drm/radeon: avoid deadlock in pm path when waiting for fence

[~andy/linux] / drivers / gpu / drm / radeon / radeon.h

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 8c42d54c2e26ad56ce7d0c7873128f466e60b801..9b9422c4403a40280e1a16af91e79b820fadad58 100644 (file)
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -109,7 +109,7 @@ extern int radeon_lockup_timeout;
 #define RADEON_BIOS_NUM_SCRATCH                        8
 
 /* max number of rings */
-#define RADEON_NUM_RINGS                       3
+#define RADEON_NUM_RINGS                       5
 
 /* fence seq are set to this number when signaled */
 #define RADEON_FENCE_SIGNALED_SEQ              0LL
@@ -122,6 +122,11 @@ extern int radeon_lockup_timeout;
 #define CAYMAN_RING_TYPE_CP1_INDEX             1
 #define CAYMAN_RING_TYPE_CP2_INDEX             2
 
+/* R600+ has an async dma ring */
+#define R600_RING_TYPE_DMA_INDEX               3
+/* cayman add a second async dma ring */
+#define CAYMAN_RING_TYPE_DMA1_INDEX            4
+
 /* hardcode those limit for now */
 #define RADEON_VA_IB_OFFSET                    (1 << 20)
 #define RADEON_VA_RESERVED_SIZE                        (8 << 20)
@@ -220,12 +225,13 @@ struct radeon_fence {
 int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring);
 int radeon_fence_driver_init(struct radeon_device *rdev);
 void radeon_fence_driver_fini(struct radeon_device *rdev);
+void radeon_fence_driver_force_completion(struct radeon_device *rdev);
 int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence **fence, int ring);
 void radeon_fence_process(struct radeon_device *rdev, int ring);
 bool radeon_fence_signaled(struct radeon_fence *fence);
 int radeon_fence_wait(struct radeon_fence *fence, bool interruptible);
 int radeon_fence_wait_next_locked(struct radeon_device *rdev, int ring);
-void radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring);
+int radeon_fence_wait_empty_locked(struct radeon_device *rdev, int ring);
 int radeon_fence_wait_any(struct radeon_device *rdev,
                          struct radeon_fence **fences,
                          bool intr);
@@ -313,6 +319,7 @@ struct radeon_bo {
        struct list_head                list;
        /* Protected by tbo.reserved */
        u32                             placements[3];
+       u32                             busy_placements[3];
        struct ttm_placement            placement;
        struct ttm_buffer_object        tbo;
        struct ttm_bo_kmap_obj          kmap;
@@ -787,6 +794,15 @@ int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *cp, unsigne
 void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *cp);
 
 
+/* r600 async dma */
+void r600_dma_stop(struct radeon_device *rdev);
+int r600_dma_resume(struct radeon_device *rdev);
+void r600_dma_fini(struct radeon_device *rdev);
+
+void cayman_dma_stop(struct radeon_device *rdev);
+int cayman_dma_resume(struct radeon_device *rdev);
+void cayman_dma_fini(struct radeon_device *rdev);
+
 /*
  * CS.
  */
@@ -824,6 +840,7 @@ struct radeon_cs_parser {
        struct radeon_cs_reloc  *relocs;
        struct radeon_cs_reloc  **relocs_ptr;
        struct list_head        validated;
+       unsigned                dma_reloc_idx;
        /* indices of various chunks */
        int                     chunk_ib_idx;
        int                     chunk_relocs_idx;
@@ -883,7 +900,9 @@ struct radeon_wb {
 #define RADEON_WB_CP_RPTR_OFFSET 1024
 #define RADEON_WB_CP1_RPTR_OFFSET 1280
 #define RADEON_WB_CP2_RPTR_OFFSET 1536
+#define R600_WB_DMA_RPTR_OFFSET   1792
 #define R600_WB_IH_WPTR_OFFSET   2048
+#define CAYMAN_WB_DMA1_RPTR_OFFSET   2304
 #define R600_WB_EVENT_OFFSET     3072
 
 /**
@@ -1539,6 +1558,8 @@ struct radeon_device {
        /* Register mmio */
        resource_size_t                 rmmio_base;
        resource_size_t                 rmmio_size;
+       /* protects concurrent MM_INDEX/DATA based register access */
+       spinlock_t mmio_idx_lock;
        void __iomem                    *rmmio;
        radeon_rreg_t                   mc_rreg;
        radeon_wreg_t                   mc_wreg;
@@ -1614,8 +1635,10 @@ int radeon_device_init(struct radeon_device *rdev,
 void radeon_device_fini(struct radeon_device *rdev);
 int radeon_gpu_wait_for_idle(struct radeon_device *rdev);
 
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg);
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
+uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg,
+                     bool always_indirect);
+void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
+                 bool always_indirect);
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg);
 void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 
@@ -1631,9 +1654,11 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 #define WREG8(reg, v) writeb(v, (rdev->rmmio) + (reg))
 #define RREG16(reg) readw((rdev->rmmio) + (reg))
 #define WREG16(reg, v) writew(v, (rdev->rmmio) + (reg))
-#define RREG32(reg) r100_mm_rreg(rdev, (reg))
-#define DREG32(reg) printk(KERN_INFO "REGISTER: " #reg " : 0x%08X\n", r100_mm_rreg(rdev, (reg)))
-#define WREG32(reg, v) r100_mm_wreg(rdev, (reg), (v))
+#define RREG32(reg) r100_mm_rreg(rdev, (reg), false)
+#define RREG32_IDX(reg) r100_mm_rreg(rdev, (reg), true)
+#define DREG32(reg) printk(KERN_INFO "REGISTER: " #reg " : 0x%08X\n", r100_mm_rreg(rdev, (reg), false))
+#define WREG32(reg, v) r100_mm_wreg(rdev, (reg), (v), false)
+#define WREG32_IDX(reg, v) r100_mm_wreg(rdev, (reg), (v), true)
 #define REG_SET(FIELD, v) (((v) << FIELD##_SHIFT) & FIELD##_MASK)
 #define REG_GET(FIELD, v) (((v) << FIELD##_SHIFT) & FIELD##_MASK)
 #define RREG32_PLL(reg) rdev->pll_rreg(rdev, (reg))
@@ -1658,7 +1683,7 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
                tmp_ |= ((val) & ~(mask));                      \
                WREG32_PLL(reg, tmp_);                          \
        } while (0)
-#define DREG32_SYS(sqf, rdev, reg) seq_printf((sqf), #reg " : 0x%08X\n", r100_mm_rreg((rdev), (reg)))
+#define DREG32_SYS(sqf, rdev, reg) seq_printf((sqf), #reg " : 0x%08X\n", r100_mm_rreg((rdev), (reg), false))
 #define RREG32_IO(reg) r100_io_rreg(rdev, (reg))
 #define WREG32_IO(reg, v) r100_io_wreg(rdev, (reg), (v))