From: Iago Toral Quiroga Date: Wed, 22 Mar 2017 07:46:04 +0000 (+0100) Subject: anv/device: keep track of 'device lost' state X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=50c8d2c1f71944688ebf753af3866981a07c4e9c;p=mesa.git anv/device: keep track of 'device lost' state The Vulkan specs say: "A logical device may become lost because of hardware errors, execution timeouts, power management events and/or platform-specific events. This may cause pending and future command execution to fail and cause hardware resources to be corrupted. When this happens, certain commands will return VK_ERROR_DEVICE_LOST (see Error Codes for a list of such commands). After any such event, the logical device is considered lost. It is not possible to reset the logical device to a non-lost state, however the lost state is specific to a logical device (VkDevice), and the corresponding physical device (VkPhysicalDevice) may be otherwise unaffected. In some cases, the physical device may also be lost, and attempting to create a new logical device will fail, returning VK_ERROR_DEVICE_LOST." This means that we need to track if a logical device has been lost so we can have the commands referenced by the spec return VK_ERROR_DEVICE_LOST immediately. Reviewed-by: Jason Ekstrand --- diff --git a/src/intel/vulkan/anv_device.c b/src/intel/vulkan/anv_device.c index 2a2a470ec99..193db01b920 100644 --- a/src/intel/vulkan/anv_device.c +++ b/src/intel/vulkan/anv_device.c @@ -929,6 +929,7 @@ anv_device_submit_simple_batch(struct anv_device *device, ret = anv_gem_wait(device, bo.gem_handle, &timeout); if (ret != 0) { /* We don't know the real error. */ + device->lost = true; result = vk_errorf(VK_ERROR_DEVICE_LOST, "execbuf2 failed: %m"); goto fail; } @@ -973,6 +974,7 @@ VkResult anv_CreateDevice( device->_loader_data.loaderMagic = ICD_LOADER_MAGIC; device->instance = physical_device->instance; device->chipset_id = physical_device->chipset_id; + device->lost = false; if (pAllocator) device->alloc = *pAllocator; @@ -1250,6 +1252,7 @@ anv_device_execbuf(struct anv_device *device, int ret = anv_gem_execbuffer(device, execbuf); if (ret != 0) { /* We don't know the real error. */ + device->lost = true; return vk_errorf(VK_ERROR_DEVICE_LOST, "execbuf2 failed: %m"); } @@ -1339,6 +1342,7 @@ out: * submit the same job again to this device. */ result = VK_ERROR_DEVICE_LOST; + device->lost = true; /* If we return VK_ERROR_DEVICE LOST here, we need to ensure that * vkWaitForFences() and vkGetFenceStatus() return a valid result @@ -1865,6 +1869,7 @@ VkResult anv_WaitForFences( return VK_TIMEOUT; } else if (ret == -1) { /* We don't know the real error. */ + device->lost = true; return vk_errorf(VK_ERROR_DEVICE_LOST, "gem wait failed: %m"); } else { fence->state = ANV_FENCE_STATE_SIGNALED; diff --git a/src/intel/vulkan/anv_private.h b/src/intel/vulkan/anv_private.h index fd82ce92df6..68f7359d71d 100644 --- a/src/intel/vulkan/anv_private.h +++ b/src/intel/vulkan/anv_private.h @@ -619,6 +619,7 @@ struct anv_device { pthread_mutex_t mutex; pthread_cond_t queue_submit; + bool lost; }; static void inline