On 9/27/20 8:43 AM, Christophe Leroy wrote:
>
>
> Le 21/09/2020 à 13:26, Will Deacon a écrit :
>> On Fri, Aug 28, 2020 at 12:14:28PM +1000, Michael Ellerman wrote:
>>> Dmitry Safonov <0x7f454...@gmail.com> writes:
[..]
>>>> I'll cook a patch for vm_special_mapping if you don't mind :-)
>>>
>>> That would be great, thanks!
>>
>> I lost track of this one. Is there a patch kicking around to resolve
>> this,
>> or is the segfault expected behaviour?
>>
>
> IIUC dmitry said he will cook a patch. I have not seen any patch yet.
Yes, sorry about the delay - I was a bit busy with xfrm patches.
I'll send patches for .close() this week, working on them now.
> AFAIKS, among the architectures having VDSO sigreturn trampolines, only
> SH, X86 and POWERPC provide alternative trampoline on stack when VDSO is
> not there.
>
> All other architectures just having a VDSO don't expect VDSO to not be
> mapped.
>
> As far as nowadays stacks are mapped non-executable, getting a segfaut
> is expected behaviour. However, I think we should really make it
> cleaner. Today it segfaults because it is still pointing to the VDSO
> trampoline that has been unmapped. But should the user map some other
> code at the same address, we'll run in the weed on signal return instead
> of segfaulting.
+1.
> So VDSO unmapping should really be properly managed, the reference
> should be properly cleared in order to segfault in a controllable manner.
>
> Only powerpc has a hook to properly clear the VDSO pointer when VDSO is
> unmapped.
Thanks,
Dmitry
