platform/kernel/linux-rpi.git
4 years agodt-bindings: dma: renesas,usb-dmac: add r8a77961 support
Yoshihiro Shimoda [Wed, 25 Mar 2020 06:18:58 +0000 (15:18 +0900)]
dt-bindings: dma: renesas,usb-dmac: add r8a77961 support

This patch adds support for r8a77961 (R-Car M3-W+).

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/1585117138-8408-1-git-send-email-yoshihiro.shimoda.uh@renesas.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: uniphier-xdmac: Remove redandant error log for platform_get_irq
Vinod Koul [Mon, 23 Mar 2020 17:19:28 +0000 (22:49 +0530)]
dmaengine: uniphier-xdmac: Remove redandant error log for platform_get_irq

platform_get_irq prints the error on failure, so there is no need to
have caller add a log.
Remove the log in uniphier_xdmac_probe() for platform_get_irq() failure

Reported-by: kbuild test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/20200323171928.424223-1-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Improve DMA synchronization
Dmitry Osipenko [Thu, 19 Mar 2020 21:23:21 +0000 (00:23 +0300)]
dmaengine: tegra-apb: Improve DMA synchronization

Boot CPU0 always handles DMA interrupts and under some rare circumstances
it could stuck in uninterruptible state for a significant time (like in a
case of KASAN + NFS root). In this case sibling CPU, which waits for DMA
transfer completion, will get a DMA transfer timeout. In order to handle
this rare condition, interrupt status needs to be polled until interrupt
is handled.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Link: https://lore.kernel.org/r/20200319212321.3297-2-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Don't save/restore IRQ flags in interrupt handler
Dmitry Osipenko [Thu, 19 Mar 2020 21:23:20 +0000 (00:23 +0300)]
dmaengine: tegra-apb: Don't save/restore IRQ flags in interrupt handler

The interrupt is already disabled while interrupt handler is running, and
thus, there is no need to save/restore the IRQ flags within the spinlock.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Link: https://lore.kernel.org/r/20200319212321.3297-1-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: mark PM functions as __maybe_unused
YueHaibing [Fri, 20 Mar 2020 07:13:37 +0000 (15:13 +0800)]
dmaengine: tegra-apb: mark PM functions as __maybe_unused

When CONFIG_PM is disabled, gcc warning this:

drivers/dma/tegra20-apb-dma.c:1587:12: warning: 'tegra_dma_runtime_resume' defined but not used [-Wunused-function]
drivers/dma/tegra20-apb-dma.c:1578:12: warning: 'tegra_dma_runtime_suspend' defined but not used [-Wunused-function]

Make it as __maybe_unused to fix the warnings,
also remove unneeded function declarations.

Fixes: ec8a1586780c ("dma: tegra: add dmaengine based dma driver")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20200320071337.59756-1-yuehaibing@huawei.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: fix spelling mistake "exceds" -> "exceeds"
Colin Ian King [Mon, 16 Mar 2020 09:16:53 +0000 (09:16 +0000)]
dmaengine: fix spelling mistake "exceds" -> "exceeds"

There are a couple of spelling mistakes in dev_err error messages.
Fix them.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20200316091653.110984-1-colin.king@canonical.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: sprd: Set request pending flag when DMA controller is active
Zhenfang Wang [Thu, 12 Mar 2020 13:26:04 +0000 (21:26 +0800)]
dmaengine: sprd: Set request pending flag when DMA controller is active

On new Spreadtrum platforms, when the CPU enters idle, it will close
the DMA controllers' clock to save power if the DMA controller is not
busy. Moreover the DMA controller's busy signal depends on the DMA
enable flag and the request pending flag.

When DMA controller starts to transfer data, which means we already
set the DMA enable flag, but now we should also set the request pending
flag, in case the DMA clock will be closed accidentally if the CPU
can not detect the DMA controller's busy signal.

Signed-off-by: Zhenfang Wang <zhenfang.wang@unisoc.com>
Signed-off-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/02adbe4364ec436ec2c5bc8fd2386bab98edd884.1584019223.git.baolin.wang7@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ppc4xx: Use scnprintf() for avoiding potential buffer overflow
Takashi Iwai [Wed, 11 Mar 2020 07:16:06 +0000 (08:16 +0100)]
dmaengine: ppc4xx: Use scnprintf() for avoiding potential buffer overflow

Since snprintf() returns the would-be-output size instead of the
actual output size, the succeeding calls may go beyond the given
buffer limit.  Fix it by replacing with scnprintf().

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: https://lore.kernel.org/r/20200311071606.4485-1-tiwai@suse.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: remove global token limit check
Dave Jiang [Tue, 10 Mar 2020 17:51:09 +0000 (10:51 -0700)]
dmaengine: idxd: remove global token limit check

The global token_limit is not tied to group tokens_reserved and
tokens_allowed parameters. Remove the check in order to allow independent
configuration.

Fixes: c52ca478233c ("dmaengine: idxd: add configuration component of driver")
Reported-by: Yixin Zhang <yixin.zhang@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/158386266911.11066.7545764533072221536.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: reflect shadow copy of traffic class programming
Dave Jiang [Tue, 10 Mar 2020 17:50:30 +0000 (10:50 -0700)]
dmaengine: idxd: reflect shadow copy of traffic class programming

The traffic class are set to -1 at initialization until the user programs
them. If the user choose not to, the driver will program appropriate
defaults. The driver also needs to update the shadowed copies of the values
after doing the programming.

Fixes: c52ca478233c ("dmaengine: idxd: add configuration component of driver")
Reported-by: Yixin Zhang <yixin.zhang@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/158386263076.10898.4586509576813094559.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: Merge definition of dsa_batch_desc into dsa_hw_desc
Tony Luck [Tue, 10 Mar 2020 22:18:02 +0000 (15:18 -0700)]
dmaengine: idxd: Merge definition of dsa_batch_desc into dsa_hw_desc

We don't need a special structure just for batch descriptors. The
layout matches the general form for other descriptors.

Merge the desc_list_addr field into the union of other aliases for
the the third quadword in the structure.

Create a union to alias "xfer_size" with "desc_count".

Signed-off-by: Tony Luck <tony.luck@intel.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/158387868208.35922.5895104426944263789.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: Create debug directories for DMA devices
Peter Ujfalusi [Fri, 6 Mar 2020 14:28:39 +0000 (16:28 +0200)]
dmaengine: Create debug directories for DMA devices

Create a placeholder directory for each registered DMA device.

DMA drivers can use the dmaengine_get_debugfs_root() call to get their
debugfs root and can populate with custom files to aim debugging.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200306142839.17910-4-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ti: k3-udma: Implement custom dbg_summary_show for debugfs
Peter Ujfalusi [Fri, 6 Mar 2020 14:28:38 +0000 (16:28 +0200)]
dmaengine: ti: k3-udma: Implement custom dbg_summary_show for debugfs

With the custom dbg_summary_show the driver can show useful information
about the used channels.

dma0 (285c0000.dma-controller): number of channels: 24

dma1 (31150000.dma-controller): number of channels: 84
 dma1chan0    | 2b00000.mcasp:tx (MEM_TO_DEV, tchan16 [0x1010 -> 0xc400], PDMA[ ACC32 BURST ], TR mode)
 dma1chan1    | 2b00000.mcasp:rx (DEV_TO_MEM, rchan16 [0x4400 -> 0x9010], PDMA[ ACC32 BURST ], TR mode)
 dma1chan2    | 2ba0000.mcasp:tx (MEM_TO_DEV, tchan17 [0x1011 -> 0xc507], PDMA[ ACC32 BURST ], TR mode)
 dma1chan3    | 2ba0000.mcasp:rx (DEV_TO_MEM, rchan17 [0x4507 -> 0x9011], PDMA[ ACC32 BURST ], TR mode)
 dma1chan4    | in-use (MEM_TO_MEM, chan0 pair [0x1000 -> 0x9000], PSI-L Native, TR mode)
 dma1chan5    | in-use (MEM_TO_MEM, chan1 pair [0x1001 -> 0x9001], PSI-L Native, TR mode)
 dma1chan6    | in-use (MEM_TO_MEM, chan4 pair [0x1004 -> 0x9004], PSI-L Native, TR mode)
 dma1chan7    | in-use (MEM_TO_MEM, chan5 pair [0x1005 -> 0x9005], PSI-L Native, TR mode)

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200306142839.17910-3-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: Add basic debugfs support
Peter Ujfalusi [Fri, 6 Mar 2020 14:28:37 +0000 (16:28 +0200)]
dmaengine: Add basic debugfs support

Via the /sys/kernel/debug/dmaengine/summary users can get information
about the DMA devices and the used channels.

Example output on am654-evm with audio using two channels and after running
dmatest on 4 channels:

dma0 (285c0000.dma-controller): number of channels: 96

dma1 (31150000.dma-controller): number of channels: 267
 dma1chan0    | 2b00000.mcasp:tx
 dma1chan1    | 2b00000.mcasp:rx
 dma1chan2    | in-use
 dma1chan3    | in-use
 dma1chan4    | in-use
 dma1chan5    | in-use

For slave channels we can show the device and the channel name a given
channel is requested.
For non slave devices the only information we know is that the channel is
in use.

DMA drivers can implement the optional dbg_summary_show callback to
provide controller specific information instead of the generic one.

It is easy to extend the generic dmaengine_summary_show() to print
additional information about the used channels.

I have taken the idea from gpiolib and clk subsystems.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200306142839.17910-2-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: fsl-dpaa2-qdma: remove set but not used variable 'dpaa2_qdma'
YueHaibing [Tue, 3 Mar 2020 13:13:47 +0000 (21:13 +0800)]
dmaengine: fsl-dpaa2-qdma: remove set but not used variable 'dpaa2_qdma'

drivers/dma/fsl-dpaa2-qdma/dpaa2-qdma.c: In function dpaa2_qdma_shutdown:
drivers/dma/fsl-dpaa2-qdma/dpaa2-qdma.c:795:28: warning: variable dpaa2_qdma set but not used [-Wunused-but-set-variable]

commit 3e0ca3c38dc2 ("dmaengine: fsl-dpaa2-qdma: Adding shutdown hook")
involved this, remove it.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20200303131347.28392-1-yuehaibing@huawei.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ti: edma: fix null dereference because of a typo in pointer name
Colin Ian King [Wed, 26 Feb 2020 18:59:21 +0000 (18:59 +0000)]
dmaengine: ti: edma: fix null dereference because of a typo in pointer name

Currently there is a dereference of the null pointer m_ddev.  This appears
to be a typo on the pointer, I believe s_ddev should be used instead.
Fix this by using the correct pointer.

Fixes: eb0249d50153 ("dmaengine: ti: edma: Support for interleaved mem to mem transfer")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Addresses-Coverity: ("Explicit null dereferenced")
Link: https://lore.kernel.org/r/20200226185921.351693-1-colin.king@canonical.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: fsl-dpaa2-qdma: Adding shutdown hook
Peng Ma [Thu, 27 Feb 2020 04:28:41 +0000 (12:28 +0800)]
dmaengine: fsl-dpaa2-qdma: Adding shutdown hook

We need to ensure DMA engine could be stopped in order for kexec
to start the next kernel.
So add the shutdown operation support.

Signed-off-by: Peng Ma <peng.ma@nxp.com>
Link: https://lore.kernel.org/r/20200227042841.18358-1-peng.ma@nxp.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: uniphier-xdmac: Add UniPhier external DMA controller driver
Kunihiko Hayashi [Fri, 21 Feb 2020 07:52:30 +0000 (16:52 +0900)]
dmaengine: uniphier-xdmac: Add UniPhier external DMA controller driver

This adds external DMA controller driver implemented in Socionext
UniPhier SoCs. This driver supports DMA_MEMCPY and DMA_SLAVE modes.

Since this driver does not support the the way to transfer size
unaligned to burst width, 'src_maxburst' or 'dst_maxburst' of
dma_slave_config must be 1 to transfer arbitrary size. If transfer
size is unaligned to burst size, the transfer isn't started and
the driver displays an error message.

Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Link: https://lore.kernel.org/r/1582271550-3403-3-git-send-email-hayashi.kunihiko@socionext.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodt-bindings: dmaengine: Add UniPhier external DMA controller bindings
Kunihiko Hayashi [Fri, 21 Feb 2020 07:52:29 +0000 (16:52 +0900)]
dt-bindings: dmaengine: Add UniPhier external DMA controller bindings

Add devicetree binding documentation for external DMA controller
implemented on Socionext UniPhier SOCs.

Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1582271550-3403-2-git-send-email-hayashi.kunihiko@socionext.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ti: k3-udma: Implement support for atype (for virtualization)
Peter Ujfalusi [Tue, 18 Feb 2020 14:31:26 +0000 (16:31 +0200)]
dmaengine: ti: k3-udma: Implement support for atype (for virtualization)

The DT for virtualized hosts have dma-cells == 2 where the second parameter
is the ATYPE for the channel.

In case of dma-cells == 1 we can configure the ATYPE as 0 (reset value).

The ATYPE defined for j721e are:
0: pointers are physical addresses (no translation)
1: pointers are intermediate addresses (PVU)
2: pointers are virtual addresses (SMMU)

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200218143126.11361-3-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodt-bindings: dma: ti: k3-udma: Update for atype support (virtualization)
Peter Ujfalusi [Tue, 18 Feb 2020 14:31:25 +0000 (16:31 +0200)]
dt-bindings: dma: ti: k3-udma: Update for atype support (virtualization)

In UDMA each channel can use different ATYPE value which tells UDMA how
the addresses in the descriptors should be treated:
0: pointers are physical addresses (no translation)
1: pointers are intermediate addresses (PVU)
2: pointers are virtual addresses (SMMU)

When virtualized environment is used then the dma binding should use
additional cell to configure the desired ATYPE for the channel.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20200218143126.11361-2-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: consistently return string literal from switch-case
Andy Shevchenko [Wed, 26 Feb 2020 10:18:41 +0000 (12:18 +0200)]
dmaengine: consistently return string literal from switch-case

There is no need to have 'break;' statement in the default case followed by
return certain string literal when all other cases have returned the string
literals. So, refactor it accordingly.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200226101842.29426-4-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: Drop redundant 'else' keyword
Andy Shevchenko [Wed, 26 Feb 2020 10:18:40 +0000 (12:18 +0200)]
dmaengine: Drop redundant 'else' keyword

It's obvious that 'else' keyword is redundant in the code like

if (foo)
return bar;
else if (baz)
...

Drop it for good.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20200226101842.29426-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: Use negative condition for better readability
Andy Shevchenko [Wed, 26 Feb 2020 10:18:39 +0000 (12:18 +0200)]
dmaengine: Use negative condition for better readability

When negative condition is in use we may decrease indentation level
and make the main part of logic better visible.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200226101842.29426-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: Refactor dmaengine_check_align() to be bit operations only
Andy Shevchenko [Wed, 26 Feb 2020 10:18:38 +0000 (12:18 +0200)]
dmaengine: Refactor dmaengine_check_align() to be bit operations only

There is no need to have branch and temporary variable in the function.
Simple convert it to be a set of bit and arithmetic operations.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200226101842.29426-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: expose general capabilities register in sysfs
Dave Jiang [Mon, 24 Feb 2020 18:01:34 +0000 (11:01 -0700)]
dmaengine: idxd: expose general capabilities register in sysfs

There are some capabilities for the device that are interesting to user
apps that are interacting directly with the device. Expose gencap register
in sysfs to allow that information.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/158256729399.55526.10842505054968710547.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: check return result from check_vma() in cdev
Dave Jiang [Tue, 25 Feb 2020 16:47:46 +0000 (09:47 -0700)]
dmaengine: idxd: check return result from check_vma() in cdev

The returned result from the check_vma() function in the cdev ->mmap() call
needs to be handled. Add the check and returning error.

Fixes: 42d279f9137a ("dmaengine: idxd: add char driver to expose submission portal to userland")
Reported-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/158264926659.9387.14325163515683047959.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Improve error message about DMA underflow
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:56 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Improve error message about DMA underflow

Technically it is possible that DMA could be misconfigured in a way that
cyclic DMA transfer is processed slower than it takes to complete the
cycle and in this case the DMA is getting aborted with a not very
informative message about the problem, let's improve it.

Suggested-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-20-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Remove unused function argument
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:55 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Remove unused function argument

Remove unused function argument from handle_continuous_head_request().

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-19-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Support COMPILE_TEST
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:54 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Support COMPILE_TEST

There is nothing arch-specific in the driver's code, so let's enable
compile-testing for the driver.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-18-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Remove MODULE_ALIAS
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:53 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Remove MODULE_ALIAS

Tegra APB DMA driver is an Open Firmware driver, so it uses OF alias
naming scheme which overrides MODULE_ALIAS, meaning that MODULE_ALIAS
does nothing and could be removed safely.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-17-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Allow to compile as a loadable kernel module
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:52 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Allow to compile as a loadable kernel module

The driver's removal was fixed by a recent commit and module load/unload
is working well now, tested on Tegra30.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-16-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Add missing of_dma_controller_free
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:51 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Add missing of_dma_controller_free

The DMA controller shall be released on driver's removal.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-15-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Clean up suspend-resume
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:50 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Clean up suspend-resume

It is enough to check whether hardware is busy on suspend and to reset
it across of suspend-resume because:

  1. Channel's configuration is fully re-programmed on each DMA
     transfer anyways.

  2. Context save-restore of an active channel won't end up well without
     pausing transfer prior to the context's saving, but note that every
     channel shall be idling at the time of suspend, so save-restore is
     not needed at all.

  3. The only case where context save-restore may be useful is when
     channel is in a paused state during suspend. But channel's pausing
     could be supported only on Tegra114+ and this functionality wasn't
     implemented by the driver for years now because there is no need for
     it in upstream kernel.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-14-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Keep clock enabled only during of DMA transfer
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:49 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Keep clock enabled only during of DMA transfer

It's a bit impractical to enable hardware's clock at the time of DMA
channel's allocation because most of DMA client drivers allocate DMA
channel at the time of the driver's probing, and thus, DMA clock is kept
always-enabled in practice, defeating the whole purpose of runtime PM.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-13-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Remove duplicated pending_sg_req checks
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:48 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Remove duplicated pending_sg_req checks

There are few place in the code which check whether pending_sg_req list is
empty despite of the check already being done. Let's remove the duplicated
checks to keep code clean.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-12-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Remove assumptions about unavailable runtime PM
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:47 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Remove assumptions about unavailable runtime PM

The runtime PM is always available on all Tegra SoCs since the commit
40b2bb1b132a ("ARM: tegra: enforce PM requirement"), so there is no
need to handle the case of unavailable RPM in the code anymore.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-11-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Remove unneeded initialization of tdc->config_init
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:46 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Remove unneeded initialization of tdc->config_init

There is no need to re-initialize the already initialized variables.
The tdc->config_init=false after driver's probe and after channel's
freeing, so there is no need to re-initialize it on the channel's
allocation.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-10-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Fix coding style problems
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:45 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Fix coding style problems

This patch fixes few dozens of coding style problems reported by
checkpatch and prettifies code where makes sense.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-9-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Use devm_request_irq
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:44 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Use devm_request_irq

Use resource-managed variant of request_irq for brevity.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-8-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Use devm_platform_ioremap_resource
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:43 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Use devm_platform_ioremap_resource

Use devm_platform_ioremap_resource to keep code cleaner a tad.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-7-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Clean up tasklet releasing
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:42 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Clean up tasklet releasing

There is no need to kill tasklet when driver's probe fails because tasklet
can't be scheduled at this time. It is also cleaner to kill tasklet on
channel's freeing rather than to kill it on driver's removal, otherwise
tasklet could perform a dummy execution after channel's releasing, which
isn't very nice.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-6-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Prevent race conditions on channel's freeing
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:41 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Prevent race conditions on channel's freeing

It's incorrect to check the channel's "busy" state without taking a lock.
That shouldn't cause any real troubles, nevertheless it's always better
not to have any race conditions in the code.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-5-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra-apb: Implement synchronization hook
Dmitry Osipenko [Sun, 9 Feb 2020 16:33:40 +0000 (19:33 +0300)]
dmaengine: tegra-apb: Implement synchronization hook

The ISR tasklet could be kept scheduled after DMA transfer termination,
let's add synchronization hook which blocks until tasklet is finished.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20200209163356.6439-4-digetx@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ti: edma: Support for interleaved mem to mem transfer
Peter Ujfalusi [Mon, 10 Feb 2020 09:44:55 +0000 (11:44 +0200)]
dmaengine: ti: edma: Support for interleaved mem to mem transfer

Add basic interleaved support via EDMA.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200210094455.3615-1-peter.ujfalusi@ti.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_xdmac: Fix locking in tasklet
Tudor Ambarus [Thu, 23 Jan 2020 14:03:17 +0000 (14:03 +0000)]
dmaengine: at_xdmac: Fix locking in tasklet

Tasklets run with all the interrupts enabled. This means that we should
replace all the (already present) spin_lock_irqsave() uses in the tasklet
with spin_lock_irq() to protect being interrupted by a IRQ which tries
to get the same lock (via calls to device_prep_dma_* for example).

spin_lock and spin_lock_bh in tasklets are not enough to protect from IRQs,
update these to spin_lock_irq().

at_xdmac_advance_work() can be called with all the interrupts enabled (when
called from tasklet), or with interrupts disabled (when called from
at_xdmac_issue_pending). Move the locking in the callers to be able to use
spin_lock_irq() and spin_lock_irqsave() for these cases.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-10-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_xdmac: GFP_KERNEL for user that can sleep
Tudor Ambarus [Thu, 23 Jan 2020 14:03:15 +0000 (14:03 +0000)]
dmaengine: at_xdmac: GFP_KERNEL for user that can sleep

device_alloc_chan_resources can sleep, use GFP_KERNEL flag.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-9-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_xdmac: Drop locking in at_xdmac_alloc_chan_resources()
Tudor Ambarus [Thu, 23 Jan 2020 14:03:14 +0000 (14:03 +0000)]
dmaengine: at_xdmac: Drop locking in at_xdmac_alloc_chan_resources()

There is no need for locking in device_alloc_chan_resources(),
the DMA core takes care of it by using a dma_list_mutex around
the DMA devices.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-8-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_xdmac: Drop always true check
Tudor Ambarus [Thu, 23 Jan 2020 14:03:12 +0000 (14:03 +0000)]
dmaengine: at_xdmac: Drop always true check

The code in cause is already in the else case of
'if (at_xdmac_chan_is_cyclic(atchan))', drop the redundant check.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-7-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_hdmac: Fix deadlocks
Tudor Ambarus [Thu, 23 Jan 2020 14:03:11 +0000 (14:03 +0000)]
dmaengine: at_hdmac: Fix deadlocks

Fix the following deadlocks:
1/ atc_handle_cyclic() and atc_chain_complete() called
dmaengine_desc_get_callback_invoke() while wrongly holding the
atchan->lock. Clients can set the callback to dmaengine_terminate_sync()
which will end up trying to get the same lock, thus a deadlock occurred.
2/ dma_run_dependencies() was called with the atchan->lock held, but the
method calls device_issue_pending() which tries to get the same lock,
and so a deadlock occurred.

The driver must not hold the lock when invoking the callback or when
running dependencies. Releasing the spinlock within a called function
before calling the callback is not a nice thing to do -> called functions
become non-atomic when called within an atomic region. Thus the lock is
now taken in the child routines whereever is needed.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-6-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_hdmac: Switch atomic allocations to GFP_NOWAIT
Tudor Ambarus [Thu, 23 Jan 2020 14:03:09 +0000 (14:03 +0000)]
dmaengine: at_hdmac: Switch atomic allocations to GFP_NOWAIT

Avoids sleeping without depleting the emergency pool.
The rationale being that in most cases a dma device is either
offloading an operation that will automatically fallback to
software when the descriptor allocation fails, or we can simply poll
and wait for the dma device to release some in use descriptors.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-5-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_hdmac: Drop description for a not defined parameter
Tudor Ambarus [Thu, 23 Jan 2020 14:03:07 +0000 (14:03 +0000)]
dmaengine: at_hdmac: Drop description for a not defined parameter

Probably a leftover, drop it.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-4-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_hdmac: Return err in case the chan is not free at alloc res time
Tudor Ambarus [Thu, 23 Jan 2020 14:03:06 +0000 (14:03 +0000)]
dmaengine: at_hdmac: Return err in case the chan is not free at alloc res time

Having a list of descriptors allocated for the channel at
device_alloc_chan_resources() time is a sign for bad free usage.
Return err and add a debug message in case the channel is not
free from a previous use.

atchan->descs_allocated becomes useless, get rid of it. More,
drop the error message in atc_desc_get() because now it would
introduce an extra if statement. The callers of atc_desc_get()
already print error messages in case the callee fails, no one
is hurt.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-3-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_hdmac: Drop locking in at_hdmac_alloc_chan_resources()
Tudor Ambarus [Thu, 23 Jan 2020 14:03:04 +0000 (14:03 +0000)]
dmaengine: at_hdmac: Drop locking in at_hdmac_alloc_chan_resources()

There is no need for locking in device_alloc_chan_resources(),
the DMA core takes care of it by using a dma_list_mutex around
the DMA devices.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-2-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: at_hdmac: Substitute kzalloc with kmalloc
Tudor Ambarus [Thu, 23 Jan 2020 14:03:02 +0000 (14:03 +0000)]
dmaengine: at_hdmac: Substitute kzalloc with kmalloc

All members of the structure are initialized below in the function,
there is no need to use kzalloc.

Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200123140237.125799-1-tudor.ambarus@microchip.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: xilinx_dma: In dma channel probe fix node order dependency
Radhey Shyam Pandey [Thu, 30 Jan 2020 12:54:25 +0000 (18:24 +0530)]
dmaengine: xilinx_dma: In dma channel probe fix node order dependency

In overlay application we noticed that dma channel node probe order is
inverted i.e s2mm channel is probed first followed by mm2s channel. The
reason for this inversion is fdtoverlay utility which uses a function
called fdt_add_subnode(*). It stores the subnodes after the properties,
this has the effect of inserting the new subnode before any others and
the end result is a reversal.

Because of this inverted channel probe order, the node probed first is
assigned a '0' index instead of Channel ID should be '0' for tx and '1'
for rx and dmatest client using the DT convention fails in dma transfer
as channel are swapped.

To fix above behavior and make channel assignment index independent
of probe order, always assign mm2s channel at '0' index and the s2mm
channel at IP specific fixed offset derived from the max_channels
count.

Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Link: https://lore.kernel.org/r/1580388865-9960-3-git-send-email-radhey.shyam.pandey@xilinx.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: xilinx_dma: Extend dma_config structure to store max channel count
Radhey Shyam Pandey [Thu, 30 Jan 2020 12:54:24 +0000 (18:24 +0530)]
dmaengine: xilinx_dma: Extend dma_config structure to store max channel count

Extend dma_config structure to store the max channel count. This input is
used to populate dma device channel nodes at the fixed offset. It serves
as a preparatory patch for removing dma channel DT node order dependency,
added in the subsequent commit.

Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Link: https://lore.kernel.org/r/1580388865-9960-2-git-send-email-radhey.shyam.pandey@xilinx.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: use vchan_terminate_vdesc() in .terminate_all
Amelie Delaunay [Wed, 29 Jan 2020 15:36:28 +0000 (16:36 +0100)]
dmaengine: stm32-dma: use vchan_terminate_vdesc() in .terminate_all

To avoid race with vchan_complete, use the race free way to terminate
running transfer.

Move vdesc->node list_del in stm32_dma_start_transfer instead of in
stm32_mdma_chan_complete to avoid another race in vchan_dma_desc_free_list.

Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-9-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: fix sleeping function called from invalid context
Amelie Delaunay [Wed, 29 Jan 2020 15:36:27 +0000 (16:36 +0100)]
dmaengine: stm32-dma: fix sleeping function called from invalid context

This patch fixes BUG: sleeping function called from invalid context in
stm32_dma_disable_chan function.

The goal of this function is to force channel disable if it has not been
disabled by hardware. This consists in clearing STM32_DMA_SCR_EN bit and
read it as 0 to ensure the channel is well disabled and the last transfer
is over.

In previous implementation, the waiting loop was based on a do...while (1)
with a call to cond_resched to give the scheduler a chance to run a higher
priority process.

But in some conditions, stm32_dma_disable_chan can be called while
preemption is disabled, on a stm32_dma_stop call for example. So
cond_resched must not be used.

To avoid this, use readl_relaxed_poll_timeout_atomic to poll
STM32_DMA_SCR_EN bit cleared.

Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-8-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: add copy_align constraint
Amelie Delaunay [Wed, 29 Jan 2020 15:36:26 +0000 (16:36 +0100)]
dmaengine: stm32-dma: add copy_align constraint

This patch adds copy_align property in accordance with hardware
restriction.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-7-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: use dma_set_max_seg_size to set the sg limit
Amelie Delaunay [Wed, 29 Jan 2020 15:36:25 +0000 (16:36 +0100)]
dmaengine: stm32-dma: use dma_set_max_seg_size to set the sg limit

This patch adds dma_set_max_seg_size to define sg dma constraint.
This constraint may be taken into account by client to scatter/gather
its buffer.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-6-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: enable descriptor_reuse
Pierre-Yves MORDRET [Wed, 29 Jan 2020 15:36:24 +0000 (16:36 +0100)]
dmaengine: stm32-dma: enable descriptor_reuse

Enable client to resubmit already processed descriptors
in order to save descriptor creation time.

Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-5-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: driver defers probe for reset
Etienne Carriere [Wed, 29 Jan 2020 15:36:23 +0000 (16:36 +0100)]
dmaengine: stm32-dma: driver defers probe for reset

Change STM32 DMA driver to defer its probe operation when reset
controller is expected but has not been probed yet when DMA
device is probed.

Changes error traces when failing to get a system resource so that
it is not printed on failure with deferred probing.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-4-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: use reset controller only at probe time
Etienne Carriere [Wed, 29 Jan 2020 15:36:22 +0000 (16:36 +0100)]
dmaengine: stm32-dma: use reset controller only at probe time

Remove reset controller reference from device instance since it is
used only at probe time.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-3-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dma: add suspend/resume power management support
Pierre-Yves MORDRET [Wed, 29 Jan 2020 15:36:21 +0000 (16:36 +0100)]
dmaengine: stm32-dma: add suspend/resume power management support

Add suspend/resume power management relying on PM Runtime engine.

Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200129153628.29329-2-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: xilinx_dma: Reset DMA channel in dma_terminate_all
Radhey Shyam Pandey [Wed, 29 Jan 2020 07:45:09 +0000 (13:15 +0530)]
dmaengine: xilinx_dma: Reset DMA channel in dma_terminate_all

Reset DMA channel after stop to ensure that pending transfers and FIFOs
in the datapath are flushed or completed. It also cleanup the terminate
path and removes stop for the cyclic mode as after the reset stop is not
required. This fixes intermittent data verification failure when xilinx
dma test the client is stressed and loaded/unloaded multiple times.

Signed-off-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Link: https://lore.kernel.org/r/1580283909-32678-1-git-send-email-radhey.shyam.pandey@xilinx.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dmamux: driver defers probe for clock and reset
Etienne Carriere [Tue, 28 Jan 2020 09:41:58 +0000 (10:41 +0100)]
dmaengine: stm32-dmamux: driver defers probe for clock and reset

Changes STM32 DMAMUX driver to defer its probe operation when
reset controller is expected but has not been probed yet.

Changes error traces when failing to get a system resource so that
it is not printed on failure with deferred probing.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200128094158.20361-5-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dmamux: use reset controller only at probe time
Etienne Carriere [Tue, 28 Jan 2020 09:41:57 +0000 (10:41 +0100)]
dmaengine: stm32-dmamux: use reset controller only at probe time

Remove reset controller reference from device instance since it is
used only at probe time.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200128094158.20361-4-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dmamux: fix clock handling in probe sequence
Etienne Carriere [Tue, 28 Jan 2020 09:41:56 +0000 (10:41 +0100)]
dmaengine: stm32-dmamux: fix clock handling in probe sequence

This change ensures the DMAMUX device is reset only once it is clocked
and that clock is released in a safe state when probe operation fails.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200128094158.20361-3-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-dmamux: add suspend/resume power management support
Pierre-Yves MORDRET [Tue, 28 Jan 2020 09:41:55 +0000 (10:41 +0100)]
dmaengine: stm32-dmamux: add suspend/resume power management support

Add suspend/resume power management relying on PM Runtime engine.

Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200128094158.20361-2-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-mdma: use vchan_terminate_vdesc() in .terminate_all
Amelie Delaunay [Mon, 27 Jan 2020 08:53:34 +0000 (09:53 +0100)]
dmaengine: stm32-mdma: use vchan_terminate_vdesc() in .terminate_all

To avoid race with vchan_complete, use the race free way to terminate
running transfer.

Move vdesc->node list_del in stm32_mdma_start_transfer instead of in
stm32_mdma_xfer_end to avoid another race in vchan_dma_desc_free_list.

Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200127085334.13163-7-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-mdma: enable descriptor_reuse
Pierre-Yves MORDRET [Mon, 27 Jan 2020 08:53:33 +0000 (09:53 +0100)]
dmaengine: stm32-mdma: enable descriptor_reuse

Enable descriptor reuse to allow client to resubmit already processed
descriptors in order to save descriptor creation time.

Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200127085334.13163-6-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-mdma: driver defers probe for clock and reset
Amelie Delaunay [Mon, 27 Jan 2020 08:53:32 +0000 (09:53 +0100)]
dmaengine: stm32-mdma: driver defers probe for clock and reset

This patch changes error log when failing to get the clock so that it is
not printed on failure with probe deferring.

It also defers probe when reset controller is expected but has not been
probed yet when MDMA device is probed.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200127085334.13163-5-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-mdma: disable clock in case of error during probe
Etienne Carriere [Mon, 27 Jan 2020 08:53:31 +0000 (09:53 +0100)]
dmaengine: stm32-mdma: disable clock in case of error during probe

This patch disables the clock in case of error during probe. The unneeded
err_unregister label is renamed err_clk instead.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200127085334.13163-4-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-mdma: use reset controller only at probe time
Etienne Carriere [Mon, 27 Jan 2020 08:53:30 +0000 (09:53 +0100)]
dmaengine: stm32-mdma: use reset controller only at probe time

Remove reset controller reference from device instance since it is
used only at probe time.

Signed-off-by: Etienne Carriere <etienne.carriere@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200127085334.13163-3-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: stm32-mdma: add suspend/resume power management support
Pierre-Yves MORDRET [Mon, 27 Jan 2020 08:53:29 +0000 (09:53 +0100)]
dmaengine: stm32-mdma: add suspend/resume power management support

Add suspend/resume power management relying on PM Runtime engine.

Signed-off-by: Pierre-Yves MORDRET <pierre-yves.mordret@st.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@st.com>
Link: https://lore.kernel.org/r/20200127085334.13163-2-amelie.delaunay@st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmanegine: ioat/dca: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Fri, 14 Feb 2020 17:13:02 +0000 (11:13 -0600)]
dmanegine: ioat/dca: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Link: https://lore.kernel.org/r/20200214171302.GA20586@embeddedor
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: tegra210-adma: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Fri, 14 Feb 2020 17:16:57 +0000 (11:16 -0600)]
dmaengine: tegra210-adma: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Link: https://lore.kernel.org/r/20200214171657.GA25663@embeddedor
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: sprd: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Fri, 14 Feb 2020 17:15:36 +0000 (11:15 -0600)]
dmaengine: sprd: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20200214171536.GA24077@embeddedor
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: sa11x0: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Fri, 14 Feb 2020 17:14:35 +0000 (11:14 -0600)]
dmaengine: sa11x0: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Link: https://lore.kernel.org/r/20200214171435.GA22930@embeddedor
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: sun4i: set the linear_mode properly
Vinod Koul [Fri, 14 Feb 2020 04:46:09 +0000 (10:16 +0530)]
dmaengine: sun4i: set the linear_mode properly

Commit 6ebb827f7aad ("dmaengine: sun4i: use 'linear_mode' in
sun4i_dma_prep_dma_cyclic") updated the condition but introduced a semi
colon this making this statement have no effect, so add the bitwise OR
to fix it"

Fixes: 6ebb827f7aad ("dmaengine: sun4i: use 'linear_mode' in sun4i_dma_prep_dma_cyclic")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Link: https://lore.kernel.org/r/20200214044609.2215861-1-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ti: dma-crossbar: convert to devm_platform_ioremap_resource()
chenqiwu [Tue, 28 Jan 2020 05:35:46 +0000 (13:35 +0800)]
dmaengine: ti: dma-crossbar: convert to devm_platform_ioremap_resource()

Use a new API devm_platform_ioremap_resource() to simplify code.

Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/1580189746-2864-1-git-send-email-qiwuchen55@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: sun4i: use 'linear_mode' in sun4i_dma_prep_dma_cyclic
YueHaibing [Fri, 7 Feb 2020 02:44:45 +0000 (10:44 +0800)]
dmaengine: sun4i: use 'linear_mode' in sun4i_dma_prep_dma_cyclic

drivers/dma/sun4i-dma.c: In function sun4i_dma_prep_dma_cyclic:
drivers/dma/sun4i-dma.c:672:24: warning:
 variable linear_mode set but not used [-Wunused-but-set-variable]

commit ffc079a4accc ("dmaengine: sun4i: Add support for cyclic requests with dedicated DMA")
involved this, explicitly using the value makes the code more readable.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20200207024445.44600-1-yuehaibing@huawei.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: ti: omap-dma: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Thu, 13 Feb 2020 00:39:25 +0000 (18:39 -0600)]
dmaengine: ti: omap-dma: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200213003925.GA6906@embeddedor.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: uniphier-mdmac: replace zero-length array with flexible-array member
Gustavo A. R. Silva [Thu, 13 Feb 2020 00:35:35 +0000 (18:35 -0600)]
dmaengine: uniphier-mdmac: replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Link: https://lore.kernel.org/r/20200213003535.GA3269@embeddedor.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: bcm-sba-raid: Replace zero-length array with flexible-array member
Gustavo A. R. Silva [Thu, 13 Feb 2020 00:37:03 +0000 (18:37 -0600)]
dmaengine: bcm-sba-raid: Replace zero-length array with flexible-array member

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Link: https://lore.kernel.org/r/20200213003703.GA4177@embeddedor.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodt-bindings: dma: ti-edma: fix example compatible property
Johan Hovold [Wed, 12 Feb 2020 10:48:40 +0000 (11:48 +0100)]
dt-bindings: dma: ti-edma: fix example compatible property

Make sure that the compatible string in the edma1_tptc1 example node
matches the binding by removing the space between the manufacturer and
model.

Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20200212104840.20393-1-johan@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: remove set but not used variable 'idxd_cdev'
YueHaibing [Mon, 10 Feb 2020 15:18:55 +0000 (23:18 +0800)]
dmaengine: idxd: remove set but not used variable 'idxd_cdev'

drivers/dma/idxd/cdev.c: In function idxd_cdev_open:
drivers/dma/idxd/cdev.c:77:20: warning:
 variable idxd_cdev set but not used [-Wunused-but-set-variable]

commit 42d279f9137a ("dmaengine: idxd: add char driver to
expose submission portal to userland") involed this.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20200210151855.55044-1-yuehaibing@huawei.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
4 years agodmaengine: idxd: remove set but not used variable 'group'
YueHaibing [Tue, 11 Feb 2020 13:53:35 +0000 (21:53 +0800)]
dmaengine: idxd: remove set but not used variable 'group'

drivers/dma/idxd/sysfs.c: In function engine_group_id_store:
drivers/dma/idxd/sysfs.c:419:29: warning: variable group set but not used [-Wunused-but-set-variable]

It is not used, so remove it.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20200211135335.55924-1-yuehaibing@huawei.com
Signed-Off-By: Vinod Koul <vkoul@kernel.org>
4 years agoLinux 5.6-rc1
Linus Torvalds [Mon, 10 Feb 2020 00:08:48 +0000 (16:08 -0800)]
Linux 5.6-rc1

4 years agoMerge tag 'kbuild-v5.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy...
Linus Torvalds [Mon, 10 Feb 2020 00:05:50 +0000 (16:05 -0800)]
Merge tag 'kbuild-v5.6-2' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull more Kbuild updates from Masahiro Yamada:

 - fix randconfig to generate a sane .config

 - rename hostprogs-y / always to hostprogs / always-y, which are more
   natual syntax.

 - optimize scripts/kallsyms

 - fix yes2modconfig and mod2yesconfig

 - make multiple directory targets ('make foo/ bar/') work

* tag 'kbuild-v5.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kbuild: make multiple directory targets work
  kconfig: Invalidate all symbols after changing to y or m.
  kallsyms: fix type of kallsyms_token_table[]
  scripts/kallsyms: change table to store (strcut sym_entry *)
  scripts/kallsyms: rename local variables in read_symbol()
  kbuild: rename hostprogs-y/always to hostprogs/always-y
  kbuild: fix the document to use extra-y for vmlinux.lds
  kconfig: fix broken dependency in randconfig-generated .config

4 years agoMerge tag 'zonefs-5.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal...
Linus Torvalds [Sun, 9 Feb 2020 23:51:46 +0000 (15:51 -0800)]
Merge tag 'zonefs-5.6-rc1' of git://git./linux/kernel/git/dlemoal/zonefs

Pull new zonefs file system from Damien Le Moal:
 "Zonefs is a very simple file system exposing each zone of a zoned
  block device as a file.

  Unlike a regular file system with native zoned block device support
  (e.g. f2fs or the on-going btrfs effort), zonefs does not hide the
  sequential write constraint of zoned block devices to the user. As a
  result, zonefs is not a POSIX compliant file system. Its goal is to
  simplify the implementation of zoned block devices support in
  applications by replacing raw block device file accesses with a richer
  file based API, avoiding relying on direct block device file ioctls
  which may be more obscure to developers.

  One example of this approach is the implementation of LSM
  (log-structured merge) tree structures (such as used in RocksDB and
  LevelDB) on zoned block devices by allowing SSTables to be stored in a
  zone file similarly to a regular file system rather than as a range of
  sectors of a zoned device. The introduction of the higher level
  construct "one file is one zone" can help reducing the amount of
  changes needed in the application while at the same time allowing the
  use of zoned block devices with various programming languages other
  than C.

  Zonefs IO management implementation uses the new iomap generic code.
  Zonefs has been successfully tested using a functional test suite
  (available with zonefs userland format tool on github) and a prototype
  implementation of LevelDB on top of zonefs"

* tag 'zonefs-5.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs:
  zonefs: Add documentation
  fs: New zonefs file system

4 years agoirqchip/gic-v4.1: Avoid 64bit division for the sake of 32bit ARM
Marc Zyngier [Sun, 9 Feb 2020 22:48:50 +0000 (22:48 +0000)]
irqchip/gic-v4.1: Avoid 64bit division for the sake of 32bit ARM

In order to allow the GICv4 code to link properly on 32bit ARM,
make sure we don't use 64bit divisions when it isn't strictly
necessary.

Fixes: 4e6437f12d6e ("irqchip/gic-v4.1: Ensure L2 vPE table is allocated at RD level")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoMerge tag '5.6-rc-smb3-plugfest-patches' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 9 Feb 2020 21:27:17 +0000 (13:27 -0800)]
Merge tag '5.6-rc-smb3-plugfest-patches' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "13 cifs/smb3 patches, most from testing at the SMB3 plugfest this week:

   - Important fix for multichannel and for modefromsid mounts.

   - Two reconnect fixes

   - Addition of SMB3 change notify support

   - Backup tools fix

   - A few additional minor debug improvements (tracepoints and
     additional logging found useful during testing this week)"

* tag '5.6-rc-smb3-plugfest-patches' of git://git.samba.org/sfrench/cifs-2.6:
  smb3: Add defines for new information level, FileIdInformation
  smb3: print warning once if posix context returned on open
  smb3: add one more dynamic tracepoint missing from strict fsync path
  cifs: fix mode bits from dir listing when mounted with modefromsid
  cifs: fix channel signing
  cifs: add SMB3 change notification support
  cifs: make multichannel warning more visible
  cifs: fix soft mounts hanging in the reconnect code
  cifs: Add tracepoints for errors on flush or fsync
  cifs: log warning message (once) if out of disk space
  cifs: fail i/o on soft mounts if sessionsetup errors out
  smb3: fix problem with null cifs super block with previous patch
  SMB3: Backup intent flag missing from some more ops

4 years agoMerge branch 'work.vboxsf' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Sun, 9 Feb 2020 20:41:00 +0000 (12:41 -0800)]
Merge branch 'work.vboxsf' of git://git./linux/kernel/git/viro/vfs

Pull vboxfs from Al Viro:
 "This is the VirtualBox guest shared folder support by Hans de Goede,
  with fixups for fs_parse folded in to avoid bisection hazards from
  those API changes..."

* 'work.vboxsf' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  fs: Add VirtualBox guest shared folder (vboxsf) support

4 years agoMerge tag 'x86-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Feb 2020 20:11:12 +0000 (12:11 -0800)]
Merge tag 'x86-urgent-2020-02-09' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Thomas Gleixner:
 "A set of fixes for X86:

   - Ensure that the PIT is set up when the local APIC is disable or
     configured in legacy mode. This is caused by an ordering issue
     introduced in the recent changes which skip PIT initialization when
     the TSC and APIC frequencies are already known.

   - Handle malformed SRAT tables during early ACPI parsing which caused
     an infinite loop anda boot hang.

   - Fix a long standing race in the affinity setting code which affects
     PCI devices with non-maskable MSI interrupts. The problem is caused
     by the non-atomic writes of the MSI address (destination APIC id)
     and data (vector) fields which the device uses to construct the MSI
     message. The non-atomic writes are mandated by PCI.

     If both fields change and the device raises an interrupt after
     writing address and before writing data, then the MSI block
     constructs a inconsistent message which causes interrupts to be
     lost and subsequent malfunction of the device.

     The fix is to redirect the interrupt to the new vector on the
     current CPU first and then switch it over to the new target CPU.
     This allows to observe an eventually raised interrupt in the
     transitional stage (old CPU, new vector) to be observed in the APIC
     IRR and retriggered on the new target CPU and the new vector.

     The potential spurious interrupts caused by this are harmless and
     can in the worst case expose a buggy driver (all handlers have to
     be able to deal with spurious interrupts as they can and do happen
     for various reasons).

   - Add the missing suspend/resume mechanism for the HYPERV hypercall
     page which prevents resume hibernation on HYPERV guests. This
     change got lost before the merge window.

   - Mask the IOAPIC before disabling the local APIC to prevent
     potentially stale IOAPIC remote IRR bits which cause stale
     interrupt lines after resume"

* tag 'x86-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/apic: Mask IOAPIC entries when disabling the local APIC
  x86/hyperv: Suspend/resume the hypercall page for hibernation
  x86/apic/msi: Plug non-maskable MSI affinity race
  x86/boot: Handle malformed SRAT tables during early ACPI parsing
  x86/timer: Don't skip PIT setup when APIC is disabled or in legacy mode

4 years agoMerge tag 'smp-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Feb 2020 20:09:43 +0000 (12:09 -0800)]
Merge tag 'smp-urgent-2020-02-09' of git://git./linux/kernel/git/tip/tip

Pull SMP fixes from Thomas Gleixner:
 "Two fixes for the SMP related functionality:

   - Make the UP version of smp_call_function_single() match SMP
     semantics when called for a not available CPU. Instead of emitting
     a warning and assuming that the function call target is CPU0,
     return a proper error code like the SMP version does.

   - Remove a superfluous check in smp_call_function_many_cond()"

* tag 'smp-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  smp/up: Make smp_call_function_single() match SMP semantics
  smp: Remove superfluous cond_func check in smp_call_function_many_cond()

4 years agoMerge tag 'perf-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Feb 2020 20:04:09 +0000 (12:04 -0800)]
Merge tag 'perf-urgent-2020-02-09' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Thomas Gleixner:
 "A set of fixes and improvements for the perf subsystem:

  Kernel fixes:

   - Install cgroup events to the correct CPU context to prevent a
     potential list double add

   - Prevent an integer underflow in the perf mlock accounting

   - Add a missing prototype for arch_perf_update_userpage()

  Tooling:

   - Add a missing unlock in the error path of maps__insert() in perf
     maps.

   - Fix the build with the latest libbfd

   - Fix the perf parser so it does not delete parse event terms, which
     caused a regression for using perf with the ARM CoreSight as the
     sink configuration was missing due to the deletion.

   - Fix the double free in the perf CPU map merging test case

   - Add the missing ustring support for the perf probe command"

* tag 'perf-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf maps: Add missing unlock to maps__insert() error case
  perf probe: Add ustring support for perf probe command
  perf: Make perf able to build with latest libbfd
  perf test: Fix test case Merge cpu map
  perf parse: Copy string to perf_evsel_config_term
  perf parse: Refactor 'struct perf_evsel_config_term'
  kernel/events: Add a missing prototype for arch_perf_update_userpage()
  perf/cgroups: Install cgroup events to correct cpuctx
  perf/core: Fix mlock accounting in perf_mmap()

4 years agoMerge tag 'timers-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 9 Feb 2020 20:00:12 +0000 (12:00 -0800)]
Merge tag 'timers-urgent-2020-02-09' of git://git./linux/kernel/git/tip/tip

Pull timer fixes from Thomas Gleixner:
 "Two small fixes for the time(r) subsystem:

   - Handle a subtle race between the clocksource watchdog and a
     concurrent clocksource watchdog stop/start sequence correctly to
     prevent a timer double add bug.

   - Fix the file path for the core time namespace file"

* tag 'timers-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  clocksource: Prevent double add_timer_on() for watchdog_timer
  MAINTAINERS: Correct path to time namespace source file

4 years agoMerge tag 'irq-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Feb 2020 19:56:41 +0000 (11:56 -0800)]
Merge tag 'irq-urgent-2020-02-09' of git://git./linux/kernel/git/tip/tip

Pull interrupt fixes from Thomas Gleixner:
 "A set of fixes for the interrupt subsystem:

   - Provision only ACPI enabled redistributors on GICv3

   - Use the proper command colums when building the INVALL command for
     the GICv3-ITS

   - Ensure the allocation of the L2 vPE table for GICv4.1

   - Correct the GICv4.1 VPROBASER programming so it uses the proper
     size

   - A set of small GICv4.1 tidy up patches

   - Configuration cleanup for C-SKY interrupt chip

   - Clarify the function documentation for irq_set_wake() to document
     that the wakeup functionality is orthogonal to the irq
     disable/enable mechanism"

* tag 'irq-urgent-2020-02-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqchip/gic-v3-its: Rename VPENDBASER/VPROPBASER accessors
  irqchip/gic-v3-its: Remove superfluous WARN_ON
  irqchip/gic-v4.1: Drop 'tmp' in inherit_vpe_l1_table_from_rd()
  irqchip/gic-v4.1: Ensure L2 vPE table is allocated at RD level
  irqchip/gic-v4.1: Set vpe_l1_base for all redistributors
  irqchip/gic-v4.1: Fix programming of GICR_VPROPBASER_4_1_SIZE
  genirq: Clarify that irq wake state is orthogonal to enable/disable
  irqchip/gic-v3-its: Reference to its_invall_cmd descriptor when building INVALL
  irqchip: Some Kconfig cleanup for C-SKY
  irqchip/gic-v3: Only provision redistributors that are enabled in ACPI