[PATCH 0/6] x86: misc testsuite reverts plus fallout thereof

[Jan Beulich]

On 17.07.2020 17:55, H.J. Lu wrote:
> On Fri, Jul 17, 2020 at 8:40 AM Jan Beulich <jbeulich@suse.com> wrote:
>>
>> On 17.07.2020 16:29, H.J. Lu wrote:
>>> On Fri, Jul 17, 2020 at 6:30 AM Jan Beulich <jbeulich@suse.com> wrote:
>>>>
>>>> Over time a number of choices I've made in testcases have been
>>>> "corrected", because of causing problems for certain targets.
>>>> The underlying issues, however, were never investigated. I've
>>>> finally taken the time to do so in two of the cases, which
>>>> interestingly has lead to fixing long-standing yet at the first
>>>> glance unrelated bugs (or shortcomings, if you like).
>>>>
>>>> Note that patch 5 is a change to common gas code; I'll try to
>>>> remember to Cc a few presumably relevant people when sending
>>>> out that one.
>>>>
>>>> 1: x86: enable more ELF tests for VxWorks
>>>> 2: Revert "x86: Replace evex-no-scale.s with evex-no-scale-[32|64].s"
>>>> 3: x86: honor absolute section when emitting code
>>>> 4: x86: revert ELF related part of "x86: Update assembler tests for non-ELF targets"
>>>
>>> Can you merge patches 2 and 4 into a single commit since they
>>> change the same file?
>>
>> I could, but this being unrelated changes I'd prefer to keep them
>> separate. Let me know if you really think merging is helpful here.
> 
> Because I expect they will/can be applied back to back.

But you realize that patch 4 functionally depends on patch 3?
It also would seem better to me for a revert to really be a
revert - when I fold both, I will need to invent different
titles (likely making the revert character of the changes
less obvious) and probably massage the then folded description.

Jan