Introduce header layering (#503)

See src/snmalloc/README.md for an explanation of the layers.

Some other cleanups on the way:

Fine-grained stats support is now gone.

It's been broken for two years, it depends on iostream (which then
causes linker failures with libstdc++) and it's collecting the wrong
stats for the new design.  After discussion with @mjp41, it's better to
remove it and introduce new stats support later, rather than keep broken
code in the main branch.

Tracing was controlled with a preprocessor macro, now there's also a
CMake option.

src/test/func/bits/bits.cc

@@ -3,7 +3,7 @@
  */
 
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/setup.h>
 
 void test_ctz()
@@ -41,4 +41,4 @@ int main(int argc, char** argv)
 
   test_clz();
   test_ctz();
-}
+}

src/test/func/domestication/domestication.cc

@@ -10,8 +10,8 @@ int main()
 
 // #  define SNMALLOC_TRACING
 
-#  include <backend/backend.h>
-#  include <snmalloc_core.h>
+#  include <snmalloc/backend/backend.h>
+#  include <snmalloc/snmalloc_core.h>
 
 // Specify type of allocator
 #  define SNMALLOC_PROVIDE_OWN_CONFIG
@@ -98,7 +98,7 @@ namespace snmalloc
 }
 
 #  define SNMALLOC_NAME_MANGLE(a) test_##a
-#  include "../../../override/malloc.cc"
+#  include <snmalloc/override/malloc.cc>
 
 int main()
 {

src/test/func/external_pagemap/external_pagemap.cc

@@ -9,7 +9,7 @@ int main()
 }
 #else
 #  define SNMALLOC_EXPOSE_PAGEMAP 1
-#  include <override/malloc.cc>
+#  include <snmalloc/override/malloc.cc>
 
 using ExternalChunkmap =
   ExternalGlobalPagemapTemplate<ChunkmapPagemap, snmalloc_chunkmap_global_get>;

src/test/func/first_operation/first_operation.cc

@@ -8,7 +8,7 @@
 #include "test/setup.h"
 
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <thread>
 
 void alloc1(size_t size)

src/test/func/fixed_region/fixed_region.cc

@@ -1,8 +1,8 @@
-#include "backend/fixedglobalconfig.h"
 #include "test/setup.h"
 
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/backend/fixedglobalconfig.h>
+#include <snmalloc/snmalloc.h>
 
 #ifdef assert
 #  undef assert

src/test/func/jemalloc/jemalloc.cc

@@ -9,8 +9,8 @@
 #define SNMALLOC_BOOTSTRAP_ALLOCATOR
 #define SNMALLOC_JEMALLOC3_EXPERIMENTAL
 #define SNMALLOC_JEMALLOC_NONSTANDARD
-#include "../../../override/jemalloc_compat.cc"
-#include "../../../override/malloc.cc"
+#include <snmalloc/override/jemalloc_compat.cc>
+#include <snmalloc/override/malloc.cc>
 
 #if __has_include(<malloc_np.h>)
 #  include <malloc_np.h>

src/test/func/malloc/malloc.cc

@@ -6,7 +6,7 @@
 #undef SNMALLOC_NO_REALLOCARRAY
 #undef SNMALLOC_NO_REALLOCARR
 #define SNMALLOC_BOOTSTRAP_ALLOCATOR
-#include "../../../override/malloc.cc"
+#include <snmalloc/override/malloc.cc>
 
 using namespace snmalloc;
 

src/test/func/memcpy/func-memcpy.cc

@@ -19,14 +19,13 @@ int main()
 #  endif
 #  define SNMALLOC_FAIL_FAST false
 #  define SNMALLOC_STATIC_LIBRARY_PREFIX my_
-#  include "ds/defines.h"
 #  ifndef SNMALLOC_PASS_THROUGH
-#    include "override/malloc.cc"
+#    include "snmalloc/override/malloc.cc"
 #  else
 #    define my_malloc(x) malloc(x)
 #    define my_free(x) free(x)
 #  endif
-#  include "override/memcpy.cc"
+#  include "snmalloc/override/memcpy.cc"
 #  include "test/helpers.h"
 
 #  include <assert.h>

src/test/func/memory/memory.cc

@@ -1,5 +1,5 @@
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/opt.h>
 #include <test/setup.h>
 #include <test/xoroshiro.h>

src/test/func/memory_usage/memory_usage.cc

@@ -7,8 +7,8 @@
 #include <vector>
 
 #define SNMALLOC_NAME_MANGLE(a) our_##a
-#include "../../../override/malloc-extensions.cc"
-#include "../../../override/malloc.cc"
+#include "../../../snmalloc/override/malloc-extensions.cc"
+#include "../../../snmalloc/override/malloc.cc"
 
 using namespace snmalloc;
 

src/test/func/pagemap/pagemap.cc

@@ -6,10 +6,8 @@
  * but no examples were using multiple levels of pagemap.
  */
 
-#include <backend/pagemap.h>
-#include <ds/bits.h>
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/setup.h>
 
 using namespace snmalloc;

src/test/func/pool/pool.cc

@@ -1,4 +1,4 @@
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/opt.h>
 #include <test/setup.h>
 #include <unordered_set>

src/test/func/redblack/redblack.cc

@@ -8,10 +8,11 @@
 #include <iostream>
 #include <vector>
 
-#define SNMALLOC_TRACING
+#ifndef SNMALLOC_TRACING
+#  define SNMALLOC_TRACING
+#endif
 // Redblack tree needs some libraries with trace enabled.
-#include "ds/redblacktree.h"
-#include "snmalloc.h"
+#include "snmalloc/snmalloc.h"
 
 struct NodeRef
 {

src/test/func/release-rounding/rounding.cc

@@ -1,5 +1,5 @@
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/setup.h>
 using namespace snmalloc;
 

src/test/func/sizeclass/sizeclass.cc

@@ -1,5 +1,5 @@
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/setup.h>
 
 NOINLINE
@@ -118,4 +118,4 @@ int main(int, char**)
     abort();
 
   test_align_size();
-}
+}

src/test/func/statistics/stats.cc

@@ -1,4 +1,4 @@
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 
 int main()
 {

src/test/func/teardown/teardown.cc

@@ -8,7 +8,7 @@
 #include "test/setup.h"
 
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <thread>
 
 void trigger_teardown()

src/test/func/thread_alloc_external/thread_alloc_external.cc

@@ -2,13 +2,13 @@
 #  undef SNMALLOC_USE_PTHREAD_DESTRUCTORS
 #endif
 
-#include <snmalloc_core.h>
+#include <snmalloc/snmalloc_core.h>
 #include <test/setup.h>
 
 // Specify using own
 #define SNMALLOC_EXTERNAL_THREAD_ALLOC
 
-#include "backend/globalconfig.h"
+#include <snmalloc/backend/globalconfig.h>
 
 namespace snmalloc
 {
@@ -32,7 +32,7 @@ public:
   }
 };
 
-#include <snmalloc_front.h>
+#include <snmalloc/snmalloc_front.h>
 
 void allocator_thread_init(void)
 {

src/test/func/two_alloc_types/alloc1.cc

@@ -1,10 +1,12 @@
-#define SNMALLOC_TRACING
+#ifndef SNMALLOC_TRACING
+#  define SNMALLOC_TRACING
+#endif
 
 // Redefine the namespace, so we can have two versions.
 #define snmalloc snmalloc_enclave
 
-#include <backend/fixedglobalconfig.h>
-#include <snmalloc_core.h>
+#include <snmalloc/backend/fixedglobalconfig.h>
+#include <snmalloc/snmalloc_core.h>
 
 // Specify type of allocator
 #define SNMALLOC_PROVIDE_OWN_CONFIG
@@ -15,7 +17,7 @@ namespace snmalloc
 }
 
 #define SNMALLOC_NAME_MANGLE(a) enclave_##a
-#include "../../../override/malloc.cc"
+#include <snmalloc/override/malloc.cc>
 
 extern "C" void oe_allocator_init(void* base, void* end)
 {

src/test/func/two_alloc_types/alloc2.cc

@@ -1,6 +1,8 @@
-#define SNMALLOC_TRACING
+#ifndef SNMALLOC_TRACING
+#  define SNMALLOC_TRACING
+#endif
 
 #define SNMALLOC_NAME_MANGLE(a) host_##a
 // Redefine the namespace, so we can have two versions.
 #define snmalloc snmalloc_host
-#include "../../../override/malloc.cc"
+#include <snmalloc/override/malloc.cc>

src/test/func/two_alloc_types/main.cc

@@ -1,6 +1,5 @@
-#include "../../../snmalloc.h"
-
 #include <iostream>
+#include <snmalloc/snmalloc.h>
 #include <stdlib.h>
 #include <string.h>
 #include <test/setup.h>

src/test/perf/contention/contention.cc

@@ -5,7 +5,7 @@
 
 #include <iomanip>
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <thread>
 #include <vector>
 

src/test/perf/external_pointer/externalpointer.cc

@@ -1,4 +1,4 @@
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/measuretime.h>
 #include <test/setup.h>
 #include <test/xoroshiro.h>

src/test/perf/low_memory/low-memory.cc

@@ -1,5 +1,5 @@
 #include <iostream>
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/opt.h>
 #include <test/setup.h>
 #include <unordered_set>
@@ -164,4 +164,4 @@ int main(int argc, char** argv)
   //   std::cout << "Release test only." << std::endl;
   // #endif
   return 0;
-}
+}

src/test/perf/memcpy/memcpy.cc

@@ -1,4 +1,4 @@
-#include "mem/memcpy.h"
+#include "snmalloc/global/memcpy.h"
 
 #include <test/measuretime.h>
 #include <test/opt.h>

src/test/perf/singlethread/singlethread.cc

@@ -1,4 +1,4 @@
-#include <snmalloc.h>
+#include <snmalloc/snmalloc.h>
 #include <test/measuretime.h>
 #include <test/setup.h>
 #include <unordered_set>

src/test/setup.h

@@ -1,10 +1,9 @@
 #if defined(SNMALLOC_CI_BUILD)
-#  include <pal/pal.h>
+#  include <snmalloc/pal/pal.h>
 #  if defined(WIN32)
-#    include <ds/bits.h>
 #    include <iostream>
-#    include <pal/pal.h>
 #    include <signal.h>
+#    include <snmalloc/ds_core/ds_Core.h>
 #    include <stdlib.h>
 // Has to come after the PAL.
 #    include <DbgHelp.h>