Hi,

I noticed hight CPU usage when using the harddrive on my Sun Blade 100
(Solaris 8), which leads to the assumtion that the IDE-DMA is switched off.
How can it be switched on? I googled two days and found nothing. Can it be
done? And how?

TIA
Alex

In <3f798203_1@news.arcor-ip.de> "Alexandros Gougousoudis" <nospam_ag@aloha-systeme.de> writes:
>I noticed hight CPU usage when using the harddrive on my Sun Blade 100
>(Solaris 8), which leads to the assumtion that the IDE-DMA is switched off.
>How can it be switched on? I googled two days and found nothing. Can it be
>done? And how?

you have to switch it off as it is on by default. check
the dcd_options property of your ide device node using
prtconf -v. it should probably be 0xa5 if everything is
set to the default values.

Hi,

<ultrasparc3@hotmail.com> schrieb

> you have to switch it off as it is on by default. check

Hm I read somewhere that it is switched off by default, because of an old
BIOS or so (or was ist Sol x86)?

> the dcd_options property of your ide device node using
> prtconf -v. it should probably be 0xa5 if everything is
> set to the default values.

I have for target0-dcd-options a 0xa4, for target1-dcd-options a 0xa2 and
for target2-dcd-options a 0xa4 and a separate dcd-options which is 0xa5.
What does it all mean?

I found no doc, man prtconf is to thin.

TIA
Alex

>> you have to switch it off as it is on by default. check
>
>Hm I read somewhere that it is switched off by default, because of an old
>BIOS or so (or was ist Sol x86)?

SPARC-based platforms do not have BIOS, you must be thinking of
Solaris x86.

>> the dcd_options property of your ide device node using
>> prtconf -v. it should probably be 0xa5 if everything is
>> set to the default values.

>I have for target0-dcd-options a 0xa4, for target1-dcd-options a 0xa2 and
>for target2-dcd-options a 0xa4 and a separate dcd-options which is 0xa5.
>What does it all mean?

various levels of UltraDMA are enabled for those devices. 0xa5 is the
highest level the controller supports, the per-target values are the
negiotiated levels for the specific devices on that controller.

>I found no doc, man prtconf is to thin.

the man page for prtconf can only describe the prtconf command and
its options. the output is the operating system's device tree which
is not appropriate to describe in the prtconf man page, the contents
vary based on operating system version, platform, and installed devices.

In article <200391.10351.8854@cable.prodigy.com>,
ultrasparc3@hotmail.com writes:
>>> you have to switch it off as it is on by default. check
>>
>>Hm I read somewhere that it is switched off by default, because of an old
>>BIOS or so (or was ist Sol x86)?
>
> SPARC-based platforms do not have BIOS, you must be thinking of
> Solaris x86.
>
>>> the dcd_options property of your ide device node using
>>> prtconf -v. it should probably be 0xa5 if everything is
>>> set to the default values.
>
>>I have for target0-dcd-options a 0xa4, for target1-dcd-options a 0xa2 and
>>for target2-dcd-options a 0xa4 and a separate dcd-options which is 0xa5.
>>What does it all mean?
>
> various levels of UltraDMA are enabled for those devices. 0xa5 is the
> highest level the controller supports, the per-target values are the
> negiotiated levels for the specific devices on that controller.

I found a breakout but I can't quote the details; however, the low order
three bits only pertain to the PIO mode, and the high bit indicates that
DMA, not PIO mode is in effect, so the difference between 0xa2, 0xa4, and
0xa5 shouldn't matter (if I understand correctly).

>>I found no doc, man prtconf is to thin.
>
> the man page for prtconf can only describe the prtconf command and
> its options. the output is the operating system's device tree which
> is not appropriate to describe in the prtconf man page, the contents
> vary based on operating system version, platform, and installed devices.

--
mailto:rlhamil@mindwarp.smart.net http://www.smart.net/~rlhamil

In <vno5sn8ou0eb2d@corp.supernews.com> Richard.L.Hamilton@mindwarp.smart.net (Richard L. Hamilton) writes:
>I found a breakout but I can't quote the details; however, the low order
>three bits only pertain to the PIO mode

not quite right as the bottom bits have different meanings depending on what
the high-order bits are set to. if the high-order bits claim "ultra dma"
then the low-order bits are the ultra-dma level. the high-order bits
need to be zero for the low-order bits to contain the PIO mode.

In article <200392.195445.18166@cable.prodigy.com>,
ultrasparc3@hotmail.com writes:
> In <vno5sn8ou0eb2d@corp.supernews.com> Richard.L.Hamilton@mindwarp.smart.net (Richard L. Hamilton) writes:
>>I found a breakout but I can't quote the details; however, the low order
>>three bits only pertain to the PIO mode
>
> not quite right as the bottom bits have different meanings depending on what
> the high-order bits are set to. if the high-order bits claim "ultra dma"
> then the low-order bits are the ultra-dma level. the high-order bits
> need to be zero for the low-order bits to contain the PIO mode.

What version of Solaris are you looking at? I admit I didn't look very
far, but I didn't see that looking at what I was looking at (early Solaris
8).

--
mailto:rlhamil@mindwarp.smart.net http://www.smart.net/~rlhamil

In <vnqn4la0921e4e@corp.supernews.com> Richard.L.Hamilton@mindwarp.smart.net (Richard L. Hamilton) writes:
>What version of Solaris are you looking at?

Solaris 9, but the meaning of dcd_options has not changed from Solaris 8.