| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
The CREDITS file was removed from barebox in 2015 by commit 6570288f2d97
("Remove the CREDITS file"). Remove references to it from several files.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
| |
Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On 12/13/19 2:12 PM, Sascha Hauer wrote:
> On Tue, Dec 10, 2019 at 03:44:52PM +0100, Hubert Feurstein wrote:
>> With v2015.06.0 the indicated progress of the copy command is very
>> smooth. Calling "cp -v /dev/zero /dev/mmc3.root" takes about 80
>> seconds for 256MB. But with v2019.12.0 the progress is very bumpy and
>> the copy takes about 280 seconds.
>>
>> I've tracked this down to this commit which destroys the performance:
>> "block: Adjust cache sizes" (b6fef20c1215c6ef0004f6af4a9c4b77af51dc43)
>
> We could just revert this patch. I can't find any workload that gets
> faster with b6fef20c1215. It's rather the other way round.
Do this by reverting commit b6fef20c1215c6ef0004f6af4a9c4b77af51dc43.
Reported-by: Hubert Feurstein <h.feurstein@gmail.com>
Suggested-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
This implements the discard_range hook. When a range of data is
discarded then we do not have to read it from the device and can
pass a zeroed buffer instead.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| | |
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When calling I/O functions of underlying block device driver we always
need to make sure that its size is small enough to not go past
device's boundary. Not only in get_chunk() and block_cache(), but in
writebuffer_flush() as well. Since the same code is used in three
different places, move it into a subroutine and adjust all of the
calls to ->write()/->read() accordingly.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
All of the functions using debug() in that file have enough info to
use dev_dbg instead. Convert all of the uses of debug() to dev_dbg()
in order to get more informative debug output.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The two functions are identical, so there's no need to keep two copies
of the same code around. Alias block_op_close() to block_op_flush()
and drop standalone definition for the former.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Only the following cdevs do not declare an .lseek() operation:
- Console devices in common/console.c
- Firmware framework in common/firmware.c
- JTAG driver in drivers/misc/jtag.c
- UBI in drivers/mtd/ubi/barebox.c
Of those four, first two are marked DEVFS_IS_CHARACTER_DEV and
implement only .write() operation and the last two don't implement
anything but .ioctl(). While there's probably no meaningful way to use
lseek() against any of those devices, there doesn't seem to be any
harm in allowing it either.
Change devfs_lseek() to ignore absense of .lseek() callback and drop
dev_lseek_default() and all references to it in the codebase.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
Getting a error from blk->ops->write() is not a very unlikely
event (happens quite often during new board bringup), so we need to
catch and propagate them to upper layers so they can be at least
reported properly. Change the code of all of the callers to bail out
as soon as blk->ops->write() fails.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Use four times more cache entries and divide the memory for each entry
by four. This lowers the linear read throughput somewhat but increases
the access speed for filesystems.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Linux also has struct file_operations which are something different.
Rename our file_operations to cdev_operations which better matches
what we have.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
In barebox the default mount path for a cdev is /mnt/<cdevname>
which can be mounted with "mount <cdevname>" without specifying
a target path explicitly. Simplify this further by creating automount
entries for the default mount pathes which makes a manual mount
completely unnecessary.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
Some block devices need an explicit callback for flushing written
blocks. Add this callback.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
Saves a few bytes of binary space.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Some drivers use blk->ops->read/write. This bypasses the caching block
layer and was never intended like this. The upper API to the block layer
is the cdev layer. This patch adds block_read and block_write functions
and uses them where appropriate.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
block_read and block_write collide with a to-be-introduced
global function, so rename the file operation functions.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
The FSF address has changed in the past. Instead of updating it
each time the address changes, just drop it completely treewide.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| | |
Drivers may use dma to transfer blocks, so put them into
dma save memory.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| | |
Next step to 64bit support: Make cdev size a 64bit type.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
| |
This is a first step for 64bit file support: Make the file sizes/offsets
64bit.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
| |
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
It does not necessarily means a bug when block_get fails here,
this can also be a failure on the underlying device.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The current caching layer only has a single buffer for
writing and reading. The FAT driver often accesses the
fat and then data again, which currently can't be cached.
Reimplement this with a list of cached chunks. The number
of chunks and their sizes are currently hardcoded, but
that could be easily made configurable.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
This adds a simple block layer to barebox. Reading and writing
to block devices can be painfully slow without caching, so
add a simple caching layer here.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|