Re: [PATCH 01/16] x86/traps: Drop last_extable_addr

[Jan Beulich <jbeulich@xxxxxxxx>]

On 11.05.2020 16:53, Andrew Cooper wrote:
> On 04/05/2020 13:44, Jan Beulich wrote:
>> On 02.05.2020 00:58, Andrew Cooper wrote:
>>> The only user of this facility is dom_crash_sync_extable() by passing 0 into
>>> asm_domain_crash_synchronous().  The common error cases are already covered
>>> with show_page_walk(), leaving only %ss/%fs selector/segment errors in the
>>> compat case.
>>>
>>> Point at dom_crash_sync_extable in the error message, which is about as good
>>> as the error hints from other users of asm_domain_crash_synchronous(), and
>>> drop last_extable_addr.
>> While I'm not entirely opposed, I'd like you to clarify that you indeed
>> mean to (mostly) revert your own improvement from 6 or 7 years back
>> (commit 8e0da8c07f4f). I'm also surprised to find this as part of the
>> series it's in - in what way does this logic get in the way of CET-SS?
> 
> It was part of the exception_fixup() cleanup.  The first 4 patches not
> specifically related to CET-SS.

"Cleanup" doesn't mean it's needed as a prereq for the CET-SS, so I'm
afraid it's still not really clear to me whether we indeed want to
effectively revert the earlier change, which had a reason to be done
the way it was done.

Jan