SegmentMap, code flushing, cleanup

Elf2Mac/CMakeLists.txt

@@ -19,7 +19,7 @@ cmake_minimum_required(VERSION 3.1)
 
 find_package(Boost COMPONENTS REQUIRED)
 
-add_executable(Elf2Mac Elf2Mac.h Elf2Mac.cc LdScript.cc)
+add_executable(Elf2Mac Elf2Mac.h Elf2Mac.cc SegmentMap.cc LdScript.cc)
 target_link_libraries(Elf2Mac ResourceFiles ${CMAKE_INSTALL_PREFIX}/lib/libelf.a -lz)
 
 target_include_directories(Elf2Mac PRIVATE ${CMAKE_INSTALL_PREFIX}/include)

Elf2Mac/Elf2Mac.cc

@@ -18,6 +18,7 @@
 */
 
 #include "Elf2Mac.h"
+#include "SegmentMap.h"
 
 #include "ResourceFork.h"
 #include "BinaryIO.h"
@@ -257,15 +258,22 @@ Section::Section(string name, int idx, SectionKind kind, Elf_Scn *elfsec)
 void Section::SetRela(Elf_Scn *scn)
 {
 	relasec = scn;
-	GElf_Shdr shdr;
+	GElf_Shdr rshdr;
-	gelf_getshdr(relasec, &shdr);
+	gelf_getshdr(relasec, &rshdr);
 
-	int nRela = shdr.sh_size / shdr.sh_entsize;
+	int nRela = rshdr.sh_size / rshdr.sh_entsize;
 	Elf_Data *data = elf_getdata(relasec, NULL);
 	for(int i = 0; i < nRela; i++)
 	{
 		GElf_Rela rela;
 		gelf_getrela(data, i, &rela);
+
+		if(rela.r_offset < shdr.sh_addr || rela.r_offset >= shdr.sh_addr + shdr.sh_size)
+		{
+			// For some reason, there sometimes are relocations beyond the end of the sections
+			// in LD output. That's bad. Let's ignore it.
+			continue;
+		}
 		relocs.push_back(rela);
 	}
 
@@ -370,7 +378,7 @@ void Section::FixRelocs()
 		{
 			case SectionKind::code:
 				relocBase = RelocBase::code;
-				if(sym.needsJT && (exceptionInfoStart == 0 || rela.r_offset < exceptionInfoStart))
+				if(sym.needsJT && (exceptionInfoStart == 0 || rela.r_offset < exceptionInfoStart || sym.GetName() == "__gxx_personality_v0"))
 				{
 					if(rela.r_addend == 0)
 					{
@@ -390,7 +398,20 @@ void Section::FixRelocs()
 					}
 				}
 				else
+				{
+					if(sym.section.get() != this)
+					{
+						std::cerr << "Invalid ref from "
+						          << name << ":" << std::hex << rela.r_offset-shdr.sh_addr << std::dec
+						          << " to " << sym.section->name
+						          << "(" << sym.GetName() << ")"
+						          << "+" << rela.r_offset << std::endl;
+						std::cerr << "needsJT: " << (sym.needsJT ? "true" : "false") << std::endl;
+						std::cerr << "from addr: " << rela.r_offset << ", exceptionInfoStart: " << exceptionInfoStart << std::endl;
+
+					}
 					assert(sym.section.get() == this);
+				}
 				break;
 			case SectionKind::data:
 				relocBase = RelocBase::data;
@@ -721,6 +742,11 @@ void MultiSegmentApp(string output)
 
 		std::cout << "CODE " << id << ": " << code.str().size() << " bytes\n";
 
+		if(code.str().size() == 80)
+		{
+			std::cout << "... empty. Skipping.\n";
+			continue;
+		}
 		rsrc.addResource(Resource(ResType("CODE"), id,
 		                          code.str()));
 
@@ -819,6 +845,7 @@ int main(int argc, char *argv[])
 			{
 				elf2mac = true;
 				flatoutput = true;
+				segments = false;
 			}
 			else if(*p == "--mac-segments")
 			{
@@ -845,8 +872,20 @@ int main(int argc, char *argv[])
 
 			{
 				ofstream out(tmpfile);
-				CreateLdScript(out, segments);
+				if(segments)
-				CreateLdScript(std::cout, segments);
+				{
+					SegmentMap map;
+					map.CreateLdScript(out);
+					map.CreateLdScript(std::cout);
+
+					ofstream f("/tmp/foo.ld");
+					map.CreateLdScript(f);
+				}
+				else
+				{
+					CreateFlatLdScript(out);
+					CreateFlatLdScript(std::cout);
+				}
 			}
 
 			args2.push_back("-o");

Elf2Mac/Elf2Mac.h

@@ -22,6 +22,6 @@
 
 #include <iosfwd>
 
-void CreateLdScript(std::ostream& out, bool segments);
+void CreateFlatLdScript(std::ostream& out);
 
 #endif // ELF2MAC_H

Elf2Mac/LdScript.cc

@@ -18,9 +18,11 @@
 */
 
 #include "Elf2Mac.h"
+#include "SegmentMap.h"
 
 #include <iostream>
 #include <boost/algorithm/string/replace.hpp>
+#include <boost/lexical_cast.hpp>
 #include <string>
 
 using std::string;
@@ -41,12 +43,12 @@ const char * textSection = R"ld(/* ld script for Elf2Mac */
         /* The entry point. */
         _entry_trampoline = .;
         SHORT(DEFINED(__break_on_entry) ? 0xA9FF : 0x4e71);
-        LONG(0x61000002);	/* bsr *+2 */
+        LONG(0x61000002); /* bsr *+2 */
         SHORT(0x0697); /* addi.l #_, (a7) */
         LONG(_start - _entry_trampoline - 6);
 
-        PROVIDE(_start = .);  /* fallback entry point to a safe spot - needed for libretro bootstrap */
+        PROVIDE(_start = .); /* fallback entry point to a safe spot - needed for libretro bootstrap */
-        Retro68InitMultisegApp = .;    /* override this for the single-segment case */
+        Retro68InitMultisegApp = .; /* override this for the single-segment case */
         SHORT(0x4e75); /* rts */
 
         *(.relocvars)
@@ -83,79 +85,6 @@ const char * textSection = R"ld(/* ld script for Elf2Mac */
     }
 )ld";
 
-const char * code1Section = R"ld(/* ld script for Elf2Mac */
-    .code1 :	{
-        _stext = . ;
-        FILL(0x4E71);
-        PROVIDE(_rsrc_start = .);
-        . = ALIGN (2);
-        _entry_trampoline = .;
-        SHORT(DEFINED(__break_on_entry) ? 0xA9FF : 0x4e71);
-        LONG(0x61000002);	/* bsr *+2 */
-        SHORT(0x0697); /* addi.l #_, (a7) */
-        LONG(_start - _entry_trampoline - 6);
-        PROVIDE(_start = .);  /* fallback entry point to a safe spot - needed for libretro bootstrap */
-        SHORT(0x4e75); /* rts */
-
-        *(.relocvars)
-        */libretrocrt.a:start.c.obj(.text*)
-        */libretrocrt.a:relocate.c.obj(.text*) 
-        */libretrocrt.a:MultiSegApp.c.obj(.text*) 
-        */libretrocrt.a:LoadSeg.s.obj(.text*) 
-        */libretrocrt.a:*(.text*) 
-        */libgcc.a:*(.text*) 
-        */libc.a:*(.text*)
-       
-        . = ALIGN (4) ;
-        __init_section = . ;
-        KEEP (*(.init))
-        __init_section_end = . ;
-        __fini_section = . ;
-        KEEP (*(.fini))
-        __fini_section_end = . ;
-
-        __EH_FRAME_BEGIN__ = .;
-        KEEP(*/libretrocrt.a:*(.eh_frame))
-        KEEP(*/libgcc.a:*(.eh_frame))
-        KEEP(*/libc.a:*(.eh_frame))
-        LONG(0);
-        KEEP(*/libretrocrt.a:*(.gcc_except_table))
-        KEEP(*/libretrocrt.a:*(.gcc_except_table.*))
-        KEEP(*/libgcc.a:*(.gcc_except_table))
-        KEEP(*/libgcc.a:*(.gcc_except_table.*))
-        KEEP(*/libc.a:*(.gcc_except_table))
-        KEEP(*/libc.a:*(.gcc_except_table.*))
-
-        . = ALIGN(0x4) ;
-        _etext = . ;
-    }
-)ld";
-const char * codeSectionTemplate = R"ld(/* ld script for Elf2Mac */
-    .code@N@ :	{
-        FILL(0x4E71);
-        @FILTER@(.text*)
-
-        @EXTRA@
-
-        . = ALIGN (4) ;
-
-        __EH_FRAME_BEGIN__@N@ = .;
-        KEEP(@FILTER@(.eh_frame))
-        LONG(0);
-        KEEP(@FILTER@(.gcc_except_table))
-        KEEP(@FILTER@(.gcc_except_table.*))
-
-        . = ALIGN(0x4);
-        . += 32;
-        LONG(__EH_FRAME_BEGIN__@N@ - .);
-    }
-)ld";
-
-const char * lastCodeExtra = R"ld(
-        *(.gnu.linkonce.t*)
-)ld";
-
-
 const char * scriptEnd = R"ld(
     .data : {
         _sdata = . ;
@@ -212,7 +141,7 @@ const char * scriptEnd = R"ld(
      * Keep them for now, they are discarded by Elf2Mac. */
 
     /DISCARD/ : { *(.note.GNU-stack) }
-    /* Stabs debugging sections.    */
+    /* Stabs debugging sections.	*/
     .stab 0 : { *(.stab) }
     .stabstr 0 : { *(.stabstr) }
     .stab.excl 0 : { *(.stab.excl) }
@@ -224,27 +153,27 @@ const char * scriptEnd = R"ld(
       Symbols in the DWARF debugging sections are relative to the beginning
       of the section so we begin them at 0.  */
     /* DWARF 1 */
-    .debug          0 : { *(.debug) }
+    .debug 0 : { *(.debug) }
-    .line           0 : { *(.line) }
+    .line 0 : { *(.line) }
     /* GNU DWARF 1 extensions */
-    .debug_srcinfo  0 : { *(.debug_srcinfo) }
+    .debug_srcinfo 0 : { *(.debug_srcinfo) }
-    .debug_sfnames  0 : { *(.debug_sfnames) }
+    .debug_sfnames 0 : { *(.debug_sfnames) }
     /* DWARF 1.1 and DWARF 2 */
-    .debug_aranges  0 : { *(.debug_aranges) }
+    .debug_aranges 0 : { *(.debug_aranges) }
     .debug_pubnames 0 : { *(.debug_pubnames) }
     /* DWARF 2 */
-    .debug_info     0 : { *(.debug_info .gnu.linkonce.wi.*) }
+    .debug_info 0 : { *(.debug_info .gnu.linkonce.wi.*) }
-    .debug_abbrev   0 : { *(.debug_abbrev) }
+    .debug_abbrev 0 : { *(.debug_abbrev) }
-    .debug_line     0 : { *(.debug_line) }
+    .debug_line 0 : { *(.debug_line) }
-    .debug_frame    0 : { *(.debug_frame) }
+    .debug_frame 0 : { *(.debug_frame) }
-    .debug_str      0 : { *(.debug_str) }
+    .debug_str 0 : { *(.debug_str) }
-    .debug_loc      0 : { *(.debug_loc) }
+    .debug_loc 0 : { *(.debug_loc) }
-    .debug_macinfo  0 : { *(.debug_macinfo) }
+    .debug_macinfo 0 : { *(.debug_macinfo) }
     /* SGI/MIPS DWARF 2 extensions */
     .debug_weaknames 0 : { *(.debug_weaknames) }
     .debug_funcnames 0 : { *(.debug_funcnames) }
     .debug_typenames 0 : { *(.debug_typenames) }
-    .debug_varnames  0 : { *(.debug_varnames) }
+    .debug_varnames 0 : { *(.debug_varnames) }
 
     /DISCARD/ : { *(*) }
 }
@@ -252,22 +181,112 @@ const char * scriptEnd = R"ld(
 )ld";
 
 
-void CreateLdScript(std::ostream& out, bool segments)
+void CreateFlatLdScript(std::ostream& out)
 {
-	if(segments)
+	out << "_MULTISEG_APP = 0;\n";
+	out << scriptStart << textSection << scriptEnd;
+}
+
+
+void SegmentInfo::WriteFilters(std::ostream &out, string section)
+{
+	for(string filter : filters)
 	{
-        out << "_MULTISEG_APP = 1;\n";
+		out << "        " << filter << "(" << section << ")\n";
-		out << scriptStart << code1Section;
+		out << "        " << filter << "(" << section << ".*)\n";
-		string code = codeSectionTemplate;
+	}
-		boost::replace_all(code, "@N@", "2");
+}
-		boost::replace_all(code, "@FILTER@", "*");
+void SegmentInfo::WriteFiltersKeep(std::ostream &out, string section)
-		boost::replace_all(code, "@EXTRA@", lastCodeExtra);
+{
-		out << code;
+	for(string filter : filters)
-		out << scriptEnd;
+	{
+		out << "\t\tKEEP(" << filter << "(" << section << "))\n";
+		out << "\t\tKEEP(" << filter << "(" << section << ".*))\n";
+	}
+}
+
+void SegmentInfo::CreateLdScript(std::ostream &out)
+{
+	out << "\t.code" << id << " : {\n";
+	out << "\t\tFILL(0x4E71);\n";
+	if(id == 1)
+	{
+		out << R"ld(
+		_stext = .;
+		FILL(0x4E71);
+		PROVIDE(_rsrc_start = .);
+		. = ALIGN (2);
+		_entry_trampoline = .;
+		SHORT(DEFINED(__break_on_entry) ? 0xA9FF : 0x4e71);
+		LONG(0x61000002);	/* bsr *+2 */
+		SHORT(0x0697); /* addi.l #_, (a7) */
+		LONG(_start - _entry_trampoline - 6);
+		PROVIDE(_start = .);  /* fallback entry point to a safe spot - needed for libretro bootstrap */
+		SHORT(0x4e75); /* rts */
+
+		FILL(0);
+		*(.relocvars)
+		FILL(0x4E71);
+)ld";
+	}
+	WriteFilters(out, ".text");
+
+	if(id == 2)
+	{
+		out << "\t\t*(.gnu.linkonce.t*)\n";
+	}
+	if(id == 1)
+	{
+		out << R"ld(
+		. = ALIGN (4) ;
+		__init_section = .;
+		KEEP (*(.init))
+		__init_section_end = .;
+		__fini_section = .;
+		KEEP (*(.fini))
+		__fini_section_end = .;
+)ld";
+	}
+
+	out << "\t\t. = ALIGN (4);\n";	// this is important, for some reason.
+	if(id == 1)
+		out << "\t\t__EH_FRAME_BEGIN__" << " = .;\n";
+	else
+		out << "\t\t__EH_FRAME_BEGIN__" << id << " = .;\n";
+	WriteFiltersKeep(out, ".eh_frame");
+	out << "\t\tLONG(0);\n";
+	WriteFiltersKeep(out, ".gcc_except_table");
+
+	if(id == 1)
+	{
+		out << R"ld(
+		. = ALIGN(0x4) ;
+		_etext = . ;
+)ld";
 	}
 	else
-    {
+	{
-        out << "_MULTISEG_APP = 0;\n";
+		out << boost::replace_all_copy<string>(R"ld(
-		out << scriptStart << textSection << scriptEnd;
+		. = ALIGN(0x4);
-    }
+		FILL(0);
+		. += 32;
+		LONG(__EH_FRAME_BEGIN__@N@ - .);
+)ld", "@N@", boost::lexical_cast<string>(id));
+	}
+
+	out << "\t}\n";
+
+}
+
+void SegmentMap::CreateLdScript(std::ostream &out)
+{
+	out << "_MULTISEG_APP = 1;\n";
+	out << scriptStart;
+
+	for(SegmentInfo& seg: segments)
+	{
+		seg.CreateLdScript(out);
+	}
+
+	out << scriptEnd;
 }

Elf2Mac/SegmentMap.cc

@@ -0,0 +1,28 @@
+#include "SegmentMap.h"
+
+
+SegmentInfo::SegmentInfo()
+{
+
+}
+
+
+SegmentMap::SegmentMap()
+{
+	segments.emplace_back(1, "Runtime",
+	                      "*/libretrocrt.a:start.c.obj",
+	                      "*/libretrocrt.a:relocate.c.obj",
+	                      "*/libretrocrt.a:MultiSegApp.c.obj",
+	                      "*/libretrocrt.a:LoadSeg.s.obj",
+	                      "*/libretrocrt.a:*",
+	                      "*/libgcc.a:*",
+	                      "*/libc.a:*"
+	                      );
+	segments.emplace_back(3, "libstdc++",
+	                      "*/libstdc++.a:*");
+	segments.emplace_back(4, "RetroConsole",
+	                      "*/libRetroConsole.a:*");
+
+	segments.emplace_back(2, "Main",
+	                      "*");
+}

Elf2Mac/SegmentMap.h

@@ -0,0 +1,35 @@
+#ifndef SEGMENTMAP_H
+#define SEGMENTMAP_H
+
+#include <vector>
+#include <string>
+
+class SegmentInfo
+{
+public:
+	int id;
+	//std::string name;
+	std::vector<std::string> filters;
+	SegmentInfo();
+
+	template<typename... Args>
+	SegmentInfo(int id, std::string name, Args... args)
+	    : id(id), filters { args... }
+	{
+	}
+
+	void WriteFilters(std::ostream& out, std::string section);
+	void WriteFiltersKeep(std::ostream& out, std::string section);
+	void CreateLdScript(std::ostream& out);
+};
+
+class SegmentMap
+{
+	std::vector<SegmentInfo> segments;
+public:
+	SegmentMap();
+
+	void CreateLdScript(std::ostream& out);
+};
+
+#endif // SEGMENTMAP_H

libretro/MultiSegApp.c

@@ -49,18 +49,7 @@ typedef struct CODEHeader
     uint32_t reserved;
 } CODEHeader;
 
-/*
+#define StripAddressCompat(p) (relocState.hasStripAddr ? StripAddress(p) : StripAddress24(p))
-   struct object is an internal data structure in libgcc.
-   Comments in unwind-dw2-fde.h imply that it will not
-   increase in size.
- */
-struct object { long space[8]; };
-
-extern void __register_frame_info (const void *, struct object *)
-				  __attribute__ ((weak));
-extern void *__deregister_frame_info (const void *)
-				     __attribute__ ((weak));
-
 
 pascal void* Retro68LoadSegment(uint8_t *p)
 {
@@ -73,12 +62,12 @@ pascal void* Retro68LoadSegment(uint8_t *p)
     Handle CODE = GetResource('CODE', id);
     HLock(CODE);
     
-    uint8_t *base = StripAddress((uint8_t *)*CODE);
+    uint8_t *base = StripAddressCompat((uint8_t *)*CODE);
     CODEHeader *header = (CODEHeader*) base;
     uint32_t codeSize = GetHandleSize(CODE);
     
         // TODO: StripAddress24
-    uint8_t * a5 = (uint8_t*) StripAddress((void*)SetCurrentA5());
+    uint8_t * a5 = (uint8_t*) StripAddressCompat((void*)SetCurrentA5());
     
     if(header->loadAddress != base || header->currentA5 != a5)
     {
@@ -114,7 +103,8 @@ pascal void* Retro68LoadSegment(uint8_t *p)
         ++jtEntry;
     }
     
-    // TODO: Flush cache
+    if(relocState.needFlushCache)
+        FlushCodeCache();
     
     /* Load Exception Information */
     if (__register_frame_info)
@@ -144,7 +134,7 @@ extern uint8_t _stext, _etext, _sdata, _edata, _sbss, _ebss;
 
 void Retro68InitMultisegApp()
 {
-    uint8_t * a5 = (uint8_t*) StripAddress((void*)SetCurrentA5());
+    uint8_t * a5 = (uint8_t*) StripAddressCompat((void*)SetCurrentA5());
 
     // CODE Segment 1 is already loaded - we are in it.
     // Update the jump table addresses.

libretro/Retro68Runtime.h

@@ -1,5 +1,5 @@
 /*
-	Copyright 2015 Wolfgang Thaller.
+	Copyright 2017 Wolfgang Thaller.
 
 	This file is part of Retro68.
 
@@ -24,6 +24,7 @@
 */
 
 #include <stdint.h>
+#include <Types.h>
 
 #define _RETRO68_GET_DISPLACEMENT(DISPLACEMENT, STRIP) \
 	do {	\
@@ -69,3 +70,27 @@ void Retro68InitMultisegApp();
 void Retro68ApplyRelocations(uint8_t *base, uint32_t size, void *relocations, uint32_t displacements[]);
 
 #define RETRO68_RELOCATE() RETRO68_CALL_UNRELOCATED(Retro68Relocate,())
+
+
+
+/*
+   struct object is an internal data structure in libgcc.
+   Comments in unwind-dw2-fde.h imply that it will not
+   increase in size.
+ */
+struct object { long space[8]; };
+
+extern void __register_frame_info (const void *, struct object *)
+				  __attribute__ ((weak));
+extern void *__deregister_frame_info (const void *)
+				     __attribute__ ((weak));
+
+typedef struct Retro68RelocState
+{
+	Ptr bssPtr;
+	Handle codeHandle;
+        char hasStripAddr;
+        char needFlushCache;
+} Retro68RelocState;
+
+extern Retro68RelocState relocState;

libretro/relocate.c

@@ -59,27 +59,9 @@ extern uint8_t _rsrc_start;
 extern voidFunction __CTOR_LIST__, __DTOR_LIST__;
 extern uint8_t __EH_FRAME_BEGIN__;
 
-/*
-   struct object is an internal data structure in libgcc.
-   Comments in unwind-dw2-fde.h imply that it will not
-   increase in size.
- */
-struct object { long space[8]; };
 
-extern void __register_frame_info (const void *, struct object *)
+Retro68RelocState relocState __attribute__ ((section(".relocvars"))) = {
-				  __attribute__ ((weak));
+	NULL, NULL, false, false
-extern void *__deregister_frame_info (const void *)
-				     __attribute__ ((weak));
-
-
-typedef struct Retro68RelocState
-{
-	Ptr bssPtr;
-	Handle codeHandle;
-} Retro68RelocState;
-
-static Retro68RelocState relocState __attribute__ ((section(".relocvars"))) = {
-	NULL, NULL
 };
 
 
@@ -162,8 +144,8 @@ void Retro68Relocate()
 
 	struct Retro68RelocState *rState = (Retro68RelocState*)
 			((char*)&relocState + displacement);
-			
 	// rState now points to the global relocState variable
+	// 
 	if(displacement == 0)
 	{
 		if(rState->bssPtr)
@@ -179,6 +161,8 @@ void Retro68Relocate()
 		}
 	}
         
+        rState->hasStripAddr = hasStripAddr;
+
 	// Locate the start of the FLT file header inside the code resource
 	uint8_t *orig_stext, *orig_etext, *orig_sdata, *orig_edata, *orig_sbss, *orig_ebss;
 	
@@ -256,7 +240,7 @@ void Retro68Relocate()
 	void *reloc;
 	Handle RELA = NULL;
 	uint32_t relocatableSize;
-	if(&_MULTISEG_APP)
+	if(&_MULTISEG_APP == (uint8_t*)1)
 	{
 		RELA = Get1Resource('RELA', 1);
 		assert(RELA);
@@ -289,6 +273,7 @@ void Retro68Relocate()
 		SysEnvirons(0, &env);
 		if(env.processor >= env68040)
 		{
+                        rState->needFlushCache = true;
 			FlushCodeCache();
 		}
 	}
@@ -355,4 +340,3 @@ void Retro68FreeGlobals()
 		relocState.bssPtr = (Ptr) -1;
 	}
 }
-