PARISC futex: special case cmpxchg NULL in kernel space
authorKyle McMartin <kyle@shortfin.cabal.ca>
Tue, 15 Apr 2008 15:45:11 +0000 (10:45 -0500)
committerChris Wright <chrisw@sous-sol.org>
Sat, 19 Apr 2008 01:53:29 +0000 (18:53 -0700)
upstream commit: c20a84c91048c76c1379011c96b1a5cee5c7d9a0

commit f9e77acd4060fefbb60a351cdb8d30fca27fe194
Author: Thomas Gleixner <tglx@linutronix.de>
Date:   Sun Feb 24 02:10:05 2008 +0000

    futex: runtime enable pi and robust functionality

which was backported to stable based on mainline Commit
a0c1e9073ef7428a14309cba010633a6cd6719ea added code to futex.c
to detect whether futex_atomic_cmpxchg_inatomic was implemented at run
time:

+       curval = cmpxchg_futex_value_locked(NULL, 0, 0);
+       if (curval == -EFAULT)
+               futex_cmpxchg_enabled = 1;

This is bogus on parisc, since page zero in kernel virtual space is the
gateway page for syscall entry, and should not be read from the kernel.
(That, and we really don't like the kernel faulting on its own address
 space...)

Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
include/asm-parisc/futex.h

index dbee6e60aa8139e66897ba65ba4ac9155bf00dab..fdc6d055ef7f2af68872f2f195d8b417ebb9558f 100644 (file)
@@ -56,6 +56,12 @@ futex_atomic_cmpxchg_inatomic(int __user *uaddr, int oldval, int newval)
        int err = 0;
        int uval;
 
+       /* futex.c wants to do a cmpxchg_inatomic on kernel NULL, which is
+        * our gateway page, and causes no end of trouble...
+        */
+       if (segment_eq(KERNEL_DS, get_fs()) && !uaddr)
+               return -EFAULT;
+
        if (!access_ok(VERIFY_WRITE, uaddr, sizeof(int)))
                return -EFAULT;
 
@@ -67,5 +73,5 @@ futex_atomic_cmpxchg_inatomic(int __user *uaddr, int oldval, int newval)
        return uval;
 }
 
-#endif
-#endif
+#endif /*__KERNEL__*/
+#endif /*_ASM_PARISC_FUTEX_H*/