implement x2APIC support

Makefile

@@ -69,7 +69,7 @@ AS = $(TOOLPREFIX)gas
 LD = $(TOOLPREFIX)ld
 OBJCOPY = $(TOOLPREFIX)objcopy
 OBJDUMP = $(TOOLPREFIX)objdump
-CFLAGS = -fno-pic -static -fno-builtin -fno-strict-aliasing -Wall -MD -g -ggdb -fno-omit-frame-pointer
+CFLAGS = -fno-pic -static -fno-builtin -fno-strict-aliasing -Wall -MD -g -ggdb -fno-omit-frame-pointer -ffunction-sections
 CFLAGS += -ffreestanding -fno-common -nostdlib -Iinclude -gdwarf-2 $(XFLAGS) $(OPT)
 CFLAGS += $(shell $(CC) -fno-stack-protector -E -x c /dev/null >/dev/null 2>&1 && echo -fno-stack-protector)
 ASFLAGS = -fno-pic -gdwarf-2 -Wa,-divide -Iinclude $(XFLAGS)
@@ -167,21 +167,21 @@ ULIB = uobj/ulib.o uobj/usys.o uobj/printf.o uobj/umalloc.o uobj/string.o
 
 fs/bin/%: uobj/%.o $(ULIB)
 	@mkdir -p fs out fs/bin
-	$(LD) $(LDFLAGS) -N -e main -Ttext 0 -o $@ $^
+	$(LD) $(LDFLAGS) --gc-sections -N -e main -Ttext 0 -o $@ $^
 	$(OBJDUMP) -S $@ > out/$*.asm
 	$(OBJDUMP) -t $@ | sed '1,/SYMBOL TABLE/d; s/ .* / /; /^$$/d' > bin/$*.sym
 
 fs/%: uobj/%.o $(ULIB)
 	@mkdir -p fs out fs/bin
-	$(LD) $(LDFLAGS) -N -e main -Ttext 0 -o $@ $^
+	$(LD) $(LDFLAGS) --gc-sections -N -e main -Ttext 0 -o $@ $^
 	$(OBJDUMP) -S $@ > out/$*.asm
 	$(OBJDUMP) -t $@ | sed '1,/SYMBOL TABLE/d; s/ .* / /; /^$$/d' > bin/$*.sym
 
 fs/forktest: uobj/forktest.o $(ULIB)
 	@mkdir -p fs
 	# forktest has less library code linked in - needs to be small
 	# in order to be able to max out the proc table.
-	$(LD) $(LDFLAGS) -N -e main -Ttext 0 -o fs/forktest uobj/forktest.o uobj/ulib.o uobj/usys.o
+	$(LD) $(LDFLAGS) --gc-sections -N -e main -Ttext 0 -o fs/forktest uobj/forktest.o uobj/ulib.o uobj/usys.o
 	$(OBJDUMP) -S fs/forktest > out/forktest.asm
 
 out/mkfs: tools/mkfs.c include/vfs.h
@@ -236,12 +236,10 @@ fs.img: out/mkfs fs/LICENSE $(UPROGS) $(SUBPROGS)
 
 	dd if=/dev/zero of=bin/fs-ext2.img bs=1k count=2000
 	mkfs -t ext2 -i 1024 -b 1024 -F bin/fs-ext2.img
-	mkdir /tmp/loop
-	mount -o loop bin/fs-ext2.img /tmp/loop
-	cp -r ./fs/ /tmp/loop/
-	umount /tmp/loop
-
-
+	mkdir -p /tmp/loop
+	sudo mount -o loop bin/fs-ext2.img /tmp/loop
+	sudo cp -r ./fs/ /tmp/loop/
+	sudo umount /tmp/loop
 -include */*.d
 
 clean:
@@ -259,7 +257,7 @@ binaries : fs.img boot.img
 	cp -r fs ./bin/
 	cd ./bin/ && tar -xvzf Xv64.vmwarevm.tar.gz && rm Xv64.vmwarevm.tar.gz && cd ..
 	qemu-img convert boot.img -O vmdk bin/Xv64.vmwarevm/boot.vmdk
-	mkdir bin/artifacts
+	mkdir -p bin/artifacts
 	cp -r uobj bin/artifacts/
 	cp -r kobj bin/artifacts/
 	cp -r out bin/artifacts/

include/defs.h

@@ -78,8 +78,10 @@ int             cpunum(void);
 extern volatile uint*    lapic;
 void            lapiceoi(void);
 void            lapicinit(void);
-void            lapicstartap(uchar, uint);
 void            microdelay(int);
+void            lapicstartap(uchar, uint);
+void            lapicx2enable(uint);
+void            lapicsendipi(uint, uint);
 
 // log.c
 void            initlog(void);

include/mmu.h

@@ -1,6 +1,8 @@
 // This file contains definitions for the
 // x86 memory management unit (MMU).
 
+#include "x86_defs.h"
+
 // Eflags register
 #define FL_CF           0x00000001      // Carry Flag
 #define FL_PF           0x00000004      // Parity Flag
@@ -24,21 +26,6 @@
 #define FL_VIP          0x00100000      // Virtual Interrupt Pending
 #define FL_ID           0x00200000      // ID flag
 
-// Control Register flags
-#define CR0_PE          0x00000001      // Protection Enable
-#define CR0_MP          0x00000002      // Monitor coProcessor
-#define CR0_EM          0x00000004      // Emulation
-#define CR0_TS          0x00000008      // Task Switched
-#define CR0_ET          0x00000010      // Extension Type
-#define CR0_NE          0x00000020      // Numeric Errror
-#define CR0_WP          0x00010000      // Write Protect
-#define CR0_AM          0x00040000      // Alignment Mask
-#define CR0_NW          0x20000000      // Not Writethrough
-#define CR0_CD          0x40000000      // Cache Disable
-#define CR0_PG          0x80000000      // Paging
-
-#define CR4_PSE         0x00000010      // Page size extension
-
 #define SEG_KCODE 1  // kernel code
 #define SEG_KDATA 2  // kernel data+stack
 #define SEG_KCPU  3  // kernel per-cpu data

include/proc.h

@@ -18,6 +18,10 @@ struct cpu {
   // Cpu-local storage variables; see below
   void *local;
   struct proc *proc;
+
+  char name[50];               // cpuid name
+  char vendor[13];             // cpuid vendor
+  uint32 model;                // cpuid model
 };
 
 extern struct cpu cpus[NCPU];

include/types.h

@@ -6,7 +6,7 @@ typedef unsigned char  uchar;
 typedef unsigned char uint8;
 typedef unsigned short uint16;
 typedef unsigned int  uint32;
-typedef unsigned long uint64;
+typedef unsigned long long uint64;
 
 //explicit-length signed ints
 typedef char int8;

include/unix/stdint.h

@@ -4,10 +4,10 @@
 typedef char int8_t;
 typedef short int16_t;
 typedef int  int32_t;
-typedef long int64_t;
+typedef long long int64_t;
 
 
 typedef unsigned char uint8_t;
 typedef unsigned short uint16_t;
 typedef unsigned int  uint32_t;
-typedef unsigned long uint64_t;
+typedef unsigned long long uint64_t;

include/x86.h

@@ -180,7 +180,18 @@ static inline unsigned int amd64_xchg(volatile unsigned int *addr, unsigned long
 	unsigned int result;
 
 	// The + in "+m" denotes a read-modify-write operand.
-	asm volatile ("lock; xchgl %0, %1" :
+	asm volatile ("lock xchgl %0, %1" :
+	              "+m" (*addr), "=a" (result) :
+	              "1" (newval) :
+	              "cc");
+	return result;
+}
+
+static inline unsigned int amd64_xadd(volatile unsigned int *addr, unsigned long newval) {
+	unsigned int result;
+
+	// The + in "+m" denotes a read-modify-write operand.
+	asm volatile ("lock xaddl %1, %0" :
 	              "+m" (*addr), "=a" (result) :
 	              "1" (newval) :
 	              "cc");
@@ -203,6 +214,31 @@ static inline void amd64_cpuid(unsigned int ax, unsigned int *p) {
 	              :  "0" (ax));
 }
 
+static inline uint64 amd64_rdmsr(uint msr) {
+    uint32 low, high;
+    asm volatile ("rdmsr"
+		  : "=a" (low), "=d" (high)
+		  : "c" (msr));
+    return (uint64)low | ((uint64)high << 32);
+}
+
+static inline void amd64_wrmsr(uint msr, uint64 newval) {
+    uint32 low = newval;
+    uint32 high = newval >> 32;
+    asm volatile ("wrmsr"
+		  :
+		  : "a" (low), "d" (high), "c" (msr));
+}
+
+static inline uint64 amd64_rdtsc()
+{
+    uint32 low, high;
+
+    asm volatile("rdtsc" : "=a" (low), "=d" (high));
+    return (uint64)low | ((uint64)high << 32);
+}
+
+
 // lie about some register names in 64bit mode to avoid
 // clunky ifdefs in proc.c and trap.c.
 struct trapframe {