Re: [patch] i386: handle_BUG(): don't print garbage if debug info unavailable

[Andrew Morton <akpm@osdl.org>]

Linus Torvalds <torvalds@osdl.org> wrote:
>
> That said, I think it's wrong to use "__get_user()", which can hang on the 
> MM semaphore if something is bogus. We should probably mark us as being 
> "in_atomic()" to make sure that the page fault handler, if it is entered, 
> will not try to get the semaphore or page anything in.

This?


--- a/include/linux/uaccess.h~add-probe_kernel_address
+++ a/include/linux/uaccess.h
@@ -19,4 +19,26 @@ static inline unsigned long __copy_from_
 
 #endif		/* ARCH_HAS_NOCACHE_UACCESS */
 
+/**
+ * probe_kernel_address(): safely attempt to read from a location
+ * @addr: address to read from - its type is type typeof(retval)*
+ * @retval: read into this variable
+ *
+ * Safely read from address @addr into variable @revtal.  If a kernel fault
+ * happens, handle that and return -EFAULT.
+ * We ensure that the __get_user() is executed in atomic context so that
+ * do_page_fault() doesn't attempt to take mmap_sem.  This makes
+ * probe_kernel_address() suitable for use within regions where the caller
+ * already holds mmap_sem, or other locks which nest inside mmap_sem.
+ */
+#define probe_kernel_address(addr, retval)		\
+	({						\
+		long ret;				\
+							\
+		inc_preempt_count();			\
+		ret = __get_user(retval, addr);		\
+		dec_preempt_count();			\
+		ret;					\
+	})
+
 #endif		/* __LINUX_UACCESS_H__ */
_