Bochs Bochs | 2 Aug 18:58 2004
Picon

Re: How about HP iPAQ h1940 with Samsung S3C2410 CPU ?

On Monday 02 August 2004 21:07, you wrote:
> On Aug 2, 2004, at 12:03 AM, Bochs Bochs wrote:
> > Does NetBSD run on HP iPAQ h1940 with Samsung processor ?
> > What files/downloads i need to install NetBSD ?
>
> I'm not sure which Samsung processor that system uses.
>
> Best to post your question to "port-arm <at> netbsd.org", rather than
> sending to the port maintainer address directly.
>
>          -- Jason R. Thorpe <thorpej <at> wasabisystems.com>
Platform: hp iPAQ h1940 (Pocket PC - ARM?)
CPU: Samsung S3C2410, system Flash ROM 32Mb, system RAM 64Mb
Equipment: TFT screen 320x240x64k, microphone, speaker,
   SD-card(with SDIO), IRDA SIR port, Bluetooth, USB.
Now running Windows CE 4.20 :-(

Richard Earnshaw | 7 Aug 13:58 2004
Picon

Adding UVMHIST tracking to arm/fault.c

There are a large number of prefetch and data abort faults on ARM that
never end up in the uvm code, so don't get logged at all in UVMHIST. 
However, it's often useful to know about all the faults that are
occurring.  I found the attached patch invaluable when tracking down a
problem recently.

Does anybody think I should *not* commit the attached?

R.

Attachment (arm.diffs): application/octet-stream, 5508 bytes
Jason Thorpe | 7 Aug 18:46 2004

Re: Adding UVMHIST tracking to arm/fault.c


On Aug 7, 2004, at 4:58 AM, Richard Earnshaw wrote:

> Does anybody think I should *not* commit the attached?

Looks good to me.  I'd say "go for it!"

         -- Jason R. Thorpe <thorpej <at> wasabisystems.com>

Richard Earnshaw | 20 Aug 11:06 2004
Picon

Heads up: Thumb code working on NetBSD

This is just a quick heads up to let folks know that I now have a NetBSD
kernel running on my integrator board (with ARM10e) that supports
applications compiled as Thumb code.

It's stable enough that I've been able to debug and complete a bootstrap
of gcc (trunk) compiling to Thumb code.  The size savings are quite
impressive: 

On my shark the size of the cc1 binary is

shark1:egcs [759] $ size gcc/cc1
   text    data     bss     dec     hex filename
3677285    6432  330916 4014633  3d4229 gcc/cc1

Whereas on the integrator we see
integrator:gcc [611] $ size gcc/cc1
   text    data     bss     dec     hex filename
2727208    6392  330496 3064096  2ec120 cc1

which represents a 30.2% saving in code size.

On v5t processors the Thumb binaries can make use of the same shared
libraries as ARM binaries (provided that the libraries are compiled for
v5t).  The limited natural inteworking ability of v4t processors means
that I don't intend to back-port this work to the older cores.

I'll be starting to commit the changes shortly; most of them are pretty
straight-forward.

R.
(Continue reading)

Charles M. Hannum | 20 Aug 14:32 2004

Re: Heads up: Thumb code working on NetBSD

On Friday 20 August 2004 09:06, Richard Earnshaw wrote:
> This is just a quick heads up to let folks know that I now have a NetBSD
> kernel running on my integrator board (with ARM10e) that supports
> applications compiled as Thumb code.

Excellent!  Maybe I'll dig out the Integrator after I move.  :-)

Steve Woodford | 20 Aug 14:59 2004
Picon

Re: Heads up: Thumb code working on NetBSD

On Friday 20 August 2004 10:06, Richard Earnshaw wrote:
> This is just a quick heads up to let folks know that I now have a
> NetBSD kernel running on my integrator board (with ARM10e) that
> supports applications compiled as Thumb code.

Thumbs up to that!

Cheers, Steve

Gavan Fantom | 20 Aug 18:08 2004

Re: Heads up: Thumb code working on NetBSD

On Fri, 20 Aug 2004, Richard Earnshaw wrote:

> It's stable enough that I've been able to debug and complete a bootstrap
> of gcc (trunk) compiling to Thumb code.  The size savings are quite
> impressive:
>
> On my shark the size of the cc1 binary is
>
> shark1:egcs [759] $ size gcc/cc1
>   text    data     bss     dec     hex filename
> 3677285    6432  330916 4014633  3d4229 gcc/cc1
>
> Whereas on the integrator we see
> integrator:gcc [611] $ size gcc/cc1
>   text    data     bss     dec     hex filename
> 2727208    6392  330496 3064096  2ec120 cc1
>
> which represents a 30.2% saving in code size.

Nice.

Does this space saving come with a performance gain or a performance hit?

--

-- 
Gillette - the best a man can forget

Richard Earnshaw | 20 Aug 18:37 2004
Picon

Re: Heads up: Thumb code working on NetBSD

On Fri, 2004-08-20 at 17:08, Gavan Fantom wrote:
> On Fri, 20 Aug 2004, Richard Earnshaw wrote:
> 
> > It's stable enough that I've been able to debug and complete a bootstrap
> > of gcc (trunk) compiling to Thumb code.  The size savings are quite
> > impressive:
> >
> > On my shark the size of the cc1 binary is
> >
> > shark1:egcs [759] $ size gcc/cc1
> >   text    data     bss     dec     hex filename
> > 3677285    6432  330916 4014633  3d4229 gcc/cc1
> >
> > Whereas on the integrator we see
> > integrator:gcc [611] $ size gcc/cc1
> >   text    data     bss     dec     hex filename
> > 2727208    6392  330496 3064096  2ec120 cc1
> >
> > which represents a 30.2% saving in code size.
> 
> Nice.
> 
> Does this space saving come with a performance gain or a performance hit?

It's normally a slight hit.  30% space saving when your instructions are
50% shorter means you must execute more instructions (about 40% more),
but on the other hand you get better I-cache utilization, so it's not
quite as bad as that in practice (in fact, some applications can end up
being faster since the entire critical code can fit in the cache).

(Continue reading)

Hal Murray | 14 Aug 09:34 2004

Shark linker status

Any news on the linker bug that breaks current?  (at least I think it's a 
linker bug)

I assume it's high on the list for getting ready for 2.0  Anything I can do 
to help?

--

-- 
The suespammers.org mail server is located in California.  So are all my
other mailboxes.  Please do not send unsolicited bulk e-mail or unsolicited
commercial e-mail to my suespammers.org address or any of my other addresses.
These are my opinions, not necessarily my employer's.  I hate spam.

Ahmad M.Afuni | 14 Aug 21:33 2004
Picon

Re: How about HP iPAQ h1940 with Samsung S3C2410 CPU ?

Hello,
The CPU is supported, but the onboard peripherals I don't know..
At the moment I'm working on the iPAQ H55xx, if I find a method of booting up NetBSD, I'll contact you.
Good luck,
Ahmad

On Monday 02 August 2004 21:07, you wrote:
> On Aug 2, 2004, at 12:03 AM, Bochs Bochs wrote:
> > Does NetBSD run on HP iPAQ h1940 with Samsung processor ?
> > What files/downloads i need to install NetBSD ?
>
> I'm not sure which Samsung processor that system uses.
>
> Best to post your question to "port-arm <at> netbsd.org", rather than
> sending to the port maintainer address directly.
>
>          -- Jason R. Thorpe <thorpej <at> wasabisystems.com>
Platform: hp iPAQ h1940 (Pocket PC - ARM?)
CPU: Samsung S3C2410, system Flash ROM 32Mb, system RAM 64Mb
Equipment: TFT screen 320x240x64k, microphone, speaker,
   SD-card(with SDIO), IRDA SIR port, Bluetooth, USB.
Now running Windows CE 4.20 :-(

_____________________________________________________________
yourname <at> digiverse.net
http://www.digiverse.net


Gmane