Re: [PATCH] x86/apic: Remove logical destination mode for 64-bit

[Thomas Gleixner <tglx@linutronix.de>]

On Fri, Sep 06 2024 at 10:34, Linux regression tracking wrote:
> On 05.09.24 16:04, Borislav Petkov wrote:
>> https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?h=x86/apic&id=838ba7733e4e3a94a928e8d0a058de1811a58621
>
> Hmmm. Please help me out here: why was that fix queued for -next and not
> for this cycle?
>
> Was that patch when it was committed considered too dangerous for
> mainlining this cycle (at this point of the cycle I guess it might)?

Yes.

> I mean, it's afaics (not totally sure here, the change is missing a Fixes:
> tag as well as Closes: tags pointing to the report) fixing a regression
> with f0551af02130 that Christian reported (see start of this thread, e.g.,
> https://lore.kernel.org/all/12df8b45-6100-4c8b-b82a-a6a75bed2e05@heusel.eu/
> ). And f0551af02130 is from v6.9-rc1, so given what Linus wrote in
> https://lore.kernel.org/all/CAHk-=wis_qQy4oDNynNKi5b7Qhosmxtoj1jxo5wmB6SRUwQUBQ@mail.gmail.com/
> that fix should likely have been (or still should be?) merged in this
> cycle, unless it's really dangerous.
>
> Or did I misunderstood something here?

I really wanted to cook it first. Aside of that f0551af02130 unearthed a
firmware bug as the reporters confirmed. So I didn't see an immediate
reason to send it to Linus.

My rationale for writing the patch was to avoid this issue in the future
for those who can't update firmware and have it in the next LTS release,
which is what distros will ship in their stable offerings.

Thanks,

        tglx