Just a few days after the report that Mandrake 9.2 can fry some LG cdroms (workaround has being issued in the Linux kernel in the meantime), now users complain that Mac OS X Panther renders unusable/fried firewire hard drives, mostly some LaCie models and models that are both USB and FW. In the meantime, MacSlash has posted a collection of bugs found on Panther here and here, while backwards compatibility with Jaguar seems to not be great either. Hopefully 10.3.1 is not too far away!
Apple’s Mac OS X Panther Fries FW/USB-Combo External Drives?
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I have a CD-RW from LaCie, should I unplug it now before I try anything like burn a CD?
Not all models seem to be affected. Mostly the dual USB/FW *hard* drives seem to have the problem.
what fries them?
are they slim line drivs that are powered off the firewire?
if so, could it be that the fire wire is sending to much current to the drive?
The LG CD-Roms that mandrake was frying were being fried by a valid series of ATAPI commands sent by the kernel. If it’s the same thing for these drives, then it doesn’t relate to the amount of data being sent to the drive.
how can software fry hardware? surely that is really dangerous? what is stoping some bitch writing code to trash all my peripherals and calling it freepr0n.bin/freepr0n.exe ?
Get the hardware to do something out of spec and erase it’s firmware. You need firmware to flash firmware. As to what prevents it? Nothing really. The next big windows worm could include this and cause physical damage.
This has already being done. There were visues in the early ’90s, where they could fry monitors. Since then, the monitor manufacturs try to be more careful in their designs and the voltage regulations and the kind of signal they get from the VGA.
It was explained as a “flush buffer” command I believe. Because the CD drive does not handle it correctly, the firmware goes bye bye. It is more the fault of the manufacturer playing light and loose on specs than any fault of Linux or OSX.
I just cannot believe that manufacturers get away with this nonsense. It is beyond reason that an ATAPI flush command can be understood to be a firmware upgrade.
I could understand that a poorly implemented ATAPI command could lead to data corruption, but frying a peripheral? Insane.
What’s next? A cp command that updates your BIOS?
What bothers me is that there will be those that try to take advantage of this to spread FUD about Linux and OS X, the only operating systems that seem to be advancing the status quo if you ask me.
“Pussycat Pussycat, I love you.
Even though none of my peripherals do….”
I still haven’t forgiven Apple for naming a previous release after Atari’s ill-fated 64-bit console… (Hi Hasbro!)
Eugenia:
This has already being done. There were visues in the early ’90s…
Don’t forget those “cool” programmes that used to play music using the heads of the floppy drives in a Commodore 64!
Hell, now that overclocking is done via software (for CPU and GPU), the potential is there to damage hardware that way, too!
Or newbies tweaking their XFree86 config files.
I’m still surprised at the lack of viruses that reflash the BIOS!
how can software fry hardware? surely that is really dangerous? what is stoping some bitch writing code to trash all my peripherals and calling it freepr0n.bin/freepr0n.exe ?
There are several *extremely* destructive things that worms could do which cause permanent, irreversable, and often destructive changes to hardware. The easiest of these to exploit would be ATAPI locking. A worm could lock hard drives of host computers it infects after a specified period of time (throught an NTLDR hack on NT-based Windows systems, and directly through DOS on 9x.) Once a drive is locked, its data is completely inaccessable without a key. Such a worm would most likely cause the drive manufacturers to divulge their master keys, which would effectively end the security of ATAPI locking.
It appears from the reports that there is a bug in panther that corrupts the partition map and sometimes more when rebooting with the firewire drive plugged in and mounted. From the looks of it, it appears that unmounting and unplugging the firewire hard disk before rebooting will prevent this. But it’s still a little early to be 100%.
i have various usb/fw hdds and opticals from a few different vendors (sony, lacie, qps). no problems whatsoever.
During ‘Night of Panther’, they told me I can get my copy in 2 weeks, because I bought a PB recently (dunno why though..)..
Since I use one of the devices affected by this bug, I’m pretty glad that I didnt get Panther yet..
I hope they get it fixed soon though..
I’m still surprised at the lack of viruses that reflash the BIOS!
Chernobyl virus used to do that. Fortunately, modern virus writers don’t do that anymore.
I’m still surprised at the lack of viruses that reflash the BIOS!
They’re not common because there are so many different types of motherboards and BIOSes out there. Either the virus would need to be huge, having compatibility modes for every known motherboard/BIOS combo, or it would only affect a small number of users.
There are also the troubles of BIOS-flashing from protected mode OSes like NT and Linux without special drivers to allow it.
Personally, I don’t regard BIOS or firmware reprogramming as “physical damage”, because really it isn’t – the correct code could be reinserted into the hardware with the right tools. “Physical damage” to me means motors being burnt out, monitors frying, read heads crashing into the disk, etc. Whilst for most people a zapped BIOS or firmware means replacing the hardware in question – because getting the right code back in there requires tools they don’t have – at the end of the day the *hardware* is still working, it just can’t be used.
A chip with firmware code (or an ASIC or FPGA, for that matter) that can be crashed (as in infinite loop etc.) is defective (but this can be fixed with a reboot). If it breaks in a more permanent way, it’s downright broken. People should sue the vendors/manufacturers, but I bet that’s very difficult because of very extensive disclaimers.
Anyway, if software can break hardware, it’s (maybe) a bug in software, but definitely a defect in hardware (and the product needs to be replaced free of charge, with a working, “repaired” version).
I once had to design a FPGA. Very simple, but I think that 75% percent of the design/implementation time was spent testing. For more complex stuff, the ratio of test time is probably lower. But the idea was that _any_ hangup was inacceptable, rebooting wasn’t an option. I guess people tend to shorten the testing time and these messed up CDROM drives are one of the results.
design a FPGA = design configuration code for a FPGA, the circuit to be implemented in the FPGA
sorry.
Its laughable to blame the makers. They produce good which work. They test it on whats available.
For Apple, or indeed Mandrake to supply a piece of software that causes failures is not acceptable. Mandrake 9.2 should not make such changes compared to Mandrake 9.1, but then when you build a system without any safety net, ie open source, and with no guarantee beyond the exellent efforts of those involved, you will sometimes get these problems.
Personally, if Panther melts down the firmware in a peripheral of mine, I expect Apple to fix the code, and pay for repaires. Now thats all questionable. You can blame the peripheral manufacturer, but if they have never said their parts work with Panther you are going to struggle to make the case.
ATAPI devices have been around for long enough that for people to start frying devices that in general work happily in most cases raises questions not just about the peripheral, but about what the hell the coders have done as well.
Thats my angle on it..
AdmV
No pun intended, but this is precisely the reason why you would like to use an Open Source OS. When you buy an OS, quality is NOT quaranteed at all! How are these unlucky, Mac OS X Panther users expected to work with their system properly? Most notably, they paid nasty dollars for it all.
Um. Did you miss the fact that Mandrake 9.2 did a similar thing with LG CDs?
No, I’m not gonna buy an OS unless people get their act together and stop producing bad code.
Define “bad” code. Bad as in virus code or bad as in low quality. Look Mac is not my thing but I would never ever claim Apple makes bad code (as in low quality).
Just go for a *nix for which you don’t have to pay. Really, you’ll be doing yourself a favor.
Oh wait I’ve got an LG CDs and that Mandrake 9.2 sure looks nice and free as well. I’ll try it. Free code is == flawless code, according to you right?
There were allot of Developers who tested panther on their systems, and even more people who were downloading it from some warez channel. Why didn’t they catch any of these bugs?
Um…the kernel and BSD subsystem layers of Mac OS X ARE Open-Source. You can go download them right now. You can go check out the development version from CVS. And guess what? None of the volunteer OS programmers caught it either. And yes, there are OS programmers who work on it. See OpenDarwin.org
But of course, Open Source programmers NEVER make mistakes, so those drive can’t ACTUALLY have been damaged.
Your angle might be right if this was a known flaw in these drives. But infact it was not. Mandrake even tested on all the models that are getting messed up, but it seems they didn’t test all the firmware versions, which would be have been insane.
Well I guess these ‘big bugs’ with Mac OS 10.3 just prove that Apple’s programmers are only human. I’m sure Apple will release a patch, or a few, to fix most, or all, of the known bugs.
This does want to make one look back at all those sites stating 10.3 was developing fast and wonder 🙂
I am glad my western digital drive is fine
As far as backwards compatibility, classic went the way of teh dinosaur for me a loooong time ago
I really don’t see how it’s fair to lay blame at Mandrakes door. Maybe this should’ve come up in testing, but if you’ve read what the problem is then how can you be anything but angry at LG for their poor firmware?
Um. Did you miss the fact that Mandrake 9.2 did a similar thing with LG CDs?
No, I did not, but Mandrake is a special all together, it is not even near the level of other *nix distros.
Free code is == flawless code, according to you right?
Not quite, but at least you’re not getting “emotional” when things don’t work as they should. This is quite a difference. Sure, I like Mac, but faulty Panther code may not be fixed within a couple of days or so. If you’ve paid for it, you feel like a child who didn’t get the ice cream. This really sucked as far as I can remember.
But of course, Open Source programmers NEVER make mistakes, so those drive can’t ACTUALLY have been damaged.
I wouldn’t say NEVER, but far less. Open Source programmers do not have a problem with $$$-driven deadlines. Their code is by definition better, in the sense that code will be eye-balled properly, and if it ain’t good enough, they simply delay a release and see to it that the code works. Sure, there are bugs, but these critters are terminated within a small time frame.
Take my word for it: Open Source programmers are better.
You really dont know what are you talking about. The new interesting frying function in OS X is entirely developed in Apple’s Labs.
That is the second mishap in very short time. Assuming that users had to paid few times already for OS X upgrade within last two years(which makes it more expensive than other OSes) OS X developers should do better. Anyway for any OS (free or not) there is obligation to release quality product unless developers say up front that the risk is on user’s side and they (developers) will hold no repsonsability whatsoever for the damages. However I bet that it will be easier to make Apple pay for damages than Mandrake
For Apple, or indeed Mandrake to supply a piece of software that causes failures is not acceptable. Mandrake 9.2 should not make such changes compared to Mandrake 9.1, but then when you build a system without any safety net, ie open source, and with no guarantee beyond the exellent efforts of those involved, you will sometimes get these problems. .
You can get those problems with any software. The fault lies squarely with LG in Mandrake’s case because they made defective hardware period. Did they issue a new firmware after confirming their bug? Yes. And if you actually did some research, you would know that, yes, Mandrake did in fact test their software with the LG CDROM models that this bug affected. However, none of the CDROM’s they tested had the old firmware and thus the issue did not come up.
Try actually reading some before throwing around the blame.
I was about to put my ThinkPad T21 on eBay and get a G4 12″ iBook, but then I read about all those problems people are having at macslash. I’m having second thoughts.
One fact provides comfort, however. It seems that people who are upgrading from old versions are having all the problems, fresh installers seem to by happy.
I was also thinking of getting an external USB2/FW hard drive. Does anyone know if there are any problems with Maxtor and Western Digital external HDDs?
I actually do know what I’m talking about, and I resent your implication that I am talking though my hat.
I have personally looked through the Darwin sources, and I have built my own kernel from them. It’s irrelevant if it was developed by Apple: the code is still open for review, just like Linux code that is developed by Red Hat or any other company. So screw off and go insult someone else.
Secondly, this does not actually present any significantly different case than in Linux. If a Linux user had this happen, he has two options: a) build his own kernel as soon as the tree is updated or b) wait for his distro to release a patch for it.
Guess what? A Mac OS X user has EXACTLY the same options. He can check out the latest Darwin sources and rebuild it for himself, or he can wait for Apple’s patch.
Before someone calls me a troll, I use Panther, FreeBSD 5.1, and Windows Server 2003 on a daily basis. I like and respect Open Source, but it is NOT a magic bullet like so many people claim.
Well, for a little perspective…
I have had zero problems with Panther and my LaCie FW drive works just fine. I did a fresh install on my iBook and everything just rocks.
All the Mac users I know (and I know quite a few) have had zero problems with Panther. From what I have read (and trust me, I read ALOT of Mac forums and discussion boards) the *VAST* majority of Panther installations have gone without a hitch. You mention Macslash — check out their poll on “First Impressions of Panther”:
http://macslash.org/pollBooth.pl?section=&qid=123&aid=-1
Obviously it’s an unscientific poll. Still, only 1& have reported “Hosed my machine” and only 2% reported “Underwhelmed.” 18% selected “Freaking phenomenal” and 39% selected “Pretty Impressed.” The point is that the vast majority of people out there have had nothing but smooth sailing.
Of course there are going to be some problems with any major upgrade to an OS. However, despite the press this FW drive problem is getting (and I’m not saying it’s not a problem — it clearly is a bug — but I’m saying it’s not a bug that affects everyone), Panther is an awesome upgrade to an already awesome OS.
Go for the switch. You won’t be sorry.
i have to agree. i know 30 plus people who have done installs and have zero problems. i know 2 with several problems. there is one thing in common, the 2 with troubles did a regular upgrade. the ones who have no problems all did “archive and install” and never looked back.
and honestly, “all those bugs?” i count a few. far less in fact than in every upgrade i ever did with windows.
sounds like a lot of people who have never seen the program, pointing at a dozen people out of thousands who have a couple of issues. more people walked out of the chandler apple store in a ten minute period (while i watched) with copies of panther than have had serious trouble. most people i have talked to have had to reinstall a few apps like stuffit, but that is about the extent of the problem.
There is big difference between FBSD kernel and Apple’s. Apple borrowed userland (mostly FBSD plus some Net) not the kernel from *BSD. LaCie models and models that are both USB and FW are even not listed on Darwin HCL. Besides Latest Darwin corresponds to Mac OS X 10.2.6 (dated from 07-2003) so the latest modification in the OS X kernel not available in Darwin.
I think that you need to wait some time before fink will get even close to BSD paskaging systems. There is coordinated effort from fink, gentoo and darwinports but at the moment OS X packaging system s nowhere close to BSD.
Anonymous wrote:
“There were allot of Developers who tested panther on their systems, and even more people who were downloading it from some warez channel. Why didn’t they catch any of these bugs?”
Actually, they did – read the very first post to the Apple Discussion Board about the problem. It was also repeated on MacInTouch early on as well – i.e., someone with access to the Developer Seeds mentioned that it had existed, then went away, then mysteriously reappeared around the time of 7B80. He was incredulous that Apple shipped 7B85 with this bug having crept back in …
Eugenia, did you actually read the articles that you referenced in your headline? I read them because I just upgraded my Dual G4 at home with Panther and I have an external FW drive I have yet to turn back on. Worried about it getting *FRIED*, I checked the articles. The problems are that on some drives the FW800 ports do not work (FW400 does), and with others the contents of the drive is getting fragged. I didn’t see a single post that said their drive was actually destroyed, just that the contents may be lost and it needs to be reformatted.
Now we have nearly 40 posts with people freaking out about Panther destroying hard disks, which is not the case (yes, Mandrake did destroy internal CD drives, but this is different)….
I have a BUSlink USB2/1394 Combo and have experience no problems.
There is big difference between FBSD kernel and Apple’s. Apple borrowed userland (mostly FBSD plus some Net) not the kernel from *BSD.
XNU, the Darwin kernel, uses FreeBSD’s Unified Buffer Cache and VFS. It’s also safe to assume that the kqueue implementation present in Panther was pulled from FreeBSD as well. I also believe Apple has introduced improvements to the XNU VMM by cribbing them from FreeBSD (FreeBSD’s VMM was derived originally from Mach’s)
It’s funny because my LG CD-RW(GCE-8160B) drive just crapped out last week while running Windows Media Player on XP. If there really is a firmware problem, could this be related?
Mac OS X kernel is based on =< 5.0 and Mach 3.0. That is true but it is still not in par with VMM performance of 4.x and or even better – current. With borrowed VMM, it is still big difference. Otherwise the approach to some aspect of memory would make OpenBSD and linux kernels similar (red-black trees). But that is not the case.
OS X problems with LaCie have nothing to do with BSD code at all. And that was first question – can one check open code for errors.
Running Panther on my G4 1.0 GHz powerbook and I use an external Firewire/USB 2.0 DATACOM hard drive (120 GB).
If anything, Panther helped speed up my transfer rate from the old Jaguar. I also just attached a new USB 2.0 external drive (small form factor) for portability sake.
Very nice, Very fast.
No problems – Yet!
Bill
FYI: 10.3 is using the development branch of *BSD. Apple apparently decide not to stick with the production version of *BSD and went to the development branch. The only reason that I know this is because I just looked at the box at a local computer store.
Not sure what version that are using for 10.3. It only says 5. Is it 5.0, 5.1 or the upcomming 5.2??? Just a heads up… Apple should have stuck with the 4.x branch….
good luck…
FYI: 10.3 is using the development branch of *BSD. Apple apparently decide not to stick with the production version of *BSD and went to the development branch. The only reason that I know this is because I just looked at the box at a local computer store.
There’s no reason not to do this, provided they code they used underwent regression testing, which one would expect from Apple.
Not sure what version that are using for 10.3. It only says 5. Is it 5.0, 5.1 or the upcomming 5.2??? Just a heads up… Apple should have stuck with the 4.x branch….
The use of code from FreeBSD 5.x would have little effect on the use of FireWire drives. OS X uses a significantly different driver architecture (IOKit) than any other operating system in existance (heritage excluded). IOKit is object oriented and written in C++. Drivers can extend/inherit code from other drivers, or lower level drivers can provide partial functionality if higher level ones cannot be found.
I never claimed that. The code that does make up Mac OS X is called Darwin, and it is open sourced. The fact that it is partially related to the FreeBSD code is irrelevant to my point.
Oh, and another thing. You’re wrong. Darwin 7.0 was released a few days ago, and it coincides with Mac OS X 10.3. In addition, the latest Darwin is ALWAYS available through CVS at OpenDarwin.org
Yes Darwin 7.0 was released 1O-24.
However,
quote:
“Um…the kernel and BSD subsystem layers of Mac OS X ARE Open-Source. You can go download them right now. You can go check out the development version from CVS. And guess what? None of the volunteer OS programmers caught it either. And yes, there are OS programmers who work on it. See OpenDarwin.org”
Frome the lines above – you are suggesting that the error in coding could be easily picked up from darwin source code. Other interpretation makes no sense in the view of original post. Now point me to the darwin or any BSD code that would cause similar problems as LaCie in 10.3.
A lot of code is originated from *BSD but assumption that Darwin CVS is directly related to OS X code is oversimplification that leads to the mistakes as above. On the other hand FreeBSD is in progres of porting msdosfs tools from Darwin. After developers port it I can look for relevant source in darwin?
Mac OS X is not opensource. Apple is using opensource and Apple is also experimenting with some BSD code using darwin. I dont see anything wrong with this approach as this is in BSD license. But you are mistaken suggesting that Darwin CVS is the same as OS X source.
After developers port it I can look for relevant source in darwin?
I meant to say:
After developers port it I can look for relevant source in darwin in case that anything (not related to msdosfs) will go wrong?