Some host controllers can set mmc->caps before sdhci_add_host().
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some host controllers such as s5pc110 support the WIDE8 feature.
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When using QUIRK_NONSTANDARD_CLOCK, it checks the set_clock() function
which is not used actually. So delete it.
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The current way of disabling it is not well tested by vendor and has all
kinds of bugs that show up on resume from ram/disk. A very good example
is a dead SDHCI controller.
Old way of disabling is still supported by continuing to use
CONFIG_MMC_RICOH_MMC.
Based on 'http://list.drzeus.cx/pipermail/sdhci-devel/2007-December/002085.html'
Therefore most of the credit for this goes to Andrew de Quincey
Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Cc: Andrew de Quincey <adq_dvb@lidskialf.net>
Acked-by: Philip Langdale <philipl@overt.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Even though many mmc host drivers pass a pm_message_t argument to
mmc_suspend_host() that argument isn't used the by MMC core. As host
drivers are converted to dev_pm_ops they'll have to construct
pm_message_t's (as they won't be passed by the PM subsystem any more) just
to appease the mmc suspend interface.
We might as well just delete the unused paramter.
Signed-off-by: Matt Fleming <matt@console-pimps.org>
Acked-by: Anton Vorontsov <cbouatmailru@gmail.com>
Acked-by: Michal Miroslaw <mirq-linux@rere.qmqm.pl>ZZ
Acked-by: Sascha Sommer <saschasommer@freenet.de>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The s3c6410 sdhci controller does not support the 'End' attribute and NOP
attribute in the same 8-Byte ADMA descriptor. This patch adds a new quirk
to identify sdhci host contollers with such behaviour. In addition to
this, for controllers using the new quirk, the last entry in the ADMA
descritor table is marked with the 'End' attribute (instead of using a NOP
descriptor with 'End' attribute).
Signed-off-by: Maurus Cuelenaere <mcuelenaere@gmail.com>
Signed-off-by: Thomas Abraham <thomas.ab@samsung.com>
Acked-by: Ben Dooks <ben-linux@fluff.org>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some hosts (e.g. as found in CNS3xxx SOCs) report wrong value in
CLOCK_BASE capability field, and currently there is no way to force the
SDHCI core to use the platform-provided base clock value.
This patch implements CAP_CLOCK_BASE_BROKEN quirk. When enabled, the
SDHCI core will always use base clock frequency provided by the platform.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Cc: Richard Röjfors <richard.rojfors@pelagicore.com>
Cc: David Vrabel <david.vrabel@csr.com>
Cc: Pierre Ossman <pierre@ossman.eu>
Cc: Ben Dooks <ben@simtec.co.uk>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Tested with an XO v1.5 from OLPC.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
sdhci_set_adma_desc() is using byte-writes to write data in a specified
order into memory. Change to using __le16 for the two byte and __le32 for
the four byte cases and use the cpu_to_{le16,le32} to do the conversion
before writing.
This will reduce the size of the code and the number of writes as we no
longer need to chop the data up before writing.
As an example on ARM S3C64XX SoC, in little-endian configuration:
000000d4 <sdhci_set_adma_desc>:
- d8: e1a0c423 lsr ip, r3, #8
- dc: e1a0ec21 lsr lr, r1, #24
- e0: e1a04821 lsr r4, r1, #16
- e4: e1a05421 lsr r5, r1, #8
- e8: e1a06442 asr r6, r2, #8
- ec: e5c0c001 strb ip, [r0, #1]
- f0: e5c0e007 strb lr, [r0, #7]
- f4: e5c04006 strb r4, [r0, #6]
- f8: e5c05005 strb r5, [r0, #5]
- fc: e5c01004 strb r1, [r0, #4]
- 100: e5c06003 strb r6, [r0, #3]
- 104: e5c02002 strb r2, [r0, #2]
- 108: e5c03000 strb r3, [r0]
+ d4: e5801004 str r1, [r0, #4]
+ d8: e1c030b0 strh r3, [r0]
+ dc: e1c020b2 strh r2, [r0, #2]
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: Pierre Ossman <pierre@ossman.eu>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The code to write the ADMA descriptor into memory is repeated several
times throughout sdhci_adma_table_pre, and thus should be moved into a
common function. This will also be useful if the patch to make the write
more efficient is accepted.
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: Pierre Ossman <pierre@ossman.eu>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On an OLPC XO-1.5 development board with Via VX855 chipset, the sdhci
controller can take up to 12ms to stabilize its clock, but the current
timeout at which we give up on the controller is 10ms.
The patch increases the timeout delay rather than using a device-specific
quirk -- since we exit the loop when the clock comes up, increasing the
timeout value will only make us mdelay() longer in the errant case of a
device with a clock that is not stabilizing, which it seems worth waiting
a little longer for in general.
Signed-off-by: Chris Ball <cjb@laptop.org>
Cc: Harald Welte <HaraldWelte@viatech.com>
Acked-by: Pierre Ossman <pierre@ossman.eu>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add support for ADMA on SDHCI hosts, not supporting SDMA.
According to the SDHCI specifications a host can support ADMA but not SDMA
Signed-off-by: Richard Röjfors <richard.rojfors@mocean-labs.com>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
eSDHC fails to recognize some SDHS cards, throwing timeout errors:
mmc0: error -110 whilst initialising SD card
That's because we calculate timeout value in a wrong way: on eSDHC hosts
the timeout clock is derivied from the SD clock, which is set dynamically.
As David Vrabel suggested, deriving timeout clock from SD clock is a
common scheme, so let's implement DATA_TIMEOUT_USES_SDCLK quirk and use it
for eSDHC hosts.
Also, from now on we don't need esdhc_get_timeout_clock() callback, so
remove it.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Cc: Pierre Ossman <pierre@ossman.eu>
Cc: Kumar Gala <galak@kernel.crashing.org>
Cc: David Vrabel <david.vrabel@csr.com>
Cc: Ben Dooks <ben@fluff.org>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
get_min_clock() makes sense only with NONSTANDARD_CLOCK quirk and when
set_clock() callback is specified.
The patch should cause no functional changes, it just makes the code
self-documented and avoids any possible misuse of get_min_clock().
Suggested-by: Pierre Ossman <pierre@ossman.eu>
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Cc: Ian Molton <ian@mnementh.co.uk>
Cc: Matt Fleming <matt@console-pimps.org>
Cc: Philip Langdale <philipl@overt.org>
Cc: Pierre Ossman <pierre@ossman.eu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
so the page will be flushed on unmap on ARCH which need it.
Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
Since commit 8dfd0374be ("MMC core: limit
minimum initialization frequency to 400kHz") MMC core checks for minimum
frequency, and that causes following messages flood when using eSDHC
controllers:
...
mmc0: Minimum clock frequency too high for identification mode
mmc0: Minimum clock frequency too high for identification mode
...
The warnings are legitimate, since if we'd use 133 MHz clocks for standard
SDHCI controllers, we'd not able to scale frequency down to 400 kHz.
But eSDHC controllers have a non-standard SD clock management, so we can
divide clock by 256 * 16, not just 256.
This patch introduces get_min_clock() callback for sdhci core and
implements it for sdhci-of driver, and thus fixes the issue.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Cc: Matt Fleming <matt@console-pimps.org>
Cc: Ian Molton <ian@mnementh.co.uk>
Cc: "Roberto A. Foglietta" <roberto.foglietta@gmail.com>
Cc: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The SDHCI controller found in the VX855ES requires 10ms
delay between applying power and applying clock.
This issue has been discovered and documented by the OLPC XO1.5 team.
Signed-off-by: Harald Welte <HaraldWelte@viatech.com>
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
Some hosts (hardware configurations, or particular SD/MMC slots) may
not support 4-bit bus. For example, on MPC8569E-MDS boards we can
switch between serial (1-bit only) and nibble (4-bit) modes, thought
we have to disable more peripherals to work in 4-bit mode.
Along with some small core changes, this patch modifies sdhci-of
driver, so that now it looks for "sdhci,1-bit-only" property in the
device-tree, and if specified we enable a proper quirk.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
Add quirk to show the controller cannot do multi-block IO.
This is mainly for the Samsung SDHCI controller that currently
cannot manage to do multi-block PIO without timing out.
Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
Update the ADMA error reporting to not only show the
overall controller state but also to print the ADMA
descriptor list.
Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
If using ADMA, then we should print the ADMA error
and current pointer in sdhci_dumpregs() when any
debug is requested.
Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
Because of granularity issues, sometimes we told the hardware to change
to the voltage we were already at. Rework the logic so this doesn't
happen.
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
FSL eSDHC controllers can support maximum block size up to 4096 bytes,
the MBL (Maximum Block Length) field in the capabilities register
extended by one bit, and is set to 0x3.
But the SDHCI core doesn't support blocks of 4096 bytes, and thus
forces blksz to the lowest value -- 512 bytes. With this patch we can
pin up the blksz to the maximum supported block size, i.e. 2048 bytes.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
FSL eSDHC controllers losing signal/interrupt enable states after
reset, so we should re-enable them.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Small udelay is needed to make eSDHC work in PIO mode. Without
the delay reading causes endless interrupt storm, and writing
corrupts data. The first guess would be that we must wait for
some bit in some register, but I didn't find any reliable bits
that change before and after the delay.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
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>
The Samsung SDHCI (and FSL eSDHC) controller block seems to fail
to generate an INT_DATA_END after the transfer has completed and
the bus busy state finished.
Changes in e809517f6f to use the
new busy method are the cause of the behaviour change.
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>
Fix the led device naming for the sdhci driver.
The led class documentation defines the led name to have the
form "devicename:colour:function" while not applicable sections
should be left blank.
To comply with the documentation the led device name is changed
from "mmc*" to "mmc*::".
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
This reverts commit a4b7619377.
It turned out that the controller had problem running at the
higher speed, so go back to trusting the hardware capability
bits.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
As reported by Randy Dunlap, having sdhci built-in and LEDs class
as a module resulted in undefined symbols. Change the code to handle
that case properly (by not having LEDs class support in sdhci).
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
CONFIG_LEDS_CLASS is defined only if led-class is built-in, otherwise
when it is a module the option is called CONFIG_LEDS_CLASS_MODULE. Led
support should also be activated in this case.
Signed-off-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
The variable 'scratch' is always initialized before it's used. The
conditional which is responsible for initialization of 'scratch' will
always evaluate 'true' when the first loop iteration occurs, and thus,
it's properly initialized. GCC doesn't see this, of course, so using
the uninitialized_var() macro seems to work for silencing this case.
Signed-off-by: Steven Noonan <steven@uplinklabs.net>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Some high speed capable controllers forget to set the high speed
capability bit. Make sure we enable the functionality anyway.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
The card detection delay was added early when the behaviour of the
card interrupt was still very much unknown (i.e. before there was a
public specification). As it is now known that it is a debounced signal,
reduce the delay to something more sensible.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
The sdhci controllers can interrupt us when the busy state from the
card has ended, saving CPU cycles and power.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
The wrong flag was manipulated when an invalid sg list was given, turning
off DMA on the next (and all subsequent) request instead of the current
one.
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Add per-device dma_mapping_ops support for CONFIG_X86_64 as POWER
architecture does:
This enables us to cleanly fix the Calgary IOMMU issue that some devices
are not behind the IOMMU (http://lkml.org/lkml/2008/5/8/423).
I think that per-device dma_mapping_ops support would be also helpful for
KVM people to support PCI passthrough but Andi thinks that this makes it
difficult to support the PCI passthrough (see the above thread). So I
CC'ed this to KVM camp. Comments are appreciated.
A pointer to dma_mapping_ops to struct dev_archdata is added. If the
pointer is non NULL, DMA operations in asm/dma-mapping.h use it. If it's
NULL, the system-wide dma_ops pointer is used as before.
If it's useful for KVM people, I plan to implement a mechanism to register
a hook called when a new pci (or dma capable) device is created (it works
with hot plugging). It enables IOMMUs to set up an appropriate
dma_mapping_ops per device.
The major obstacle is that dma_mapping_error doesn't take a pointer to the
device unlike other DMA operations. So x86 can't have dma_mapping_ops per
device. Note all the POWER IOMMUs use the same dma_mapping_error function
so this is not a problem for POWER but x86 IOMMUs use different
dma_mapping_error functions.
The first patch adds the device argument to dma_mapping_error. The patch
is trivial but large since it touches lots of drivers and dma-mapping.h in
all the architecture.
This patch:
dma_mapping_error() doesn't take a pointer to the device unlike other DMA
operations. So we can't have dma_mapping_ops per device.
Note that POWER already has dma_mapping_ops per device but all the POWER
IOMMUs use the same dma_mapping_error function. x86 IOMMUs use device
argument.
[akpm@linux-foundation.org: fix sge]
[akpm@linux-foundation.org: fix svc_rdma]
[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: fix bnx2x]
[akpm@linux-foundation.org: fix s2io]
[akpm@linux-foundation.org: fix pasemi_mac]
[akpm@linux-foundation.org: fix sdhci]
[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: fix sparc]
[akpm@linux-foundation.org: fix ibmvscsi]
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Avi Kivity <avi@qumranet.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Kyungmin Park