diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c | 153 |
1 files changed, 134 insertions, 19 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c index 6409d6b1b2df..1105a09e55dc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c @@ -20,12 +20,12 @@ * OTHER DEALINGS IN THE SOFTWARE. */ #include <linux/module.h> -#include <linux/fdtable.h> #include <linux/uaccess.h> #include <linux/firmware.h> #include "amdgpu.h" #include "amdgpu_amdkfd.h" #include "amdgpu_amdkfd_arcturus.h" +#include "amdgpu_reset.h" #include "sdma0/sdma0_4_2_2_offset.h" #include "sdma0/sdma0_4_2_2_sh_mask.h" #include "sdma1/sdma1_4_2_2_offset.h" @@ -48,6 +48,8 @@ #include "amdgpu_amdkfd_gfx_v9.h" #include "gfxhub_v1_0.h" #include "mmhub_v9_4.h" +#include "gc/gc_9_0_offset.h" +#include "gc/gc_9_0_sh_mask.h" #define HQD_N_REGS 56 #define DUMP_REG(addr) do { \ @@ -57,11 +59,6 @@ (*dump)[i++][1] = RREG32(addr); \ } while (0) -static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) -{ - return (struct amdgpu_device *)kgd; -} - static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd) { return (struct v9_sdma_mqd *)mqd; @@ -123,10 +120,9 @@ static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev, return sdma_rlc_reg_offset; } -int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, +int kgd_arcturus_hqd_sdma_load(struct amdgpu_device *adev, void *mqd, uint32_t __user *wptr, struct mm_struct *mm) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; unsigned long end_jiffies; @@ -193,18 +189,17 @@ int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd, return 0; } -int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd, +int kgd_arcturus_hqd_sdma_dump(struct amdgpu_device *adev, uint32_t engine_id, uint32_t queue_id, uint32_t (**dump)[2], uint32_t *n_regs) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, engine_id, queue_id); uint32_t i = 0, reg; #undef HQD_N_REGS #define HQD_N_REGS (19+6+7+10) - *dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL); + *dump = kmalloc_array(HQD_N_REGS, sizeof(**dump), GFP_KERNEL); if (*dump == NULL) return -ENOMEM; @@ -225,9 +220,9 @@ int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd, return 0; } -bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) +bool kgd_arcturus_hqd_sdma_is_occupied(struct amdgpu_device *adev, + void *mqd) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t sdma_rlc_rb_cntl; @@ -244,10 +239,9 @@ bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd) return false; } -int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, +int kgd_arcturus_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd, unsigned int utimeout) { - struct amdgpu_device *adev = get_amdgpu_device(kgd); struct v9_sdma_mqd *m; uint32_t sdma_rlc_reg_offset; uint32_t temp; @@ -284,6 +278,117 @@ int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd, return 0; } +/* + * Helper used to suspend/resume gfx pipe for image post process work to set + * barrier behaviour. + */ +static int suspend_resume_compute_scheduler(struct amdgpu_device *adev, bool suspend) +{ + int i, r = 0; + + for (i = 0; i < adev->gfx.num_compute_rings; i++) { + struct amdgpu_ring *ring = &adev->gfx.compute_ring[i]; + + if (!amdgpu_ring_sched_ready(ring)) + continue; + + /* stop secheduler and drain ring. */ + if (suspend) { + drm_sched_stop(&ring->sched, NULL); + r = amdgpu_fence_wait_empty(ring); + if (r) + goto out; + } else { + drm_sched_start(&ring->sched, 0); + } + } + +out: + /* return on resume or failure to drain rings. */ + if (!suspend || r) + return r; + + return amdgpu_device_ip_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GFX); +} + +static void set_barrier_auto_waitcnt(struct amdgpu_device *adev, bool enable_waitcnt) +{ + uint32_t data; + + WRITE_ONCE(adev->barrier_has_auto_waitcnt, enable_waitcnt); + + if (!down_read_trylock(&adev->reset_domain->sem)) + return; + + amdgpu_amdkfd_suspend(adev, true); + + if (suspend_resume_compute_scheduler(adev, true)) + goto out; + + data = RREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CONFIG)); + data = REG_SET_FIELD(data, SQ_CONFIG, DISABLE_BARRIER_WAITCNT, + !enable_waitcnt); + WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CONFIG), data); + +out: + suspend_resume_compute_scheduler(adev, false); + + amdgpu_amdkfd_resume(adev, true); + + up_read(&adev->reset_domain->sem); +} + +/* + * restore_dbg_registers is ignored here but is a general interface requirement + * for devices that support GFXOFF and where the RLC save/restore list + * does not support hw registers for debugging i.e. the driver has to manually + * initialize the debug mode registers after it has disabled GFX off during the + * debug session. + */ +static uint32_t kgd_arcturus_enable_debug_trap(struct amdgpu_device *adev, + bool restore_dbg_registers, + uint32_t vmid) +{ + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + set_barrier_auto_waitcnt(adev, true); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} + +/* + * keep_trap_enabled is ignored here but is a general interface requirement + * for devices that support multi-process debugging where the performance + * overhead from trap temporary setup needs to be bypassed when the debug + * session has ended. + */ +static uint32_t kgd_arcturus_disable_debug_trap(struct amdgpu_device *adev, + bool keep_trap_enabled, + uint32_t vmid) +{ + + mutex_lock(&adev->grbm_idx_mutex); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, true); + + set_barrier_auto_waitcnt(adev, false); + + WREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_GDBG_TRAP_MASK), 0); + + kgd_gfx_v9_set_wave_launch_stall(adev, vmid, false); + + mutex_unlock(&adev->grbm_idx_mutex); + + return 0; +} const struct kfd2kgd_calls arcturus_kfd2kgd = { .program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings, .set_pasid_vmid_mapping = kgd_gfx_v9_set_pasid_vmid_mapping, @@ -297,13 +402,23 @@ const struct kfd2kgd_calls arcturus_kfd2kgd = { .hqd_sdma_is_occupied = kgd_arcturus_hqd_sdma_is_occupied, .hqd_destroy = kgd_gfx_v9_hqd_destroy, .hqd_sdma_destroy = kgd_arcturus_hqd_sdma_destroy, - .address_watch_disable = kgd_gfx_v9_address_watch_disable, - .address_watch_execute = kgd_gfx_v9_address_watch_execute, .wave_control_execute = kgd_gfx_v9_wave_control_execute, - .address_watch_get_offset = kgd_gfx_v9_address_watch_get_offset, .get_atc_vmid_pasid_mapping_info = kgd_gfx_v9_get_atc_vmid_pasid_mapping_info, .set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base, - .get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy + .enable_debug_trap = kgd_arcturus_enable_debug_trap, + .disable_debug_trap = kgd_arcturus_disable_debug_trap, + .validate_trap_override_request = kgd_gfx_v9_validate_trap_override_request, + .set_wave_launch_trap_override = kgd_gfx_v9_set_wave_launch_trap_override, + .set_wave_launch_mode = kgd_gfx_v9_set_wave_launch_mode, + .set_address_watch = kgd_gfx_v9_set_address_watch, + .clear_address_watch = kgd_gfx_v9_clear_address_watch, + .get_iq_wait_times = kgd_gfx_v9_get_iq_wait_times, + .build_dequeue_wait_counts_packet_info = kgd_gfx_v9_build_dequeue_wait_counts_packet_info, + .get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy, + .program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings, + .hqd_get_pq_addr = kgd_gfx_v9_hqd_get_pq_addr, + .hqd_reset = kgd_gfx_v9_hqd_reset, + .hqd_sdma_get_doorbell = kgd_gfx_v9_hqd_sdma_get_doorbell }; |