hi all
Does anyone install qnx on PowerPc .
thanks
You mean the CPU sure, got one running just beside me at the moment.
As for MAC that’s another story 
“yair” <yshemama@rtview.co.il> wrote in message
news:8vvscg$5o0$1@inn.qnx.com…
hi all
Does anyone install qnx on PowerPc .
thanks
“Mario Charest” <mcharest@void_zinformatic.com> wrote in message
news:90096k$hub$1@nntp.qnx.com…
|
| You mean the CPU sure, got one running just beside me at the moment.
|
| As for MAC that’s another story
|
| “yair” <yshemama@rtview.co.il> wrote in message
| news:8vvscg$5o0$1@inn.qnx.com…
| > hi all
| >
| > Does anyone install qnx on PowerPc .
| >
| > thanks
That’s what I’d really like to see, Nto running on a Mac G4. That might even
cause me to buy my first Mac ever!
-Warren
“Warren Peece” <warren@nospam.com> wrote in message
news:903gpc$1sv$1@inn.qnx.com…
“Mario Charest” <mcharest@void_zinformatic.com> wrote in message
news:90096k$hub$> 1@nntp.qnx.com> …
|
| You mean the CPU sure, got one running just beside me at the moment.
|
| As for MAC that’s another story >
|
| “yair” <> yshemama@rtview.co.il> > wrote in message
| news:8vvscg$5o0$> 1@inn.qnx.com> …
| > hi all
|
| > Does anyone install qnx on PowerPc .
|
| > thanksThat’s what I’d really like to see, Nto running on a Mac G4. That might
even
cause me to buy my first Mac ever!
Me too. I’d even take a loss on my Nasdaq stocks to get the cash
to buy one, if RtP would run on it, and yes it’d be my first apple of
any ilk 
John Doe <john@csical.com> wrote in message news:903p28$6vc$1@inn.qnx.com…
“Warren Peece” <> warren@nospam.com> > wrote in message
news:903gpc$1sv$> 1@inn.qnx.com> …“Mario Charest” <mcharest@void_zinformatic.com> wrote in message
news:90096k$hub$> 1@nntp.qnx.com> …
|
| You mean the CPU sure, got one running just beside me at the moment.
|
| As for MAC that’s another story >
|
| “yair” <> yshemama@rtview.co.il> > wrote in message
| news:8vvscg$5o0$> 1@inn.qnx.com> …
| > hi all
|
| > Does anyone install qnx on PowerPc .
|
| > thanksThat’s what I’d really like to see, Nto running on a Mac G4. That might
even
cause me to buy my first Mac ever!Me too. I’d even take a loss on my Nasdaq stocks to get the cash
to buy one, if RtP would run on it, and yes it’d be my first apple of
any ilk >
Seb, here is your chance to get rich! Start a business and sell iMacs to
those guys! Oh yeah, you need USB drivers first…
- igor
In article <903vin$arp$1@inn.qnx.com>, “Igor Kovalenko”
<Igor.Kovalenko@motorola.com> wrote:
John Doe <> john@csical.com> > wrote in message
news:903p28$6vc$> 1@inn.qnx.com> …“Warren Peece” <> warren@nospam.com> > wrote in message
news:903gpc$1sv$> 1@inn.qnx.com> …“Mario Charest” <mcharest@void_zinformatic.com> wrote in message
news:90096k$hub$> 1@nntp.qnx.com> …
That’s what I’d really like to see, Nto running on a Mac G4. That
might
even
cause me to buy my first Mac ever!Me too. I’d even take a loss on my Nasdaq stocks to get the cash
to buy one, if RtP would run on it, and yes it’d be my first apple of
any ilk >
Seb, here is your chance to get rich! Start a business and sell iMacs to
those guys! Oh yeah, you need USB drivers first… >
- igor
Well, I have a bunch of Macs already, and I am dreading the day that I
won’t be able to buy MacOS 9.x anymore and will have to install MacOS X.
I would kill to be able to run Neutrino on these machines. In fact, I’d
even be willing to work on the necessary drivers if I could get
appropriate documentation for and a roadmap to bringing up Neutrino on a
given desktop system.
Regards,
Eric
Eric Berdahl <berdahl@intelligentparadigm.com> wrote:
In article <903vin$arp$> 1@inn.qnx.com> >, “Igor Kovalenko”
Igor.Kovalenko@motorola.com> > wrote:John Doe <> john@csical.com> > wrote in message
news:903p28$6vc$> 1@inn.qnx.com> …“Warren Peece” <> warren@nospam.com> > wrote in message
news:903gpc$1sv$> 1@inn.qnx.com> …“Mario Charest” <mcharest@void_zinformatic.com> wrote in message
news:90096k$hub$> 1@nntp.qnx.com> …
That’s what I’d really like to see, Nto running on a Mac G4. That
might
even
cause me to buy my first Mac ever!Me too. I’d even take a loss on my Nasdaq stocks to get the cash
to buy one, if RtP would run on it, and yes it’d be my first apple of
any ilk >
Seb, here is your chance to get rich! Start a business and sell iMacs to
those guys! Oh yeah, you need USB drivers first… >
- igor
Well, I have a bunch of Macs already, and I am dreading the day that I
won’t be able to buy MacOS 9.x anymore and will have to install MacOS X.
I would kill to be able to run Neutrino on these machines. In fact, I’d
even be willing to work on the necessary drivers if I could get
appropriate documentation for and a roadmap to bringing up Neutrino on a
given desktop system.Regards,
Eric
Depending on the install you wanted (like the Windows one) a file system
driver would be needed first…
http://www-sccm.stanford.edu/Students/hargrove/HFS/
Any takers?
I just remember a particular gentleman who was working with the Linux
filesystem driver visiting everytime he toasted his Linux partition
and would visit while it was being restored (which was somtimes 2,
3 times a day)
Erick
Eric Berdahl <berdahl@intelligentparadigm.com> wrote in message
news:berdahl-D854ED.22032329112000@inn.qnx.com…
In article <903vin$arp$> 1@inn.qnx.com> >, “Igor Kovalenko”
Igor.Kovalenko@motorola.com> > wrote:John Doe <> john@csical.com> > wrote in message
news:903p28$6vc$> 1@inn.qnx.com> …“Warren Peece” <> warren@nospam.com> > wrote in message
news:903gpc$1sv$> 1@inn.qnx.com> …“Mario Charest” <mcharest@void_zinformatic.com> wrote in message
news:90096k$hub$> 1@nntp.qnx.com> …
That’s what I’d really like to see, Nto running on a Mac G4. That
might
even
cause me to buy my first Mac ever!Me too. I’d even take a loss on my Nasdaq stocks to get the cash
to buy one, if RtP would run on it, and yes it’d be my first apple of
any ilk >
Seb, here is your chance to get rich! Start a business and sell iMacs to
those guys! Oh yeah, you need USB drivers first… >
- igor
Well, I have a bunch of Macs already, and I am dreading the day that I
won’t be able to buy MacOS 9.x anymore and will have to install MacOS X.
I would kill to be able to run Neutrino on these machines. In fact, I’d
even be willing to work on the necessary drivers if I could get
appropriate documentation for and a roadmap to bringing up Neutrino on a
given desktop system.
A reasonable plan would be to get some ATX form-factor PCI PPC motherboard,
like Motorola MTX600 series. If you have or can afford cPCI stuff you could
go with cPCI boards too, that’s even better because you can have an Intel
SBC in system slot and G4 PPC board in non-system slot. Install QNX4 on
Intel box, then install Neutrino for PPC target. It boots easily using
PPCBug and TFTP. You’re ready to do cross-development.
Speaking about RTP, for what I know the only major missing piece is USB
support, the rest of stuff was made to work on G4 Macs, even on dual ones
I’ve heard. There are not many other divers, no SCSI for example, but they
should be portable from x86 once QNX finally open the source. That’s of
course is just unofficial rumor and I doubt QNX will really support the idea
given the behavior of Apple wrt specs.
Keep in mind, even if you succeed you’re on your own, pretty much. PPC
version probably won’t be free and there won’t be any reasonable user
community to share anything with. You can play with it as much as you want,
but without support from either hardware or OS vendor you’ll soon find
yourself very lonely, unless you have a realistic business case and going to
release a product. Of course if that’s a case, go ahead, call QNX sales and
present your plan.
- Igor
“Igor Kovalenko” <Igor.Kovalenko@motorola.com> wrote in message
news:906jd7$h2$1@inn.qnx.com…
| A reasonable plan would be to get some ATX form-factor PCI PPC motherboard,
| like Motorola MTX600 series. If you have or can afford cPCI stuff you could
| go with cPCI boards too, that’s even better because you can have an Intel
| SBC in system slot and G4 PPC board in non-system slot. Install QNX4 on
| Intel box, then install Neutrino for PPC target. It boots easily using
| PPCBug and TFTP. You’re ready to do cross-development.
[yoink]
Hey Igor, you seem to be knowledgeable about cPCI hardware. I understand the
boards are hot-swappable, is this true? You mentioned the “system slot”. What
happens if the board in the system slot fries- do you lose the entire chassis
or do you yank the dead board and slap in a new one, just like any other slot?
I was daydreaming about a few different ways I could use a hot-swappable
multi-CPU backplane-communicating rack like that, but if blowing one CPU card
would take it out of action then it won’t work for me. Some basic info like
that would be appreciated, when you have a chance…
Thanks in advance,
-Warren
“Warren Peece” <warren@nospam.com> wrote in message
news:906mci$267$1@inn.qnx.com…
“Igor Kovalenko” <> Igor.Kovalenko@motorola.com> > wrote in message
news:906jd7$h2$> 1@inn.qnx.com> …
| A reasonable plan would be to get some ATX form-factor PCI PPC
motherboard,
| like Motorola MTX600 series. If you have or can afford cPCI stuff you
could
| go with cPCI boards too, that’s even better because you can have an
Intel
| SBC in system slot and G4 PPC board in non-system slot. Install QNX4 on
| Intel box, then install Neutrino for PPC target. It boots easily using
| PPCBug and TFTP. You’re ready to do cross-development.[yoink]
Hey Igor, you seem to be knowledgeable about cPCI hardware. I understand
the
boards are hot-swappable, is this true? You mentioned the “system slot”.
What
happens if the board in the system slot fries- do you lose the entire
chassis
or do you yank the dead board and slap in a new one, just like any other
slot?
I was daydreaming about a few different ways I could use a hot-swappable
multi-CPU backplane-communicating rack like that, but if blowing one CPU
card
would take it out of action then it won’t work for me. Some basic info
like
that would be appreciated, when you have a chance…
That depends on particular hardware. Don’t consider it a plug, but I’m
mostly knowleageable about Motorola stuff. They (Motorola Computer Group -
MCG) do different kinds of cPCI chassis. There are simpler/cheaper
chassis/SBCs, which don’t do what you want and there are pretty damned
expensive ones, which do. There is no approved formal standard about how to
do hot-swap stuff, but there are few industry standards and MCG is trying to
establish theirs as de-facto one.
Here is how it works.
- A hot swappable system must use hot-swappable chassis, which means it
will have 2 sets of slots, called ‘domains’. Each domain has 1 system slot
and 6 or 9 non-system ones. - System slots are ‘double-width’, but SBCs must be single width. The other
half of each double width system slot is occupied by HotSwap Controller
(HSC). - HSC is basically a PCI-to-PCI bridge connecting 2 domains, plus it has
capability to power-up/down all slots and light-up/down LEDs on boards and
chassis. - BIOS of SBC in system slot must know how to initialize HSCs. MCG ships
modified version of Phoenix BIOS, which allows to enable/disable and
configure HSC. Configuring means assigning domains to system slots. Each
system slot can ‘own’ either or both domains.
So, the idea is that if one SBC goes to nirvana, the other can take over its
domain, assuming both domain have identical sets of hardware. Other SBC must
be in some sort of hot standby state, which implies some checkpoint system
between them. An OS must have hot-swap support too, which would allow it to
detect hot-swap events and trigger appropriate actions.
Note, while all compliant boards are phisycally hot-swappable (which means
they have microswitch in the lock/release hatch and LED) and can generate
HSC events when you push the hatch, it will only make sense if there is
support for such events in drivers. Those drivers are supposed to take
whatever action is appropriate and then light-up “removal allowed” LED on
the board. For some drivers it is reasonably easy to take an action, for
some it is very complicated (e.g., SCSI board controlling drive with root
filesystem on it). In the simplest case a generic HSC “framework” could just
kill corresponding hardware driver, but that may too be complicated because
one driver may serve multiply boards. In latter case the corresponding OS
subsystem must support loading/unloading drivers on per-board basis (e.g.,
io-net). Similarly, when a board is inserted, the HSC framework must rescan
PCI bus and assign resources for it, which is not easy task too because a
board can contain a PCI-PCI bridge and another PCI bus behind it (or even
couple of them) which need to be initialized…
MCG ships version of Linux which supports at least some of that stuff, they
call it ‘HA Linux’ where HA stands for High Availability. QSSL is working on
HSC framework and they have made some progress. There is no built-in support
in their drivers yet. We (Motorola iDEN group) have independent HSC support
for Neutrino on Motorola hardware and we’re closely cooperating with QSSL.
There is no high-level framework for HA-aware application development yet,
but both MCG and QSSL have plans involving a 3rd party, which I’m not sure
they’re ready to announce yet. MCG is doing their development with Linux and
VxWorks so far.
Speaking about back-plane communications, keep in mind that communications
over cPCI backplane are usually CPU intensive and may not really be very
fast. MCG is going diferent way - their latest chassis has built-in ethernet
mesh. They are marketing it for carrier-grade systems. For some aplications
you can also use H.110 (telephony) bus - that’s what we do.
Hope this helps to navigate through that bizzare stuff
- igor
Igor Kovalenko wrote:
“Warren Peece” <> warren@nospam.com> > wrote in message
news:906mci$267$> 1@inn.qnx.com> …
“Igor Kovalenko” <> Igor.Kovalenko@motorola.com> > wrote in message
news:906jd7$h2$> 1@inn.qnx.com> …
| A reasonable plan would be to get some ATX form-factor PCI PPC
motherboard,
[ clip … ]Speaking about back-plane communications, keep in mind that communications
over cPCI backplane are usually CPU intensive
… what about DMA ?
and may not really be very fast.
Newer PC-to-PCI bridges have in burst mode a
performance of ~200MegaBytes/s .
In shared memory configurations ( slave CPUs
behind a non transparent PCI-to-PCI bridge using
shared memory of a master CPU ) is the transfer
rate normaly in the range of 30-60 MBytes/s.
Doesn’t use Motorola shared memory communication
with their multi processor cPCI systems ?
Armin
“Igor Kovalenko” <kovalenko@home.com> wrote in message
news:907feg$ffg$1@inn.qnx.com…
[ detailed description deleted ]
Hope this helps to navigate through that bizzare stuff >
- igor
Wow, I didn’t think it would be that complicated. Yes that was extremely
helpful. Sounds like I should hang tight and see what you & QNX come up for
more of a total package solution. That way I can leave all of the “fun
stuff” for you to figure out (I won’t need my helmet)…
Thanks again.
-Warren
“Armin Steinhoff” <A-Steinhoff@web_.de> wrote in message
news:3A275793.33B95C69@web_.de…
Igor Kovalenko wrote:
Speaking about back-plane communications, keep in mind that
communications
over cPCI backplane are usually CPU intensive… what about DMA ?
and may not really be very fast.
Newer PC-to-PCI bridges have in burst mode a
performance of ~200MegaBytes/s .
In shared memory configurations ( slave CPUs
behind a non transparent PCI-to-PCI bridge using
shared memory of a master CPU ) is the transfer
rate normaly in the range of 30-60 MBytes/s.Doesn’t use Motorola shared memory communication
with their multi processor cPCI systems ?Armin
I thought there was some sort of message passing coprocessor on cPCI
designed specifically for this sort of thing, rather along the lines of
Multibus II? A strictly CPU oriented IPC scheme would be way too slow to be
useful. Any sort of DMA solution as Armin suggested would probably do okay,
but I thought the backplane communication was being touted as a major
feature of cPCI so it seems odd to me (knowing very little about cPCI
admittedly) that they would go with an Ethernet mesh…
-Warren
Warren Peece <Warren@nospam.com> wrote in message
news:908b6q$2gd$1@inn.qnx.com…
“Armin Steinhoff” <A-Steinhoff@web_.de> wrote in message
news:3A275793.33B95C69@web_.de…Igor Kovalenko wrote:
Speaking about back-plane communications, keep in mind that
communications
over cPCI backplane are usually CPU intensive… what about DMA ?
and may not really be very fast.
Newer PC-to-PCI bridges have in burst mode a
performance of ~200MegaBytes/s .
In shared memory configurations ( slave CPUs
behind a non transparent PCI-to-PCI bridge using
shared memory of a master CPU ) is the transfer
rate normaly in the range of 30-60 MBytes/s.Doesn’t use Motorola shared memory communication
with their multi processor cPCI systems ?Armin
I thought there was some sort of message passing coprocessor on cPCI
designed specifically for this sort of thing, rather along the lines of
Multibus II?
Not that I know of.
A strictly CPU oriented IPC scheme would be way too slow to be
useful. Any sort of DMA solution as Armin suggested would probably do
okay,
How DMA is gonna make it less CPU intensive? DMA allows you to offload
transfer work from central CPU to dedicated CPUs on DMA capable boards. Now,
if all your boards are CPU boards, where you’re offloading? You could have
dedicated ‘message passing coprocessor’ on each board, but that’s
effectively what Motorola does anyway - they just use ethernet instead of
backplane for transport media (and there is a reason for that), but
logically it is the same picture. Ethernet boards are ‘message passing
coprocessors’ then.
but I thought the backplane communication was being touted as a major
feature of cPCI
Major features of cPCI are hot swap and ability to use spare backplane pins
to implement various transport mechanisms optimized for particular tasks,
like H.110 bus.
so it seems odd to me (knowing very little about cPCI
admittedly) that they would go with an Ethernet mesh…
Pumping data from each board to each through PCI bus will produce terabytes
of traffic and that will make PCI a bottleneck. It is one thing to pump data
between host memory and PCI boards (1 to many), and another thing to pump
data between ALL boards (many to many).
The ethernet mesh is effectively offloading traffic from PCI bus. It
actually does use spare pins of cPCI to implement full mesh (192
point-2-point 100Mbit/sec connections), so it can be switched media, not
shared. That model has more complex HSC controller, which acts like 1000/100
Ethernet switch/router as well. It has 16100mbit channels wired to every
slot, plus 4100mbit and 2*1000Mbit channels wired to front panels to
connect multiply chassises. Each slot in the chassis can be populated with
'IP Recourse Board" which is non-system slot CPU board with 128Mb or memory
and slots for 2 PCI mezzanine cards (PMC). One of those PMC slots has
connectors wired to backplane pins used for ethernet mesh and could be
populated with either normal ethernet PMC or a full router/switch PMC to
create fully switched system.
The extra benefit is, most of traffic passing through the system could be IP
traffic anyway and you don’t have to tickle it much if all
inter/extra-communications are done over ethernet. You just set up routing,
and let packets fly through without having to do packets
assembling/disassembling. It is similar concept to H.110 bus, just H.110 was
designed for voice traffic while ethernet mesh is good for IP traffic.
- Igor
Igor Kovalenko wrote:
Warren Peece <> Warren@nospam.com> > wrote in message
news:908b6q$2gd$> 1@inn.qnx.com> …
“Armin Steinhoff” <A-Steinhoff@web_.de> wrote in message
news:3A275793.33B95C69@web_.de…Igor Kovalenko wrote:
Speaking about back-plane communications, keep in mind that
communications
over cPCI backplane are usually CPU intensive… what about DMA ?
and may not really be very fast.
Newer PC-to-PCI bridges have in burst mode a
performance of ~200MegaBytes/s .
In shared memory configurations ( slave CPUs
behind a non transparent PCI-to-PCI bridge using
shared memory of a master CPU ) is the transfer
rate normaly in the range of 30-60 MBytes/s.Doesn’t use Motorola shared memory communication
with their multi processor cPCI systems ?Armin
I thought there was some sort of message passing coprocessor on cPCI
designed specifically for this sort of thing, rather along the lines of
Multibus II?
Such co-processing is typical used with
‘reflective memory’ hardware.
Not that I know of.
A strictly CPU oriented IPC scheme would be way too slow to be
useful. Any sort of DMA solution as Armin suggested would probably do
okay,How DMA is gonna make it less CPU intensive?
The DMA controller overtakes the bus as bus master
(cycle stealing at memory level) and does the
transfer independend at hardware level … the CPU
isn’t active involved in that transfer. The CPU
shares the resources main bus and memory … and
this could be an issue for the real-time behavior
of a real-time system.
DMA allows you to offload transfer work from central CPU to dedicated CPUs on >DMA capable boards.
Direct Memory Access (DMA) deals with memory …
the CPUs are only used for the setup of the DMA
controllers.
Now,
if all your boards are CPU boards, where you’re offloading? You could have
dedicated ‘message passing coprocessor’ on each board, but that’s
effectively what Motorola does anyway - they just use ethernet instead of
backplane for transport media (and there is a reason for that), but
logically it is the same picture. Ethernet boards are ‘message passing
coprocessors’ then.but I thought the backplane communication was being touted as a major
feature of cPCIMajor features of cPCI are hot swap and ability to use spare backplane pins
to implement various transport mechanisms optimized for particular tasks,
like H.110 bus.so it seems odd to me (knowing very little about cPCI
admittedly) that they would go with an Ethernet mesh…Pumping data from each board to each through PCI bus will produce terabytes
of traffic and that will make PCI a bottleneck. It is one thing to pump data
between host memory and PCI boards (1 to many), and another thing to pump
data between ALL boards (many to many).The ethernet mesh is effectively offloading traffic from PCI bus. It
actually does use spare pins of cPCI to implement full mesh (192
point-2-point 100Mbit/sec connections), so it can be switched media, not
shared. That model has more complex HSC controller, which acts like 1000/100
Ethernet switch/router as well. It has 16100mbit channels wired to every
slot, plus 4100mbit and 2*1000Mbit channels wired to front panels to
connect multiply chassises. Each slot in the chassis can be populated with
'IP Recourse Board" which is non-system slot CPU board with 128Mb or memory
and slots for 2 PCI mezzanine cards (PMC). One of those PMC slots has
connectors wired to backplane pins used for ethernet mesh and could be
populated with either normal ethernet PMC or a full router/switch PMC to
create fully switched system.The extra benefit is, most of traffic passing through the system could be IP
traffic anyway and you don’t have to tickle it much if all
inter/extra-communications are done over ethernet. You just set up routing,
and let packets fly through without having to do packets
assembling/disassembling. It is similar concept to H.110 bus, just H.110 was
designed for voice traffic while ethernet mesh is good for IP traffic.
Yes … replacing of parallel bus systems by
serial bus systems is an established trend. See
USB (far too slooow), FireWire, Gigabit-Ethernet
… or the industrial PROFIBUS as a replacement of
PLC backplanes.
Greetings
Armin
Igor:
A reasonable plan would be to get some ATX form-factor PCI PPC motherboard,
like Motorola MTX600 series. If you have or can afford cPCI stuff you could
go with cPCI boards too, that’s even better because you can have an Intel
SBC in system slot and G4 PPC board in non-system slot. Install QNX4 on
Intel box, then install Neutrino for PPC target. It boots easily using
PPCBug and TFTP. You’re ready to do cross-development.Speaking about RTP, for what I know the only major missing piece is USB
support, the rest of stuff was made to work on G4 Macs, even on dual ones
I’ve heard. There are not many other divers, no SCSI for example, but they
should be portable from x86 once QNX finally open the source. That’s of
course is just unofficial rumor and I doubt QNX will really support the idea
given the behavior of Apple wrt specs.
i Know were you heard that Ogor LOL, and its true folks, Dan (dodge) the
main Man has openly said that IF an OEM PPC vendor were willing to
make a valid comsumer leval machine that could become popular than
he would be willing to talk about an RTP port if the HW happened.
now for the readers that dont know, there was an HW OEM PPC machine in the
works over in the AMIGA market that was to get this PPC RTP, alas the
situation didn`t happen due to the OEM folding.
that being said, there is a srong posibility that IF a new OEM PPC
motherboard/machine maker were to take the plunge and create a consumer
priced PPC product that Dan would take a very close look at this
and Perhaps arrange something !, the question is Though, can there
be a large market for this new CL PPC motherboard ?, is so we need
to find and convince the right people to DO IT so we can buy it.
personaly i think a good price for a consumer 5 PCi/ 1AGP/ Zif PPC slot/
3 DIMM/ USB/ optional Firewire ?/ eIDE/ flash Bios (autoconfig/non PC style)
ATX formfacture motherboard would be around the 300UKP mark, what ya think ?.
and/or a consumer leval PPC on a PCi card running RTP would be a very
good thing too and open up the posability of makeing any wintel motherboard
a slave to the real PPC RTP you all seem to like the idea of 8}.
the PPC on a PCi card is called the InsideOut by Mick Tinker of
Amiga Boxer motherboard fame, alas it seems that the IO had to take
a back seat for a long time now, so any other takers out there
than might be up to the challenge of OEM`ing real soon now ?.
Keep in mind, even if you succeed you’re on your own, pretty much. PPC
version probably won’t be free and there won’t be any reasonable user
community to share anything with. You can play with it as much as you want,
but without support from either hardware or OS vendor you’ll soon find
yourself very lonely, unless you have a realistic business case and going to
release a product. Of course if that’s a case, go ahead, call QNX sales and
present your plan.
- Igor
–
Paul May, Manchester, UK
Team AMIGA Central, Phoenix Core
“Paul May” <paul@phinixi.com> wrote in message
news:2996.377T1044T5214766paul@phinixi.com…
Igor:
A reasonable plan would be to get some ATX form-factor PCI PPC
motherboard,
like Motorola MTX600 series. If you have or can afford cPCI stuff you
could
go with cPCI boards too, that’s even better because you can have an Intel
SBC in system slot and G4 PPC board in non-system slot. Install QNX4 on
Intel box, then install Neutrino for PPC target. It boots easily using
PPCBug and TFTP. You’re ready to do cross-development.Speaking about RTP, for what I know the only major missing piece is USB
support, the rest of stuff was made to work on G4 Macs, even on dual ones
I’ve heard. There are not many other divers, no SCSI for example, but
they
should be portable from x86 once QNX finally open the source. That’s of
course is just unofficial rumor and I doubt QNX will really support the
idea
given the behavior of Apple wrt specs.i Know were you heard that Ogor LOL, and its true folks, Dan (dodge) the
main Man has openly said that IF an OEM PPC vendor were willing to
make a valid comsumer leval machine that could become popular than
he would be willing to talk about an RTP port if the HW happened.now for the readers that dont know, there was an HW OEM PPC machine in the
works over in the AMIGA market that was to get this PPC RTP, alas the
situation didn`t happen due to the OEM folding.that being said, there is a srong posibility that IF a new OEM PPC
motherboard/machine maker were to take the plunge and create a consumer
priced PPC product that Dan would take a very close look at this
and Perhaps arrange something !, the question is Though, can there
be a large market for this new CL PPC motherboard ?, is so we need
to find and convince the right people toDO ITso we can buy it.personaly i think a good price for a consumer 5 PCi/ 1AGP/ Zif PPC slot/
3 DIMM/ USB/ optional Firewire ?/ eIDE/ flash Bios (autoconfig/non PC
style)
ATX formfacture motherboard would be around the 300UKP mark, what ya think
?.and/or a consumer leval PPC on a PCi card running RTP would be a very
good thing too and open up the posability of makeing any wintel
motherboard
a slave to the real PPC RTP you all seem to like the idea of 8}.the PPC on a PCi card is called the
InsideOutby Mick Tinker of
AmigaBoxermotherboard fame, alas it seems that the IO had to take
a back seat for a long time now, so any other takers out there
than might be up to the challenge of OEM`ing real soon now ?.
Mac already has a hard time making a living with a full fledge OS.
Commodore when belly up doing it, and even with all the software
bases off the Amiga it was never enough to really resurect it.
So QRTP situation is even worse: no serious OEM is going to
build a “personal” PPC machine for which you have little if not any
software.
Mario Charest wrote:
Mac already has a hard time making a living with a full fledge OS.
Commodore when belly up doing it, and even with all the software
bases off the Amiga it was never enough to really resurect it.
So QRTP situation is even worse: no serious OEM is going to
build a “personal” PPC machine for which you have little if not any
software.
Yeah, I think if Sony will perhaps show us an example by resurrecting
Beta, then who knows
I don’t know if Apple has in fact such a hard time. They appear to be
doing good with iMacs and if MacOS X is successfull, they’ll be just
fine I believe. If only they weren’t such a morons to hide specs…
- igor
Previously, Igor Kovalenko wrote in qdn.public.qnxrtp.installation:
I don’t know if Apple has in fact such a hard time. They appear to be
doing good with iMacs and if MacOS X is successfull, they’ll be just
fine I believe. If only they weren’t such a morons to hide specs…
Thanks to the darwin open source thing going on, the newer macs are at least documented in code.
As for the older macs (pre-pci)… Well, you really don’t want to know what goes on inside those computers. Trust me, I’ve seen it, you don’t want to know.
New ones are nice though, very nice. RTP on an smp PMac G4 would rock some serious ass. And I’m not just saying that because I own apple stock (don’t worry, bought it after the hit they took a few months ago).
\
Tony Mantler | Proud ---- Days since the last
QNX Consulting | of our | 27 |
tony@astra.mb.ca | Record ---- “Gerbil Incident”
Eric Berdahl wrote:
Well, I have a bunch of Macs already, and I am dreading the day that I
won’t be able to buy MacOS 9.x anymore and will have to install MacOS X.
I would kill to be able to run Neutrino on these machines. In fact, I’d
even be willing to work on the necessary drivers if I could get
appropriate documentation for and a roadmap to bringing up Neutrino on a
given desktop system.
The best you can do is pore over the Linux PPC sources. They’ve had to
reverse engineer and hack a whole lot, plus they had “help” from Apple’s
aborted linuxish stuff for Mac. QSSL is in the same boat vis-a-vis Mac
as BeOS. They could spend the time to hack their OS to run on Macs
but at the flip of one of Steve Jobs’ warped braincells Apple could
easily make the next batch of Macs block out non MacOS and OSX systems.
Even the older ones aren’t “safe”, at least the ones with the “ROM” in
a file on disk instead of a ROM chip. If you want to work on QNX for
Mac, your best bet is to start with the pre-G3 PCI PowerMacs. Those
are from the days when Open Firmware was truely open and the info
is available. BeOS will work fine on them even with Gx CPU upgrades.
Even the 7500 works once you swap the stock 601 for a 603 or better.
The 7200 will not run BeOS due to the non-removable 601 CPU.
But hey, as unbloated as QNX RTP is it should really fly on a 9600!
(When/if its ever ported to Mac.)