bcm2711, Rasberry Pi 4's SoC, shares one interrupt for multiple
instances of the bcm2835 SPI controller. So this enables shared
interrupt support for them.
The early bail out in the interrupt routine avoids messing with buffers
of transfers being done by other means. Otherwise, the driver can handle
receiving interrupts asserted by other controllers during an IRQ based
transfer.
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20200528185805.28991-1-nsaenzjulienne@suse.de
Signed-off-by: Mark Brown <broonie@kernel.org>
Most of blk-mq drivers depend on managed IRQ's auto-affinity to setup
up queue mapping. Thomas mentioned the following point[1]:
"That was the constraint of managed interrupts from the very beginning:
The driver/subsystem has to quiesce the interrupt line and the associated
queue _before_ it gets shutdown in CPU unplug and not fiddle with it
until it's restarted by the core when the CPU is plugged in again."
However, current blk-mq implementation doesn't quiesce hw queue before
the last CPU in the hctx is shutdown. Even worse, CPUHP_BLK_MQ_DEAD is a
cpuhp state handled after the CPU is down, so there isn't any chance to
quiesce the hctx before shutting down the CPU.
Add new CPUHP_AP_BLK_MQ_ONLINE state to stop allocating from blk-mq hctxs
where the last CPU goes away, and wait for completion of in-flight
requests. This guarantees that there is no inflight I/O before shutting
down the managed IRQ.
Add a BLK_MQ_F_STACKING and set it for dm-rq and loop, so we don't need
to wait for completion of in-flight requests from these drivers to avoid
a potential dead-lock. It is safe to do this for stacking drivers as those
do not use interrupts at all and their I/O completions are triggered by
underlying devices I/O completion.
[1] https://lore.kernel.org/linux-block/alpine.DEB.2.21.1904051331270.1802@nanos.tec.linutronix.de/
[hch: different retry mechanism, merged two patches, minor cleanups]
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Daniel Wagner <dwagner@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
kobject_init_and_add() takes reference even when it fails.
If this function returns an error, kobject_put() must be called to
properly clean up the memory associated with the object.
Previous commit "b8eb718348b8" fixed a similar problem.
Signed-off-by: Qiushi Wu <wu000273@umn.edu>
[ rjw: Subject ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Versions of VMD with the Host Physical Address shadow register use this
register to calculate the bus address offset needed to do guest
passthrough of the domain. This register shadows the Host Physical
Address registers including the resource type bits. After calculating
the offset, the extra resource type bits lead to the VMD resources being
over-provisioned at the front and under-provisioned at the back.
Example:
pci 10000:80:02.0: reg 0x10: [mem 0xf801fffc-0xf803fffb 64bit]
Expected:
pci 10000:80:02.0: reg 0x10: [mem 0xf8020000-0xf803ffff 64bit]
If other devices are mapped in the over-provisioned front, it could lead
to resource conflict issues with VMD or those devices.
Link: https://lore.kernel.org/r/20200528030240.16024-3-jonathan.derrick@intel.com
Fixes: a1a3017013 ("PCI: vmd: Fix shadow offsets to reflect spec changes")
Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
The name of pm_runtime_callbacks_present() is confusing, because
it suggests that the device has PM-runtime callbacks if 'true' is
returned by that function, but in fact that may not be the case,
so replace it with pm_runtime_has_no_callbacks() which is not
ambiguous.
No functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Use 'struct domain_pgtable' instead to free_pagetable(). This solves
the problem that amd_iommu_domain_direct_map() needs to restore
domain->pt_root after the device table has been updated just to make
free_pagetable release the domain page-table.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Link: https://lore.kernel.org/r/20200527115313.7426-4-joro@8bytes.org
This function is internal to the AMD IOMMU driver and only exported
because the amd_iommu_v2 modules calls it. But the reason it is called
from there could better be handled by amd_iommu_is_attach_deferred().
So unexport get_dev_data() and use amd_iommu_is_attach_deferred()
instead.
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Link: https://lore.kernel.org/r/20200527115313.7426-3-joro@8bytes.org
get_user_pages_fast() doesn't need the caller to check that.
NB: reachable only from ioctl(2) and only under USER_DS
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
pin_user_pages_fast() doesn't need that from its caller.
NB: only reachable from ->ioctl(), and only under USER_DS
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
in all affected cases addresses are passed only to
copy_from()_user or copy_to_user().
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
No, you do NOT need to "protect copy from user" that way.
Incidentally, your userland ABI stinks. I understand that you
wanted to accept "reset" and "reset\n" as equivalent, but I suspect
that accepting "reset this, you !@^!@!" had been an accident.
Nothing to do about that now - it is a userland ABI...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Contrary to the comments, those do *NOT* verify anything about
writability of memory, etc.
In all cases addresses are passed only to copy_to_user().
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Address is passed to get_user_pages_fast(), which does access_ok().
NB: this is called only from ->ioctl(), and only under USER_DS.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
In general each DMA-based SPI transfer can be split up into two stages:
DMA data transmission/reception and SPI-bus transmission/reception. DMA
asynchronous transactions completion can be tracked by means of the
DMA async Tx-descriptor completion callback. But that callback being
called indicates that the DMA transfer has been finished, it doesn't
mean that SPI data transmission is also done. Moreover in fact it isn't
for at least Tx-only SPI transfers. Upon DMA transfer completion some
data is left in the Tx FIFO and being pushed out by the SPI controller.
So in order to make sure that an SPI transfer is completely pushed to the
SPI-bus, the driver has to wait for both DMA transaction and the SPI-bus
transmission/reception are finished. Note if there is a way to
asynchronously track the former event by means of the DMA async Tx
callback, there isn't easy one for the later (IRQ-based solution won't
work since SPI controller doesn't notify about Rx FIFO being empty).
The DMA transfer completion callback isn't suitable to wait for the
SPI controller activity finish either. The callback might (in case of DW
DMAC it will) be called in the tasklet context. Waiting for the SPI
controller to complete the transfer might take a considerable amount of
time since SPI-bus might be pretty slow. In this case delaying the
execution in the tasklet atomic context might cause significant system
performance drop.
So to speak the best option we've got to solve the problem is to
consequently wait for both stages being finished in the locally
implemented SPI transfer execution procedure even if it costs us of the
local wait-function re-implementation. In this case we don't need to use
the SPI-core transfer-wait functionality, but we'll make sure that
all DMA and SPI-bus transactions are completely finished before the
SPI-core transfer_one callback returns. In this commit we provide an
implementation of the DMA-transfers completion wait functionality.
The DW APB SSI DMA-specific SPI transfer_one function waits for both
Tx and Rx DMA transfers being finished, and only then exits with zero
returned signalling to the SPI core that the SPI transfer is finished.
This implementation is fully equivalent to the currently used
DMA-execution-SPI-core-wait algorithm. The SPI-bus transmission/reception
wait methods will be added in the follow-up commits.
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Cc: Georgy Vlasov <Georgy.Vlasov@baikalelectronics.ru>
Cc: Ramil Zaripov <Ramil.Zaripov@baikalelectronics.ru>
Cc: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Feng Tang <feng.tang@intel.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: linux-mips@vger.kernel.org
Cc: devicetree@vger.kernel.org
Link: https://lore.kernel.org/r/20200529131205.31838-4-Sergey.Semin@baikalelectronics.ru
Signed-off-by: Mark Brown <broonie@kernel.org>
