On Thu, Jul 10, 2014 at 8:33 AM, Christian König
<deathsimple@xxxxxxxxxxx> wrote:
> From: Christian König <christian.koenig@xxxxxxx>
>
> v2: use flag instead of boolean
>
> Signed-off-by: Christian König <christian.koenig@xxxxxxx>
> ---
> drivers/gpu/drm/radeon/r100.c | 2 +-
> drivers/gpu/drm/radeon/r300.c | 8 ++++++--
> drivers/gpu/drm/radeon/radeon.h | 12 +++++++-----
> drivers/gpu/drm/radeon/radeon_asic.h | 8 ++++----
> drivers/gpu/drm/radeon/radeon_gart.c | 9 +++++----
> drivers/gpu/drm/radeon/radeon_ttm.c | 4 ++--
> drivers/gpu/drm/radeon/rs400.c | 9 +++++++--
> drivers/gpu/drm/radeon/rs600.c | 8 ++++++--
> 8 files changed, 38 insertions(+), 22 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
> index a6f7d61..fa71121d7 100644
> --- a/drivers/gpu/drm/radeon/r100.c
> +++ b/drivers/gpu/drm/radeon/r100.c
> @@ -650,7 +650,7 @@ void r100_pci_gart_disable(struct radeon_device *rdev)
> }
>
> void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
> - uint64_t addr)
> + uint64_t addr, uint32_t flags)
> {
> u32 *gtt = rdev->gart.ptr;
> gtt[i] = cpu_to_le32(lower_32_bits(addr));
> diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
> index 8d14e66..b947f42 100644
> --- a/drivers/gpu/drm/radeon/r300.c
> +++ b/drivers/gpu/drm/radeon/r300.c
> @@ -73,13 +73,17 @@ void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev)
> #define R300_PTE_READABLE (1 << 3)
>
> void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
> - uint64_t addr)
> + uint64_t addr, uint32_t flags)
> {
> void __iomem *ptr = rdev->gart.ptr;
>
> addr = (lower_32_bits(addr) >> 8) |
> ((upper_32_bits(addr) & 0xff) << 24) |
> - R300_PTE_WRITEABLE | R300_PTE_READABLE;
> + R300_PTE_READABLE;
> +
> + if (!(flags & RADEON_GART_PAGE_READONLY))
> + addr |= R300_PTE_WRITEABLE;
> +
> /* on x86 we want this to be CPU endian, on powerpc
> * on powerpc without HW swappers, it'll get swapped on way
> * into VRAM - so no need for cpu_to_le32 on VRAM tables */
> diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
> index c5318cac..86fdfa30 100644
> --- a/drivers/gpu/drm/radeon/radeon.h
> +++ b/drivers/gpu/drm/radeon/radeon.h
> @@ -592,6 +592,8 @@ struct radeon_mc;
> #define RADEON_GPU_PAGE_SHIFT 12
> #define RADEON_GPU_PAGE_ALIGN(a) (((a) + RADEON_GPU_PAGE_MASK) & ~RADEON_GPU_PAGE_MASK)
>
> +#define RADEON_GART_PAGE_READONLY 1
> +
> struct radeon_gart {
> dma_addr_t table_addr;
> struct radeon_bo *robj;
> @@ -616,7 +618,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
> int pages);
> int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
> int pages, struct page **pagelist,
> - dma_addr_t *dma_addr);
> + dma_addr_t *dma_addr, uint32_t flags);
>
>
> /*
> @@ -854,8 +856,8 @@ struct radeon_mec {
> #define R600_PTE_FRAG_256KB (6 << 7)
>
> /* flags used for GART page table entries on R600+ */
> -#define R600_PTE_GART ( R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED \
> - | R600_PTE_READABLE | R600_PTE_WRITEABLE)
> +#define R600_PTE_GART ( R600_PTE_VALID | R600_PTE_SYSTEM | \
> + R600_PTE_SNOOPED | R600_PTE_READABLE )
>
This will disable the gart/GPUVM update optimization in most cases.
Might be better to just leave this define as is and use the individual
flags in the set_page() functions. I realize this is just a proof of
concept. Just pointing it out for future reference.
Alex
> struct radeon_vm_pt {
> struct radeon_bo *bo;
> @@ -1775,7 +1777,7 @@ struct radeon_asic {
> struct {
> void (*tlb_flush)(struct radeon_device *rdev);
> void (*set_page)(struct radeon_device *rdev, unsigned i,
> - uint64_t addr);
> + uint64_t addr, uint32_t flags);
> } gart;
> struct {
> int (*init)(struct radeon_device *rdev);
> @@ -2703,7 +2705,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
> #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state))
> #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev))
> #define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev))
> -#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p))
> +#define radeon_gart_set_page(rdev, i, p, r) (rdev)->asic->gart.set_page((rdev), (i), (p), (r))
> #define radeon_asic_vm_init(rdev) (rdev)->asic->vm.init((rdev))
> #define radeon_asic_vm_fini(rdev) (rdev)->asic->vm.fini((rdev))
> #define radeon_asic_vm_set_page(rdev, ib, pe, addr, count, incr, flags) ((rdev)->asic->vm.set_page((rdev), (ib), (pe), (addr), (count), (incr), (flags)))
> diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
> index 7531b5e..f7d7c33 100644
> --- a/drivers/gpu/drm/radeon/radeon_asic.h
> +++ b/drivers/gpu/drm/radeon/radeon_asic.h
> @@ -68,7 +68,7 @@ int r100_asic_reset(struct radeon_device *rdev);
> u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc);
> void r100_pci_gart_tlb_flush(struct radeon_device *rdev);
> void r100_pci_gart_set_page(struct radeon_device *rdev, unsigned i,
> - uint64_t addr);
> + uint64_t addr, uint32_t flags);
> void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
> int r100_irq_set(struct radeon_device *rdev);
> int r100_irq_process(struct radeon_device *rdev);
> @@ -172,7 +172,7 @@ extern void r300_fence_ring_emit(struct radeon_device *rdev,
> extern int r300_cs_parse(struct radeon_cs_parser *p);
> extern void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev);
> extern void rv370_pcie_gart_set_page(struct radeon_device *rdev, unsigned i,
> - uint64_t addr);
> + uint64_t addr, uint32_t flags);
> extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes);
> extern int rv370_get_pcie_lanes(struct radeon_device *rdev);
> extern void r300_set_reg_safe(struct radeon_device *rdev);
> @@ -208,7 +208,7 @@ extern int rs400_suspend(struct radeon_device *rdev);
> extern int rs400_resume(struct radeon_device *rdev);
> void rs400_gart_tlb_flush(struct radeon_device *rdev);
> void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
> - uint64_t addr);
> + uint64_t addr, uint32_t flags);
> uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg);
> void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
> int rs400_gart_init(struct radeon_device *rdev);
> @@ -232,7 +232,7 @@ void rs600_irq_disable(struct radeon_device *rdev);
> u32 rs600_get_vblank_counter(struct radeon_device *rdev, int crtc);
> void rs600_gart_tlb_flush(struct radeon_device *rdev);
> void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
> - uint64_t addr);
> + uint64_t addr, uint32_t flags);
> uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg);
> void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
> void rs600_bandwidth_update(struct radeon_device *rdev);
> diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
> index b7d3e84..b26c0b2 100644
> --- a/drivers/gpu/drm/radeon/radeon_gart.c
> +++ b/drivers/gpu/drm/radeon/radeon_gart.c
> @@ -243,7 +243,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
> page_base = rdev->gart.pages_addr[p];
> for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
> if (rdev->gart.ptr) {
> - radeon_gart_set_page(rdev, t, page_base);
> + radeon_gart_set_page(rdev, t, page_base, 0);
> }
> page_base += RADEON_GPU_PAGE_SIZE;
> }
> @@ -266,8 +266,9 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
> * (all asics).
> * Returns 0 for success, -EINVAL for failure.
> */
> -int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
> - int pages, struct page **pagelist, dma_addr_t *dma_addr)
> +int radeon_gart_bind(struct radeon_device *rdev, unsigned offset, int pages,
> + struct page **pagelist, dma_addr_t *dma_addr,
> + uint32_t flags)
> {
> unsigned t;
> unsigned p;
> @@ -287,7 +288,7 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
> if (rdev->gart.ptr) {
> page_base = rdev->gart.pages_addr[p];
> for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
> - radeon_gart_set_page(rdev, t, page_base);
> + radeon_gart_set_page(rdev, t, page_base, flags);
> page_base += RADEON_GPU_PAGE_SIZE;
> }
> }
> diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
> index c8a8a51..19d662f 100644
> --- a/drivers/gpu/drm/radeon/radeon_ttm.c
> +++ b/drivers/gpu/drm/radeon/radeon_ttm.c
> @@ -528,8 +528,8 @@ static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
> WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
> ttm->num_pages, bo_mem, ttm);
> }
> - r = radeon_gart_bind(gtt->rdev, gtt->offset,
> - ttm->num_pages, ttm->pages, gtt->ttm.dma_address);
> + r = radeon_gart_bind(gtt->rdev, gtt->offset, ttm->num_pages,
> + ttm->pages, gtt->ttm.dma_address, 0);
> if (r) {
> DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
> ttm->num_pages, (unsigned)gtt->offset);
> diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
> index 4519f9c..5164544 100644
> --- a/drivers/gpu/drm/radeon/rs400.c
> +++ b/drivers/gpu/drm/radeon/rs400.c
> @@ -211,14 +211,19 @@ void rs400_gart_fini(struct radeon_device *rdev)
> #define RS400_PTE_WRITEABLE (1 << 2)
> #define RS400_PTE_READABLE (1 << 3)
>
> -void rs400_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr)
> +void rs400_gart_set_page(struct radeon_device *rdev, unsigned i,
> + uint64_t addr, uint32_t flags)
> {
> uint32_t entry;
> u32 *gtt = rdev->gart.ptr;
>
> entry = (lower_32_bits(addr) & PAGE_MASK) |
> ((upper_32_bits(addr) & 0xff) << 4) |
> - RS400_PTE_WRITEABLE | RS400_PTE_READABLE;
> + RS400_PTE_READABLE;
> +
> + if (!(flags & RADEON_GART_PAGE_READONLY))
> + entry |= RS400_PTE_WRITEABLE;
> +
> entry = cpu_to_le32(entry);
> gtt[i] = entry;
> }
> diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
> index cd7b4b0..3cba999 100644
> --- a/drivers/gpu/drm/radeon/rs600.c
> +++ b/drivers/gpu/drm/radeon/rs600.c
> @@ -611,15 +611,19 @@ static void rs600_gart_fini(struct radeon_device *rdev)
> radeon_gart_table_vram_free(rdev);
> }
>
> -void rs600_gart_set_page(struct radeon_device *rdev, unsigned i, uint64_t addr)
> +void rs600_gart_set_page(struct radeon_device *rdev, unsigned i,
> + uint64_t addr, uint32_t flags)
> {
> void __iomem *ptr = (void *)rdev->gart.ptr;
>
> addr = addr & 0xFFFFFFFFFFFFF000ULL;
> if (addr == rdev->dummy_page.addr)
> addr |= R600_PTE_SYSTEM | R600_PTE_SNOOPED;
> - else
> + else {
> addr |= R600_PTE_GART;
> + if (!(flags & RADEON_GART_PAGE_READONLY))
> + addr |= R600_PTE_WRITEABLE;
> + }
> writeq(addr, ptr + (i * 8));
> }
>
> --
> 1.9.1
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@xxxxxxxxxxxxxxxxxxxxx
> http://lists.freedesktop.org/mailman/listinfo/dri-devel
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/dri-devel
