Implement new execbuf3 ioctl (I915_GEM_EXECBUFFER3) which only
works in vm_bind mode. The vm_bind mode only works with
this new execbuf3 ioctl.
The new execbuf3 ioctl will not have any list of objects to validate
bind as all required objects binding would have been requested by the
userspace befor
The new execbuf3 ioctl path and the legacy execbuf ioctl
paths have many common functionalities.
Abstract out the common execbuf functionalities into a
separate file where possible, thus allowing code sharing.
Reviewed-by: Andi Shyti
Reviewed-by: Matthew Auld
Signed-off-by: Niranjana Vishwanatha
Support eviction by maintaining a list of evicted persistent vmas
for rebinding during next submission. Ensure the list do not
include persistent vmas that are being purged.
v2: Remove unused I915_VMA_PURGED definition.
v3: Properly handle __i915_vma_unbind_async() case.
Reviewed-by: Matthew Auld
Add function __i915_sw_fence_await_reservation() for
asynchronous wait on a dma-resv object with specified
dma_resv_usage. This is required for async vma unbind
with vm_bind.
Reviewed-by: Matthew Auld
Signed-off-by: Niranjana Vishwanathapura
---
drivers/gpu/drm/i915/i915_sw_fence.c | 28 +++
Ensure i915_vma_verify_bind_complete() handles case where bind
is not initiated. Also make it non static, add documentation
and move it out of CONFIG_DRM_I915_DEBUG_GEM.
v2: Fix fence leak
Reviewed-by: Matthew Auld
Signed-off-by: Niranjana Vishwanathapura
Signed-off-by: Andi Shyti
---
drivers
DRM_I915_GEM_VM_BIND/UNBIND ioctls allows UMD to bind/unbind GEM
buffer objects (BOs) or sections of a BOs at specified GPU virtual
addresses on a specified address space (VM). Multiple mappings can map
to the same physical pages of an object (aliasing). These mappings (also
referred to as persiste
Add i915_vma_instance_persistent() to create persistent vmas.
Persistent vmas will use i915_gtt_view to support partial binding.
vma_lookup is tied to segment of the object instead of section
of VA space. Hence, it do not support aliasing. ie., multiple
mappings (at different VA) point to the same
Expose i915_gem_object_max_page_size() function non-static
which will be used by the vm_bind feature.
Reviewed-by: Matthew Auld
Signed-off-by: Niranjana Vishwanathapura
Signed-off-by: Andi Shyti
---
drivers/gpu/drm/i915/gem/i915_gem_create.c | 18 +-
drivers/gpu/drm/i915/gem/i9
Only support vm_bind mode with non-recoverable contexts.
With new vm_bind mode with eb3 submission path, we need not
support older recoverable contexts.
Signed-off-by: Niranjana Vishwanathapura
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 6 ++
1 file changed, 6 insertions(+)
diff --gi
Add getparam support for VM_BIND capability version.
Add VM creation time flag to enable vm_bind_mode for the VM.
v2: update kernel-doc
v3: create vm->root_obj only upon I915_VM_CREATE_FLAGS_USE_VM_BIND
v4: replace vm->vm_bind_mode check with i915_gem_vm_is_vm_bind_mode()
Reviewed-by: Matthew Aul
Each VM creates a root_obj and shares it with all of its private objects
to use it as dma_resv object. This has a performance advantage as it
requires a single dma_resv object update for all private BOs vs list of
dma_resv objects update for shared BOs, in the execbuf path.
VM private BOs can be o
Add support for handling out fence for vm_bind call.
v2: Reset vma->vm_bind_fence.syncobj to NULL at the end
of vm_bind call.
v3: Remove vm_unbind out fence uapi which is not supported yet.
v4: Return error if I915_TIMELINE_FENCE_WAIT fence flag is set.
Wait for bind to complete iff I915_T
For persistent (vm_bind) vmas of userptr BOs, handle the user
page pinning by using the i915_gem_object_userptr_submit_init()
/done() functions
v2: Do not double add vma to vm->userptr_invalidated_list
Signed-off-by: Niranjana Vishwanathapura
Signed-off-by: Andi Shyti
---
.../gpu/drm/i915/gem/
Update the execbuf path to use common execbuf functions to
reduce code duplication with the newer execbuf3 path.
Reviewed-by: Matthew Auld
Signed-off-by: Niranjana Vishwanathapura
---
.../gpu/drm/i915/gem/i915_gem_execbuffer.c| 507 ++
1 file changed, 38 insertions(+), 469 d
On 2022/10/25 3:11, Jani Nikula wrote:
On Tue, 25 Oct 2022, Akihiko Odaki wrote:
That aside, the first patch in this series can be applied without the
later patches so you may have a look at it. It's fine if you don't merge
it though since it does not fix really a pragmatic bug as its message s
Make i915_gem_vm_lookup() function non-static as it will be
used by the vm_bind feature.
Reviewed-by: Matthew Auld
Signed-off-by: Niranjana Vishwanathapura
Signed-off-by: Andi Shyti
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 11 ++-
drivers/gpu/drm/i915/gem/i915_gem_context.h |
Rename __i915_request_await_bind() as i915_request_await_bind()
and make it non-static as it will be used in execbuf3 ioctl path.
v2: add documentation
Reviewed-by: Matthew Auld
Reviewed-by: Andi Shyti
Signed-off-by: Niranjana Vishwanathapura
---
drivers/gpu/drm/i915/i915_vma.c | 8 +---
Do not use i915_vma activeness tracking for persistent vmas.
As persistent vmas are part of working set for each execbuf
submission on that address space (VM), a persistent vma is
active if the VM active. As vm->root_obj->base.resv will be
updated for each submission on that VM, it correctly
repre
Handle persistent (VM_BIND) mappings during the request submission
in the execbuf3 path.
v2: Ensure requests wait for bindings to complete.
v3: Remove short term pinning with PIN_VALIDATE flag.
Individualize fences before adding to dma_resv obj.
v4: Fix bind completion check, use PIN_NOEVICT,
Hi dri-devel,
I have two machines with ASPEED GPU's (ast Linux driver). One machine
is x86_64, running an ASRock Rack Tommy 90-SC02P1-00UBNZ GPU (AST2510
chipset) with KDE Plasma Wayland; the other is ppc64le, running an
integrated AST2500 GPU with KDE Plasma X11. Both the AST2510 and
AST25
Hi
Am 24.10.22 um 18:48 schrieb Alex Deucher:
On Mon, Oct 24, 2022 at 3:33 AM Thomas Zimmermann wrote:
Hi
Am 24.10.22 um 08:20 schrieb Quan, Evan:
[AMD Official Use Only - General]
Reviewed-by: Evan Quan
-Original Message-
From: amd-gfx On Behalf Of Alex
Deucher
Sent: Thursday,
./drivers/gpu/drm/amd/amdkfd/kfd_migrate.c:985:58-62: ERROR: p is NULL but
dereferenced.
Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=2549
Reported-by: Abaci Robot
Signed-off-by: Yang Li
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-
On Fri, 9 Sept 2022 at 19:18, Daniel Stone wrote:
>
> Hi,
>
> On Fri, 9 Sept 2022 at 15:15, Tomeu Vizoso wrote:
>>
>> Also include a configuration file that points to the out-of-tree CI
>> scripts.
>
>
> I think this para is outdated given ...
>
>> v8:
>> - Move all files specific to testing t
Hi
Am 25.10.22 um 09:12 schrieb Jeremy Rand:
Hi dri-devel,
I have two machines with ASPEED GPU's (ast Linux driver). One machine
is x86_64, running an ASRock Rack Tommy 90-SC02P1-00UBNZ GPU (AST2510
chipset) with KDE Plasma Wayland; the other is ppc64le, running an
integrated AST2500 GPU wi
tree: git://people.freedesktop.org/~airlied/linux.git 01.01-gsp-rm
head: 6be95d5e52818808565790c5ee3fd5569263bd36
commit: beea5ab213b28f7abf7c3405439ac9abcd1b8415 [132/180] drm/nouveau/disp:
add output aux xfer method
config: i386-allyesconfig (attached as .config)
compiler: gcc-11 (Debian 11.
Hi Andreas
Am 24.10.22 um 18:19 schrieb Andreas Thalhammer:
Am 24.10.22 um 13:31 schrieb Thomas Zimmermann:
Hi
Am 24.10.22 um 13:27 schrieb Greg KH:
On Mon, Oct 24, 2022 at 12:41:43PM +0200, Thorsten Leemhuis wrote:
Hi! Thx for the reply.
On 24.10.22 12:26, Thomas Zimmermann wrote:
Am 23.1
Am 25.10.22 um 10:16 schrieb Thomas Zimmermann:
Hi Andreas
Am 24.10.22 um 18:19 schrieb Andreas Thalhammer:
Am 24.10.22 um 13:31 schrieb Thomas Zimmermann:
Hi
Am 24.10.22 um 13:27 schrieb Greg KH:
On Mon, Oct 24, 2022 at 12:41:43PM +0200, Thorsten Leemhuis wrote:
Hi! Thx for the reply.
On
Hi
Am 25.10.22 um 10:45 schrieb Andreas Thalhammer:
[...]
Yeah, it's also part of a larger changeset. But I wouldn't want to
backport all those changes either.
Attached is a simple patch for linux-stable that adds the necessary fix.
If this still doesn't work, we should probably revert the prob
tree: git://people.freedesktop.org/~airlied/linux.git 01.01-gsp-rm
head: 6be95d5e52818808565790c5ee3fd5569263bd36
commit: e7be44d4385351aaf43f09786faa38d153722673 [87/180] drm/nouveau/acr: use
common falcon HS FW code for ACR FWs
config: microblaze-randconfig-r001-20221023 (attached as .config
From: Jason A. Donenfeld
> Sent: 24 October 2022 17:29
> To: linux-ker...@vger.kernel.org
>
> With char becoming unsigned by default, and with `char` alone being
> ambiguous and based on architecture, signed chars need to be marked
> explicitly as such. This fixes warnings like:
>
...
> ---
> dr
Add {begin,end}_fb_access helpers to run at the beginning and end of
an atomic commit. The begin_fb_access helper aquires resources that
are necessary to perform the atomic commit. It it similar to prepare_fb,
except that the resources are to be released at the end of the commit.
Resources acquired
This patchset adds the callbacks begin_fb_access and end_fb_access
to struct drm_plane_helper_funcs. They provide hooks to acquire and
release resources that are only held during the commit. It adds
related simple-KMS helpers and converts shadow-plane helpers.
As resource allocation often can fail
Move the vmap code for shadow-plane helpers from prepare_fb to
begin_fb_access helpers. Vunmap is now performed at the end of
the current pageflip, instead of the end of the following pageflip.
Reduces the duration of the mapping from while the framebuffer is
being displayed to just the atomic com
On 9/28/22 14:01, Alistair Popple wrote:
> This series aims to fix a number of page reference counting issues in
> drivers dealing with device private ZONE_DEVICE pages. These result in
> use-after-free type bugs, either from accessing a struct page which no
> longer exists because it has been remo
On Tue, Oct 25, 2022 at 11:21:57AM +0200, Thomas Zimmermann wrote:
> Hi
>
> Am 25.10.22 um 10:45 schrieb Andreas Thalhammer:
> [...]
> > > Yeah, it's also part of a larger changeset. But I wouldn't want to
> > > backport all those changes either.
> > >
> > > Attached is a simple patch for linux-s
Pushed to drm-misc-next.
Christian.
Am 24.10.22 um 23:26 schrieb Luben Tuikov:
The currently default Round-Robin GPU scheduling can result in starvation
of entities which have a large number of jobs, over entities which have
a very small number of jobs (single digit).
This can be illustrated i
Call sysfb_disable() from aperture_remove_conflicting_pci_devices()
before removing PCI devices. Without, simpledrm can still bind to
simple-framebuffer devices after the hardware driver has taken over
the hardware. Both drivers interfere with each other and results are
undefined.
Reported modeset
On Tue, Oct 25, 2022 at 12:27:11PM +1000, Dave Airlie wrote:
> The userspace for those is normally bespoke like ROCm, which uses
> amdkfd, and amdkfd doesn't operate like most device files from what I
> know, so I'm not sure we'd want it to operate as an accel device.
I intensely dislike this dir
On Tue, Oct 25, 2022 at 01:04:53PM +0200, Thomas Zimmermann wrote:
> Call sysfb_disable() from aperture_remove_conflicting_pci_devices()
> before removing PCI devices. Without, simpledrm can still bind to
> simple-framebuffer devices after the hardware driver has taken over
> the hardware. Both dri
Hello Dmitry Osipenko,
The patch 809d9c72c2f8: "dma-buf: Move dma_buf_attach() to dynamic
locking specification" from Oct 17, 2022, leads to the following
Smatch static checker warning:
drivers/dma-buf/dma-buf.c:957 dma_buf_dynamic_attach()
error: double unlocked 'dmabuf->resv' (o
Hi, thanks for detailed review. My responses inline.
On 10/25/2022 1:00 AM, Jeffrey Hugo wrote:
> On 9/24/2022 9:11 AM, Jacek Lawrynowicz wrote:
>> VPU stands for Versatile Processing Unit and it's a CPU-integrated
>> inference accelerator for Computer Vision and Deep Learning
>> applications.
>>
Hello Pin-yen Lin,
The patch 1051d302: "drm/bridge: it6505: Adapt runtime power
management framework" from Oct 4, 2022, leads to the following Smatch
static checker warning:
drivers/gpu/drm/bridge/ite-it6505.c:2712 it6505_extcon_work()
warn: pm_runtime_get_sync() also returns
On 10/25/22 14:41, Dan Carpenter wrote:
> Hello Dmitry Osipenko,
>
> The patch 809d9c72c2f8: "dma-buf: Move dma_buf_attach() to dynamic
> locking specification" from Oct 17, 2022, leads to the following
> Smatch static checker warning:
>
> drivers/dma-buf/dma-buf.c:957 dma_buf_dynamic_attac
Hi
Am 25.10.22 um 13:42 schrieb Jacek Lawrynowicz:
Hi, thanks for detailed review. My responses inline.
On 10/25/2022 1:00 AM, Jeffrey Hugo wrote:
On 9/24/2022 9:11 AM, Jacek Lawrynowicz wrote:
VPU stands for Versatile Processing Unit and it's a CPU-integrated
inference accelerator for Comput
Hi Niranjana,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on drm-tip/drm-tip]
url:
https://github.com/intel-lab-lkp/linux/commits/Niranjana-Vishwanathapura/drm-i915-vm_bind-Add-VM_BIND-functionality/20221025-150246
base: git://anongit.freedesktop.org/drm/drm
tree: git://people.freedesktop.org/~airlied/linux.git 01.01-gsp-rm
head: 6be95d5e52818808565790c5ee3fd5569263bd36
commit: 2428d9aef24a6a497b8740afadbb028c17b5e697 [173/180]
drm/nouveau/gsp/tu102-: add support for booting GSP-RM
config: microblaze-randconfig-r001-20221023 (attached as .config)
Hi,
please find some review comments below.
Am 24.09.22 um 17:11 schrieb Jacek Lawrynowicz:
VPU stands for Versatile Processing Unit and it's a CPU-integrated
inference accelerator for Computer Vision and Deep Learning
applications.
The VPU device consist of following componensts:
- Buttres
Hi
Am 24.09.22 um 17:11 schrieb Jacek Lawrynowicz:
Adds four types of GEM-based BOs for the VPU:
- shmem
- userptr
- internal
- prime
All types are implemented as struct ivpu_bo, based on
struct drm_gem_object. VPU address is allocated when buffer is created
except for imported prim
On Mon, Oct 24, 2022 at 8:29 PM Helge Deller wrote:
>
> On 10/24/22 18:29, Jason A. Donenfeld wrote:
> > With char becoming unsigned by default, and with `char` alone being
> > ambiguous and based on architecture, signed chars need to be marked
> > explicitly as such. This fixes warnings like:
> >
On 10/25/22 14:55, Jason A. Donenfeld wrote:
On Mon, Oct 24, 2022 at 8:29 PM Helge Deller wrote:
On 10/24/22 18:29, Jason A. Donenfeld wrote:
With char becoming unsigned by default, and with `char` alone being
ambiguous and based on architecture, signed chars need to be marked
explicitly as s
On Mon, Oct 24, 2022 at 08:43:58PM +0300, Oded Gabbay wrote:
> On Mon, Oct 24, 2022 at 6:21 PM Jeffrey Hugo wrote:
> >
> > On 10/22/2022 3:46 PM, Oded Gabbay wrote:
> > > The accelerator devices are exposed to user-space using a dedicated
> > > major. In addition, they are represented in /dev with
+ Luben
On Tue, Oct 25, 2022 at 2:55 AM brolerliew wrote:
>
> When entity move from one rq to another, current_entity will be set to NULL
> if it is the moving entity. This make entities close to rq head got
> selected more frequently, especially when doing load balance between
> multiple drm_gpu
On Wed 2022-10-19 17:01:46, John Ogness wrote:
> Since the console_lock is not being used for anything other than
> safe console list traversal, use srcu console list iteration instead.
>
> Signed-off-by: John Ogness
Reviewed-by: Petr Mladek
> ---
> drivers/video/fbdev/xen-fbfront.c | 8 +
On 9/11/2022 3:14 PM, Michał Winiarski wrote:
IDR is deprecated, and since XArray manages its own state with internal
locking, it simplifies the locking on DRM side.
Additionally, don't use the IRQ-safe variant, since operating on drm
minor is not done in IRQ context.
Signed-off-by: Michał Winia
On Tue, Oct 11, 2022 at 01:55:01PM +0200, Christian König wrote:
> Am 11.10.22 um 13:39 schrieb Simon Ser:
> > On Tuesday, October 11th, 2022 at 13:04, Christian König
> > wrote:
> >
> > > --- a/include/drm/drm_file.h
> > > +++ b/include/drm/drm_file.h
> > > @@ -54,7 +54,6 @@ struct file;
> > >
On 10/25/2022 5:42 AM, Jacek Lawrynowicz wrote:
Hi, thanks for detailed review. My responses inline.
On 10/25/2022 1:00 AM, Jeffrey Hugo wrote:
On 9/24/2022 9:11 AM, Jacek Lawrynowicz wrote:
VPU stands for Versatile Processing Unit and it's a CPU-integrated
inference accelerator for Computer V
Series is
Reviewed-by: Harry Wentland
Harry
On 2022-10-24 15:22, Hamza Mahfooz wrote:
> This reverts commit 55eea8ef98641f6e1e1c202bd3a49a57c1dd4059.
>
> This quirk is now handled in the DRM core, so we can drop all of
> the internal code that was added to handle it.
>
> Signed-off-by: Hamza
On Tue, Oct 25, 2022 at 7:15 AM Jason Gunthorpe wrote:
>
> On Tue, Oct 25, 2022 at 12:27:11PM +1000, Dave Airlie wrote:
>
> > The userspace for those is normally bespoke like ROCm, which uses
> > amdkfd, and amdkfd doesn't operate like most device files from what I
> > know, so I'm not sure we'd w
@Daniel Vetter
, @Dave Airlie
Any objections taking this through the AMD tree or would you rather it
landed via drm-misc?
Thanks,
Alex
On Tue, Oct 25, 2022 at 10:21 AM Harry Wentland wrote:
>
> Series is
>
> Reviewed-by: Harry Wentland
>
> Harry
>
> On 2022-10-24 15:22, Hamza Mahfooz wrote:
>
On Tue, Oct 25, 2022 at 10:21:34AM -0400, Alex Deucher wrote:
> E.g., the kfd node provides platform level compute
> topology information; e.g., the NUMA details for connected GPUs and
> CPUs, non-GPU compute node information, cache level topologies, etc.
See, this is exactly what I'm talking abo
On Tue, Oct 25, 2022 at 10:34 AM Jason Gunthorpe wrote:
>
> On Tue, Oct 25, 2022 at 10:21:34AM -0400, Alex Deucher wrote:
>
> > E.g., the kfd node provides platform level compute
> > topology information; e.g., the NUMA details for connected GPUs and
> > CPUs, non-GPU compute node information, cac
Hi all,
On Tue, 25 Oct 2022 at 08:32, Daniel Vetter wrote:
> On Fri, 9 Sept 2022 at 19:18, Daniel Stone wrote:
> > But equally - and sorry for not jumping on the IRC (?) discussion as I was
> > in the middle of other stuff when it came up - I'm don't think this is the
> > right plan.
> >
> > M
On Mon, 24 Oct 2022 15:54:53 -0700, Vinay Belgaumkar wrote:
>
> GuC will set the min/max frequencies to theoretical max on
> ATS-M. This will break kernel ABI, so limit min/max frequency
> to RP0(platform max) instead.
>
> Also modify the SLPC selftest to update the min frequency
> when we have a s
Hello,
These patches add support for the JDI LPM102A188A display panel,
found in the Google Pixel C.
Patch 1 adds the DT bindings for the panel.
Patch 2 adds a register clear to the Tegra DSI driver, needed for the
panel initialization commands to be properly sent.
Patch 3 adds the panel driver
The LPM102A188A is a 10.2" 2560x1800 IPS panel found in
the Google Pixel C.
Signed-off-by: Diogo Ivo
---
Changes in v2:
- removed the touch screen property
.../display/panel/jdi,lpm102a188a.yaml| 94 +++
1 file changed, 94 insertions(+)
create mode 100644
Documentatio
In cases where the DSI module is left on by the bootloader
some panels may fail to initialize if the enable register is not cleared
before the panel's initialization sequence is sent, so clear it if that
is the case.
Signed-off-by: Diogo Ivo
---
Changes in v2:
- detect if the DSI module is on b
The JDI LPM102A188A is a 2560x1800 IPS panel found in the Google Pixel C.
This driver is based on the downstream GPLv2 driver released by Google
written by Sean Paul [1], which was then adapted to the newer kernel APIs.
[1]:
https://android.googlesource.com/kernel/tegra/+/refs/heads/android-tegra
The Google Pixel C has a JDI LPM102A188A display panel. Add a
DT node for it. Tested on Pixel C.
Signed-off-by: Diogo Ivo
---
Changes in v2:
- renamed backlight node to a generic name
- removed underscores
arch/arm64/boot/dts/nvidia/tegra210-smaug.dts | 70 +++
1 file changed,
change-id: 20221025-ts050-timings-2fb4b034a268
Best regards,
--
Neil Armstrong
Change ttm_resource structure from num_pages to size_t size in bytes.
v1 -> v2: change PFN_UP(dst_mem->size) to ttm->num_pages
v1 -> v2: change bo->resource->size to bo->base.size at some places
v1 -> v2: remove the local variable
v1 -> v2: cleanup cmp_size_smaller_first()
Signed-off-by: Somalapur
https://bugzilla.kernel.org/show_bug.cgi?id=216625
Bug ID: 216625
Summary: [regression] GPU lockup on Radeon R7 Kaveri
Product: Drivers
Version: 2.5
Kernel Version: 5.19.16-100.fc35.x86_64
Hardware: All
OS: Linux
https://bugzilla.kernel.org/show_bug.cgi?id=216625
Pierre Ossman (pierre-bugzi...@ossman.eu) changed:
What|Removed |Added
Tree|Mainline|Fedora
--
Yo
Am 25.10.22 um 15:59 schrieb Michał Winiarski:
On Tue, Oct 11, 2022 at 01:55:01PM +0200, Christian König wrote:
Am 11.10.22 um 13:39 schrieb Simon Ser:
On Tuesday, October 11th, 2022 at 13:04, Christian König
wrote:
--- a/include/drm/drm_file.h
+++ b/include/drm/drm_file.h
@@ -54,7 +54,6 @@
On Sat, Oct 22, 2022 at 7:06 PM Chris Healy wrote:
>
> I can't speak to why you are experiencing issues when using the GPU,
> but in the examples you gave, the example that is working is using a
> SW based GL implementation instead of the real GPU. This can be
> determined by looking at the GL_RE
https://bugzilla.kernel.org/show_bug.cgi?id=216625
Alex Deucher (alexdeuc...@gmail.com) changed:
What|Removed |Added
CC||alexdeuc...@gmail.c
Am 25.10.22 um 17:50 schrieb Somalapuram Amaranath:
Change ttm_resource structure from num_pages to size_t size in bytes.
v1 -> v2: change PFN_UP(dst_mem->size) to ttm->num_pages
v1 -> v2: change bo->resource->size to bo->base.size at some places
v1 -> v2: remove the local variable
v1 -> v2: clea
On Tue, 25 Oct 2022 at 16:51, Somalapuram Amaranath
wrote:
>
> Change ttm_resource structure from num_pages to size_t size in bytes.
> v1 -> v2: change PFN_UP(dst_mem->size) to ttm->num_pages
> v1 -> v2: change bo->resource->size to bo->base.size at some places
> v1 -> v2: remove the local variabl
Looking...
Regards,
Luben
On 2022-10-25 09:35, Alex Deucher wrote:
> + Luben
>
> On Tue, Oct 25, 2022 at 2:55 AM brolerliew wrote:
>>
>> When entity move from one rq to another, current_entity will be set to NULL
>> if it is the moving entity. This make entities close to rq head got
>> selected
This legacy module knob has become uAPI, when set on the vfio_iommu_type1
it disables some security protections in the iommu drivers. Move the
storage for this knob to vfio_main.c so that iommufd can access it too.
Signed-off-by: Jason Gunthorpe
---
drivers/vfio/vfio.h | 2 ++
driver
These functions don't really assign anything anymore, they just increment
some refcounts and do a sanity check. Call them
vfio_group_[un]use_container()
Signed-off-by: Jason Gunthorpe
---
drivers/vfio/container.c | 14 ++
drivers/vfio/vfio.h | 4 ++--
drivers/vfio/vfio_main.c |
This series provides an alternative container layer for VFIO implemented
using iommufd. This is optional, if CONFIG_IOMMUFD is not set then it will
not be compiled in.
At this point iommufd can be injected by passing in a iommfd FD to
VFIO_GROUP_SET_CONTAINER which will use the VFIO compat layer i
iommufd doesn't establish the iommu_domains until after the device FD is
opened, even if the container has been set. This design is part of moving
away from the group centric iommu APIs.
This is fine, except that the normal sequence of establishing the kvm
wbindv won't work:
group = open("/dev
The only thing this function does is assert the group has an assigned
container and incrs refcounts.
The overall model we have is that once a conatiner_users refcount is
incremented it cannot be de-assigned from the group -
vfio_group_ioctl_unset_container() will fail and the group FD cannot be
cl
This error unwind is getting complicated. Move all the code into two
pair'd function. The functions should be called when the open_count == 1
after incrementing/before decrementing.
Signed-off-by: Jason Gunthorpe
---
drivers/vfio/vfio_main.c | 95 ++--
1 file
If the VFIO container is compiled out, give a kconfig option for iommufd
to provide the miscdev node with the same name and permissions as vfio
uses.
The compatibility node supports the same ioctls as VFIO and automatically
enables the VFIO compatible pinned page accounting mode.
Signed-off-by: J
Emulated VFIO devices are calling vfio_register_emulated_iommu_dev() and
consist of all the mdev drivers.
Like the physical drivers, support for iommufd is provided by the driver
supplying the correct correct standard ops. Provide ops from the core that
duplicate what vfio_register_emulated_iommu_
This makes VFIO_GROUP_SET_CONTAINER accept both a vfio container FD and an
iommufd.
In iommufd mode an IOAS will exist after the SET_CONTAINER, but it will
not be attached to any groups.
>From a VFIO perspective this means that the VFIO_GROUP_GET_STATUS and
VFIO_GROUP_FLAGS_VIABLE works subtly di
Hi,
On 10/24/22 22:30, Matthew Garrett wrote:
> On Tue, Sep 27, 2022 at 01:04:52PM +0200, Hans de Goede wrote:
>
>> So to fix this we need to make acpi_video_get_backlight_type()
>> return native on the Acer Chromebook Spin 713.
>
> Isn't the issue broader than that? Unless the platform is Windo
This creates the iommufd_device for the physical VFIO drivers. These are
all the drivers that are calling vfio_register_group_dev() and expect the
type1 code to setup a real iommu_domain against their parent struct
device.
The design gives the driver a choice in how it gets connected to iommufd
by
Add a kconfig CONFIG_VFIO_CONTAINER that controls compiling the container
code. If 'n' then only iommufd will provide the container service. All the
support for vfio iommu drivers, including type1, will not be built.
This allows a compilation check that no inappropriate dependencies between
the de
On Tue, Oct 25, 2022 at 08:50:54PM +0200, Hans de Goede wrote:
> That is a valid point, but keep in mind that this is only used on ACPI
> platforms and then only on devices with a builtin LCD panel and then
> only by GPU drivers which actually call acpi_video_get_backlight_type(),
> so e.g. not by
Am 2022-10-25 um 03:28 schrieb Yang Li:
./drivers/gpu/drm/amd/amdkfd/kfd_migrate.c:985:58-62: ERROR: p is NULL but
dereferenced.
Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=2549
Reported-by: Abaci Robot
Signed-off-by: Yang Li
---
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c | 3 ++-
1
Am 2022-10-25 um 02:09 schrieb Yang Li:
drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_migrate.c:331
svm_migrate_copy_to_vram() warn: inconsistent indenting
Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=2537
Reported-by: Abaci Robot
Signed-off-by: Yang Li
This patch doesn't apply to our am
Am 2022-10-25 um 05:12 schrieb Paulo Miguel Almeida:
kfd_pm4_headers_diq.h header is a leftover from the old H/W debugger
module support added on commit . That implementation
was removed after a while and the last file that included that header
was removed on commit <5bdd3eb253544b1>.
This patch
From: Nathan Chancellor
When booting with CONFIG_CFI_CLANG, there are numerous violations when
accessing the files under
/sys/devices/pci:00/:00:02.0/drm/card0/gt/gt0:
$ cd /sys/devices/pci:00/:00:02.0/drm/card0/gt/gt0
$ grep . *
id:0
punit_req_freq_mhz:350
rc6_enable:
Simplify the MSM IOMMU code a bit. This moves iommu_domain_alloc() and
iommu_set_pgtable_quirks() calls to msm_iommu_new() to get rid of the
disbalance, when the iommu domain is allocated by the caller of
msm_iommu_new() and then it is freed by the msm_iommu code itself.
Changes since v2:
- Reorde
After the msm_iommu instance is created, the IOMMU domain is completely
handled inside the msm_iommu code. Move the iommu_domain_alloc() call
into the msm_iommu_new() to simplify callers code.
Reported-by: kernel test robot
Signed-off-by: Dmitry Baryshkov
---
drivers/gpu/drm/msm/adreno/a6xx_gmu
The function a6xx_create_address_space() is mostly a copy of
adreno_iommu_create_address_space() with added quirk setting. Rework
these two functions to be a thin wrappers around a common helper.
Signed-off-by: Dmitry Baryshkov
---
drivers/gpu/drm/msm/adreno/a3xx_gpu.c | 2 +-
drivers/gpu/drm
Hi Matthew,
On 10/25/22 21:32, Matthew Garrett wrote:
> On Tue, Oct 25, 2022 at 08:50:54PM +0200, Hans de Goede wrote:
>
>> That is a valid point, but keep in mind that this is only used on ACPI
>> platforms and then only on devices with a builtin LCD panel and then
>> only by GPU drivers which a
On 10/25/2022 15:25, Hans de Goede wrote:
Hi Matthew,
On 10/25/22 21:32, Matthew Garrett wrote:
On Tue, Oct 25, 2022 at 08:50:54PM +0200, Hans de Goede wrote:
That is a valid point, but keep in mind that this is only used on ACPI
platforms and then only on devices with a builtin LCD panel and
1 - 100 of 147 matches
Mail list logo
