Using a mask to represent bus DMA constraints has a set of limitations.
The biggest one being it can only hold a power of two (minus one). The
DMA mapping code is already aware of this and treats dev->bus_dma_mask
as a limit. This quirk is already used by some architectures although
still rare.

With the introduction of the Raspberry Pi 4 we've found a new contender
for the use of bus DMA limits, as its PCIe bus can only address the
lower 3GB of memory (of a total of 4GB). This is impossible to represent
with a mask. To make things worse the device-tree code rounds non power
of two bus DMA limits to the next power of two, which is unacceptable in
this case.

In the light of this, rename dev->bus_dma_mask to dev->bus_dma_limit all
over the tree and treat it as such. Note that dev->bus_dma_limit should
contain the higher accessible DMA address.
Signed-off-by: NNicolas Saenz Julienne <nsaenzjulienne@suse.de>
Reviewed-by: NRobin Murphy <robin.murphy@arm.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

As we've seen from USB and other areas[1], we need to always do runtime
checks for DMA operating on memory regions that might be remapped. This
adds vmap checks (similar to those already in USB but missing in other
places) into dma_map_single() so all callers benefit from the checking.

[1] https://git.kernel.org/linus/3840c5b78803b2b6cc1ff820100a74a092c40cbbSuggested-by: NLaura Abbott <labbott@redhat.com>
Signed-off-by: NKees Cook <keescook@chromium.org>
[hch: fixed the printk message]
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Daniele reported that issue previously fixed in c41f9ea9
("drivers: dma-coherent: Account dma_pfn_offset when used with device
tree") reappear shortly after 43fc509c ("dma-coherent: introduce
interface for default DMA pool") where fix was accidentally dropped.

Lets put fix back in place and respect dma-ranges for reserved memory.

Fixes: 43fc509c ("dma-coherent: introduce interface for default DMA pool")
Reported-by: NDaniele Alessandrelli <daniele.alessandrelli@gmail.com>
Tested-by: NDaniele Alessandrelli <daniele.alessandrelli@gmail.com>
Tested-by: NAlexandre Torgue <alexandre.torgue@st.com>
Signed-off-by: NVladimir Murzin <vladimir.murzin@arm.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

A helper to find the backing page array based on a virtual address.
This also ensures we do the same vm_flags check everywhere instead
of slightly different or missing ones in a few places.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Currently the generic dma remap allocator gets a vm_flags passed by
the caller that is a little confusing.  We just introduced a generic
vmalloc-level flag to identify the dma coherent allocations, so use
that everywhere and remove the now pointless argument.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This function is entirely unused given that declared memory is
generally provided by platform setup code.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

We can already use DMA_ATTR_WRITE_COMBINE or the _wc prefixed version,
so remove the third way of doing things.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add a helper to check if DMA allocations for a specific device can be
mapped to userspace using dma_mmap_*.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This patch adds a new DMA API "dma_get_merge_boundary". This function
returns the DMA merge boundary if the DMA layer can merge the segments.
This patch also adds the implementation for a new dma_map_ops pointer.
Signed-off-by: NYoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: NSimon Horman <horms+renesas@verge.net.au>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

The memory allocated for the atomic pool needs to have the same
mapping attributes that we use for remapping, so use
pgprot_dmacoherent instead of open coding it.  Also deduct a
suitable zone to allocate the memory from based on the presence
of the DMA zones.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

No users left.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20190816062435.881-6-hch@lst.deSigned-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

We currently have cases where the dma_addressing_limited() gets
called with dma_mask unset. This causes a NULL pointer dereference.

Use dma_get_mask() accessor to prevent the crash.

Fixes: b8664554 ("dma-mapping: add a dma_addressing_limited helper")
Signed-off-by: NEric Auger <eric.auger@redhat.com>
Acked-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This helper returns if the device has issues addressing all present
memory in the system.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

These days, the DMA mapping code must bounce buffers for any unsupported
address. If the driver needs to optimize for natively supported ranges,
then it should use dma_get_required_mask.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NMarc Gonzalez <marc.w.gonzalez@free.fr>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Most dma_map_ops implementations already had some issues with a NULL
device, or did simply crash if one was fed to them.  Now that we have
cleaned up all the obvious offenders we can stop to pretend we
support this mode.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

The function returns the maximum size that can be mapped
using DMA-API functions. The patch also adds the
implementation for direct DMA and a new dma_map_ops pointer
so that other implementations can expose their limit.

Cc: stable@vger.kernel.org
Reviewed-by: NKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJoerg Roedel <jroedel@suse.de>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

All users of dma_declare_coherent want their allocations to be
exclusive, so default to exclusive allocations.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This API is not used anywhere, so remove it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This API is primarily used through DT entries, but two architectures
and two drivers call it directly.  So instead of selecting the config
symbol for random architectures pull it in implicitly for the actual
users.  Also rename the Kconfig option to describe the feature better.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: Paul Burton <paul.burton@mips.com> # MIPS
Acked-by: NLee Jones <lee.jones@linaro.org>
Reviewed-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: Catalin Marinas <catalin.marinas@arm.com> # arm64

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: Paul Burton <paul.burton@mips.com> # MIPS
Acked-by: Catalin Marinas <catalin.marinas@arm.com> # arm64

Use WARN_ON_ONCE to print a stack trace and return a proper error
code instead.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NRobin Murphy <robin.murphy@arm.com>
Tested-by: NMarek Szyprowski <m.szyprowski@samsung.com>

Instead provide a proper implementation in the direct mapping code, and
also wire it up for arm and powerpc, leaving an error return for all the
IOMMU or virtual mapping instances for which we'd have to wire up an
actual implementation
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Tested-by: NMarek Szyprowski <m.szyprowski@samsung.com>

dma_zalloc_coherent() is no longer needed as it has no users because
dma_alloc_coherent() already zeroes out memory for us.

The Coccinelle grammar rule that used to check for dma_alloc_coherent()
+ memset() is modified so that it just tells the user that the memset is
not needed anymore.
Suggested-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NLuis Chamberlain <mcgrof@kernel.org>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This avoids link failures in drivers using the DMA API, when they
are compiled for user mode Linux with CONFIG_COMPILE_TEST=y.

Fixes: 356da6d0 ("dma-mapping: bypass indirect calls for dma-direct")
Signed-off-by: NChristoph Hellwig <hch@lst.de>

These functions have never been used.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

dmam_alloc_coherent is just the default no-flags case of
dmam_alloc_attrs, so take advantage of this similar to the non-managed
version.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

And also switch the way we implement the unmap side around to stay
consistent.  This ensures dma-debug works again because it records which
function we used for mapping to ensure it is also used for unmapping,
and also reduces further code duplication.  Last but not least this
also officially allows calling dma_sync_single_* for mappings created
using dma_map_page, which is perfectly fine given that the sync calls
only take a dma_addr_t, but not a virtual address or struct page.

Fixes: 7f0fee24 ("dma-mapping: merge dma_unmap_page_attrs and dma_unmap_single_attrs")
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Tested-by: NLABBE Corentin <clabbe.montjoie@gmail.com>

We really need the writecombine flag in dma_alloc_wc, fix a stupid
oversight.

Fixes: 7ed1d91a ("dma-mapping: translate __GFP_NOFAIL to DMA_ATTR_NO_WARN")
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

We now always return zeroed memory from dma_alloc_coherent.  Note that
simply passing GFP_ZERO to dma_alloc_coherent wasn't always doing the
right thing to start with given that various allocators are not backed
by the page allocator and thus would ignore GFP_ZERO.
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Avoid expensive indirect calls in the fast path DMA mapping
operations by directly calling the dma_direct_* ops if we are using
the directly mapped DMA operations.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NTony Luck <tony.luck@intel.com>

The dummy DMA ops are currently used by arm64 for any device which has
an invalid ACPI description and is thus barred from using DMA due to not
knowing whether is is cache-coherent or not. Factor these out into
general dma-mapping code so that they can be referenced from other
common code paths. In the process, we can prune all the optional
callbacks which just do the same thing as the default behaviour, and
fill in .map_resource for completeness.
Signed-off-by: NRobin Murphy <robin.murphy@arm.com>
[hch: moved to a separate source file]
Reviewed-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NTony Luck <tony.luck@intel.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

This isn't exactly a slow path routine, but it is not super critical
either, and moving it out of line will help to keep the include chain
clean for the following DMA indirection bypass work.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NTony Luck <tony.luck@intel.com>

There is no need to have all setup and coherent allocation / freeing
routines inline.  Move them out of line to keep the implemeation
nicely encapsulated and save some kernel text size.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NTony Luck <tony.luck@intel.com>

The two functions are exactly the same, so don't bother implementing
them twice.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NTony Luck <tony.luck@intel.com>

We can just call the regular calls after adding offset the the address instead
of reimplementing them.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NJesper Dangaard Brouer <brouer@redhat.com>
Tested-by: NTony Luck <tony.luck@intel.com>

Currently dma_mapping_error returns a boolean as int, with 1 meaning
error.  This is rather unusual and many callers have to convert it to
errno value.  The callers are highly inconsistent with error codes
ranging from -ENOMEM over -EIO, -EINVAL and -EFAULT ranging to -EAGAIN.
Return -ENOMEM which seems to be what the largest number of callers
convert it to, and which also matches the typical error case where
we are out of resources.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NRussell King <rmk+kernel@armlinux.org.uk>
Acked-by: NLinus Torvalds <torvalds@linux-foundation.org>

No users left except for vmd which just forwards it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Acked-by: NLinus Torvalds <torvalds@linux-foundation.org>

Error handling of the dma_map_single and dma_map_page APIs is a little
problematic at the moment, in that we use different encodings in the
returned dma_addr_t to indicate an error.  That means we require an
additional indirect call to figure out if a dma mapping call returned
an error, and a lot of boilerplate code to implement these semantics.

Instead return the maximum addressable value as the error.  As long
as we don't allow mapping single-byte ranges with single-byte alignment
this value can never be a valid return.  Additionaly if drivers do
not check the return value from the dma_map* routines this values means
they will generally not be pointed to actual memory.

Once the default value is added here we can start removing the
various mapping_error methods and just rely on this generic check.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NRobin Murphy <robin.murphy@arm.com>
Acked-by: NRussell King <rmk+kernel@armlinux.org.uk>
Acked-by: NLinus Torvalds <torvalds@linux-foundation.org>

The arm64 codebase to implement coherent dma allocation for architectures
with non-coherent DMA is a good start for a generic implementation, given
that is uses the generic remap helpers, provides the atomic pool for
allocations that can't sleep and still is realtively simple and well
tested.  Move it to kernel/dma and allow architectures to opt into it
using a config symbol.  Architectures just need to provide a new
arch_dma_prep_coherent helper to writeback an invalidate the caches
for any memory that gets remapped for uncached access.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NWill Deacon <will.deacon@arm.com>
Reviewed-by: NRobin Murphy <robin.murphy@arm.com>