[PATCH] uml: maintain own LDT entries

Patch imlements full LDT handling in SKAS:
 * UML holds it's own LDT table, used to deliver data on
   modify_ldt(READ)
 * UML disables the default_ldt, inherited from the host (SKAS3)
   or resets LDT entries, set by host's clib and inherited in
   SKAS0
 * A new global variable skas_needs_stub is inserted, that
   can be used to decide, whether stub-pages must be supported
   or not.
 * Uses the syscall-stub to replace missing PTRACE_LDT (therefore,
   write_ldt_entry needs to be modified)
Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

arch/um/kernel/skas/include/mmu-skas.h

@@ -8,6 +8,7 @@
 
 #include "linux/config.h"
 #include "mm_id.h"
+#include "asm/ldt.h"
 
 struct mmu_context_skas {
 	struct mm_id id;
@@ -15,6 +16,7 @@ struct mmu_context_skas {
 #ifdef CONFIG_3_LEVEL_PGTABLES
         unsigned long last_pmd;
 #endif
+	uml_ldt_t ldt;
 };
 
 extern void switch_mm_skas(struct mm_id * mm_idp);

arch/um/kernel/skas/include/skas.h

@@ -10,7 +10,8 @@
 #include "sysdep/ptrace.h"
 
 extern int userspace_pid[];
-extern int proc_mm, ptrace_faultinfo;
+extern int proc_mm, ptrace_faultinfo, ptrace_ldt;
+extern int skas_needs_stub;
 
 extern void switch_threads(void *me, void *next);
 extern void thread_wait(void *sw, void *fb);

arch/um/kernel/skas/mem.c

@@ -20,7 +20,7 @@ unsigned long set_task_sizes_skas(int arg, unsigned long *host_size_out,
 	*task_size_out = CONFIG_HOST_TASK_SIZE;
 #else
 	*host_size_out = top;
-	if (proc_mm && ptrace_faultinfo)
+	if (!skas_needs_stub)
 		*task_size_out = top;
 	else *task_size_out = CONFIG_STUB_START & PGDIR_MASK;
 #endif

arch/um/kernel/skas/mmu.c

@@ -15,6 +15,7 @@
 #include "asm/mmu.h"
 #include "asm/pgalloc.h"
 #include "asm/pgtable.h"
+#include "asm/ldt.h"
 #include "os.h"
 #include "skas.h"
 
@@ -74,13 +75,12 @@ static int init_stub_pte(struct mm_struct *mm, unsigned long proc,
 
 int init_new_context_skas(struct task_struct *task, struct mm_struct *mm)
 {
-	struct mm_struct *cur_mm = current->mm;
-	struct mm_id *cur_mm_id = &cur_mm->context.skas.id;
-	struct mm_id *mm_id = &mm->context.skas.id;
+ 	struct mmu_context_skas *from_mm = NULL;
+	struct mmu_context_skas *to_mm = &mm->context.skas;
 	unsigned long stack = 0;
-	int from, ret = -ENOMEM;
+	int from_fd, ret = -ENOMEM;
 
-	if(!proc_mm || !ptrace_faultinfo){
+	if(skas_needs_stub){
 		stack = get_zeroed_page(GFP_KERNEL);
 		if(stack == 0)
 			goto out;
@@ -102,33 +102,43 @@ int init_new_context_skas(struct task_struct *task, struct mm_struct *mm)
 
 		mm->nr_ptes--;
 	}
-	mm_id->stack = stack;
+
+	to_mm->id.stack = stack;
+	if(current->mm != NULL && current->mm != &init_mm)
+		from_mm = &current->mm->context.skas;
 
 	if(proc_mm){
-		if((cur_mm != NULL) && (cur_mm != &init_mm))
-			from = cur_mm_id->u.mm_fd;
-		else from = -1;
+		if(from_mm)
+			from_fd = from_mm->id.u.mm_fd;
+		else from_fd = -1;
 
-		ret = new_mm(from, stack);
+		ret = new_mm(from_fd, stack);
 		if(ret < 0){
 			printk("init_new_context_skas - new_mm failed, "
 			       "errno = %d\n", ret);
 			goto out_free;
 		}
-		mm_id->u.mm_fd = ret;
+		to_mm->id.u.mm_fd = ret;
 	}
 	else {
-		if((cur_mm != NULL) && (cur_mm != &init_mm))
-			mm_id->u.pid = copy_context_skas0(stack,
-							  cur_mm_id->u.pid);
-		else mm_id->u.pid = start_userspace(stack);
+		if(from_mm)
+			to_mm->id.u.pid = copy_context_skas0(stack,
+							     from_mm->id.u.pid);
+		else to_mm->id.u.pid = start_userspace(stack);
+	}
+
+	ret = init_new_ldt(to_mm, from_mm);
+	if(ret < 0){
+		printk("init_new_context_skas - init_ldt"
+		       " failed, errno = %d\n", ret);
+		goto out_free;
 	}
 
 	return 0;
 
  out_free:
-	if(mm_id->stack != 0)
-		free_page(mm_id->stack);
+	if(to_mm->id.stack != 0)
+		free_page(to_mm->id.stack);
  out:
 	return ret;
 }

arch/um/kernel/skas/process.c

@@ -381,9 +381,9 @@ int copy_context_skas0(unsigned long new_stack, int pid)
 }
 
 /*
- * This is used only, if proc_mm is available, while PTRACE_FAULTINFO
- * isn't. Opening /proc/mm creates a new mm_context, which lacks the stub-pages
- * Thus, we map them using /proc/mm-fd
+ * This is used only, if stub pages are needed, while proc_mm is
+ * availabl. Opening /proc/mm creates a new mm_context, which lacks
+ * the stub-pages. Thus, we map them using /proc/mm-fd
  */
 void map_stub_pages(int fd, unsigned long code,
 		    unsigned long data, unsigned long stack)

arch/um/kernel/skas/process_kern.c

@@ -145,7 +145,7 @@ int new_mm(int from, unsigned long stack)
 			       "err = %d\n", -n);
 	}
 
-	if(!ptrace_faultinfo)
+	if(skas_needs_stub)
 		map_stub_pages(fd, CONFIG_STUB_CODE, CONFIG_STUB_DATA, stack);
 
 	return(fd);

arch/um/os-Linux/start_up.c

@@ -135,7 +135,9 @@ static int stop_ptraced_child(int pid, void *stack, int exitcode,
 }
 
 int ptrace_faultinfo = 1;
+int ptrace_ldt = 1;
 int proc_mm = 1;
+int skas_needs_stub = 0;
 
 static int __init skas0_cmd_param(char *str, int* add)
 {
@@ -352,14 +354,26 @@ __uml_setup("noptracefaultinfo", noptracefaultinfo_cmd_param,
 "    it. To support PTRACE_FAULTINFO, the host needs to be patched\n"
 "    using the current skas3 patch.\n\n");
 
+static int __init noptraceldt_cmd_param(char *str, int* add)
+{
+	ptrace_ldt = 0;
+	return 0;
+}
+
+__uml_setup("noptraceldt", noptraceldt_cmd_param,
+"noptraceldt\n"
+"    Turns off usage of PTRACE_LDT, even if host supports it.\n"
+"    To support PTRACE_LDT, the host needs to be patched using\n"
+"    the current skas3 patch.\n\n");
+
 #ifdef UML_CONFIG_MODE_SKAS
-static inline void check_skas3_ptrace_support(void)
+static inline void check_skas3_ptrace_faultinfo(void)
 {
 	struct ptrace_faultinfo fi;
 	void *stack;
 	int pid, n;
 
-	printf("Checking for the skas3 patch in the host...");
+	printf("  - PTRACE_FAULTINFO...");
 	pid = start_ptraced_child(&stack);
 
 	n = ptrace(PTRACE_FAULTINFO, pid, 0, &fi);
@@ -381,9 +395,49 @@ static inline void check_skas3_ptrace_support(void)
 	stop_ptraced_child(pid, stack, 1, 1);
 }
 
-int can_do_skas(void)
+static inline void check_skas3_ptrace_ldt(void)
+{
+#ifdef PTRACE_LDT
+	void *stack;
+	int pid, n;
+	unsigned char ldtbuf[40];
+	struct ptrace_ldt ldt_op = (struct ptrace_ldt) {
+		.func = 2, /* read default ldt */
+		.ptr = ldtbuf,
+		.bytecount = sizeof(ldtbuf)};
+
+	printf("  - PTRACE_LDT...");
+	pid = start_ptraced_child(&stack);
+
+	n = ptrace(PTRACE_LDT, pid, 0, (unsigned long) &ldt_op);
+	if (n < 0) {
+		if(errno == EIO)
+			printf("not found\n");
+		else {
+			perror("not found");
+		}
+		ptrace_ldt = 0;
+	}
+	else {
+		if(ptrace_ldt)
+			printf("found\n");
+		else
+			printf("found, but use is disabled\n");
+	}
+
+	stop_ptraced_child(pid, stack, 1, 1);
+#else
+	/* PTRACE_LDT might be disabled via cmdline option.
+	 * We want to override this, else we might use the stub
+	 * without real need
+	 */
+	ptrace_ldt = 1;
+#endif
+}
+
+static inline void check_skas3_proc_mm(void)
 {
-	printf("Checking for /proc/mm...");
+	printf("  - /proc/mm...");
 	if (os_access("/proc/mm", OS_ACC_W_OK) < 0) {
  		proc_mm = 0;
 		printf("not found\n");
@@ -394,8 +448,19 @@ int can_do_skas(void)
 		else
 			printf("found\n");
 	}
+}
+
+int can_do_skas(void)
+{
+	printf("Checking for the skas3 patch in the host:\n");
+
+	check_skas3_proc_mm();
+	check_skas3_ptrace_faultinfo();
+	check_skas3_ptrace_ldt();
+
+	if(!proc_mm || !ptrace_faultinfo || !ptrace_ldt)
+		skas_needs_stub = 1;
 
-	check_skas3_ptrace_support();
 	return 1;
 }
 #else

arch/um/scripts/Makefile.rules

@@ -26,8 +26,13 @@ define unprofile
 	$(patsubst -pg,,$(patsubst -fprofile-arcs -ftest-coverage,,$(1)))
 endef
 
+# cmd_make_link checks to see if the $(foo-dir) variable starts with a /.  If
+# so, it's considered to be a path relative to $(srcdir) rather than
+# $(srcdir)/arch/$(SUBARCH).  This is because x86_64 wants to get ldt.c from
+# arch/um/sys-i386 rather than arch/i386 like the other borrowed files.  So,
+# it sets $(ldt.c-dir) to /arch/um/sys-i386.
 quiet_cmd_make_link = SYMLINK $@
-cmd_make_link       = ln -sf $(srctree)/arch/$(SUBARCH)/$($(notdir $@)-dir)/$(notdir $@) $@
+cmd_make_link       = rm -f $@; ln -sf $(srctree)$(if $(filter-out /%,$($(notdir $@)-dir)),/arch/$(SUBARCH))/$($(notdir $@)-dir)/$(notdir $@) $@
 
 # this needs to be before the foreach, because targets does not accept
 # complete paths like $(obj)/$(f). To make sure this works, use a := assignment

arch/um/sys-x86_64/Makefile

@@ -5,7 +5,7 @@
 #
 
 #XXX: why into lib-y?
-lib-y = bitops.o bugs.o csum-partial.o delay.o fault.o mem.o memcpy.o \
+lib-y = bitops.o bugs.o csum-partial.o delay.o fault.o ldt.o mem.o memcpy.o \
 	ptrace.o ptrace_user.o sigcontext.o signal.o stub.o \
 	stub_segv.o syscalls.o syscall_table.o sysrq.o thunk.o
 
@@ -14,7 +14,7 @@ obj-$(CONFIG_MODULES) += module.o um_module.o
 
 USER_OBJS := ptrace_user.o sigcontext.o
 
-SYMLINKS = bitops.c csum-copy.S csum-partial.c csum-wrappers.c memcpy.S \
+SYMLINKS = bitops.c csum-copy.S csum-partial.c csum-wrappers.c ldt.c memcpy.S \
 	thunk.S module.c
 
 include arch/um/scripts/Makefile.rules
@@ -23,6 +23,7 @@ bitops.c-dir = lib
 csum-copy.S-dir = lib
 csum-partial.c-dir = lib
 csum-wrappers.c-dir = lib
+ldt.c-dir = /arch/um/sys-i386
 memcpy.S-dir = lib
 thunk.S-dir = lib
 module.c-dir = kernel

arch/um/sys-x86_64/syscalls.c

@@ -28,81 +28,6 @@ asmlinkage long sys_uname64(struct new_utsname __user * name)
 	return err ? -EFAULT : 0;
 }
 
-#ifdef CONFIG_MODE_TT
-extern int modify_ldt(int func, void *ptr, unsigned long bytecount);
-
-long sys_modify_ldt_tt(int func, void *ptr, unsigned long bytecount)
-{
-	/* XXX This should check VERIFY_WRITE depending on func, check this
-	 * in i386 as well.
-	 */
-	if (!access_ok(VERIFY_READ, ptr, bytecount))
-		return -EFAULT;
-	return(modify_ldt(func, ptr, bytecount));
-}
-#endif
-
-#ifdef CONFIG_MODE_SKAS
-extern int userspace_pid[];
-
-#include "skas_ptrace.h"
-
-long sys_modify_ldt_skas(int func, void *ptr, unsigned long bytecount)
-{
-	struct ptrace_ldt ldt;
-        void *buf;
-        int res, n;
-
-        buf = kmalloc(bytecount, GFP_KERNEL);
-        if(buf == NULL)
-                return(-ENOMEM);
-
-        res = 0;
-
-        switch(func){
-        case 1:
-        case 0x11:
-                res = copy_from_user(buf, ptr, bytecount);
-                break;
-        }
-
-        if(res != 0){
-                res = -EFAULT;
-                goto out;
-        }
-
-	ldt = ((struct ptrace_ldt) { .func	= func,
-				     .ptr	= buf,
-				     .bytecount = bytecount });
-#warning Need to look up userspace_pid by cpu
-	res = ptrace(PTRACE_LDT, userspace_pid[0], 0, (unsigned long) &ldt);
-        if(res < 0)
-                goto out;
-
-        switch(func){
-        case 0:
-        case 2:
-                n = res;
-                res = copy_to_user(ptr, buf, n);
-                if(res != 0)
-                        res = -EFAULT;
-                else
-                        res = n;
-                break;
-        }
-
- out:
-        kfree(buf);
-        return(res);
-}
-#endif
-
-long sys_modify_ldt(int func, void *ptr, unsigned long bytecount)
-{
-        return(CHOOSE_MODE_PROC(sys_modify_ldt_tt, sys_modify_ldt_skas, func,
-                                ptr, bytecount));
-}
-
 #ifdef CONFIG_MODE_TT
 extern long arch_prctl(int code, unsigned long addr);
 

include/asm-um/ldt-i386.h

+/*
+ * Copyright (C) 2004 Fujitsu Siemens Computers GmbH
+ * Licensed under the GPL
+ *
+ * Author: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
+ */
+
+#ifndef __ASM_LDT_I386_H
+#define __ASM_LDT_I386_H
+
+#include "asm/semaphore.h"
+#include "asm/arch/ldt.h"
+
+struct mmu_context_skas;
+extern void ldt_host_info(void);
+extern long init_new_ldt(struct mmu_context_skas * to_mm,
+			 struct mmu_context_skas * from_mm);
+extern void free_ldt(struct mmu_context_skas * mm);
+
+#define LDT_PAGES_MAX \
+	((LDT_ENTRIES * LDT_ENTRY_SIZE)/PAGE_SIZE)
+#define LDT_ENTRIES_PER_PAGE \
+	(PAGE_SIZE/LDT_ENTRY_SIZE)
+#define LDT_DIRECT_ENTRIES \
+	((LDT_PAGES_MAX*sizeof(void *))/LDT_ENTRY_SIZE)
+
+struct ldt_entry {
+	__u32 a;
+	__u32 b;
+};
+
+typedef struct uml_ldt {
+	int entry_count;
+	struct semaphore semaphore;
+	union {
+		struct ldt_entry * pages[LDT_PAGES_MAX];
+		struct ldt_entry entries[LDT_DIRECT_ENTRIES];
+	};
+} uml_ldt_t;
+
+/*
+ * macros stolen from include/asm-i386/desc.h
+ */
+#define LDT_entry_a(info) \
+	((((info)->base_addr & 0x0000ffff) << 16) | ((info)->limit & 0x0ffff))
+
+#define LDT_entry_b(info) \
+	(((info)->base_addr & 0xff000000) | \
+	(((info)->base_addr & 0x00ff0000) >> 16) | \
+	((info)->limit & 0xf0000) | \
+	(((info)->read_exec_only ^ 1) << 9) | \
+	((info)->contents << 10) | \
+	(((info)->seg_not_present ^ 1) << 15) | \
+	((info)->seg_32bit << 22) | \
+	((info)->limit_in_pages << 23) | \
+	((info)->useable << 20) | \
+	0x7000)
+
+#define LDT_empty(info) (\
+	(info)->base_addr	== 0	&& \
+	(info)->limit		== 0	&& \
+	(info)->contents	== 0	&& \
+	(info)->read_exec_only	== 1	&& \
+	(info)->seg_32bit	== 0	&& \
+	(info)->limit_in_pages	== 0	&& \
+	(info)->seg_not_present	== 1	&& \
+	(info)->useable		== 0	)
+
+#endif

include/asm-um/ldt.h

+/*
+ * Copyright (C) 2004 Fujitsu Siemens Computers GmbH
+ * Licensed under the GPL
+ *
+ * Author: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
+ */
+
+#ifndef __ASM_LDT_I386_H
+#define __ASM_LDT_I386_H
+
+#include "asm/semaphore.h"
+#include "asm/arch/ldt.h"
+
+struct mmu_context_skas;
+extern void ldt_host_info(void);
+extern long init_new_ldt(struct mmu_context_skas * to_mm,
+			 struct mmu_context_skas * from_mm);
+extern void free_ldt(struct mmu_context_skas * mm);
+
+#define LDT_PAGES_MAX \
+	((LDT_ENTRIES * LDT_ENTRY_SIZE)/PAGE_SIZE)
+#define LDT_ENTRIES_PER_PAGE \
+	(PAGE_SIZE/LDT_ENTRY_SIZE)
+#define LDT_DIRECT_ENTRIES \
+	((LDT_PAGES_MAX*sizeof(void *))/LDT_ENTRY_SIZE)
+
+struct ldt_entry {
+	__u32 a;
+	__u32 b;
+};
+
+typedef struct uml_ldt {
+	int entry_count;
+	struct semaphore semaphore;
+	union {
+		struct ldt_entry * pages[LDT_PAGES_MAX];
+		struct ldt_entry entries[LDT_DIRECT_ENTRIES];
+	};
+} uml_ldt_t;
+
+/*
+ * macros stolen from include/asm-i386/desc.h
+ */
+#define LDT_entry_a(info) \
+	((((info)->base_addr & 0x0000ffff) << 16) | ((info)->limit & 0x0ffff))
+
+#define LDT_entry_b(info) \
+	(((info)->base_addr & 0xff000000) | \
+	(((info)->base_addr & 0x00ff0000) >> 16) | \
+	((info)->limit & 0xf0000) | \
+	(((info)->read_exec_only ^ 1) << 9) | \
+	((info)->contents << 10) | \
+	(((info)->seg_not_present ^ 1) << 15) | \
+	((info)->seg_32bit << 22) | \
+	((info)->limit_in_pages << 23) | \
+	((info)->useable << 20) | \
+	0x7000)
+
+#define LDT_empty(info) (\
+	(info)->base_addr	== 0	&& \
+	(info)->limit		== 0	&& \
+	(info)->contents	== 0	&& \
+	(info)->read_exec_only	== 1	&& \
+	(info)->seg_32bit	== 0	&& \
+	(info)->limit_in_pages	== 0	&& \
+	(info)->seg_not_present	== 1	&& \
+	(info)->useable		== 0	)
+
+#endif
 #ifndef __UM_LDT_H
 #define __UM_LDT_H