When write a file into FAT file system, it will search a match file in
root dir. So the find_directory_entry() will get the first cluster of
root dir content and search the directory item one by one. If the file
is not found, we will call get_fatent_value() to get next cluster of root
dir via lookup the FAT table and continue the search.
The issue is in FAT16/12 system, we cannot get root dir's next clust
from FAT table. The FAT table only be use to find the clust of data
aera in FAT16/12.
In FAT16/12 if the clust is in root dir, the clust number is a negative
number or 0, 1. Since root dir is located in front of the data area.
Data area start clust #2. So the root dir clust number should < 2.
This patch will check above situation before call get_fatenv_value().
If curclust is < 2, include minus number, we just increase one on the
curclust since root dir is in continous cluster.
The patch also add a sanity check for entry in get_fatenv_value().
Signed-off-by: Josh Wu <josh.wu@atmel.com>
In fat_write.c, the last clust condition check is incorrect:
if ((curclust >= 0xffffff8) || (curclust >= 0xfff8)) {
... ...
}
For example, in FAT32 if curclust is 0x11000. It is a valid clust.
But on above condition check, it will be think as a last clust.
So the correct last clust check should be:
in fat32, curclust >= 0xffffff8
in fat16, curclust >= 0xfff8
in fat12, curclust >= 0xff8
This patch correct the last clust check.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
This bug shows up when file stored on the ext4 file system is updated.
The ext4fs_delete_file() is responsible for deleting file's (e.g. uImage)
data.
However some global data (especially ext4fs_indir2_block), which is used
during file deletion are left unchanged.
The ext4fs_indir2_block pointer stores reference to old ext4 double
indirect allocated blocks. When it is unchanged, after file deletion,
ext4fs_write_file() uses the same pointer (since it is already initialized
- i.e. not NULL) to return number of blocks to write. This trunks larger
file when previous one was smaller.
Lets consider following scenario:
1. Flash target with ext4 formatted boot.img (which has uImage [*] on itself)
2. Developer wants to upload their custom uImage [**]
- When new uImage [**] is smaller than the [*] - everything works
correctly - we are able to store the whole smaller file with corrupted
ext4fs_indir2_block pointer
- When new uImage [**] is larger than the [*] - theCRC is corrupted,
since truncation on data stored at eMMC was done.
3. When uImage CRC error appears, then reboot and LTHOR/DFU reflashing causes
proper setting of ext4fs_indir2_block() and after that uImage[**]
is successfully stored (correct uImage [*] metadata is stored at an
eMMC on the first flashing).
Due to above the bug was very difficult to reproduce.
This patch sets default values for all ext4fs_indir* pointers/variables.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Code responsible for handling situation when ext4 has block size of 1024B
can be ordered to take less space.
This patch does that for ext4 common and write files.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Use of malloc of do_fat_write() causes cache error on ARM v7 platforms.
Perhaps, the same problem will occur at any other CPUs.
This replaces malloc with memalign to fix cache buffer alignment.
Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Yoshiyuki Ito <yoshiyuki.ito.ub@renesas.com>
Tested-by: Hector Palacios <hector.palacios@digi.com>
U-Boot already has a list implementation, and files which include both
that and the yaffs implementation will get errors:
In file included from ydirectenv.h:80:0,
from yportenv.h:81,
from yaffs_guts.h:19,
from yaffs_allocator.h:19,
from yaffs_allocator.c:14:
yaffs_list.h:32:8: error: redefinition of ‘struct list_head’
struct list_head {
^
Remove the yaffs implementation.
Signed-off-by: Simon Glass <sjg@chromium.org>
When we tell the compiler to optimize for ARMv7 (and ARMv6 for that
matter) it assumes a default of SCTRL.A being cleared and unaligned
accesses being allowed and fast at the hardware level. We set this bit
and must pass along -mno-unaligned-access so that the compiler will
still breakdown accesses and not trigger a data abort.
To better help understand the requirements of the project with respect
to unaligned memory access, the
Documentation/unaligned-memory-access.txt file has been added as
doc/README.unaligned-memory-access.txt and is taken from the v3.14-rc1
tag of the kernel.
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Mans Rullgard <mans@mansr.com>
Signed-off-by: Tom Rini <trini@ti.com>
This reverts commit fc0fc50f38.
The author has asked on the mailing list that we revert this for now as
it breaks write support.
Reported-by: Łukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Tom Rini <trini@ti.com>
In an ext4 filesystem, the inode corresponding to a file has a 60-byte
area which contains an extent header structure and up to 4 extent
structures (5 x 12 bytes).
For files that need more than 4 extents to be represented (either files
larger than 4 x 128MB = 512MB or smaller files but very fragmented),
ext4 creates extent index structures. Each extent index points to a 4KB
physical block where one extent header and additional 340 extents could
be stored.
The current u-boot ext4 code is very inefficient when it tries to load a
file which has extent indexes. For each logical file block the code will
read over and over again the same blocks of 4096 bytes from the disk.
Since the extent tree in a file is always the same, we can cache the
extent structures in memory before actually starting to read the file.
This patch creates a simple linked list of structures holding information
about all the extents used to represent a file. The list is sorted by
the logical block number (ee_block) so that we can easily find the
proper extent information for any file block.
Without this patch, a 69MB file which had just one extent index pointing
to a block with another 6 extents was read in approximately 3 minutes.
With this patch applied the same file can be read in almost 20 seconds.
Signed-off-by: Ionut Nicu <ioan.nicu.ext@nsn.com>
Now we are ready to switch over to real Kbuild.
This commit disables temporary scripts:
scripts/{Makefile.build.tmp, Makefile.host.tmp}
and enables real Kbuild scripts:
scripts/{Makefile.build,Makefile.host,Makefile.lib}.
This switch is triggered by the line in scripts/Kbuild.include
-build := -f $(if $(KBUILD_SRC),$(srctree)/)scripts/Makefile.build.tmp obj
+build := -f $(if $(KBUILD_SRC),$(srctree)/)scripts/Makefile.build obj
We need to adjust some build scripts for U-Boot.
But smaller amount of modification is preferable.
Additionally, we need to fix compiler flags which are
locally added or removed.
In Kbuild, it is not allowed to change CFLAGS locally.
Instead, ccflags-y, asflags-y, cppflags-y,
CFLAGS_$(basetarget).o, CFLAGS_REMOVE_$(basetarget).o
are prepared for that purpose.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Tested-by: Gerhard Sittig <gsi@denx.de>
This commit changes the working directory
where the build process occurs.
Before this commit, build process occurred under the source
tree for both in-tree and out-of-tree build.
That's why we needed to add $(obj) prefix to all generated
files in makefiles like follows:
$(obj)u-boot.bin: $(obj)u-boot
Here, $(obj) is empty for in-tree build, whereas it points
to the output directory for out-of-tree build.
And our old build system changes the current working directory
with "make -C <sub-dir>" syntax when descending into the
sub-directories.
On the other hand, Kbuild uses a different idea
to handle out-of-tree build and directory descending.
The build process of Kbuild always occurs under the output tree.
When "O=dir/to/store/output/files" is given, the build system
changes the current working directory to that directory and
restarts the make.
Kbuild uses "make -f $(srctree)/scripts/Makefile.build obj=<sub-dir>"
syntax for descending into sub-directories.
(We can write it like "make $(obj)=<sub-dir>" with a shorthand.)
This means the current working directory is always the top
of the output directory.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Tested-by: Gerhard Sittig <gsi@denx.de>
This hooks into the generic "file exists" support added in an earlier
patch, and provides an implementation for the FAT filesystem.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
This hooks into the generic "file exists" support added in an earlier
patch, and provides an implementation for the ext4 filesystem.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
This hooks into the generic "file exists" support added in an earlier
patch, and provides an implementation for the sandbox test environment.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
FAT and ext4 expect that the passed in block device descriptor not be
NULL. This causes problems on sandbox, where get_device_and_partition()
succeeds for the "host" device, yet passes back a NULL device descriptor.
Add special handling for this situation, so that the generic filesystem
commands operate as expected on sandbox.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
This could be used in scripts such as:
if test -e mmc 0:1 /boot/boot.scr; then
load mmc 0:1 ${scriptaddr} /boot/boot.scr
source ${scriptaddr}
fi
rather than:
if load mmc 0:1 ${scriptaddr} /boot/boot.scr; then
source ${scriptaddr}
fi
This prevents errors being printed by attempts to load non-existent
files, which can be important when checking for a large set of files,
such as /boot/boot.scr.uimg, /boot/boot.scr, /boot/extlinux.conf,
/boot.scr.uimg, /boot.scr, /extlinux.conf.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
Fix a few issues with the generic "save" shell command, and fs_write()
function.
1) fstypes[].write wasn't filled in for some file-systems, and isn't
checked when used, which could cause crashes/... if executing save
on e.g. fat/ext filesystems.
2) fs_write() requires the length argument to be non-zero, since it needs
to know exactly how many bytes to write. Adjust the comments and code
according to this.
3) fs_write() wasn't prototyped in <fs.h> like other generic functions;
other code should be able to call this directly rather than invoking
the "save" shell command.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Simon Glass <sjg@chromium.org>
The summary already has other verification. This one is not needed.
The check caused summaries to be ignored if they were not on the
numbered block. This caused problems when a summary was embedded in an
image and the image is written to a flash with bad blocks.
Signed-off-by: Charles Manning <cdhmanning@gmail.com>
For files where we actually have extent indexes following
an extent header (ext_block->eh_depth != 0), the do/while
loop from ext4fs_get_extent_block() does not select the
proper extent index structure.
For example, if we have:
ext_block->eh_depth = 1
ext_block->eh_entries = 1
fileblock = 0
index[0].ei_block = 0
the do/while loop will exit with i set to 0 and the
ext4fs_get_extent_block() function will return 0, even if
there was a valid extent index structure following the
header.
Signed-off-by: Ionut Nicu <ioan.nicu.ext@nsn.com>
Signed-off-by: Mathias Rulf <mathias.rulf@nsn.com>
Using fs->blksz in ext4fs_get_extent_block() is not
correct since fs->blksz is not initialized on the
read path. Use EXT2_BLOCK_SIZE() instead which will
produce the desired output.
Signed-off-by: Ionut Nicu <ioan.nicu.ext@nsn.com>
Signed-off-by: Mathias Rulf <mathias.rulf@nsn.com>
Curently, we are using 32 bit multiplication to calculate the offset,
so the result will always be 32 bit.
This can silently cause file system corruption when performing a write
operation on partition larger than 4 GiB.
This patch address the issue by simply promoting the terms to 64 bit,
and let compilers decide how to do the multiplication efficiently.
Signed-off-by: Ma Haijun <mahaijuns@gmail.com>
It may cause file system corruption when do a write operation.
This issue only affects boards that use 32 bit lbaint_t.
Signed-off-by: Ma Haijun <mahaijuns@gmail.com>
Curently memcpy copies string without null terminating char because
function strlen returns only number of characters excluding
null terminating character. Replace memcpy with strcpy.
Signed-off-by: Piotr Wilczek <p.wilczek@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
CC: Tom Rini <trini@ti.com>
This commit moves some subdirectories of fs
from the toplevel Makefile to fs/Makefile
using Kbuild descending feature.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: Simon Glass <sjg@chromium.org>
commit 39ac34473f ("cmd_mtdparts: use 64
bits for flash size, partition size & offset") introduced warnings
in a couple places due to printf formats or pointer casting.
This patch fixes the warnings pointed out here:
http://lists.denx.de/pipermail/u-boot/2013-October/164981.html
Signed-off-by: Scott Wood <scottwood@freescale.com>
Cc: York Sun <yorksun@freescale.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Tom Rini <trini@ti.com>
As documented, almost all U-Boot commands expect numbers to be entered
in hexadecimal input format. (Exception: for historical reasons, the
"sleep" command takes its argument in decimal input format.)
This rule was broken for the "load" command; for details please see
especially commits 045fa1e "fs: add filesystem switch libary,
implement ls and fsload commands" and 3f83c87 "fs: fix number base
behaviour change in fatload/ext*load". In the result, the load
command would always require an explicit "0x" prefix for regular
(i. e. base 16 formatted) input.
Change this to use the standard notation of base 16 input format.
While strictly speaking this is a change of the user interface, we
hope that it will not cause trouble. Stephen Warren comments (see
[1]):
I suppose you can change the behaviour if you want; anyone
writing "0x..." for their values presumably won't be
affected, and if people really do assume all values in U-Boot
are in hex, presumably nobody currently relies upon using
non-prefixed values with the generic load command, since it
doesn't work like that right now.
[1] http://article.gmane.org/gmane.comp.boot-loaders.u-boot/171172
Acked-by: Tom Rini <trini@ti.com>
Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>
In the set_cluster() function, it will convert the buffer size to sector
numbers. Then call disk_write() to write by sector.
For remaining buffer, the size is less than a sector, call disk_write()
again to write them in one sector.
But if the total buffer size is less then one sector, the original code
will call disk_write() with zero sector number. It is unnecessary.
So this patch fix this. Now it will not call disk_write() if total buffer size
is less than one sector.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
Fix reading ext4_extent_header struture on BE machines. Some 16 bit
fields where converted to 32 bit fields, due to the byte swap on BE
machines the containing value was corrupted. Therefore reading ext4
filesystems on BE machines where broken before.
Signed-off-by: Rommel Custodio <sessyargc+uboot@gmail.com>
[sent via git-send-email; rework commit message]
Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Simon Glass <sjg@chromium.org>
Tested-by: Lukasz Majewski <l.majewski@samsung.com>
With CONFIG_SYS_64BIT_LBA, lbaint_t gets defined as a 64-bit type,
which is required to represent block numbers for storage devices that
exceed 2TiB (the block size usually is 512B), e.g. recent hard drives
We now use lbaint_t for partition offset to reflect the lbaint_t change,
and access partitions beyond or crossing the 2.1TiB limit.
This required changes to signature of ext4fs_devread(), and type of all
variables relatives to block sector.
ext2/ext4 fs uses logical block represented by a 32 bit value. Logical
block is a multiple of device block sector. To avoid overflow problem
when calling ext4fs_devread(), we need to cast the sector parameter.
Signed-off-by: Frédéric Leroy <fredo@starox.org>
"cramfsload uImage_1" succeeds even though the actual file is named
"uImage".
Fix file name comparison when one name is the prefix of the other.
Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
cc: Wolfgang Denk <wd@denx.de>
cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
This patch is essentially an update of u-boot MTD subsystem to
the state of Linux-3.7.1 with exclusion of some bits:
- the update is concentrated on NAND, no onenand or CFI/NOR/SPI
flashes interfaces are updated EXCEPT for API changes.
- new large NAND chips support is there, though some updates
have got in Linux-3.8.-rc1, (which will follow on top of this patch).
To produce this update I used tag v3.7.1 of linux-stable repository.
The update was made using application of relevant patches,
with changes relevant to U-Boot-only stuff sticked together
to keep bisectability. Then all changes were grouped together
to this patch.
Signed-off-by: Sergey Lapin <slapin@ossfans.org>
[scottwood@freescale.com: some eccstrength and build fixes]
Signed-off-by: Scott Wood <scottwood@freescale.com>
Commit 50ce4c0 "fs/ext4: Support device block sizes != 512 bytes"
modified ext4fs_set_blk_dev() to calculate total_sect based on
get_fs()->dev_desc->log2blksz rather than SECTOR_SIZE. However, this
value wasn't yet assigned. Move the assignment earlier so the code
doesn't crash or hang.
Cc: Egbert Eich <eich@suse.com>
Tested-by: Tom Rini <trini@ti.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
The 512 byte block size was hard coded in the ext4 file systems.
Large harddisks today support bigger block sizes typically 4096
bytes.
This patch removes this limitation.
Signed-off-by: Egbert Eich <eich@suse.com>
Bugfix:
Here at this place we need the fat size in sectors not bytes.
This was found during code review when adding support for storage
devices with blocksizes != 512.
Signed-off-by: Egbert Eich <eich@suse.com>
This allows write of files from the host filesystem in sandbox. There is
currently no concept of overwriting the file and removing its existing
contents - all writing is done on top of what is there. This means that
writing 10 bytes to the start of a 1KB file will only update those 10
bytes, not truncate the file to 10 byte slong.
If the file does not exist it is created.
Signed-off-by: Simon Glass <sjg@chromium.org>
