xv6-65oo2/Makefile

218 lines
5.3 KiB
Makefile
Raw Normal View History

2006-07-16 15:35:18 +00:00
OBJS = \
2019-05-31 13:45:59 +00:00
start.o \
console.o \
uart.o \
kalloc.o \
spinlock.o \
string.o \
main.o \
vm.o \
proc.o \
swtch.o \
trampoline.o \
trap.o \
syscall.o \
2019-05-31 15:45:42 +00:00
sysproc.o \
bio.o \
fs.o \
log.o \
sleeplock.o \
file.o \
pipe.o \
ramdisk.o \
exec.o \
sysfile.o \
kernelvec.o \
plic.o
2019-05-31 13:45:59 +00:00
# riscv64-unknown-elf- or riscv64-linux-gnu-
# perhaps in /opt/riscv/bin
#TOOLPREFIX =
2006-07-11 01:07:40 +00:00
2010-08-31 19:01:26 +00:00
# Try to infer the correct TOOLPREFIX if not set
ifndef TOOLPREFIX
2019-05-31 13:45:59 +00:00
TOOLPREFIX := $(shell if riscv64-unknown-elf-objdump -i 2>&1 | grep 'elf64-big' >/dev/null 2>&1; \
then echo 'riscv64-unknown-elf-'; \
elif riscv64-linux-gnu-objdump -i 2>&1 | grep 'elf64-big' >/dev/null 2>&1; \
then echo 'riscv64-linux-gnu-'; \
2010-08-31 19:01:26 +00:00
else echo "***" 1>&2; \
2019-05-31 13:45:59 +00:00
echo "*** Error: Couldn't find an riscv64 version of GCC/binutils." 1>&2; \
2010-08-31 19:01:26 +00:00
echo "*** To turn off this error, run 'gmake TOOLPREFIX= ...'." 1>&2; \
echo "***" 1>&2; exit 1; fi)
endif
2019-05-31 13:45:59 +00:00
QEMU = qemu-system-riscv64
2010-08-31 19:05:27 +00:00
2006-07-11 01:07:40 +00:00
CC = $(TOOLPREFIX)gcc
2007-08-28 04:20:40 +00:00
AS = $(TOOLPREFIX)gas
2006-07-11 01:07:40 +00:00
LD = $(TOOLPREFIX)ld
OBJCOPY = $(TOOLPREFIX)objcopy
OBJDUMP = $(TOOLPREFIX)objdump
Checkpoint port of xv6 to x86-64. Passed usertests on 2 processors a few times. The x86-64 doesn't just add two levels to page tables to support 64 bit addresses, but is a different processor. For example, calling conventions, system calls, and segmentation are different from 32-bit x86. Segmentation is basically gone, but gs/fs in combination with MSRs can be used to hold a per-core pointer. In general, x86-64 is more straightforward than 32-bit x86. The port uses code from sv6 and the xv6 "rsc-amd64" branch. A summary of the changes is as follows: - Booting: switch to grub instead of xv6's bootloader (pass -kernel to qemu), because xv6's boot loader doesn't understand 64bit ELF files. And, we don't care anymore about booting. - Makefile: use -m64 instead of -m32 flag for gcc, delete boot loader, xv6.img, bochs, and memfs. For now dont' use -O2, since usertests with -O2 is bigger than MAXFILE! - Update gdb.tmpl to be for i386 or x86-64 - Console/printf: use stdarg.h and treat 64-bit addresses different from ints (32-bit) - Update elfhdr to be 64 bit - entry.S/entryother.S: add code to switch to 64-bit mode: build a simple page table in 32-bit mode before switching to 64-bit mode, share code for entering boot processor and APs, and tweak boot gdt. The boot gdt is the gdt that the kernel proper also uses. (In 64-bit mode, the gdt/segmentation and task state mostly disappear.) - exec.c: fix passing argv (64-bit now instead of 32-bit). - initcode.c: use syscall instead of int. - kernel.ld: load kernel very high, in top terabyte. 64 bits is a lot of address space! - proc.c: initial return is through new syscall path instead of trapret. - proc.h: update struct cpu to have some scratch space since syscall saves less state than int, update struct context to reflect x86-64 calling conventions. - swtch: simplify for x86-64 calling conventions. - syscall: add fetcharg to handle x86-64 calling convetions (6 arguments are passed through registers), and fetchaddr to read a 64-bit value from user space. - sysfile: update to handle pointers from user space (e.g., sys_exec), which are 64 bits. - trap.c: no special trap vector for sys calls, because x86-64 has a different plan for system calls. - trapasm: one plan for syscalls and one plan for traps (interrupt and exceptions). On x86-64, the kernel is responsible for switching user/kernel stacks. To do, xv6 keeps some scratch space in the cpu structure, and uses MSR GS_KERN_BASE to point to the core's cpu structure (using swapgs). - types.h: add uint64, and change pde_t to uint64 - usertests: exit() when fork fails, which helped in tracking down one of the bugs in the switch from 32-bit to 64-bit - vectors: update to make them 64 bits - vm.c: use bootgdt in kernel too, program MSRs for syscalls and core-local state (for swapgs), walk 4 levels in walkpgdir, add DEVSPACETOP, use task segment to set kernel stack for interrupts (but simpler than in 32-bit mode), add an extra argument to freevm (size of user part of address space) to avoid checking all entries till KERNBASE (there are MANY TB before the top 1TB). - x86: update trapframe to have 64-bit entries, which is what the processor pushes on syscalls and traps. simplify lgdt and lidt, using struct desctr, which needs the gcc directives packed and aligned. TODO: - use int32 instead of int? - simplify curproc(). xv6 has per-cpu state again, but this time it must have it. - avoid repetition in walkpgdir - fix validateint() in usertests.c - fix bugs (e.g., observed one a case of entering kernel with invalid gs or proc
2018-09-23 12:24:42 +00:00
2019-06-05 18:31:13 +00:00
# CFLAGS = -fno-pic -static -fno-builtin -fno-strict-aliasing -Wall -MD -ggdb -Werror -fno-omit-frame-pointer -O
CFLAGS = -Wall -Werror -O -fno-omit-frame-pointer -ggdb
2019-06-05 18:31:13 +00:00
CFLAGS += -mcmodel=medany
2019-05-31 13:45:59 +00:00
CFLAGS += -ffreestanding -fno-common -nostdlib -mno-relax
2007-08-28 04:20:40 +00:00
CFLAGS += $(shell $(CC) -fno-stack-protector -E -x c /dev/null >/dev/null 2>&1 && echo -fno-stack-protector)
2006-06-12 15:22:12 +00:00
2018-08-30 18:09:03 +00:00
# Disable PIE when possible (for Ubuntu 16.10 toolchain)
ifneq ($(shell $(CC) -dumpspecs 2>/dev/null | grep -e '[^f]no-pie'),)
CFLAGS += -fno-pie -no-pie
endif
ifneq ($(shell $(CC) -dumpspecs 2>/dev/null | grep -e '[^f]nopie'),)
CFLAGS += -fno-pie -nopie
endif
2019-05-31 13:45:59 +00:00
LDFLAGS = -z max-page-size=4096
2019-05-31 15:45:42 +00:00
kernel: $(OBJS) entry.o kernel.ld initcode
2019-05-31 13:45:59 +00:00
$(LD) $(LDFLAGS) -T kernel.ld -o kernel entry.o $(OBJS)
Checkpoint port of xv6 to x86-64. Passed usertests on 2 processors a few times. The x86-64 doesn't just add two levels to page tables to support 64 bit addresses, but is a different processor. For example, calling conventions, system calls, and segmentation are different from 32-bit x86. Segmentation is basically gone, but gs/fs in combination with MSRs can be used to hold a per-core pointer. In general, x86-64 is more straightforward than 32-bit x86. The port uses code from sv6 and the xv6 "rsc-amd64" branch. A summary of the changes is as follows: - Booting: switch to grub instead of xv6's bootloader (pass -kernel to qemu), because xv6's boot loader doesn't understand 64bit ELF files. And, we don't care anymore about booting. - Makefile: use -m64 instead of -m32 flag for gcc, delete boot loader, xv6.img, bochs, and memfs. For now dont' use -O2, since usertests with -O2 is bigger than MAXFILE! - Update gdb.tmpl to be for i386 or x86-64 - Console/printf: use stdarg.h and treat 64-bit addresses different from ints (32-bit) - Update elfhdr to be 64 bit - entry.S/entryother.S: add code to switch to 64-bit mode: build a simple page table in 32-bit mode before switching to 64-bit mode, share code for entering boot processor and APs, and tweak boot gdt. The boot gdt is the gdt that the kernel proper also uses. (In 64-bit mode, the gdt/segmentation and task state mostly disappear.) - exec.c: fix passing argv (64-bit now instead of 32-bit). - initcode.c: use syscall instead of int. - kernel.ld: load kernel very high, in top terabyte. 64 bits is a lot of address space! - proc.c: initial return is through new syscall path instead of trapret. - proc.h: update struct cpu to have some scratch space since syscall saves less state than int, update struct context to reflect x86-64 calling conventions. - swtch: simplify for x86-64 calling conventions. - syscall: add fetcharg to handle x86-64 calling convetions (6 arguments are passed through registers), and fetchaddr to read a 64-bit value from user space. - sysfile: update to handle pointers from user space (e.g., sys_exec), which are 64 bits. - trap.c: no special trap vector for sys calls, because x86-64 has a different plan for system calls. - trapasm: one plan for syscalls and one plan for traps (interrupt and exceptions). On x86-64, the kernel is responsible for switching user/kernel stacks. To do, xv6 keeps some scratch space in the cpu structure, and uses MSR GS_KERN_BASE to point to the core's cpu structure (using swapgs). - types.h: add uint64, and change pde_t to uint64 - usertests: exit() when fork fails, which helped in tracking down one of the bugs in the switch from 32-bit to 64-bit - vectors: update to make them 64 bits - vm.c: use bootgdt in kernel too, program MSRs for syscalls and core-local state (for swapgs), walk 4 levels in walkpgdir, add DEVSPACETOP, use task segment to set kernel stack for interrupts (but simpler than in 32-bit mode), add an extra argument to freevm (size of user part of address space) to avoid checking all entries till KERNBASE (there are MANY TB before the top 1TB). - x86: update trapframe to have 64-bit entries, which is what the processor pushes on syscalls and traps. simplify lgdt and lidt, using struct desctr, which needs the gcc directives packed and aligned. TODO: - use int32 instead of int? - simplify curproc(). xv6 has per-cpu state again, but this time it must have it. - avoid repetition in walkpgdir - fix validateint() in usertests.c - fix bugs (e.g., observed one a case of entering kernel with invalid gs or proc
2018-09-23 12:24:42 +00:00
$(OBJDUMP) -S kernel > kernel.asm
$(OBJDUMP) -t kernel | sed '1,/SYMBOL TABLE/d; s/ .* / /; /^$$/d' > kernel.sym
2006-06-12 15:22:12 +00:00
2007-08-28 04:20:40 +00:00
initcode: initcode.S
$(CC) $(CFLAGS) -nostdinc -I. -c initcode.S
2019-05-31 15:45:42 +00:00
$(LD) $(LDFLAGS) -N -e start -Ttext 0 -o initcode.out initcode.o
$(OBJCOPY) -S -O binary initcode.out initcode
$(OBJDUMP) -S initcode.o > initcode.asm
2007-08-28 04:20:40 +00:00
tags: $(OBJS) entryother.S _init
etags *.S *.c
2007-08-28 04:20:40 +00:00
vectors.S: vectors.pl
./vectors.pl > vectors.S
2006-06-13 15:50:06 +00:00
ULIB = ulib.o usys.o printf.o umalloc.o
2006-07-11 01:07:40 +00:00
2007-08-28 04:20:40 +00:00
_%: %.o $(ULIB)
$(LD) $(LDFLAGS) -N -e main -Ttext 0 -o $@ $^
2007-08-28 04:20:40 +00:00
$(OBJDUMP) -S $@ > $*.asm
2007-08-28 12:48:33 +00:00
$(OBJDUMP) -t $@ | sed '1,/SYMBOL TABLE/d; s/ .* / /; /^$$/d' > $*.sym
2007-08-08 08:38:55 +00:00
2019-05-31 15:45:42 +00:00
usys.S : usys.pl
perl ./usys.pl > usys.S
2007-08-24 20:20:23 +00:00
_forktest: forktest.o $(ULIB)
# 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 _forktest forktest.o ulib.o usys.o
2007-08-24 20:20:23 +00:00
$(OBJDUMP) -S _forktest > forktest.asm
2007-08-28 04:20:40 +00:00
mkfs: mkfs.c fs.h
gcc -Werror -Wall -o mkfs mkfs.c
2007-08-28 04:20:40 +00:00
# Prevent deletion of intermediate files, e.g. cat.o, after first build, so
# that disk image changes after first build are persistent until clean. More
# details:
# http://www.gnu.org/software/make/manual/html_node/Chained-Rules.html
.PRECIOUS: %.o
2007-08-28 04:20:40 +00:00
UPROGS=\
_cat\
2007-08-28 04:26:34 +00:00
_echo\
2007-08-28 04:20:40 +00:00
_forktest\
2007-08-28 04:26:34 +00:00
_grep\
2007-08-28 04:20:40 +00:00
_init\
_kill\
_ln\
_ls\
_mkdir\
_rm\
_sh\
_stressfs\
2007-08-28 04:20:40 +00:00
_usertests\
_wc\
_zombie\
fs.img: mkfs README $(UPROGS)
2007-08-08 08:38:55 +00:00
./mkfs fs.img README $(UPROGS)
2006-06-15 19:58:01 +00:00
-include *.d
2007-08-28 04:20:40 +00:00
clean:
2007-08-30 18:33:48 +00:00
rm -f *.tex *.dvi *.idx *.aux *.log *.ind *.ilg \
2011-09-05 19:21:40 +00:00
*.o *.d *.asm *.sym vectors.S bootblock entryother \
Checkpoint port of xv6 to x86-64. Passed usertests on 2 processors a few times. The x86-64 doesn't just add two levels to page tables to support 64 bit addresses, but is a different processor. For example, calling conventions, system calls, and segmentation are different from 32-bit x86. Segmentation is basically gone, but gs/fs in combination with MSRs can be used to hold a per-core pointer. In general, x86-64 is more straightforward than 32-bit x86. The port uses code from sv6 and the xv6 "rsc-amd64" branch. A summary of the changes is as follows: - Booting: switch to grub instead of xv6's bootloader (pass -kernel to qemu), because xv6's boot loader doesn't understand 64bit ELF files. And, we don't care anymore about booting. - Makefile: use -m64 instead of -m32 flag for gcc, delete boot loader, xv6.img, bochs, and memfs. For now dont' use -O2, since usertests with -O2 is bigger than MAXFILE! - Update gdb.tmpl to be for i386 or x86-64 - Console/printf: use stdarg.h and treat 64-bit addresses different from ints (32-bit) - Update elfhdr to be 64 bit - entry.S/entryother.S: add code to switch to 64-bit mode: build a simple page table in 32-bit mode before switching to 64-bit mode, share code for entering boot processor and APs, and tweak boot gdt. The boot gdt is the gdt that the kernel proper also uses. (In 64-bit mode, the gdt/segmentation and task state mostly disappear.) - exec.c: fix passing argv (64-bit now instead of 32-bit). - initcode.c: use syscall instead of int. - kernel.ld: load kernel very high, in top terabyte. 64 bits is a lot of address space! - proc.c: initial return is through new syscall path instead of trapret. - proc.h: update struct cpu to have some scratch space since syscall saves less state than int, update struct context to reflect x86-64 calling conventions. - swtch: simplify for x86-64 calling conventions. - syscall: add fetcharg to handle x86-64 calling convetions (6 arguments are passed through registers), and fetchaddr to read a 64-bit value from user space. - sysfile: update to handle pointers from user space (e.g., sys_exec), which are 64 bits. - trap.c: no special trap vector for sys calls, because x86-64 has a different plan for system calls. - trapasm: one plan for syscalls and one plan for traps (interrupt and exceptions). On x86-64, the kernel is responsible for switching user/kernel stacks. To do, xv6 keeps some scratch space in the cpu structure, and uses MSR GS_KERN_BASE to point to the core's cpu structure (using swapgs). - types.h: add uint64, and change pde_t to uint64 - usertests: exit() when fork fails, which helped in tracking down one of the bugs in the switch from 32-bit to 64-bit - vectors: update to make them 64 bits - vm.c: use bootgdt in kernel too, program MSRs for syscalls and core-local state (for swapgs), walk 4 levels in walkpgdir, add DEVSPACETOP, use task segment to set kernel stack for interrupts (but simpler than in 32-bit mode), add an extra argument to freevm (size of user part of address space) to avoid checking all entries till KERNBASE (there are MANY TB before the top 1TB). - x86: update trapframe to have 64-bit entries, which is what the processor pushes on syscalls and traps. simplify lgdt and lidt, using struct desctr, which needs the gcc directives packed and aligned. TODO: - use int32 instead of int? - simplify curproc(). xv6 has per-cpu state again, but this time it must have it. - avoid repetition in walkpgdir - fix validateint() in usertests.c - fix bugs (e.g., observed one a case of entering kernel with invalid gs or proc
2018-09-23 12:24:42 +00:00
initcode initcode.out kernel fs.img kernelmemfs \
mkfs .gdbinit \
2007-08-08 08:38:55 +00:00
$(UPROGS)
2006-09-07 20:06:15 +00:00
# make a printout
2007-08-30 14:12:19 +00:00
FILES = $(shell grep -v '^\#' runoff.list)
2011-09-02 20:36:08 +00:00
PRINT = runoff.list runoff.spec README toc.hdr toc.ftr $(FILES)
2006-09-07 20:06:15 +00:00
2007-08-28 04:20:40 +00:00
xv6.pdf: $(PRINT)
2006-09-07 20:06:15 +00:00
./runoff
ls -l xv6.pdf
2006-09-07 20:06:15 +00:00
2007-08-28 04:20:40 +00:00
print: xv6.pdf
2006-09-07 20:06:15 +00:00
2009-09-15 21:15:36 +00:00
# try to generate a unique GDB port
GDBPORT = $(shell expr `id -u` % 5000 + 25000)
# QEMU's gdb stub command line changed in 0.11
2010-08-31 19:05:27 +00:00
QEMUGDB = $(shell if $(QEMU) -help | grep -q '^-gdb'; \
then echo "-gdb tcp::$(GDBPORT)"; \
else echo "-s -p $(GDBPORT)"; fi)
ifndef CPUS
CPUS := 3
endif
2019-05-31 13:45:59 +00:00
QEMUOPTS = -machine virt -kernel kernel -m 3G -smp $(CPUS) -nographic
2019-05-31 15:45:42 +00:00
QEMUOPTS += -initrd fs.img
2006-09-07 20:06:15 +00:00
2019-06-05 15:42:03 +00:00
qemu: kernel fs.img
2019-05-31 13:45:59 +00:00
$(QEMU) $(QEMUOPTS)
2009-09-15 21:15:36 +00:00
2019-05-31 13:45:59 +00:00
.gdbinit: .gdbinit.tmpl-riscv
sed "s/:1234/:$(GDBPORT)/" < $^ > $@
2009-09-15 21:15:36 +00:00
2019-06-05 15:42:03 +00:00
qemu-gdb: kernel .gdbinit fs.img
2009-09-15 21:15:36 +00:00
@echo "*** Now run 'gdb'." 1>&2
Checkpoint port of xv6 to x86-64. Passed usertests on 2 processors a few times. The x86-64 doesn't just add two levels to page tables to support 64 bit addresses, but is a different processor. For example, calling conventions, system calls, and segmentation are different from 32-bit x86. Segmentation is basically gone, but gs/fs in combination with MSRs can be used to hold a per-core pointer. In general, x86-64 is more straightforward than 32-bit x86. The port uses code from sv6 and the xv6 "rsc-amd64" branch. A summary of the changes is as follows: - Booting: switch to grub instead of xv6's bootloader (pass -kernel to qemu), because xv6's boot loader doesn't understand 64bit ELF files. And, we don't care anymore about booting. - Makefile: use -m64 instead of -m32 flag for gcc, delete boot loader, xv6.img, bochs, and memfs. For now dont' use -O2, since usertests with -O2 is bigger than MAXFILE! - Update gdb.tmpl to be for i386 or x86-64 - Console/printf: use stdarg.h and treat 64-bit addresses different from ints (32-bit) - Update elfhdr to be 64 bit - entry.S/entryother.S: add code to switch to 64-bit mode: build a simple page table in 32-bit mode before switching to 64-bit mode, share code for entering boot processor and APs, and tweak boot gdt. The boot gdt is the gdt that the kernel proper also uses. (In 64-bit mode, the gdt/segmentation and task state mostly disappear.) - exec.c: fix passing argv (64-bit now instead of 32-bit). - initcode.c: use syscall instead of int. - kernel.ld: load kernel very high, in top terabyte. 64 bits is a lot of address space! - proc.c: initial return is through new syscall path instead of trapret. - proc.h: update struct cpu to have some scratch space since syscall saves less state than int, update struct context to reflect x86-64 calling conventions. - swtch: simplify for x86-64 calling conventions. - syscall: add fetcharg to handle x86-64 calling convetions (6 arguments are passed through registers), and fetchaddr to read a 64-bit value from user space. - sysfile: update to handle pointers from user space (e.g., sys_exec), which are 64 bits. - trap.c: no special trap vector for sys calls, because x86-64 has a different plan for system calls. - trapasm: one plan for syscalls and one plan for traps (interrupt and exceptions). On x86-64, the kernel is responsible for switching user/kernel stacks. To do, xv6 keeps some scratch space in the cpu structure, and uses MSR GS_KERN_BASE to point to the core's cpu structure (using swapgs). - types.h: add uint64, and change pde_t to uint64 - usertests: exit() when fork fails, which helped in tracking down one of the bugs in the switch from 32-bit to 64-bit - vectors: update to make them 64 bits - vm.c: use bootgdt in kernel too, program MSRs for syscalls and core-local state (for swapgs), walk 4 levels in walkpgdir, add DEVSPACETOP, use task segment to set kernel stack for interrupts (but simpler than in 32-bit mode), add an extra argument to freevm (size of user part of address space) to avoid checking all entries till KERNBASE (there are MANY TB before the top 1TB). - x86: update trapframe to have 64-bit entries, which is what the processor pushes on syscalls and traps. simplify lgdt and lidt, using struct desctr, which needs the gcc directives packed and aligned. TODO: - use int32 instead of int? - simplify curproc(). xv6 has per-cpu state again, but this time it must have it. - avoid repetition in walkpgdir - fix validateint() in usertests.c - fix bugs (e.g., observed one a case of entering kernel with invalid gs or proc
2018-09-23 12:24:42 +00:00
$(QEMU) $(QEMUOPTS) -S $(QEMUGDB)
Checkpoint port of xv6 to x86-64. Passed usertests on 2 processors a few times. The x86-64 doesn't just add two levels to page tables to support 64 bit addresses, but is a different processor. For example, calling conventions, system calls, and segmentation are different from 32-bit x86. Segmentation is basically gone, but gs/fs in combination with MSRs can be used to hold a per-core pointer. In general, x86-64 is more straightforward than 32-bit x86. The port uses code from sv6 and the xv6 "rsc-amd64" branch. A summary of the changes is as follows: - Booting: switch to grub instead of xv6's bootloader (pass -kernel to qemu), because xv6's boot loader doesn't understand 64bit ELF files. And, we don't care anymore about booting. - Makefile: use -m64 instead of -m32 flag for gcc, delete boot loader, xv6.img, bochs, and memfs. For now dont' use -O2, since usertests with -O2 is bigger than MAXFILE! - Update gdb.tmpl to be for i386 or x86-64 - Console/printf: use stdarg.h and treat 64-bit addresses different from ints (32-bit) - Update elfhdr to be 64 bit - entry.S/entryother.S: add code to switch to 64-bit mode: build a simple page table in 32-bit mode before switching to 64-bit mode, share code for entering boot processor and APs, and tweak boot gdt. The boot gdt is the gdt that the kernel proper also uses. (In 64-bit mode, the gdt/segmentation and task state mostly disappear.) - exec.c: fix passing argv (64-bit now instead of 32-bit). - initcode.c: use syscall instead of int. - kernel.ld: load kernel very high, in top terabyte. 64 bits is a lot of address space! - proc.c: initial return is through new syscall path instead of trapret. - proc.h: update struct cpu to have some scratch space since syscall saves less state than int, update struct context to reflect x86-64 calling conventions. - swtch: simplify for x86-64 calling conventions. - syscall: add fetcharg to handle x86-64 calling convetions (6 arguments are passed through registers), and fetchaddr to read a 64-bit value from user space. - sysfile: update to handle pointers from user space (e.g., sys_exec), which are 64 bits. - trap.c: no special trap vector for sys calls, because x86-64 has a different plan for system calls. - trapasm: one plan for syscalls and one plan for traps (interrupt and exceptions). On x86-64, the kernel is responsible for switching user/kernel stacks. To do, xv6 keeps some scratch space in the cpu structure, and uses MSR GS_KERN_BASE to point to the core's cpu structure (using swapgs). - types.h: add uint64, and change pde_t to uint64 - usertests: exit() when fork fails, which helped in tracking down one of the bugs in the switch from 32-bit to 64-bit - vectors: update to make them 64 bits - vm.c: use bootgdt in kernel too, program MSRs for syscalls and core-local state (for swapgs), walk 4 levels in walkpgdir, add DEVSPACETOP, use task segment to set kernel stack for interrupts (but simpler than in 32-bit mode), add an extra argument to freevm (size of user part of address space) to avoid checking all entries till KERNBASE (there are MANY TB before the top 1TB). - x86: update trapframe to have 64-bit entries, which is what the processor pushes on syscalls and traps. simplify lgdt and lidt, using struct desctr, which needs the gcc directives packed and aligned. TODO: - use int32 instead of int? - simplify curproc(). xv6 has per-cpu state again, but this time it must have it. - avoid repetition in walkpgdir - fix validateint() in usertests.c - fix bugs (e.g., observed one a case of entering kernel with invalid gs or proc
2018-09-23 12:24:42 +00:00
qemu-nox-gdb: fs.img kernel .gdbinit
2009-10-01 02:09:48 +00:00
@echo "*** Now run 'gdb'." 1>&2
2010-08-31 19:05:27 +00:00
$(QEMU) -nographic $(QEMUOPTS) -S $(QEMUGDB)
2009-10-01 02:09:48 +00:00
2006-09-08 14:41:06 +00:00
# CUT HERE
# prepare dist for students
# after running make dist, probably want to
# rename it to rev0 or rev1 or so on and then
# check in that version.
2007-08-30 14:12:19 +00:00
EXTRA=\
2009-11-23 22:50:58 +00:00
mkfs.c ulib.c user.h cat.c echo.c forktest.c grep.c kill.c\
ln.c ls.c mkdir.c rm.c stressfs.c usertests.c wc.c zombie.c\
printf.c umalloc.c\
2007-08-30 14:12:19 +00:00
README dot-bochsrc *.pl toc.* runoff runoff1 runoff.list\
2009-11-23 22:50:58 +00:00
.gdbinit.tmpl gdbutil\
2007-08-30 14:12:19 +00:00
2007-08-28 04:20:40 +00:00
dist:
2006-09-08 14:41:06 +00:00
rm -rf dist
mkdir dist
2009-11-23 22:50:58 +00:00
for i in $(FILES); \
2006-09-08 14:41:06 +00:00
do \
grep -v PAGEBREAK $$i >dist/$$i; \
done
sed '/CUT HERE/,$$d' Makefile >dist/Makefile
2007-08-30 14:12:19 +00:00
echo >dist/runoff.spec
cp $(EXTRA) dist
2006-09-08 14:41:06 +00:00
2007-08-28 04:20:40 +00:00
dist-test:
2007-08-30 14:12:19 +00:00
rm -rf dist
make dist
2006-09-08 14:41:06 +00:00
rm -rf dist-test
mkdir dist-test
cp dist/* dist-test
cd dist-test; $(MAKE) print
cd dist-test; $(MAKE) bochs || true
cd dist-test; $(MAKE) qemu
2006-09-08 14:41:06 +00:00
# update this rule (change rev#) when it is time to
2006-09-08 14:41:06 +00:00
# make a new revision.
2007-08-28 04:20:40 +00:00
tar:
2006-09-08 14:41:06 +00:00
rm -rf /tmp/xv6
mkdir -p /tmp/xv6
2009-09-15 23:15:59 +00:00
cp dist/* dist/.gdbinit.tmpl /tmp/xv6
(cd /tmp; tar cf - xv6) | gzip >xv6-rev10.tar.gz # the next one will be 10 (9/17)
.PHONY: dist-test dist