mbox

[GIT,PULL] Please pull powerpc/linux.git powerpc-4.9-6 tag

Message ID 87inraslrk.fsf@concordia.ellerman.id.au (mailing list archive)
State Not Applicable
Headers show

Pull-request

git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git powerpc-4.9-6

Message

Michael Ellerman Nov. 26, 2016, 8:11 a.m. UTC
Hi Linus,

Please pull some more powerpc fixes for 4.9:

The following changes since commit cac4a185405d4415eca269cae976438b44a37ae0:

  powerpc/mm: Fix missing update of HID register on secondary CPUs (2016-11-18 23:16:58 +1100)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git powerpc-4.9-6

for you to fetch changes up to 984d7a1ec67ce3a46324fa4bcb4c745bbc266cf2:

  powerpc/mm: Fixup kernel read only mapping (2016-11-25 14:18:25 +1100)

----------------------------------------------------------------
powerpc fixes for 4.9 #6

Fixes marked for stable:
 - Set missing wakeup bit in LPCR on POWER9 (Benjamin Herrenschmidt)
 - Fix the early OPAL console wrappers (Oliver O'Halloran)
 - Fixup kernel read only mapping (Aneesh Kumar K.V)

Fixes for code merged this cycle:
 - Fix missing CRCs, add more asm-prototypes.h declarations (Nicholas Piggin)

----------------------------------------------------------------
Aneesh Kumar K.V (1):
      powerpc/mm: Fixup kernel read only mapping

Benjamin Herrenschmidt (1):
      powerpc: Set missing wakeup bit in LPCR on POWER9

Nicholas Piggin (1):
      powerpc: Fix missing CRCs, add more asm-prototypes.h declarations

Oliver O'Halloran (1):
      powerpc/boot: Fix the early OPAL console wrappers

 arch/powerpc/boot/main.c                  |  8 ++++++--
 arch/powerpc/boot/opal-calls.S            | 13 +++++++++++++
 arch/powerpc/boot/opal.c                  | 11 +++++++++++
 arch/powerpc/boot/ops.h                   |  1 +
 arch/powerpc/include/asm/asm-prototypes.h | 12 ++++++++++++
 arch/powerpc/include/asm/mmu.h            | 14 ++++++++++----
 arch/powerpc/include/asm/reg.h            |  1 +
 arch/powerpc/kernel/cpu_setup_power.S     |  8 ++++----
 arch/powerpc/mm/hash_utils_64.c           |  8 ++++++--
 9 files changed, 64 insertions(+), 12 deletions(-)

Comments

Linus Torvalds Nov. 26, 2016, 7:29 p.m. UTC | #1
On Sat, Nov 26, 2016 at 12:11 AM, Michael Ellerman <mpe@ellerman.id.au> wrote:
>
> powerpc fixes for 4.9 #6
>
> Fixes marked for stable:
>  - Set missing wakeup bit in LPCR on POWER9 (Benjamin Herrenschmidt)
>  - Fix the early OPAL console wrappers (Oliver O'Halloran)
>  - Fixup kernel read only mapping (Aneesh Kumar K.V)
>
> Fixes for code merged this cycle:
>  - Fix missing CRCs, add more asm-prototypes.h declarations (Nicholas Piggin)

Pulled, but I wanted to talk about your merge "summary".

Your merge summaries seem to be entirely automatically generated,
which makes them less than great. I can see all that stuff in the git
tree already, just formatting it differently isn't all that useful.

For something like this late-rc pull when there are only a couple of
commits, the end result actually ends up looking almost like a summary
and all I did was remove the names that don't add to the description
(and are in the git commits).

For some of the bigger pull requests, the summary is almost anything
but, and the only real value-add is the grouping by subject area.

I really prefer a _summary_. Something that is human-legible. So that
when people read the merge commit log, they get an overview of what
changed, not a list of the details.

Again, when there are four commits, a "list of the details" pretty
much works. So the reason I react to _this_ pull request is mainly
that I have way more time to react to it during the late rc series
than I do earlier in the cycle..

                     Linus
Michael Ellerman Nov. 28, 2016, 9:46 a.m. UTC | #2
Linus Torvalds <torvalds@linux-foundation.org> writes:

> On Sat, Nov 26, 2016 at 12:11 AM, Michael Ellerman <mpe@ellerman.id.au> wrote:
>>
>> powerpc fixes for 4.9 #6
>>
>> Fixes marked for stable:
>>  - Set missing wakeup bit in LPCR on POWER9 (Benjamin Herrenschmidt)
>>  - Fix the early OPAL console wrappers (Oliver O'Halloran)
>>  - Fixup kernel read only mapping (Aneesh Kumar K.V)
>>
>> Fixes for code merged this cycle:
>>  - Fix missing CRCs, add more asm-prototypes.h declarations (Nicholas Piggin)
>
> Pulled, but I wanted to talk about your merge "summary".
>
> Your merge summaries seem to be entirely automatically generated,
> which makes them less than great. I can see all that stuff in the git
> tree already, just formatting it differently isn't all that useful.

OK. The starting point is obviously an automatically generated list of
commits, but I have been editing that a fair bit to drop boring commits
and combine multiple commits into a single line, and then sort it by
topic area etc.

But obviously I'm not editing it enough, so I'll try to summarise it
much more heavily.

> For something like this late-rc pull when there are only a couple of
> commits, the end result actually ends up looking almost like a summary
> and all I did was remove the names that don't add to the description
> (and are in the git commits).

I actually do like to include the names, just to give people a bit of
acknowledgment in the pull request, but I can drop them if you prefer.
Or maybe I'll just include a credits section at the bottom of the tag
with everyone's name once, and you can drop that from the commit?

> For some of the bigger pull requests, the summary is almost anything
> but, and the only real value-add is the grouping by subject area.
>
> I really prefer a _summary_. Something that is human-legible. So that
> when people read the merge commit log, they get an overview of what
> changed, not a list of the details.

Right.

cheers