter_npages(const struct iov_iter *ii_p) { + return iov_iter_npages(ii_p, FUSE_MAX_PAGES_PER_REQ); +} + +ssize_t fuse_direct_io(struct fuse_io_priv *io, struct iov_iter *iter, + loff_t *ppos, int flags) +{ + int write = flags & FUSE_DIO_WRITE; + int cuse = flags & FUSE_DIO_CUSE; + struct file *file = io->file; + struct inode *inode = file->f_mapping->host; struct fuse_file *ff = file->private_data; struct fuse_conn *fc = ff->fc; size_t nmax = write ? fc->max_write : fc->max_read; loff_t pos = *ppos; + size_t count = iov_iter_count(iter); + pgoff_t idx_from = pos >> PAGE_CACHE_SHIFT; + pgoff_t idx_to = (pos + count - 1) >> PAGE_CACHE_SHIFT; ssize_t res = 0; struct fuse_req *req; - req = fuse_get_req(fc); + if (io->async) + req = fuse_get_req_for_background(fc, fuse_iter_npages(iter)); + else + req = fuse_get_req(fc, fuse_iter_npages(iter)); if (IS_ERR(req)) return PTR_ERR(req); + if (!cuse && fuse_range_is_writeback(inode, idx_from, idx_to)) { + if (!write) + mutex_lock(&inode->i_mutex); + fuse_sync_writes(inode); + if (!write) + mutex_unlock(&inode->i_mutex); + } + while (count) { size_t nres; fl_owner_t owner = current->files; size_t nbytes = min(count, nmax); - int err = fuse_get_user_pages(req, buf, &nbytes, write); + int err = fuse_get_user_pages(req, iter, &nbytes, write); if (err) { res = err; break; } if (write) - nres = fuse_send_write(req, file, pos, nbytes, owner); + nres = fuse_send_write(req, io, pos, nbytes, owner); else - nres = fuse_send_read(req, file, pos, nbytes, owner); + nres = fuse_send_read(req, io, pos, nbytes, owner); - fuse_release_user_pages(req, !write); + if (!io->async) + fuse_release_user_pages(req, !write); if (req->out.h.error) { if (!res) res = req->out.h.error; @@ -1064,12 +1399,15 @@ ssize_t fuse_direct_io(struct file *file, const char __user *buf, count -= nres; res += nres; pos += nres; - buf += nres; if (nres != nbytes) break; if (count) { fuse_put_request(fc, req); - req = fuse_get_req(fc); + if (io->async) + req = fuse_get_req_for_background(fc, + fuse_iter_npages(iter)); + else + req = fuse_get_req(fc, fuse_iter_npages(iter)); if (IS_ERR(req)) break; } @@ -1083,16 +1421,49 @@ ssize_t fuse_direct_io(struct file *file, const char __user *buf, } EXPORT_SYMBOL_GPL(fuse_direct_io); -static ssize_t fuse_direct_read(struct file *file, char __user *buf, - size_t count, loff_t *ppos) +static ssize_t __fuse_direct_read(struct fuse_io_priv *io, + struct iov_iter *iter, + loff_t *ppos) { ssize_t res; - struct inode *inode = file->f_path.dentry->d_inode; + struct file *file = io->file; + struct inode *inode = file_inode(file); if (is_bad_inode(inode)) return -EIO; - res = fuse_direct_io(file, buf, count, ppos, 0); + res = fuse_direct_io(io, iter, ppos, 0); + + fuse_invalidate_attr(inode); + + return res; +} + +static ssize_t fuse_direct_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct fuse_io_priv io = { .async = 0, .file = file }; + struct iovec iov = { .iov_base = buf, .iov_len = count }; + struct iov_iter ii; + iov_iter_init(&ii, READ, &iov, 1, count); + return __fuse_direct_read(&io, &ii, ppos); +} + +static ssize_t __fuse_direct_write(struct fuse_io_priv *io, + struct iov_iter *iter, + loff_t *ppos) +{ + struct file *file = io->file; + struct inode *inode = file_inode(file); + size_t count = iov_iter_count(iter); + ssize_t res; + + + res = generic_write_checks(file, ppos, &count, 0); + if (!res) { + iov_iter_truncate(iter, count); + res = fuse_direct_io(io, iter, ppos, FUSE_DIO_WRITE); + } fuse_invalidate_attr(inode); @@ -1102,31 +1473,35 @@ static ssize_t fuse_direct_read(struct file *file, char __user *buf, static ssize_t fuse_direct_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos) { - struct inode *inode = file->f_path.dentry->d_inode; + struct iovec iov = { .iov_base = (void __user *)buf, .iov_len = count }; + struct inode *inode = file_inode(file); ssize_t res; + struct fuse_io_priv io = { .async = 0, .file = file }; + struct iov_iter ii; + iov_iter_init(&ii, WRITE, &iov, 1, count); if (is_bad_inode(inode)) return -EIO; /* Don't allow parallel writes to the same file */ mutex_lock(&inode->i_mutex); - res = generic_write_checks(file, ppos, &count, 0); - if (!res) { - res = fuse_direct_io(file, buf, count, ppos, 1); - if (res > 0) - fuse_write_update_size(inode, *ppos); - } + res = __fuse_direct_write(&io, &ii, ppos); + if (res > 0) + fuse_write_update_size(inode, *ppos); mutex_unlock(&inode->i_mutex); - fuse_invalidate_attr(inode); - return res; } static void fuse_writepage_free(struct fuse_conn *fc, struct fuse_req *req) { - __free_page(req->pages[0]); - fuse_file_put(req->ff); + int i; + + for (i = 0; i < req->num_pages; i++) + __free_page(req->pages[i]); + + if (req->ff) + fuse_file_put(req->ff, false); } static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req) @@ -1134,30 +1509,34 @@ static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req) struct inode *inode = req->inode; struct fuse_inode *fi = get_fuse_inode(inode); struct backing_dev_info *bdi = inode->i_mapping->backing_dev_info; + int i; list_del(&req->writepages_entry); - dec_bdi_stat(bdi, BDI_WRITEBACK); - dec_zone_page_state(req->pages[0], NR_WRITEBACK_TEMP); - bdi_writeout_inc(bdi); + for (i = 0; i < req->num_pages; i++) { + dec_bdi_stat(bdi, BDI_WRITEBACK); + dec_zone_page_state(req->pages[i], NR_WRITEBACK_TEMP); + bdi_writeout_inc(bdi); + } wake_up(&fi->page_waitq); } /* Called under fc->lock, may release and reacquire it */ -static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req) +static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req, + loff_t size) __releases(fc->lock) __acquires(fc->lock) { struct fuse_inode *fi = get_fuse_inode(req->inode); - loff_t size = i_size_read(req->inode); struct fuse_write_in *inarg = &req->misc.write.in; + __u64 data_size = req->num_pages * PAGE_CACHE_SIZE; if (!fc->connected) goto out_free; - if (inarg->offset + PAGE_CACHE_SIZE <= size) { - inarg->size = PAGE_CACHE_SIZE; + if (inarg->offset + data_size <= size) { + inarg->size = data_size; } else if (inarg->offset < size) { - inarg->size = size & (PAGE_CACHE_SIZE - 1); + inarg->size = size - inarg->offset; } else { /* Got truncated off completely */ goto out_free; @@ -1188,12 +1567,13 @@ __acquires(fc->lock) { struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_inode *fi = get_fuse_inode(inode); + size_t crop = i_size_read(inode); struct fuse_req *req; while (fi->writectr >= 0 && !list_empty(&fi->queued_writes)) { req = list_entry(fi->queued_writes.next, struct fuse_req, list); list_del_init(&req->list); - fuse_send_writepage(fc, req); + fuse_send_writepage(fc, req, crop); } } @@ -1204,12 +1584,85 @@ static void fuse_writepage_end(struct fuse_conn *fc, struct fuse_req *req) mapping_set_error(inode->i_mapping, req->out.h.error); spin_lock(&fc->lock); + while (req->misc.write.next) { + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_write_in *inarg = &req->misc.write.in; + struct fuse_req *next = req->misc.write.next; + req->misc.write.next = next->misc.write.next; + next->misc.write.next = NULL; + next->ff = fuse_file_get(req->ff); + list_add(&next->writepages_entry, &fi->writepages); + + /* + * Skip fuse_flush_writepages() to make it easy to crop requests + * based on primary request size. + * + * 1st case (trivial): there are no concurrent activities using + * fuse_set/release_nowrite. Then we're on safe side because + * fuse_flush_writepages() would call fuse_send_writepage() + * anyway. + * + * 2nd case: someone called fuse_set_nowrite and it is waiting + * now for completion of all in-flight requests. This happens + * rarely and no more than once per page, so this should be + * okay. + * + * 3rd case: someone (e.g. fuse_do_setattr()) is in the middle + * of fuse_set_nowrite..fuse_release_nowrite section. The fact + * that fuse_set_nowrite returned implies that all in-flight + * requests were completed along with all of their secondary + * requests. Further primary requests are blocked by negative + * writectr. Hence there cannot be any in-flight requests and + * no invocations of fuse_writepage_end() while we're in + * fuse_set_nowrite..fuse_release_nowrite section. + */ + fuse_send_writepage(fc, next, inarg->offset + inarg->size); + } fi->writectr--; fuse_writepage_finish(fc, req); spin_unlock(&fc->lock); fuse_writepage_free(fc, req); } +static struct fuse_file *__fuse_write_file_get(struct fuse_conn *fc, + struct fuse_inode *fi) +{ + struct fuse_file *ff = NULL; + + spin_lock(&fc->lock); + if (!list_empty(&fi->write_files)) { + ff = list_entry(fi->write_files.next, struct fuse_file, + write_entry); + fuse_file_get(ff); + } + spin_unlock(&fc->lock); + + return ff; +} + +static struct fuse_file *fuse_write_file_get(struct fuse_conn *fc, + struct fuse_inode *fi) +{ + struct fuse_file *ff = __fuse_write_file_get(fc, fi); + WARN_ON(!ff); + return ff; +} + +int fuse_write_inode(struct inode *inode, struct writeback_control *wbc) +{ + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_inode *fi = get_fuse_inode(inode); + struct fuse_file *ff; + int err; + + ff = __fuse_write_file_get(fc, fi); + err = fuse_flush_times(inode, ff); + if (ff) + fuse_file_put(ff, 0); + + return err; +} + static int fuse_writepage_locked(struct page *page) { struct address_space *mapping = page->mapping; @@ -1217,39 +1670,40 @@ static int fuse_writepage_locked(struct page *page) struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_req *req; - struct fuse_file *ff; struct page *tmp_page; + int error = -ENOMEM; set_page_writeback(page); - req = fuse_request_alloc_nofs(); + req = fuse_request_alloc_nofs(1); if (!req) goto err; + req->background = 1; /* writeback always goes to bg_queue */ tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM); if (!tmp_page) goto err_free; - spin_lock(&fc->lock); - BUG_ON(list_empty(&fi->write_files)); - ff = list_entry(fi->write_files.next, struct fuse_file, write_entry); - req->ff = fuse_file_get(ff); - spin_unlock(&fc->lock); + error = -EIO; + req->ff = fuse_write_file_get(fc, fi); + if (!req->ff) + goto err_nofile; - fuse_write_fill(req, ff, page_offset(page), 0); + fuse_write_fill(req, req->ff, page_offset(page), 0); copy_highpage(tmp_page, page); req->misc.write.in.write_flags |= FUSE_WRITE_CACHE; + req->misc.write.next = NULL; req->in.argpages = 1; req->num_pages = 1; req->pages[0] = tmp_page; - req->page_offset = 0; + req->page_descs[0].offset = 0; + req->page_descs[0].length = PAGE_SIZE; req->end = fuse_writepage_end; req->inode = inode; inc_bdi_stat(mapping->backing_dev_info, BDI_WRITEBACK); inc_zone_page_state(tmp_page, NR_WRITEBACK_TEMP); - end_page_writeback(page); spin_lock(&fc->lock); list_add(&req->writepages_entry, &fi->writepages); @@ -1257,25 +1711,344 @@ static int fuse_writepage_locked(struct page *page) fuse_flush_writepages(inode); spin_unlock(&fc->lock); + end_page_writeback(page); + return 0; +err_nofile: + __free_page(tmp_page); err_free: fuse_request_free(req); err: end_page_writeback(page); - return -ENOMEM; + return error; } static int fuse_writepage(struct page *page, struct writeback_control *wbc) { int err; + if (fuse_page_is_writeback(page->mapping->host, page->index)) { + /* + * ->writepages() should be called for sync() and friends. We + * should only get here on direct reclaim and then we are + * allowed to skip a page which is already in flight + */ + WARN_ON(wbc->sync_mode == WB_SYNC_ALL); + + redirty_page_for_writepage(wbc, page); + return 0; + } + err = fuse_writepage_locked(page); unlock_page(page); return err; } +struct fuse_fill_wb_data { + struct fuse_req *req; + struct fuse_file *ff; + struct inode *inode; + struct page **orig_pages; +}; + +static void fuse_writepages_send(struct fuse_fill_wb_data *data) +{ + struct fuse_req *req = data->req; + struct inode *inode = data->inode; + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_inode *fi = get_fuse_inode(inode); + int num_pages = req->num_pages; + int i; + + req->ff = fuse_file_get(data->ff); + spin_lock(&fc->lock); + list_add_tail(&req->list, &fi->queued_writes); + fuse_flush_writepages(inode); + spin_unlock(&fc->lock); + + for (i = 0; i < num_pages; i++) + end_page_writeback(data->orig_pages[i]); +} + +static bool fuse_writepage_in_flight(struct fuse_req *new_req, + struct page *page) +{ + struct fuse_conn *fc = get_fuse_conn(new_req->inode); + struct fuse_inode *fi = get_fuse_inode(new_req->inode); + struct fuse_req *tmp; + struct fuse_req *old_req; + bool found = false; + pgoff_t curr_index; + + BUG_ON(new_req->num_pages != 0); + + spin_lock(&fc->lock); + list_del(&new_req->writepages_entry); + list_for_each_entry(old_req, &fi->writepages, writepages_entry) { + BUG_ON(old_req->inode != new_req->inode); + curr_index = old_req->misc.write.in.offset >> PAGE_CACHE_SHIFT; + if (curr_index <= page->index && + page->index < curr_index + old_req->num_pages) { + found = true; + break; + } + } + if (!found) { + list_add(&new_req->writepages_entry, &fi->writepages); + goto out_unlock; + } + + new_req->num_pages = 1; + for (tmp = old_req; tmp != NULL; tmp = tmp->misc.write.next) { + BUG_ON(tmp->inode != new_req->inode); + curr_index = tmp->misc.write.in.offset >> PAGE_CACHE_SHIFT; + if (tmp->num_pages == 1 && + curr_index == page->index) { + old_req = tmp; + } + } + + if (old_req->num_pages == 1 && (old_req->state == FUSE_REQ_INIT || + old_req->state == FUSE_REQ_PENDING)) { + struct backing_dev_info *bdi = page->mapping->backing_dev_info; + + copy_highpage(old_req->pages[0], page); + spin_unlock(&fc->lock); + + dec_bdi_stat(bdi, BDI_WRITEBACK); + dec_zone_page_state(page, NR_WRITEBACK_TEMP); + bdi_writeout_inc(bdi); + fuse_writepage_free(fc, new_req); + fuse_request_free(new_req); + goto out; + } else { + new_req->misc.write.next = old_req->misc.write.next; + old_req->misc.write.next = new_req; + } +out_unlock: + spin_unlock(&fc->lock); +out: + return found; +} + +static int fuse_writepages_fill(struct page *page, + struct writeback_control *wbc, void *_data) +{ + struct fuse_fill_wb_data *data = _data; + struct fuse_req *req = data->req; + struct inode *inode = data->inode; + struct fuse_conn *fc = get_fuse_conn(inode); + struct page *tmp_page; + bool is_writeback; + int err; + + if (!data->ff) { + err = -EIO; + data->ff = fuse_write_file_get(fc, get_fuse_inode(inode)); + if (!data->ff) + goto out_unlock; + } + + /* + * Being under writeback is unlikely but possible. For example direct + * read to an mmaped fuse file will set the page dirty twice; once when + * the pages are faulted with get_user_pages(), and then after the read + * completed. + */ + is_writeback = fuse_page_is_writeback(inode, page->index); + + if (req && req->num_pages && + (is_writeback || req->num_pages == FUSE_MAX_PAGES_PER_REQ || + (req->num_pages + 1) * PAGE_CACHE_SIZE > fc->max_write || + data->orig_pages[req->num_pages - 1]->index + 1 != page->index)) { + fuse_writepages_send(data); + data->req = NULL; + } + err = -ENOMEM; + tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM); + if (!tmp_page) + goto out_unlock; + + /* + * The page must not be redirtied until the writeout is completed + * (i.e. userspace has sent a reply to the write request). Otherwise + * there could be more than one temporary page instance for each real + * page. + * + * This is ensured by holding the page lock in page_mkwrite() while + * checking fuse_page_is_writeback(). We already hold the page lock + * since clear_page_dirty_for_io() and keep it held until we add the + * request to the fi->writepages list and increment req->num_pages. + * After this fuse_page_is_writeback() will indicate that the page is + * under writeback, so we can release the page lock. + */ + if (data->req == NULL) { + struct fuse_inode *fi = get_fuse_inode(inode); + + err = -ENOMEM; + req = fuse_request_alloc_nofs(FUSE_MAX_PAGES_PER_REQ); + if (!req) { + __free_page(tmp_page); + goto out_unlock; + } + + fuse_write_fill(req, data->ff, page_offset(page), 0); + req->misc.write.in.write_flags |= FUSE_WRITE_CACHE; + req->misc.write.next = NULL; + req->in.argpages = 1; + req->background = 1; + req->num_pages = 0; + req->end = fuse_writepage_end; + req->inode = inode; + + spin_lock(&fc->lock); + list_add(&req->writepages_entry, &fi->writepages); + spin_unlock(&fc->lock); + + data->req = req; + } + set_page_writeback(page); + + copy_highpage(tmp_page, page); + req->pages[req->num_pages] = tmp_page; + req->page_descs[req->num_pages].offset = 0; + req->page_descs[req->num_pages].length = PAGE_SIZE; + + inc_bdi_stat(page->mapping->backing_dev_info, BDI_WRITEBACK); + inc_zone_page_state(tmp_page, NR_WRITEBACK_TEMP); + + err = 0; + if (is_writeback && fuse_writepage_in_flight(req, page)) { + end_page_writeback(page); + data->req = NULL; + goto out_unlock; + } + data->orig_pages[req->num_pages] = page; + + /* + * Protected by fc->lock against concurrent access by + * fuse_page_is_writeback(). + */ + spin_lock(&fc->lock); + req->num_pages++; + spin_unlock(&fc->lock); + +out_unlock: + unlock_page(page); + + return err; +} + +static int fuse_writepages(struct address_space *mapping, + struct writeback_control *wbc) +{ + struct inode *inode = mapping->host; + struct fuse_fill_wb_data data; + int err; + + err = -EIO; + if (is_bad_inode(inode)) + goto out; + + data.inode = inode; + data.req = NULL; + data.ff = NULL; + + err = -ENOMEM; + data.orig_pages = kcalloc(FUSE_MAX_PAGES_PER_REQ, + sizeof(struct page *), + GFP_NOFS); + if (!data.orig_pages) + goto out; + + err = write_cache_pages(mapping, wbc, fuse_writepages_fill, &data); + if (data.req) { + /* Ignore errors if we can write at least one page */ + BUG_ON(!data.req->num_pages); + fuse_writepages_send(&data); + err = 0; + } + if (data.ff) + fuse_file_put(data.ff, false); + + kfree(data.orig_pages); +out: + return err; +} + +/* + * It's worthy to make sure that space is reserved on disk for the write, + * but how to implement it without killing performance need more thinking. + */ +static int fuse_write_begin(struct file *file, struct address_space *mapping, + loff_t pos, unsigned len, unsigned flags, + struct page **pagep, void **fsdata) +{ + pgoff_t index = pos >> PAGE_CACHE_SHIFT; + struct fuse_conn *fc = get_fuse_conn(file->f_dentry->d_inode); + struct page *page; + loff_t fsize; + int err = -ENOMEM; + + WARN_ON(!fc->writeback_cache); + + page = grab_cache_page_write_begin(mapping, index, flags); + if (!page) + goto error; + + fuse_wait_on_page_writeback(mapping->host, page->index); + + if (PageUptodate(page) || len == PAGE_CACHE_SIZE) + goto success; + /* + * Check if the start this page comes after the end of file, in which + * case the readpage can be optimized away. + */ + fsize = i_size_read(mapping->host); + if (fsize <= (pos & PAGE_CACHE_MASK)) { + size_t off = pos & ~PAGE_CACHE_MASK; + if (off) + zero_user_segment(page, 0, off); + goto success; + } + err = fuse_do_readpage(file, page); + if (err) + goto cleanup; +success: + *pagep = page; + return 0; + +cleanup: + unlock_page(page); + page_cache_release(page); +error: + return err; +} + +static int fuse_write_end(struct file *file, struct address_space *mapping, + loff_t pos, unsigned len, unsigned copied, + struct page *page, void *fsdata) +{ + struct inode *inode = page->mapping->host; + + if (!PageUptodate(page)) { + /* Zero any unwritten bytes at the end of the page */ + size_t endoff = (pos + copied) & ~PAGE_CACHE_MASK; + if (endoff) + zero_user_segment(page, endoff, PAGE_CACHE_SIZE); + SetPageUptodate(page); + } + + fuse_write_update_size(inode, pos + copied); + set_page_dirty(page); + unlock_page(page); + page_cache_release(page); + + return copied; +} + static int fuse_launder_page(struct page *page) { int err = 0; @@ -1315,38 +2088,32 @@ static void fuse_vma_close(struct vm_area_struct *vma) static int fuse_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) { struct page *page = vmf->page; - /* - * Don't use page->mapping as it may become NULL from a - * concurrent truncate. - */ - struct inode *inode = vma->vm_file->f_mapping->host; + struct inode *inode = file_inode(vma->vm_file); + + file_update_time(vma->vm_file); + lock_page(page); + if (page->mapping != inode->i_mapping) { + unlock_page(page); + return VM_FAULT_NOPAGE; + } fuse_wait_on_page_writeback(inode, page->index); - return 0; + return VM_FAULT_LOCKED; } static const struct vm_operations_struct fuse_file_vm_ops = { .close = fuse_vma_close, .fault = filemap_fault, + .map_pages = filemap_map_pages, .page_mkwrite = fuse_page_mkwrite, + .remap_pages = generic_file_remap_pages, }; static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma) { - if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE)) { - struct inode *inode = file->f_dentry->d_inode; - struct fuse_conn *fc = get_fuse_conn(inode); - struct fuse_inode *fi = get_fuse_inode(inode); - struct fuse_file *ff = file->private_data; - /* - * file may be written through mmap, so chain it onto the - * inodes's write_file list - */ - spin_lock(&fc->lock); - if (list_empty(&ff->write_entry)) - list_add(&ff->write_entry, &fi->write_files); - spin_unlock(&fc->lock); - } + if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE)) + fuse_link_write_file(file); + file_accessed(file); vma->vm_ops = &fuse_file_vm_ops; return 0; @@ -1392,7 +2159,7 @@ static void fuse_lk_fill(struct fuse_req *req, struct file *file, const struct file_lock *fl, int opcode, pid_t pid, int flock) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_file *ff = file->private_data; struct fuse_lk_in *arg = &req->misc.lk_in; @@ -1414,13 +2181,13 @@ static void fuse_lk_fill(struct fuse_req *req, struct file *file, static int fuse_getlk(struct file *file, struct file_lock *fl) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_req *req; struct fuse_lk_out outarg; int err; - req = fuse_get_req(fc); + req = fuse_get_req_nopages(fc); if (IS_ERR(req)) return PTR_ERR(req); @@ -1439,14 +2206,14 @@ static int fuse_getlk(struct file *file, struct file_lock *fl) static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_req *req; int opcode = (fl->fl_flags & FL_SLEEP) ? FUSE_SETLKW : FUSE_SETLK; pid_t pid = fl->fl_type != F_UNLCK ? current->tgid : 0; int err; - if (fl->fl_lmops && fl->fl_lmops->fl_grant) { + if (fl->fl_lmops && fl->fl_lmops->lm_grant) { /* NLM needs asynchronous locks, which we don't support yet */ return -ENOLCK; } @@ -1455,7 +2222,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) if (fl->fl_flags & FL_CLOSE) return 0; - req = fuse_get_req(fc); + req = fuse_get_req_nopages(fc); if (IS_ERR(req)) return PTR_ERR(req); @@ -1471,7 +2238,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) static int fuse_file_lock(struct file *file, int cmd, struct file_lock *fl) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); int err; @@ -1494,15 +2261,17 @@ static int fuse_file_lock(struct file *file, int cmd, struct file_lock *fl) static int fuse_file_flock(struct file *file, int cmd, struct file_lock *fl) { - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); int err; - if (fc->no_lock) { + if (fc->no_flock) { err = flock_lock_file_wait(file, fl); } else { + struct fuse_file *ff = file->private_data; + /* emulate flock with POSIX locks */ - fl->fl_owner = (fl_owner_t) file; + ff->flock = true; err = fuse_setlk(file, fl, 1); } @@ -1521,7 +2290,7 @@ static sector_t fuse_bmap(struct address_space *mapping, sector_t block) if (!inode->i_sb->s_bdev || fc->no_bmap) return 0; - req = fuse_get_req(fc); + req = fuse_get_req_nopages(fc); if (IS_ERR(req)) return 0; @@ -1545,32 +2314,21 @@ static sector_t fuse_bmap(struct address_space *mapping, sector_t block) return err ? 0 : outarg.block; } -static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin) +static loff_t fuse_file_llseek(struct file *file, loff_t offset, int whence) { loff_t retval; - struct inode *inode = file->f_path.dentry->d_inode; + struct inode *inode = file_inode(file); + + /* No i_mutex protection necessary for SEEK_CUR and SEEK_SET */ + if (whence == SEEK_CUR || whence == SEEK_SET) + return generic_file_llseek(file, offset, whence); mutex_lock(&inode->i_mutex); - switch (origin) { - case SEEK_END: - retval = fuse_update_attributes(inode, NULL, file, NULL); - if (retval) - goto exit; - offset += i_size_read(inode); - break; - case SEEK_CUR: - offset += file->f_pos; - } - retval = -EINVAL; - if (offset >= 0 && offset <= inode->i_sb->s_maxbytes) { - if (offset != file->f_pos) { - file->f_pos = offset; - file->f_version = 0; - } - retval = offset; - } -exit: + retval = fuse_update_attributes(inode, NULL, file, NULL); + if (!retval) + retval = generic_file_llseek(file, offset, whence); mutex_unlock(&inode->i_mutex); + return retval; } @@ -1583,7 +2341,7 @@ static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov, if (!bytes) return 0; - iov_iter_init(&ii, iov, nr_segs, bytes, 0); + iov_iter_init(&ii, to_user ? READ : WRITE, iov, nr_segs, bytes); while (iov_iter_count(&ii)) { struct page *page = pages[page_idx++]; @@ -1618,6 +2376,94 @@ static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov, } /* + * CUSE servers compiled on 32bit broke on 64bit kernels because the + * ABI was defined to be 'struct iovec' which is different on 32bit + * and 64bit. Fortunately we can determine which structure the server + * used from the size of the reply. + */ +static int fuse_copy_ioctl_iovec_old(struct iovec *dst, void *src, + size_t transferred, unsigned count, + bool is_compat) +{ +#ifdef CONFIG_COMPAT + if (count * sizeof(struct compat_iovec) == transferred) { + struct compat_iovec *ciov = src; + unsigned i; + + /* + * With this interface a 32bit server cannot support + * non-compat (i.e. ones coming from 64bit apps) ioctl + * requests + */ + if (!is_compat) + return -EINVAL; + + for (i = 0; i < count; i++) { + dst[i].iov_base = compat_ptr(ciov[i].iov_base); + dst[i].iov_len = ciov[i].iov_len; + } + return 0; + } +#endif + + if (count * sizeof(struct iovec) != transferred) + return -EIO; + + memcpy(dst, src, transferred); + return 0; +} + +/* Make sure iov_length() won't overflow */ +static int fuse_verify_ioctl_iov(struct iovec *iov, size_t count) +{ + size_t n; + u32 max = FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT; + + for (n = 0; n < count; n++, iov++) { + if (iov->iov_len > (size_t) max) + return -ENOMEM; + max -= iov->iov_len; + } + return 0; +} + +static int fuse_copy_ioctl_iovec(struct fuse_conn *fc, struct iovec *dst, + void *src, size_t transferred, unsigned count, + bool is_compat) +{ + unsigned i; + struct fuse_ioctl_iovec *fiov = src; + + if (fc->minor < 16) { + return fuse_copy_ioctl_iovec_old(dst, src, transferred, + count, is_compat); + } + + if (count * sizeof(struct fuse_ioctl_iovec) != transferred) + return -EIO; + + for (i = 0; i < count; i++) { + /* Did the server supply an inappropriate value? */ + if (fiov[i].base != (unsigned long) fiov[i].base || + fiov[i].len != (unsigned long) fiov[i].len) + return -EIO; + + dst[i].iov_base = (void __user *) (unsigned long) fiov[i].base; + dst[i].iov_len = (size_t) fiov[i].len; + +#ifdef CONFIG_COMPAT + if (is_compat && + (ptr_to_compat(dst[i].iov_base) != fiov[i].base || + (compat_size_t) dst[i].iov_len != fiov[i].len)) + return -EIO; +#endif + } + + return 0; +} + + +/* * For ioctls, there is no generic way to determine how much memory * needs to be read and/or written. Furthermore, ioctls are allowed * to dereference the passed pointer, so the parameter requires deep @@ -1677,18 +2523,25 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, struct fuse_ioctl_out outarg; struct fuse_req *req = NULL; struct page **pages = NULL; - struct page *iov_page = NULL; + struct iovec *iov_page = NULL; struct iovec *in_iov = NULL, *out_iov = NULL; unsigned int in_iovs = 0, out_iovs = 0, num_pages = 0, max_pages; size_t in_size, out_size, transferred; int err; +#if BITS_PER_LONG == 32 + inarg.flags |= FUSE_IOCTL_32BIT; +#else + if (flags & FUSE_IOCTL_COMPAT) + inarg.flags |= FUSE_IOCTL_32BIT; +#endif + /* assume all the iovs returned by client always fits in a page */ - BUILD_BUG_ON(sizeof(struct iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE); + BUILD_BUG_ON(sizeof(struct fuse_ioctl_iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE); err = -ENOMEM; - pages = kzalloc(sizeof(pages[0]) * FUSE_MAX_PAGES_PER_REQ, GFP_KERNEL); - iov_page = alloc_page(GFP_KERNEL); + pages = kcalloc(FUSE_MAX_PAGES_PER_REQ, sizeof(pages[0]), GFP_KERNEL); + iov_page = (struct iovec *) __get_free_page(GFP_KERNEL); if (!pages || !iov_page) goto out; @@ -1697,7 +2550,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, * RETRY from server is not allowed. */ if (!(flags & FUSE_IOCTL_UNRESTRICTED)) { - struct iovec *iov = page_address(iov_page); + struct iovec *iov = iov_page; iov->iov_base = (void __user *)arg; iov->iov_len = _IOC_SIZE(cmd); @@ -1735,7 +2588,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, num_pages++; } - req = fuse_get_req(fc); + req = fuse_get_req(fc, num_pages); if (IS_ERR(req)) { err = PTR_ERR(req); req = NULL; @@ -1743,6 +2596,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, } memcpy(req->pages, pages, sizeof(req->pages[0]) * num_pages); req->num_pages = num_pages; + fuse_page_descs_length_init(req, 0, req->num_pages); /* okay, let's send it to the client */ req->in.h.opcode = FUSE_IOCTL; @@ -1778,7 +2632,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, /* did it ask for retry? */ if (outarg.flags & FUSE_IOCTL_RETRY) { - char *vaddr; + void *vaddr; /* no retry if in restricted mode */ err = -EIO; @@ -1798,18 +2652,25 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, in_iovs + out_iovs > FUSE_IOCTL_MAX_IOV) goto out; - err = -EIO; - if ((in_iovs + out_iovs) * sizeof(struct iovec) != transferred) + vaddr = kmap_atomic(pages[0]); + err = fuse_copy_ioctl_iovec(fc, iov_page, vaddr, + transferred, in_iovs + out_iovs, + (flags & FUSE_IOCTL_COMPAT) != 0); + kunmap_atomic(vaddr); + if (err) goto out; - /* okay, copy in iovs and retry */ - vaddr = kmap_atomic(pages[0], KM_USER0); - memcpy(page_address(iov_page), vaddr, transferred); - kunmap_atomic(vaddr, KM_USER0); - - in_iov = page_address(iov_page); + in_iov = iov_page; out_iov = in_iov + in_iovs; + err = fuse_verify_ioctl_iov(in_iov, in_iovs); + if (err) + goto out; + + err = fuse_verify_ioctl_iov(out_iov, out_iovs); + if (err) + goto out; + goto retry; } @@ -1821,8 +2682,7 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, out: if (req) fuse_put_request(fc, req); - if (iov_page) - __free_page(iov_page); + free_page((unsigned long) iov_page); while (num_pages) __free_page(pages[--num_pages]); kfree(pages); @@ -1831,13 +2691,13 @@ long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg, } EXPORT_SYMBOL_GPL(fuse_do_ioctl); -static long fuse_file_ioctl_common(struct file *file, unsigned int cmd, - unsigned long arg, unsigned int flags) +long fuse_ioctl_common(struct file *file, unsigned int cmd, + unsigned long arg, unsigned int flags) { - struct inode *inode = file->f_dentry->d_inode; + struct inode *inode = file_inode(file); struct fuse_conn *fc = get_fuse_conn(inode); - if (!fuse_allow_task(fc, current)) + if (!fuse_allow_current_process(fc)) return -EACCES; if (is_bad_inode(inode)) @@ -1849,13 +2709,13 @@ static long fuse_file_ioctl_common(struct file *file, unsigned int cmd, static long fuse_file_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { - return fuse_file_ioctl_common(file, cmd, arg, 0); + return fuse_ioctl_common(file, cmd, arg, 0); } static long fuse_file_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { - return fuse_file_ioctl_common(file, cmd, arg, FUSE_IOCTL_COMPAT); + return fuse_ioctl_common(file, cmd, arg, FUSE_IOCTL_COMPAT); } /* @@ -1899,7 +2759,7 @@ static void fuse_register_polled_file(struct fuse_conn *fc, { spin_lock(&fc->lock); if (RB_EMPTY_NODE(&ff->polled_node)) { - struct rb_node **link, *parent; + struct rb_node **link, *uninitialized_var(parent); link = fuse_find_polled_node(fc, ff->kh, &parent); BUG_ON(*link); @@ -1922,6 +2782,7 @@ unsigned fuse_file_poll(struct file *file, poll_table *wait) return DEFAULT_POLLMASK; poll_wait(file, &ff->poll_wait, wait); + inarg.events = (__u32)poll_requested_events(wait); /* * Ask for notification iff there's someone waiting for it. @@ -1932,7 +2793,7 @@ unsigned fuse_file_poll(struct file *file, poll_table *wait) fuse_register_polled_file(fc, ff); } - req = fuse_get_req(fc); + req = fuse_get_req_nopages(fc); if (IS_ERR(req)) return POLLERR; @@ -1982,12 +2843,198 @@ int fuse_notify_poll_wakeup(struct fuse_conn *fc, return 0; } +static void fuse_do_truncate(struct file *file) +{ + struct inode *inode = file->f_mapping->host; + struct iattr attr; + + attr.ia_valid = ATTR_SIZE; + attr.ia_size = i_size_read(inode); + + attr.ia_file = file; + attr.ia_valid |= ATTR_FILE; + + fuse_do_setattr(inode, &attr, file); +} + +static inline loff_t fuse_round_up(loff_t off) +{ + return round_up(off, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT); +} + +static ssize_t +fuse_direct_IO(int rw, struct kiocb *iocb, struct iov_iter *iter, + loff_t offset) +{ + ssize_t ret = 0; + struct file *file = iocb->ki_filp; + struct fuse_file *ff = file->private_data; + bool async_dio = ff->fc->async_dio; + loff_t pos = 0; + struct inode *inode; + loff_t i_size; + size_t count = iov_iter_count(iter); + struct fuse_io_priv *io; + + pos = offset; + inode = file->f_mapping->host; + i_size = i_size_read(inode); + + if ((rw == READ) && (offset > i_size)) + return 0; + + /* optimization for short read */ + if (async_dio && rw != WRITE && offset + count > i_size) { + if (offset >= i_size) + return 0; + count = min_t(loff_t, count, fuse_round_up(i_size - offset)); + iov_iter_truncate(iter, count); + } + + io = kmalloc(sizeof(struct fuse_io_priv), GFP_KERNEL); + if (!io) + return -ENOMEM; + spin_lock_init(&io->lock); + io->reqs = 1; + io->bytes = -1; + io->size = 0; + io->offset = offset; + io->write = (rw == WRITE); + io->err = 0; + io->file = file; + /* + * By default, we want to optimize all I/Os with async request + * submission to the client filesystem if supported. + */ + io->async = async_dio; + io->iocb = iocb; + + /* + * We cannot asynchronously extend the size of a file. We have no method + * to wait on real async I/O requests, so we must submit this request + * synchronously. + */ + if (!is_sync_kiocb(iocb) && (offset + count > i_size) && rw == WRITE) + io->async = false; + + if (rw == WRITE) + ret = __fuse_direct_write(io, iter, &pos); + else + ret = __fuse_direct_read(io, iter, &pos); + + if (io->async) { + fuse_aio_complete(io, ret < 0 ? ret : 0, -1); + + /* we have a non-extending, async request, so return */ + if (!is_sync_kiocb(iocb)) + return -EIOCBQUEUED; + + ret = wait_on_sync_kiocb(iocb); + } else { + kfree(io); + } + + if (rw == WRITE) { + if (ret > 0) + fuse_write_update_size(inode, pos); + else if (ret < 0 && offset + count > i_size) + fuse_do_truncate(file); + } + + return ret; +} + +static long fuse_file_fallocate(struct file *file, int mode, loff_t offset, + loff_t length) +{ + struct fuse_file *ff = file->private_data; + struct inode *inode = file->f_inode; + struct fuse_inode *fi = get_fuse_inode(inode); + struct fuse_conn *fc = ff->fc; + struct fuse_req *req; + struct fuse_fallocate_in inarg = { + .fh = ff->fh, + .offset = offset, + .length = length, + .mode = mode + }; + int err; + bool lock_inode = !(mode & FALLOC_FL_KEEP_SIZE) || + (mode & FALLOC_FL_PUNCH_HOLE); + + if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE)) + return -EOPNOTSUPP; + + if (fc->no_fallocate) + return -EOPNOTSUPP; + + if (lock_inode) { + mutex_lock(&inode->i_mutex); + if (mode & FALLOC_FL_PUNCH_HOLE) { + loff_t endbyte = offset + length - 1; + err = filemap_write_and_wait_range(inode->i_mapping, + offset, endbyte); + if (err) + goto out; + + fuse_sync_writes(inode); + } + } + + if (!(mode & FALLOC_FL_KEEP_SIZE)) + set_bit(FUSE_I_SIZE_UNSTABLE, &fi->state); + + req = fuse_get_req_nopages(fc); + if (IS_ERR(req)) { + err = PTR_ERR(req); + goto out; + } + + req->in.h.opcode = FUSE_FALLOCATE; + req->in.h.nodeid = ff->nodeid; + req->in.numargs = 1; + req->in.args[0].size = sizeof(inarg); + req->in.args[0].value = &inarg; + fuse_request_send(fc, req); + err = req->out.h.error; + if (err == -ENOSYS) { + fc->no_fallocate = 1; + err = -EOPNOTSUPP; + } + fuse_put_request(fc, req); + + if (err) + goto out; + + /* we could have extended the file */ + if (!(mode & FALLOC_FL_KEEP_SIZE)) { + bool changed = fuse_write_update_size(inode, offset + length); + + if (changed && fc->writeback_cache) + file_update_time(file); + } + + if (mode & FALLOC_FL_PUNCH_HOLE) + truncate_pagecache_range(inode, offset, offset + length - 1); + + fuse_invalidate_attr(inode); + +out: + if (!(mode & FALLOC_FL_KEEP_SIZE)) + clear_bit(FUSE_I_SIZE_UNSTABLE, &fi->state); + + if (lock_inode) + mutex_unlock(&inode->i_mutex); + + return err; +} + static const struct file_operations fuse_file_operations = { .llseek = fuse_file_llseek, - .read = do_sync_read, - .aio_read = fuse_file_aio_read, - .write = do_sync_write, - .aio_write = fuse_file_aio_write, + .read = new_sync_read, + .read_iter = fuse_file_read_iter, + .write = new_sync_write, + .write_iter = fuse_file_write_iter, .mmap = fuse_file_mmap, .open = fuse_open, .flush = fuse_flush, @@ -1999,6 +3046,7 @@ static const struct file_operations fuse_file_operations = { .unlocked_ioctl = fuse_file_ioctl, .compat_ioctl = fuse_file_compat_ioctl, .poll = fuse_file_poll, + .fallocate = fuse_file_fallocate, }; static const struct file_operations fuse_direct_io_file_operations = { @@ -2015,18 +3063,21 @@ static const struct file_operations fuse_direct_io_file_operations = { .unlocked_ioctl = fuse_file_ioctl, .compat_ioctl = fuse_file_compat_ioctl, .poll = fuse_file_poll, + .fallocate = fuse_file_fallocate, /* no splice_read */ }; static const struct address_space_operations fuse_file_aops = { .readpage = fuse_readpage, .writepage = fuse_writepage, + .writepages = fuse_writepages, .launder_page = fuse_launder_page, - .write_begin = fuse_write_begin, - .write_end = fuse_write_end, .readpages = fuse_readpages, .set_page_dirty = __set_page_dirty_nobuffers, .bmap = fuse_bmap, + .direct_IO = fuse_direct_IO, + .write_begin = fuse_write_begin, + .write_end = fuse_write_end, }; void fuse_init_file_inode(struct inode *inode) |