kernel/linux.git/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c, branch v7.0-rc7

Convert more 'alloc_obj' cases to default GFP_KERNEL arguments

2026-02-22T04:03:00+00:00

This converts some of the visually simpler cases that have been split over multiple lines. I only did the ones that are easy to verify the resulting diff by having just that final GFP_KERNEL argument on the next line. Somebody should probably do a proper coccinelle script for this, but for me the trivial script actually resulted in an assertion failure in the middle of the script. I probably had made it a bit _too_ trivial. So after fighting that far a while I decided to just do some of the syntactically simpler cases with variations of the previous 'sed' scripts. The more syntactically complex multi-line cases would mostly really want whitespace cleanup anyway. Signed-off-by: Linus Torvalds

Convert 'alloc_obj' family to use the new default GFP_KERNEL argument

2026-02-22T01:09:51+00:00

This was done entirely with mindless brute force, using git grep -l '\

treewide: Replace kmalloc with kmalloc_obj for non-scalar types

2026-02-21T09:02:28+00:00

This is the result of running the Coccinelle script from scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to avoid scalar types (which need careful case-by-case checking), and instead replace kmalloc-family calls that allocate struct or union object instances: Single allocations: kmalloc(sizeof(TYPE), ...) are replaced with: kmalloc_obj(TYPE, ...) Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...) are replaced with: kmalloc_objs(TYPE, COUNT, ...) Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...) are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...) (where TYPE may also be *VAR) The resulting allocations no longer return "void *", instead returning "TYPE *". Signed-off-by: Kees Cook

drm/amdgpu: Skip loading SDMA_RS64 in VF

2026-01-10T19:06:33+00:00

VFs use the PF SDMA ucode and are unable to load SDMA_RS64. Signed-off-by: YuBiao Wang Signed-off-by: Victor Skvortsov Reviewed-by: Gavin Wan Signed-off-by: Alex Deucher

drm/amdgpu: RLC-G VF Register Access Interface

2026-01-05T21:59:57+00:00

- Implement Gfx v12.1 VFi interface under SRIOV - Redirect all RLCG interface access to new function after Gfx v12.1 v2: squash in register updates Signed-off-by: Bokun Zhang Signed-off-by: Alex Deucher

drm/amd/ras: Compatible with legacy sriov host

2025-12-08T18:56:34+00:00

If sriov host is legacy, the guest uniras will be disabled. Signed-off-by: YiPeng Chai Reviewed-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher

drm/amdgpu: Add virt command to send VF ras command

2025-12-08T18:56:33+00:00

Add virt command and interface to send VF ras command. Signed-off-by: YiPeng Chai Reviewed-by: Tao Zhou Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher

drm/amdgpu: Add lock to serialize sriov command execution

2025-11-14T16:28:07+00:00

Add lock to serialize sriov command execution. Signed-off-by: YiPeng Chai Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher

drm/amdgpu: Fix pointer casts when reading dynamic region sizes

2025-10-28T13:55:16+00:00

The function amdgpu_virt_get_dynamic_data_info() writes a 64-bit size value. In two places (amdgpu_bios.c and amdgpu_discovery.c), the code passed the address of a smaller variable by casting it to u64 *, which is unsafe. This could make the function write more bytes than the smaller variable can hold, possibly overwriting nearby memory. Reported by static analysis tools. v2: Dynamic region size comes from the host (SR-IOV setup) and is always fixed to 5 MB. (Lijo/Ellen) 5 MB easily fits inside a 32-bit value, so using a 64-bit type is not needed. It also avoids extra type casts Fixes: b4a8fcc7826a ("drm/amdgpu: Add logic for VF ipd and VF bios to init from dynamic crit_region offsets") Reported by: Dan Carpenter Cc: Ellen Pan Cc: Christian König Cc: Alex Deucher Signed-off-by: Srinivasan Shanmugam Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher

drm/amdgpu: Make SR-IOV critical region checks overflow-safe

2025-10-28T13:52:44+00:00

The function amdgpu_virt_init_critical_region() contained an invalid check for a negative init_hdr_offset value: if (init_hdr_offset < 0) Since init_hdr_offset is an unsigned 32-bit integer, this condition can never be true and triggers a Smatch warning: warn: unsigned 'init_hdr_offset' is never less than zero In addition, the subsequent bounds check: if ((init_hdr_offset + init_hdr_size) > vram_size) was vulnerable to integer overflow when adding the two unsigned values. Thus, by promoting offset and size to 64-bit and using check_add_overflow() to safely validate the sum against VRAM size. Fixes: 07009df6494d ("drm/amdgpu: Introduce SRIOV critical regions v2 during VF init") Reported by: Dan Carpenter Cc: Ellen Pan Cc: Christian König Cc: Alex Deucher Signed-off-by: Srinivasan Shanmugam Reviewed-by: Ellen Pan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher