Re: Powerbook Install
Chris Tillman wrote: nvsetenv boot-device hd:2,\\:tbxi This tells OpenFirmware to look on the second partition of the main hard disk for a blessed startup file. Blessed has nothing to do with it. Segher
Re: Bad sound from iMac headphone jacks
Matthew Buckett wrote: I'm running debian woody on a iMac 350 (slot loading) and I seem to get problems with the sound when it is sent through the headphones socket on the right of the machine. The sound is rough, not crackly, it's That is not a headphone socket -- it is a line-level socket. Segher
Re: Bad sound from iMac headphone jacks
Matthew Buckett wrote: That is not a headphone socket -- it is a line-level socket. Ok, not being an audio person, what is the differnce between a line level socket and a headphones socket? Different voltage, different impedance. Is a headphone socket for devices without build in amplifiers? and a line level socket the opposite? Somewhat like that, yes. Presumably using headphones isn't the cause of the very poor sound on the line level socket? It is one cause -- but if it sounds differently in MacOS and in Linux, there's another problem, too. Segher
Re: Sid, mp3blaster and ogg Files ...
Michel Dänzer wrote: (IMHO libvorbis is broken for not defaulting to native endianness, but that's another story) Why don't you file a bug at bugs.xiph.org? Please do. Segher
Re: cpufreq on TiBook 400 MHz
Krisztian Mark Szentes wrote:
Hi
I wonder why there is no /proc/cpufreq on my Titanium PowerBook G4 (400 MHz).
Because no-one filled in the right values yet.
In linux-2.4.20/arch/ppc/platforms/pmac_cpufreq.c it says:
/* Currently, we support the following machines:
*
* - Titanium PowerBook 800 (PMU based, 667Mhz & 800Mhz)
* - Titanium PowerBook 500 (PMU based, 300Mhz & 500Mhz)
* - iBook2 500 (PMU based, 400Mhz & 500Mhz)
> * - iBook2 700 (CPU based, 400Mhz & 700Mhz, support low voltage)
> */
From Apple's spec sheets:
PowerBook4,1: high speed is 500 or 600, low speed is 400.
PowerBook4,2: no spec available.
PowerBook4,3: high speed is 600 or 700, low speed not mentioned.
These are the iBooks, I don't have the sheets for the TiBooks
readily available; maybe look in there?
I seem to remember the PowerBook3,2 @ 400 has a low speed of
300MHz, but I'm not fully sure about that -- better check it
yourself (apple.com -> developer -> hardware -> all products
or something like that, it's not too hard to find).
Or ask the System Profiler (in MacOS), under heading "hardware overview".
and:
/* Else check for TiPb 500 */
else if (machine_is_compatible("PowerBook3,2")) {
/* We only know about 500Mhz model */
if (cur_freq < 45 || cur_freq > 55)
goto out;
Try to remove these last three lines (your own risk, of course).
hi_freq = cur_freq;
low_freq = 30;
has_freq_ctl = 1;
cpufreq_uses_pmu = 1;
}
Is my machine missing some crucial hardware? I have a TiPb 400, don't I?
It doesn't miss anything. You do have a PowerBook G4 "Mercury" @ 400MHz.
Have fun,
Segher
Re: kernel compiling problems on iMac CRT
Michel Dänzer wrote: I wonder why i2c-keywest doesn't get loaded for you. Because an iMac G3 doesn't need it? ;-) If I was him, I'd try disabling the emu10k crap he enabled... Segher
Re: Apple PowerMac G5
Sven Luther wrote: > So you think that the north-bridge could be an Apple-IBM common design ? It's an Apple design. > The rest is not as important, as it is connected trough a HyperTransport > bus, so any HT chip will do. The funny thing is that they have a special > PCI-X bridge sitting between the north bridge and the super IO chip. I No. The "super-io chip" (a KeyLargo derivative) is connected to HT. > guess this one is just a off-the-shelve piece, while the IO chip is of Correct, the HT<->PCI-X bridges for the slots are AMD parts. > more common apple lineage. It has serial ATA, standard ATA for the I believe the serial ATA is a separate (non-Apple) part. > SuperDrive (this one is a guess), Correct. > USB 2.0, A separate, non-Apple part (hi Ben). > Firewire 800, I'm not sure about this. The FireWire 400 is the same as before, though. > networking and Still SUN GEM. > audio. Still TAS3004, but with added optical links. > Naturally, one could even build a PPC 970 motherboard with This > IBM/Apple northbridge, some PCI bridge, and Nvidia's HT connected > southbridges, not that we have much drivers for them though. Yes, but why? And engineering adequate cooling for the 970 seems to be non-trivial, to say the least... Segher
Re: some basic problems with ppc assembler
Brad Boyer wrote: Do you actually have a document for the official ABI? I found a link to a document in some old messages, but it was a dead link. A generic search of the web didn't turn up anything, although I might have been searching for the wrong words... System V application binary interface I got mine from http://www.caldera.com/developers/devspecs/gabi41.pdf System V application binary interface PowerPC processsor supplement I got mine from http://www.esofta.com/pdfs/SVR4abippc.pdf Segher
Re: some basic problems with ppc assembler
Benjamin Herrenschmidt wrote: The stwu will do r1=r1-32 and store old r1 there (stack pointer). This will basically sets up the stackframe for the function. The LR is backed up in r0 (the return address) and stored in the stackframe. r31 is a non-volatile register, since the compiler thinks it will need it, it backs it up (though the reason it uses it is unclear, looks like lack of optimisations). Finally, it bask up r1 (stack pointer) to r31, I'm not sure why at this point. The stwu r1,-32(r1) makes room on the stack, but r31 is the frame pointer. Not that this function uses it, and of course it can be optimized away to use only r1 (the stack pointer). Segher
Re: some basic problems with ppc assembler
Oliver Ripka wrote: .LC0://I guess this means .LC0 .LC0 is a label. .string "Hello World" //is a global string Outputs a string to object code. It can be accessed via the label .LC0 . .globl main //within the main label. Declares 'main' as a global symbol (it will be marked thus in the object file). .type main, @function //whats that for? Marks main as a function. main: A label again. Segher
Re: ibook2
digger vermont wrote: On the alsa mailing list I've asked if the reason for no recording is due a lack of specs for the sound chip or a lack of desire and time on the developer end. I've yet to get a response. Does anyone here know? The docs for the current Mac sound chips are publicly available from Texas Instruments, and very readable. Search their site for the specs to the TAS3001C and/or TAS3004. Have fun, Segher
Re: problem with sound driver in iBook2
Michel Dänzer wrote: Yes, I was just looking what happens. Deep inside I was hoping that recording capability might perhaps be added with not too much effort. How about Darwin, couldn't we borrow some code of their driver, or is this code not freely available? I doubt it, but as Segher pointed out, the specs are available. Good luck. :) Well, you not only need the sound chip spec, but you also need to know how certain function are connected in the particular machines, for example, which GPIOs are the input/output selects and the function detects (headphone connect detect, etc.); the Darwin driver source can help here. Some (most?) of this stuff is put into properties in the Open Firmware device tree, but not all. Also note that older Mac sound systems (DACA, Screamer) are less "clean". And for some systems, there are separate extra chips for output stages or whatnot (S/PDIF receiver? :-) ) All this asks for a a heavily modularized sound driver design, not the big blob that dmasound is. The current ALSA drivers are better in this regard, I think. Cheers, Segher
Re: PC graphic adapters on Mac Hardware
Jeroen Roovers wrote: As I said before, if your Mac has Open Firmware, you can (theoretically, which means I've read about it :-) use a serial It works fine. Do note to use 57600 baud on newer machines, not 38400 as the Apple technotes say. connection to get to a boot prompt, and if Linux is set to use the Open Firmware interface at boot time, one ought to be able to see what's going on and interact with the machine properly. On my B&W G3, I ought to be able to use the USB interface as a serial connection, USB is not serial. Repeat that: USB is not serial. USB is a wacko very much over-complicated over-engineered "standard" with everything but the kitchen sink included. Heck, it has four wires, that is at least one too many ;-) It would take more code to implement functional early-boot- console-over-USB than it takes code to implement all of the early booting. Much more. Forget about USB. It's not worth it. You really want to use a simple RS-nnn or SCC or whatever 2- or 3-wire connection. Cheers, Segher
Re: PC graphic adapters on Mac Hardware
Benjamin Herrenschmidt wrote: What we would need now is for somebody to port uBoot's emulator to some userland app in Linux, and actually try to see if we can get an x86 typical Radeon card to boot in a Mac or not... While that might be a reasonable short-time plan, what we *really* need is for graphics card manufacturers to just include OF drivers on each and every card manufactured, and recognize x86 as the oddball it is, not treat cards with a proper firmware as special expensive "Mac cards". I'm a bit hazy on how to get them to do that, though -- the only way might be to make the OF-booting market so big that it'll be cheaper for the manufacturers to do so. Still striving for World Domination, Segher
Re: Newbie (to ppc), but glad to be here.
Benjamin Herrenschmidt wrote: I suspect we should probably just expose the PMU i2c bus to userland some way and have a userland tool to deal with that, I'm not too fan of doing a kernel driver... I also have not much time. We _already_ expose it to userland, via the generic PMU interface. And *please* don't add an easier-to-use userland interface to the PMU IIC bus; it's too easy to lock it up / cause other bad interactions, so if userland needs any more, a proper driver should be written. We have enough rope already :-) The same applies to the KL and the UNI-N IIC busses. Segher
Re: CPU and Memory usage of XFree86
Michel Dänzer wrote: > On Wed, 2003-10-01 at 00:42, Nirmal Govind wrote: > > Hi.. I've noticed that the XFree86 process on my ibook is taking up a > > decent chunk of memory and some CPU and this seems to be slowing down my > > machine a bit .. when I do 'top', XFree86 is a process that's constantly > > at the top of the list.. > > X server memory consumption is a FAQ and has recently been discussed > here, but the X server should normally never hog the CPU (and it doesn't > here) unless there are clients flooding it with requests. Of course, notice that the Xterm (or worse, kterm etc.) you're running the top command in _is_ flooding the X server... :-) (I've actually measured this some ten years ago -- the result (on *that* old system, your mileage will be totally different I'm sure) was that top+xterm took 10%-20% cpu. top itself *still* takes ~2% cpu on busy systems (lots of processes), and the Xterm situation has not considerably improved either, I feel confident in saying). Segher
Re: Installing onto external FireWire drive.
> Oh, and the yaboot config is a bit of a pain because Apple can't seem to > make up their minds about what to call the devalias for the firewire > controler. On the iBook it's fw, the G4 fwx, I think it's back to fw on the > G5, maybe. It only was fwx on some older systems, for which the built-in to KeyLargo FireWire controller doesn't work. On these systems, there's an extra (separate chip) FireWire controller, called fwx, _and_ the internal (but disabled) controller, called fw. And yaboot config is a pain anyway. Whoever decided to model it after lilo config needs a bit of punching (hi Benjamin ;-) ) Segher
Re: CPU and Memory usage of XFree86
Michel Dänzer wrote: > FWIW, top uses about 0.3% here (the 2.6 kernel might help there), > gnome-terminal and the X server less than 1% together. Okay, another datapoint... PowerBook3,2 500MHz, 2.4.20-ben10, 30 processes, no load. top -d1 takes 2.9% cpu, top -d0.5 takes 5.6%. Starting X... 37 processes, top -d1 takes 1.2% cpu, xterm takes 0.4%, XFree86 takes 3.9%. top -d0.5: top takes 2.3%, xterm takes 0.5%, XFree86 takes 7.4%. This system is not the newest, and my X runs in framebuffer mode. Btw, the reason that top itself on the console takes more time than in the xterm, is that on the console, all the fb console output time is calculated towards the top process. > I guess the bottom line is that the X server obviously does need some > CPU cycles to do its work. :) Unless it's constantly hogging the CPU, > that's normal and shouldn't have a significant impact on overall system > performance. The kernel scheduler and VM are much more likely to matter > there. Certainly. Just be aware that tools like top will very measurably affect the data that they themselves are collecting. I prefer using vmstat for this reason (and it gives you more and more useful data, anyway). Segher
Re: Would setting the CONFIG_RTC option break the powerpc kernel on your machine ?
No, that's VIA as in Versatile Interface Adapter, not as in VIA Technologies. It used to be a separate chip in the old 68000 days, but now it's integrated in some Mac fabric (e.g. Mac I/O). Actually, there is no RTC chip _at all_ in NewWorld Macs. The PMU itself keeps the time (from a 32768Hz clock). Segher
Re: Would setting the CONFIG_RTC option break the powerpc kernel on your machine ?
Actually, there is no RTC chip _at all_ in NewWorld Macs. The PMU itself keeps the time (from a 32768Hz clock). So from a technical point of view, the PMU is the RTC chip, right? Not really, it's a very different mechanism. But you could say it is yeah, if you don't mind being sloppy ;-) Segher
Re: Key repeat (kernel Bug?)
> i get that too. i'll be scrolling down, or whatever, and it keeps > scrolling down. only if i'm scrolling down, though, which i find > strange. It happens for me on other keys, too. And it misses the shift/ctrl/whatever release sometimes, too. It's just that you keep key-down pressed more often than other keys ;) Segher
Re: unsupported apple hardware and darwin
> Actually, even Apple's own drivers are not fully open source. Some of > them are, some of them aren't (the PMU driver isn't, the temp control > and fan drivers aren't, the new machines ATA/100 controller driver > isn't, etc...) The temperature and fan controllers have nice specs; can't find the urls right now, but one of'em is for the Analog Devices ADM1030 (I believe that's the one in the rackmac and the windtunnel, but this is all from memory, so ymmv). Just ask if you need to know more. Regarding the ata100: which one? Cheers, Segher
Re: date is 1934 on reboot
> My TiBook relativeléy often resets the time to 1904, which is also too > far off for ntp to correct it. I then need to manually adjust it, before > I can sync it with ntp. > > It seems this happens every time the PMU gets reset, regardless of > battery availability. At least this TiBook _does_ have an internal bakup > battery (glued to the top cover together with the PMU, above the DVD > drive). If the PMU gets reset that often, beacuse you press the little reset key at the back, the solution is easy: just don't do that. There's hardly ever a reason to rest the PMU. If your box crashes, you can also bring it down by keeping the power key pressed down for about 5 seconds; this will not reset the PMU, and so will keep your time intact. Now why doesn't the manual mention this? Cheers, Segher
Re: Sound on g3 beige 233
marco wrote:
>
> Hi,
> I have no sound here.
> Added my user to audio. Could you tel me what to do ?
>
> ## ls -la /dev/dsp
> crw-r--r--1 root audio 14, 3 Mar 14 2002 /dev/dsp
>
> ## ls -la /dev/mixer
> crw-r--r--1 root audio 14, 0 Mar 14 2002 /dev/mixer
chmod g+w /dev/{dsp,mixer}
Cheers,
Segher
Re: G4 Cube CD Eject
"Albert D. Cahalan" wrote: > Don't add an option. Just do the eject. Even a PC user will like that. > It's also nice to notice the new disk immediately, instead of waiting > for the user to press enter or similar. *Please* don't eject without asking first; some people have laptops, and you too easily end up with a broken cd (or drive) when the machine tries to eject the cd while the machine is in your lap. Bad bad idea. Segher
Re: memory interleaving
Geert Uytterhoeven wrote: > > On Mon, 9 Dec 2002, David Ulrich wrote: > > Does the debian kernel support for the memory interleaving on a > > PowerMac 9500? (2.4.18 SMP) > > Isn't memory interleaving a hardware feature that's independent of software? Depends. Usually it's configurable on your host bridge, and the firmware takes care of setting it up correctly. Sorry, I don't know much about any OldWorld machine, so I can't help any further. Segher
Re: I'd like open the CD-ROM unit from Firmware
Gabriel Paubert wrote: > > On Thu, 19 Dec 2002, Heissu wrote: > > >Hi! > > > > I want to do that is in the subject. Everyone can help me? > > > > eject cd > > Gabriel. > Or, if you just want to boot from a different cdrom, hold OPTION-EJECT. Segher
Re: Brightness and Contrast controls on iMacs
Benjamin Herrenschmidt wrote: > > On Fri, 2003-01-10 at 11:14, Scott McMahan wrote: > > I'm looking for a way to adjust the screen settings on my iMac. I've > > search the list/newsgroups and the answers I've found don't work on CRTs > > (fblevel) or don't really address the right problem (xgamma). Is there > > a utility available to handle this problem? If not, is anyone working > > on one? If not, can anyone help point me in the right direction so I > > can start working on one? > > My understanding is that the screen (brightness, contrast, power > management and geometry) on iMacs is handled by tapping an i2c > device. On some iMacs, it's on one of the PMU i2c busses, I'm not > too sure about older iMacs, it's either on Cuda or on the video > card's i2c bus. > > There is no real support for PMU i2c currently in the kernel, > though you can eventually try to feed PMU i2c commands via > /dev/adb, though I'm not completely sure you'll get the replies > properly, that's something I have to investigate. You can send a request and then poll for a reply. Not sure if it works via /dev/adb, but I think it will (i2c isn't very timing sensitive :) ) Bit 0x08 of byte 0x0e of the system config says if you have such a crt controller (all UMA iMacs and eMacs do). The controller is at i2c address 8c, with an eeprom at a6; the eMac has an additional thingy at 98. All at pmu i2c bus 2. The commands you send it are two bytes: register, value. That's about all I can remember offhand. So you send the pmu something like 9a 09 02 00 00 8c 00 00 02 reg val and read back two bytes for status. > Most of the informations about the device itself can be found > in Darwin's source AppleOnboardDisplay module. Thanks! I'll go read that. Cheers, Segher
Re: vi oriented mailer .. was Re: Suggest a really good email client
micro-people wrote: > > just curious, but > > is there a vi friendly (or oriented) mailer out there ? > have been a debianppc newbie for years @ iBook-Tangerine > and wanna keep this machine promt-mode only You might like elm (elm-me). Segher
Re: g4 and switching consoles
J Q Private wrote: > > Minor note: > > I read, early on, that I'd have to type the alt/option > key first, then Fn, then F[1-7] to get the different > consoles, and that worked. Hitting Fn before > alt/option does not. Cmd-left and Cmd-right is easier. > But once X starts, it does not work anymore. I found > that hitting alt/option + ctrl + fn and then F[1-7] > does work. > > Switching to any getty'd console [1-6] to any any > other [1-7] requires only three buttons, but from the > X console (traditionally 7) only the four button way > works. Can't you change this in your X keymap? I suppose if all else fails you can always have your windowmanager run chvt on certain key events. Ah well, I don't run X11 at all, so I'm very happy :) Segher
Re: bloat tolerance?
Albert D. Cahalan wrote: This is kind of a survey for PowerPC Linux users. Is it OK to make 32-bit PowerPC apps bigger and slower to allow for running them on IBM's 64-bit hardware? This would hurt everyone running Linux on a Mac. My vote is *NO*. Why not have the people running such systems use a different binary? Segher
Re: Trouble with Dual boot MacOSX/Woody - Yaboot on G4 Gigabit
Clive Menzies wrote: My G4 is at least a couple of years old (Sawtooth?) AGP Gigabit - so it That's "Mystic", one revision newer than Sawtooth. Segher
Re: What is pbuttonsd ?
J Q Private wrote: I do _not_ have pbbuttonsd running. Pardon my ignorance, but I thought I needed pbbuttonsd to get F1 and F2 to change brightness. The PMU handles F1 and F2 itself, unless that function is disabled. Works fine for me. I believe this only works for builtin displays, though, and certainly only for "ADB" keyboards (that is, builtin keyboards). Segher
Re: Booting the Debian installation CD on Powerbook G4 12"
Benjamin Herrenschmidt wrote: On Mon, 2003-02-17 at 02:07, paubert wrote: Well, the timebase and decremeter are part of the PPC architecture. I'm not sure things like 4xx implement them "as-is"... It is part of the user-level 32-bit PEM -- they better implement it :) BTW for the "53 MHz PB", what are the following properties of cpus/PowerPC,G4: It is 53_300_000 Hz, and it is most probably a typo in the OF. All other such values are rounded to exact millions. If you're going to reimplement the delays using the time base, why not use the KeyLargo timer instead? It runs at the same frequency (18.432MHz) on all machines, so that'll save a multiply ;) Too bad for the non-Macs, but else you'll have to special-case the 601 (and maybe others) anyway. Segher
Re: Booting the Debian installation CD on Powerbook G4 12"
Benjamin Herrenschmidt wrote: If you're going to reimplement the delays using the time base, why not use the KeyLargo timer instead? It runs at the same frequency (18.432MHz) on all machines, so that'll save a multiply ;) Too bad for the non-Macs, but else you'll have to special-case the 601 (and maybe others) anyway. I don't want to rely on external HW, Why not? > IOs are slow, Duh... this is the delay loop we're talking about! > KL doesn't exist on all macs, Not interesting ;) > and other PPCs would benefit from TB based delays as well anyway. That's a very good reason though. KL timer would be useful for better calibration of TB during boot on pmacs though (Darwin recent versions use it for that as well). I know (you told me :) ) Segher
Re: Booting the Debian installation CD on Powerbook G4 12"
Gabriel Paubert wrote: If you're going to reimplement the delays using the time base, why not use the KeyLargo timer instead? And tie the bus on repeated I/O trying to read the timer, fighting perhaps with another processor on SMP? No thanks. The pci/33 bus is much slower than the arbus/maxbus/whatever you want to call it; furthermore, there's a maximum length you can "tie" the bus. This is not really a problem. It runs at the same frequency (18.432MHz) on all machines, so that'll save a multiply ;) Too bad for the non-Macs, but else you'll have to special-case the 601 (and maybe others) anyway. I don't know how you'll avoid a multiply unless you want to pass delays in strange units. It's better to have the The compiler can fold it. Segher
Re: review wanted of new Debian X FAQ entry
Martin Kuball wrote: The problem of course is to find out what kind of bus your machine uses if it's a laptop. Well, trial and error obviously revealed that my iBook uses ADB. All (reasonably new) Apple laptops use ADB. Segher
Re: some help here please
José Salavert Torres wrote: Well this is not an error or question about linux but I thought that ths was the best place to ask. I'm searching for information about powermacs architecture in order to make a work for the University and It's eery difficult for me to find information related to last powermacs and PPC 7455: -Memory map (directionable bytes...), It's PCI, you can map (most) stuff where you want it. > memory achitecture (CAS/RAS etc this is like pc's DDR I hope where to find). It _is_ the same memory as used in most x86 machines. See the specs from Apple on what exactly on what machine. -Input/output, interruptions treatment. It uses OpenPIC; if anyone can find me a spec for that... -Altivec. See simdtech.org -Caché behaivour(write through, write back) for the three levels. Mentioned in the 7455 User Manual. -DMA controllers and other pheripheral controllers (keylargo). There is no (public) spec for KeyLargo; you might want to search the web for a document called "MacTech.pdf" though, it describes some older Apple I/O chip used in CHRP. I just ask because I think that people that works in the kernel may know. I've found some information in developer.apple.com and in motorola's webpage (semiconductors section). You can read the kernel for some info; also try the Darwin kernel. I will join all the urls to make a source of all about this. I think this is more necesary than my work. I need to the name of a document that apple made about making clone macs from oher companies, I don't remembeer the name and I lost it :-( Apple Technology in the Common Hardware Reference Platform, perhaps? It's the MacTech.pdf mentioned above. Segher
Re: recording with alsa on tibook
Brett Carter wrote: Nope, I believe recording works on the 'clipper' chipset (tibooks v1 & v2), but not yet on the 'snapper' chipset (ti v3 & above). "v2" (PowerBook3,3) uses the "Tumbler" chipset (TI TAS3001C), "v3" and up (PowerBook3,4 and up) use the "Snapper" (TI TAS3004). I didn't hear the name "Clipper" before; any references for that? Segher
Re: Install debian on a iBook with external FireWire hard drive
[EMAIL PROTECTED] wrote: I try to install debian on a iBook. I have an external firewire hard drive on which I want to put debian. I created a HFS+ partition and put the following files: yaboot, yaboot.conf, linux.bin and root.bin. Now I try to boot on this partition. What is the openfirmwire command to do this ? I tried 0> boot fw:1,yaboot but it does not work... Maybe you're trying the wrong partition number. Try dir fw:1,\ to see if that is the case; if so, vary the number until you found the right one (2 or 8 are nice tries). Segher
