- Modified the psximfs01 test to validate the functionality
- Modified the IMFS_fs_info_t keeping the jnode counter
- Added imfs_statvfs.c which sets the statvfs struct fields for imfs
- Remove the `fsmount_me_t` handler moving its signature to `mount_t`.
- The `mount_t` signature with the `data` lets file systems handle
options and that means the shell `mount` command can mount
more file systems.
- Clean up the `mount` call's handling of the `mount_t` and
`fsmount_me_t` structures.
This is an update of many lingering embedded brains GmbH licenses from
the RTEMS License to the 2-BSD license where possible depending on the
joint copyrights.
Updates #3053.
This is an update of IMD (precursor to embedded brains) licenses from
the RTEMS License to the 2-BSD license where possible depending on the
joint copyrights.
Updates #3053.
The rtime decompression routine does not fully check bounds during the
entirety of the decompression pass and can corrupt memory outside the
decompression buffer if the compressed data is corrupted. This adds the
required check to prevent this failure mode.
Updates #5072
Add filename length checks to prevent the creation of files and folders
that cannot be listed, unlinked or renamed by subsequent calls.
These changes are sufficient to make the jffs2fsrenamelongname and
jffs2nandfsrenamelongname testcases pass.
Updates #5073
Previously the first cluster of the file was not updated when
truncating to 0 but was in any case freed.
Now the first cluster of the file is updated during the truncation,
leaving the directory entry in a consistent state.
Closes#2757
rename now actively changes the inode of the file being renamed.
This change has been done to fix an issue when the file being renamed is already opened before the rename, in particular failing to update the inode number (and the two lookup tables) could cause at least two problems:
1) A new file created in the same source folder of the file being renamed may get the same fat file descriptor of the renamed file
2) An open to the renamed file opens a new fat file descriptor, thus putting out of sync the operations on the file (two distinct fat file descriptor working on the same clusters and entries)
Previously the inode was in any case changed once the file was closed.
Another possible solution would have been to separate the inode number from the lookup table, but this solution would have required much more memory because it needed to keep track of all files seen in the file system up to the last operation and not only the opened ones.
Closes#5044
When unmounting a JFFS2 filesystem, any outstanding write buffers must
be flushed to disk. In some circumstances, these write buffers are
instantiated by a garbage collection pass and as such no inode number is
associated with it. Due to the way that JFFS2 processes these garbage
collection passes, a write buffer without any associated inodes will not
be flushed unless it is forced with jffs2_flush_wbuf_pad().
On JFFS2 file systems on NOR flash or dataflash that does not have spare
area for metadata and thus does not invoke delayed writes, it is
possible to put the file system into a state where all blocks have been
written to and all files have been deleted from the filesystem. There is
a bug in the JFFS2 file system scan routine that detects this situation
as a corrupted file system since the scan routine relies on "used" space
to discriminate a valid file system when there are blocks that need to
be erased. The correct fix would require a partial rewrite of the scan
routine, so instead this patch tracks the space marked as obsolete along
with space at the end of each block that is otherwise too small to
contain a JFFS2 node so that it can me compared with the dirty space.
Corrupted data (or otherwise non-JFFS2 blocks) will still cause this
check to fail as corrupted data isn't recognized as obsoleted (deleted)
nodes.
When encountering an error during filesystem creation, fat_fd must be
cleaned up appropriately once the file is opened. There was an
opportunity for a resource leak due to jumping to the incorrect error
handling label.
Revert JFFS2 to non-granular locking. This makes the superblock
available to the delayed work that it owns so that delayed work
processing can ensure that shared resources are being exclusively
accessed. This change only affects NAND systems. Moving forward with
granular locking will require a significant investment of time in
producing a test harness that doesn't require hardware such that this
can get a reasonable amount of test coverage.
This provides a non-noop implementation of spinlocks for JFFS2 in terms
of RTEMS mutexes. POSIX spinlocks are not appropriate for the types of
actions that occur during a held JFFS2 spinlock and can cause bad
dispatch level conditions.
This reworks the JFFS2 delayed work queue to avoid use of
on-chain/off-chain semantics since they vary in behavior under
RTEMS_DEBUG and are not guaranteed to be safe to use in SMP systems.
This adds all delayed work structs to the chain on FS init and does not
remove them until umount.
Updates #4956
This moves delayed work to a temporary chain to prevent a locking
inversion between the delayed work lock and the alloc_sem lock. Delayed
work is now processed after the delayed work lock is released. Locking
order is any JFFS2 locks before the delayed work lock.
This adds write buffer and bad block support required for JFFS2
operation on NAND devices. This also adds the minor modifications
necessary for RTEMS support in the Linux header stubs and in wbuf.c.
Memory and NOR backed applications should experience no difference in
operation since they do not expose the callbacks required for write
buffer support.
The RTEMS and Linux JFFS2 compression support have a slightly different
interface. This fixes a bug introduced by:
commit 17c3d536ca
Author: Yang Yang <yang.yang29@zte.com.cn>
Date: Thu Jan 28 02:55:35 2021 -0800
jffs2: check the validity of dstlen in jffs2_zlib_compress()
KASAN reports a BUG when download file in jffs2 filesystem.It is
because when dstlen == 1, cpage_out will write array out of bounds.
Actually, data will not be compressed in jffs2_zlib_compress() if
data's length less than 4.
There is a regular need in the kernel to provide a way to declare
having a dynamically sized set of trailing elements in a structure.
Kernel code should always use “flexible array members”[1] for these
cases. The older style of one-element or zero-length arrays should
no longer be used[2].
This code was transformed with the help of Coccinelle:
(linux-5.19-rc2$ spatch --jobs $(getconf _NPROCESSORS_ONLN) --sp-file script.cocci --include-headers --dir . > output.patch)
@@
identifier S, member, array;
type T1, T2;
@@
struct S {
...
T1 member;
T2 array[
- 0
];
};
-fstrict-flex-arrays=3 is coming and we need to land these changes
to prevent issues like these in the short future:
../fs/minix/dir.c:337:3: warning: 'strcpy' will always overflow; destination buffer has size 0,
but the source string has length 2 (including NUL byte) [-Wfortify-source]
strcpy(de3->name, ".");
^
Since these are all [0] to [] changes, the risk to UAPI is nearly zero. If
this breaks anything, we can use a union with a new member name.
[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://www.kernel.org/doc/html/v5.16/process/deprecated.html#zero-length-and-one-element-arrays
Link: https://github.com/KSPP/linux/issues/78
Build-tested-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/lkml/62b675ec.wKX6AOZ6cbE71vtF%25lkp@intel.com/
Acked-by: Dan Williams <dan.j.williams@intel.com> # For ndctl.h
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Use kzalloc rather than duplicating its implementation, which
makes code simple and easy to understand.
Signed-off-by: Haowen Bai <baihaowen@meizu.com>
[rw: Fixed printk string]
Signed-off-by: Richard Weinberger <richard@nod.at>
By making filler_t the same as read_folio, we can use the same function
for both in gfs2. We can push the use of folios down one more level
in jffs2 and nfs. We also increase type safety for future users of the
various read_cache_page() family of functions by forcing the parameter
to be a pointer to struct file (or NULL).
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
In preparation for unifying the read_cache_page() and read_folio()
implementations, make jffs2_do_readpage_unlock() get the inode
from the page instead of passing it in from read_cache_page().
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Joel Sherrill
Sepehr Ganji
Bhavya Shah
Chris Johns
Amar Takhar
Gedare Bloom
Kinsey Moore
yang.zhang
Loris Nardo
Sebastian Huber
Gustavo A. R. Silva
Haowen Bai
Matthew Wilcox (Oracle)
Baokun Li