Status of -rt for linux-2.6.20/arch/powerpc

DISCUSSION Wolfgang Grandegger 3 participants 8 messages Feb 11, 2007

Share & tools

Wolfgang Grandegger

author

2007-02-11 11:09 UTC #0 lore raw

Hello,

what is the status of the Preemptive Kernel and the High Resolution Timers for Linux 2.6.20 for PowerPC (arch/powerpc or arch/ppc)?

I realized build problems when trying to get it working for my Lite5200 (MPC5200) board with patch-2-6-20-rt5. The patch seems to contain reminders of merge conflicts, e.g.:

  +++ linux/include/asm-powerpc/hw_irq.h
  ...
  +<<<<<<< delete extern unsigned long local_get_flags(void);
  +<<<<<<< delete extern unsigned long local_irq_disable(void);

and the make fails in arch/powerpc/kernel/time.c as well (see attachment).

Thanks for help.

Wolfgang.

Robert Schwebel

2007-02-11 11:29 UTC #1 lore raw

Hello Wolfgang,

Quoted 7 lines · Wolfgang Grandegger

what is the status of the Preemptive Kernel and the High Resolution Timers for Linux 2.6.20 for PowerPC (arch/powerpc or arch/ppc)?
I realized build problems when trying to get it working for my Lite5200 (MPC5200) board with patch-2-6-20-rt5. The patch seems to contain reminders of merge conflicts, e.g.:

Quoted 6 lines

  +++ linux/include/asm-powerpc/hw_irq.h
  ...
  +<<<<<<< delete extern unsigned long local_get_flags(void);
  +<<<<<<< delete extern unsigned long local_irq_disable(void);

and the make fails in arch/powerpc/kernel/time.c as well (see attachment).

We have 2.6.20-rt3 running on the phyCORE-MPC5200B, so that one should work and the merge problems did probably come in later (don't have the time to look right now because of the Embedded World activities for next week).

Robert

Signature · 5 lines

 Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de
 Pengutronix - Linux Solutions for Science and Industry
   Handelsregister:  Amtsgericht Hildesheim, HRA 2686
     Hannoversche Str. 2, 31134 Hildesheim, Germany
   Phone: +49-5121-206917-0 |  Fax: +49-5121-206917-9

Wolfgang Grandegger

author

2007-02-11 12:39 UTC #2 lore raw

Hi Robert,

Quoted 12 lines · Robert Schwebel

Hello Wolfgang,
Quoted 9 lines · Wolfgang Grandegger
what is the status of the Preemptive Kernel and the High Resolution Timers for Linux 2.6.20 for PowerPC (arch/powerpc or arch/ppc)?
I realized build problems when trying to get it working for my Lite5200 (MPC5200) board with patch-2-6-20-rt5. The patch seems to contain reminders of merge conflicts, e.g.:
  +++ linux/include/asm-powerpc/hw_irq.h
  ...
  +<<<<<<< delete extern unsigned long local_get_flags(void);
  +<<<<<<< delete extern unsigned long local_irq_disable(void);

Quoted 6 lines

Quoted 1 line
and the make fails in arch/powerpc/kernel/time.c as well (see attachment).
We have 2.6.20-rt3 running on the phyCORE-MPC5200B, so that one should work and the merge problems did probably come in later (don't have the time to look right now because of the Embedded World activities for next week).

As I see it, 2.6.20-rt3 has the same issues (have now attached the missing make output). I fixed them manually and are now able to boot and run the 2.6.20-rt5 on my Icecube board. Nevertheless, cyclictest returns high values of around 8000-10000 (8-10ms). The fact that the unknown field 'is_continuous' in arch/powerpc/kernel/time.c is still present gives me the impression, that the code in there is not up-to-date. See you next week on the Embedded World.

Thanks.

Wolfgang.

CC arch/powerpc/kernel/time.o
In file included from arch/powerpc/kernel/time.c:924:
include/linux/clocksource.h: In function 'ns2cyc':
include/linux/clocksource.h:174: warning: comparison of distinct pointer types lacks a cast include/linux/clocksource.h:174: warning: right shift count >= width of type include/linux/clocksource.h:174: warning: passing argument 1 of '__div64_32' from incompatible pointer type arch/powerpc/kernel/time.c: At top level:
arch/powerpc/kernel/time.c:938: error: unknown field 'is_continuous' specified in initializer make[1]: *** [arch/powerpc/kernel/time.o] Error 1
make[1]: Target `__build' not remade because of errors.
make: *** [arch/powerpc/kernel] Error 2

Thomas Gleixner

2007-02-11 15:57 UTC #3 lore raw

Wolfgang,

Quoted 7 lines · Wolfgang Grandegger

As I see it, 2.6.20-rt3 has the same issues (have now attached the missing make output). I fixed them manually and are now able to boot and run the 2.6.20-rt5 on my Icecube board. Nevertheless, cyclictest returns high values of around 8000-10000 (8-10ms). The fact that the unknown field 'is_continuous' in arch/powerpc/kernel/time.c is still present gives me the impression, that the code in there is not up-to-date. See you next week on the Embedded World.

Yup, I did not come around to do the fixups. Is top on my todo list though.

tglx

Wolfgang Grandegger

author

2007-02-11 16:37 UTC #4 lore raw

Quoted 7 lines · Thomas Gleixner

Wolfgang,
Quoted 1 line · Wolfgang Grandegger
As I see it, 2.6.20-rt3 has the same issues (have now attached the missing make output). I fixed them manually and are now able to boot and run the 2.6.20-rt5 on my Icecube board. Nevertheless, cyclictest returns high values of around 8000-10000 (8-10ms). The fact that the unknown field 'is_continuous' in arch/powerpc/kernel/time.c is still present gives me the impression, that the code in there is not up-to-date. See you next week on the Embedded World.
Yup, I did not come around to do the fixups. Is top on my todo list though.

Apart from that, I also realized that it's not possible to enable HIGH_RES_TIMERS, because GENERIC_CLOCKEVENTS is not selected or available for PowerPC.

Wolfgang.

Thomas Gleixner

2007-02-11 16:39 UTC #5 lore raw

Quoted 6 lines · Wolfgang Grandegger

Quoted 2 lines
Yup, I did not come around to do the fixups. Is top on my todo list though.
Apart from that, I also realized that it's not possible to enable HIGH_RES_TIMERS, because GENERIC_CLOCKEVENTS is not selected or available for PowerPC.

Yup, the whole bunch is missing. It got dropped at some point.

tglx

Robert Schwebel

2007-02-11 15:59 UTC #6 lore raw

In reply to Wolfgang Grandegger

Quoted 5 lines · Wolfgang Grandegger

Quoted 4 lines
We have 2.6.20-rt3 running on the phyCORE-MPC5200B, so that one should work and the merge problems did probably come in later (don't have the time to look right now because of the Embedded World activities for next week).

Quoted 7 lines

As I see it, 2.6.20-rt3 has the same issues (have now attached the missing make output). I fixed them manually and are now able to boot and run the 2.6.20-rt5 on my Icecube board. Nevertheless, cyclictest returns high values of around 8000-10000 (8-10ms). The fact that the unknown field 'is_continuous' in arch/powerpc/kernel/time.c is still present gives me the impression, that the code in there is not up-to-date. See you next week on the Embedded World.

I have uploaded the latest results to this place:

http://www.pengutronix.de/oselas/realtime/results/20070211-1/phyCORE-MPC5200B-tiny-20070211-1.png

Note that this is still no waterproof measurement, with 100k data points and only ad-hoc stressing. We'll do more testing after the tradeshow.

The 2.6.20-rt3 is a little bit patched, but I don't have the time right now to dig into the details. We can discuss it on the Embedded World next week, I'll take the patch stack with me.

Robert

Signature · 5 lines

 Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de
 Pengutronix - Linux Solutions for Science and Industry
   Handelsregister:  Amtsgericht Hildesheim, HRA 2686
     Hannoversche Str. 2, 31134 Hildesheim, Germany
   Phone: +49-5121-206917-0 |  Fax: +49-5121-206917-9

Wolfgang Grandegger

author

2007-02-11 16:39 UTC #7 lore raw

Quoted 11 lines · Robert Schwebel

Quoted 5 lines · Wolfgang Grandegger
Quoted 4 lines
We have 2.6.20-rt3 running on the phyCORE-MPC5200B, so that one should work and the merge problems did probably come in later (don't have the time to look right now because of the Embedded World activities for next week).
As I see it, 2.6.20-rt3 has the same issues (have now attached the missing make output). I fixed them manually and are now able to boot and run the 2.6.20-rt5 on my Icecube board. Nevertheless, cyclictest returns high values of around 8000-10000 (8-10ms). The fact that the unknown field 'is_continuous' in arch/powerpc/kernel/time.c is still present gives me the impression, that the code in there is not up-to-date. See you next week on the Embedded World.
I have uploaded the latest results to this place:
http://www.pengutronix.de/oselas/realtime/results/20070211-1/phyCORE-MPC5200B-tiny-20070211-1.png

Quoted 7 lines

Note that this is still no waterproof measurement, with 100k data points and only ad-hoc stressing. We'll do more testing after the tradeshow.
The 2.6.20-rt3 is a little bit patched, but I don't have the time right now to dig into the details. We can discuss it on the Embedded World next week, I'll take the patch stack with me.

OK, great, thanks.

See you next week.

Wolfgang.