vulkan: Pull the device lost framework from ANV

It's a bit on the over-complicated side but the objective is to make the
debug log messages show up in the same thread as the first
VK_ERROR_DEVICE_LOST so we don't massively confuse the app.  It's
unknown if this is actually ever a problem but, with submit happening
off on its own thread, logging errors from threads the client doesn't
know about doesn't seem like a massively great plan.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13427>

docs/envvars.rst

@@ -189,6 +189,10 @@ Core Mesa environment variables
    overrides the WSI present mode clients specify in
    ``VkSwapchainCreateInfoKHR::presentMode``. Values can be ``fifo``,
    ``relaxed``, ``mailbox`` or ``immediate``.
+:envvar:`MESA_VK_ABORT_ON_DEVICE_LOSS`
+   causes the Vulkan driver to call abort() immediately after detecting a
+   lost device.  This is extremely useful when testing as it prevents the
+   test suite from continuing on with a lost device.
 :envvar:`MESA_LOADER_DRIVER_OVERRIDE`
    chooses a different driver binary such as ``etnaviv`` or ``zink``.
 

src/vulkan/runtime/vk_device.c

@@ -29,6 +29,7 @@
 #include "vk_physical_device.h"
 #include "vk_queue.h"
 #include "vk_util.h"
+#include "util/debug.h"
 #include "util/hash_table.h"
 #include "util/ralloc.h"
 
@@ -117,6 +118,45 @@ vk_device_finish(UNUSED struct vk_device *device)
    vk_object_base_finish(&device->base);
 }
 
+void
+_vk_device_report_lost(struct vk_device *device)
+{
+   assert(p_atomic_read(&device->_lost.lost) > 0);
+
+   device->_lost.reported = true;
+
+   vk_foreach_queue(queue, device) {
+      if (queue->_lost.lost) {
+         __vk_errorf(queue, VK_ERROR_DEVICE_LOST,
+                     queue->_lost.error_file, queue->_lost.error_line,
+                     "%s", queue->_lost.error_msg);
+      }
+   }
+}
+
+VkResult
+_vk_device_set_lost(struct vk_device *device,
+                    const char *file, int line,
+                    const char *msg, ...)
+{
+   /* This flushes out any per-queue device lost messages */
+   if (vk_device_is_lost(device))
+      return VK_ERROR_DEVICE_LOST;
+
+   p_atomic_inc(&device->_lost.lost);
+   device->_lost.reported = true;
+
+   va_list ap;
+   va_start(ap, msg);
+   __vk_errorv(device, VK_ERROR_DEVICE_LOST, file, line, msg, ap);
+   va_end(ap);
+
+   if (env_var_as_boolean("MESA_VK_ABORT_ON_DEVICE_LOSS", false))
+      abort();
+
+   return VK_ERROR_DEVICE_LOST;
+}
+
 PFN_vkVoidFunction
 vk_device_get_proc_addr(const struct vk_device *device,
                         const char *name)

src/vulkan/runtime/vk_device.h

@@ -28,6 +28,7 @@
 #include "vk_object.h"
 
 #include "util/list.h"
+#include "util/u_atomic.h"
 
 #ifdef __cplusplus
 extern "C" {
@@ -47,6 +48,11 @@ struct vk_device {
 
    struct list_head queues;
 
+   struct {
+      int lost;
+      bool reported;
+   } _lost;
+
 #ifdef ANDROID
    mtx_t swapchain_private_mtx;
    struct hash_table *swapchain_private;
@@ -66,6 +72,31 @@ vk_device_init(struct vk_device *device,
 void
 vk_device_finish(struct vk_device *device);
 
+VkResult PRINTFLIKE(4, 5)
+_vk_device_set_lost(struct vk_device *device,
+                    const char *file, int line,
+                    const char *msg, ...);
+
+#define vk_device_set_lost(device, ...) \
+   _vk_device_set_lost(device, __FILE__, __LINE__, __VA_ARGS__)
+
+void _vk_device_report_lost(struct vk_device *device);
+
+static inline bool
+vk_device_is_lost_no_report(struct vk_device *device)
+{
+   return p_atomic_read(&device->_lost.lost) > 0;
+}
+
+static inline bool
+vk_device_is_lost(struct vk_device *device)
+{
+   int lost = vk_device_is_lost_no_report(device);
+   if (unlikely(lost && !device->_lost.reported))
+      _vk_device_report_lost(device);
+   return lost;
+}
+
 PFN_vkVoidFunction
 vk_device_get_proc_addr(const struct vk_device *device,
                         const char *name);

src/vulkan/runtime/vk_queue.c

@@ -23,6 +23,8 @@
 
 #include "vk_queue.h"
 
+#include "util/debug.h"
+
 #include "vk_device.h"
 
 VkResult
@@ -54,3 +56,30 @@ vk_queue_finish(struct vk_queue *queue)
    list_del(&queue->link);
    vk_object_base_finish(&queue->base);
 }
+
+VkResult
+_vk_queue_set_lost(struct vk_queue *queue,
+                   const char *file, int line,
+                   const char *msg, ...)
+{
+   if (queue->_lost.lost)
+      return VK_ERROR_DEVICE_LOST;
+
+   queue->_lost.lost = true;
+   queue->_lost.error_file = file;
+   queue->_lost.error_line = line;
+
+   va_list ap;
+   va_start(ap, msg);
+   vsnprintf(queue->_lost.error_msg, sizeof(queue->_lost.error_msg), msg, ap);
+   va_end(ap);
+
+   p_atomic_inc(&queue->base.device->_lost.lost);
+
+   if (env_var_as_boolean("MESA_VK_ABORT_ON_DEVICE_LOSS", false)) {
+      _vk_device_report_lost(queue->base.device);
+      abort();
+   }
+
+   return VK_ERROR_DEVICE_LOST;
+}

src/vulkan/runtime/vk_queue.h

@@ -48,6 +48,14 @@ struct vk_queue {
    /* Which queue this is within the queue family */
    uint32_t index_in_family;
 
+   struct {
+      /* Only set once atomically by the queue */
+      int lost;
+      int error_line;
+      const char *error_file;
+      char error_msg[80];
+   } _lost;
+
    /**
     * VK_EXT_debug_utils
     *
@@ -99,6 +107,20 @@ vk_queue_init(struct vk_queue *queue, struct vk_device *device,
 void
 vk_queue_finish(struct vk_queue *queue);
 
+VkResult PRINTFLIKE(4, 5)
+_vk_queue_set_lost(struct vk_queue *queue,
+                   const char *file, int line,
+                   const char *msg, ...);
+
+#define vk_queue_set_lost(queue, ...) \
+   _vk_queue_set_lost(queue, __FILE__, __LINE__, __VA_ARGS__)
+
+static inline bool
+vk_queue_is_lost(struct vk_queue *queue)
+{
+   return queue->_lost.lost;
+}
+
 #define vk_foreach_queue(queue, device) \
    list_for_each_entry(struct vk_queue, queue, &(device)->queues, link)