linux.git/kernel/dma/remap.c, branch v6.12.80

dma-mapping: use bit masking to check VM_DMA_COHERENT

2024-08-22T04:15:35+00:00

In dma_common_find_pages(), area->flags are compared directly with
VM_DMA_COHERENT. This works because VM_DMA_COHERENT is the only set
flag.

During development of a new feature (ASI [1]), a new VM flag is
introduced, and that flag can be injected into VM_DMA_COHERENT mappings
(among others).  The presence of that flag caused
dma_common_find_pages() to return NULL for VM_DMA_COHERENT addresses,
leading to a lot of problems ending in crashing during boot. It took a
bit of time to figure this problem out.

It was a mistake to inject a VM flag to begin with, but it took a
significant amount of debugging to figure out the problem. Most users of
area->flags use bitmasking rather than equivalency to check for flags.
Update dma_common_find_pages() and dma_common_free_remap() to do the
same, which would have avoided the boot crashing. Instead, add a warning
in dma_common_find_pages() if any extra VM flags are set to catch such
problems more easily during development.

No functional change intended.

[1]https://lore.kernel.org/lkml/20240712-asi-rfc-24-v1-0-144b319a40d8@google.com/

Signed-off-by: Yosry Ahmed 
Signed-off-by: Christoph Hellwig

dma-remap: use kvmalloc_array/kvfree for larger dma memory remap

2023-06-07T13:06:28+00:00

If dma_direct_alloc() alloc memory in size of 64MB, the inner function
dma_common_contiguous_remap() will allocate 128KB memory by invoking
the function kmalloc_array(). and the kmalloc_array seems to fail to try to
allocate 128KB mem.

Call trace:
[14977.928623] qcrosvm: page allocation failure: order:5, mode:0x40cc0
[14977.928638] dump_backtrace.cfi_jt+0x0/0x8
[14977.928647] dump_stack_lvl+0x80/0xb8
[14977.928652] warn_alloc+0x164/0x200
[14977.928657] __alloc_pages_slowpath+0x9f0/0xb4c
[14977.928660] __alloc_pages+0x21c/0x39c
[14977.928662] kmalloc_order+0x48/0x108
[14977.928666] kmalloc_order_trace+0x34/0x154
[14977.928668] __kmalloc+0x548/0x7e4
[14977.928673] dma_direct_alloc+0x11c/0x4f8
[14977.928678] dma_alloc_attrs+0xf4/0x138
[14977.928680] gh_vm_ioctl_set_fw_name+0x3c4/0x610 [gunyah]
[14977.928698] gh_vm_ioctl+0x90/0x14c [gunyah]
[14977.928705] __arm64_sys_ioctl+0x184/0x210

work around by doing kvmalloc_array instead.

Signed-off-by: Gao Xu 
Reviewed-by: Suren Baghdasaryan 
Signed-off-by: Christoph Hellwig

kernel/dma: remove unnecessary unmap_kernel_range

2021-04-30T18:20:40+00:00

vunmap will remove ptes.

Link: https://lkml.kernel.org/r/20210322021806.892164-3-npiggin@gmail.com
Signed-off-by: Nicholas Piggin 
Reviewed-by: Christoph Hellwig 
Cc: Cédric Le Goater 
Cc: Uladzislau Rezki 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

dma-mapping: move more functions to dma-map-ops.h

2020-10-20T08:41:07+00:00

Due to a mismerge a bunch of prototypes that should have moved to
dma-map-ops.h are still in dma-mapping.h, fix that up.

Signed-off-by: Christoph Hellwig

dma-remap: align the size in dma_common_*_remap()

2020-06-23T12:14:41+00:00

Running a guest with a virtio-iommu protecting virtio devices
is broken since commit 515e5b6d90d4 ("dma-mapping: use vmap insted
of reimplementing it"). Before the conversion, the size was
page aligned in __get_vm_area_node(). Doing so fixes the
regression.

Fixes: 515e5b6d90d4 ("dma-mapping: use vmap insted of reimplementing it")
Signed-off-by: Eric Auger 
Signed-off-by: Christoph Hellwig

Merge tag 'dma-mapping-5.8' of git://git.infradead.org/users/hch/dma-mapping

2020-06-06T18:43:23+00:00

Pull dma-mapping updates from Christoph Hellwig:

 - enhance the dma pool to allow atomic allocation on x86 with AMD SEV
   (David Rientjes)

 - two small cleanups (Jason Yan and Peter Collingbourne)

* tag 'dma-mapping-5.8' of git://git.infradead.org/users/hch/dma-mapping:
  dma-contiguous: fix comment for dma_release_from_contiguous
  dma-pool: scale the default DMA coherent pool size with memory capacity
  x86/mm: unencrypted non-blocking DMA allocations use coherent pools
  dma-pool: add pool sizes to debugfs
  dma-direct: atomic allocations must come from atomic coherent pools
  dma-pool: dynamically expanding atomic pools
  dma-pool: add additional coherent pools to map to gfp mask
  dma-remap: separate DMA atomic pools from direct remap code
  dma-debug: make __dma_entry_alloc_check_leak() static

dma-mapping: use vmap insted of reimplementing it

2020-06-02T17:59:10+00:00

Replace the open coded instance of vmap with the actual function.  In
the non-contiguous (IOMMU) case this requires an extra find_vm_area,
but given that this isn't a fast path function that is a small price
to pay.

Signed-off-by: Christoph Hellwig 
Signed-off-by: Andrew Morton 
Acked-by: Peter Zijlstra (Intel) 
Cc: Christian Borntraeger 
Cc: Christophe Leroy 
Cc: Daniel Vetter 
Cc: David Airlie 
Cc: Gao Xiang 
Cc: Greg Kroah-Hartman 
Cc: Haiyang Zhang 
Cc: Johannes Weiner 
Cc: "K. Y. Srinivasan" 
Cc: Laura Abbott 
Cc: Mark Rutland 
Cc: Michael Kelley 
Cc: Minchan Kim 
Cc: Nitin Gupta 
Cc: Robin Murphy 
Cc: Sakari Ailus 
Cc: Stephen Hemminger 
Cc: Sumit Semwal 
Cc: Wei Liu 
Cc: Benjamin Herrenschmidt 
Cc: Catalin Marinas 
Cc: Heiko Carstens 
Cc: Paul Mackerras 
Cc: Vasily Gorbik 
Cc: Will Deacon 
Link: http://lkml.kernel.org/r/20200414131348.444715-6-hch@lst.de
Signed-off-by: Linus Torvalds

dma-remap: separate DMA atomic pools from direct remap code

2020-04-20T10:08:45+00:00

DMA atomic pools will be needed beyond only CONFIG_DMA_DIRECT_REMAP so
separate them out into their own file.

This also adds a new Kconfig option that can be subsequently used for
options, such as CONFIG_AMD_MEM_ENCRYPT, that will utilize the coherent
pools but do not have a dependency on direct remapping.

For this patch alone, there is no functional change introduced.

Reviewed-by: Christoph Hellwig 
Signed-off-by: David Rientjes 
[hch: fixup copyrights and remove unused includes]
Signed-off-by: Christoph Hellwig

lib/genalloc.c: rename addr_in_gen_pool to gen_pool_has_addr

2019-12-05T03:44:13+00:00

Follow the kernel conventions, rename addr_in_gen_pool to
gen_pool_has_addr.

[sjhuang@iluvatar.ai: fix Documentation/ too]
 Link: http://lkml.kernel.org/r/20181229015914.5573-1-sjhuang@iluvatar.ai
Link: http://lkml.kernel.org/r/20181228083950.20398-1-sjhuang@iluvatar.ai
Signed-off-by: Huang Shijie 
Reviewed-by: Andrew Morton 
Cc: Russell King 
Cc: Arnd Bergmann 
Cc: Greg Kroah-Hartman 
Cc: Christoph Hellwig 
Cc: Marek Szyprowski 
Cc: Robin Murphy 
Cc: Stephen Rothwell 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

dma-mapping: merge the generic remapping helpers into dma-direct

2019-11-11T09:52:18+00:00

Integrate the generic dma remapping implementation into the main flow.
This prepares for architectures like xtensa that use an uncached
segment for pages in the kernel mapping, but can also remap highmem
from CMA.  To simplify that implementation we now always deduct the
page from the physical address via the DMA address instead of the
virtual address.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Max Filippov