* x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi
@ 2007-01-27 1:24 Andrew Morton
2007-01-27 9:29 ` Len Brown
0 siblings, 1 reply; 6+ messages in thread
From: Andrew Morton @ 2007-01-27 1:24 UTC (permalink / raw)
To: Len Brown, Andi Kleen, linux-acpi
The new stuff which just landed in Len's tree caused a huge mess in
arch/x86_64/kernel/genapic.c:clustered_apic_check() when applying
ftp://ftp.firstfloor.org/pub/ak/x86_64/quilt/patches/always-use-physical-delivery-mode-on-8-cpus
on top of it.
In fact the ACPI change has trashed a fair slice of Andi's pending tree.
I think I'll revert to yesterday's git-acpi, let you guys sort it all out.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi
2007-01-27 1:24 x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi Andrew Morton
@ 2007-01-27 9:29 ` Len Brown
2007-01-27 9:38 ` Andrew Morton
2007-01-27 22:20 ` Andrew Morton
0 siblings, 2 replies; 6+ messages in thread
From: Len Brown @ 2007-01-27 9:29 UTC (permalink / raw)
To: Andrew Morton; +Cc: Andi Kleen, linux-acpi
On Friday 26 January 2007 20:24, Andrew Morton wrote:
>
> The new stuff which just landed in Len's tree caused a huge mess in
> arch/x86_64/kernel/genapic.c:clustered_apic_check() when applying
> ftp://ftp.firstfloor.org/pub/ak/x86_64/quilt/patches/always-use-physical-delivery-mode-on-8-cpus
> on top of it.
The ACPI change this cleanup patch is conflicting with is quite small:
------------------------- arch/x86_64/kernel/genapic.c -------------------------
index b007433..0b3603a 100644
@@ -58,8 +58,8 @@ void __init clustered_apic_check(void)
* Some x86_64 machines use physical APIC mode regardless of how many
* procs/clusters are present (x86_64 ES7000 is an example).
*/
- if (acpi_fadt.revision > FADT2_REVISION_ID)
- if (acpi_fadt.force_apic_physical_destination_mode) {
+ if (acpi_gbl_FADT.header.revision > FADT2_REVISION_ID)
+ if (acpi_gbl_FADT.flags & ACPI_FADT_APIC_PHYSICAL) {
genapic = &apic_cluster;
goto print;
}
> In fact the ACPI change has trashed a fair slice of Andi's pending tree.
>
> I think I'll revert to yesterday's git-acpi, let you guys sort it all out.
Send me a version of the always-use-physical-delivery-mode-on-8-cpus patch
that applies to Linus' tree (the one above doesn't), and I'll be happy to apply
the 2-line diff above to it.
-Len
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi
2007-01-27 9:29 ` Len Brown
@ 2007-01-27 9:38 ` Andrew Morton
2007-01-27 22:20 ` Andrew Morton
1 sibling, 0 replies; 6+ messages in thread
From: Andrew Morton @ 2007-01-27 9:38 UTC (permalink / raw)
To: Len Brown; +Cc: Andi Kleen, linux-acpi
On Sat, 27 Jan 2007 04:29:57 -0500
Len Brown <lenb@kernel.org> wrote:
> On Friday 26 January 2007 20:24, Andrew Morton wrote:
> >
> > The new stuff which just landed in Len's tree caused a huge mess in
> > arch/x86_64/kernel/genapic.c:clustered_apic_check() when applying
> > ftp://ftp.firstfloor.org/pub/ak/x86_64/quilt/patches/always-use-physical-delivery-mode-on-8-cpus
> > on top of it.
>
> The ACPI change this cleanup patch is conflicting with is quite small:
Oh, OK. I'll take another lok at it if/when I manage to get next -mm out
the door (over 40 compilation fixes needed so far).
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi
2007-01-27 9:29 ` Len Brown
2007-01-27 9:38 ` Andrew Morton
@ 2007-01-27 22:20 ` Andrew Morton
2007-01-29 2:58 ` Len Brown
1 sibling, 1 reply; 6+ messages in thread
From: Andrew Morton @ 2007-01-27 22:20 UTC (permalink / raw)
To: Len Brown; +Cc: Andi Kleen, linux-acpi
On Sat, 27 Jan 2007 04:29:57 -0500
Len Brown <lenb@kernel.org> wrote:
> On Friday 26 January 2007 20:24, Andrew Morton wrote:
> >
> > The new stuff which just landed in Len's tree caused a huge mess in
> > arch/x86_64/kernel/genapic.c:clustered_apic_check() when applying
> > ftp://ftp.firstfloor.org/pub/ak/x86_64/quilt/patches/always-use-physical-delivery-mode-on-8-cpus
> > on top of it.
>
> The ACPI change this cleanup patch is conflicting with is quite small:
>
> ------------------------- arch/x86_64/kernel/genapic.c -------------------------
> index b007433..0b3603a 100644
> @@ -58,8 +58,8 @@ void __init clustered_apic_check(void)
> * Some x86_64 machines use physical APIC mode regardless of how many
> * procs/clusters are present (x86_64 ES7000 is an example).
> */
> - if (acpi_fadt.revision > FADT2_REVISION_ID)
> - if (acpi_fadt.force_apic_physical_destination_mode) {
> + if (acpi_gbl_FADT.header.revision > FADT2_REVISION_ID)
> + if (acpi_gbl_FADT.flags & ACPI_FADT_APIC_PHYSICAL) {
> genapic = &apic_cluster;
> goto print;
> }
> > In fact the ACPI change has trashed a fair slice of Andi's pending tree.
> >
> > I think I'll revert to yesterday's git-acpi, let you guys sort it all out.
>
> Send me a version of the always-use-physical-delivery-mode-on-8-cpus patch
> that applies to Linus' tree (the one above doesn't), and I'll be happy to apply
> the 2-line diff above to it.
>
OK, so I took another look. This ACPI update does extensive damage to
Andi's pending queue. I fixed four patches, dropped four or five more, hit
more problems then gave up again. I'll go back to Thursday's git-acpi
again.
Longer-term, I expect Andi will merge before acpi does, and you're looking
at a fairly large amount of fixing after that happens.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi
2007-01-27 22:20 ` Andrew Morton
@ 2007-01-29 2:58 ` Len Brown
2007-01-29 6:38 ` Andrew Morton
0 siblings, 1 reply; 6+ messages in thread
From: Len Brown @ 2007-01-29 2:58 UTC (permalink / raw)
To: Andrew Morton; +Cc: Andi Kleen, linux-acpi
On Saturday 27 January 2007 17:20, Andrew Morton wrote:
> On Sat, 27 Jan 2007 04:29:57 -0500
> Len Brown <lenb@kernel.org> wrote:
>
> > On Friday 26 January 2007 20:24, Andrew Morton wrote:
> > >
> > > The new stuff which just landed in Len's tree caused a huge mess in
> > > arch/x86_64/kernel/genapic.c:clustered_apic_check() when applying
> > > ftp://ftp.firstfloor.org/pub/ak/x86_64/quilt/patches/always-use-physical-delivery-mode-on-8-cpus
> > > on top of it.
> >
> > The ACPI change this cleanup patch is conflicting with is quite small:
> >
> > ------------------------- arch/x86_64/kernel/genapic.c -------------------------
> > index b007433..0b3603a 100644
> > @@ -58,8 +58,8 @@ void __init clustered_apic_check(void)
> > * Some x86_64 machines use physical APIC mode regardless of how many
> > * procs/clusters are present (x86_64 ES7000 is an example).
> > */
> > - if (acpi_fadt.revision > FADT2_REVISION_ID)
> > - if (acpi_fadt.force_apic_physical_destination_mode) {
> > + if (acpi_gbl_FADT.header.revision > FADT2_REVISION_ID)
> > + if (acpi_gbl_FADT.flags & ACPI_FADT_APIC_PHYSICAL) {
> > genapic = &apic_cluster;
> > goto print;
> > }
> > > In fact the ACPI change has trashed a fair slice of Andi's pending tree.
> > >
> > > I think I'll revert to yesterday's git-acpi, let you guys sort it all out.
> >
> > Send me a version of the always-use-physical-delivery-mode-on-8-cpus patch
> > that applies to Linus' tree (the one above doesn't), and I'll be happy to apply
> > the 2-line diff above to it.
> >
>
> OK, so I took another look. This ACPI update does extensive damage to
> Andi's pending queue. I fixed four patches, dropped four or five more, hit
> more problems then gave up again. I'll go back to Thursday's git-acpi
> again.
>
> Longer-term, I expect Andi will merge before acpi does, and you're looking
> at a fairly large amount of fixing after that happens.
Exactly what patches from Andi do I need to merge on top of to get the ACPI tree into -mm?
thanks,
-Len
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi
2007-01-29 2:58 ` Len Brown
@ 2007-01-29 6:38 ` Andrew Morton
0 siblings, 0 replies; 6+ messages in thread
From: Andrew Morton @ 2007-01-29 6:38 UTC (permalink / raw)
To: Len Brown; +Cc: Andi Kleen, linux-acpi
On Sun, 28 Jan 2007 21:58:12 -0500
Len Brown <lenb@kernel.org> wrote:
> > OK, so I took another look. This ACPI update does extensive damage to
> > Andi's pending queue. I fixed four patches, dropped four or five more, hit
> > more problems then gave up again. I'll go back to Thursday's git-acpi
> > again.
> >
> > Longer-term, I expect Andi will merge before acpi does, and you're looking
> > at a fairly large amount of fixing after that happens.
>
> Exactly what patches from Andi do I need to merge on top of to get the ACPI tree into -mm?
ftp://ftp.firstfloor.org/pub/ak/x86_64/quilt/patches/ is everything.
In rc6-mm1 that's everything between #X86_64-START and #X86_64-END in the
series file. I have another 35-odd x86 patches on top of that lot.
Note that I normally stage the x86_64 tree _after_ the ACPI tree, but I can
change that.
I don't think that making the acpi tree dependent upon the x86_64 tree
would be a smart idea, frankly. It'll cause you heaps of pain.
Perhaps Andi could stick his tree into git for you to base off, dunno. But
he doesn't seem to be around at present.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2007-01-29 6:39 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-01-27 1:24 x86_64-mm-always-use-physical-delivery-mode-on-8-cpus vs git-acpi Andrew Morton
2007-01-27 9:29 ` Len Brown
2007-01-27 9:38 ` Andrew Morton
2007-01-27 22:20 ` Andrew Morton
2007-01-29 2:58 ` Len Brown
2007-01-29 6:38 ` Andrew Morton
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox