The official 2.2.18 should work pretty fine on most boxes. I remember a few things slipped past 2.2.18 release and will be in official 2.2.19. However, there may be some SMP troubles with 2.2.x in general on SMP, I would suggest using a 2.4 for that.
For 3rd party IDE, There should be no problem with 2.4 provided that you enable support for the card's chipset when compiling the kernel. For 2.2, the reason is that most of these cards use chipsets not supported by plain 2.2 and Andre's big IDE patch for 2.2 lacks some stuffs for PPC. That's my fault, I did the fixes and never took the time to make the proper patches for Andre (it's a bit painful to make a patch of a patch).
Well, I don't think important bug fixes are beeing ignored. And if they are, then you didn't spam us hard enough. There used to be few kernel developers (fortunately, that number is increasing) and even less who had knowledge of some of the badly (or un-) documented chipsets in macs, since I beleive your rants are about Apple HW.
The fbdev's are not broken. Maybe _some_ fbdevs have bugs, but that's an issue with the fbdev maintainers which are not the same as the kernel maintainers. A given person can hardly have knowledge of the entire set of hardware out there today.
The IDE init code is not broken AFAIK. If it is, please send me a detailed report. The only breakage I know with it so far are with some very odd portables with weird old hard disks that may timeout trying to detect the disk (no disk detected), and the usual breakage that happened regulary with CMD646 chips (broken hardware). AFAIK, only the first one remains and those rare affected users already know how to work around it.
The lost interrupts are a symptom of either a wrong bootloader or a too old kernel used on newer hardware.
I can't speak for the floppy driver as I hardly ever touched that beast, but it's definitely a very weird and sensible (and badnly or undocumented) chip that have some bizarre timing constraints. It works well on some machines. I don't think anybody ever posted a patch, and I at least don't own a machine with a floppy anymore to give it a try.
Serial port drivers suffered from HW bugs on some machines that have finally been worked around, and AFAIK, they are working now.
I don't either know what you call bells & whistles, but please don't underestimate the importance of a strong lower layer. The large variety of machines requires it.
Right. And that's one reason why the PPC tree itself is maintained as a BitKeeper repository.
There's one fed from Linus official patches which is the "parent" of the linuxppc_2_4 tree which is our current "stable" tree. We also have a _2_5 tree with our "work in progress" things that will ultimately be rooted on Linus 2.5 when it comes to life.
The fact is that we always try to avoid as much as possible changing non PPC-specific code, and when such has to be done (like endian fixes), we usually separate this clearly and go thru the specific maintainers.
But I also think bringing this whole thing to/. was not a good idea at the very beginning;)
Well, I won't even try to rely to most of the crap I've read here, sorry guys, but some of your comments are just so...
There are several points about the PPC arch. First of all, it's huge and very quickly growing. Why ? Well, in a single arch, it handles all PCI PowerMacs (with new hardware coming out of Apple every 6 monthes), CHRP boxes (including some RS6k IBM machines), PReP boxes, APUS, NuBus PowerMac is coming in, Gemini, and I'm forgetting some...
Add to that the embedded hardware (8xx CPUs, 4xx CPUs coming in soon) with the zillion of hardware variations.
So it's _huge_, it's quickly moving forward (remember USB, we needed to get that working for kbd and mouse on iMacs while most x86 boxes didn't even had Windoze drivers for their USB controller), and the necessary consequence is that patches are huge. Fortunately, most of them just touch arch/ppc, include/asm-ppc or PPC specific drivers.
But I'm not telling you all here;) The truth is that we are in fact paid by Microsoft to write crappy code, flood Linus with huge and unmergeable patches in the clear intend to cause a Linux fork !
With recent 2.2.18pre's, all that support has been merged in the main tree. You can either pick Alan latest 2.2.18pre or use Paul Mackerras usual rsync repository.
Hehe. Of course, Apple had it working in the labs for quite some time. The difference here is that Apple engineers did have the informations about how it was implemented in HW (especially how to start the second CPU), while we didn't and we had to reverse engineer it;)
It was just "fun" to announce it just before MacOS X Beta;)
I certainly don't want to start any kind of NetBSD vs. Linux war, nor do I want to minimize the great work done by NetBSD coders for which I have great respect. NetBSD may have a lot of strengths over Linux (or may not, I don't know and I don't want to know now;) but I don't think they are more "mature" in their support of the PowerMac hardware. Most low-level hacking and reverse engineering needed to adapt "free" OSes on those platforms was done on Linux/PPC (except for the bits that filtered out of Apple via the original MkLinux and now with Darwin). Linux/PPC, afaik, is still unique to support sleep and power management features on PowerBooks (not all of them yet), and a variety of other hardware features I won't enumerate now.
The ROM file is part of MacOS. It replaces the "old" >4Mb ROM that "oldworld" Macs had. It's loaded from disk by the firmware (and in our case, by the emulator which simulates the firmware).
For example, if you buy a MacOS 9 "box", on the CD, you'll find the ROM file. You just need to copy it over to the linux filesystem, strip it, and then, boot MacOnLinux from the CD in order to do the MacOS install.
I don't know where you found those informations, probably in some ultra-liberal activist site. I tend not to answer to sensible political comments like yours, but I really can't let you say such aberrations while having a score of 5 !
But maybe was it intentionally ?
Anyway, the _fact_ is that Jose Bove has always been non-violent, and his actions has always been more symbolic than anything else. The "confederation paysanne" never destroyed the MacDonald's restaurant with a bulldozer, but actually "unmounted" it cleanly, without damaging parts. This action was made after the officials were made aware of it, without any form of violence.
The attack on MacDonald's your are talking about, where unfortunately a 28 years old waitress died, was a completely different terrorist action issued by Britain separatists, and can in no way be compared or assimilated to Jose protests.
He (and his movement) has always been non-violent.
The interesting thing here is that I finally see how much non-sense (since I know this specific subject) can an anti-Katz actvist say;)
That is just plain wrong, I don't know where you did get your infos from, but: - MaxOS X is based on Mach 3.0 plus some additions (mostly real time) from mach 4.0 (which is _not_ a newer version of Mach but a separate branch) and some Apple custom additions. This _is_ the Mach microkernel. However, they have made all sorts of optimisations to reduce the overhead of having a separate BSD layer, and so BSD is not implement as a server process but is somewhat "wired" into Mach. (Mach IPC-based API is turned into direct function calls and BSD is co-located in the kernel address space). Apple did already experiment with this technique on MkLinux. - Darwin and MacOS X are not distinct. Darwin is the foundation of MacOS X, there are not 2 different kernels. - They do, indeed, have nice copy-on-write features, and a bunch of other cool stuffs like the IOKit
For completeness, add to the powerbook feature list:
- USB - video is a R128 mobility with external CRT output on an AGP bus - 16bits/44KHz sound ouput & input (+built in mic) - hot-swappable media bays - Cardbus & zoom video - irda
That's unfortunately not so simple. For example, the way GM food is distributed makes it almost impossible for people to have a choice. There is almost no tracability, local grown beef can be fed with industrial food made from GM stuff without beeing marked GM, and more...
Actually, one representant of Monsanto agreed, in a public interview, that if GM food was not prohibited, it would be difficult and certainly very expensive to buy no-GM guaranteed food.
The idea that people have choice and vote with their money is a myth. A strong myth, especially in the US. People buy what they are offered and what they can afford.
Re:Consistency: Where's The Specs?
on
iBook boots Linux
·
· Score: 1
Did you actually read those technotes ? They don't really contain anything more than an overall block diagram, nothing like a register layout of the new chipset. Fortunately, there's Open Firmware which contains a few more details available to who can read Forth, and parts of those new ASICs are similar enough to the old ones so that extrapolating was not difficult enough. Also, those new machines are closer to CHRP than any other previous Mac;)
There is an interesting law for this in France: You can only do comparative ads if your ad is backed up by some kind of real benchmark-like comparison, or based on an independant comparison (done by the press for example). I don't know the exact details, but the result in reality is that we almost never see comparative advertising here...
This is an interesting point. Actually, European countries are trying to block hormon meat, and European population appears to be strongly against GM modified food.
On another hand, it appears that USA Gov. is trying to force european countries to accept those products via mondial commerce organisation. (we already have to pay huge penalities for refusing meat with hormones).
You can't turn it off. The PMU does a lot more than sleep mode, it replaces Cuda (ADB controller) which manages the keyboard and the trackpad, and it handles the nvram on those machines.
What you did probably turn off is the compile option for Power Management. This option enables the sleep code and should be harmless unless you actually try to sleep. For sleeping, there is the snooze command, and there's a little daemon (pmud) written by Paul, which monitors your machine and can excute scripts when the Power cable is connecter or disconnected. pmud also handles sleep when the lid is closed.
Slashdot Mirror
Sorry, I meant some 2.2.x SMP problems on PPC... looks like it's time for me to reach my bed.
Thanks ;)
The official 2.2.18 should work pretty fine on most boxes. I remember a few things slipped past 2.2.18 release and will be in official 2.2.19. However, there may be some SMP troubles with 2.2.x in general on SMP, I would suggest using a 2.4 for that.
For 3rd party IDE, There should be no problem with 2.4 provided that you enable support for the card's chipset when compiling the kernel. For 2.2, the reason is that most of these cards use chipsets not supported by plain 2.2 and Andre's big IDE patch for 2.2 lacks some stuffs for PPC. That's my fault, I did the fixes and never took the time to make the proper patches for Andre (it's a bit painful to make a patch of a patch).
Ben.
Well, I don't think important bug fixes are beeing ignored. And if they are, then you didn't spam us hard enough. There used to be few kernel developers (fortunately, that number is increasing) and even less who had knowledge of some of the badly (or un-) documented chipsets in macs, since I beleive your rants are about Apple HW.
The fbdev's are not broken. Maybe _some_ fbdevs have bugs, but that's an issue with the fbdev maintainers which are not the same as the kernel maintainers. A given person can hardly have knowledge of the entire set of hardware out there today.
The IDE init code is not broken AFAIK. If it is, please send me a detailed report. The only breakage I know with it so far are with some very odd portables with weird old hard disks that may timeout trying to detect the disk (no disk detected), and the usual breakage that happened regulary with CMD646 chips (broken hardware). AFAIK, only the first one remains and those rare affected users already know how to work around it.
The lost interrupts are a symptom of either a wrong bootloader or a too old kernel used on newer hardware.
I can't speak for the floppy driver as I hardly ever touched that beast, but it's definitely a very weird and sensible (and badnly or undocumented) chip that have some bizarre timing constraints. It works well on some machines. I don't think anybody ever posted a patch, and I at least don't own a machine with a floppy anymore to give it a try.
Serial port drivers suffered from HW bugs on some machines that have finally been worked around, and AFAIK, they are working now.
I don't either know what you call bells & whistles, but please don't underestimate the importance of a strong lower layer. The large variety of machines requires it.
Thanks Michel ;)
..
Actually, I _was_, I've been pretty lazy lately
Ben.
Right. And that's one reason why the PPC tree itself is maintained as a BitKeeper repository.
There's one fed from Linus official patches which is the "parent" of the linuxppc_2_4 tree which is our current "stable" tree. We also have a _2_5 tree with our "work in progress" things that will ultimately be rooted on Linus 2.5 when it comes to life.
Heh, well, don't worry for my feeling s;)
/. was not a good idea at the very beginning ;)
The fact is that we always try to avoid as much as possible changing non PPC-specific code, and when such has to be done (like endian fixes), we usually separate this clearly and go thru the specific maintainers.
But I also think bringing this whole thing to
Ben.
Hrm... Where the hell did you see that we were changing a _lot_ of non-PPC specific code ?
In your dreams maybe ?
How do you feel bashing other people's work like this ?
Ben.
Well, I won't even try to rely to most of the crap I've read here, sorry guys, but some of your comments are just so...
;) The truth is that we are in fact paid by Microsoft to write crappy code, flood Linus with huge and unmergeable patches in the clear intend to cause a Linux fork !
There are several points about the PPC arch. First of all, it's huge and very quickly growing. Why ? Well, in a single arch, it handles all PCI PowerMacs (with new hardware coming out of Apple every 6 monthes), CHRP boxes (including some RS6k IBM machines), PReP boxes, APUS, NuBus PowerMac is coming in, Gemini, and I'm forgetting some...
Add to that the embedded hardware (8xx CPUs, 4xx CPUs coming in soon) with the zillion of hardware variations.
So it's _huge_, it's quickly moving forward (remember USB, we needed to get that working for kbd and mouse on iMacs while most x86 boxes didn't even had Windoze drivers for their USB controller), and the necessary consequence is that patches are huge. Fortunately, most of them just touch arch/ppc, include/asm-ppc or PPC specific drivers.
But I'm not telling you all here
Ben.
With recent 2.2.18pre's, all that support has been merged in the main tree. You can either pick Alan latest 2.2.18pre or use Paul Mackerras usual rsync repository.
Hehe. Of course, Apple had it working in the labs for quite some time. The difference here is that Apple engineers did have the informations about how it was implemented in HW (especially how to start the second CPU), while we didn't and we had to reverse engineer it ;)
;)
It was just "fun" to announce it just before MacOS X Beta
Well, you may be interested to know, then, that the announce was posted to Slashdot by a TerraSoft employee ;)
In this case, you should read "Linux/PPC" as the "Linux running on the PPC platform", not LinuxPPC Inc. (the distro).
I certainly don't want to start any kind of NetBSD vs. Linux war, nor do I want to minimize the great work done by NetBSD coders for which I have great respect. ;) but I don't think they are more "mature" in their support of the PowerMac hardware. Most low-level hacking and reverse engineering needed to adapt "free" OSes on those platforms was done on Linux/PPC (except for the bits that filtered out of Apple via the original MkLinux and now with Darwin). Linux/PPC, afaik, is still unique to support sleep and power management features on PowerBooks (not all of them yet), and a variety of other hardware features I won't enumerate now.
NetBSD may have a lot of strengths over Linux (or may not, I don't know and I don't want to know now
Ben.
The ROM file is part of MacOS. It replaces the "old" >4Mb ROM that "oldworld" Macs had. It's loaded from disk by the firmware (and in our case, by the emulator which simulates the firmware).
For example, if you buy a MacOS 9 "box", on the CD, you'll find the ROM file. You just need to copy it over to the linux filesystem, strip it, and then, boot MacOnLinux from the CD in order to do the MacOS install.
I don't know where you found those informations, probably in some ultra-liberal activist site. I tend not to answer to sensible political comments like yours, but I really can't let you say such aberrations while having a score of 5 !
;)
But maybe was it intentionally ?
Anyway, the _fact_ is that Jose Bove has always been non-violent, and his actions has always been more symbolic than anything else. The "confederation paysanne" never destroyed the MacDonald's restaurant with a bulldozer, but actually "unmounted" it cleanly, without damaging parts. This action was made after the officials were made aware of it, without any form of violence.
The attack on MacDonald's your are talking about, where unfortunately a 28 years old waitress died, was a completely different terrorist action issued by Britain separatists, and can in no way be compared or assimilated to Jose protests.
He (and his movement) has always been non-violent.
The interesting thing here is that I finally see how much non-sense (since I know this specific subject) can an anti-Katz actvist say
Ben.
That is just plain wrong, I don't know where you did get your infos from, but: - MaxOS X is based on Mach 3.0 plus some additions (mostly real time) from mach 4.0 (which is _not_ a newer version of Mach but a separate branch) and some Apple custom additions. This _is_ the Mach microkernel. However, they have made all sorts of optimisations to reduce the overhead of having a separate BSD layer, and so BSD is not implement as a server process but is somewhat "wired" into Mach. (Mach IPC-based API is turned into direct function calls and BSD is co-located in the kernel address space). Apple did already experiment with this technique on MkLinux. - Darwin and MacOS X are not distinct. Darwin is the foundation of MacOS X, there are not 2 different kernels. - They do, indeed, have nice copy-on-write features, and a bunch of other cool stuffs like the IOKit
For completeness, add to the powerbook feature list:
- USB
- video is a R128 mobility with external CRT output on an AGP bus
- 16bits/44KHz sound ouput & input (+built in mic)
- hot-swappable media bays
- Cardbus & zoom video
- irda
I just tried and all I got is:
No match for "MASSLINUX.COM@WHOIS.NETWORKSOLUTIONS.COM".
And
No match for "209.219.204.3@WHOIS.ARIN.NET".
But this may be my broken setup...
That's unfortunately not so simple. For example, the way GM food is distributed makes it almost impossible for people to have a choice. There is almost no tracability, local grown beef can be fed with industrial food made from GM stuff without beeing marked GM, and more...
Actually, one representant of Monsanto agreed, in a public interview, that if GM food was not prohibited, it would be difficult and certainly very expensive to buy no-GM guaranteed food.
The idea that people have choice and vote with their money is a myth. A strong myth, especially in the US. People buy what they are offered and what they can afford.
Did you actually read those technotes ? They don't really contain anything more than an overall block diagram, nothing like a register layout of the new chipset. ;)
Fortunately, there's Open Firmware which contains a few more details available to who can read Forth, and parts of those new ASICs are similar enough to the old ones so that extrapolating was not difficult enough.
Also, those new machines are closer to CHRP than any other previous Mac
There is an interesting law for this in France: You can only do comparative ads if your ad is backed up by some kind of real benchmark-like comparison, or based on an independant comparison (done by the press for example). I don't know the exact details, but the result in reality is that we almost never see comparative advertising here...
This is an interesting point. Actually, European countries are trying to block hormon meat, and European population appears to be strongly against GM modified food.
On another hand, it appears that USA Gov. is trying to force european countries to accept those products via mondial commerce organisation. (we already have to pay huge penalities for refusing meat with hormones).
You can't turn it off. The PMU does a lot more than sleep mode, it replaces Cuda (ADB controller) which manages the keyboard and the trackpad, and it handles the nvram on those machines.
What you did probably turn off is the compile option for Power Management. This option enables the sleep code and should be harmless unless you actually try to sleep. For sleeping, there is the snooze command, and there's a little daemon (pmud) written by Paul, which monitors your machine and can excute scripts when the Power cable is connecter or disconnected. pmud also handles sleep when the lid is closed.
I received it but I'm not a US citizen
;-)
too bad...
Uncle Sam still giving you one hand while keeping the other securing the wallet