idebus=133 ?

idebus=133 ?

[Dave Ray]

It's my understanding the Mac G4 towers run at a bus
speed of 133 MHz. However the kernel argument for
idebus=xx apparently only allows values up to 66. Is
this intentional?

I checked where this was is the kernel source and
manually changed the limit from 66 to to 133 to see
what happens. The machine boots fine with no errors.
The file is: drivers/ide/ide.c. On lines 3159 and 3162
I simply changed "66" to "133". Then I boot with
"idebus=133".

Is the 66 limit intentional or should it be changed to
133?

-Dave


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Tom Rini]

On Sun, Oct 21, 2001 at 03:42:11PM -0700, Dave Ray wrote:

> It's my understanding the Mac G4 towers run at a bus
> speed of 133 MHz. However the kernel argument for
> idebus=xx apparently only allows values up to 66. Is
> this intentional?

idebus != 'bus'.  The 133 referrs to the speed the memory and cpu can
talk at.  Your idebus setting shouldn't change from the default.
>From Documentation/ide.txt:
"idebus=xx"            : inform IDE driver of VESA/PCI bus speed in MHz,
                                where "xx" is between 20 and 66 inclusive,
                                used when tuning chipset PIO modes.
                                For PCI bus, 25 is correct for a P75 system,
                                30 is correct for P90,P120,P180 systems,
                                and 33 is used for P100,P133,P166 systems.
                                If in doubt, use idebus=33 for PCI.
                                As for VLB, it is safest to not specify it.
                                Bigger values are safer than smaller ones.

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Benjamin Herrenschmidt]

>
>It's my understanding the Mac G4 towers run at a bus
>speed of 133 MHz. However the kernel argument for
>idebus=xx apparently only allows values up to 66. Is
>this intentional?

I don't think this value has any meaning on macs. It's
related to the IDE bus frequency, which is 66Mhz on
Apple "UDMA capable" HW and 33 on older chipsets, but
the ide-pmac.c code does all the calculations internally
and shouldn't rely on the generic idebus value.

>I checked where this was is the kernel source and
>manually changed the limit from 66 to to 133 to see
>what happens. The machine boots fine with no errors.
>The file is: drivers/ide/ide.c. On lines 3159 and 3162
>I simply changed "66" to "133". Then I boot with
>"idebus=133".
>
>Is the 66 limit intentional or should it be changed to
>133?

You don't have to care about it.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Dave Ray]

Ben and Tom,

Thanks for the replies. The reason I ask is that when
I use "idebus=66" in the boot argument, the syslog
shows bootup message "using 66MHz bus speed.." rather
than "assuming 33MHz bus speed..". So I assumed the
inclusion of the boot argument sped up the computer.
Are the appearances of these messages in the syslog
fictitious? Or should I simply not go over 66?

-Dave

--- Benjamin Herrenschmidt <benh@kernel.crashing.org>
wrote:
> >
> >It's my understanding the Mac G4 towers run at a
> bus
> >speed of 133 MHz. However the kernel argument for
> >idebus=xx apparently only allows values up to 66.
> Is
> >this intentional?
>
> I don't think this value has any meaning on macs.
> It's
> related to the IDE bus frequency, which is 66Mhz on
> Apple "UDMA capable" HW and 33 on older chipsets,
> but
> the ide-pmac.c code does all the calculations
> internally
> and shouldn't rely on the generic idebus value.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Benjamin Herrenschmidt]

>
>Thanks for the replies. The reason I ask is that when
>I use "idebus=66" in the boot argument, the syslog
>shows bootup message "using 66MHz bus speed.." rather
>than "assuming 33MHz bus speed..". So I assumed the
>inclusion of the boot argument sped up the computer.
>Are the appearances of these messages in the syslog
>fictitious? Or should I simply not go over 66?

Actually, it's a bit more annoying than what I though,
as some drivers will use the speed. It's used by those
drivers to calc. some timings, and so should be properly
defined to the speed of the bus those controllers are
connected to.

The Apple controllers don't need it, they already know
the bus timings of the various revision of Apple chips.
However, the CMD 6xx, aec, and maybe other drivers
use that informations.

For now, those are PCI based exclusively and Apple's
PCI bus connectors are always 33Mhz. But we should
keep an eye on that.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Tony 'Nicoya' Mantler]

At 2:29 PM -0500 10/22/01, Benjamin Herrenschmidt wrote:
[...]
>For now, those are PCI based exclusively and Apple's
>PCI bus connectors are always 33Mhz. But we should
>keep an eye on that.
[...]

I thought Apple put a 66MHz PCI slot in some G3 systems for the video card,
before switching to AGP.

Of course, it's usually filled with a video card rather than an IDE card,
but who knows how some people might want to reconfigure their systems.


Cheers - Tony 'Nicoya' Mantler :)


--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nicoya@apia.dhs.org
Winnipeg, Manitoba, Canada           --           http://nicoya.feline.pp.se/


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Benjamin Herrenschmidt]

>I thought Apple put a 66MHz PCI slot in some G3 systems for the video card,
>before switching to AGP.

Hehe... in fact, when I wrote the previous email, I had that, then
decided it was just bloat to say that as nobody use it for anything
but video ;)

>Of course, it's usually filled with a video card rather than an IDE card,
>but who knows how some people might want to reconfigure their systems.

Well. I don't even know if it's really a 66Mhz slot, or some kind of
hacked 33Mhz with a doubled clock (I don't know if it really follow
the PCI spec for 66Mhz). I heard noise when the machine was released
about that slot not beeing completely standard.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Derrik Pates]

On Mon, 22 Oct 2001, Benjamin Herrenschmidt wrote:

> Well. I don't even know if it's really a 66Mhz slot, or some kind of
> hacked 33Mhz with a doubled clock (I don't know if it really follow
> the PCI spec for 66Mhz). I heard noise when the machine was released
> about that slot not beeing completely standard.

The PCI bus spec allows for up to 66 MHz bus clock with a 64-bit-wide bus.
Some systems (most notably Alpha-based systems) actually implemented such
slots.

Derrik Pates      |   Sysadmin, Douglas School   |    #linuxOS on EFnet
dpates@dsdk12.net |     District (dsdk12.net)    |    #linuxOS on OPN


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Geert Uytterhoeven]

On Mon, 22 Oct 2001, Benjamin Herrenschmidt wrote:
> >I thought Apple put a 66MHz PCI slot in some G3 systems for the video card,
> >before switching to AGP.
>
> Hehe... in fact, when I wrote the previous email, I had that, then
> decided it was just bloat to say that as nobody use it for anything
> but video ;)
>
> >Of course, it's usually filled with a video card rather than an IDE card,
> >but who knows how some people might want to reconfigure their systems.
>
> Well. I don't even know if it's really a 66Mhz slot, or some kind of
> hacked 33Mhz with a doubled clock (I don't know if it really follow
> the PCI spec for 66Mhz). I heard noise when the machine was released
> about that slot not beeing completely standard.

I once heard the electrical stuff is 66 MHz, but the connector is mounted
differently so only the custom OEM cards from ATI will fit. That's what you can
call `user lock-in' :-)

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Tom Rini]

On Mon, Oct 22, 2001 at 11:11:06AM -0700, Dave Ray wrote:


> Thanks for the replies. The reason I ask is that when
> I use "idebus=66" in the boot argument, the syslog
> shows bootup message "using 66MHz bus speed.." rather
> than "assuming 33MHz bus speed..".

Because you overrode it, that's why. :)  I don't think the kernel ever
tries to check things that the user tell it.

> So I assumed the
> inclusion of the boot argument sped up the computer.
> Are the appearances of these messages in the syslog
> fictitious? Or should I simply not go over 66?

You shouldn't change it at all.  Again, this has to do with the speed of
the PCI bus, _not_ weather it's ATA/33, ATA/66, ATA/100, and not the
speed of the memory bus.

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: idebus=133 ?

[Tony 'Nicoya' Mantler]

At 2:47 PM -0500 10/22/01, Benjamin Herrenschmidt wrote:
[...]
>Well. I don't even know if it's really a 66Mhz slot, or some kind of
>hacked 33Mhz with a doubled clock (I don't know if it really follow
>the PCI spec for 66Mhz). I heard noise when the machine was released
>about that slot not beeing completely standard.

Just for the sake of trivia, I pulled up apple's docs on the B&W G3, and it
had this to say about the slot:


>The primary PCI bus includes slot 1, which accommodates only 32-bit 66 MHz
>+3.3 V PCI cards. Slot 1 conforms to the PCI V2.1 specification with the
>exception that its clock speed is fixed at 66 MHz so it does not accept
>33 MHz cards.
>
>The Power Macintosh G3 computer is always configured with an Apple 2D/3D
>accelerated graphics card installed in slot 1, so that slot is not
>available for PCI card expansion unless the card is removed. Slot 1 is
>keyed for 3.3V only operation, so older 5V cards cannot be installed in
>that slot.


So the fuss was likely about the different keying for the 3.3v cards (this
is also an issue in some SGI x86 machines), and the 66MHz-only restriction
(most 66MHz slots I've seen will automatically go down to 33MHz when you
insert an older card).

3.3v/universal cards seem to be becoming more widespread recently, though
66MHz cards seem to still be thin on the ground, or not labled very
obviously for me to notice them.


Seems kinda silly the way everyone's talking nowadays about PCI-X and
whatnot when you can't even get 64/66 PCI slots on anything sub-$15k, and
even 64/33 is a rarity outside macs and servers. Ah well.


Cheers - Tony 'Nicoya' Mantler :)


--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nicoya@apia.dhs.org
Winnipeg, Manitoba, Canada           --           http://nicoya.feline.pp.se/


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/