diff options
| author | Simona Vetter <simona.vetter@ffwll.ch> | 2025-10-31 19:07:39 +0100 |
|---|---|---|
| committer | Simona Vetter <simona.vetter@ffwll.ch> | 2025-10-31 19:07:39 +0100 |
| commit | 3d8d35bf8dc8893e34d92212d413d7cfd89c560e (patch) | |
| tree | 0fbf0710a112181778f2f540fe2a397e7f8eb91c /drivers/gpu/drm/msm/msm_gem_vma.c | |
| parent | 3a9f6bd54e20d0eaf89ca3ec6d82e5263cac93e8 (diff) | |
| parent | f5d079564c44baaeedf5e25f4b943aa042ea0eb1 (diff) | |
Merge tag 'drm-msm-fixes-2025-10-29' of https://gitlab.freedesktop.org/drm/msm into drm-fixes
Fixes for v6.18-rc4
CI
- Disable broken sanity job
GEM
- Fix vm_bind prealloc error path
- Fix dma-buf import free
- Fix last-fence update
- Reject MAP_NULL if PRR is unsupported
- Ensure vm is created in VM_BIND ioctl
GPU
- GMU fw parsing fix
DPU:
- Fixed mode_valid callback
- Fixed planes on DPU 1.x devices.
Signed-off-by: Simona Vetter <simona.vetter@ffwll.ch>
From: Rob Clark <rob.clark@oss.qualcomm.com>
Link: https://patch.msgid.link/CACSVV03kUm1ms7FBg0m9U4ZcyickSWbnayAWqYqs0XH4UjWf+A@mail.gmail.com
Diffstat (limited to 'drivers/gpu/drm/msm/msm_gem_vma.c')
| -rw-r--r-- | drivers/gpu/drm/msm/msm_gem_vma.c | 8 |
1 files changed, 7 insertions, 1 deletions
diff --git a/drivers/gpu/drm/msm/msm_gem_vma.c b/drivers/gpu/drm/msm/msm_gem_vma.c index 8316af1723c2..89a95977f41e 100644 --- a/drivers/gpu/drm/msm/msm_gem_vma.c +++ b/drivers/gpu/drm/msm/msm_gem_vma.c @@ -971,6 +971,7 @@ static int lookup_op(struct msm_vm_bind_job *job, const struct drm_msm_vm_bind_op *op) { struct drm_device *dev = job->vm->drm; + struct msm_drm_private *priv = dev->dev_private; int i = job->nr_ops++; int ret = 0; @@ -1017,6 +1018,11 @@ lookup_op(struct msm_vm_bind_job *job, const struct drm_msm_vm_bind_op *op) break; } + if ((op->op == MSM_VM_BIND_OP_MAP_NULL) && + !adreno_smmu_has_prr(priv->gpu)) { + ret = UERR(EINVAL, dev, "PRR not supported\n"); + } + return ret; } @@ -1421,7 +1427,7 @@ msm_ioctl_vm_bind(struct drm_device *dev, void *data, struct drm_file *file) * Maybe we could allow just UNMAP ops? OTOH userspace should just * immediately close the device file and all will be torn down. */ - if (to_msm_vm(ctx->vm)->unusable) + if (to_msm_vm(msm_context_vm(dev, ctx))->unusable) return UERR(EPIPE, dev, "context is unusable"); /* |