FSL eSDHC hosts have incompatible register map to manage the SDCLK.
This patch adds set_clock callback so that drivers could overwrite
set_clock behaviour.
Similar patch[1] was posted by Ben Dooks, though in Ben's version the
callback is named change_clock, plus the patch has some unrelated bits
that makes the patch difficult to reuse.
[1] http://lkml.org/lkml/2008/12/2/160
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some controllers do not provide clock information in their capabilities
(in the Samsung case, it is because there are multiple clock sources
available to the controller). Add hooks to allow the system to supply
clock information.
p.s.
In the original Ben's patch there was a bug that makes sdhci_add_host()
return -ENODEV even if callbacks were specified. This is fixed now.
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
This patch adds SDHCI_QUIRK_INVERTED_WRITE_PROTECT quirk. When
specified, the sdhci driver will invert WP state.
p.s. Actually, the quirk is more board-specific than
controller-specific.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
This patch adds SDHCI_QUIRK_BROKEN_CARD_DETECTION quirk. When specified,
sdhci driver will set MMC_CAP_NEEDS_POLL MMC host capability, and won't
enable card insert/remove interrupts.
This is needed for hosts with unreliable card detection, such as FSL
eSDHC. The original eSDHC driver was tring to "debounce" card-detection
IRQs by reading present state and disabling particular interrupts. But
with this debouncing scheme I noticed that sometimes we miss card
insertion/removal events.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some hosts (that is, FSL eSDHC) throw PIO interrupts during DMA
transfers, this causes tons of unneeded interrupts, and thus highly
degraded speed.
This patch modifies the driver so that now we only enable relevant
(DMA or PIO) interrupts during transfers.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Card detection interrupts should be handled separately as they should
not be enabled before mmc_add_host() returns and should be disabled
before calling mmc_remove_host(). The same is for suspend and resume
routines.
sdhci_init() no longer enables card-detection irqs. Instead, two new
functions implemented: sdhci_enable_card_detection() and
sdhci_disable_card_detection().
New sdhci_reinit() call implemented to behave the same way as the old
sdhci_init().
Also, this patch implements and uses few new helpers to manage IRQs in
a more conveinient way, that is:
- sdhci_clear_set_irqs()
- sdhci_unmask_irqs()
- sdhci_mask_irqs()
- SDHCI_INT_ALL_MASK constant
sdhci_enable_sdio_irq() converted to these new helpers, plus the
helpers will be used by the subsequent patches.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Currently the SDHCI driver works with PCI accessors (write{l,b,w} and
read{l,b,w}).
With this patch drivers may change memory accessors, so that we can
support hosts with "weird" IO memory access requirments.
For example, in "FSL eSDHC" SDHCI hardware all registers are 32 bit
width, with big-endian addressing. That is, readb(0x2f) should turn
into readb(0x2c), and readw(0x2c) should be translated to
le16_to_cpu(readw(0x2e)).
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some cards are not able to send the data token in time, but
miss the time frame for some bits(!). So synchronize to the
start of the token.
Signed-off-by: Wolfgang Muees <wolfgang.mues@auerswald.de>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
It is necessary to wait for bus power before sending
any commands.
Signed-off-by: Adrian Hunter <adrian.hunter@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Data transfers on third OMAP3 MMC controller don't work
because DMA line numbers are only defined for MMC1 and MMC2.
Fix that and store line numbers in mmc_omap_host structure
to reduce code size.
Tested on OMAP3 pandora board.
Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Turn off the bus power at suspend.
Signed-off-by: Jarkko Lavinen <jarkko.lavinen@nokia.com>
Signed-off-by: Adrian Hunter <adrian.hunter@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Allow slot_name to be the same as the other OMAP
driver, by removing the redundant "slot:" prefix.
Signed-off-by: Adrian Hunter <adrian.hunter@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Allow a cover switch to be used to cause a rescan of the
MMC slot.
Signed-off-by: Adrian Hunter <adrian.hunter@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Instead of using the bounce buffer, using scatter-gather emulation
(as in the OMAP1/2 MMC driver) removes the need of one extra memory
copy and improves performance.
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some MMC commands result in the card becoming busy after
the response is received. This needs to be specified
for the omap_hsmmc host controller, which is what this
patch does. However, the effect is that some commands
with no data will cause a Transfer Complete (TC) interrupt
in addition to the Command Complete (CC) interrupt.
In order to deal with that, the irq handler has needed
a few changes also.
The benefit of this change is that the omap_hsmmc host
controller driver now waits for the TC interrupt while
the card is busy, so the mmc_block driver needs to poll
the card status just once instead of repeatedly.
i.e. the net result is more sleep and less cpu.
The command sequence for open-ended multi-block write
with DMA is now:
Issue write command CMD25
Receive CC interrupt
Data is sent
Receive TC interrupt (DMA is done)
Issue stop command CMD12
Receive CC interrupt
Card is busy
Receive TC interrupt
Card is now ready for next transfer
Signed-off-by: Adrian Hunter <adrian.hunter@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
This supports MMC/SD/SDIO currently found on the Kirkwood 88F6281 and
88F6192 SoC controllers.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
drivers/mmc/host/tmio_mmc.h: In function 'tmio_mmc_kmap_atomic':
drivers/mmc/host/tmio_mmc.h:147: error: implicit declaration of function 'kmap_atomic'
drivers/mmc/host/tmio_mmc.h:147: error: 'KM_BIO_SRC_IRQ' undeclared (first use in this function)
drivers/mmc/host/tmio_mmc.h: In function 'tmio_mmc_kunmap_atomic':
drivers/mmc/host/tmio_mmc.h:153: error: implicit declaration of function 'kunmap_atomic'
drivers/mmc/host/tmio_mmc.h:153: error: 'KM_BIO_SRC_IRQ' undeclared (first use in this function)
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Update the tmio_mmc code to call mmc_free_host() when
done using the private data. Without this fix the driver
frees memory and then keeps on using it as private data.
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Ian Molton <ian@mnementh.co.uk>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Update the tmio_mmc code to use resource_size(). With this
patch applied the correct resource size is passed to ioremap().
Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Ian Molton <ian@mnementh.co.uk>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some SD cards have very high timeouts in SPI mode.
So adjust the timeouts from theory to practice.
Signed-off-by: Wolfgang Muees <wolfgang.mues@auerswald.de>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some cards are slower than the standard allows and need more
time to respond to a command. Max. observed number of bytes
was 12.
Signed-off-by: Wolfgang Muees <wolfgang.mues@auerswald.de>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Allow the platform data structures to specify spi mode 3
(if there is a pullup on the clock line or the spi hardware
is not able to serve spi mode 0).
Signed-off-by: Wolfgang Muees <wolfgang.mues@auerswald.de>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
During mmc unsafe resume, choose the right voltage for the card after
powerup.
Although this has not seen to cause trouble, it's the wrong behaviour.
Signed-off-by: Balaji Rao <balajirrao@openmoko.org>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Most registers lose its state when the processor wakes up from sleep state.
Thus registers should be initialized, when the processor wakes up. However the
current hsmmc 'resume' function doesn't consider this issue and finally makes
deadlock. So this patch fixes this problem.
Signed-off-by: Kim Kyuwon <chammoru@gmail.com>
Signed-off-by: Adrian Hunter <ext-adrian.hunter@nokia.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Remove code that turns MMC1 power back on after it
has been powered off (when the voltage is 1.8V).
The offending code is not necessary because the
host controller bus voltage is initialized to
3V when probing or resuming. Note that MMC powers up
with the highest voltage available (see mmc_power_up())
which will be 3V also.
Signed-off-by: Adrian Hunter <ext-adrian.hunter@nokia.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Extended CSD is a MMC card register. As increasingly interesting
fields are being added to Extended CSD, it is helpful to see its
value. Note that SD cards do not have an Extended CSD
register, so it is MMC only.
Signed-off-by: Adrian Hunter <ext-adrian.hunter@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Currently we are using an explicit udev rule to trigger loading of the
mmc-block module when an MMC or SD card is detected:
SUBSYSTEM=="mmc", RUN+="/sbin/modprobe -Qba mmc-block"
It makes much more sense for the mmc bus driver and the mmc-block module to
share MODALIAS information so that they are linked automatically.
There is no real information of use in the MMC system at the current time.
All devices inserted require us to load the mmc-block device. Until such
time as useful parameters exist simply reflect the module linkage via
the module alias below:
mmc:block
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
The delayed work item mmc_host.detect is now cancelled before flushing
the work queue. This takes care of cases when delayed_work was scheduled
for mmc_host.detect, but not yet placed in the work queue.
Signed-off-by: Jorg Schummer <ext-jorg.2.schummer@nokia.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
On m68k:
| drivers/net/dnet.c: In function 'dnet_readw_mac':
| drivers/net/dnet.c:36: error: implicit declaration of function 'writel'
| drivers/net/dnet.c:43: error: implicit declaration of function 'readl'
| drivers/net/dnet.c: In function 'dnet_probe':
| drivers/net/dnet.c:873: error: implicit declaration of function 'ioremap'
| drivers/net/dnet.c:873: warning: assignment makes pointer from integer without a cast
| drivers/net/dnet.c:939: error: implicit declaration of function 'iounmap'
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The fsync log has code to make sure all of the parents of a file are in the
log along with the file. It uses a minimal log of the parent directory
inodes, just enough to get the parent directory on disk.
If the transaction that originally created a file is fully on disk,
and the file hasn't been renamed or linked into other directories, we
can safely skip the parent directory walk. We know the file is on disk
somewhere and we can go ahead and just log that single file.
This is more important now because unrelated unlinks in the parent directory
might make us force a commit if we try to log the parent.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The tree logging code allows individual files or directories to be logged
without including operations on other files and directories in the FS.
It tries to commit the minimal set of changes to disk in order to
fsync the single file or directory that was sent to fsync or O_SYNC.
The tree logging code was allowing files and directories to be unlinked
if they were part of a rename operation where only one directory
in the rename was in the fsync log. This patch adds a few new rules
to the tree logging.
1) on rename or unlink, if the inode being unlinked isn't in the fsync
log, we must force a full commit before doing an fsync of the directory
where the unlink was done. The commit isn't done during the unlink,
but it is forced the next time we try to log the parent directory.
Solution: record transid of last unlink/rename per directory when the
directory wasn't already logged. For renames this is only done when
renaming to a different directory.
mkdir foo/some_dir
normal commit
rename foo/some_dir foo2/some_dir
mkdir foo/some_dir
fsync foo/some_dir/some_file
The fsync above will unlink the original some_dir without recording
it in its new location (foo2). After a crash, some_dir will be gone
unless the fsync of some_file forces a full commit
2) we must log any new names for any file or dir that is in the fsync
log. This way we make sure not to lose files that are unlinked during
the same transaction.
2a) we must log any new names for any file or dir during rename
when the directory they are being removed from was logged.
2a is actually the more important variant. Without the extra logging
a crash might unlink the old name without recreating the new one
3) after a crash, we must go through any directories with a link count
of zero and redo the rm -rf
mkdir f1/foo
normal commit
rm -rf f1/foo
fsync(f1)
The directory f1 was fully removed from the FS, but fsync was never
called on f1, only its parent dir. After a crash the rm -rf must
be replayed. This must be able to recurse down the entire
directory tree. The inode link count fixup code takes care of the
ugly details.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
During log replay, inodes are copied from the log to the main filesystem
btrees. Sometimes they have a zero link count in the log but they actually
gain links during the replay or have some in the main btree.
This patch updates the link count to be at least one after copying the
inode out of the log. This makes sure the inode is deleted during an
iput while the rest of the replay code is still working on it.
The log replay has fixup code to make sure that link counts are correct
at the end of the replay, so we could use any non-zero number here and
it would work fine.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The delayed reference mechanism is responsible for all updates to the
extent allocation trees, including those updates created while processing
the delayed references.
This commit tries to limit the amount of work that gets created during
the final run of delayed refs before a commit. It avoids cowing new blocks
unless it is required to finish the commit, and so it avoids new allocations
that were not really required.
The goal is to avoid infinite loops where we are always making more work
on the final run of delayed refs. Over the long term we'll make a
special log for the last delayed ref updates as well.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This reads in blocks in the checksum btree before starting the
transaction in btrfs_finish_ordered_io. It makes it much more likely
we'll be able to do operations inside the transaction without
needing any btree reads, which limits transaction latencies overall.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
btrfs_mark_buffer dirty would set dirty bits in the extent_io tree
for the buffers it was dirtying. This may require a kmalloc and it
was not atomic. So, anyone who called btrfs_mark_buffer_dirty had to
set any btree locks they were holding to blocking first.
This commit changes dirty tracking for extent buffers to just use a flag
in the extent buffer. Now that we have one and only one extent buffer
per page, this can be safely done without losing dirty bits along the way.
This also introduces a path->leave_spinning flag that callers of
btrfs_search_slot can use to indicate they will properly deal with a
path returned where all the locks are spinning instead of blocking.
Many of the btree search callers now expect spinning paths,
resulting in better btree concurrency overall.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Commits are fairly expensive, and so btrfs has code to sit around for a while
during the commit and let new writers come in.
But, while we're sitting there, new delayed refs might be added, and those
can be expensive to process as well. Unless the transaction is very very
young, it makes sense to go ahead and let the commit finish without hanging
around.
The commit grow loop isn't as important as it used to be, the fsync logging
code handles most performance critical syncs now.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This reduces contention on the extent buffer spin locks by testing for a
blocking lock before trying to take the spinlock.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The fs/btrfs/inode.c code to run delayed allocation during writout
needed some stack usage optimization. This is the first pass, it does
the check for compression earlier on, which allows us to do the common
(no compression) case higher up in the call chain.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
To avoid deadlocks and reduce latencies during some critical operations, some
transaction writers are allowed to jump into the running transaction and make
it run a little longer, while others sit around and wait for the commit to
finish.
This is a bit unfair, especially when the callers that jump in do a bunch
of IO that makes all the others procs on the box wait. This commit
reduces the stalls this produces by pre-reading file extent pointers
during btrfs_finish_ordered_io before the transaction is joined.
It also tunes the drop_snapshot code to politely wait for transactions
that have started writing out their delayed refs to finish. This avoids
new delayed refs being flooded into the queue while we're trying to
close off the transaction.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
The delayed reference queue maintains pending operations that need to
be done to the extent allocation tree. These are processed by
finding records in the tree that are not currently being processed one at
a time.
This is slow because it uses lots of time searching through the rbtree
and because it creates lock contention on the extent allocation tree
when lots of different procs are running delayed refs at the same time.
This commit changes things to grab a cluster of refs for processing,
using a cursor into the rbtree as the starting point of the next search.
This way we walk smoothly through the rbtree.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
When extents are freed, it is likely that we've removed the last
delayed reference update for the extent. This checks the delayed
ref tree when things are freed, and if no ref updates area left it
immediately processes the delayed ref.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
