Re: [PATCH 6/7] mm: kill vma flag VM_EXECUTABLE

From: Konstantin Khlebnikov
Date: Mon Apr 02 2012 - 13:14:49 EST


Cyrill Gorcunov wrote:
On Mon, Apr 02, 2012 at 08:19:59PM +0400, Konstantin Khlebnikov wrote:
Oleg Nesterov wrote:
On 04/02, Konstantin Khlebnikov wrote:

In this patch I leave mm->exe_file lockless.
After exec/fork we can change it only for current task and only if mm->mm_users == 1.

something like this:

task_lock(current);

OK, this protects against the race with get_task_mm()

if (atomic_read(&current->mm->mm_users) == 1)

this means PR_SET_MM_EXE_FILE can fail simply because someone did
get_task_mm(). Or the caller is multithreaded.

This is sad, seems like we should keep mm->exe_file protection by mm->mmap_sem.
So, I'll rework this patch...

Ah, it's about locking. I misundertand it at first.
Oleg, forget about my email then.

Yes, it's about locking. Please review patch for your code from attachment. diff --git a/include/linux/sched.h b/include/linux/sched.h
index cff94cd..4a41270 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -437,6 +437,7 @@ extern int get_dumpable(struct mm_struct *mm);
/* leave room for more dump flags */
#define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */
#define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */
+#define MMF_EXE_FILE_CHANGED 18 /* see prctl(PR_SET_MM_EXE_FILE) */

#define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK)

diff --git a/kernel/sys.c b/kernel/sys.c
index da660f3..b217069 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1714,17 +1714,11 @@ static bool vma_flags_mismatch(struct vm_area_struct *vma,

static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd)
{
+ struct vm_area_struct *vma;
struct file *exe_file;
struct dentry *dentry;
int err;

- /*
- * Setting new mm::exe_file is only allowed when no VM_EXECUTABLE vma's
- * remain. So perform a quick test first.
- */
- if (mm->num_exe_file_vmas)
- return -EBUSY;
-
exe_file = fget(fd);
if (!exe_file)
return -EBADF;
@@ -1745,17 +1739,28 @@ static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd)
if (err)
goto exit;

+ down_write(&mm->mmap_sem);
+ /*
+ * Forbid mm->exe_file change if there are mapped some other files.
+ */
+ err = -EEXIST;
+ for (vma = mm->mmap; vma; vma = vma->vm_next) {
+ if (vma->vm_file &&
+ !path_equal(&vma->vm_file->f_path, &exe_file->f_path))
+ goto out_unlock;
+ }
/*
* The symlink can be changed only once, just to disallow arbitrary
* transitions malicious software might bring in. This means one
* could make a snapshot over all processes running and monitor
* /proc/pid/exe changes to notice unusual activity if needed.
*/
- down_write(&mm->mmap_sem);
- if (likely(!mm->exe_file))
- set_mm_exe_file(mm, exe_file);
- else
- err = -EBUSY;
+ err = -EBUSY;
+ if (test_and_set_bit(MMF_EXE_FILE_CHANGED, &mm->flags))
+ goto out_unlock;
+ set_mm_exe_file(mm, exe_file);
+ err = 0;
+out_unlock:
up_write(&mm->mmap_sem);

exit: