Revision e91467ecd1ef381377fd327c0ded922835ec52ab authored by Christian Borntraeger on 05 August 2006, 19:13:52 UTC, committed by Linus Torvalds on 06 August 2006, 15:57:46 UTC
This patch adds a barrier() in futex unqueue_me to avoid aliasing of two
pointers.

On my s390x system I saw the following oops:

Unable to handle kernel pointer dereference at virtual kernel address
0000000000000000
Oops: 0004 [#1]
CPU:    0    Not tainted
Process mytool (pid: 13613, task: 000000003ecb6ac0, ksp: 00000000366bdbd8)
Krnl PSW : 0704d00180000000 00000000003c9ac2 (_spin_lock+0xe/0x30)
Krnl GPRS: 00000000ffffffff 000000003ecb6ac0 0000000000000000 0700000000000000
           0000000000000000 0000000000000000 000001fe00002028 00000000000c091f
           000001fe00002054 000001fe00002054 0000000000000000 00000000366bddc0
           00000000005ef8c0 00000000003d00e8 0000000000144f91 00000000366bdcb8
Krnl Code: ba 4e 20 00 12 44 b9 16 00 3e a7 84 00 08 e3 e0 f0 88 00 04
Call Trace:
([<0000000000144f90>] unqueue_me+0x40/0xe4)
 [<0000000000145a0c>] do_futex+0x33c/0xc40
 [<000000000014643e>] sys_futex+0x12e/0x144
 [<000000000010bb00>] sysc_noemu+0x10/0x16
 [<000002000003741c>] 0x2000003741c

The code in question is:

static int unqueue_me(struct futex_q *q)
{
        int ret = 0;
        spinlock_t *lock_ptr;

        /* In the common case we don't take the spinlock, which is nice. */
 retry:
        lock_ptr = q->lock_ptr;
        if (lock_ptr != 0) {
                spin_lock(lock_ptr);
		/*
                 * q->lock_ptr can change between reading it and
                 * spin_lock(), causing us to take the wrong lock.  This
                 * corrects the race condition.
[...]

and my compiler (gcc 4.1.0) makes the following out of it:

00000000000003c8 <unqueue_me>:
     3c8:       eb bf f0 70 00 24       stmg    %r11,%r15,112(%r15)
     3ce:       c0 d0 00 00 00 00       larl    %r13,3ce <unqueue_me+0x6>
                        3d0: R_390_PC32DBL      .rodata+0x2a
     3d4:       a7 f1 1e 00             tml     %r15,7680
     3d8:       a7 84 00 01             je      3da <unqueue_me+0x12>
     3dc:       b9 04 00 ef             lgr     %r14,%r15
     3e0:       a7 fb ff d0             aghi    %r15,-48
     3e4:       b9 04 00 b2             lgr     %r11,%r2
     3e8:       e3 e0 f0 98 00 24       stg     %r14,152(%r15)
     3ee:       e3 c0 b0 28 00 04       lg      %r12,40(%r11)
		/* write q->lock_ptr in r12 */
     3f4:       b9 02 00 cc             ltgr    %r12,%r12
     3f8:       a7 84 00 4b             je      48e <unqueue_me+0xc6>
		/* if r12 is zero then jump over the code.... */
     3fc:       e3 20 b0 28 00 04       lg      %r2,40(%r11)
		/* write q->lock_ptr in r2 */
     402:       c0 e5 00 00 00 00       brasl   %r14,402 <unqueue_me+0x3a>
                        404: R_390_PC32DBL      _spin_lock+0x2
		/* use r2 as parameter for spin_lock */

So the code becomes more or less:
if (q->lock_ptr != 0) spin_lock(q->lock_ptr)
instead of
if (lock_ptr != 0) spin_lock(lock_ptr)

Which caused the oops from above.
After adding a barrier gcc creates code without this problem:
[...] (the same)
     3ee:       e3 c0 b0 28 00 04       lg      %r12,40(%r11)
     3f4:       b9 02 00 cc             ltgr    %r12,%r12
     3f8:       b9 04 00 2c             lgr     %r2,%r12
     3fc:       a7 84 00 48             je      48c <unqueue_me+0xc4>
     400:       c0 e5 00 00 00 00       brasl   %r14,400 <unqueue_me+0x38>
                        402: R_390_PC32DBL      _spin_lock+0x2

As a general note, this code of unqueue_me seems a bit fishy. The retry logic
of unqueue_me only works if we can guarantee, that the original value of
q->lock_ptr is always a spinlock (Otherwise we overwrite kernel memory). We
know that q->lock_ptr can change. I dont know what happens with the original
spinlock, as I am not an expert with the futex code.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Ingo Molnar <mingo@redhat.com>
Cc: Thomas Gleixner <tglx@timesys.com>
Signed-off-by: Christian Borntraeger <borntrae@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
1 parent 72f0b4e
History
File Mode Size
8253pit.h -rw-r--r-- 129 bytes
Kbuild -rw-r--r-- 123 bytes
a.out.h -rw-r--r-- 2.5 KB
agp.h -rw-r--r-- 575 bytes
agp_backend.h -rw-r--r-- 909 bytes
atomic.h -rw-r--r-- 5.3 KB
auxvec.h -rw-r--r-- 651 bytes
barrier.h -rw-r--r-- 693 bytes
bitops.h -rw-r--r-- 8.1 KB
bug.h -rw-r--r-- 439 bytes
bugs.h -rw-r--r-- 294 bytes
byteorder.h -rw-r--r-- 1.3 KB
cache.h -rw-r--r-- 542 bytes
cacheflush.h -rw-r--r-- 2.5 KB
checksum.h -rw-r--r-- 2.2 KB
compiler.h -rw-r--r-- 3.7 KB
console.h -rw-r--r-- 2.0 KB
core_apecs.h -rw-r--r-- 16.9 KB
core_cia.h -rw-r--r-- 15.4 KB
core_irongate.h -rw-r--r-- 6.6 KB
core_lca.h -rw-r--r-- 11.3 KB
core_marvel.h -rw-r--r-- 9.1 KB
core_mcpcia.h -rw-r--r-- 11.4 KB
core_polaris.h -rw-r--r-- 2.9 KB
core_t2.h -rw-r--r-- 19.6 KB
core_titan.h -rw-r--r-- 11.2 KB
core_tsunami.h -rw-r--r-- 8.4 KB
core_wildfire.h -rw-r--r-- 8.4 KB
cputime.h -rw-r--r-- 118 bytes
current.h -rw-r--r-- 201 bytes
delay.h -rw-r--r-- 225 bytes
div64.h -rw-r--r-- 31 bytes
dma-mapping.h -rw-r--r-- 2.6 KB
dma.h -rw-r--r-- 12.3 KB
elf.h -rw-r--r-- 5.7 KB
emergency-restart.h -rw-r--r-- 149 bytes
err_common.h -rw-r--r-- 3.2 KB
err_ev6.h -rw-r--r-- 116 bytes
err_ev7.h -rw-r--r-- 4.3 KB
errno.h -rw-r--r-- 5.1 KB
fcntl.h -rw-r--r-- 1.1 KB
floppy.h -rw-r--r-- 3.2 KB
fpu.h -rw-r--r-- 6.1 KB
futex.h -rw-r--r-- 82 bytes
gct.h -rw-r--r-- 1003 bytes
gentrap.h -rw-r--r-- 1.4 KB
hardirq.h -rw-r--r-- 828 bytes
hw_irq.h -rw-r--r-- 216 bytes
hwrpb.h -rw-r--r-- 6.9 KB
ide.h -rw-r--r-- 1002 bytes
io.h -rw-r--r-- 14.6 KB
io_trivial.h -rw-r--r-- 2.9 KB
ioctl.h -rw-r--r-- 2.2 KB
ioctls.h -rw-r--r-- 3.6 KB
ipcbuf.h -rw-r--r-- 575 bytes
irq.h -rw-r--r-- 2.4 KB
jensen.h -rw-r--r-- 8.3 KB
kmap_types.h -rw-r--r-- 477 bytes
linkage.h -rw-r--r-- 86 bytes
local.h -rw-r--r-- 1.4 KB
machvec.h -rw-r--r-- 3.5 KB
mc146818rtc.h -rw-r--r-- 641 bytes
md.h -rw-r--r-- 251 bytes
mman.h -rw-r--r-- 2.3 KB
mmu.h -rw-r--r-- 164 bytes
mmu_context.h -rw-r--r-- 7.0 KB
mmzone.h -rw-r--r-- 3.0 KB
module.h -rw-r--r-- 448 bytes
msgbuf.h -rw-r--r-- 859 bytes
mutex.h -rw-r--r-- 308 bytes
namei.h -rw-r--r-- 376 bytes
page.h -rw-r--r-- 2.5 KB
pal.h -rw-r--r-- 935 bytes
param.h -rw-r--r-- 646 bytes
parport.h -rw-r--r-- 517 bytes
pci.h -rw-r--r-- 9.0 KB
percpu.h -rw-r--r-- 114 bytes
pgalloc.h -rw-r--r-- 1.5 KB
pgtable.h -rw-r--r-- 13.3 KB
poll.h -rw-r--r-- 473 bytes
posix_types.h -rw-r--r-- 3.2 KB
processor.h -rw-r--r-- 2.1 KB
ptrace.h -rw-r--r-- 2.0 KB
reg.h -rw-r--r-- 942 bytes
regdef.h -rw-r--r-- 1004 bytes
resource.h -rw-r--r-- 698 bytes
rtc.h -rw-r--r-- 142 bytes
rwsem.h -rw-r--r-- 5.6 KB
scatterlist.h -rw-r--r-- 381 bytes
sections.h -rw-r--r-- 128 bytes
segment.h -rw-r--r-- 132 bytes
semaphore.h -rw-r--r-- 3.1 KB
sembuf.h -rw-r--r-- 607 bytes
serial.h -rw-r--r-- 1006 bytes
setup.h -rw-r--r-- 87 bytes
sfp-machine.h -rw-r--r-- 2.9 KB
shmbuf.h -rw-r--r-- 1.0 KB
shmparam.h -rw-r--r-- 152 bytes
sigcontext.h -rw-r--r-- 828 bytes
siginfo.h -rw-r--r-- 169 bytes
signal.h -rw-r--r-- 3.8 KB
smp.h -rw-r--r-- 1.3 KB
socket.h -rw-r--r-- 1.5 KB
sockios.h -rw-r--r-- 385 bytes
spinlock.h -rw-r--r-- 3.0 KB
spinlock_types.h -rw-r--r-- 370 bytes
stat.h -rw-r--r-- 1012 bytes
statfs.h -rw-r--r-- 89 bytes
string.h -rw-r--r-- 2.1 KB
suspend.h -rw-r--r-- 107 bytes
sysinfo.h -rw-r--r-- 857 bytes
system.h -rw-r--r-- 16.8 KB
termbits.h -rw-r--r-- 4.2 KB
termios.h -rw-r--r-- 4.7 KB
thread_info.h -rw-r--r-- 2.8 KB
timex.h -rw-r--r-- 758 bytes
tlb.h -rw-r--r-- 394 bytes
tlbflush.h -rw-r--r-- 3.6 KB
topology.h -rw-r--r-- 835 bytes
types.h -rw-r--r-- 1.3 KB
uaccess.h -rw-r--r-- 14.3 KB
ucontext.h -rw-r--r-- 309 bytes
unaligned.h -rw-r--r-- 100 bytes
unistd.h -rw-r--r-- 19.9 KB
user.h -rw-r--r-- 2.1 KB
vga.h -rw-r--r-- 1.2 KB
xor.h -rw-r--r-- 21.7 KB

back to top