| Age | Commit message (Collapse) | Author |
|---|
|
Fix fastpath issues
Since mmci_request() can be called from a non-interrupt
context, and does, during kernel init, causing a host
of debug messages during boot if you enable spinlock debugging,
we need to use the spinlock calls that save IRQ flags and
restore them.
Signed-off-by: Linus Walleij <triad@df.lth.se>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
Relax requirements on host controllers and only require that they do not
report a transfer count than is larger than the actual one (i.e. a lower
value is okay). This is how many other parts of the kernel behaves so
upper layers should already be prepared to handle that scenario. This
gives us a performance boost on MMC cards.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
|
|
This fixes some two minor clk issues.
The first is a comparison where a byte will probably wrap around to 0 instead of being saturated to 255, shouldn't be triggered very often but need fixing.
The second is an attempt by the driver to adjust MCLK down to the maximum frequency according to the spec, so we don't accidentally overclock the PL18x block. None of the mach-{versatile|integrator|lh7a40x} that use it in-tree seem to have a problem with this (all are well below 100MHz, typically 33MHz), but some day there will be a problem.
This is not applied on top of the earlier mmci patch for race condition but rather a clean 2.6.25, but I guess it applies without major protests anyway.
Signed-off-by: Linus Walleij <triad@df.lth.se>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
Updated version of 4446/1. This also drops the suggested comparison
of host_remain for == 0, since that doesn't make sense (still works
for us, too). We have verified that this patch solve race problems
on atleast 2 archs at high frequencies.
(Verbatim copy of old patch text below.)
The patch below fixes a race condition in the ARM MMCI PL18x driver.
If new data arrives in the FIFO while existing data is being read then
we get a second iteration of the loop in mmci_pio_read.
However host->size is not updated until after mmci_pio_read returns,
so we get count = number of new bytes PLUS number of bytes already
copied in the first iteration. This results in a FIFO underrun as
we try and read mode data than is available.
The fix is to compensating for data read on previous iterations
when calculating the amount of data in the FIFO.
Signed-off-by: Linus Walleij <triad@df.lth.se>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
Do a full scan of the directory to try and be a bit more proactive,
instead of waiting for things to break.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
|
|
Using weight32() to determine if a value is a power of 2 is a rather
heavi weight solution. The classic idiom is (x & (x - 1)) == 0, but
the kernel already provide a is_power_of_2 function for it.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
|
|
Remove the BYTEBLOCK capability and let the broken hosts fail the
requests with -EINVAL instead.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
|
|
Convert the MMC layer to use standard error codes and not its own,
incompatible values.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
|
|
Make sure all headers in the files reflect their true position
in the tree.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
|
|
Clean up the drivers/mmc directory by moving card and host drivers
into subdirectories.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
|
