search for: gk20a_instmem

...emory) @@ -68,7 +65,6 @@ struct gk20a_instobj { struct gk20a_instobj_dma { struct gk20a_instobj base; - u32 *cpuaddr; dma_addr_t handle; struct nvkm_mm_node r; }; @@ -81,6 +77,11 @@ struct gk20a_instobj_dma { struct gk20a_instobj_iommu { struct gk20a_instobj base; + /* to link into gk20a_instmem::vaddr_lru */ + struct list_head vaddr_node; + /* how many clients are using vaddr? */ + u32 use_cpt; + /* will point to the higher half of pages */ dma_addr_t *dma_addrs; /* array of base.mem->size pages (+ dma_addr_ts) */ @@ -109,8 +110,6 @@ struct gk20a_instmem { /* Only used by DMA...

...ct gk20a_instobj_dma { > struct gk20a_instobj base; > > - u32 *cpuaddr; > dma_addr_t handle; > struct nvkm_mm_node r; > }; > @@ -81,6 +77,11 @@ struct gk20a_instobj_dma { > struct gk20a_instobj_iommu { > struct gk20a_instobj base; > > + /* to link into gk20a_instmem::vaddr_lru */ > + struct list_head vaddr_node; > + /* how many clients are using vaddr? */ > + u32 use_cpt; > + > /* will point to the higher half of pages */ > dma_addr_t *dma_addrs; > /* array of base.mem->size pages (+ dma_addr_ts) */ > @@ -109,8 +110,6 @@ struct...

...diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index a6a7fa0d7679..7f5244d57d2f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -94,7 +94,7 @@ struct gk20a_instmem { struct nvkm_instmem base; /* protects vaddr_* and gk20a_instobj::vaddr* */ - spinlock_t lock; + struct mutex lock; /* CPU mappings LRU */ unsigned int vaddr_use; @@ -184,11 +184,10 @@ gk20a_instobj_acquire_iommu(struct nvkm_memory *memory) struct gk20a_instmem *imem = node->base....

.../nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c >> index a6a7fa0d7679..7f5244d57d2f 100644 >> --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c >> +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c >> @@ -94,7 +94,7 @@ struct gk20a_instmem { >> struct nvkm_instmem base; >> >> /* protects vaddr_* and gk20a_instobj::vaddr* */ >> - spinlock_t lock; >> + struct mutex lock; >> >> /* CPU mappings LRU */ >> unsigned int vaddr_use; >> @@ -184,11 +184,10 @@ gk20a_instobj_acquire_io...

...u32 *vaddr; struct list_head vaddr_node; + /* How many clients are using vaddr? */ + u32 use_cpt; }; #define gk20a_instobj(p) container_of((p), struct gk20a_instobj, memory) @@ -158,27 +160,35 @@ gk20a_instobj_cpu_map_iommu(struct nvkm_memory *memory) } /* - * Must be called while holding gk20a_instmem_lock + * Recycle the vaddr of obj. Must be called with gk20a_instmem::lock held. + */ +static void +gk20a_instobj_recycle_vaddr(struct gk20a_instobj *obj) +{ + struct gk20a_instmem *imem = obj->imem; + /* there should not be any user left... */ + WARN_ON(obj->use_cpt); + list_del(&obj-&gt...

...file changed, 6 insertions(+), 9 deletions(-) diff --git a/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drm/nouveau/nvkm/subdev/instmem/gk20a.c index fc419bb8eab7..d015633b8edd 100644 --- a/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -92,6 +92,7 @@ struct gk20a_instmem { /* protects vaddr_* and gk20a_instobj::vaddr* */ spinlock_t lock; + unsigned long flags; /* CPU mappings LRU */ unsigned int vaddr_use; @@ -188,12 +189,11 @@ gk20a_instobj_acquire(struct nvkm_memory *memory) struct gk20a_instobj *node = gk20a_instobj(memory); struct gk20a_instmem...

.../gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index a6a7fa0d7679..7f5244d57d2f 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -94,7 +94,7 @@ struct gk20a_instmem { > struct nvkm_instmem base; > > /* protects vaddr_* and gk20a_instobj::vaddr* */ > - spinlock_t lock; > + struct mutex lock; > > /* CPU mappings LRU */ > unsigned int vaddr_use; > @@ -184,11 +184,10 @@ gk20a_instobj_acquire_iommu(struct nvkm_memory *memory)...

.../* How many clients are using vaddr? */ > + u32 use_cpt; > }; > #define gk20a_instobj(p) container_of((p), struct gk20a_instobj, memory) > > @@ -158,27 +160,35 @@ gk20a_instobj_cpu_map_iommu(struct nvkm_memory *memory) > } > > /* > - * Must be called while holding gk20a_instmem_lock > + * Recycle the vaddr of obj. Must be called with gk20a_instmem::lock held. > + */ > +static void > +gk20a_instobj_recycle_vaddr(struct gk20a_instobj *obj) > +{ > + struct gk20a_instmem *imem = obj->imem; > + /* there should not be any user left... */ > + WARN_ON(o...

...ff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 6b8f2a19b2d9..a6a7fa0d7679 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -109,7 +109,7 @@ struct gk20a_instmem { u16 iommu_bit; /* Only used by DMA API */ - struct dma_attrs attrs; + unsigned long attrs; }; #define gk20a_instmem(p) container_of((p), struct gk20a_instmem, base) @@ -293,7 +293,7 @@ gk20a_instobj_dtor_dma(struct nvkm_memory *memory) goto out; dma_free_attrs(dev, node->base....

...ff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 6b8f2a19b2d9..a6a7fa0d7679 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -109,7 +109,7 @@ struct gk20a_instmem { u16 iommu_bit; /* Only used by DMA API */ - struct dma_attrs attrs; + unsigned long attrs; }; #define gk20a_instmem(p) container_of((p), struct gk20a_instmem, base) @@ -293,7 +293,7 @@ gk20a_instobj_dtor_dma(struct nvkm_memory *memory) goto out; dma_free_attrs(dev, node->base....

...ff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 6b8f2a19b2d9..a6a7fa0d7679 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -109,7 +109,7 @@ struct gk20a_instmem { u16 iommu_bit; /* Only used by DMA API */ - struct dma_attrs attrs; + unsigned long attrs; }; #define gk20a_instmem(p) container_of((p), struct gk20a_instmem, base) @@ -293,7 +293,7 @@ gk20a_instobj_dtor_dma(struct nvkm_memory *memory) goto out; dma_free_attrs(dev, node->base....

...ae9214 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -49,14 +49,14 @@ #include <subdev/mmu.h> struct gk20a_instobj { - struct nvkm_memory memory; + struct nvkm_instobj base; struct nvkm_mm_node *mn; struct gk20a_instmem *imem; /* CPU mapping */ u32 *vaddr; }; -#define gk20a_instobj(p) container_of((p), struct gk20a_instobj, memory) +#define gk20a_instobj(p) container_of((p), struct gk20a_instobj, base.memory) /* * Used for objects allocated using the DMA API @@ -148,7 +148,7 @@ gk20a_instobj_iommu_recy...

..., 9 deletions(-) > > diff --git a/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index fc419bb8eab7..d015633b8edd 100644 > --- a/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -92,6 +92,7 @@ struct gk20a_instmem { > > /* protects vaddr_* and gk20a_instobj::vaddr* */ > spinlock_t lock; > + unsigned long flags; > > /* CPU mappings LRU */ > unsigned int vaddr_use; > @@ -188,12 +189,11 @@ gk20a_instobj_acquire(struct nvkm_memory *memory) > struct gk20a_instobj *node = g...

...ubdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -49,14 +49,14 @@ > #include <subdev/mmu.h> > > struct gk20a_instobj { > - struct nvkm_memory memory; > + struct nvkm_instobj base; > struct nvkm_mm_node *mn; > struct gk20a_instmem *imem; > > /* CPU mapping */ > u32 *vaddr; > }; > -#define gk20a_instobj(p) container_of((p), struct gk20a_instobj, memory) > +#define gk20a_instobj(p) container_of((p), struct gk20a_instobj, base.memory) > > /* > * Used for objects allocated using the...

Hi, This is fourth approach for replacing struct dma_attrs with unsigned long. The main patch (1/44) doing the change is split into many subpatches for easier review (2-42). They should be squashed together when applying. *Important:* Patchset is tested on my ARM platforms and *only* build tested on allyesconfigs: ARM, ARM64, i386, x86_64 and powerpc. Please kindly provide reviewes and tests

These 4 patches fix two issues that existed on Tegra regarding DMA: 1) The bit indicating whether to use an IOMMU or not was hardcoded ; make this a platform property and use it in instmem 2) The DMA mask was not set for platform devices. Fix this by converting more pci_dma* to the DMA API, and use that more generic code to set the DMA mask properly for all platforms. Tested on both x86

.../gk20a.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 6b8f2a19b2d9..4c3c744c5c6b 100644 --- a/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -576,6 +576,7 @@ gk20a_instmem_new(struct nvkm_device *device, int index, { struct nvkm_device_tegra *tdev = device->func->tegra(device); struct gk20a_instmem *imem; + int ret; if (!(imem = kzalloc(sizeof(*imem), GFP_KERNEL))) return -ENOMEM; @@ -583,6 +584,10 @@ gk20a_instmem_new(struct nvkm_device *device, in...

...a/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 4c20fec64d96..6b8f2a19b2d9 100644 --- a/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -228,6 +228,8 @@ gk20a_instobj_release_dma(struct nvkm_memory *memory) struct gk20a_instmem *imem = node->imem; struct nvkm_ltc *ltc = imem->base.subdev.device->ltc; + /* in case we got a write-combined mapping */ + wmb(); nvkm_ltc_invalidate(ltc); } -- 2.7.2

...GE_SIZE); + iommu_unmap_sync(imem->domain, + (r->offset + i) << imem->iommu_pgshift, + PAGE_SIZE); dma_unmap_page(dev, node->dma_addrs[i], PAGE_SIZE, DMA_BIDIRECTIONAL); __free_page(node->pages[i]); @@ -458,14 +459,15 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, for (i = 0; i < npages; i++) { u32 offset = (r->offset + i) << imem->iommu_pgshift; - ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map_sync(imem->domain, offset, node...

From: Thierry Reding <treding at nvidia.com> Hi, the GPU on Jetson TX2 (GP10B) does not work properly on all devices. Why exactly is not clear, but there are slight differences between the SKUs that were tested. It turns out that the biggest issue is that on some devices (e.g. the one that I have), pulsing the GPU reset twice as is done in the current code (once as part of the power-ungate