[RESEND,0/4] Rework sparc32 page-table layout

Hi all,

This is a reposting of the patch series I sent previously to rework the
sparc32 page-table layout so that 'pmd_t' can be used safely with
READ_ONCE():

https://lore.kernel.org/lkml/20200324104005.11279-1-will@kernel.org

This is blocking the READ_ONCE() rework, which in turn allows us to
bump the minimum GCC version for building the kernel up to 4.8.

Cheers,

Will

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Cc: Matt Fleming <matt@codeblueprint.co.uk>
Cc: sparclinux@vger.kernel.org
Cc: kernel-team@android.com

--->8

Will Deacon (4):
  sparc32: mm: Fix argument checking in __srmmu_get_nocache()
  sparc32: mm: Restructure sparc32 MMU page-table layout
  sparc32: mm: Change pgtable_t type to pte_t * instead of struct page *
  sparc32: mm: Reduce allocation size for PMD and PTE tables

 arch/sparc/include/asm/page_32.h    | 12 ++--
 arch/sparc/include/asm/pgalloc_32.h | 11 ++--
 arch/sparc/include/asm/pgtable_32.h | 40 +++++++-----
 arch/sparc/include/asm/pgtsrmmu.h   | 36 +----------
 arch/sparc/include/asm/viking.h     |  5 +-
 arch/sparc/kernel/head_32.S         |  8 +--
 arch/sparc/mm/hypersparc.S          |  3 +-
 arch/sparc/mm/srmmu.c               | 95 ++++++++++-------------------
 arch/sparc/mm/viking.S              |  5 +-
 9 files changed, 85 insertions(+), 130 deletions(-)

From: Will Deacon <will@kernel.org>
Date: Tue, 14 Apr 2020 22:40:07 +0100

> This is a reposting of the patch series I sent previously to rework the
> sparc32 page-table layout so that 'pmd_t' can be used safely with
> READ_ONCE():
> 
> https://lore.kernel.org/lkml/20200324104005.11279-1-will@kernel.org
> 
> This is blocking the READ_ONCE() rework, which in turn allows us to
> bump the minimum GCC version for building the kernel up to 4.8.

Series applied to sparc-next, thanks Will.

Hi David,

On Wed, May 13, 2020 at 03:32:26PM -0700, David Miller wrote:
> From: Will Deacon <will@kernel.org>
> Date: Tue, 14 Apr 2020 22:40:07 +0100
> 
> > This is a reposting of the patch series I sent previously to rework the
> > sparc32 page-table layout so that 'pmd_t' can be used safely with
> > READ_ONCE():
> > 
> > https://lore.kernel.org/lkml/20200324104005.11279-1-will@kernel.org
> > 
> > This is blocking the READ_ONCE() rework, which in turn allows us to
> > bump the minimum GCC version for building the kernel up to 4.8.
> 
> Series applied to sparc-next, thanks Will.

Thanks, although this was picked up by -tip a couple of days ago [1] since
the READ_ONCE() rework relies on it to avoid breaking the sparc32 build.

It's obviously harmless to go in both trees, but if you don't need it for
anything then you may want to drop these to avoid duplicate commits in
mainline.

Cheers,

Will

[1] https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/log/?h=locking/kcsan

From: Will Deacon <will@kernel.org>
Date: Thu, 14 May 2020 08:26:24 +0100

> Hi David,
> 
> On Wed, May 13, 2020 at 03:32:26PM -0700, David Miller wrote:
>> From: Will Deacon <will@kernel.org>
>> Date: Tue, 14 Apr 2020 22:40:07 +0100
>> 
>> > This is a reposting of the patch series I sent previously to rework the
>> > sparc32 page-table layout so that 'pmd_t' can be used safely with
>> > READ_ONCE():
>> > 
>> > https://lore.kernel.org/lkml/20200324104005.11279-1-will@kernel.org
>> > 
>> > This is blocking the READ_ONCE() rework, which in turn allows us to
>> > bump the minimum GCC version for building the kernel up to 4.8.
>> 
>> Series applied to sparc-next, thanks Will.
> 
> Thanks, although this was picked up by -tip a couple of days ago [1] since
> the READ_ONCE() rework relies on it to avoid breaking the sparc32 build.
> 
> It's obviously harmless to go in both trees, but if you don't need it for
> anything then you may want to drop these to avoid duplicate commits in
> mainline.
 ...
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/log/?h=locking/kcsan

Thanks for the heads up but I'll leave it in the sparc-next tree for now.