- Nov 11, 2023
-
-
There are two problems with the current method of determining the virtio-gpu debug name. 1) TASK_COMM_LEN is defined to be 16 bytes only, and this is a Linux kernel idiom (see PR_SET_NAME + PR_GET_NAME). Though, Android/FreeBSD get around this via setprogname(..)/getprogname(..) in libc. On Android, names longer than 16 bytes are common. For example, one often encounters a program like "com.android.systemui". The virtio-gpu spec allows the debug name to be up to 64 bytes, so ideally userspace should be able to set debug names up to 64 bytes. 2) The current implementation determines the debug name using whatever task initiated virtgpu. This is could be a "RenderThread" of a larger program, when we actually want to propagate the debug name of the program. To fix these issues, add a new CONTEXT_INIT param that allows userspace to set the debug name when creating a context. It takes a null-terminated C-string as the param value. The length of the string (excluding the terminator) **should** be <= 64 bytes. Otherwise, the debug_name will be truncated to 64 bytes. Link to open-source userspace: https://android-review.googlesource.com/c/platform/hardware/google/gfxstream/+/2787176 Signed-off-by:
Gurchetan Singh <gurchetansingh@chromium.org> Reviewed-by:
Josh Simonot <josh.simonot@gmail.com> Reviewed-by:
Dmitry Osipenko <dmitry.osipenko@collabora.com> Signed-off-by:
-
drm_virtgpu_context_set_param defines both param and value to be u64s. Signed-off-by:
-
- Jun 03, 2023
-
-
Dmitry Osipenko authored
Move virtio_gpu_execbuffer_ioctl() into separate virtgpu_submit.c file, refactoring and optimizing the code along the way to ease addition of new features to the ioctl. The optimization is done by using optimal ordering of the job's submission steps, reducing code path from the start of the ioctl to the point of pushing job to virtio queue. Job's initialization is now performed before in-fence is awaited and out-fence setup is made after sending out job to virtio. Reviewed-by:
Rob Clark <robdclark@gmail.com> Reviewed-by:
Emil Velikov <emil.velikov@collabora.com> Tested-by:
Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Signed-off-by:
-
- Feb 06, 2023
-
-
An interrupted dma_fence_wait() becomes an -ERESTARTSYS returned to userspace ioctl(DRM_IOCTL_VIRTGPU_EXECBUFFER) calls, prompting to retry the ioctl(), but the passed exbuf->fence_fd has been reset to -1, making the retry attempt fail at sync_file_get_fence(). The uapi for DRM_IOCTL_VIRTGPU_EXECBUFFER is changed to retain the passed value for exbuf->fence_fd when returning anything besides a successful result from the ioctl. Fixes: 2cd7b6f0 ("drm/virtio: add in/out fence support for explicit synchronization") Signed-off-by:
Ryan Neph <ryanneph@chromium.org> Reviewed-by:
-
- Jan 09, 2023
-
-
Userspace can guess the handle value and try to race GEM object creation with handle close, resulting in a use-after-free if we dereference the object after dropping the handle's reference. For that reason, dropping the handle's reference must be done *after* we are done dereferencing the object. Signed-off-by:
Rob Clark <robdclark@chromium.org> Reviewed-by:
Chia-I Wu <olvaffe@gmail.com> Fixes: 62fb7a5e ("virtio-gpu: add 3d/virgl support") Cc: stable@vger.kernel.org Signed-off-by:
-
- Sep 23, 2022
-
-
The ->ring_idx_mask variable is a u64 so static checkers, Smatch in this case, complain if the BIT() is not also a u64. drivers/gpu/drm/virtio/virtgpu_ioctl.c:50 virtio_gpu_fence_event_create() warn: should '(1 << ring_idx)' be a 64 bit type? Fixes: cd7f5ca3 ("drm/virtio: implement context init: add virtio_gpu_fence_event") Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by:
Gerd Hoffmann <kraxel@redhat.com>
-
- Aug 25, 2022
-
-
When VIRTGPU_EXECBUF_RING_IDX is used, we should be considering the timeline that the EB if running on rather than the global driver fence context. Fixes: 85c83ea9 ("drm/virtio: implement context init: allocate an array of fence contexts") Signed-off-by:
-
- Jun 09, 2022
-
-
'cache_ent' could be set NULL inside virtio_gpu_cmd_get_capset() and it will lead to a NULL dereference by a lately use of it (i.e., ptr = cache_ent->caps_cache). Fix it with a NULL check. Fixes: 62fb7a5e ("virtio-gpu: add 3d/virgl support") Signed-off-by:
Xiaomeng Tong <xiam0nd.tong@gmail.com> Reviewed-by:
-
- Apr 07, 2022
-
-
Christian König authored
This change adds the dma_resv_usage enum and allows us to specify why a dma_resv object is queried for its containing fences. Additional to that a dma_resv_usage_rw() helper function is added to aid retrieving the fences for a read or write userspace submission. This is then deployed to the different query functions of the dma_resv object and all of their users. When the write paratermer was previously true we now use DMA_RESV_USAGE_WRITE and DMA_RESV_USAGE_READ otherwise. v2: add KERNEL/OTHER in separate patch v3: some kerneldoc suggestions by Daniel v4: some more kerneldoc suggestions by Daniel, fix missing cases lost in the rebase pointed out by Bas. Signed-off-by:Christian König <christian.koenig@amd.com> Reviewed-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20220407085946.744568-2-christian.koenig@amd.com
-
- Mar 25, 2022
-
-
With native userspace drivers in guest, a lot of GEM objects need to be neither shared nor mappable. And in fact making everything mappable and/or sharable results in unreasonably high fd usage in host VMM. Signed-off-by:
-
- Nov 29, 2021
-
-
The current virtgpu implementation of poll(..) drops events when VIRTGPU_CONTEXT_PARAM_POLL_RINGS_MASK is enabled (otherwise it's like a normal DRM driver). This is because paravirtualized userspaces receives responses in a buffer of type BLOB_MEM_GUEST, not by read(..). To be in line with other DRM drivers and avoid specialized behavior, it is possible to define a dummy event for virtgpu. Paravirtualized userspace will now have to call read(..) on the DRM fd to receive the dummy event. Fixes: b1079043 ("drm/virtgpu api: create context init feature") Reported-by:
-
- Oct 29, 2021
-
-
The left shift of unsigned int 32 bit integer constant 1 is evaluated using 32 bit arithmetic and then assigned to a signed 64 bit integer. In the case where value is 32 or more this can lead to an overflow (value can be in range 0..MAX_CAPSET_ID (63). Fix this by shifting the value 1ULL instead. Addresses-Coverity: ("Uninitentional integer overflow") Fixes: 4fb530e5 ("drm/virtio: implement context init: support init ioctl") Signed-off-by:Colin Ian King <colin.king@canonical.com> Link: http://patchwork.freedesktop.org/patch/msgid/20210930102748.16922-1-colin.king@canonical.com Signed-off-by:
-
The left shift of unsigned int 32 bit integer constant 1 is evaluated using 32 bit arithmetic and then assigned to a signed 64 bit integer. In the case where i is 32 or more this can lead to an overflow. Fix this by shifting the value 1ULL instead. Addresses-Coverity: ("Uninitentional integer overflow") Fixes: 8d6b006e ("drm/virtio: implement context init: handle VIRTGPU_CONTEXT_PARAM_POLL_RINGS_MASK") Signed-off-by:
-
- Sep 29, 2021
-
-
This advertises the context init feature to userspace, along with a mask of supported capabilities. Signed-off-by:
Lingfeng Yang <lfy@google.com> Link: http://patchwork.freedesktop.org/patch/msgid/20210921232024.817-13-gurchetansingh@chromium.org Signed-off-by:
-
Similar to DRM_VMW_EVENT_FENCE_SIGNALED. Sends a pollable event to the DRM file descriptor when a fence on a specific ring is signaled. One difference is the event is not exposed via the UAPI -- this is because host responses are on a shared memory buffer of type BLOB_MEM_GUEST [this is the common way to receive responses with virtgpu]. As such, there is no context specific read(..) implementation either -- just a poll(..) implementation. Signed-off-by:
Nicholas Verne <nverne@chromium.org> Link: http://patchwork.freedesktop.org/patch/msgid/20210921232024.817-12-gurchetansingh@chromium.org Signed-off-by:
-
For the Sommelier guest Wayland proxy, it's desirable for the DRM fd to be pollable in response to an host compositor event. This can also be used by the 3D driver to poll events on a CPU timeline. This enables the DRM fd associated with a particular 3D context to be polled independent of KMS events. The parameter VIRTGPU_CONTEXT_PARAM_POLL_RINGS_MASK specifies the pollable rings. Signed-off-by:
-
We don't want fences from different 3D contexts (virgl, gfxstream, venus) to be on the same timeline. With explicit context creation, we can specify the number of ring each context wants. Execbuffer can specify which ring to use. Signed-off-by:
-
These were defined in the previous commit. We'll need these parameters when allocating a dma_fence. The use case for this is multiple synchronizations timelines. The maximum number of timelines per 3D instance will be 32. Usually, only 2 are needed -- one for CPU commands, and another for GPU commands. As such, we'll need to specify these parameters when allocating a dma_fence. vgdev->fence_drv.context is the "default" fence context for 2D mode and old userspace. Signed-off-by:
-
This implements the context initialization ioctl. A list of params is passed in by userspace, and kernel driver validates them. The only currently supported param is VIRTGPU_CONTEXT_PARAM_CAPSET_ID. If the context has already been initialized, -EEXIST is returned. This happens after Linux userspace does dumb_create + followed by opening the Mesa virgl driver with the same virtgpu instance. However, for most applications, 3D contexts will be explicitly initialized when the feature is available. Signed-off-by:
Anthoine Bourgeois <anthoine.bourgeois@gmail.com> Acked-by:
-
- Jun 06, 2021
-
-
Christian König authored
The functions can be called both in _rcu context as well as while holding the lock. v2: add some kerneldoc as suggested by Daniel v3: fix indentation Signed-off-by:
Jason Ekstrand <jason@jlekstrand.net> Acked-by:
-
- Mar 09, 2021
-
-
virtio_gpu_object array is not freed or unlocked in some failed cases. Signed-off-by:
xndcn <xndchn@gmail.com> Link: http://patchwork.freedesktop.org/patch/msgid/20210305151819.14330-1-xndchn@gmail.com Signed-off-by:
-
- Nov 20, 2020
-
-
For coherency, all ioctls are suffixed Signed-off-by:
-
- Sep 29, 2020
-
-
New api changes are now available to userspace. Also, the comparison to true is redundant, so remove it. Signed-off-by:
Tomeu Vizoso <tomeu.vizoso@collabora.com> Link: http://patchwork.freedesktop.org/patch/msgid/20200924003214.662-19-gurchetansingh@chromium.org Signed-off-by:
-
Gerd Hoffmann authored
Implement resource create blob as specified. Signed-off-by:
-
The stride field has never been used, so repurpose it to be "blob_mem". This way, userspace can know the memory properties of the blob if it's passed between userspace processes and no suitable userspace API exists to transmit that knowledge. Signed-off-by:
-
The old transfer ioctls may work on blob resources, and there is no TRANSFER_BLOB hypercall now for simplicity. The guest may have a image view on the blob resources such that the stride is not equal to width * bytes_per_pixel. For host-only blobs, we can repurpose the transfer ioctls to synchronize caches as well. For guest-only blobs, these operations are undefined for now so leave them out. Also, with seamless Wayland integration between guest/host looking increasingly attractive, it also makes sense to keep track of one value for stride. Signed-off-by:
-
- Aug 07, 2020
-
-
We should put the reference count of the fence after calling virtio_gpu_cmd_submit(). So add the missing dma_fence_put(). Fixes: 2cd7b6f0 ("drm/virtio: add in/out fence support for explicit synchronization") Co-developed-by:
Xin He <hexin.op@bytedance.com> Signed-off-by:
Qi Liu <liuqi.16@bytedance.com> Reviewed-by:
Muchun Song <songmuchun@bytedance.com> Link: http://patchwork.freedesktop.org/patch/msgid/20200721101647.42653-1-hexin.op@bytedance.com Signed-off-by:
-
- May 19, 2020
-
-
Spelling out _unlocked for each and every driver is a annoying. Especially if we consider how many drivers, do not know (or need to) about the horror stories involving struct_mutex. Just drop the suffix. It makes the API cleaner. Done via the following script: __from=drm_gem_object_put_unlocked __to=drm_gem_object_put for __file in $(git grep --name-only $__from); do sed -i "s/$__from/$__to/g" $__file; done Cc: David Airlie <airlied@linux.ie> Cc: Gerd Hoffmann <kraxel@redhat.com> Cc: Daniel Vetter <daniel@ffwll.ch> Signed-off-by:
Sam Ravnborg <sam@ravnborg.org> Acked-by:
Thomas Zimmermann <tzimmermann@suse.de> Link: https://patchwork.freedesktop.org/patch/msgid/20200515095118.2743122-36-emil.l.velikov@gmail.com
-
- May 04, 2020
-
-
If 3D is enabled, but userspace requests a dumb buffer, we will call CTX_ATTACH_RESOURCE before actually creating the context. Fixes: 72b48ae8 ("drm/virtio: enqueue virtio_gpu_create_context after the first 3D ioctl") Signed-off-by:
-
- Apr 17, 2020
-
-
Michael S. Tsirkin authored
In preparation to virtio header changes, include uaccess.h directly as this file is using copy to/from user. Signed-off-by:Michael S. Tsirkin <mst@redhat.com>
-
- Apr 03, 2020
-
-
The first 3D ioctl will take care of notification. Signed-off-by:
-
- Feb 25, 2020
-
-
For old userspace, initialization will still be implicit. For backwards compatibility, enqueue virtio_gpu_cmd_context_create after the first 3D ioctl. v3: staticify virtio_gpu_create_context remove notify to batch vm-exit v6: Remove nested 3D checks (emil.velikov): - unify 3D check in resource create v7: Remove check when getting capabilities Reviewed-by: -
Use an boolean variable to track whether a context has been initiated. v5: Fix possible race and sleep via mutex (olv) Reviewed-by:
-
We currently create an OpenGL context when opening the DRM fd if 3D is available. We may need other context types (VK,..) in the future, and the plan is to have explicit initialization for that. For explicit initialization to work, we need to factor out virtio_gpu_create_context from driver initialization. v2: Move context handle initialization too (olv) v6: Remove redundant 3D check (emil.velikov) Reviewed-by:
-
Minor cleanup, change: - file_priv--> file, - drm_file --> file. Reviewed-by:
-
- Feb 17, 2020
-
-
Gerd Hoffmann authored
Move all remaining virtio_gpu_notify() calls from virtio_gpu_cmd_* to the callers, for consistency reasons. Signed-off-by:
-
Gerd Hoffmann authored
Move virtio_gpu_notify() to higher-level functions for virtio_gpu_cmd_resource_flush(), virtio_gpu_cmd_set_scanout() and virtio_gpu_cmd_transfer_to_host_{2d,3d}(). virtio_gpu_primary_plane_update() will notify only once for a series of commands (restores plane update command batching). Signed-off-by:
-
- Feb 13, 2020
-
-
Gerd Hoffmann authored
Lockdep says we can't call vmemdup() while having objects reserved because it needs the mmap semaphore. So reorder the calls reserve the objects later. Signed-off-by:
-
- Nov 20, 2019
-
-
Gerd Hoffmann authored
Be consistent with the rest of the code base. No functional change. v2: - fix sparse warnings for virtio_gpu_cmd_transfer_to_host_2d call. - move convert_to_hw_box helper function. Signed-off-by:
-
- Sep 12, 2019
-
-
Userspace requested command buffer allocations could be too large to make as a contiguous allocation. Use vmalloc if necessary to satisfy those allocations. Signed-off-by:
David Riley <davidriley@chromium.org> Link: http://patchwork.freedesktop.org/patch/msgid/20190911181403.40909-3-davidriley@chromium.org Signed-off-by:
-
