Re: [PATCH] x86_64, asm: Work around AMD SYSRET SS descriptor attribute issue

[Andy Lutomirski]

On Fri, Apr 24, 2015 at 1:46 PM, Denys Vlasenko
<vda.linux@xxxxxxxxxxxxxx> wrote:
> On Fri, Apr 24, 2015 at 10:21 PM, Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote:
>> On Thu, Apr 23, 2015 at 7:15 PM, Andy Lutomirski <luto@xxxxxxxxxx> wrote:
>>> AMD CPUs don't reinitialize the SS descriptor on SYSRET, so SYSRET
>>> with SS == 0 results in an invalid usermode state in which SS is
>>> apparently equal to __USER_DS but causes #SS if used.
>>>
>>> Work around the issue by replacing NULL SS values with __KERNEL_DS
>>> in __switch_to, thus ensuring that SYSRET never happens with SS set
>>> to NULL.
>>>
>>> This was exposed by a recent vDSO cleanup.
>>>
>>> Fixes: e7d6eefaaa44 x86/vdso32/syscall.S: Do not load __USER32_DS to %ss
>>> Signed-off-by: Andy Lutomirski <luto@xxxxxxxxxx>
>>> ---
>>>
>>> Tested only on Intel, which isn't very interesting.  I'll tidy up
>>> and send a test case, too, once Borislav confirms that it works.
>>>
>>> Please don't actually apply this until we're sure we understand the
>>> scope of the issue.  If this doesn't affect SYSRETQ, then we might
>>> to fix it on before SYSRETL to avoid impacting 64-bit processes
>>> at all.
>>>
>>
>> After sleeping on it, I think I want to offer a different, more
>> complicated approach.  AFAIK there are really only two ways that this
>> issue can be visible:
>>
>> 1. SYSRETL.  We can fix that up in the AMD SYSRETL path.  I think
>> there's a decent argument that that path is less performance-critical
>> than context switches.
>>
>> 2. SYSRETQ.  The only way that I know of to see the problem is SYSRETQ
>> followed by a far jump or return.  This is presumably *extremely*
>> rare.
>>
>> What if we fixed #2 up in do_stack_segment.  We should double-check
>> the docs, but I think that this will only ever manifest as #SS(0) with
>> regs->ss == __USER_DS and !user_mode_64bit(regs).  We need to avoid
>> infinite retry looks, but this might be okay.  I think that #SS(0)
>> from userspace under those conditions can *only* happen as a result of
>> this issue.  Even if not, we could come up with a way to only retry
>> once per syscall (e.g. set some ti->status flag in the 64-bit syscall
>> path on AMD and clear it in do_stack_segment).
>>
>> This might be way more trouble than it's worth.
>
> Exactly my feeling. What are you trying to save? About four CPU
> cycles of checking %ss != __KERNEL_DS on each switch_to?
> That's not worth bothering about. Your last patch seems to be perfect.

We'll have to do the write to ss almost every time an AMD CPU sleeps
in a syscall.  Maybe that's still not a big deal.

--Andy

-- 
Andy Lutomirski
AMA Capital Management, LLC
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/