[PATCH] x86_64: TASK_SIZE fixes for compatibility mode processes

Appended patch will setup compatibility mode TASK_SIZE properly.  This will
fix atleast three known bugs that can be encountered while running
compatibility mode apps.

a) A malicious 32bit app can have an elf section at 0xffffe000.  During
   exec of this app, we will have a memory leak as insert_vm_struct() is
   not checking for return value in syscall32_setup_pages() and thus not
   freeing the vma allocated for the vsyscall page.  And instead of exec
   failing (as it has addresses > TASK_SIZE), we were allowing it to
   succeed previously.

b) With a 32bit app, hugetlb_get_unmapped_area/arch_get_unmapped_area
   may return addresses beyond 32bits, ultimately causing corruption
   because of wrap-around and resulting in SEGFAULT, instead of returning
   ENOMEM.

c) 32bit app doing this below mmap will now fail.

  mmap((void *)(0xFFFFE000UL), 0x10000UL, PROT_READ|PROT_WRITE,
	MAP_FIXED|MAP_PRIVATE|MAP_ANON, 0, 0);

Signed-off-by: Zou Nan hai <nanhai.zou@intel.com>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

3 files changed, 16 insertions, 19 deletions

diff --git a/arch/x86_64/kernel/process.c b/arch/x86_64/kernel/process.c
index 761b6d3..dce8bab 100644
--- a/arch/x86_64/kernel/process.c
+++ b/arch/x86_64/kernel/process.c
@@ -656,7 +656,7 @@ long do_arch_prctl(struct task_struct *task, int code, unsigned long addr)
 
 	switch (code) { 
 	case ARCH_SET_GS:
-		if (addr >= TASK_SIZE) 
+		if (addr >= TASK_SIZE_OF(task))
 			return -EPERM; 
 		cpu = get_cpu();
 		/* handle small bases via the GDT because that's faster to 
@@ -682,7 +682,7 @@ long do_arch_prctl(struct task_struct *task, int code, unsigned long addr)
 	case ARCH_SET_FS:
 		/* Not strictly needed for fs, but do it for symmetry
 		   with gs */
-		if (addr >= TASK_SIZE)
+		if (addr >= TASK_SIZE_OF(task))
 			return -EPERM; 
 		cpu = get_cpu();
 		/* handle small bases via the GDT because that's faster to 
diff --git a/arch/x86_64/kernel/ptrace.c b/arch/x86_64/kernel/ptrace.c
index 525f6a1..bbf64b5 100644
--- a/arch/x86_64/kernel/ptrace.c
+++ b/arch/x86_64/kernel/ptrace.c
@@ -257,12 +257,12 @@ static int putreg(struct task_struct *child,
 			value &= 0xffff;
 			return 0;
 		case offsetof(struct user_regs_struct,fs_base):
-			if (value >= TASK_SIZE)
+			if (value >= TASK_SIZE_OF(child))
 				return -EIO;
 			child->thread.fs = value;
 			return 0;
 		case offsetof(struct user_regs_struct,gs_base):
-			if (value >= TASK_SIZE)
+			if (value >= TASK_SIZE_OF(child))
 				return -EIO;
 			child->thread.gs = value;
 			return 0;
@@ -279,7 +279,7 @@ static int putreg(struct task_struct *child,
 			break;
 		case offsetof(struct user_regs_struct, rip):
 			/* Check if the new RIP address is canonical */
-			if (value >= TASK_SIZE)
+			if (value >= TASK_SIZE_OF(child))
 				return -EIO;
 			break;
 	}
@@ -419,6 +419,8 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data
 		break;
 
 	case PTRACE_POKEUSR: /* write the word at location addr in the USER area */
+	{
+		int dsize = test_tsk_thread_flag(child, TIF_IA32) ? 3 : 7;
 		ret = -EIO;
 		if ((addr & 7) ||
 		    addr > sizeof(struct user) - 7)
@@ -430,22 +432,22 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data
 			break;
 		/* Disallows to set a breakpoint into the vsyscall */
 		case offsetof(struct user, u_debugreg[0]):
-			if (data >= TASK_SIZE-7) break;
+			if (data >= TASK_SIZE_OF(child) - dsize) break;
 			child->thread.debugreg0 = data;
 			ret = 0;
 			break;
 		case offsetof(struct user, u_debugreg[1]):
-			if (data >= TASK_SIZE-7) break;
+			if (data >= TASK_SIZE_OF(child) - dsize) break;
 			child->thread.debugreg1 = data;
 			ret = 0;
 			break;
 		case offsetof(struct user, u_debugreg[2]):
-			if (data >= TASK_SIZE-7) break;
+			if (data >= TASK_SIZE_OF(child) - dsize) break;
 			child->thread.debugreg2 = data;
 			ret = 0;
 			break;
 		case offsetof(struct user, u_debugreg[3]):
-			if (data >= TASK_SIZE-7) break;
+			if (data >= TASK_SIZE_OF(child) - dsize) break;
 			child->thread.debugreg3 = data;
 			ret = 0;
 			break;
@@ -469,6 +471,7 @@ asmlinkage long sys_ptrace(long request, long pid, unsigned long addr, long data
 		  break;
 		}
 		break;
+	}
 	case PTRACE_SYSCALL: /* continue and stop at next (return from) syscall */
 	case PTRACE_CONT:    /* restart after signal. */
 
diff --git a/arch/x86_64/kernel/sys_x86_64.c b/arch/x86_64/kernel/sys_x86_64.c
index dbebd5c..d9798dd 100644
--- a/arch/x86_64/kernel/sys_x86_64.c
+++ b/arch/x86_64/kernel/sys_x86_64.c
@@ -68,13 +68,7 @@ out:
 static void find_start_end(unsigned long flags, unsigned long *begin,
 			   unsigned long *end)
 {
-#ifdef CONFIG_IA32_EMULATION
-	if (test_thread_flag(TIF_IA32)) { 
-		*begin = TASK_UNMAPPED_32;
-		*end = IA32_PAGE_OFFSET; 
-	} else 
-#endif
-	if (flags & MAP_32BIT) { 
+	if (!test_thread_flag(TIF_IA32) && (flags & MAP_32BIT)) {
 		/* This is usually used needed to map code in small
 		   model, so it needs to be in the first 31bit. Limit
 		   it to that.  This means we need to move the
@@ -84,10 +78,10 @@ static void find_start_end(unsigned long flags, unsigned long *begin,
 		   of playground for now. -AK */ 
 		*begin = 0x40000000; 
 		*end = 0x80000000;		
-	} else { 
-		*begin = TASK_UNMAPPED_64; 
+	} else {
+		*begin = TASK_UNMAPPED_BASE;
 		*end = TASK_SIZE; 
-		}
+	}
 } 
 
 unsigned long