From bd5bcbb17667e86b7225710b8bb4018ae942205d Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Thu, 24 May 2018 14:53:40 -0400 Subject: [PATCH 5622/5725] drm/amd: Update KFD-Thunk ioctl ABI to match upstream - Clean up and renumber scratch memory ioctl - Renumber get_tile_config ioctl - Renumber set_trap_handler ioctl - Update KFD_IOC_ALLOC_MEM_FLAGS - Renumber GPUVM memory management ioctls - Remove unused SEP_PROCESS_DGPU_APERTURE ioctl - Update memory management ioctls Replace device_ids_array_size (in bytes) with n_devices. Fix error handling and use n_success to update device_id arrays in objects. This commit breaks the ABI and requires a corresponding Thunk change. Change-Id: I62149841f1603ec36143836d2eb5ab0fcaf37cf5 Signed-off-by: Felix Kuehling Signed-off-by: Kalyan Alle --- drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 131 ++++++++++----------------- drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c | 10 -- drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 2 - drivers/gpu/drm/amd/amdkfd/kfd_process.c | 11 ++- 4 files changed, 53 insertions(+), 101 deletions(-) mode change 100644 => 100755 drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index 33a5793..1b96337 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -1066,17 +1066,14 @@ static int kfd_ioctl_wait_events(struct file *filp, struct kfd_process *p, return err; } -static int kfd_ioctl_alloc_scratch_memory(struct file *filep, +static int kfd_ioctl_set_scratch_backing_va(struct file *filep, struct kfd_process *p, void *data) { - struct kfd_ioctl_alloc_memory_of_scratch_args *args = data; + struct kfd_ioctl_set_scratch_backing_va_args *args = data; struct kfd_process_device *pdd; struct kfd_dev *dev; long err; - if (args->size == 0) - return -EINVAL; - dev = kfd_device_by_id(args->gpu_id); if (!dev) return -EINVAL; @@ -1352,31 +1349,30 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep, void *mem; struct kfd_dev *dev, *peer; long err = 0; - int i, num_dev = 0; + int i; uint32_t *devices_arr = NULL; dev = kfd_device_by_id(GET_GPU_ID(args->handle)); if (!dev) return -EINVAL; - if (args->device_ids_array_size == 0) { - pr_debug("Device ID array size is 0\n"); + if (!args->n_devices) { + pr_debug("Device IDs array empty\n"); return -EINVAL; } - - if (args->device_ids_array_size % sizeof(uint32_t)) { - pr_debug("Node IDs array size %u\n", - args->device_ids_array_size); + if (args->n_success > args->n_devices) { + pr_debug("n_success exceeds n_devices\n"); return -EINVAL; } - devices_arr = kmalloc(args->device_ids_array_size, GFP_KERNEL); + devices_arr = kmalloc_array(args->n_devices, sizeof(*devices_arr), + GFP_KERNEL); if (!devices_arr) return -ENOMEM; err = copy_from_user(devices_arr, - (void __user *)args->device_ids_array_ptr, - args->device_ids_array_size); + (void __user *)args->device_ids_array_ptr, + args->n_devices * sizeof(*devices_arr)); if (err != 0) { err = -EFAULT; goto copy_from_user_failed; @@ -1397,12 +1393,11 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep, goto get_mem_obj_from_handle_failed; } - num_dev = args->device_ids_array_size / sizeof(uint32_t); - for (i = 0 ; i < num_dev; i++) { + for (i = args->n_success; i < args->n_devices; i++) { peer = kfd_device_by_id(devices_arr[i]); if (!peer) { pr_debug("Getting device by id failed for 0x%x\n", - devices_arr[i]); + devices_arr[i]); err = -EINVAL; goto get_mem_obj_from_handle_failed; } @@ -1413,12 +1408,13 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep, goto get_mem_obj_from_handle_failed; } err = peer->kfd2kgd->map_memory_to_gpu( - peer->kgd, (struct kgd_mem *)mem, peer_pdd->vm); - if (err != 0) { - pr_err("Failed to map to gpu %d, num_dev=%d\n", - i, num_dev); + peer->kgd, (struct kgd_mem *)mem, peer_pdd->vm); + if (err) { + pr_err("Failed to map to gpu %d/%d\n", + i, args->n_devices); goto map_memory_to_gpu_failed; } + args->n_success = i+1; } mutex_unlock(&p->mutex); @@ -1430,7 +1426,7 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep, } /* Flush TLBs after waiting for the page table updates to complete */ - for (i = 0; i < num_dev; i++) { + for (i = 0; i < args->n_devices; i++) { peer = kfd_device_by_id(devices_arr[i]); if (WARN_ON_ONCE(!peer)) continue; @@ -1463,30 +1459,29 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, void *mem; struct kfd_dev *dev, *peer; long err = 0; - uint32_t *devices_arr = NULL, num_dev, i; + uint32_t *devices_arr = NULL, i; dev = kfd_device_by_id(GET_GPU_ID(args->handle)); if (!dev) return -EINVAL; - if (args->device_ids_array_size == 0) { - pr_debug("Device ID array size is 0\n"); + if (!args->n_devices) { + pr_debug("Device IDs array empty\n"); return -EINVAL; } - - if (args->device_ids_array_size % sizeof(uint32_t)) { - pr_debug("Node IDs array size %u\n", - args->device_ids_array_size); + if (args->n_success > args->n_devices) { + pr_debug("n_success exceeds n_devices\n"); return -EINVAL; } - devices_arr = kmalloc(args->device_ids_array_size, GFP_KERNEL); + devices_arr = kmalloc_array(args->n_devices, sizeof(*devices_arr), + GFP_KERNEL); if (!devices_arr) return -ENOMEM; err = copy_from_user(devices_arr, - (void __user *)args->device_ids_array_ptr, - args->device_ids_array_size); + (void __user *)args->device_ids_array_ptr, + args->n_devices * sizeof(*devices_arr)); if (err != 0) { err = -EFAULT; goto copy_from_user_failed; @@ -1496,8 +1491,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, pdd = kfd_get_process_device_data(dev, p); if (!pdd) { - pr_debug("Process device data doesn't exist\n"); - err = -ENODEV; + err = -EINVAL; goto bind_process_to_device_failed; } @@ -1508,8 +1502,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, goto get_mem_obj_from_handle_failed; } - num_dev = args->device_ids_array_size / sizeof(uint32_t); - for (i = 0 ; i < num_dev; i++) { + for (i = args->n_success; i < args->n_devices; i++) { peer = kfd_device_by_id(devices_arr[i]); if (!peer) { err = -EINVAL; @@ -1525,9 +1518,10 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, peer->kgd, (struct kgd_mem *)mem, peer_pdd->vm); if (err) { pr_err("Failed to unmap from gpu %d/%d\n", - i, num_dev); + i, args->n_devices); goto unmap_memory_from_gpu_failed; } + args->n_success = i+1; } kfree(devices_arr); @@ -1544,34 +1538,6 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, return err; } -static int kfd_ioctl_set_process_dgpu_aperture(struct file *filep, - struct kfd_process *p, void *data) -{ - struct kfd_ioctl_set_process_dgpu_aperture_args *args = data; - struct kfd_dev *dev; - struct kfd_process_device *pdd; - long err; - - dev = kfd_device_by_id(args->gpu_id); - if (!dev) - return -EINVAL; - - mutex_lock(&p->mutex); - - pdd = kfd_bind_process_to_device(dev, p); - if (IS_ERR(pdd)) { - err = PTR_ERR(pdd); - goto exit; - } - - err = kfd_set_process_dgpu_aperture(pdd, args->dgpu_base, - args->dgpu_limit); - -exit: - mutex_unlock(&p->mutex); - return err; -} - static int kfd_ioctl_get_dmabuf_info(struct file *filep, struct kfd_process *p, void *data) { @@ -2012,6 +1978,21 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { AMDKFD_IOCTL_DEF(AMDKFD_IOC_DBG_WAVE_CONTROL, kfd_ioctl_dbg_wave_control, 0), + AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_SCRATCH_BACKING_VA, + kfd_ioctl_set_scratch_backing_va, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_TILE_CONFIG, + kfd_ioctl_get_tile_config, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_TRAP_HANDLER, + kfd_ioctl_set_trap_handler, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_PROCESS_APERTURES_NEW, + kfd_ioctl_get_process_apertures_new, 0), + + AMDKFD_IOCTL_DEF(AMDKFD_IOC_ACQUIRE_VM, + kfd_ioctl_acquire_vm, 0), + AMDKFD_IOCTL_DEF(AMDKFD_IOC_ALLOC_MEMORY_OF_GPU, kfd_ioctl_alloc_memory_of_gpu, 0), @@ -2024,30 +2005,15 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { AMDKFD_IOCTL_DEF(AMDKFD_IOC_UNMAP_MEMORY_FROM_GPU, kfd_ioctl_unmap_memory_from_gpu, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_ALLOC_MEMORY_OF_SCRATCH, - kfd_ioctl_alloc_scratch_memory, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_CU_MASK, kfd_ioctl_set_cu_mask, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_PROCESS_DGPU_APERTURE, - kfd_ioctl_set_process_dgpu_aperture, 0), - - AMDKFD_IOCTL_DEF(AMDKFD_IOC_SET_TRAP_HANDLER, - kfd_ioctl_set_trap_handler, 0), - - AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_PROCESS_APERTURES_NEW, - kfd_ioctl_get_process_apertures_new, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_DMABUF_INFO, kfd_ioctl_get_dmabuf_info, 0), AMDKFD_IOCTL_DEF(AMDKFD_IOC_IMPORT_DMABUF, kfd_ioctl_import_dmabuf, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_TILE_CONFIG, - kfd_ioctl_get_tile_config, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_IPC_IMPORT_HANDLE, kfd_ioctl_ipc_import_handle, 0), @@ -2060,9 +2026,6 @@ static const struct amdkfd_ioctl_desc amdkfd_ioctls[] = { AMDKFD_IOCTL_DEF(AMDKFD_IOC_GET_QUEUE_WAVE_STATE, kfd_ioctl_get_queue_wave_state, 0), - AMDKFD_IOCTL_DEF(AMDKFD_IOC_ACQUIRE_VM, - kfd_ioctl_acquire_vm, 0) - }; #define AMDKFD_CORE_IOCTL_COUNT ARRAY_SIZE(amdkfd_ioctls) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c old mode 100644 new mode 100755 index 2c00711..be376d93 --- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c @@ -313,16 +313,6 @@ #define SVM_CWSR_BASE (SVM_USER_BASE - KFD_CWSR_TBA_TMA_SIZE) #define SVM_IB_BASE (SVM_CWSR_BASE - PAGE_SIZE) -int kfd_set_process_dgpu_aperture(struct kfd_process_device *pdd, - uint64_t base, uint64_t limit) -{ - if (base < SVM_USER_BASE) { - pr_err("Set dgpu vm base 0x%llx failed.\n", base); - return -EINVAL; - } - return 0; -} - void kfd_init_apertures_vi(struct kfd_process_device *pdd, uint8_t id) { /* diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index b2ef0f5..be66eae 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -875,8 +875,6 @@ int kgd2kfd_post_reset(struct kfd_dev *kfd); /* amdkfd Apertures */ int kfd_init_apertures(struct kfd_process *process); -int kfd_set_process_dgpu_aperture(struct kfd_process_device *pdd, - uint64_t base, uint64_t limit); /* Queue Context Management */ int init_queue(struct queue **q, const struct queue_properties *properties); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index 5fa3559..2b4c5bd 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -182,8 +182,10 @@ static int kfd_process_alloc_gpuvm(struct kfd_process_device *pdd, static int kfd_process_device_reserve_ib_mem(struct kfd_process_device *pdd) { struct qcm_process_device *qpd = &pdd->qpd; - uint32_t flags = ALLOC_MEM_FLAGS_GTT | ALLOC_MEM_FLAGS_NONPAGED | - ALLOC_MEM_FLAGS_NO_SUBSTITUTE | ALLOC_MEM_FLAGS_EXECUTE_ACCESS; + uint32_t flags = ALLOC_MEM_FLAGS_GTT | + ALLOC_MEM_FLAGS_NO_SUBSTITUTE | + ALLOC_MEM_FLAGS_WRITABLE | + ALLOC_MEM_FLAGS_EXECUTABLE; void *kaddr; int ret; @@ -503,9 +505,8 @@ static int kfd_process_device_init_cwsr_dgpu(struct kfd_process_device *pdd) { struct kfd_dev *dev = pdd->dev; struct qcm_process_device *qpd = &pdd->qpd; - uint32_t flags = ALLOC_MEM_FLAGS_GTT | ALLOC_MEM_FLAGS_NONPAGED | - ALLOC_MEM_FLAGS_NO_SUBSTITUTE | ALLOC_MEM_FLAGS_READONLY | - ALLOC_MEM_FLAGS_EXECUTE_ACCESS; + uint32_t flags = ALLOC_MEM_FLAGS_GTT | + ALLOC_MEM_FLAGS_NO_SUBSTITUTE | ALLOC_MEM_FLAGS_EXECUTABLE; void *kaddr; int ret; -- 2.7.4