Linaro-mm-sig

linaro-mm-sig@lists.linaro.org

5 participants
2540 discussions

[PATCH] drm:amd:amdgpu: Fix missing bo unlock in failure path

by Sukrut Bellary

smatch warning - inconsistent handling of buffer object reserve and unreserve. Signed-off-by: Sukrut Bellary <sukrut.bellary(a)linux.com> --- drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 278416acf060..5de44d7e92de 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -4686,8 +4686,10 @@ static int gfx_v8_0_kiq_resume(struct amdgpu_device *adev) return r; r = amdgpu_bo_kmap(ring->mqd_obj, &ring->mqd_ptr); - if (unlikely(r != 0)) + if (unlikely(r != 0)) { + amdgpu_bo_unreserve(ring->mqd_obj); return r; + } gfx_v8_0_kiq_init_queue(ring); amdgpu_bo_kunmap(ring->mqd_obj); -- 2.34.1

1 year, 8 months

[PATCH v10 0/2] drm: add kms driver for loongson display controller

by Sui Jingfeng

Loongson display controller IP has been integrated in both Loongson north bridge chipsets(ls7a1000/ls7a2000) and SoCs(ls2k1000/ls2k2000), it has been included in Loongson self-made BMC products. This display controller is a PCI device in all of chips mentiond, it has two display pipes which support primary planes and cursor plane. For the DC in ls7a1000 and ls2k1000, each display pipe has a DVO output interface which provide RGB888 signals, vertical & horizontal synchronisations and the pixel clock. Each CRTC is able to support 1920x1080@60Hz, the maximum resolution is 2048x2048 according to the hardware spec. For the DC in LS7A2000, each display pipe is equipped with a built-in HDMI encoder which is compliant with the HDMI 1.4 specification, thus it support 3840x2160@30Hz. The first display pipe is also equipped with a transparent vga encoder which is parallel with the HDMI encoder. The DC in LS7A2000 is more complete compare with the one in old chips, besides above feature, it has two hardware cursors, two hardware vblank counter and two scanout position recorders unit. It also support tiled framebuffer format which can be used to scan out the framebuffer rendered by the LoongGPU directly. v1 -> v2: 1) Use hpd status reg when polling for ls7a2000 2) Fix all warnings emerged when compile with W=1 v2 -> v3: 1) Add COMPILE_TEST in Kconfig and make the driver off by default 2) Alphabetical sorting headers (Thomas) 3) Untangle register access functions as much as possible (Thomas) 4) Switch to TTM based memory manager and prefer cached mapping for Loongson SoC (Thomas) 5) Add chip id detection method, now all models are distinguishable. 6) Revise builtin HDMI phy driver, nearly all main stream mode below 4K@30Hz is tested, this driver supported these mode very well including clone display mode and extend display mode. v3 -> v4: 1) Quickly fix a small mistake. v4 -> v5: 1) Drop potential support for Loongson 2K series SoC temporary, this part should be resend with the DT binding patch in the future. 2) Add per display pipe debugfs support to the builtin HDMI encoder. 3) Rewrite atomic_update() for hardware cursors plane(Thomas) 4) Rewrite encoder and connector initialization part, untangle it according to the chip(Thomas). v5 -> v6: 1) Remove stray code which didn't get used, say lsdc_of_get_reserved_ram 2) Fix all typos I could found, make sentences and code more readable 3) Untangle lsdc_hdmi*_connector_detect() function according to the pipe 4) After a serious consideration, we rename this driver as loongson. Because we also have drivers toward the LoongGPU IP in LS7A2000 and LS2K2000. Besides, there are also drivers about the external encoder, HDMI audio driver and vbios support etc. This patch only provide DC driver part, my teammate Li Yi believe that loongson will be more suitable for loongson graphics than lsdc in the long run. loongson.ko = LSDC + LoongGPU + encoders driver + vbios/DT ... v6 -> v7: 1) Add prime support, self-sharing is works. sharing buffer with etnaviv is also tested, and its works with limitation. 2) Implement buffer objects tracking with list_head. 3) S3(sleep to RAM) is tested on ls3a5000+ls7a2000 evb and it works. 4) Rewrite lsdc_bo_move, since ttm core stop allocating resources during BO creation. Patch V1 ~ V6 of this series no longer works on latest kernel. Thus, we send V7 to revival them. v7 -> v8: 1) Zero a compile warnnings on 32-bit platform, compile with W=1 2) Revise lsdc_bo_gpu_offset() and minor cleanup 3) Pageflip tested on the virtual terminal with following commands modetest -M loongson -s 32:1920x1080 -v modetest -M loongson -s 34:1920x1080 -v -F tiles It works like a charm, when running pageflip test with dual screnn configuration, another two additional bo created by the modetest emerged, VRAM usage up to 40+MB, well we have at least 64MB, still enough. # cat bos bo[0000]: size: 8112kB VRAM bo[0001]: size: 16kB VRAM bo[0002]: size: 16kB VRAM bo[0003]: size: 16208kB VRAM bo[0004]: size: 8112kB VRAM bo[0005]: size: 8112kB VRAM v8 -> v9: 1) Select I2C and I2C_ALGOBIT in Kconfig and should depend on MMU. 2) Using pci_get_domain_bus_and_slot to get the GPU device. 3) Other minor improvements. Those patches are tested on ls3a5000 + ls7a1000 CRB, ls3a5000 + ls7a2000 evb, and lemote a1901 board(ls3a4000 + ls7a1000). On loongson mips CPU, the write combine support should be enabled, to get a decent performance for writing framebuffer data to the VRAM. v9 -> v10: 1) Revise lsdc_drm_freeze() to implement S3 completely and correctly. I suddenly realized that pinned buffer can not move and VRAM lost power when sleep to RAM. Thus, the data in the buffer who is pinned in VRAM will get lost when resume. Yet it's not big problem because we are software rendering solution which relay on the CPU update the front framebuffer. We can see the garbage data when resume from S3, but the screen will show correct image as I move the cursor which is due to the cpu repaint. v10 of this patch make S3 perfect by unpin all of BOs in VRAM, evict them all to system RAM. Sui Jingfeng (2): MAINTAINERS: add maintainers for DRM LOONGSON driver drm: add kms driver for loongson display controller MAINTAINERS | 7 + drivers/gpu/drm/Kconfig | 2 + drivers/gpu/drm/Makefile | 1 + drivers/gpu/drm/loongson/Kconfig | 17 + drivers/gpu/drm/loongson/Makefile | 16 + drivers/gpu/drm/loongson/lsdc_crtc.c | 381 ++++++++++++++++ drivers/gpu/drm/loongson/lsdc_debugfs.c | 261 +++++++++++ drivers/gpu/drm/loongson/lsdc_drv.c | 526 ++++++++++++++++++++++ drivers/gpu/drm/loongson/lsdc_drv.h | 324 ++++++++++++++ drivers/gpu/drm/loongson/lsdc_gem.c | 294 +++++++++++++ drivers/gpu/drm/loongson/lsdc_gem.h | 26 ++ drivers/gpu/drm/loongson/lsdc_i2c.c | 171 +++++++ drivers/gpu/drm/loongson/lsdc_irq.c | 86 ++++ drivers/gpu/drm/loongson/lsdc_irq.h | 12 + drivers/gpu/drm/loongson/lsdc_output.c | 563 ++++++++++++++++++++++++ drivers/gpu/drm/loongson/lsdc_output.h | 14 + drivers/gpu/drm/loongson/lsdc_plane.c | 432 ++++++++++++++++++ drivers/gpu/drm/loongson/lsdc_pll.c | 338 ++++++++++++++ drivers/gpu/drm/loongson/lsdc_pll.h | 76 ++++ drivers/gpu/drm/loongson/lsdc_probe.c | 86 ++++ drivers/gpu/drm/loongson/lsdc_probe.h | 11 + drivers/gpu/drm/loongson/lsdc_regs.h | 370 ++++++++++++++++ drivers/gpu/drm/loongson/lsdc_ttm.c | 438 ++++++++++++++++++ drivers/gpu/drm/loongson/lsdc_ttm.h | 72 +++ 24 files changed, 4524 insertions(+) create mode 100644 drivers/gpu/drm/loongson/Kconfig create mode 100644 drivers/gpu/drm/loongson/Makefile create mode 100644 drivers/gpu/drm/loongson/lsdc_crtc.c create mode 100644 drivers/gpu/drm/loongson/lsdc_debugfs.c create mode 100644 drivers/gpu/drm/loongson/lsdc_drv.c create mode 100644 drivers/gpu/drm/loongson/lsdc_drv.h create mode 100644 drivers/gpu/drm/loongson/lsdc_gem.c create mode 100644 drivers/gpu/drm/loongson/lsdc_gem.h create mode 100644 drivers/gpu/drm/loongson/lsdc_i2c.c create mode 100644 drivers/gpu/drm/loongson/lsdc_irq.c create mode 100644 drivers/gpu/drm/loongson/lsdc_irq.h create mode 100644 drivers/gpu/drm/loongson/lsdc_output.c create mode 100644 drivers/gpu/drm/loongson/lsdc_output.h create mode 100644 drivers/gpu/drm/loongson/lsdc_plane.c create mode 100644 drivers/gpu/drm/loongson/lsdc_pll.c create mode 100644 drivers/gpu/drm/loongson/lsdc_pll.h create mode 100644 drivers/gpu/drm/loongson/lsdc_probe.c create mode 100644 drivers/gpu/drm/loongson/lsdc_probe.h create mode 100644 drivers/gpu/drm/loongson/lsdc_regs.h create mode 100644 drivers/gpu/drm/loongson/lsdc_ttm.c create mode 100644 drivers/gpu/drm/loongson/lsdc_ttm.h -- 2.25.1

1 year, 8 months

[PATCH v1 0/3] Add ioctls to map grant refs on the external backing storage

by Oleksii Moisieiev

Hello, Let me introduce the new ioctls, which are intended to allow gntdev to map scatter-gather table on top of the existing dmabuf, referenced by file descriptor. When using dma-buf exporter to create dma-buf with backing storage and map it to the grant refs, provided from the domain, we've met a problem, that several HW (i.MX8 gpu in our case) do not support external buffer and requires backing storage to be created using it's native tools. That's why new ioctls were added to be able to pass existing dma-buffer fd as input parameter and use it as backing storage to export to refs. Following calls were added: IOCTL_GNTDEV_DMABUF_MAP_REFS_TO_BUF - map existing buffer as the backing storage and export it to the provided grant refs; IOCTL_GNTDEV_DMABUF_MAP_RELEASE - detach buffer from the grant table and set notification to unmap grant refs before releasing the external buffer. After this call the external buffer should be detroyed. IOCTL_GNTDEV_DMABUF_MAP_WAIT_RELEASED - wait for timeout until buffer is completely destroyed and gnt refs unmapped so domain could free grant pages. Should be called after buffer was destoyed. Our setup is based on IMX8QM board. We're trying to implement zero-copy support for DomU graphics using Wayland zwp_linux_dmabuf_v1_interface implementation. For dma-buf exporter we used i.MX8 gpu native tools to create backing storage grant-refs, received from DomU. Buffer for the backing storage was allocated using gbm_bo_create call because gpu do not support external buffer and requires backing storage to be created using it's native tools (eglCreateImageKHR returns EGL_NO_IMAGE_KHR for buffers, which were not created using gbm_bo_create). This behaviour was also tested on Qemu setup using DRM_IOCTL_MODE_CREATE_DUMB call to create backing storage buffer. --- Oleksii Moisieiev (3): xen/grant-table: save page_count on map and use if during async unmapping dma-buf: add dma buffer release notifier callback xen/grant-table: add new ioctls to map dmabuf to existing fd drivers/dma-buf/dma-buf.c | 44 ++++ drivers/xen/gntdev-common.h | 8 +- drivers/xen/gntdev-dmabuf.c | 416 +++++++++++++++++++++++++++++++++++- drivers/xen/gntdev-dmabuf.h | 7 + drivers/xen/gntdev.c | 101 ++++++++- drivers/xen/grant-table.c | 73 +++++-- include/linux/dma-buf.h | 15 ++ include/uapi/xen/gntdev.h | 62 ++++++ include/xen/grant_table.h | 8 + 9 files changed, 703 insertions(+), 31 deletions(-) -- 2.25.1

1 year, 8 months

[PATCH 2/2] drm/amdgpu: Fix integer overflow in amdgpu_cs_pass1

by hackyzh002

The type of size is unsigned int, if size is 0x40000000, there will be an integer overflow, size will be zero after size *= sizeof(uint32_t), will cause uninitialized memory to be referenced later. Signed-off-by: hackyzh002 <hackyzh002(a)gmail.com> --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 08eced097..c17b3af85 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -192,7 +192,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, uint64_t *chunk_array_user; uint64_t *chunk_array; uint32_t uf_offset = 0; - unsigned int size; + uint64_t int size; int ret; int i; @@ -235,7 +235,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, size = p->chunks[i].length_dw; cdata = u64_to_user_ptr(user_chunk.chunk_data); - p->chunks[i].kdata = kvmalloc_array(size, sizeof(uint32_t), + p->chunks[i].kdata = kvcalloc(size, sizeof(uint32_t), GFP_KERNEL); if (p->chunks[i].kdata == NULL) { ret = -ENOMEM; -- 2.34.1

1 year, 8 months

[PATCH v2 1/2] drm/radeon: Fix integer overflow in radeon_cs_parser_init

by hackyzh002

The type of size is unsigned, if size is 0x40000000, there will be an integer overflow, size will be zero after size *= sizeof(uint32_t), will cause uninitialized memory to be referenced later Signed-off-by: hackyzh002 <hackyzh002(a)gmail.com> --- drivers/gpu/drm/radeon/radeon_cs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 46a27ebf4..8e12b406e 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -270,7 +270,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) { struct drm_radeon_cs *cs = data; uint64_t *chunk_array_ptr; - unsigned size, i; + u64 size, i; u32 ring = RADEON_CS_RING_GFX; s32 priority = 0; -- 2.34.1

1 year, 8 months

[PATCH V2 2/2] drm/amdgpu: Fix integer overflow in amdgpu_cs_pass1

by hackyzh002

The type of size is unsigned int, if size is 0x40000000, there will be an integer overflow, size will be zero after size *= sizeof(uint32_t), will cause uninitialized memory to be referenced later. Signed-off-by: hackyzh002 <hackyzh002(a)gmail.com> --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 08eced097..c17b3af85 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -192,7 +192,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, uint64_t *chunk_array_user; uint64_t *chunk_array; uint32_t uf_offset = 0; - unsigned int size; + size_t size; int ret; int i; @@ -235,7 +235,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, size = p->chunks[i].length_dw; cdata = u64_to_user_ptr(user_chunk.chunk_data); - p->chunks[i].kdata = kvmalloc_array(size, sizeof(uint32_t), + p->chunks[i].kdata = kvcalloc(size, sizeof(uint32_t), GFP_KERNEL); if (p->chunks[i].kdata == NULL) { ret = -ENOMEM; -- 2.34.1

1 year, 8 months

[PATCH 1/2] drm/radeon: Fix integer overflow in radeon_cs_parser_init

by hackyzh002

The type of size is unsigned, if size is 0x40000000, there will be an integer overflow, size will be zero after size *= sizeof(uint32_t), will cause uninitialized memory to be referenced later Signed-off-by: hackyzh002 <hackyzh002(a)gmail.com> --- drivers/gpu/drm/radeon/radeon_cs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 46a27ebf4..472c29050 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -270,7 +270,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) { struct drm_radeon_cs *cs = data; uint64_t *chunk_array_ptr; - unsigned size, i; + u64 size, i; u32 ring = RADEON_CS_RING_GFX; s32 priority = 0; @@ -347,7 +347,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) continue; } - p->chunks[i].kdata = kvmalloc_array(size, sizeof(uint32_t), GFP_KERNEL); + p->chunks[i].kdata = kvcalloc(size, sizeof(uint32_t), GFP_KERNEL); size *= sizeof(uint32_t); if (p->chunks[i].kdata == NULL) { return -ENOMEM; -- 2.34.1

1 year, 8 months

[PATCH RFC 00/18] Rust DRM subsystem abstractions (& preview AGX driver)

by Asahi Lina

Hi everyone! This is my first take on the Rust abstractions for the DRM subsystem. It includes the abstractions themselves, some minor prerequisite changes to the C side, as well as the drm-asahi GPU driver (for reference on how the abstractions are used, but not necessarily intended to land together). These patches apply on top of the tree at [1], which is based on 6.3-rc1 with a large number of Rust abstraction/support commits added on top. Most of these are not prerequisites for the DRM abstractions themselves, but rather only of the driver. * #1-12 introduce the abstractions, module by module, with minor C changes before the dependent abstraction. * Patch 10 is a little addition to drm_sched that I ended up needing, but I can pull it out of the abstraction into its own patch if needed. * #13-14 add a minor feature to drm/gem and its abstraction used by the driver. * #15-16 introduce the (unstable) asahi UAPI. This is obviously not ready for merge yet, but comments are welcome! * #17 adds a Rust helper macro to handle GPU core/firmware differences. This probably belongs in the driver at this point, but right now it has to live in rust/macros since there is no mechanism for per-driver proc macros. * #18 adds the driver proper, in one big commit, for reference purposes. I've been working since mid last year on an Apple AGX GPU driver for Linux, using the (at the time) out-of-tree Rust support. As part of this effort, I've been writing safe Rust abstractions for portions of the DRM subsystem. Now that Rust itself is upstream, I'd like to get all the abstractions upstreamed so we can eventually get the driver upstreamed! These abstractions have been used by the driver since our release in December [2], in a simpler synchronous-submission form: * drm::ioctl * drm::device * drm::drv * drm::file * drm::{gem, gem::shmem} * drm::mm This series adds these too, which are used by the explicit sync refactor of the driver (the version in this series): * drm::syncobj * drm::sched * dma_fence The major dependencies for the DRM abstractions themselves are: * [3] rust: error: Add missing wrappers to convert to/from kernel error codes * [4] rust: Miscellaneous macro improvements * [5] rust: Add a Sealed trait * [6] rust: device: Add a minimal RawDevice trait * [7] rust: Enable the new_uninit feature for kernel and driver crates * [8] rust: ioctl: Add ioctl number manipulation functions * [9] rust: sync: Arc: Any downcasting and assume_init() * rust: Add `container_of` and `offset_of` macros * kernel::sync::mutex and dependencies Most of these (the ones with links) have already been submitted, and I expect all of them to land for 6.4 (the mutex one will likely be last, since there is some refactoring that will happen over the current state to make it more ergonomic to use). The mutex dep is only necessary for drm::mm and dma_fence, and transitively drm::syncobj and drm::sched. Things work! We've had most of the abstractions in production edge kernels with the driver, and the new explicit sync stuff has passed quite a few torture tests (this is how we found the drm_sched issue, patch 11). The abstractions are intended to be safe (safety review very welcome!). While writing them, I tried to avoid making any changes to the C side unless absolutely necessary. I understand that it will probably make sense to adjust the C side to make some things easier, but I wanted to start from this as a baseline. Known issues: - The existing Rust integration does not currently allow building abstractions as modules, so the Rust abstractions are only available for DRM components that are built in. I added some extra Kconfig symbols to deal with this, so a driver built as a module can depende on having those built in. This should go away in the future (but may not be ready in time for submission... I understand this probably shouldn't be a blocker though?). - DRM relies heavily on the "subclassing" pattern for driver objects, and this doesn't map well to Rust. I tried several approaches for various bits, so we can see how they work out. In particular, whether wrapper types should pretend to be smart pointers and Deref to their inner driver-specific types, and whether they should be marked as method receivers (Yuck, internal rustc implementation hacks! But Arc<T> already does the same thing and it makes usage in driver-implemented callbacks as `self` possible) are things I'd love to discuss ^^. - Only what I need for my driver is implemented (plus a small amount of obvious extras where better API completeness makes sense). I think the general idea with Rust abstractions is that we add things as they become necessary. - The plain GEM vs. GEM-shmem duality ended up with quite a hairy type hierarchy. I'd love to figure out how to make this simpler... - drm::mm ends up requiring a built-in mutex in the abstraction, instead of delegating that to the user with the usual Rust mutability rules. This is because nodes can be dropped at any time, and those operations need to be synchronized. We could try to avoid forbidding those drops or mark the node type !Send, but that would make it a lot less ergonomic to use... I'm looking for feedback on the abstractions of all kinds, so we can move towards an upstreamable version. Optimistically, I'd love to get this upstream for 6.5, and the driver for 6.6. Please feel free to ask any questions about the Rust bits, since I know a lot of this is new to many of the C folks! This is a fairly complete driver for Apple AGX G13 and G14 series GPUs. The driver today supports the Apple M1, M1 Pro, M1 Max, M1 Ultra, and M2 SoCs, across two firmware revisions each. It has an explicit sync UAPI heavily inspired by the upcoming Intel Xe UAPI, designed with Vulkan support in mind. On the Mesa side we currently have a Gallium driver that is mostly already upstream (missing the UAPI bits mostly) and passes the dEQP GLES2/EGL tests, with most of GLES3.0 passing in downstream work-in-progress branches. This is a reverse engineered community driver (we have no hardware documentation of any kind, other than some hints from aspects shared with PowerVR). While developing the driver, I tried to make use of Rust's safety and lifetime features to provide not just CPU-side safety, but also partial firmware-ABI safety. Thanks to this, it has turned out to be a very stable driver even though GPU firmware crashes are fatal (no restart capability, need to reboot!) and the FW/driver interface is a huge mess of unsafe shared memory structures with complex pointer chains. There are over 70 ABI types and 3000+ lines of firmware ABI type definitions that vary between firmware builds and GPU cores... In a simpler blocking-submission form, it has been shipping in Asahi Linux edge kernels since December [2], with lots of users and zero (!) reported oopses (and only a couple reports of GPU firmware crashes, though that issue should now be fixed). It has survived OOM scenarios (Rust makes error cleanup easy!), UAPI-level fuzzing, countless broken Mesa builds, uptimes of 40+ days, and more. The explicit sync refactor significantly increases performance (and potential problems), but this version has survived a lot of torture with dEQP/piglit tests and some manual corner case testing. In other words, Rust works! ^^ There are some design notes on the driver and further links at [10]. [1] https://github.com/AsahiLinux/linux.git drm-rfc-base-20230307 [2] https://asahilinux.org/2022/12/gpu-drivers-now-in-asahi-linux/ [3] https://lore.kernel.org/rust-for-linux/20230224-rust-error-v1-0-f8f9a9a8730… [4] https://lore.kernel.org/rust-for-linux/20230224-rust-macros-v1-0-b39fae46e1… [5] https://lore.kernel.org/rust-for-linux/20230224-rust-iopt-rtkit-v1-0-49ced3… [6] https://lore.kernel.org/rust-for-linux/20230224-rust-iopt-rtkit-v1-0-49ced3… [7] https://lore.kernel.org/rust-for-linux/CQV7ZNT6LMXI.1XG4YXSH8I7JK@vincent-a… [8] https://lore.kernel.org/rust-for-linux/61f734d6-1497-755f-3632-3f261b890846… [9] https://lore.kernel.org/rust-for-linux/20230224-rust-arc-v1-0-568eea613a41@… [10] https://github.com/AsahiLinux/docs/wiki/SW:AGX-driver-notes Signed-off-by: Asahi Lina <lina(a)asahilina.net> --- Asahi Lina (18): rust: drm: ioctl: Add DRM ioctl abstraction rust: drm: Add Device and Driver abstractions rust: drm: file: Add File abstraction rust: drm: gem: Add GEM object abstraction drm/gem-shmem: Export VM ops functions rust: drm: gem: shmem: Add DRM shmem helper abstraction rust: drm: mm: Add DRM MM Range Allocator abstraction rust: dma_fence: Add DMA Fence abstraction rust: drm: syncobj: Add DRM Sync Object abstraction drm/scheduler: Add can_run_job callback drm/scheduler: Clean up jobs when the scheduler is torn down rust: drm: sched: Add GPU scheduler abstraction drm/gem: Add a flag to control whether objects can be exported rust: drm: gem: Add set_exportable() method drm/asahi: Add the Asahi driver UAPI [DO NOT MERGE] rust: bindings: Bind the Asahi DRM UAPI rust: macros: Add versions macro drm/asahi: Add the Asahi driver for Apple AGX GPUs drivers/gpu/drm/Kconfig | 19 + drivers/gpu/drm/Makefile | 1 + drivers/gpu/drm/asahi/Kconfig | 35 + drivers/gpu/drm/asahi/Makefile | 3 + drivers/gpu/drm/asahi/alloc.rs | 1046 ++++++++++++++++++++++++++ drivers/gpu/drm/asahi/asahi.rs | 53 ++ drivers/gpu/drm/asahi/buffer.rs | 694 ++++++++++++++++++ drivers/gpu/drm/asahi/channel.rs | 542 ++++++++++++++ drivers/gpu/drm/asahi/debug.rs | 129 ++++ drivers/gpu/drm/asahi/driver.rs | 166 +++++ drivers/gpu/drm/asahi/event.rs | 229 ++++++ drivers/gpu/drm/asahi/file.rs | 718 ++++++++++++++++++ drivers/gpu/drm/asahi/float.rs | 381 ++++++++++ drivers/gpu/drm/asahi/fw/buffer.rs | 170 +++++ drivers/gpu/drm/asahi/fw/channels.rs | 385 ++++++++++ drivers/gpu/drm/asahi/fw/compute.rs | 107 +++ drivers/gpu/drm/asahi/fw/event.rs | 100 +++ drivers/gpu/drm/asahi/fw/fragment.rs | 276 +++++++ drivers/gpu/drm/asahi/fw/initdata.rs | 1264 ++++++++++++++++++++++++++++++++ drivers/gpu/drm/asahi/fw/job.rs | 56 ++ drivers/gpu/drm/asahi/fw/microseq.rs | 384 ++++++++++ drivers/gpu/drm/asahi/fw/mod.rs | 15 + drivers/gpu/drm/asahi/fw/types.rs | 233 ++++++ drivers/gpu/drm/asahi/fw/vertex.rs | 177 +++++ drivers/gpu/drm/asahi/fw/workqueue.rs | 168 +++++ drivers/gpu/drm/asahi/gem.rs | 301 ++++++++ drivers/gpu/drm/asahi/gpu.rs | 1088 +++++++++++++++++++++++++++ drivers/gpu/drm/asahi/hw/mod.rs | 522 +++++++++++++ drivers/gpu/drm/asahi/hw/t600x.rs | 140 ++++ drivers/gpu/drm/asahi/hw/t8103.rs | 80 ++ drivers/gpu/drm/asahi/hw/t8112.rs | 82 +++ drivers/gpu/drm/asahi/initdata.rs | 777 ++++++++++++++++++++ drivers/gpu/drm/asahi/mem.rs | 133 ++++ drivers/gpu/drm/asahi/microseq.rs | 61 ++ drivers/gpu/drm/asahi/mmu.rs | 1249 +++++++++++++++++++++++++++++++ drivers/gpu/drm/asahi/object.rs | 704 ++++++++++++++++++ drivers/gpu/drm/asahi/place.rs | 343 +++++++++ drivers/gpu/drm/asahi/queue/common.rs | 52 ++ drivers/gpu/drm/asahi/queue/compute.rs | 371 ++++++++++ drivers/gpu/drm/asahi/queue/mod.rs | 725 ++++++++++++++++++ drivers/gpu/drm/asahi/queue/render.rs | 1173 +++++++++++++++++++++++++++++ drivers/gpu/drm/asahi/regs.rs | 387 ++++++++++ drivers/gpu/drm/asahi/slotalloc.rs | 292 ++++++++ drivers/gpu/drm/asahi/util.rs | 44 ++ drivers/gpu/drm/asahi/workqueue.rs | 880 ++++++++++++++++++++++ drivers/gpu/drm/drm_gem.c | 1 + drivers/gpu/drm/drm_gem_shmem_helper.c | 9 +- drivers/gpu/drm/drm_prime.c | 5 + drivers/gpu/drm/scheduler/sched_main.c | 37 +- include/drm/drm_gem.h | 8 + include/drm/drm_gem_shmem_helper.h | 3 + include/drm/gpu_scheduler.h | 8 + include/uapi/drm/asahi_drm.h | 556 ++++++++++++++ rust/bindings/bindings_helper.h | 14 + rust/helpers.c | 168 +++++ rust/kernel/dma_fence.rs | 532 ++++++++++++++ rust/kernel/drm/device.rs | 76 ++ rust/kernel/drm/drv.rs | 342 +++++++++ rust/kernel/drm/file.rs | 113 +++ rust/kernel/drm/gem/mod.rs | 384 ++++++++++ rust/kernel/drm/gem/shmem.rs | 381 ++++++++++ rust/kernel/drm/ioctl.rs | 147 ++++ rust/kernel/drm/mm.rs | 309 ++++++++ rust/kernel/drm/mod.rs | 13 + rust/kernel/drm/sched.rs | 358 +++++++++ rust/kernel/drm/syncobj.rs | 77 ++ rust/kernel/lib.rs | 4 + rust/macros/lib.rs | 7 + rust/macros/versions.rs | 267 +++++++ 69 files changed, 20569 insertions(+), 5 deletions(-) --- base-commit: c9eb15274c9861026682a6b3e645891fccf88e07 change-id: 20230307-rust-drm-b5af3c2a9e55 Thank you, ~~ Lina

1 year, 8 months

[PATCH] dma-buf/dma-resv.c: fix a typo

by Sui Jingfeng

The dma_resv_wait_timeout() function return a value greater than zero on success. Signed-off-by: Sui Jingfeng <suijingfeng(a)loongson.cn> --- drivers/dma-buf/dma-resv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c index 2a594b754af1..b6f71eb00866 100644 --- a/drivers/dma-buf/dma-resv.c +++ b/drivers/dma-buf/dma-resv.c @@ -660,7 +660,7 @@ EXPORT_SYMBOL_GPL(dma_resv_get_singleton); * dma_resv_lock() already * RETURNS * Returns -ERESTARTSYS if interrupted, 0 if the wait timed out, or - * greater than zer on success. + * greater than zero on success. */ long dma_resv_wait_timeout(struct dma_resv *obj, enum dma_resv_usage usage, bool intr, unsigned long timeout) -- 2.25.1

1 year, 8 months

Re: [PATCH] misc: sram: Add dma-heap-export reserved SRAM area type

by Christian Gmeiner

Hi Andrew > > > Okay, will split for v2. > > Was there a follow-up v2 of this patchset? AFAICT this series did not make it into the mainline kernel. Do you have any plans to work on it? If not I would like to help out as we have a use case where we want to use a dma-buf sram exporter. -- greets -- Christian Gmeiner, MSc https://christian-gmeiner.info/privacypolicy

1 year, 9 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

Linaro-mm-sig