All of lore.kernel.org
 help / color / mirror / Atom feed
* [Adeos-main] ADEOS with dcache write through enabled on ARM
@ 2007-02-08  5:14 lobby
  2007-02-08  8:52 ` Jan Kiszka
  0 siblings, 1 reply; 2+ messages in thread
From: lobby @ 2007-02-08  5:14 UTC (permalink / raw)
  To: adeos-main

Hello,

I read a suggestion Michael Neuhauser made on a web page to enable dcache
write through on arm9 processors to improve real time latency on the ARM
CPU.

I tried this on a cirrus EP9302 (ARM 920T) on a TS-7400 board and it does
improve the worst case latency dramaticaly but it causes ramdom
segmentation violations (init process on a idle machine and also user
started processes)

I was wondering if any one has heard of problems on this specific processor
when CONFIG_CPU_DCACHE_WRITETHROUGH=y with linux kernel 2.4.26 and ADEOS

The problem only happens when ADEOS patches are applied together with
CONFIG_CPU_DCACHE_WRITETHROUGH=y. Having the dcache write through enabled
without adeos patches works fine.

Please let me know if I missed something.


Thank You

Patrick Robin



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [Adeos-main] ADEOS with dcache write through enabled on ARM
  2007-02-08  5:14 [Adeos-main] ADEOS with dcache write through enabled on ARM lobby
@ 2007-02-08  8:52 ` Jan Kiszka
  0 siblings, 0 replies; 2+ messages in thread
From: Jan Kiszka @ 2007-02-08  8:52 UTC (permalink / raw)
  To: lobby; +Cc: adeos-main

[-- Attachment #1: Type: text/plain, Size: 1233 bytes --]

lobby@domain.hid wrote:
> Hello,
> 
> I read a suggestion Michael Neuhauser made on a web page to enable dcache
> write through on arm9 processors to improve real time latency on the ARM
> CPU.
> 
> I tried this on a cirrus EP9302 (ARM 920T) on a TS-7400 board and it does
> improve the worst case latency dramaticaly but it causes ramdom
> segmentation violations (init process on a idle machine and also user
> started processes)
> 
> I was wondering if any one has heard of problems on this specific processor
> when CONFIG_CPU_DCACHE_WRITETHROUGH=y with linux kernel 2.4.26 and ADEOS

That means you are using a really old and outdated version of Adeos. Do
you have the chance to test a recent 2.6.19 I-pipe patch as well?

I guess many people are interested in reducing the poor ARM latencies
and would be happy to work on fixing remaining quirks, but development
now takes place around I-pipe, using Xenomai.

> 
> The problem only happens when ADEOS patches are applied together with
> CONFIG_CPU_DCACHE_WRITETHROUGH=y. Having the dcache write through enabled
> without adeos patches works fine.
> 
> Please let me know if I missed something.
> 
> 
> Thank You
> 
> Patrick Robin
> 

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 250 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2007-02-08  8:52 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-02-08  5:14 [Adeos-main] ADEOS with dcache write through enabled on ARM lobby
2007-02-08  8:52 ` Jan Kiszka

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.