The call to spi_master_put() in img_spfi_remove() is redundant since
the master is registered using devm_spi_register_master() and no
reference hold by using spi_master_get() in img_spfi_remove().
This is detected by Coccinelle semantic patch.
Signed-off-by: Wei Yongjun <weiyj.lk@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The call to spi_master_put() in mtk_spi_remove() is redundant since
the master is registered using devm_spi_register_master() and no
reference hold by using spi_master_get() in mtk_spi_remove().
This is detected by Coccinelle semantic patch.
Signed-off-by: Wei Yongjun <weiyj.lk@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The call to spi_master_put() in spi_qup_remove() is redundant since
the master is registered using devm_spi_register_master() and no
reference hold by using spi_master_get() in spi_qup_remove().
This is detected by Coccinelle semantic patch.
Signed-off-by: Wei Yongjun <weiyj.lk@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The current SPI code attempts to use bus_lock_mutex for two purposes. One
is to implement spi_bus_lock() which grants exclusive access to the bus.
The other is to serialize access to the physical hardware. This duplicate
purpose causes confusion which leads to cases where access is not locked
when a caller holds the bus lock mutex. Fix this by splitting out the I/O
functionality into a new io_mutex.
This means taking both mutexes in the DMA path, replacing the existing
mutex with the new I/O one in the message pump (the mutex now always
being taken in the message pump) and taking the bus lock mutex in
spi_sync(), allowing __spi_sync() to have no mutex handling.
While we're at it hoist the mutex further up the message pump before we
power up the device so that all power up/down of the block is covered by
it and there are no races with in-line pumping of messages.
Reported-by: Rich Felker <dalias@libc.org>
Tested-by: Rich Felker <dalias@libc.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Separate driver probing from SPI transfer functions.
Signed-off-by: Jan Glauber <jglauber@cavium.com>
Tested-by: Steven J. Hill <steven.hill@cavium.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Move the register definitions to the drivers directory because they
are only used there.
Signed-off-by: Jan Glauber <jglauber@cavium.com>
Tested-by: Steven J. Hill <steven.hill@cavium.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of hard-coding the register offsets put them into a struct
and set them in the probe function.
Signed-off-by: Jan Glauber <jglauber@cavium.com>
Tested-by: Steven J. Hill <steven.hill@cavium.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Storing the system clock frequency in struct octeon_spi avoids
calling the MIPS specific octeon_get_io_clock_rate() for every transfer.
Signed-off-by: Jan Glauber <jglauber@cavium.com>
Tested-by: Steven J. Hill <steven.hill@cavium.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Remove all calls to cvmx_read_csr()/cvmx_write_csr() and use
the portable readq()/writeq() functions.
Signed-off-by: Steven J. Hill <steven.hill@cavium.com>
Signed-off-by: Jan Glauber <jglauber@cavium.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
wait_for_completion_timeout returns unsigned long not int so the check for
<= 0 should be == 0 here, and the type unsigned long. The function return
is set to -ETIMEDOUT to reflect the actual problem.
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
wait_for_completion_timeout returns unsigned long not int so the check for
<= 0 should be == 0 here.
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The Rockchip SPI controller's length register only supports 16-bits,
yielding a maximum length of 64KiB (the CTRLR1 register holds "length -
1"). Trying to transfer more than that (e.g., with a large SPI flash
read) will cause the driver to hang.
Now, it seems that while theoretically we should be able to program
CTRLR1 with 0xffff, and get a 64KiB transfer, but that also seems to
cause the core to choke, so stick with a maximum of 64K - 1 bytes --
i.e., 0xffff.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Return IRQ_NONE from the interrupt handler if the handler is running, but
no interrupt was detected. This allows the system to recover in case of an
interrupt storm due to an invalid interrupt configuration or faulty
hardware.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The Xilinx SPI driver can operate without an IRQ, but not every error
returned by platform_get_irq() means that no IRQ was specified. It will
also return an error if the IRQ specification is invalid or the IRQ
provider is not yet available (EPROBE_DEFER).
So instead of ignoring all errors only ignore ENXIO, which means no IRQ was
specified, and propagate all other errors to device driver core.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Patch a9e93e8 has erroneously removed some comments which are
important to understand why the bus frequency is multiplied by
two during the spi transfer.
Reword the previous comment to a more appropriate message.
Suggested-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Reviewed-by: Michael Turquette <mturquette@baylibre.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The new compatible is related to the Samsung Exynos5433 SoC. The
difference between the previous is that in the exynos5433 the SPI
controller is driven by three clocks instead of only one.
The new clock (ioclk) is controlling the input/output clock
whenever the controller is slave or master.
The presence of the clock line is detected from the compatibility
structure (exynos5433_spi_port_config) as a boolean value.
The probe function checks whether the ioclk is present and if so,
it acquires.
The runtime suspend and resume functions will handle the clock
enabling and disabling as well.
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Reviewed-by: Michael Turquette <mturquette@baylibre.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
If clk_prepare_enable() fails do not return -EBUSY but use the
value provided by the function itself.
Suggested-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Reviewed-by: Michael Turquette <mturquette@baylibre.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The goto labels of the style of
err4:
err3:
err2:
err1:
are complex to insert in between new errors without renaming all
the goto statements. Replace the errX naming style to meaningful
names in order to make it easier to insert new goto exit points.
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Michael Turquette <mturquette@baylibre.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Once a spi_master_get() call succeeds, we need an additional
spi_master_put() call to free the memory, otherwise we will
leak a reference to master. Fix by removing the unnecessary
spi_master_get() call.
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Reviewed-by: Andi Shyti <andi.shyti@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The workqueue has a single workitem(&ss->ws) and hence doesn't require
ordering. Also, it is not being used on a memory reclaim path. Hence, the
singlethreaded workqueue has been replaced with the use of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in spi_sh_remove() to ensure that
there are no pending tasks while disconnecting the driver.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The workqueue "wk" serves as a queue for carrying out execution
of requests. It has a single work item(&drv_data->work) and hence doesn't
require ordering. Also, it is not being used on a memory reclaim path.
Hence, the singlethreaded workqueue has been replaced with the use of
system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in pch_spi_free_resources() to ensure that
there are no pending tasks while disconnecting the driver.
Also dropped the label 'err_return' since it's not being used anymore.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently, the driver handles mapping buffers to be used by the DMA.
However, there are times that the current mapping implementation will
fail for certain buffers. Fortunately, the SPI framework can detect
and map buffers so its usable by the DMA.
Update the driver to utilize the SPI framework for buffer
mapping instead. Also incorporate hooks that the framework uses to
determine if the DMA can or can not be used.
This will result in the original omap2_mcspi_transfer_one function being
deleted and omap2_mcspi_work_one being renamed to
omap2_mcspi_transfer_one. Previously transfer_one was only responsible
for mapping and work_one handled the transfer. But now only transferring
needs to be handled by the driver.
Signed-off-by: Franklin S Cooper Jr <fcooper@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The function sg_split will be used by spi-omap2-mcspi to handle a SoC
workaround in the SPI driver. Therefore, select SG_SPLIT so this function
is available to the driver.
Signed-off-by: Franklin S Cooper Jr <fcooper@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The commit 30f3a6ab44 ("spi: pxa2xx: Add support for both chip selects on
Intel Braswell") introduces a support of chipselects for Intel Braswell SPI
host controller. Though it missed to convert the PCI part of the driver.
Do conversion here which enables both chipselects on Intel Braswell when
enumerated via PCI.
We don't care about num_chipselect value since it is overrided inside core
driver.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
It seems the commit e5262d0568 ("spi: spi-pxa2xx: SPI support for Intel Quark
X1000") misses one place to be adapted for Intel Quark, i.e. in reset_sccr1().
Clear all RFT bits when call reset_sccr1() on Intel Quark.
Fixes: e5262d0568 ("spi: spi-pxa2xx: SPI support for Intel Quark X1000")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
This is a complex patch for refactoring CLPS711X SPI driver.
This change adds devicetree support and removes board support.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
When the clock is coming from the cmu it is not required to be
disabled and then re-enabled in order to change the rate.
Besides, some exynos chipsets (e.g. exynos5433) do not deliver
any to the SFR if one from the pclk ("spi" in this case) or sclk
("busclk") is disabled.
Remove the clock disabling/enabling to avoid falling into this
situation.
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
OMAP35x and OMAP37x mentions in the McSPI End-of-Transfer Sequences section
that if the McSPI is configured as a Master and only DMA RX is being
performed then the DMA transfer size needs to be reduced by 1 or 2.
This was originally implemented by:
commit 57c5c28dbc ("spi: omap2_mcspi rxdma bugfix")
This patch adds comments to clarify what is going on in the code since its
not obvious what problem its addressing.
Signed-off-by: Franklin S Cooper Jr <fcooper@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Some IoT and maker software stacks are using spidev to perform raw access
to the SPI bus instead of relying existing drivers provided by the kernel.
They then implement their own "drivers" in userspace on top of the spidev
raw interface. This is far from being an ideal solution but we do not want
to prevent using mainline Linux in these devices.
Now, it turns out that Windows has similar SPI devices than spidev which
allow raw access on the SPI bus to userspace programs as described in the
link below:
https://msdn.microsoft.com/windows/hardware/drivers/spb/spi-tests-in-mitt
These SPI test devices are also meant to be used during development and
testing.
In order to allow usage of spidev for development and testing in Linux, add
those same ACPI IDs to the spidev driver (which is Linux counterpart of the
Windows SPI test devices), but complain loudly so that users know it is not
good idea to use it in production systems. Instead they should be using
proper drivers for peripherals connected to the SPI bus.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Simply sort header block alphabetically.
While here, sort devices by PCI ID and add a copyright line for Intel.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
pcim_iomap_table() can't fail when called after pcim_iomap_regions(). Moreover,
we already dereference returned value and kernel will crash if it is not
correct.
Remove obvious leftover of commit 0202775bc3 ("spi/pxa2xx-pci: switch to use
pcim_* interfaces").
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The SPI controllers used on Intel Merrifield are PXA2XX compatible. This patch
enables them.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Move LPSS specific setup to a separate function. It makes ->probe() cleaner as
well as allows extend the driver for different variation of hardware in the
future, e.g. for Intel Merrifield.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Kaby Lake PCH-H has the same SPI host controller as Skylake. Add these new
PCI IDs to the list of supported devices.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
It generates a static checker warning if an if statement isn't indented.
I think the code is fine except for the white space issue.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The workqueue "workqueue" serves as a driver message queue.
It has a single work item(&drv_data->pump_messages) and hence doesn't
require ordering. Also, it is not being used on a memory reclaim path.
Hence, the singlethreaded workqueue has been replaced with the use of
system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in bfin_sport_spi_destroy_queue() to ensure
that there are no pending tasks while disconnecting the driver.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The workqueue "workqueue" has a single work item(&mps->work)
doesn't require ordering. Also, it is not being used on a memory reclaim
path. Hence, the singlethreaded workqueue has been replaced with the use
of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in mpc52xx_psc_spi_of_remove() to ensure that
nothing is pending while disconnecting the driver.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The workqueue "workqueue" has a single work item(&c->work) and hence
doesn't require ordering. Also, it is not being used on a memory reclaim
path. Hence, the singlethreaded workqueue has been replaced with the use of
system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in txx9spi_remove() to ensure that
nothing is pending while disconnecting the driver.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The workqueue "workqueue" serves as a driver message queue.
It has a single work item(&drv_data->pump_messages) and hence doesn't
require ordering. Also, it is not being used on a memory reclaim path.
Hence, the singlethreaded workqueue has been replaced with the use of
system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in bfin_spi_destroy_queue() to ensure that there
are no pending tasks while disconnecting the driver.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The whole function is inside an 'if' statement
("!is_polling(sdd)").
Check the opposite of that statement at the beginning and exit,
this way we can have one level less of indentation.
Remove the goto paths as they are redundant.
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
At the start of the transfer, the spi_config function is called
twice, the first time when the 3c64xx_spi_prepare_message is
called and the second time with the s3c64xx_spi_transfer_one,
both called from the spi framework.
Remove the first call at the prepare message because in that
point we don't have the imformation about "bit per word" and
frequency.
Signed-off-by: Andi Shyti <andi.shyti@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
