Now that mv_xor_slot_cleanup() has no remaining callers, we remove it
and rename __mv_xor_slot_cleanup() to mv_xor_slot_cleanup().

We take this opportunity to add a comment that makes it clear that the
channel spinlock should be held before calling mv_xor_slot_cleanup().
Signed-off-by: NThomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

In order to simplify the code, remove all the calls to the locked
mv_xor_slot_cleanup() and instead use the unlocked version only,

It's less error prone to have just one function, and require the caller
to ensure proper locking.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

In mv_xor_status(), we are currently calling mv_xor_clean_completed_slots()
when the transaction is complete (the cookie status is DMA_COMPLETE).
However, a completed status means that mv_xor_slot_cleanup() was called,
which cleans the completed slots.

In other words, there's nothing to cleanup for a completed transaction in
mv_xor_status(). Remove the unneeded call to mv_xor_clean_completed_slots().
Reported-by: NDan Williams <dan.j.williams@intel.com>
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range()  or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.

Function pci_enable_msix() returns a tri-state value while
pci_enable_msi_exact() is a canonical zero/-errno variant.
The former is being phased out in favor of the latter.
In case of 'ioat' there (should be) no difference.

Cc: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: NAlexander Gordeev <agordeev@redhat.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Includes an appropriate header file dma_v2.h in ioat/dca.c because
functions ioat2_dca_init() and ioat3_dca_init() have their function
declarations in dma_v2.h.

This eliminates the following warning in ioat/dca.c:
drivers/dma/ioat/dca.c:410:22: warning: no previous prototype for ‘ioat2_dca_init’ [-Wmissing-prototypes]
drivers/dma/ioat/dca.c:624:22: warning: no previous prototype for ‘ioat3_dca_init’ [-Wmissing-prototypes]
Signed-off-by: NRashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: NJosh Triplett <josh@joshtriplett.org>
Acked-by: NVinod Koul <vinod.koul@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Mark the functions ioat3_prep_xor_val(), ioat3_prep_pq_val() and
ioat3_prep_pqxor_val() as static in dma_v3.c because they are not used
outside this file.

This eliminates the following warnings in dma_v3.c:
drivers/dma/ioat/dma_v3.c:741:1: warning: no previous prototype for ‘ioat3_prep_xor_val’ [-Wmissing-prototypes]
drivers/dma/ioat/dma_v3.c:1092:1: warning: no previous prototype for ‘ioat3_prep_pq_val’ [-Wmissing-prototypes]
drivers/dma/ioat/dma_v3.c:1134:1: warning: no previous prototype for ‘ioat3_prep_pqxor_val’ [-Wmissing-prototypes]
Signed-off-by: NRashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: NJosh Triplett <josh@joshtriplett.org>
Acked-by: NVinod Koul <vinod.koul@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

This commit adds proper error checking for various DMA API calls,
as reported by DMA_API_DEBUG=y.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Use PCI standard marco dev_is_pci() instead of directly compare
pci_bus_type to check whether it is pci device.
Signed-off-by: NYijing Wang <wangyijing@huawei.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

Since commit 77873803 "net_dma: mark broken" we no longer pin dma
engines active for the network-receive-offload use case.  As a result
the ->free_chan_resources() that occurs after the driver self test no
longer has a NET_DMA induced ->alloc_chan_resources() to back it up.  A
late firing irq can lead to ksoftirqd spinning indefinitely due to the
tasklet_disable() performed by ->free_chan_resources().  Only
->alloc_chan_resources() can clear this condition in affected kernels.

This problem has been present since commit 3e037454 "I/OAT: Add
support for MSI and MSI-X" in 2.6.24, but is now exposed. Given the
NET_DMA use case is deprecated we can revisit moving the driver to use
threaded irqs.  For now, just tear down the irq and tasklet properly by:

1/ Disable the irq from triggering the tasklet

2/ Disable the irq from re-arming

3/ Flush inflight interrupts

4/ Flush the timer

5/ Flush inflight tasklets

References:
https://lkml.org/lkml/2014/1/27/282
https://lkml.org/lkml/2014/2/19/672

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: <stable@vger.kernel.org>
Reported-by: NMike Galbraith <bitbucket@online.de>
Reported-by: NStanislav Fomichev <stfomichev@yandex-team.ru>
Tested-by: NMike Galbraith <bitbucket@online.de>
Tested-by: NStanislav Fomichev <stfomichev@yandex-team.ru>
Reviewed-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

imx25 did not work without a firmware previously.

This patch adds a DT compatible to pass the correct data with the
default script addresses for imx25.

Add imx25 compatible to the list of compatibles in the binding
documentation.
Signed-off-by: NMarkus Pargmann <mpa@pengutronix.de>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

It appears that in the DMA40 driver the DMA tasklet will very
often dereference memory for a descriptor just free:d from the
DMA40 slab. Nothing happens because no other part of the driver
has yet had a chance to claim this memory, but it's really
nasty to dereference free:d memory, so let's check the flag
before the descriptor is free and store it in a bool variable.

Cc: stable@vger.kernel.org
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Enabling some of the mvebu platforms in the multiplatform config for ARM
enabled these drivers, which also triggered a bunch of warnings when LPAE
is enabled (thus making phys_addr_t 64-bit).

Most changes are switching printk formats, but also a bit of changes to what
used to be array-based pointer arithmetic that could just be done with the
address types instead.

The warnings were:

drivers/dma/mv_xor.c: In function 'mv_xor_tx_submit':
drivers/dma/mv_xor.c:500:3: warning: format '%x' expects argument of type
    'unsigned int', but argument 4 has type 'dma_addr_t' [-Wformat]
drivers/dma/mv_xor.c: In function 'mv_xor_alloc_chan_resources':
drivers/dma/mv_xor.c:553:13: warning: cast to pointer from integer of
    different size [-Wint-to-pointer-cast]
drivers/dma/mv_xor.c:555:4: warning: cast from pointer to integer of
    different size [-Wpointer-to-int-cast]
drivers/dma/mv_xor.c: In function 'mv_xor_prep_dma_memcpy':
drivers/dma/mv_xor.c:584:2: warning: format '%x' expects argument of type
    'unsigned int', but argument 5 has type 'dma_addr_t' [-Wformat]
drivers/dma/mv_xor.c:584:2: warning: format '%x' expects argument of type
    'unsigned int', but argument 6 has type 'dma_addr_t' [-Wformat]
drivers/dma/mv_xor.c: In function 'mv_xor_prep_dma_xor':
drivers/dma/mv_xor.c:628:2: warning: format '%u' expects argument of type
    'unsigned int', but argument 7 has type 'dma_addr_t' [-Wformat]
Acked-by: NVinod Koul <vinod.koul@intel.com>
Acked-by: NJason Cooper <jason@lakedaemon.net>
Signed-off-by: NOlof Johansson <olof@lixom.net>

Move the DMA_OF selection to the DMA driver to fix kconfig warning:

warning: (ARCH_MOXART) selects DMA_OF which has unmet direct dependencies (DMADEVICES && OF)
Signed-off-by: NRob Herring <robh@kernel.org>
Signed-off-by: NOlof Johansson <olof@lixom.net>

The merge between 2b7f65b1 "mmp_pdma: Style neatening" and
8010dad5 "dma: add dma_get_any_slave_channel(), for use in of_xlate()"
caused a build error by leaving obsolete code in place:

  mmp_pdma.c: In function 'mmp_pdma_dma_xlate':
  mmp_pdma.c:909:31: error: 'candidate' undeclared
  mmp_pdma.c:912:3: error: label 'retry' used but not defined
  mmp_pdma.c:901:24: warning: unused variable 'c' [-Wunused-variable]

This removes the extraneous lines.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Export the symbol so that it is accessible to modules.
Fixes the following error:
ERROR: "pl08x_filter_id" [sound/soc/samsung/snd-soc-s3c-dma.ko] undefined!
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

It mostly fixes the "RETURN" sections in the resulting manual page.

There is no functional change.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

vd->tx.cookie is set zero on dma_cookie_complete(),
save to local before printing it.
Signed-off-by: NJonas Jensen <jonas.jensen@gmail.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

With Device Tree a typical DMA controller device name can look like
10000000.dma-controller, which extends the current size of the string,
allocated for this parameter. This patch extends its size from 20 to 32
characters.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Set the return variable to an error code as done elsewhere in the function.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
(
if@p1 (\(ret < 0\|ret != 0\))
 { ... return ret; }
|
ret@p1 = 0
)
... when != ret = e1
    when != &ret
*if(...)
{
  ... when != ret = e2
      when forall
 return ret;
}

// </smpl>
Signed-off-by: NJulia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

The channel allocated/released messages are just informative and
not really interesting to users. Change them to "debug" level.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Tegra148 introduces a few changes to the APB DMA HW registers. Update the
driver to cope with them. Tegra124 inherits these changes.

* The register address stride between DMA channels increases.

* A new per-channel WCOUNT register is introduced.
Signed-off-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NKunal Agrawal <kunala@nvidia.com>
[swarren, remove .dts file change, rewrote commit description, removed
some duplicate/unused code and register IO]
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Reviewed-by: NThierry Reding <treding@nvidia.com>
Tested-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Since we have nice helper to print dma_addr_t values by reference we may use it
instead of explicit casting to a longest type.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

The joined messages are better to grep when debugging.

There is no functional change.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Simple fix a style of the multiline comment.

There is no functional change.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Fix sparse warnings:
drivers/dma/k3dma.c:480:20: warning: Using plain integer as NULL pointer
drivers/dma/k3dma.c:820:1: warning: symbol 'k3_dma_pmops' was not declared. Should it be static?
Signed-off-by: NZhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Currently the driver uses dma_request_channel() with a custom filter function to
find the requested channel. This will loop over all available channels until the
one we want has been found, but we already know which channel we want to
request, so we can dma_get_slave_channel(). This also makes the code a bit
shorter cleaner.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

The pl330 dmaengine driver currently does not differentiate between submitted
and issued descriptors. It won't start transferring a newly submitted
descriptor until issue_pending() is called, but only if it is idle. If it is
active and a new descriptor is submitted before it goes idle it will happily
start the newly submitted descriptor once all earlier submitted descriptors have
been completed. This is not a 100% correct with regards to the dmaengine
interface semantics. A descriptor is not supposed to be started until the next
issue_pending() call after the descriptor has been submitted. This patch adds a
second per channel list that keeps track of the submitted descriptors. Once
issue_pending() is called the submitted descriptors are moved to the working
list and only descriptors on the working list are started.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

this patch adds device_slave_caps() callback as SiRF SoC sound drivers
depend on it.
Signed-off-by: NRongjun Ying <rongjun.ying@csr.com>
Signed-off-by: NBarry Song <Barry.Song@csr.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Freescale DMA has a feature of BandWidth Control (ab. BWC), which is currently
256 bytes and should be changed to 1024 bytes for best DMA throughput.
Changing BWC from 256 to 1024 will improve DMA performance much, in cases
whatever one channel is running or multi channels are running simultanously,
large or small buffers are copied.  And this change doesn't impact memory
access performance remarkably, lmbench tests show that for some cases the
memory performance are decreased very slightly, while the others are even
better.
Tested on T4240.
Signed-off-by: NHongbo Zhang <hongbo.zhang@freescale.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

The MOXA ART SoC has a DMA controller capable of offloading expensive
memory operations, such as large copies. This patch adds support for
the controller including four channels. Two of these are used to
handle MMC copy on the UC-7112-LX hardware. The remaining two can be
used in a future audio driver or client application.
Signed-off-by: NJonas Jensen <jonas.jensen@gmail.com>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Without DMA_PRIVATE the driver is not able to allocate more than one channel.
Since it uses dma_get_any_slave_channel that calls private_candidate,
the second allocation fails at
/* some channels are already publicly allocated */
Maybe it should be fixed in the core, but at least this fixes the bug.
Signed-off-by: NFlorian Meier <florian.meier@koalo.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

The pl330 driver currently does not support residue reporting, so set the
residue granularity to DMA_RESIDUE_GRANULARITY_DESCRIPTOR.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Acked-by: NVinod Koul <vinod.koul@intel.com>
Signed-off-by: NMark Brown <broonie@linaro.org>

i.MX series have inner firmware in its ROM code: when SDMA isn't provided
any firmware from Kernel or rootfs, the default inner ROM firmware will be
activated. However the current driver doesn't assign any script number to
this situation, and those platform running in this case would be broken.

Thus this patch adds a default script number when no external firmware being
loaded so that people would continue to be able to use basic scripts to run
their platform without any firmware.
Reported-by: NFabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: NNicolin Chen <Guangyu.Chen@freescale.com>
Tested-by: NFabio Estevam <fabio.estevam@freescale.com>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

    Support to get sram pool from generic device tree binding. The
existing way of get sram poll, directly call sram_get_gpool(), still
work here.
Signed-off-by: NNenghua Cao <nhcao@marvell.com>
Acked-by: NZhangfei Gao <zhangfei.gao@gmail.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Add support for DMA controller of BCM2835 as used in the Raspberry Pi.
Currently it only supports cyclic DMA.
Signed-off-by: NFlorian Meier <florian.meier@koalo.de>
Reviewed-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Check DMA mapping return values in function ioat_dma_self_test() to get
rid of following warning message.

  ------------[ cut here ]------------
  WARNING: CPU: 0 PID: 1203 at lib/dma-debug.c:937 check_unmap+0x4c0/0x9a0()
  ioatdma 0000:00:04.0: DMA-API: device driver failed to check map error[device address=0x000000085191b000] [size=2000 bytes] [mapped as single]
  Modules linked in: ioatdma(+) mac_hid wmi acpi_pad lp parport hidd_generic usbhid hid ixgbe isci dca libsas ahci ptp libahci scsi_transport_sas meegaraid_sas pps_core mdio
  CPU: 0 PID: 1203 Comm: systemd-udevd Not tainted 3.13.0-rc4+ #8
  Hardware name: Intel Corporation BRICKLAND/BRICKLAND, BIOS BRIVTIIN1.86B.0044.L09.1311181644 11/18/2013
  Call Trace:
    dump_stack+0x4d/0x66
    warn_slowpath_common+0x7d/0xa0
    warn_slowpath_fmt+0x4c/0x50
    check_unmap+0x4c0/0x9a0
    debug_dma_unmap_page+0x81/0x90
    ioat_dma_self_test+0x3d2/0x680 [ioatdma]
    ioat3_dma_self_test+0x12/0x30 [ioatdma]
    ioat_probe+0xf4/0x110 [ioatdma]
    ioat3_dma_probe+0x268/0x410 [ioatdma]
    ioat_pci_probe+0x122/0x1b0 [ioatdma]
    local_pci_probe+0x45/0xa0
    pci_device_probe+0xd9/0x130
    driver_probe_device+0x171/0x490
    __driver_attach+0x93/0xa0
    bus_for_each_dev+0x6b/0xb0
    driver_attach+0x1e/0x20
    bus_add_driver+0x1f8/0x2b0
    driver_register+0x81/0x110
    __pci_register_driver+0x60/0x70
    ioat_init_module+0x89/0x1000 [ioatdma]
    do_one_initcall+0xe2/0x250
    load_module+0x2313/0x2a00
    SyS_init_module+0xd9/0x130
    system_call_fastpath+0x1a/0x1f
  ---[ end trace 990c591681d27c31 ]---
  Mapped at:
    debug_dma_map_page+0xbe/0x180
    ioat_dma_self_test+0x1ab/0x680 [ioatdma]
    ioat3_dma_self_test+0x12/0x30 [ioatdma]
    ioat_probe+0xf4/0x110 [ioatdma]
    ioat3_dma_probe+0x268/0x410 [ioatdma]
Signed-off-by: NJiang Liu <jiang.liu@linux.intel.com>
Cc: Vinod Koul <vinod.koul@intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

net_dma can cause data to be copied to a stale mapping if a
copy-on-write fault occurs during dma.  The application sees missing
data.

The following trace is triggered by modifying the kernel to WARN if it
ever triggers copy-on-write on a page that is undergoing dma:

 WARNING: CPU: 24 PID: 2529 at lib/dma-debug.c:485 debug_dma_assert_idle+0xd2/0x120()
 ioatdma 0000:00:04.0: DMA-API: cpu touching an active dma mapped page [pfn=0x16bcd9]
 Modules linked in: iTCO_wdt iTCO_vendor_support ioatdma lpc_ich pcspkr dca
 CPU: 24 PID: 2529 Comm: linbug Tainted: G        W    3.13.0-rc1+ #353
  00000000000001e5 ffff88016f45f688 ffffffff81751041 ffff88017ab0ef70
  ffff88016f45f6d8 ffff88016f45f6c8 ffffffff8104ed9c ffffffff810f3646
  ffff8801768f4840 0000000000000282 ffff88016f6cca10 00007fa2bb699349
 Call Trace:
  [<ffffffff81751041>] dump_stack+0x46/0x58
  [<ffffffff8104ed9c>] warn_slowpath_common+0x8c/0xc0
  [<ffffffff810f3646>] ? ftrace_pid_func+0x26/0x30
  [<ffffffff8104ee86>] warn_slowpath_fmt+0x46/0x50
  [<ffffffff8139c062>] debug_dma_assert_idle+0xd2/0x120
  [<ffffffff81154a40>] do_wp_page+0xd0/0x790
  [<ffffffff811582ac>] handle_mm_fault+0x51c/0xde0
  [<ffffffff813830b9>] ? copy_user_enhanced_fast_string+0x9/0x20
  [<ffffffff8175fc2c>] __do_page_fault+0x19c/0x530
  [<ffffffff8175c196>] ? _raw_spin_lock_bh+0x16/0x40
  [<ffffffff810f3539>] ? trace_clock_local+0x9/0x10
  [<ffffffff810fa1f4>] ? rb_reserve_next_event+0x64/0x310
  [<ffffffffa0014c00>] ? ioat2_dma_prep_memcpy_lock+0x60/0x130 [ioatdma]
  [<ffffffff8175ffce>] do_page_fault+0xe/0x10
  [<ffffffff8175c862>] page_fault+0x22/0x30
  [<ffffffff81643991>] ? __kfree_skb+0x51/0xd0
  [<ffffffff813830b9>] ? copy_user_enhanced_fast_string+0x9/0x20
  [<ffffffff81388ea2>] ? memcpy_toiovec+0x52/0xa0
  [<ffffffff8164770f>] skb_copy_datagram_iovec+0x5f/0x2a0
  [<ffffffff8169d0f4>] tcp_rcv_established+0x674/0x7f0
  [<ffffffff816a68c5>] tcp_v4_do_rcv+0x2e5/0x4a0
  [..]
 ---[ end trace e30e3b01191b7617 ]---
 Mapped at:
  [<ffffffff8139c169>] debug_dma_map_page+0xb9/0x160
  [<ffffffff8142bf47>] dma_async_memcpy_pg_to_pg+0x127/0x210
  [<ffffffff8142cce9>] dma_memcpy_pg_to_iovec+0x119/0x1f0
  [<ffffffff81669d3c>] dma_skb_copy_datagram_iovec+0x11c/0x2b0
  [<ffffffff8169d1ca>] tcp_rcv_established+0x74a/0x7f0:

...the problem is that the receive path falls back to cpu-copy in
several locations and this trace is just one of the areas.  A few
options were considered to fix this:

1/ sync all dma whenever a cpu copy branch is taken

2/ modify the page fault handler to hold off while dma is in-flight

Option 1 adds yet more cpu overhead to an "offload" that struggles to compete
with cpu-copy.  Option 2 adds checks for behavior that is already documented as
broken when using get_user_pages().  At a minimum a debug mode is warranted to
catch and flag these violations of the dma-api vs get_user_pages().

Thanks to David for his reproducer.

Cc: <stable@vger.kernel.org>
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: Vinod Koul <vinod.koul@intel.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Reported-by: NDavid Whipple <whipple@securedatainnovations.ch>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

I see the following splat with 3.13-rc1 when attempting to perform DMA:

[  253.004516] Alignment trap: not handling instruction e1902f9f at [<c0204b40>]
[  253.004583] Unhandled fault: alignment exception (0x221) at 0xdfdfdfd7
[  253.004646] Internal error: : 221 [#1] PREEMPT SMP ARM
[  253.004691] Modules linked in: dmatest(+) [last unloaded: dmatest]
[  253.004798] CPU: 0 PID: 671 Comm: kthreadd Not tainted 3.13.0-rc1+ #2
[  253.004864] task: df9b0900 ti: df03e000 task.ti: df03e000
[  253.004937] PC is at dmaengine_unmap_put+0x14/0x34
[  253.005010] LR is at pl330_tasklet+0x3c8/0x550
[  253.005087] pc : [<c0204b44>]    lr : [<c0207478>]    psr: a00e0193
[  253.005087] sp : df03fe48  ip : 00000000  fp : df03bf18
[  253.005178] r10: bf00e108  r9 : 00000001  r8 : 00000000
[  253.005245] r7 : df837040  r6 : dfb41800  r5 : df837048  r4 : df837000
[  253.005316] r3 : dfdfdfcf  r2 : dfb41f80  r1 : df837048  r0 : dfdfdfd7
[  253.005384] Flags: NzCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
[  253.005459] Control: 30c5387d  Table: 9fb9ba80  DAC: fffffffd
[  253.005520] Process kthreadd (pid: 671, stack limit = 0xdf03e248)

This is due to desc->txd.unmap containing garbage (uninitialised memory).

Rather than add another dummy initialisation to _init_desc, instead
ensure that the descriptors are zero-initialised during allocation and
remove the dummy, per-field initialisation.

Cc: Andriy Shevchenko <andriy.shevchenko@intel.com>
Acked-by: NJassi Brar <jassisinghbrar@gmail.com>
Signed-off-by: NWill Deacon <will.deacon@arm.com>
Acked-by: NVinod Koul <vinod.koul@intel.com>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>

It might happen that a platform wants to use its DMA engine for
memcpy only, and then we have zero slave channels to initialize,
so allow the slave initialization to return zero.
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

The channel allocated/released messages are very spammy and not really
interesting to users. Change them to "debug" level.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Acked-by: NMatt Porter <mporter@linaro.org>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>