Skip to contentSkip to conversation

x86: fix savesegment() bug causing crashes on 64-bit

PATCH In review Ingo Molnar 2 participants 2 messages Jul 11, 2008

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86: fix savesegment() bug causing crashes on 64-bit

i spent a fair amount of time chasing a 64-bit bootup crash in tip/master that manifested itself as bootup segfaults:

      S10network[1825]: segfault at 7f3e2b5d16b8 ip 00000031108748c9 sp 00007fffb9c14c70 error 4 in libc-2.7.so[3110800000+14d000]

eventually causing init to die and panic the system:

      Kernel panic - not syncing: Attempted to kill init!
      Pid: 1, comm: init Not tainted 2.6.26-rc9-tip #13878

after a marathonic bisection session, the bad commit turned out to be this one in tip/x86/*:

    | b7675791859075418199c7af86a116ea34eaf5bd is first bad commit
    | commit b7675791859075418199c7af86a116ea34eaf5bd
    | Author: Jeremy Fitzhardinge <[email protected]>
    | Date:   Wed Jun 25 00:19:00 2008 -0400
    |
    |     x86: remove open-coded save/load segment operations
    |
    |     This removes a pile of buggy open-coded implementations of savesegment
    |     and loadsegment.

after some more bisection of this patch itself, it turns out that what makes the difference are the savesegment() changes to __switch_to().

Taking a look at this portion of arch/x86/kernel/process_64.o revealed this crutial difference:

    | good:    99c:       8c e0                   mov    %fs,%eax
    |          99e:       89 45 cc                mov    %eax,-0x34(%rbp)
    |
    | bad:     99c:       8c 65 cc                mov    %fs,-0x34(%rbp)

which is due to:

    |                 unsigned fsindex;
    | -               asm volatile("movl %%fs,%0" : "=r" (fsindex));
    | +               savesegment(fs, fsindex);

savesegment() is implemented as:

     #define savesegment(seg, value)                                \
              asm("mov %%" #seg ",%0":"=rm" (value) : : "memory")

note the "m" modifier - it allows GCC to generate the segment move into a memory operand as well.

But regarding segment operands there's a subtle detail in the x86 instruction set: the above 16-bit moves are zero-extend, but only

If it goes to a memory operand, -0x34(%rbp) in the above case, there's no zero-extend to 32-bit and the instruction will only save 16 bits

The other 16 bits is random data - which can cause problems when that value is used later on.

The solution is to only allow segment operands to go to registers. This fix allows my test-system to boot up without crashing.

---
 include/asm-x86/system.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
include/asm-x86/system.h +1 -1
index 45641bc..929345a 100644
@@ -164,7 +164,7 @@ extern void native_load_gs_index(unsigned);
  * Save a segment register away
  */
 #define savesegment(seg, value)				\
-	asm("mov %%" #seg ",%0":"=rm" (value) : : "memory")
+	asm("mov %%" #seg ",%0":"=r" (value) : : "memory")
 
 static inline unsigned long get_limit(unsigned long segment)
 {
Acked-by on v1 1/1
#1 lore raw
Quoted 3 lines · Ingo Molnar

i spent a fair amount of time chasing a 64-bit bootup crash in tip/master that manifested itself as bootup segfaults:

Crap, I'm sorry. I fixed this locally a couple of days ago, but I hadn't actually observed it causing any problems, so I didn't rush to push it out.

Quoted 4 lines

The solution is to only allow segment operands to go to registers. This fix allows my test-system to boot up without crashing.

J