Re: [PATCH 4.4 00/37] 4.4.110-stable review

[Andy Lutomirski]

> On Jan 4, 2018, at 12:57 PM, Hugh Dickins <hughd@xxxxxxxxxx> wrote:
> 
>> On Thu, Jan 4, 2018 at 12:43 PM, Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote:
>> 
>>>> On Jan 4, 2018, at 12:29 PM, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>>>> 
>>>> On Thu, Jan 4, 2018 at 12:16 PM, Thomas Voegtle <tv@xxxxxxxx> wrote:
>>>> 
>>>> Attached a screenshot.
>>>> Is that useful? Are there some debug options I can add?
>>> 
>>> Not much of an oops, because the SIGSEGV happens in user space. The
>>> only reason you get any kernel stack printout at all is because 'init'
>>> dying will make the kernel print that out.
>>> 
>>> The segfault address for init looks like the fixmap area to me (first
>>> byte in the last page of the fixmap?). "Error 5" means that it's a
>>> user-space read that got a protection fault. So it's not a LDT of GDT
>>> update or anything like that, it's a normal access from user space (or
>>> a qemu emulation bug, but that sounds unlikely).
>>> 
>>> Is that the vsyscall page?
>>> 
>>> Adding Luto to the participants. I think he noticed one of the
>>> vsyscall patches missing earlier in the 4.9 series. Maybe the 4.4
>>> series had something similar..
>>> 
>> 
>> That's almost certainly it.
> 
> I'm hopeless on the FIXMAP arithmetic, but I'm pretty sure that
> ffffffffff5ff000 is either VSYSCALL page or PVCLOCK page (I think it
> was VVAR page when init segfaulted on it in my 3.2).

Nah, that's one page below VSYSCALL.  Vvar is 0x7fff...

I don't have the actual screenshot, I think.

> 
> I'll forward Borislav's suggested 4.4 VSYSCALL patch from the kaiser
> backports ml to Thomas, to see if that sorts his crash (forwarding in
> the hope that gmail doesn't mess up the patch).
> 
> Seems odd that 4.4 should be broken but 4.9 not broken here, I'd
> expect them to be equally known broken with respect to VSYSCALL; but
> perhaps it's a matter of userspace trying different fallbacks
> according to what kernel supports, and only hitting this on 4.4.

I don't think any current userspace is that dumb.  But Go was still using vsyscall fairly recently.

I may be able to look for real tonight.

> 
> Hugh
> 
>> 
>> I'll try to find some time today  Thomnor tomorrow to add a proper selftest.
>> 
>>>             Linus