linux-stable.git/fs/ext2, branch v6.10

ext2: Remove LEGACY_DIRECT_IO dependency

2024-05-03T09:50:28+00:00

commit fb5de4358e1a ("ext2: Move direct-io to use iomap"), converted
ext2 direct-io to iomap which killed the call to blockdev_direct_IO().
So let's remove LEGACY_DIRECT_IO config dependency from ext2 Kconfig.

Signed-off-by: Ritesh Harjani (IBM) 
Signed-off-by: Jan Kara 
Message-Id:

ext2: Remove call to folio_set_error()

2024-04-23T21:22:04+00:00

Nobody checks this flag on ext2 folios, stop setting it.

Cc: Jan Kara 
Cc: linux-ext4@vger.kernel.org
Signed-off-by: Matthew Wilcox (Oracle) 
Signed-off-by: Jan Kara 
Message-Id: <20240420025029.2166544-10-willy@infradead.org>

ext2: set FMODE_CAN_ODIRECT instead of a dummy direct_IO method

2024-04-04T14:27:54+00:00

Since commit a2ad63daa88b ("VFS: add FMODE_CAN_ODIRECT file flag") file
systems can just set the FMODE_CAN_ODIRECT flag at open time instead of
wiring up a dummy direct_IO method to indicate support for direct I/O.

Signed-off-by: Ritesh Harjani (IBM) 
Signed-off-by: Jan Kara 
Message-Id: <94f78492f55c3f421359fb6e0d8fab6e79ea17b2.1709215665.git.ritesh.list@gmail.com>

Merge tag 'fs_for_v6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

2024-03-13T21:30:58+00:00

Pull ext2, isofs, udf, and quota updates from Jan Kara:
 "A lot of material this time:

   - removal of a lot of GFP_NOFS usage from ext2, udf, quota (either it
     was legacy or replaced with scoped memalloc_nofs_*() API)

   - removal of BUG_ONs in quota code

   - conversion of UDF to the new mount API

   - tightening quota on disk format verification

   - fix some potentially unsafe use of RCU pointers in quota code and
     annotate everything properly to make sparse happy

   - a few other small quota, ext2, udf, and isofs fixes"

* tag 'fs_for_v6.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs: (26 commits)
  udf: remove SLAB_MEM_SPREAD flag usage
  quota: remove SLAB_MEM_SPREAD flag usage
  isofs: remove SLAB_MEM_SPREAD flag usage
  ext2: remove SLAB_MEM_SPREAD flag usage
  ext2: mark as deprecated
  udf: convert to new mount API
  udf: convert novrs to an option flag
  MAINTAINERS: add missing git address for ext2 entry
  quota: Detect loops in quota tree
  quota: Properly annotate i_dquot arrays with __rcu
  quota: Fix rcu annotations of inode dquot pointers
  isofs: handle CDs with bad root inode but good Joliet root directory
  udf: Avoid invalid LVID used on mount
  quota: Fix potential NULL pointer dereference
  quota: Drop GFP_NOFS instances under dquot->dq_lock and dqio_sem
  quota: Set nofs allocation context when acquiring dqio_sem
  ext2: Remove GFP_NOFS use in ext2_xattr_cache_insert()
  ext2: Drop GFP_NOFS use in ext2_get_blocks()
  ext2: Drop GFP_NOFS allocation from ext2_init_block_alloc_info()
  udf: Remove GFP_NOFS allocation in udf_expand_file_adinicb()
  ...

mm, slab: remove last vestiges of SLAB_MEM_SPREAD

2024-03-13T03:32:19+00:00

Yes, yes, I know the slab people were planning on going slow and letting
every subsystem fight this thing on their own.  But let's just rip off
the band-aid and get it over and done with.  I don't want to see a
number of unnecessary pull requests just to get rid of a flag that no
longer has any meaning.

This was mainly done with a couple of 'sed' scripts and then some manual
cleanup of the end result.

Link: https://lore.kernel.org/all/CAHk-=wji0u+OOtmAOD-5JV3SXcRJF___k_+8XNKmak0yd5vW1Q@mail.gmail.com/
Signed-off-by: Linus Torvalds

ext2: remove SLAB_MEM_SPREAD flag usage

2024-03-05T14:09:41+00:00

The SLAB_MEM_SPREAD flag is already a no-op after removal of SLAB
allocator and in [1] it was fully deprecated. Remove its usage so we can
delete it from slab. No functional change.

Signed-off-by: Chengming Zhou 
Signed-off-by: Jan Kara 
Link: https://lore.kernel.org/all/20240223-slab-cleanup-flags-v2-1-02f1753e8303@suse.cz/
Message-Id: <20240224134816.829424-1-chengming.zhou@linux.dev>

ext2: mark as deprecated

2024-02-22T10:14:42+00:00

Add a DEPRECATED keyword to the kernel parameter description,
to warn users that this filesystem doesn't support dates beyond 2038.

Signed-off-by: Michael Opdenacker 
Signed-off-by: Jan Kara 
Message-Id: <20240222095001.137660-1-michael.opdenacker@bootlin.com>

quota: Properly annotate i_dquot arrays with __rcu

2024-02-08T11:04:59+00:00

Dquots pointed to from i_dquot arrays in inodes are protected by
dquot_srcu. Annotate them as such and change .get_dquots callback to
return properly annotated pointer to make sparse happy.

Fixes: b9ba6f94b238 ("quota: remove dqptr_sem")
Signed-off-by: Jan Kara

ext2: Remove GFP_NOFS use in ext2_xattr_cache_insert()

2024-01-23T18:21:11+00:00

ext2_xattr_cache_insert() calls mb_cache_entry_create() with GFP_NOFS
because it is called under EXT2_I(inode)->xattr_sem. However xattr_sem
or any higher ranking lock is not acquired on fs reclaim path for ext2
at least since we don't do page writeback from direct reclaim.  Thus
GFP_NOFS is not needed.

Signed-off-by: Jan Kara

ext2: Drop GFP_NOFS use in ext2_get_blocks()

2024-01-23T18:21:10+00:00

ext2_get_blocks() calls sb_issue_zeroout() with GFP_NOFS flag. However
the call is performed under inode->i_rwsem and
EXT2_I(inode)->i_truncate_mutex neither of which is acquired during
direct fs reclaim. So it is safe to change the gfp mask to GFP_KERNEL.

Signed-off-by: Jan Kara