[compiler-rt] Mark symbolize_stack_fp test unsupported on Arm Thumb

[DavidSpickett]

Summary

The new test symbolize_stack_fp.cpp added in
https://reviews.llvm.org/D102046 assumes that
we can fall back to the fast unwinder.

This is not the case for Thumb and the test is currently
failing on our v7 thumb bot:
https://lab.llvm.org/buildbot/#/builders/26/builds/2096

Skip the test if we're building for a Thumb target.

[MaskRay]

compiler-rt/test/sanitizer_common/TestCases/Linux/symbolize_stack_fp.cpp

10

Does UNSUPPORTED: thumb work? I think it is too early to generalize this to a new feature.

[DavidSpickett]

Use thumb feature instead. (still needs a config change, there's no existing thumb feature)

[MaskRay]

Thanks!

compiler-rt/test/sanitizer_common/TestCases/Linux/symbolize_stack_fp.cpp

 /// When the main program doesn't use .eh_frame, the slow unwinder does not work.
 /// Test that we can fall back to the fast unwinder.
 // RUN: %clangxx -O0 -g1 -fno-exceptions -fno-unwind-tables -fno-asynchronous-unwind-tables -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer %s -o %t
 // RUN: llvm-readelf -S %t | FileCheck %s --check-prefix=SEC
 // RUN: %run %t 2>&1 | FileCheck %s
 
 // On android %t is a wrapper python script so llvm-readelf will fail.
 // UNSUPPORTED: android
 
+// REQUIRES: fast-unwinder-works
+
 /// No .eh_frame && -g => .debug_frame
 // SEC: .debug_frame
 
 #include <sanitizer/common_interface_defs.h>
 #include <vector>
 
 template <int N>
 struct A {

compiler-rt/test/sanitizer_common/lit.common.cfg.py

 else:
   lit_config.fatal("Unknown tool for sanitizer_common tests: %r" % config.tool_name)
 
 config.available_features.add(config.tool_name)
 
 if config.host_os == 'Linux' and config.tool_name == "lsan" and config.target_arch == 'i386':
   config.available_features.add("lsan-x86")
 
+# Fast unwinder doesn't work with Thumb
+if not config.arm_thumb:
+  config.available_features.add('fast-unwinder-works')
+
 if config.host_os == 'Darwin':
   # On Darwin, we default to `abort_on_error=1`, which would make tests run
   # much slower. Let's override this and run lit tests with 'abort_on_error=0'.
   default_tool_options += ['abort_on_error=0']
   if config.tool_name == "tsan":
     default_tool_options += ['ignore_interceptors_accesses=0']
 elif config.android:
   # The same as on Darwin, we default to "abort_on_error=1" which slows down

compiler-rt/test/sanitizer_common/lit.site.cfg.py.in

 @LIT_SITE_CFG_IN_HEADER@
 
 # Tool-specific config options.
 config.name_suffix = "@CONFIG_NAME@"
 config.tool_name = "@SANITIZER_COMMON_LIT_TEST_MODE@"
 config.target_cflags = "@SANITIZER_COMMON_TEST_TARGET_CFLAGS@"
 config.target_arch = "@SANITIZER_COMMON_TEST_TARGET_ARCH@"
 
+config.arm_thumb = "@COMPILER_RT_ARM_THUMB@"
+
 # Load common config for all compiler-rt lit tests.
 lit_config.load_config(config, "@COMPILER_RT_BINARY_DIR@/test/lit.common.configured")
 
 # Load tool-specific config that would do the real work.
 lit_config.load_config(config, "@SANITIZER_COMMON_LIT_SOURCE_DIR@/lit.common.cfg.py")