i'm pleased to announce the v2.6.21-rt3 kernel, which can be downloaded 
from the usual place:

    http://redhat.com/~mingo/realtime-preempt/
 
more info about the -rt patchset can be found in the RT wiki:
 
    http://rt.wiki.kernel.org

This is a fixes-only release assembled by Thomas Gleixner. Many PPC 
fixes from Sergei Shtylyov, dynticks softirq fixes from Mikulas Patocka, 
a CONFIG_IRQSOFF_TIMING fix from Daniel Walker and x86_64 hrt fixes from 
Thomas Gleixner.

to build a 2.6.21-rt3 tree, the following patches should be applied:

  http://kernel.org/pub/linux/kernel/v2.6/linux-2.6.21.tar.bz2
  http://redhat.com/~mingo/realtime-preempt/patch-2.6.21-rt3

	Ingo
-

[ message continues ]

Hi,
I have tried to compile -rt3 patch for ppc (mpc8540ads) and it failed
with undefined referance errors on ppc/kernel/asm-offsets.c. Here is
the compiler outputs:

  HOSTCC  scripts/basic/fixdep
  HOSTCC  scripts/basic/docproc
  HOSTCC  scripts/kconfig/conf.o
  HOSTCC  scripts/kconfig/kxgettext.o
  SHIPPED scripts/kconfig/zconf.tab.c
  SHIPPED scripts/kconfig/lex.zconf.c
  SHIPPED scripts/kconfig/zconf.hash.c
  HOSTCC  scripts/kconfig/zconf.tab.o
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf -s arch/ppc/Kconfig
  CHK     include/linux/version.h
  UPD     include/linux/version.h
  CHK     include/linux/utsrelease.h
  UPD     include/linux/utsrelease.h
  SYMLINK include/asm -> include/asm-ppc
  CC      arch/ppc/kernel/asm-offsets.s
In file included from arch/ppc/include/asm/hw_irq.h:110,
                 from include/asm/system.h:10,
                 from include/linux/list.h:9,
                 from include/linux/signal.h:8,
                 from arch/ppc/kernel/asm-offsets.c:11:
include/linux/irqflags.h:92:1: warning: "raw_local_irq_save" redefined
In file included from include/asm/system.h:10,
                 from include/linux/list.h:9,
                 from include/linux/signal.h:8,
                 from arch/ppc/kernel/asm-offsets.c:11:
arch/ppc/include/asm/hw_irq.h:103:1: warning: this is the location of
the previous definition
In file included from arch/ppc/include/asm/hw_irq.h:110,
                 from include/asm/system.h:10,
                 from include/linux/list.h:9,
                 from include/linux/signal.h:8,
                 from arch/ppc/kernel/asm-offsets.c:11:
include/linux/irqflags.h:97:1: warning: "raw_local_irq_restore" redefined
In file included from include/asm/system.h:10,
                 from include/linux/list.h:9,
                 from include/linux/signal.h:8,
                 from arch/ppc/kernel/asm-offsets.c:11:
arch/ppc/include/asm/hw_irq.h:59:1: warning: this is the location of
the previous definition
In file included ...
[ message continues ]

arch/ppc is known to be broken. Does it even compile on 2.6.21 vanilla ?

	tglx


-

[ message continues ]

Yes, 2.6.21 was compiled and boot successfully for ppc, Is 85xx family
supported for powerpc arch?
Thanks.

-

[ message continues ]

Please do not top post !


I don't know. You might ask the folks on the ppc mailing list.

	tglx


-

[ message continues ]

E.g. check http://lxr.free-electrons.com/source/arch/powerpc/configs/. 
Even your board seems supported. But you need a boot loader aware of the 
  Open Firmware Flat Device Tree.

Wolfgang.

-

[ message continues ]

This is actually regarding v2.6.21-rt5 but I don't remember seeing an
announcement for that one. The attached patch is necessary if you happen
to have RTC_HISTOGRAM enabled, which I'm guessing most folks don't.

BTW, what was the consensus on pagefault_enable and pagefault_disable?

-- 
   kr
[ message continues ]

Where did the arm ep93xx changes come from? It looks like patch below
came in with them in 2.6.21-rt6 ..

--- linux.orig/include/asm-arm/system.h
+++ linux/include/asm-arm/system.h
@@ -235,8 +235,9 @@ static inline void set_copro_access(unsi
  * so enable interrupts over the context switch to avoid high
  * latency.
  */
+#if 0 // preempt_rt hack, because warn_on in shedule.c triggers
 #define __ARCH_WANT_INTERRUPTS_ON_CTXSW
-
+#endif
 /*
  * switch_to(prev, next) should switch from task `prev' to `next'
  * `prev' will never be the same as `next'.  schedule() itself


-

[ message continues ]

So what ?

	tglx


-

[ message continues ]

It's silencing a WARN_ON() with no real reason why it's happening, or
why/if it's safe to silence it .. Why is that there?

Daniel

-

[ message continues ]

Care to look at the code ? 

	tglx


-

[ message continues ]

I copied the patch into the first email, I did look at it which is why
you got an email. Your introducing the patch , it's for you to explain
why it needs to be there.. considering that the comment says "hack" in
it, I have to wonder what the non-hack version is ..

Daniel

-

[ message continues ]

you might not have realized it, but the problem i think was the attitude 

AFAICS Thomas just mirrored that snotty comment back to you, and yes, 
when you are on the receiving end it feels nasty, doesnt it? So please 
restart and rephrase that into a more polite form and you might get an 

You are asking Thomas a favor after all... Be a bit more polite when 
asking others, and a completely new world might open up for you :-)

	Ingo
-

[ message continues ]

I was not intending any attitude. I had no idea where those changes came

I don't read the sentence you quoted as snotty at all ..  I had no idea
where the changes came from, and I was simply asking that.. Since there
isn't a history for the -rt patch, I honestly don't know where some of
the code comes from ..

Since I had no idea that Thomas wrote that code, there's no premeditated

I don't see that as me asking a favor.. If a patch crosses LKML and I
ask "Why does this hunk exist?" that's not a favor. That is pure code
review, which is exactly what I was doing in this thread. Which is
reviewing code and asking questions..

I hope you don't read my comments above as snotty cause it's not suppose
to be, just conversation..

Daniel

-

[ message continues ]