SquashFS supports sprase blocks in files - that is, if a given block is
composed only of zeros, it's not written to the output file to save
space and it's on-disk length field is set to zero to indicate that.
Previously the squashfs driver did not recognise that, and would attempt
to read and decompress a zero-sized block, which obviously failed.
The following command may be used to create a file for testing:
cat <(dd if=/dev/urandom of=/dev/stdout bs=1M count=1) \
<(dd if=/dev/zero of=/dev/stdout bs=1M count=1) \
<(dd if=/dev/urandom of=/dev/stdout bs=1k count=200) >test_file
Signed-off-by: Campbell Suter <campbell@snapit.group>
In __btrfs_map_block() we do a int * int and assign it to u64.
This is not safe as the result (int * int) is still evaluated as (int)
thus it can overflow.
Convert one of the multiplier to u64 to prevent such problem.
In real world, this should not cause problem as we have device number
limit thus it won't go beyond 4G for a single stripe.
But it's harder to teach coverity about all these hidden limits, so just
fix the possible overflow.
Reported-by: Coverity CID 312957
Reported-by: Coverity CID 312948
Signed-off-by: Qu Wenruo <wqu@suse.com>
Up to now the bootefi command used the last file loaded to determine the
boot partition. This has led to errors when the fdt had been loaded from
another partition after the EFI binary.
Before setting the boot device from a loaded file check if it is a PE-COFF
image or a FIT image.
For a PE-COFF image remember address and size, boot device and path.
For a FIT image remember boot device and path.
If the PE-COFF image is overwritten by loading another file, forget it.
Do not allow to start an image via bootefi which is not the last loaded
PE-COFF image.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The FAT file system does not have a UUID but a 4 byte volume ID.
Let the fsuuid command show it in XXXX-XXXX format.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The FAT filesystem implementation uses several marcros referring to a magic
variable name mydata which renders the code less readable. Eliminate one of
them which is only used for a debug() statement.
Use log_debug() instead of debug().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Long file names are stored in multiple directory entries. When deleting a
file we must delete all of them.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
A long name is split over multiple directory entries. When deleting a file
with a long name we need the first directory entry to be able to delete the
whole chain.
Add the necessary fields to the FAT iterator:
* cluster of first directory entry
* address of first directory entry
* remaining entries in cluster
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When deleting a directory entry 0xe5 is written to name[0].
We have a constant for this value and should use it consistently.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Searching for a file is not a write operation. So it should not lead to the
allocation of a new cluster to the directory.
If we reuse deleted entries, we might not even use the new cluster and due
to not flushing it the directory could be corrupted.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When creating new directory entries try to reuse entries marked as deleted.
In fill_dir_slot() do not allocate new clusters as this has already been
done in fat_find_empty_dentries().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Provide a function to find a series of empty directory entries.
The current directory is scanned for deleted entries.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When handling long file names directory entries may be split over multiple
clusters. We must make sure that new clusters are zero filled on disk.
When allocating a new cluster for a directory flush it.
The flushing should be executed before updating the FAT. This way if
flushing fails, we still have a valid directory structure.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
When iterating over a child directory we set itr->start_clust.
Do the same when over the root directory.
When looking for deleted directory entries or existing short names we will
have to iterate over directories a second and third time. With this patch
we do not need any special logic for the root directory.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
For reusing deleted directory entries we have to adjust the function called
to step to the next directory entry.
This patch alone is not enough to actually reuse deleted directory entries
as the fill_dir_slot() is still called with first never used directory
entry.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
File names must be unique within their directory. So before assigning a
short name we must check that it is unique.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
In set_name() we select the short name. Once this is correctly implemented
this will be a performance intensive operation because we need to check
that the name does not exist yet. So set_name should only be called once.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Currently we pass the short name via the directory iterator.
Pass it explicitly as a parameter.
This removes the requirement to set the short name in the iterator before
writing the long name.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The current function set_name() used to create short names has the
following deficiencies resolved by this patch:
* Long names (e.g. FOO.TXT) are stored even if a short name is enough.
* Short names with spaces are created, e.g. "A ~1.TXT".
* Short names with illegal characters are created, e.g. "FOO++BAR".
* Debug output does not not consider that the short file name has no
concluding '\0'.
The solution for the following bug is split of into a separate patch:
* Short file names must be unique.
This patch only provides the loop over possible short file names.
Fixes: c30a15e590 ("FAT: Add FAT write feature")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Rename function next_cluster() to fat_next_cluster() and export it.
When creating a new directory entries we should reuse deleted entries.
This requires re-scanning the directory.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
The FAT specification [1] requires that for a '..' directory entry pointing
to the root directory the fields DIR_FstClusHi and DIR_FstClusLo are 0.
[1] Microsoft FAT Specification, Microsoft Corporation, August 30 2005
Fixes: 31a18d570d ("fs: fat: support mkdir")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Using constants instead of anonymous numbers increases code readability.
Fixes: 704df6aa0a ("fs: fat: refactor write interface for a file offset")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
0x05 is used as replacement letter for 0xe5 at the first position of short
file names. We must not skip over directory entries starting with 0x05.
Cf. Microsoft FAT Specification, August 30 2005
Fixes: 39606d462c ("fs: fat: handle deleted directory entries correctly")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When trying to create a file in the full root directory of a FAT32
filesystem a NULL dereference can be observed.
When the root directory of a FAT16 filesystem is full fill_dir_slot() must
return -1 to signal that a new directory entry could not be allocated.
Fixes: cd2d727fff ("fs: fat: allocate a new cluster for root directory of fat32")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
This permits to find a file and use the distro_bootcmd
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
offset is the offset in the file read, not the offset in the destination
buffer.
If the offset is not null, this will lead to a memory corruption.
So, for now, we are returning an error if the offset is used.
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
sqfs_decompressor_init() returns a value, so it's better to use it than
to force the return value to EINVAL (it could be ENOMEM)
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
Resetting the context on error will prevent some checks like:
if (!ctx.cur_dev)
To pass when the probe method has failed
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
If SquashFS magic number is invalid, there's a memory leak.
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
finfo.blk_sizes may not be freed in case of error in the for loop
Setting it to null and freeing it at the end makes prevents that from
happening.
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
if sqfs_tokenize(rel_tokens, rc, rel); fails, the function exits
without freeing the array base_tokens.
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
the return value of sqfs_tokenize(rel_tokens, rc, rel); wasn't checked.
(but "ret" value was !)
This is obviouly a typo.
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
sqfs_closedir() should be called to free memory allocated by
sqfs_opendir()
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
memory allocation should always be checked
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
path, target, res, rem and sym_tokens were not free on error nor success.
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
Using only one label permits to prevents bugs when moving code around.
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
add missing squashfs function to prevent dangling or null pointers.
For exemple, when calling test [ -e somefile ], squashfs.exists may be
called.
Signed-off-by: Richard Genoud <richard.genoud@posteo.net>
Reviewed-by: Joao Marcos Costa <jmcosta944@gmail.com>
In show_dir() if we hit a ROOT_ITEM, we can exit with uninitialized
@ret.
Fix it by initializing it to 0.
Reported-by: Coverity CID 312955
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
In btrfs_lookup_path() the local variable @type should always be updated
after we hit any file/dir.
But if @filename is NULL from the very beginning, then we don't
initialize it and return it directly.
To prevent such problem from happening, we initialize @type to
BTRFS_FT_UNKNOWN.
For normal execution route, it will get updated for each filename we
resolved.
Buf if we didn't find any path, we check if the type is still FT_UNKNOWN
and ret == 0. If true we know there is something wrong, just return
-EUCLEAN to inform the caller.
Reported-by: Coverity CID 312958
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
I've created a squashfs file system with Yocto (it use squashfs-tools)
and u-boot command sqfsls give the error:'Error while searching inode:
unknown type.'
After some digging in the code I found that the index is off by 1.
This patch fix this issue and I can successful use the sqfsls command.
After search for the squashfs format I found a link talk about a
similar issue but this time in the documentation. The link is:
https://github.com/AgentD/squashfs-tools-ng/commit/e6588526838caece9529
Signed-off-by: Gerard Koskamp <gerard.koskamp@nedap.com>
Tested-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
We should check if the incoming parameter file_mapping is not NULL instead
of checking after adding an offset.
Reported-by: Coverity CID 307210
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
This cleans up the now unneeded code from the old btrfs implementation.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This patch introduces a new function, list_one_subvol(), which will
resolve the path to FS_TREE of one subvolume.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This patch introduces a new function, get_path_in_subvolume(), which
resolves inode number into path inside a subvolume.
This function will be later used for btrfs subvolume list functionality.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This version of btrfs_file_read() has the following new features:
- Tries all mirrors
- More handling on unaligned size
- Better compressed extent handling
The old implementation doesn't handle compressed extent with offset
properly: we need to read out the whole compressed extent, then
decompress the whole extent, and only then copy the requested part.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This implements lookup_data_extent() function for the incoming
new implementation of btrfs_file_read().
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
These two functions are used to do sector aligned read, which will be
later used to implement btrfs_file_read().
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Rename btrfs_file_read() and its callees to avoid name conflicts with
the incoming new code.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
After this the only remaining function that still utilizes
__btrfs_lookup_path() is btrfs_read().
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Use extent buffer based infrastructure to re-implement btrfs_readdir().
Along this rework, some small corner cases fixed:
- Subvolume tree mtime
Mtime of a subvolume tree is recorded in its root item, since there is
no INODE_ITEM for it.
This needs extra search from tree root.
- Output the unknown type
If the DIR_ITEM is corrupted, at least don't try to access the memory
out of boundary.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This is the extent buffer based path lookup routine.
To implement this, btrfs_lookup_dir_item() is crossported from
btrfs-progs, and implements btrfs_lookup_path() from scratch.
Unlike the existing __btrfs_lookup_path(), since btrfs_read_fs_root()
will check whether a root is a orphan at read time, there is no need to
check root backref, this makes the code a little easier to read.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
All existing next_length() caller handles return value > BTRFS_NAME_LEN,
so there is no need to do BTRFS_NAME_LEN check in next_length().
But still, we want to exit early if we're beyond BTRFS_NAME_LEN, so this
patch makes next_length() exit as soon as we're beyond BTRFS_NAME_LEN.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
The existing __btrfs_readlink() can be easily re-implemented using the
extent buffer based btrfs_readlink().
This is the first step to re-implement U-Boot's btrfs code.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Since the old code is using __btrfs_path/__btrfs_root which is different
from the regular extent buffer based one, we add "__" prefix for the old
implementation to avoid name conflicts for the incoming crossport.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
open_ctree_fs_info() is the main entry point to open btrfs.
This version is a simplfied version of __open_ctree_fd() of btrfs-progs,
the main differences are:
- Parameters on how to specify a block device
Instead of @fd and @path, U-Boot uses blk_desc and disk_partition_t.
- Remove open_ctree flags
There won't be multiple open ctree modes in U-Boot.
Otherwise functions structures are all kept the same.
With open_ctree_fs_info() implemented, also introduce the global
current_fs_info pointer to show the current opened btrfs.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
These two functions play a big role in btrfs bootstrap.
The following function is removed:
- Seed device support
Although in theory we can still support multiple devices, we don't have
a facility in U-Boot to do device scan without opening them.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This patch copies the core function, btrfs_search_slot(), from
btrfs-progs.
This version has the following functionality removed:
- The ability to COW tree block
Related code is commented out, and can be enabled in the future.
- The readahead functionality
This is abused in kernel. Remove it completely.
With the core function in place, btrfs developers should feel at home now.
This also crossports supporting code like btrfs_previous_item() to
ctree.[ch].
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Crossport struct btrfs_root to ctree.h from btrfs-progs, with write
related members deleted.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This is to avoid naming conflicts between extent buffer based
btrfs_root.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
To avoid name conflicting between the extent buffer based btrfs_path
from btrfs-progs, rename struct btrfs_path to struct __btrfs_path.
Also rename btrfs_free_path() to __btrfs_free_path() to avoid conflicts.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This is the one of the basic stone function for btrfs, which:
- Resolves the chunk mappings
- Reads data from disk
- Does various sanity check
With read_tree_block(), we can finally crossport needed btrfs btree
operations to U-Boot.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This patch crossports volumes.[ch] from btrfs-progs, including:
- btrfs_map_block()
The core mechanism to map btrfs logical address to physical address.
This version includes multi-device support, along with RAID56 support.
- btrfs_scan_one_device()
This is the function to register one btrfs device to the list.
This is the main part of the multi-device btrfs assembling process.
Although we're not going to support multiple devices until U-Boot
allows us to scan one device without actually opening it.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
[trini: Use %zu in a debug print to avoid warning]
Signed-off-by: Tom Rini <trini@konsulko.com>
This brings all structure accessors from btrfs-progs/ctree.h, as in
kernel's ctree.h.
All these accessors handle the endian convert at runtime, and since all
of them are defined as static inline functions, those which aren't used
won't take space in resulting binary.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This brings the extent_io_tree infrastructure, with which we can finally
bring in proper btrfs_fs_info structure to ctree.h.
With read/write_extent_buffer() implemented we also backport
read/write_eb_member() to ctree.h.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This patch implements an infrastructure to insert/search/merge an extent
range (with variable length).
This provides the basis for later extent buffer cache used in btrfs.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This patch uses generic code from btrfs-progs to read one super block
from block device.
To support the btrfs-progs coding style, the following is also
crossported:
- BTRFS_SETGET_FUNC for btrfs_super_block
- btrfs_check_super() function
- Move btrfs_read_superblock() to disk-io.[ch]
Since super.c only contains pretty small amount of code, and
the extra check will be covered in later root read patches.
Differences between this implementation and btrfs-progs:
- No sbflags/sb_bytenr support
Since we only need to read the primary super block (like kernel),
sbflags/sb_bytenr used by super block recovery is not needed.
This also changes the following behavior of U-Boot btrfs:
- Only reads the primary super block
The old implementation reads all 3 super blocks, and also one
non-existing backup.
This is not correct, especially if there is another filesystem created
on the device but old superblocks are not rewritten.
Just like kernel, we only check the primary super block.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
[trini: Change error to be a define in compat.h]
Signed-off-by: Tom Rini <trini@konsulko.com>
This mostly crossports crypto/hash.[ch] from btrfs-progs.
The differences are:
- No blake2 support
No blake2 related library in U-Boot yet.
- Use uboot xxhash/sha256 directly
No need to implement the code as U-Boot has already provided the
interface.
This adds the support for the following csums:
- SHA256
- XXHASH
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
This version includes all needed on-disk format from kernel.
Only need to modify the include headers for U-Boot, everything else is
untouched.
Also, since U-Boot btrfs is using a different endian convert timing (at
tree block read time), it needs some forced type conversion before
proper crossport.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Use log functions for error and debug messages of the file-system.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Fix defects such as uninitialized variables and untrusted pointer
operations. Most part of the tainted variables and the related defects
actually comes from Linux's macro get_unaligned_le**, extensively used
in SquashFS code. Add sanity checks for those variables.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Add call to lzo's lzo1x_decompress_safe() into sqfs_decompress().
U-Boot's LZO sources may still have some unsolved issues that could make the
decompression crash when dealing with fragmented files, so those should be
avoided. The "-no-fragments" option can be passed to mksquashfs.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Add call to ZSTD's ZSTD_decompressDCtx(). In this use case, the caller
can upper bound the decompressed size, which will be the SquashFS data
block (or metadata block) size, so there is no need to use streaming
API. Add ZSTD's worskpace to squashfs_ctxt structure.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Add sqfs_decompressor_init() and sqfs_decompressor_cleanup(). These
functions are called respectively in sqfs_probe() and sqfs_close(). For
now, only ZSTD requires an initialization logic. ZSTD support will be
added in a follow-up commit.
Move squashfs_ctxt definition to sqfs_filesystem.h. This structure is
passed to sqfs_decompressor_init() and sqfs_decompressor_cleanup(), so
it can no longer be local to sqfs.c.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Add call to zlib's 'uncompress' function. Add function to display the
right error message depending on the decompression's return value.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Add zlib (v1.2.11) uncompr() function to U-Boot. SquashFS depends on
this function to decompress data from a raw disk image. The actual
support for zlib into SquashFS sources will be added in a follow-up
commit.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
Add support for SquashFS filesystem. Right now, it does not support
compression but support for zlib will be added in a follow-up commit.
Signed-off-by: Joao Marcos Costa <joaomarcos.costa@bootlin.com>
While using u-boot with qemu's virtio driver I stumbled across a
problem reading files less than sector size. On the real hardware the
block reader seems ok with reading zero blocks, and while we could fix
the virtio host side of qemu to deal with a zero block read instead of
crashing, the u-boot fat driver should not be doing zero block reads
in the first place. If you ask hardware to read zero blocks you are
just going to get zero data. There may also be other hardware that
responds similarly to the virtio interface so this is worth fixing.
Without the patch I get the following and have to restart qemu because
it dies.
---------------------------------
=> fatls virtio 0:1
30 cmdline.txt
=> fatload virtio 0:1 ${loadaddr} cmdline.txt
qemu-system-aarch64: virtio: zero sized buffers are not allowed
---------------------------------
With the patch I get the expected results.
---------------------------------
=> fatls virtio 0:1
30 cmdline.txt
=> fatload virtio 0:1 ${loadaddr} cmdline.txt
30 bytes read in 11 ms (2 KiB/s)
=> md.b ${loadaddr} 0x1E
40080000: 64 77 63 5f 6f 74 67 2e 6c 70 6d 5f 65 6e 61 62 dwc_otg.lpm_enab
40080010: 6c 65 3d 30 20 72 6f 6f 74 77 61 69 74 0a le=0 rootwait.
---------------------------------
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
If a file cannot be loaded, show an error message.
Set the EFI boot device only after successfully loading a file.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Allocated tmpbuf_cluster dynamically to reduce the data size added by
compiling with CONFIG_FAT_WRITE.
Reported-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Truncate file names if the buffer size is exceeded to avoid a buffer
overflow.
Use Sphinx style function description.
Add a TODO comment.
Reported-by: CID 303779
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Added command "fstypes" to list supported/included filesystems.
Signed-off-by: Niel Fourie <lusus@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
[trini: Limit to sandbox]
Signed-off-by: Tom Rini <trini@konsulko.com>
According to the FAT specification it is valid to have files with an
attribute value of 0x0. This fixes a regression where different U-Boot
versions are showing different amount of files on the same storage
device. With this change U-Boot shows the same number of files and folders
as Linux and Windows.
Fixes: 39606d462c ("fs: fat: handle deleted directory entries correctly")
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
The size is not actually used since it is present in the header. Drop this
parameter. Also tidy up error handling while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Currently we support reading a file from CBFS given the address of the end
of the ROM. Sometimes we only know the start of the CBFS. Add a function
to find a file given that.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
This function currently returns a node pointer so there is no way to know
the error code. Also it uses data in BSS which seems unnecessary since the
caller might prefer to use a local variable.
Update the function and split its body out into a separate function so we
can use it later.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
We may as well return the error code and use it directly in the command
code. CBFS still uses its own error enum which we may be able to remove,
but leave it for now.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
The start address of the CBFS is used when scanning for files. It makes
sense to put this in our cbfs_priv struct and calculate it when we read
the header.
Update the code accordingly.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
It doesn't make sense to use u8 * as the pointer type for accessing the
CBFS since we do not access it as bytes, but via structures. Change it to
void *, which allows us to avoid a cast.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
These two functions have mostly the same code. Pull this out into a common
function.
Also make this function zero the private data so that callers don't have
to do it. Finally, update cbfs_load_header_ptr() to take the base of the
ROM as its parameter, which makes more sense than passing the address of
the header within the ROM.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
This function is strange at the moment in that it takes a header pointer
but then accesses the cbfs_s global. Currently clients have their own priv
pointer, so update the function to take that as a parameter instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
This function is strange at the moment in that it takes a header pointer
but then accesses the cbfs_s global. Currently clients have their own priv
pointer, so update the function to take that as a parameter instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
At present this uses a true return to indicate it found a file. Adjust it
to use 0 for this, so it is consistent with other functions.
Update its callers accordingly and add a check for malloc() failure in
file_cbfs_fill_cache().
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
At present this uses an int type. U-Boot now supports bool so use this
instead.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
U-Boot uses ulong for addresses but there are a few places in this driver
that don't use it. Convert this driver over to follow this convention
fully.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
At present the result variable in the cbfs_priv is called 'result' as is
the local variable in a few functions. Change the latter to 'ret' which is
more common in U-Boot and avoids confusion.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
We should not use typedefs in U-Boot. They cannot be used as forward
declarations which means that header files must include the full header to
access them.
Drop the typedef and rename the struct to remove the _s suffix which is
now not useful.
This requires quite a few header-file additions.
Signed-off-by: Simon Glass <sjg@chromium.org>
We should not be using typedefs and these make it harder to use
forward declarations (to reduce header file inclusions). Drop the typedef.
Signed-off-by: Simon Glass <sjg@chromium.org>
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.
Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.
Signed-off-by: Simon Glass <sjg@chromium.org>
Move this uncommon header out of the common header.
Fix up some style problems in flash.h while we are here.
Signed-off-by: Simon Glass <sjg@chromium.org>
The inode list uses version and ino, the dirent list uses version and pino.
This information is collected during scanning, reducing accesses to flash
and significantly speeding up ls and read.
Signed-off-by: Petr Borsodi <petr.borsodi@i.cz>
Obsolete nodes (ie. without the JFFS2_NODE_ACCURATE flag) were ignored
because they had seemingly invalid crc. This could lead to finding
the phantom node header in obsolete node data.
Signed-off-by: Petr Borsodi <petr.borsodi@i.cz>
As u-boot doesn't support the metadata_csum feature, writing to a
filesystem with this feature enabled will fail, as expected. However,
during the process, a journal state check is performed, which could
result in:
- a fs recovery if the fs wasn't umounted properly
- the fs being marked dirty
Both these cases result in a superblock change, leading to a mismatch
between the superblock checksum and its contents. Therefore, Linux will
consider the filesystem heavily corrupted and will require e2fsck to be
run manually to boot.
By bypassing the journal state check, this patch ensures the superblock
won't be corrupted if the filesystem has metadata_csum feature enabled.
Signed-off-by: Arnaud Ferraris <arnaud.ferraris@collabora.com>
When logical address of a regular extent is 0, the extent is sparse and
consists of all zeros.
Without this when sparse extents are used in a file reading fails with
Cannot map logical address 0 to physical
Signed-off-by: Marek Behún <marek.behun@nic.cz>
For certain btrfs files with compressed file extent, uboot will fail to
load it:
btrfs_read_extent_reg: disk_bytenr=14229504 disk_len=73728 offset=0 nr_bytes=131
072
decompress_lzo: tot_len=70770
decompress_lzo: in_len=1389
decompress_lzo: in_len=2400
decompress_lzo: in_len=3002
decompress_lzo: in_len=1379
decompress_lzo: in_len=88539136
decompress_lzo: header error, in_len=88539136 clen=65534 tot_len=62580
NOTE: except the last line, all other lines are debug output.
Btrfs lzo compression uses its own format to record compressed size
(segment header, LE32).
However to make decompression easier, we never put such segment header
across page boundary.
In above case, the xxd dump of the lzo compressed data looks like this:
00001fe0: 4cdc 02fc 0bfd 02c0 dc02 0d13 0100 0001 L...............
00001ff0: 0000 0008 0300 0000 0000 0011 0000|0000 ................
00002000: 4705 0000 0001 cc02 0000 0000 0000 1e01 G...............
'|' is the "expected" segment header start position.
But in that page, there are only 2 bytes left, can't contain the 4 bytes
segment header.
So btrfs compression will skip that 2 bytes, put the segment header in
next page directly.
Uboot doesn't have such check, and read the header with 2 bytes offset,
result 0x05470000 (88539136), other than the expected result
0x00000547 (1351), resulting above error.
Follow the btrfs-progs restore implementation, by introducing tot_in to
record total processed bytes (including headers), and do proper page
boundary skip to fix it.
Please note that, current code base doesn't parse fs_info thus we can't
grab sector size easily, so it uses PAGE_SIZE, and relying on fs open
time check to exclude unsupported sector size.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Cc: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Although in theory u-boot fs driver could easily support more sector
sizes, current code base doesn't have good enough way to grab sector
size yet.
This would cause problem for later LZO fixes which rely on sector size.
And considering that most u-boot boards are using 4K page size, which is
also the most common sector size for btrfs, rejecting fs with
non-page-sized sector size shouldn't cause much problem.
This should only be a quick fix before we implement better sector size
support.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Cc: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Just a cleanup. These immediate numbers make my eyes hurt.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Cc: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
We need to align the cache buffer to ARCH_DMA_MINALIGN in order to avoid
access errors like
CACHE: Misaligned operation at range [be0231e0, be0235e0]
seen on the MCIMX7SABRE.
Fixes: d5aee659f2 ("fs: ext4: cache extent data")
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
TPM TEE driver
Various minor sandbox video enhancements
New driver model core utility functions
-----BEGIN PGP SIGNATURE-----
iQEzBAABCgAdFiEEslwAIq+Gp8wWVbYnfxc6PpAIreYFAl48iogACgkQfxc6PpAI
reaVzAf/an3/yKe6r3CVWlcRV6H/dVg1ApnnLpX7jS0p0b++oCVvOiy7z1WPXj3k
b1SSgENDeeZ/8EHio+Gf7ZidH/TGEj7L6YEFwd1t60GMkZiWEkNf4Z53tw482YG+
96hoPD+ySTW+ddIdVHWAFG2I4aEiKHANJAp/ItNdD+rLbrEwNQy+eiK5JTOk80B6
/X8AJCLZeAC1s7vs+2+WolgjT78QGzA9HHalMiublcqh0ivKKk0QeQiOKKPe8JYJ
om5YY1TxayQ60Xmo5f39/SBfzEEklxw83sU9o1tBeYzyVUpu7fQdkxiDbWdsij77
DgwLdeYQJGbN+hdSWE0gjTqyhW+lWA==
=KRoA
-----END PGP SIGNATURE-----
Merge tag 'dm-pull-6feb20' of https://gitlab.denx.de/u-boot/custodians/u-boot-dm
sandbox conversion to SDL2
TPM TEE driver
Various minor sandbox video enhancements
New driver model core utility functions
The code for handing file overwrite incorrectly assumed that the file on
disk is always contiguous. This resulted in corrupting disk structure
every time when write to existing fragmented file happened. Fix this
by adding proper check for cluster discontinuity and adjust chunk size
on each partial write.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
This patch partially fixes the issue revealed by the following test
script:
--->8-fat_test1.sh---
#!/bin/bash
make sandbox_defconfig
make
dd if=/dev/zero of=/tmp/10M.img bs=1024 count=10k
mkfs.vfat -v /tmp/10M.img
cat >/tmp/cmds <<EOF
x
host bind 0 /tmp/10M.img
fatls host 0
mw 0x1000000 0x0a434241 0x1000 # "ABC\n"
mw 0x1100000 0x0a464544 0x8000 # "DEF\n"
fatwrite host 0 0x1000000 file0001.raw 0x1000
fatwrite host 0 0x1000000 file0002.raw 0x1000
fatwrite host 0 0x1000000 file0003.raw 0x1000
fatwrite host 0 0x1000000 file0004.raw 0x1000
fatwrite host 0 0x1000000 file0005.raw 0x1000
fatrm host 0 file0002.raw
fatrm host 0 file0004.raw
fatls host 0
fatwrite host 0 0x1100000 file0007.raw 0x4000
fatwrite host 0 0x1100000 file0007.raw 0x4000
reset
EOF
./u-boot </tmp/cmds
#verify
rm -r /tmp/result /tmp/model
mkdir /tmp/result
mkdir /tmp/model
yes ABC | head -c 4096 >/tmp/model/file0001.raw
yes ABC | head -c 4096 >/tmp/model/file0003.raw
yes ABC | head -c 4096 >/tmp/model/file0005.raw
yes DEF | head -c 16384 >/tmp/model/file0007.raw
mcopy -n -i /tmp/10M.img ::file0001.raw /tmp/result
mcopy -n -i /tmp/10M.img ::file0003.raw /tmp/result
mcopy -n -i /tmp/10M.img ::file0005.raw /tmp/result
mcopy -n -i /tmp/10M.img ::file0007.raw /tmp/result
hd /tmp/10M.img
if diff -urq /tmp/model /tmp/result
then
echo Test okay
else
echo Test fail
fi
--->8---
Overwritting a discontiguous test file (file0007.raw) no longer causes
corruption to file0003.raw, which's data lies between the chunks of the
test file. The amount of data written to disk is still incorrect, what
causes damage to the file (file0005.raw), which's data lies next to the
test file. This will be fixed by the next patch.
Feel free to prepare a proper sandbox/py_test based tests based on the
provided test scripts.
At present dm/device.h includes the linux-compatible features. This
requires including linux/compat.h which in turn includes a lot of headers.
One of these is malloc.h which we thus end up including in every file in
U-Boot. Apart from the inefficiency of this, it is problematic for sandbox
which needs to use the system malloc() in some files.
Move the compatibility features into a separate header file.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present devres.h is included in all files that include dm.h but few
make use of it. Also this pulls in linux/compat which adds several more
headers. Drop the automatic inclusion and require files to include devres
themselves. This provides a good indication of which files use devres.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Anatolij Gustschin <agust@denx.de>
linux_compat.c is the best place for kmemdup(), which is currenly used
only in ubifs.c, but will also be used when other kernel files
(in my case, lib/crypto/x509_cert_parser.c and pkcs7_parser.c) will be
imported. So just move it.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Unlink test for FAT file system seems to fail at test_unlink2.
(When I added this test, I haven't seen any errors though.)
for example,
===8<===
fs_obj_unlink = ['fat', '/home/akashi/tmp/uboot_sandbox_test/128MB.fat32.img']
def test_unlink2(self, u_boot_console, fs_obj_unlink):
"""
Test Case 2 - delete many files
"""
fs_type,fs_img = fs_obj_unlink
with u_boot_console.log.section('Test Case 2 - unlink (many)'):
output = u_boot_console.run_command('host bind 0 %s' % fs_img)
for i in range(0, 20):
output = u_boot_console.run_command_list([
'%srm host 0:0 dir2/0123456789abcdef%02x' % (fs_type, i),
'%sls host 0:0 dir2/0123456789abcdef%02x' % (fs_type, i)])
assert('' == ''.join(output))
output = u_boot_console.run_command(
'%sls host 0:0 dir2' % fs_type)
> assert('0 file(s), 2 dir(s)' in output)
E AssertionError: assert '0 file(s), 2 dir(s)' in ' ./\r\r\n ../\r\r\n 0 0123456789abcdef11\r\r\n\r\r\n1 file(s), 2 dir(s)'
test/py/tests/test_fs/test_unlink.py:52: AssertionError
===>8===
This can happen when fat_itr_next() wrongly detects an already-
deleted directory entry.
File deletion, which was added in the commit f8240ce95d ("fs: fat:
support unlink"), is implemented by marking its entry for a short name
with DELETED_FLAG, but related entry slots for a long file name are kept
unmodified. (So entries will never be actually deleted from media.)
To handle this case correctly, an additional check for a directory slot
will be needed in fat_itr_next().
In addition, I added extra comments about long file name and short file
name format in FAT file system. Although they are not directly related
to the issue, I hope it will be helpful for better understandings
in general.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
These don't need to be in common.h so move them out into a new header.
Also add some missing comments.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Drop inclusion of crc.h in common.h and use the correct header directly
instead.
With this we can drop the conflicting definition in fw_env.h and rely on
the crc.h header, which is already included.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
This function is a variant of fs_get_type_name() and returns a filesystem
type with which the current device is associated.
We don't want to export fs_type variable directly because we have to take
care of it consistently within fs.c.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
fs_ls(), fs_mkdir() and fs_unlink() sets fs_type to FS_TYPE_ANY
explicitly, but it is redundant as they call fs_close().
So just remove those lines.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
fs_close() closes the connection to a file system which opened with
either fs_set_blk_dev() or fs_set_dev_with_part(). Many file system
functions implicitly call fs_close(), e.g. fs_closedir(), fs_exist(),
fs_ln(), fs_ls(), fs_mkdir(), fs_read(), fs_size(), fs_write()
and fs_unlink().
So just export it.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
