[PATCH] uml: skas0 stubs now check system call return values

Change syscall-stub's data to include a "expected retval".

Stub now checks syscalls retval and aborts execution of syscall list, if
retval != expected retval.

run_syscall_stub prints the data of the failed syscall, using the data pointer
and retval written by the stub to the beginning of the stack.

one_syscall_stub is removed, to simplify code, because only some instructions
are saved by one_syscall_stub, no host-syscall.

Using the stub with additional data (modify_ldt via stub)
is prepared also.
Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

arch/um/include/tlb.h

@@ -38,9 +38,9 @@ extern void mprotect_kernel_vm(int w);
 extern void force_flush_all(void);
 extern void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
                              unsigned long end_addr, int force,
-			     void *(*do_ops)(union mm_context *,
+			     int (*do_ops)(union mm_context *,
 					   struct host_vm_op *, int, int,
-					     void *));
+					   void **));
 extern int flush_tlb_kernel_range_common(unsigned long start,
 					 unsigned long end);
 

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

@@ -24,14 +24,14 @@ extern void new_thread_proc(void *stack, void (*handler)(int sig));
 extern void remove_sigstack(void);
 extern void new_thread_handler(int sig);
 extern void handle_syscall(union uml_pt_regs *regs);
-extern void *map(struct mm_id * mm_idp, unsigned long virt,
+extern int map(struct mm_id * mm_idp, unsigned long virt,
 	       unsigned long len, int r, int w, int x, int phys_fd,
-		 unsigned long long offset, int done, void *data);
-extern void *unmap(struct mm_id * mm_idp, void *addr,
-		   unsigned long len, int done, void *data);
-extern void *protect(struct mm_id * mm_idp, unsigned long addr,
+	       unsigned long long offset, int done, void **data);
+extern int unmap(struct mm_id * mm_idp, void *addr, unsigned long len,
+		 int done, void **data);
+extern int protect(struct mm_id * mm_idp, unsigned long addr,
 		   unsigned long len, int r, int w, int x, int done,
-		     void *data);
+		   void **data);
 extern void user_signal(int sig, union uml_pt_regs *regs, int pid);
 extern int new_mm(int from, unsigned long stack);
 extern int start_userspace(unsigned long stub_stack);
@@ -39,16 +39,11 @@ extern int copy_context_skas0(unsigned long stack, int pid);
 extern void get_skas_faultinfo(int pid, struct faultinfo * fi);
 extern long execute_syscall_skas(void *r);
 extern unsigned long current_stub_stack(void);
+extern long run_syscall_stub(struct mm_id * mm_idp,
+                             int syscall, unsigned long *args, long expected,
+                             void **addr, int done);
+extern long syscall_stub_data(struct mm_id * mm_idp,
+                              unsigned long *data, int data_count,
+                              void **addr, void **stub_addr);
 
 #endif
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */

arch/um/kernel/skas/tlb.c

@@ -18,27 +18,28 @@
 #include "os.h"
 #include "tlb.h"
 
-static void *do_ops(union mm_context *mmu, struct host_vm_op *ops, int last,
-		    int finished, void *flush)
+static int do_ops(union mm_context *mmu, struct host_vm_op *ops, int last,
+		  int finished, void **flush)
 {
 	struct host_vm_op *op;
-	int i;
+        int i, ret = 0;
 
-	for(i = 0; i <= last; i++){
+        for(i = 0; i <= last && !ret; i++){
 		op = &ops[i];
 		switch(op->type){
 		case MMAP:
-			flush = map(&mmu->skas.id, op->u.mmap.addr,
+			ret = map(&mmu->skas.id, op->u.mmap.addr,
 				  op->u.mmap.len, op->u.mmap.r, op->u.mmap.w,
 				  op->u.mmap.x, op->u.mmap.fd,
 				  op->u.mmap.offset, finished, flush);
 			break;
 		case MUNMAP:
-			flush = unmap(&mmu->skas.id, (void *) op->u.munmap.addr,
+			ret = unmap(&mmu->skas.id,
+				    (void *) op->u.munmap.addr,
 				    op->u.munmap.len, finished, flush);
 			break;
 		case MPROTECT:
-			flush = protect(&mmu->skas.id, op->u.mprotect.addr,
+			ret = protect(&mmu->skas.id, op->u.mprotect.addr,
 				      op->u.mprotect.len, op->u.mprotect.r,
 				      op->u.mprotect.w, op->u.mprotect.x,
 				      finished, flush);
@@ -49,7 +50,7 @@ static void *do_ops(union mm_context *mmu, struct host_vm_op *ops, int last,
 		}
 	}
 
-	return flush;
+	return ret;
 }
 
 extern int proc_mm;

arch/um/kernel/tlb.c

@@ -16,34 +16,34 @@
 #include "os.h"
 
 static int add_mmap(unsigned long virt, unsigned long phys, unsigned long len,
-		    int r, int w, int x, struct host_vm_op *ops, int index,
+ 		    int r, int w, int x, struct host_vm_op *ops, int *index,
 		    int last_filled, union mm_context *mmu, void **flush,
-		    void *(*do_ops)(union mm_context *, struct host_vm_op *,
-				    int, int, void *))
+		    int (*do_ops)(union mm_context *, struct host_vm_op *,
+				  int, int, void **))
 {
         __u64 offset;
 	struct host_vm_op *last;
-	int fd;
+	int fd, ret = 0;
 
 	fd = phys_mapping(phys, &offset);
-	if(index != -1){
-		last = &ops[index];
+	if(*index != -1){
+		last = &ops[*index];
 		if((last->type == MMAP) &&
 		   (last->u.mmap.addr + last->u.mmap.len == virt) &&
 		   (last->u.mmap.r == r) && (last->u.mmap.w == w) &&
 		   (last->u.mmap.x == x) && (last->u.mmap.fd == fd) &&
 		   (last->u.mmap.offset + last->u.mmap.len == offset)){
 			last->u.mmap.len += len;
-			return index;
+			return 0;
 		}
 	}
 
-	if(index == last_filled){
-		*flush = (*do_ops)(mmu, ops, last_filled, 0, *flush);
-		index = -1;
+	if(*index == last_filled){
+		ret = (*do_ops)(mmu, ops, last_filled, 0, flush);
+		*index = -1;
 	}
 
-	ops[++index] = ((struct host_vm_op) { .type	= MMAP,
+	ops[++*index] = ((struct host_vm_op) { .type	= MMAP,
 			     			.u = { .mmap = {
 						       .addr	= virt,
 						       .len	= len,
@@ -53,78 +53,80 @@ static int add_mmap(unsigned long virt, unsigned long phys, unsigned long len,
 						       .fd	= fd,
 						       .offset	= offset }
 			   } });
-	return index;
+	return ret;
 }
 
 static int add_munmap(unsigned long addr, unsigned long len,
-		      struct host_vm_op *ops, int index, int last_filled,
+		      struct host_vm_op *ops, int *index, int last_filled,
 		      union mm_context *mmu, void **flush,
-		      void *(*do_ops)(union mm_context *, struct host_vm_op *,
-				      int, int, void *))
+		      int (*do_ops)(union mm_context *, struct host_vm_op *,
+				    int, int, void **))
 {
 	struct host_vm_op *last;
+	int ret = 0;
 
-	if(index != -1){
-		last = &ops[index];
+	if(*index != -1){
+		last = &ops[*index];
 		if((last->type == MUNMAP) &&
 		   (last->u.munmap.addr + last->u.mmap.len == addr)){
 			last->u.munmap.len += len;
-			return index;
+			return 0;
 		}
 	}
 
-	if(index == last_filled){
-		*flush = (*do_ops)(mmu, ops, last_filled, 0, *flush);
-		index = -1;
+	if(*index == last_filled){
+		ret = (*do_ops)(mmu, ops, last_filled, 0, flush);
+		*index = -1;
 	}
 
-	ops[++index] = ((struct host_vm_op) { .type	= MUNMAP,
+	ops[++*index] = ((struct host_vm_op) { .type	= MUNMAP,
 			     		       .u = { .munmap = {
 						        .addr	= addr,
 							.len	= len } } });
-	return index;
+	return ret;
 }
 
 static int add_mprotect(unsigned long addr, unsigned long len, int r, int w,
-			int x, struct host_vm_op *ops, int index,
+			int x, struct host_vm_op *ops, int *index,
 			int last_filled, union mm_context *mmu, void **flush,
-			void *(*do_ops)(union mm_context *,
-				       struct host_vm_op *, int, int, void *))
+ 			int (*do_ops)(union mm_context *, struct host_vm_op *,
+				      int, int, void **))
 {
 	struct host_vm_op *last;
+	int ret = 0;
 
-	if(index != -1){
-		last = &ops[index];
+	if(*index != -1){
+		last = &ops[*index];
 		if((last->type == MPROTECT) &&
 		   (last->u.mprotect.addr + last->u.mprotect.len == addr) &&
 		   (last->u.mprotect.r == r) && (last->u.mprotect.w == w) &&
 		   (last->u.mprotect.x == x)){
 			last->u.mprotect.len += len;
-			return index;
+			return 0;
 		}
 	}
 
-	if(index == last_filled){
-		*flush = (*do_ops)(mmu, ops, last_filled, 0, *flush);
-		index = -1;
+	if(*index == last_filled){
+		ret = (*do_ops)(mmu, ops, last_filled, 0, flush);
+		*index = -1;
 	}
 
-	ops[++index] = ((struct host_vm_op) { .type	= MPROTECT,
+	ops[++*index] = ((struct host_vm_op) { .type	= MPROTECT,
 			     		       .u = { .mprotect = {
 						       .addr	= addr,
 						       .len	= len,
 						       .r	= r,
 						       .w	= w,
 						       .x	= x } } });
-	return index;
+	return ret;
 }
 
 #define ADD_ROUND(n, inc) (((n) + (inc)) & ~((inc) - 1))
 
 void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
                       unsigned long end_addr, int force,
-		      void *(*do_ops)(union mm_context *, struct host_vm_op *,
-				      int, int, void *))
+		      int (*do_ops)(union mm_context *, struct host_vm_op *,
+				    int, int, void **))
 {
         pgd_t *npgd;
         pud_t *npud;
@@ -136,19 +138,20 @@ void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
         struct host_vm_op ops[1];
         void *flush = NULL;
         int op_index = -1, last_op = sizeof(ops) / sizeof(ops[0]) - 1;
+        int ret = 0;
 
         if(mm == NULL) return;
 
         ops[0].type = NONE;
-        for(addr = start_addr; addr < end_addr;){
+        for(addr = start_addr; addr < end_addr && !ret;){
                 npgd = pgd_offset(mm, addr);
                 if(!pgd_present(*npgd)){
                         end = ADD_ROUND(addr, PGDIR_SIZE);
                         if(end > end_addr)
                                 end = end_addr;
                         if(force || pgd_newpage(*npgd)){
-                                op_index = add_munmap(addr, end - addr, ops,
-                                                      op_index, last_op, mmu,
+                                ret = add_munmap(addr, end - addr, ops,
+                                                 &op_index, last_op, mmu,
                                                  &flush, do_ops);
                                 pgd_mkuptodate(*npgd);
                         }
@@ -162,8 +165,8 @@ void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
                         if(end > end_addr)
                                 end = end_addr;
                         if(force || pud_newpage(*npud)){
-                                op_index = add_munmap(addr, end - addr, ops,
-                                                      op_index, last_op, mmu,
+                                ret = add_munmap(addr, end - addr, ops,
+                                                 &op_index, last_op, mmu,
                                                  &flush, do_ops);
                                 pud_mkuptodate(*npud);
                         }
@@ -177,8 +180,8 @@ void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
                         if(end > end_addr)
                                 end = end_addr;
                         if(force || pmd_newpage(*npmd)){
-                                op_index = add_munmap(addr, end - addr, ops,
-                                                      op_index, last_op, mmu,
+                                ret = add_munmap(addr, end - addr, ops,
+                                                 &op_index, last_op, mmu,
                                                  &flush, do_ops);
                                 pmd_mkuptodate(*npmd);
                         }
@@ -198,24 +201,32 @@ void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
                 }
                 if(force || pte_newpage(*npte)){
                         if(pte_present(*npte))
-                                op_index = add_mmap(addr,
+			  ret = add_mmap(addr,
 					 pte_val(*npte) & PAGE_MASK,
 					 PAGE_SIZE, r, w, x, ops,
-                                                    op_index, last_op, mmu,
+					 &op_index, last_op, mmu,
 					 &flush, do_ops);
-                        else op_index = add_munmap(addr, PAGE_SIZE, ops,
-                                                   op_index, last_op, mmu,
+			else ret = add_munmap(addr, PAGE_SIZE, ops,
+					      &op_index, last_op, mmu,
 					      &flush, do_ops);
                 }
                 else if(pte_newprot(*npte))
-                        op_index = add_mprotect(addr, PAGE_SIZE, r, w, x, ops,
-                                                op_index, last_op, mmu,
+			ret = add_mprotect(addr, PAGE_SIZE, r, w, x, ops,
+					   &op_index, last_op, mmu,
 					   &flush, do_ops);
 
                 *npte = pte_mkuptodate(*npte);
                 addr += PAGE_SIZE;
         }
-	flush = (*do_ops)(mmu, ops, op_index, 1, flush);
+
+	if(!ret)
+		ret = (*do_ops)(mmu, ops, op_index, 1, &flush);
+
+	/* This is not an else because ret is modified above */
+	if(ret) {
+		printk("fix_range_common: failed, killing current process\n");
+		force_sig(SIGKILL, current);
+	}
 }
 
 int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)

arch/um/kernel/tt/tlb.c

@@ -17,26 +17,31 @@
 #include "os.h"
 #include "tlb.h"
 
-static void *do_ops(union mm_context *mmu, struct host_vm_op *ops, int last,
-		    int finished, void *flush)
+static int do_ops(union mm_context *mmu, struct host_vm_op *ops, int last,
+		    int finished, void **flush)
 {
 	struct host_vm_op *op;
-	int i;
+        int i, ret=0;
 
-	for(i = 0; i <= last; i++){
+        for(i = 0; i <= last && !ret; i++){
 		op = &ops[i];
 		switch(op->type){
 		case MMAP:
-                        os_map_memory((void *) op->u.mmap.addr, op->u.mmap.fd,
-				      op->u.mmap.offset, op->u.mmap.len,
-				      op->u.mmap.r, op->u.mmap.w,
-				      op->u.mmap.x);
+                        ret = os_map_memory((void *) op->u.mmap.addr,
+                                            op->u.mmap.fd, op->u.mmap.offset,
+                                            op->u.mmap.len, op->u.mmap.r,
+                                            op->u.mmap.w, op->u.mmap.x);
 			break;
 		case MUNMAP:
-			os_unmap_memory((void *) op->u.munmap.addr,
+                        ret = os_unmap_memory((void *) op->u.munmap.addr,
                                               op->u.munmap.len);
 			break;
 		case MPROTECT:
+                        ret = protect_memory(op->u.mprotect.addr,
+                                             op->u.munmap.len,
+                                             op->u.mprotect.r,
+                                             op->u.mprotect.w,
+                                             op->u.mprotect.x, 1);
 			protect_memory(op->u.mprotect.addr, op->u.munmap.len,
 				       op->u.mprotect.r, op->u.mprotect.w,
 				       op->u.mprotect.x, 1);
@@ -47,7 +52,7 @@ static void *do_ops(union mm_context *mmu, struct host_vm_op *ops, int last,
 		}
 	}
 
-	return NULL;
+	return ret;
 }
 
 static void fix_range(struct mm_struct *mm, unsigned long start_addr, 

arch/um/sys-i386/stub.S

@@ -2,24 +2,50 @@
 
 	.globl syscall_stub
 .section .__syscall_stub, "x"
-syscall_stub:
-	int 	$0x80
-	mov	%eax, UML_CONFIG_STUB_DATA
-	int3
 
 	.globl batch_syscall_stub
 batch_syscall_stub:
-	mov	$UML_CONFIG_STUB_DATA, %esp
-again:	pop	%eax
+	/* load pointer to first operation */
+	mov	$(UML_CONFIG_STUB_DATA+8), %esp
+
+again:
+	/* load length of additional data */
+	mov	0x0(%esp), %eax
+
+	/* if(length == 0) : end of list */
+	/* write possible 0 to header */
+	mov	%eax, UML_CONFIG_STUB_DATA+4
 	cmpl	$0, %eax
 	jz	done
+
+	/* save current pointer */
+	mov	%esp, UML_CONFIG_STUB_DATA+4
+
+	/* skip additional data */
+	add	%eax, %esp
+
+	/* load syscall-# */
+	pop	%eax
+
+	/* load syscall params */
 	pop	%ebx
 	pop	%ecx
 	pop	%edx
 	pop	%esi
  	pop	%edi
 	pop	%ebp
+
+	/* execute syscall */
 	int	$0x80
+
+	/* check return value */
+	pop	%ebx
+	cmp	%ebx, %eax
+	je	again
+
+done:
+	/* save return value */
 	mov	%eax, UML_CONFIG_STUB_DATA
-	jmp	again
-done:	int3
+
+	/* stop */
+	int3

arch/um/sys-x86_64/stub.S

@@ -16,21 +16,51 @@ syscall_stub:
 
 	.globl batch_syscall_stub
 batch_syscall_stub:
-	movq	$(UML_CONFIG_STUB_DATA >> 32), %rbx
-	salq	$32, %rbx
-	movq	$(UML_CONFIG_STUB_DATA & 0xffffffff), %rcx
-	or	%rcx, %rbx
-	movq	%rbx, %rsp
-again:	pop	%rax
-	cmpq	$0, %rax
-jz	done
+	mov	$(UML_CONFIG_STUB_DATA >> 32), %rbx
+	sal	$32, %rbx
+	mov	$(UML_CONFIG_STUB_DATA & 0xffffffff), %rax
+	or	%rax, %rbx
+	/* load pointer to first operation */
+	mov	%rbx, %rsp
+	add	$0x10, %rsp
+again:
+	/* load length of additional data */
+	mov	0x0(%rsp), %rax
+
+	/* if(length == 0) : end of list */
+	/* write possible 0 to header */
+	mov	%rax, 8(%rbx)
+	cmp	$0, %rax
+	jz	done
+
+	/* save current pointer */
+	mov	%rsp, 8(%rbx)
+
+	/* skip additional data */
+	add	%rax, %rsp
+
+	/* load syscall-# */
+	pop	%rax
+
+	/* load syscall params */
 	pop	%rdi
 	pop	%rsi
 	pop	%rdx
 	pop	%r10
  	pop	%r8
 	pop	%r9
+
+	/* execute syscall */
 	syscall
+
+	/* check return value */
+	pop	%rcx
+	cmp	%rcx, %rax
+	je	again
+
+done:
+	/* save return value */
 	mov	%rax, (%rbx)
-	jmp	again
-done:	int3
+
+	/* stop */
+	int3