vulkan: keep a failing shader's dump from ending the process

When a graphics pipeline's translation throws, its handler walks each
program again to dump it. A program whose flow cannot be analyzed (an
indirect branch Nintendo Switch Sports Resort uses) throws again there,
the exception leaves the handler and the process ends. Contain it, so the
pipeline fails, the error is logged and the title goes on.

1 files changed, 6 insertions, 1 deletions

diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
index ac878ddfef..bbc4f3c625 100644
--- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp
@@ -981,7 +981,12 @@ std::unique_ptr<GraphicsPipeline> PipelineCache::CreateGraphicsPipeline(
         ++env_index;
 
         const u32 cfg_offset{static_cast<u32>(env.StartAddress() + sizeof(Shader::ProgramHeader))};
-        Shader::Maxwell::Flow::CFG cfg(env, pools.flow_block, cfg_offset, index == 0);
+        try {
+            // Walks the program for the dump; the one that failed may fail here again, and an
+            // exception leaving this handler would end the process.
+            Shader::Maxwell::Flow::CFG cfg(env, pools.flow_block, cfg_offset, index == 0);
+        } catch (const Shader::Exception&) {
+        }
         env.Dump(hash, key.unique_hashes[index]);
     }
     LOG_ERROR(Render_Vulkan, "{}", exception.what());