* [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths
@ 2011-05-02 9:13 Tejun Heo
2011-05-02 10:18 ` Ingo Molnar
0 siblings, 1 reply; 6+ messages in thread
From: Tejun Heo @ 2011-05-02 9:13 UTC (permalink / raw)
To: Ingo Molnar; +Cc: linux-kernel, x86, Yinghai Lu
Hello, Ingo.
Please pull from the following branch to receive unification of 32 and
64bit NUMA init paths.
git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
ssh://master.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
Other than addition of two patches from Yinghai and rebasing on top of
x86/urgent, nothing has changed from the initial posting[1].
Please note that the tree is on top of
x86/mm c7a7b814c9dca9ee01b38e63b4a46de87156d3b6
+ x86/numa 993ba1585cbb03fab012e41d1a5d24330a283b31
+ x86/urgent 2be19102b71c1a45d37fec50303791daa1a06869
x86/urgent was needed for the following two NUMA fix commits.
2be19102b7 "x86, NUMA: Fix empty memblk detection in numa_cleanup_meminfo()"
765af22da8 "x86-32, NUMA: Fix ACPI NUMA init broken by recent x86-64 change"
I can regenerate tree such that it's x86/mm + x86/numa + cherry pick
of the above two patches. If that's preferred, please let me know.
Thank you.
Tejun Heo (25):
x86-64, NUMA: Simplify hotadd memory handling
x86-64, NUMA: trivial cleanups for setup_node_bootmem()
x86-64, NUMA: simplify nodedata allocation
x86-32, NUMA: Automatically set apicid -> node in setup_local_APIC()
x86, NUMA: Unify 32/64bit numa_cpu_node() implementation
x86-32, NUMA: Make apic->x86_32_numa_cpu_node() optional
x86-32, NUMA: use sparse_memory_present_with_active_regions()
x86, NUMA: trivial cleanups
x86, NUMA: rename srat_64.c to srat.c
x86, NUMA: make srat.c 32bit safe
x86-32, NUMA: Move get_memcfg_numa() into numa_32.c
x86, NUMA: Move numa_nodes_parsed to numa.[hc]
x86-32, NUMA: implement temporary NUMA init shims
x86-32, NUMA: Replace srat_32.c with srat.c
x86-32, NUMA: Update numaq to use new NUMA init protocol
x86, NUMA: Move NUMA init logic from numa_64.c to numa.c
x86, NUMA: Enable build of generic NUMA init code on 32bit
x86, NUMA: Remove long 64bit assumption from numa.c
x86-32, NUMA: Add @start and @end to init_alloc_remap()
x86, NUMA: Initialize and use remap allocator from setup_node_bootmem()
x86, NUMA: Make 32bit use common NUMA init path
x86, NUMA: Make numa_init_array() static
x86, NUMA: Rename amdtopology_64.c to amdtopology.c
x86, NUMA: Enable CONFIG_AMD_NUMA on 32bit too
x86, NUMA: Enable emulation on 32bit too
Yinghai Lu (2):
x86, NUMA: Rename setup_node_bootmem() to setup_node_data()
x86, NUMA: Trim numa meminfo with max_pfn in a separate loop
arch/x86/Kconfig | 4 +-
arch/x86/include/asm/acpi.h | 2 -
arch/x86/include/asm/amd_nb.h | 1 -
arch/x86/include/asm/apic.h | 9 +-
arch/x86/include/asm/dma.h | 12 +-
arch/x86/include/asm/mmzone_32.h | 20 -
arch/x86/include/asm/numa.h | 32 ++-
arch/x86/include/asm/numa_32.h | 10 -
arch/x86/include/asm/numa_64.h | 36 --
arch/x86/include/asm/numaq.h | 7 +-
arch/x86/include/asm/srat.h | 39 --
arch/x86/include/asm/topology.h | 7 -
arch/x86/kernel/apic/apic.c | 26 +-
arch/x86/kernel/apic/apic_noop.c | 9 -
arch/x86/kernel/apic/bigsmp_32.c | 1 -
arch/x86/kernel/apic/es7000_32.c | 7 -
arch/x86/kernel/apic/numaq_32.c | 30 +-
arch/x86/kernel/apic/probe_32.c | 1 -
arch/x86/kernel/apic/summit_32.c | 1 -
arch/x86/mm/Makefile | 4 +-
arch/x86/mm/{amdtopology_64.c => amdtopology.c} | 21 +-
arch/x86/mm/init_32.c | 1 +
arch/x86/mm/init_64.c | 8 -
arch/x86/mm/numa.c | 550 ++++++++++++++++++++++-
arch/x86/mm/numa_32.c | 165 +-------
arch/x86/mm/numa_64.c | 519 +---------------------
arch/x86/mm/numa_emulation.c | 16 +-
arch/x86/mm/numa_internal.h | 8 +
arch/x86/mm/{srat_64.c => srat.c} | 82 +----
arch/x86/mm/srat_32.c | 287 ------------
30 files changed, 655 insertions(+), 1260 deletions(-)
delete mode 100644 arch/x86/include/asm/srat.h
rename arch/x86/mm/{amdtopology_64.c => amdtopology.c} (90%)
rename arch/x86/mm/{srat_64.c => srat.c} (67%)
delete mode 100644 arch/x86/mm/srat_32.c
--
tejun
[1] http://thread.gmane.org/gmane.linux.kernel/1133076
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths
2011-05-02 9:13 [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths Tejun Heo
@ 2011-05-02 10:18 ` Ingo Molnar
2011-05-02 12:23 ` Tejun Heo
0 siblings, 1 reply; 6+ messages in thread
From: Ingo Molnar @ 2011-05-02 10:18 UTC (permalink / raw)
To: Tejun Heo; +Cc: Ingo Molnar, linux-kernel, x86, Yinghai Lu
* Tejun Heo <tj@kernel.org> wrote:
> Hello, Ingo.
>
> Please pull from the following branch to receive unification of 32 and
> 64bit NUMA init paths.
>
> git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
> ssh://master.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
>
> Other than addition of two patches from Yinghai and rebasing on top of
> x86/urgent, nothing has changed from the initial posting[1].
>
> Please note that the tree is on top of
>
> x86/mm c7a7b814c9dca9ee01b38e63b4a46de87156d3b6
> + x86/numa 993ba1585cbb03fab012e41d1a5d24330a283b31
> + x86/urgent 2be19102b71c1a45d37fec50303791daa1a06869
>
> x86/urgent was needed for the following two NUMA fix commits.
>
> 2be19102b7 "x86, NUMA: Fix empty memblk detection in numa_cleanup_meminfo()"
> 765af22da8 "x86-32, NUMA: Fix ACPI NUMA init broken by recent x86-64 change"
>
> I can regenerate tree such that it's x86/mm + x86/numa + cherry pick
> of the above two patches. If that's preferred, please let me know.
Yeah, it's fine to me if you do merges, but could you please do a x86/mm tree
where all your merge commit messages are amended with a 'Merge reason'
explanation?
See commit a4c98f8bbeafee12c979c90743f6fda94f7515c7 for an example of how Merge
reason lines look like.
This is what we do for all merge commits in the -tip tree - that way it's clear
in retrospect why a particular merge was done.
Thanks,
Ingo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths
2011-05-02 10:18 ` Ingo Molnar
@ 2011-05-02 12:23 ` Tejun Heo
2011-05-02 15:04 ` Ingo Molnar
0 siblings, 1 reply; 6+ messages in thread
From: Tejun Heo @ 2011-05-02 12:23 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Ingo Molnar, linux-kernel, x86, Yinghai Lu
Hello, Ingo.
On Mon, May 02, 2011 at 12:18:28PM +0200, Ingo Molnar wrote:
> Yeah, it's fine to me if you do merges, but could you please do a x86/mm tree
> where all your merge commit messages are amended with a 'Merge reason'
> explanation?
>
> See commit a4c98f8bbeafee12c979c90743f6fda94f7515c7 for an example of how Merge
> reason lines look like.
>
> This is what we do for all merge commits in the -tip tree - that way it's clear
> in retrospect why a particular merge was done.
Branches updated as suggested. The new HEAD is 356aec8b6e "x86, NUMA:
Trim numa meminfo with max_pfn in a separate loop".
git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
ssh://master.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
And FYI, the following descriptions were used for the two merge
commits.
Author: Tejun Heo <tj@kernel.org> 2011-05-02 14:08:43
Committer: Tejun Heo <tj@kernel.org> 2011-05-02 14:08:47
Parent: c7a7b814c9dca9ee01b38e63b4a46de87156d3b6 (ioremap: Delay sanity check until after a successful mapping)
Parent: 993ba1585cbb03fab012e41d1a5d24330a283b31 (x86-32, numa: Update remap allocator comments)
Child: ba67cf5cf2ce10ad86a212b70f8c7c75d93a5016 (Merge branch 'x86/urgent' into x86-mm)
Branches: remotes/origin/x86-mm, x86-mm
Follows: v2.6.39-rc2
Precedes:
Merge branch 'x86/numa' into x86-mm
Merge reason: Pick up x86-32 remap allocator cleanup changes - 14
commits, 3fe14ab541^..993ba1585c.
3fe14ab541: x86-32, numa: Fix failure condition check in alloc_remap()
993ba1585c: x86-32, numa: Update remap allocator comments
Scheduled NUMA init 32/64bit unification changes depend on them.
Signed-off-by: Tejun Heo <tj@kernel.org>
Author: Tejun Heo <tj@kernel.org> 2011-05-02 14:16:37
Committer: Tejun Heo <tj@kernel.org> 2011-05-02 14:16:47
Parent: aff364860aa105b2deacc6f21ec8ef524460e3fc (Merge branch 'x86/numa' into x86-mm)
Parent: 2be19102b71c1a45d37fec50303791daa1a06869 (x86, NUMA: Fix empty memblk detection in numa_cleanup_meminfo())
Child: 9688678a6670c7f0ae3872450a8047c0ad401efb (x86-64, NUMA: Simplify hotadd memory handling)
Branches: remotes/origin/x86-mm, x86-mm
Follows: v2.6.39-rc5
Precedes:
Merge branch 'x86/urgent' into x86-mm
Merge reason: Pick up the following two fix commits.
2be19102b7: x86, NUMA: Fix empty memblk detection in numa_cleanup_meminfo()
765af22da8: x86-32, NUMA: Fix ACPI NUMA init broken by recent x86-64 change
Scheduled NUMA init 32/64bit unification changes depend on these.
Signed-off-by: Tejun Heo <tj@kernel.org>
Thank you.
--
tejun
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths
2011-05-02 12:23 ` Tejun Heo
@ 2011-05-02 15:04 ` Ingo Molnar
2011-05-02 15:09 ` Tejun Heo
0 siblings, 1 reply; 6+ messages in thread
From: Ingo Molnar @ 2011-05-02 15:04 UTC (permalink / raw)
To: Tejun Heo; +Cc: Ingo Molnar, linux-kernel, x86, Yinghai Lu
* Tejun Heo <tj@kernel.org> wrote:
> Hello, Ingo.
>
> On Mon, May 02, 2011 at 12:18:28PM +0200, Ingo Molnar wrote:
> > Yeah, it's fine to me if you do merges, but could you please do a x86/mm tree
> > where all your merge commit messages are amended with a 'Merge reason'
> > explanation?
> >
> > See commit a4c98f8bbeafee12c979c90743f6fda94f7515c7 for an example of how Merge
> > reason lines look like.
> >
> > This is what we do for all merge commits in the -tip tree - that way it's clear
> > in retrospect why a particular merge was done.
>
> Branches updated as suggested. The new HEAD is 356aec8b6e "x86, NUMA:
> Trim numa meminfo with max_pfn in a separate loop".
>
> git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
Doesnt build on 32-bit allyesconfig:
ERROR: "physnode_map" [drivers/net/wan/dscc4.ko] undefined!
ERROR: "physnode_map" [crypto/tcrypt.ko] undefined!
Thanks,
Ingo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths
2011-05-02 15:04 ` Ingo Molnar
@ 2011-05-02 15:09 ` Tejun Heo
2011-05-02 15:30 ` Tejun Heo
0 siblings, 1 reply; 6+ messages in thread
From: Tejun Heo @ 2011-05-02 15:09 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Ingo Molnar, linux-kernel, x86, Yinghai Lu
On Mon, May 02, 2011 at 05:04:01PM +0200, Ingo Molnar wrote:
> Doesnt build on 32-bit allyesconfig:
>
> ERROR: "physnode_map" [drivers/net/wan/dscc4.ko] undefined!
> ERROR: "physnode_map" [crypto/tcrypt.ko] undefined!
Dang, I only checked allmodconfig. Testing allyesconfig now.
Thanks.
--
tejun
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths
2011-05-02 15:09 ` Tejun Heo
@ 2011-05-02 15:30 ` Tejun Heo
0 siblings, 0 replies; 6+ messages in thread
From: Tejun Heo @ 2011-05-02 15:30 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Ingo Molnar, linux-kernel, x86, Yinghai Lu
Hello, again.
On Mon, May 02, 2011 at 05:09:45PM +0200, Tejun Heo wrote:
> On Mon, May 02, 2011 at 05:04:01PM +0200, Ingo Molnar wrote:
> > Doesnt build on 32-bit allyesconfig:
> >
> > ERROR: "physnode_map" [drivers/net/wan/dscc4.ko] undefined!
> > ERROR: "physnode_map" [crypto/tcrypt.ko] undefined!
>
> Dang, I only checked allmodconfig. Testing allyesconfig now.
Nope, it was a last minute edit while tree generation which dropped
#include <linux/module.h> in the unification patch. This broke
EXPORT_SYMBOL(physnode_map) and thus led to the above failure.
I regenerated from the reunification patch up. The new HEAD is
e5a10c1bd12a5d71bbb6406c1b0dbbc9d8958397. Sorry about the fuss.
git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
ssh://master.kernel.org/pub/scm/linux/kernel/git/tj/misc.git x86-mm
Thanks.
--
tejun
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2011-05-02 15:31 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-05-02 9:13 [GIT PULL tip:x86/mm] Unify 32 and 64bit NUMA init paths Tejun Heo
2011-05-02 10:18 ` Ingo Molnar
2011-05-02 12:23 ` Tejun Heo
2011-05-02 15:04 ` Ingo Molnar
2011-05-02 15:09 ` Tejun Heo
2011-05-02 15:30 ` Tejun Heo
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox