2006-10-21 16:55:46

by Andi Kleen

[permalink] [raw]
Subject: Please pull x86 tree


Linus, please pull from

git://one.firstfloor.org/home/andi/git/linux-2.6 for-linus

These are all accumulated bug fixes for x86-64 and i386 and should
be all pretty safe.

The only thing that isn't a clear bug fix is the dwarf2 unwinder
speedup -- i'm including that on popular demand because it fixes
a serious performance regression with lockdep.

There are also some reverts included where it turned out that
the change wasn't that great an idea.

Andi Kleen:
i386: Update defconfig
x86: Use -maccumulate-outgoing-args
x86-64: Revert interrupt backlink changes
i386: Disable nmi watchdog on all ThinkPads
x86: Revert new unwind kernel stack termination
x86-64: Revert timer routing behaviour back to 2.6.16 state

Andrew Morton:
i386: fix .cfi_signal_frame copy-n-paste error

bibo,mao:
x86-64: x86_64 add NX mask for PTE entry

Corey Minyard:
x86-64: Fix for arch/x86_64/pci/Makefile CFLAGS

Eric W. Biederman:
x86-64: Use irq_domain in ioapic_retrigger_irq
x86-64: Put more than one cpu in TARGET_CPUS

Jan Beulich:
x86-64: Speed up dwarf2 unwinder
x86-64: Fix ENOSYS in system call tracing

Jeremy Fitzhardinge:
i386: Fix fake return address

keith mannthey:
x86-64: x86_64 hot-add memory srat.c fix

Vivek Goyal:
x86-64: fix page align in e820 allocator
x86-64: Overlapping program headers in physical addr space fix

Yinghai Lu:
x86-64: typo in __assign_irq_vector when updating pos for vector and offset

Makefile | 1
arch/i386/Makefile | 8 +
arch/i386/defconfig | 30 ++-
arch/i386/kernel/head.S | 2
arch/i386/kernel/nmi.c | 10 +
arch/i386/kernel/process.c | 6 -
arch/x86_64/Makefile | 4
arch/x86_64/kernel/e820.c | 14 +-
arch/x86_64/kernel/early-quirks.c | 9 +
arch/x86_64/kernel/entry.S | 10 -
arch/x86_64/kernel/genapic_flat.c | 2
arch/x86_64/kernel/io_apic.c | 15 +-
arch/x86_64/kernel/vmlinux.lds.S | 3
arch/x86_64/mm/srat.c | 4
arch/x86_64/pci/Makefile | 2
drivers/firmware/dmi_scan.c | 20 ++
include/asm-generic/vmlinux.lds.h | 16 ++
include/asm-x86_64/pgtable.h | 1
include/asm-x86_64/proto.h | 2
include/linux/dmi.h | 2
include/linux/unwind.h | 2
init/main.c | 1
kernel/unwind.c | 318 ++++++++++++++++++++++++++++++++-----
23 files changed, 385 insertions(+), 97 deletions(-)


2006-10-21 17:51:29

by Muli Ben-Yehuda

[permalink] [raw]
Subject: Re: [discuss] Please pull x86 tree

On Sat, Oct 21, 2006 at 06:44:06PM +0200, Andi Kleen wrote:
>
> Linus, please pull from
>
> git://one.firstfloor.org/home/andi/git/linux-2.6 for-linus
>
> These are all accumulated bug fixes for x86-64 and i386 and should
> be all pretty safe.
>
> The only thing that isn't a clear bug fix is the dwarf2 unwinder
> speedup -- i'm including that on popular demand because it fixes
> a serious performance regression with lockdep.

It would've been good to get the Calgary bug-fix in this
series... next one, please?

Cheers,
Muli

2006-10-21 18:15:43

by Andi Kleen

[permalink] [raw]
Subject: Re: [discuss] Please pull x86 tree

On Saturday 21 October 2006 19:51, Muli Ben-Yehuda wrote:
> On Sat, Oct 21, 2006 at 06:44:06PM +0200, Andi Kleen wrote:
> >
> > Linus, please pull from
> >
> > git://one.firstfloor.org/home/andi/git/linux-2.6 for-linus
> >
> > These are all accumulated bug fixes for x86-64 and i386 and should
> > be all pretty safe.
> >
> > The only thing that isn't a clear bug fix is the dwarf2 unwinder
> > speedup -- i'm including that on popular demand because it fixes
> > a serious performance regression with lockdep.
>
> It would've been good to get the Calgary bug-fix in this
> series... next one, please?

You said it wasn't critical and I only sent critical issues

-Andi

2006-10-21 19:41:27

by Muli Ben-Yehuda

[permalink] [raw]
Subject: Re: [discuss] Please pull x86 tree

On Sat, Oct 21, 2006 at 08:14:48PM +0200, Andi Kleen wrote:
> On Saturday 21 October 2006 19:51, Muli Ben-Yehuda wrote:
> > On Sat, Oct 21, 2006 at 06:44:06PM +0200, Andi Kleen wrote:
> > >
> > > Linus, please pull from
> > >
> > > git://one.firstfloor.org/home/andi/git/linux-2.6 for-linus
> > >
> > > These are all accumulated bug fixes for x86-64 and i386 and should
> > > be all pretty safe.
> > >
> > > The only thing that isn't a clear bug fix is the dwarf2 unwinder
> > > speedup -- i'm including that on popular demand because it fixes
> > > a serious performance regression with lockdep.
> >
> > It would've been good to get the Calgary bug-fix in this
> > series... next one, please?
>
> You said it wasn't critical and I only sent critical issues

You asked if it should go into 2.6.19 and I said it should. I also
said it's not critical, by which I meant "it doesn't need to go in
*right now*, it can wait for the next regularly-scheduled
patchset". Sorry about the confusion.

Thanks,
Muli

2006-10-21 19:56:59

by Jeremy Fitzhardinge

[permalink] [raw]
Subject: Re: Please pull x86 tree

Andi Kleen wrote:
> Jeremy Fitzhardinge:
> i386: Fix fake return address
>

I think this should only make a difference with the PDA tree applied.

J