Changes in ext2.h

eab61d32 - ext2: convert to the new mount API

Sun, 23 Feb 2025 19:57:40 +0000

ext2: convert to the new mount APIConvert ext2 to the new mount API.Note that this makes the sb= option more accepting than it was before;previosly, sb= was only accepted if it was the first specified option.Now it can exist anywhere, and if respecified, the last specified valueis used.Parse-time messages here are sent to ext2_msg with a NULL sb, andext2_msg is adjusted to accept that, as ext4 does today as well.Signed-off-by: Eric Sandeen Signed-off-by: Jan Kara Link: https://patch.msgid.link/20250223201014.7541-2-sandeen@redhat.com List of files: /linux-6.15/fs/ext2/ext2.h

ccb49011 - quota: Properly annotate i_dquot arrays with __rcu

Tue, 06 Feb 2024 14:08:19 +0000

quota: Properly annotate i_dquot arrays with __rcuDquots pointed to from i_dquot arrays in inodes are protected bydquot_srcu. Annotate them as such and change .get_dquots callback toreturn properly annotated pointer to make sparse happy.Fixes: b9ba6f94b238 ("quota: remove dqptr_sem")Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

c2d20492 - ext2: Convert ext2_unlink() and ext2_rename() to use folios

Thu, 21 Sep 2023 20:07:45 +0000

ext2: Convert ext2_unlink() and ext2_rename() to use foliosThis involves changing ext2_find_entry(), ext2_dotdot(),ext2_inode_by_name(), ext2_set_link() and ext2_delete_entry() totake a folio. These were also the last users of ext2_get_page() andext2_put_page(), so remove those at the same time.Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Jan Kara Message-Id: <20230921200746.3303942-8-willy@infradead.org> List of files: /linux-6.15/fs/ext2/ext2.h

2ebc736c - ext2: improve consistency of ext2_fsblk_t datatype usage

Thu, 17 Aug 2023 19:59:25 +0000

ext2: improve consistency of ext2_fsblk_t datatype usageThe ext2 block allocation/deallocation functions and their respectivecalls use a mixture of unsigned long and ext2_fsblk_t datatypes toindex the desired ext2 block. This commit replaces occurrences ofunsigned long with ext2_fsblk_t, covering the functionsext2_new_block(), ext2_new_blocks(), ext2_free_blocks(),ext2_free_data() and ext2_free_branches(). This commit is ratherconservative, and only replaces unsigned long with ext2_fsblk_t ifthe variable is used to index a specific ext2 block.Signed-off-by: Georg Ottinger Signed-off-by: Jan Kara Message-Id: <20230817195925.10268-1-g.ottinger@gmx.at> List of files: /linux-6.15/fs/ext2/ext2.h

b450159d - ext2: introduce new flags argument for ext2_new_blocks()

Tue, 15 Aug 2023 11:26:10 +0000

ext2: introduce new flags argument for ext2_new_blocks()This patch introduces a new flags argument for ext2_new_blocks() and alsoa new EXT2_ALLOC_NORESERVE flag.Signed-off-by: Ye Bin Message-Id: <20230815112612.221145-3-yebin10@huawei.com>Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

2445a8a1 - ext2: remove ext2_new_block()

Tue, 15 Aug 2023 11:26:09 +0000

ext2: remove ext2_new_block()Now, only xattr allocate block use ext2_new_block(), so just opencode it inthe xattr code.Signed-off-by: Ye Bin Message-Id: <20230815112612.221145-2-yebin10@huawei.com>Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

404615d7 - ext2: Drop fragment support

Tue, 13 Jun 2023 10:25:52 +0000

ext2: Drop fragment supportExt2 has fields in superblock reserved for subblock allocation support.However that never landed. Drop the many years dead code.Reported-by: syzbot+af5e10f73dbff48f70af@syzkaller.appspotmail.comSigned-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

b8b9e8b3 - ext2_find_entry()/ext2_dotdot(): callers don't need page_addr anymore

Wed, 14 Dec 2022 01:53:47 +0000

ext2_find_entry()/ext2_dotdot(): callers don't need page_addr anymore... and that's how it should've been done in the first placeSigned-off-by: Al Viro Reviewed-by: Fabio M. De Francesco Tested-by: Fabio M. De Francesco Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

dae42837 - ext2_{set_link,delete_entry}(): don't bother with page_addr

Wed, 14 Dec 2022 01:31:39 +0000

ext2_{set_link,delete_entry}(): don't bother with page_addrext2_set_link() simply doesn't use it anymore and ext2_delete_entry()can easily obtain it from the directory entry pointer...Signed-off-by: Al Viro Reviewed-by: Fabio M. De Francesco Tested-by: Fabio M. De Francesco Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

fb5de435 - ext2: Move direct-io to use iomap

Fri, 21 Apr 2023 09:46:15 +0000

ext2: Move direct-io to use iomapThis patch converts ext2 direct-io path to iomap interface.- This also takes care of DIO_SKIP_HOLES part in which we return -ENOTBLK from ext2_iomap_begin(), in case if the write is done on a hole.- This fallbacks to buffered-io in case of DIO_SKIP_HOLES or in case of a partial write or if any error is detected in ext2_iomap_end(). We try to return -ENOTBLK in such cases.- For any unaligned or extending DIO writes, we pass IOMAP_DIO_FORCE_WAIT flag to ensure synchronous writes.- For extending writes we set IOMAP_F_DIRTY in ext2_iomap_begin because otherwise with dsync writes on devices that support FUA, generic_write_sync won't be called and we might miss inode metadata updates.- Since ext2 already now uses _nolock vartiant of sync write. Hence there is no inode lock problem with iomap in this patch.- ext2_iomap_ops are now being shared by DIO, DAX & fiemap pathTested-by: Disha Goel Reviewed-by: Christoph Hellwig Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Jan Kara Message-Id: <610b672a52f2a7ff6dc550fd14d0f995806232a5.1682069716.git.ritesh.list@gmail.com> List of files: /linux-6.15/fs/ext2/ext2.h

62aeb944 - ext2: Check block size validity during mount

Wed, 01 Mar 2023 10:59:39 +0000

ext2: Check block size validity during mountCheck that log of block size stored in the superblock has sensiblevalue. Otherwise the shift computing the block size can overflow leadingto undefined behavior.Reported-by: syzbot+4fec412f59eba8c01b77@syzkaller.appspotmail.comSigned-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

e9cd1d9a - ext2: Correct maximum ext2 filesystem block size

Wed, 01 Mar 2023 11:02:08 +0000

ext2: Correct maximum ext2 filesystem block sizeExt2 has traditionally supported filesystem block sizes upto page sizeor upto 65536. Macro EXT2_MAX_BLOCK_SIZE is set to 4096, however that isnever used in ext2 so practically we always allowed whateversb_set_blocksize() accepted. Fix value of EXT2_MAX_BLOCK_SIZE because itwill be used in the next patch.Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

8782a9ae - fs: port ->fileattr_set() to pass mnt_idmap

Fri, 13 Jan 2023 11:49:21 +0000

fs: port ->fileattr_set() to pass mnt_idmapConvert to struct mnt_idmap.Last cycle we merged the necessary infrastructure in256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").This is just the conversion to struct mnt_idmap.Currently we still pass around the plain namespace that was attached to amount. This is in general pretty convenient but it makes it easy toconflate namespaces that are relevant on the filesystem with namespacesthat are relevent on the mount level. Especially for non-vfs developerswithout detailed knowledge in this area this can be a potential source forbugs.Once the conversion to struct mnt_idmap is done all helpers down to thereally low-level helpers will take a struct mnt_idmap argument instead oftwo namespace arguments. This way it becomes impossible to conflate the twoeliminating the possibility of any bugs. All of the vfs and all filesystemsonly operate on struct mnt_idmap.Acked-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Christian Brauner (Microsoft) List of files: /linux-6.15/fs/ext2/ext2.h

b74d24f7 - fs: port ->getattr() to pass mnt_idmap

Fri, 13 Jan 2023 11:49:12 +0000

fs: port ->getattr() to pass mnt_idmapConvert to struct mnt_idmap.Last cycle we merged the necessary infrastructure in256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").This is just the conversion to struct mnt_idmap.Currently we still pass around the plain namespace that was attached to amount. This is in general pretty convenient but it makes it easy toconflate namespaces that are relevant on the filesystem with namespacesthat are relevent on the mount level. Especially for non-vfs developerswithout detailed knowledge in this area this can be a potential source forbugs.Once the conversion to struct mnt_idmap is done all helpers down to thereally low-level helpers will take a struct mnt_idmap argument instead oftwo namespace arguments. This way it becomes impossible to conflate the twoeliminating the possibility of any bugs. All of the vfs and all filesystemsonly operate on struct mnt_idmap.Acked-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Christian Brauner (Microsoft) List of files: /linux-6.15/fs/ext2/ext2.h

c1632a0f - fs: port ->setattr() to pass mnt_idmap

Fri, 13 Jan 2023 11:49:11 +0000

fs: port ->setattr() to pass mnt_idmapConvert to struct mnt_idmap.Last cycle we merged the necessary infrastructure in256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").This is just the conversion to struct mnt_idmap.Currently we still pass around the plain namespace that was attached to amount. This is in general pretty convenient but it makes it easy toconflate namespaces that are relevant on the filesystem with namespacesthat are relevent on the mount level. Especially for non-vfs developerswithout detailed knowledge in this area this can be a potential source forbugs.Once the conversion to struct mnt_idmap is done all helpers down to thereally low-level helpers will take a struct mnt_idmap argument instead oftwo namespace arguments. This way it becomes impossible to conflate the twoeliminating the possibility of any bugs. All of the vfs and all filesystemsonly operate on struct mnt_idmap.Acked-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Christian Brauner (Microsoft) List of files: /linux-6.15/fs/ext2/ext2.h

7a5fa171 - ext2: propagate errors from ext2_prepare_chunk

Mon, 16 Jan 2023 08:52:05 +0000

ext2: propagate errors from ext2_prepare_chunkPropagate errors from ext2_prepare_chunk to the callers and handle themthere. While touching the prototype also turn update_times into a boolfrom the current int used as bool.[JK: fixed up error recovery path in ext2_rename()]Signed-off-by: Christoph Hellwig Signed-off-by: Jan Kara Message-Id: <20230116085205.2342975-1-hch@lst.de> List of files: /linux-6.15/fs/ext2/ext2.h

0cc5b4ce - ext2: remove nobh support

Mon, 13 Jun 2022 05:37:11 +0000

ext2: remove nobh supportThe nobh mode is an obscure feature to save lowlevel for large memory32-bit configurations while trading for much slower performance andhas been long obsolete. Remove it.Signed-off-by: Christoph Hellwig Reviewed-by: Jan Kara Signed-off-by: Matthew Wilcox (Oracle) List of files: /linux-6.15/fs/ext2/ext2.h

cd913c76 - dax: return the partition offset from fs_dax_get_by_bdev

Mon, 29 Nov 2021 10:21:59 +0000

dax: return the partition offset from fs_dax_get_by_bdevPrepare for the removal of the block_device from the DAX I/O path byreturning the partition offset from fs_dax_get_by_bdev so that the filesystems have it at hand for use during I/O.Signed-off-by: Christoph Hellwig Reviewed-by: Dan Williams Reviewed-by: Darrick J. Wong Link: https://lore.kernel.org/r/20211129102203.2243509-26-hch@lst.deSigned-off-by: Dan Williams List of files: /linux-6.15/fs/ext2/ext2.h

728d392f - fs/ext2: Avoid page_address on pages returned by ext2_get_page

Wed, 14 Jul 2021 16:54:48 +0000

fs/ext2: Avoid page_address on pages returned by ext2_get_pageCommit 782b76d7abdf02b12c46ed6f1e9bf715569027f7 ("fs/ext2: Replacekmap() with kmap_local_page()") replaced the kmap/kunmap calls inext2_get_page/ext2_put_page with kmap_local_page/kunmap_local forefficiency reasons. As a necessary side change, the commit alsomade ext2_get_page (and ext2_find_entry and ext2_dotdot) returnthe mapping address along with the page itself, as it is requiredfor kunmap_local, and converted uses of page_address on such pagesto use the newly returned address instead. However, uses ofpage_address on such pages were missed in ext2_check_page andext2_delete_entry, which triggers oopses if kmap_local_page happensto return an address from high memory. Fix this now by convertingthe remaining uses of page_address to use the right address, asreturned by kmap_local_page.Link: https://lore.kernel.org/r/20210714185448.8707ac239e9f12b3a7f5b9f9@urjc.esReviewed-by: Ira Weiny Signed-off-by: Javier Pello Fixes: 782b76d7abdf ("fs/ext2: Replace kmap() with kmap_local_page()")Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h

70f3bad8 - ext2: Convert to using invalidate_lock

Mon, 12 Apr 2021 16:43:11 +0000

ext2: Convert to using invalidate_lockExt2 has its private dax_sem used for synchronizing page faults andtruncation. Use mapping->invalidate_lock instead as it is meant for thispurpose.CC: Reviewed-by: Christoph Hellwig Signed-off-by: Jan Kara List of files: /linux-6.15/fs/ext2/ext2.h