From: Andy Whitcroft <apw@shadowen.org>

If a fault in the kernel leads to an unexpected protection fault whilst in a
code path which holds mmap_sem we will deadlock in do_page_fault() while
trying to classify the fault.  By carefully testing the source of the fault we
can detect and OOPS on the vast majority of these, greatly enhancing diagnosis
of such bugs.


---

 25-akpm/arch/i386/mm/extable.c |   12 ++++++++++++
 25-akpm/arch/i386/mm/fault.c   |   22 +++++++++++++++++++++-
 2 files changed, 33 insertions(+), 1 deletion(-)

diff -puN arch/i386/mm/extable.c~ia32-fault-deadlock-fix arch/i386/mm/extable.c
--- 25/arch/i386/mm/extable.c~ia32-fault-deadlock-fix	Wed May 19 14:41:01 2004
+++ 25-akpm/arch/i386/mm/extable.c	Wed May 19 14:41:01 2004
@@ -34,3 +34,15 @@ int fixup_exception(struct pt_regs *regs
 
 	return 0;
 }
+
+int check_exception(struct pt_regs *regs)
+{
+	const struct exception_table_entry *fixup;
+
+	fixup = search_exception_tables(regs->eip);
+	if (fixup) {
+		return 1;
+	}
+
+	return 0;
+}
diff -puN arch/i386/mm/fault.c~ia32-fault-deadlock-fix arch/i386/mm/fault.c
--- 25/arch/i386/mm/fault.c~ia32-fault-deadlock-fix	Wed May 19 14:41:01 2004
+++ 25-akpm/arch/i386/mm/fault.c	Wed May 19 14:41:01 2004
@@ -197,6 +197,7 @@ static inline int is_prefetch(struct pt_
 } 
 
 asmlinkage void do_invalid_op(struct pt_regs *, unsigned long);
+int check_exception(struct pt_regs *regs);
 
 /*
  * This routine handles page faults.  It determines the address,
@@ -261,7 +262,26 @@ asmlinkage void do_page_fault(struct pt_
 	if (in_atomic() || !mm)
 		goto bad_area_nosemaphore;
 
-	down_read(&mm->mmap_sem);
+	/* When running in the kernel we expect faults to occur only to
+	 * addresses in user space.  All other faults represent errors in the
+	 * kernel and should generate an OOPS.  Unfortunatly, in the case of an
+	 * erroneous fault occuring in a code path which already holds mmap_sem
+	 * we will deadlock attempting to validate the fault against the
+	 * address space.  Luckily the kernel only validly references user
+	 * space from well defined areas of code, which are listed in the
+	 * exceptions table.
+	 *
+	 * As the vast majority of faults will be valid we will only perform
+	 * the source reference check when there is a possibilty of a deadlock.
+	 * Attempt to lock the address space, if we cannot we then validate the
+	 * source.  If this is invalid we can skip the address space check,
+	 * thus avoiding the deadlock.
+	 */
+	if (!down_read_trylock(&mm->mmap_sem)) {
+		if ((error_code & 4) == 0 && !check_exception(regs))
+			goto bad_area_nosemaphore;
+		down_read(&mm->mmap_sem);
+	}
 
 	vma = find_vma(mm, address);
 	if (!vma)

_