From: kernel test robot <lkp@intel.com>
To: Qu Wenruo <wqu@suse.com>, linux-btrfs@vger.kernel.org
Cc: kbuild-all@lists.01.org
Subject: Re: [PATCH v2 16/18] btrfs: introduce btrfs_subpage for data inodes
Date: Fri, 11 Dec 2020 08:43:14 +0800 [thread overview]
Message-ID: <202012110846.Zo8Tb8bu-lkp@intel.com> (raw)
In-Reply-To: <20201210063905.75727-17-wqu@suse.com>
[-- Attachment #1: Type: text/plain, Size: 7566 bytes --]
Hi Qu,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on kdave/for-next]
[also build test WARNING on next-20201210]
[cannot apply to v5.10-rc7]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Qu-Wenruo/btrfs-add-read-only-support-for-subpage-sector-size/20201210-144442
base: https://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux.git for-next
config: i386-randconfig-m021-20201209 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
New smatch warnings:
fs/btrfs/inode.c:8361 btrfs_page_mkwrite() warn: unsigned 'ret' is never less than zero.
Old smatch warnings:
include/linux/fs.h:862 i_size_write() warn: statement has no effect 31
vim +/ret +8361 fs/btrfs/inode.c
8283
8284 /*
8285 * btrfs_page_mkwrite() is not allowed to change the file size as it gets
8286 * called from a page fault handler when a page is first dirtied. Hence we must
8287 * be careful to check for EOF conditions here. We set the page up correctly
8288 * for a written page which means we get ENOSPC checking when writing into
8289 * holes and correct delalloc and unwritten extent mapping on filesystems that
8290 * support these features.
8291 *
8292 * We are not allowed to take the i_mutex here so we have to play games to
8293 * protect against truncate races as the page could now be beyond EOF. Because
8294 * truncate_setsize() writes the inode size before removing pages, once we have
8295 * the page lock we can determine safely if the page is beyond EOF. If it is not
8296 * beyond EOF, then the page is guaranteed safe against truncation until we
8297 * unlock the page.
8298 */
8299 vm_fault_t btrfs_page_mkwrite(struct vm_fault *vmf)
8300 {
8301 struct page *page = vmf->page;
8302 struct inode *inode = file_inode(vmf->vma->vm_file);
8303 struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
8304 struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
8305 struct btrfs_ordered_extent *ordered;
8306 struct extent_state *cached_state = NULL;
8307 struct extent_changeset *data_reserved = NULL;
8308 char *kaddr;
8309 unsigned long zero_start;
8310 loff_t size;
8311 vm_fault_t ret;
8312 int ret2;
8313 int reserved = 0;
8314 u64 reserved_space;
8315 u64 page_start;
8316 u64 page_end;
8317 u64 end;
8318
8319 reserved_space = PAGE_SIZE;
8320
8321 sb_start_pagefault(inode->i_sb);
8322 page_start = page_offset(page);
8323 page_end = page_start + PAGE_SIZE - 1;
8324 end = page_end;
8325
8326 /*
8327 * Reserving delalloc space after obtaining the page lock can lead to
8328 * deadlock. For example, if a dirty page is locked by this function
8329 * and the call to btrfs_delalloc_reserve_space() ends up triggering
8330 * dirty page write out, then the btrfs_writepage() function could
8331 * end up waiting indefinitely to get a lock on the page currently
8332 * being processed by btrfs_page_mkwrite() function.
8333 */
8334 ret2 = btrfs_delalloc_reserve_space(BTRFS_I(inode), &data_reserved,
8335 page_start, reserved_space);
8336 if (!ret2) {
8337 ret2 = file_update_time(vmf->vma->vm_file);
8338 reserved = 1;
8339 }
8340 if (ret2) {
8341 ret = vmf_error(ret2);
8342 if (reserved)
8343 goto out;
8344 goto out_noreserve;
8345 }
8346
8347 ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */
8348 again:
8349 lock_page(page);
8350 size = i_size_read(inode);
8351
8352 if ((page->mapping != inode->i_mapping) ||
8353 (page_start >= size)) {
8354 /* page got truncated out from underneath us */
8355 goto out_unlock;
8356 }
8357 wait_on_page_writeback(page);
8358
8359 lock_extent_bits(io_tree, page_start, page_end, &cached_state);
8360 ret = set_page_extent_mapped(page);
> 8361 if (ret < 0)
8362 goto out_unlock;
8363
8364 /*
8365 * we can't set the delalloc bits if there are pending ordered
8366 * extents. Drop our locks and wait for them to finish
8367 */
8368 ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start,
8369 PAGE_SIZE);
8370 if (ordered) {
8371 unlock_extent_cached(io_tree, page_start, page_end,
8372 &cached_state);
8373 unlock_page(page);
8374 btrfs_start_ordered_extent(ordered, 1);
8375 btrfs_put_ordered_extent(ordered);
8376 goto again;
8377 }
8378
8379 if (page->index == ((size - 1) >> PAGE_SHIFT)) {
8380 reserved_space = round_up(size - page_start,
8381 fs_info->sectorsize);
8382 if (reserved_space < PAGE_SIZE) {
8383 end = page_start + reserved_space - 1;
8384 btrfs_delalloc_release_space(BTRFS_I(inode),
8385 data_reserved, page_start,
8386 PAGE_SIZE - reserved_space, true);
8387 }
8388 }
8389
8390 /*
8391 * page_mkwrite gets called when the page is firstly dirtied after it's
8392 * faulted in, but write(2) could also dirty a page and set delalloc
8393 * bits, thus in this case for space account reason, we still need to
8394 * clear any delalloc bits within this page range since we have to
8395 * reserve data&meta space before lock_page() (see above comments).
8396 */
8397 clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, end,
8398 EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING |
8399 EXTENT_DEFRAG, 0, 0, &cached_state);
8400
8401 ret2 = btrfs_set_extent_delalloc(BTRFS_I(inode), page_start, end, 0,
8402 &cached_state);
8403 if (ret2) {
8404 unlock_extent_cached(io_tree, page_start, page_end,
8405 &cached_state);
8406 ret = VM_FAULT_SIGBUS;
8407 goto out_unlock;
8408 }
8409
8410 /* page is wholly or partially inside EOF */
8411 if (page_start + PAGE_SIZE > size)
8412 zero_start = offset_in_page(size);
8413 else
8414 zero_start = PAGE_SIZE;
8415
8416 if (zero_start != PAGE_SIZE) {
8417 kaddr = kmap(page);
8418 memset(kaddr + zero_start, 0, PAGE_SIZE - zero_start);
8419 flush_dcache_page(page);
8420 kunmap(page);
8421 }
8422 ClearPageChecked(page);
8423 set_page_dirty(page);
8424 SetPageUptodate(page);
8425
8426 BTRFS_I(inode)->last_trans = fs_info->generation;
8427 BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid;
8428 BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->root->last_log_commit;
8429
8430 unlock_extent_cached(io_tree, page_start, page_end, &cached_state);
8431
8432 btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
8433 sb_end_pagefault(inode->i_sb);
8434 extent_changeset_free(data_reserved);
8435 return VM_FAULT_LOCKED;
8436
8437 out_unlock:
8438 unlock_page(page);
8439 out:
8440 btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
8441 btrfs_delalloc_release_space(BTRFS_I(inode), data_reserved, page_start,
8442 reserved_space, (ret != 0));
8443 out_noreserve:
8444 sb_end_pagefault(inode->i_sb);
8445 extent_changeset_free(data_reserved);
8446 return ret;
8447 }
8448
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 36210 bytes --]
next prev parent reply other threads:[~2020-12-11 0:45 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-10 6:38 [PATCH v2 00/18] btrfs: add read-only support for subpage sector size Qu Wenruo
2020-12-10 6:38 ` [PATCH v2 01/18] btrfs: extent_io: rename @offset parameter to @disk_bytenr for submit_extent_page() Qu Wenruo
2020-12-17 15:44 ` Josef Bacik
2020-12-10 6:38 ` [PATCH v2 02/18] btrfs: extent_io: refactor __extent_writepage_io() to improve readability Qu Wenruo
2020-12-10 12:12 ` Nikolay Borisov
2020-12-10 12:53 ` Qu Wenruo
2020-12-10 12:58 ` Nikolay Borisov
2020-12-17 15:43 ` Josef Bacik
2020-12-10 6:38 ` [PATCH v2 03/18] btrfs: file: update comment for btrfs_dirty_pages() Qu Wenruo
2020-12-10 12:16 ` Nikolay Borisov
2020-12-10 6:38 ` [PATCH v2 04/18] btrfs: extent_io: introduce a helper to grab an existing extent buffer from a page Qu Wenruo
2020-12-10 13:51 ` Nikolay Borisov
2020-12-17 15:50 ` Josef Bacik
2020-12-10 6:38 ` [PATCH v2 05/18] btrfs: extent_io: introduce the skeleton of btrfs_subpage structure Qu Wenruo
2020-12-17 15:52 ` Josef Bacik
2020-12-10 6:38 ` [PATCH v2 06/18] btrfs: extent_io: make attach_extent_buffer_page() to handle subpage case Qu Wenruo
2020-12-10 15:30 ` Nikolay Borisov
2020-12-17 6:48 ` Qu Wenruo
2020-12-10 16:09 ` Nikolay Borisov
2020-12-17 16:00 ` Josef Bacik
2020-12-18 0:44 ` Qu Wenruo
2020-12-18 15:41 ` Josef Bacik
2020-12-19 0:24 ` Qu Wenruo
2020-12-21 10:15 ` Qu Wenruo
2020-12-10 6:38 ` [PATCH v2 07/18] btrfs: extent_io: make grab_extent_buffer_from_page() " Qu Wenruo
2020-12-10 15:39 ` Nikolay Borisov
2020-12-17 6:55 ` Qu Wenruo
2020-12-17 16:02 ` Josef Bacik
2020-12-18 0:49 ` Qu Wenruo
2020-12-10 6:38 ` [PATCH v2 08/18] btrfs: extent_io: support subpage for extent buffer page release Qu Wenruo
2020-12-10 16:13 ` Nikolay Borisov
2020-12-10 6:38 ` [PATCH v2 09/18] btrfs: subpage: introduce helper for subpage uptodate status Qu Wenruo
2020-12-11 10:10 ` Nikolay Borisov
2020-12-11 10:48 ` Qu Wenruo
2020-12-11 11:41 ` Nikolay Borisov
2020-12-11 11:56 ` Qu Wenruo
2020-12-10 6:38 ` [PATCH v2 10/18] btrfs: subpage: introduce helper for subpage error status Qu Wenruo
2020-12-10 6:38 ` [PATCH v2 11/18] btrfs: extent_io: make set/clear_extent_buffer_uptodate() to support subpage size Qu Wenruo
2020-12-10 6:38 ` [PATCH v2 12/18] btrfs: extent_io: implement try_release_extent_buffer() for subpage metadata support Qu Wenruo
2020-12-11 12:00 ` Nikolay Borisov
2020-12-11 12:11 ` Qu Wenruo
2020-12-11 16:57 ` Nikolay Borisov
2020-12-12 1:28 ` Qu Wenruo
2020-12-12 9:26 ` Nikolay Borisov
2020-12-12 10:26 ` Qu Wenruo
2020-12-12 5:44 ` Qu Wenruo
2020-12-12 10:30 ` Nikolay Borisov
2020-12-12 10:31 ` Qu Wenruo
2020-12-10 6:39 ` [PATCH v2 13/18] btrfs: extent_io: introduce read_extent_buffer_subpage() Qu Wenruo
2020-12-10 6:39 ` [PATCH v2 14/18] btrfs: extent_io: make endio_readpage_update_page_status() to handle subpage case Qu Wenruo
2020-12-14 9:57 ` Nikolay Borisov
2020-12-14 10:46 ` Qu Wenruo
2020-12-10 6:39 ` [PATCH v2 15/18] btrfs: disk-io: introduce subpage metadata validation check Qu Wenruo
2020-12-10 13:24 ` kernel test robot
2020-12-10 13:39 ` kernel test robot
2020-12-14 10:21 ` Nikolay Borisov
2020-12-14 10:50 ` Qu Wenruo
2020-12-14 11:17 ` Nikolay Borisov
2020-12-14 11:32 ` Qu Wenruo
2020-12-14 12:40 ` Nikolay Borisov
2020-12-10 6:39 ` [PATCH v2 16/18] btrfs: introduce btrfs_subpage for data inodes Qu Wenruo
2020-12-10 9:44 ` kernel test robot
2020-12-11 0:43 ` kernel test robot [this message]
2020-12-14 12:46 ` Nikolay Borisov
2020-12-10 6:39 ` [PATCH v2 17/18] btrfs: integrate page status update for read path into begin/end_page_read() Qu Wenruo
2020-12-14 13:59 ` Nikolay Borisov
2020-12-10 6:39 ` [PATCH v2 18/18] btrfs: allow RO mount of 4K sector size fs on 64K page system Qu Wenruo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=202012110846.Zo8Tb8bu-lkp@intel.com \
--to=lkp@intel.com \
--cc=kbuild-all@lists.01.org \
--cc=linux-btrfs@vger.kernel.org \
--cc=wqu@suse.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox