The Annnoying Affair has No eXcuses - Bulletin #3

The Annnoying Affair has No eXcuses - Bulletin #3

[Thomas Gleixner]

[-- Attachment #1: Type: text/plain, Size: 348 bytes --]

Folks,

just a quick update.

The TAA fix from Josh is now applied to master and to the stable branches,
which are still waiting for Paolos NX updates. Hint, hint...

The changes are available as usual from:

  cvs.ou.linutronix.de:linux/speck/linux  

Git bundle for the master branch is attached. Applies on top of 5.4-rc6.

Thanks,

        tglx

[-- Attachment #2: Type: application/octet-stream, Size: 43106 bytes --]

[MODERATED] Re: The Annnoying Affair has No eXcuses - Bulletin #3

[Paolo Bonzini]

[-- Attachment #1.1: Type: text/plain, Size: 279 bytes --]

On 07/11/19 16:16, speck for Thomas Gleixner wrote:
> The TAA fix from Josh is now applied to master and to the stable branches,
> which are still waiting for Paolos NX updates. Hint, hint...

Looks like I have No Excuses for procrastinating. :)  Bundles attached.

Paolo

[-- Attachment #1.2: nx-4.14-bundle --]
[-- Type: application/octet-stream, Size: 19884 bytes --]

[-- Attachment #1.3: nx-4.19-bundle --]
[-- Type: application/octet-stream, Size: 19906 bytes --]

[-- Attachment #1.4: nx-5.3-bundle --]
[-- Type: application/octet-stream, Size: 12472 bytes --]

[MODERATED] Re: The Annnoying Affair has No eXcuses - Bulletin #3

[mark gross]

On Thu, Nov 07, 2019 at 06:10:17PM +0100, speck for Paolo Bonzini wrote:
> On 07/11/19 16:16, speck for Thomas Gleixner wrote:
> > The TAA fix from Josh is now applied to master and to the stable branches,
> > which are still waiting for Paolos NX updates. Hint, hint...
> 
> Looks like I have No Excuses for procrastinating. :)  Bundles attached.
> 
I'm unbundlign these now and we will start testing an all the branches today.

--mark

Re: The Annnoying Affair has No eXcuses - Bulletin #3

[Thomas Gleixner]

Paolo,

On Thu, 7 Nov 2019, speck for Paolo Bonzini wrote:
> On 07/11/19 16:16, speck for Thomas Gleixner wrote:
> > The TAA fix from Josh is now applied to master and to the stable branches,
> > which are still waiting for Paolos NX updates. Hint, hint...
> 
> Looks like I have No Excuses for procrastinating. :)  Bundles attached.

First of all thanks for providing that!

But now I really have to ask whether you actually ever looked how backports
into stable kernels are done? Obviously not, because your series is just
buggered:

 1) All backport commits have to have

    commit LOOOONGSHAAAAAA upstream

    Right at the top of the changelog, not some randomly defined
    (cherrypicked ....) thingy in the middle of the SOB chain

 2) All backport commits have to be the same as the upstream
    commit including change logs, except for eventual conflicts.

    Instead of using the commits from the repo or the bundle you picked
    your V9 version and backported it and of course you still have the
    wrong default in the documentation of the command line parameter along
    with different change logs.

 3) All backport commits have to preserve the SOB/tag chain of the
    upstream commit, plus your own if the committer was not yourself.

    Of course all the ones which are in the master branch do not have that
    because you used your V9 version as a base. Aside of missing my SOB's
    the broken SOB chain from 'kvm: mmu: ITLB_MULTIHIT mitigation' which I
    fixed up when merging the lot is still in your backports.

 4) Random 'RHEL:' tags which are nowhere in the upstream commit have
    really no place in stable kernels.

Groan, you are doing kernel development for how long now?

Sorry for my lousy mood, but after staring into futex code 3 days straight,
this was really what I needed - NOT. I really was hoping just to do a pull
and write a merge message and not wasting an hour for untinkering.

I fixed it all up and I'm going to send out a bulletin with complete
TAA/NX bundles in a few.

Yours grumpy

      tglx