SAS (Installer doesn't see hard-drives)
![](https://seccdn.libravatar.org/avatar/6eab2c4a7f74247d7445c9cc51f74c35.jpg?s=120&d=mm&r=g)
Hi, I try to install SuSE Linux 10.1 on a Dell 2950. The problem is, that the installer just doesn't see the hard-drives. This is a system with two dual-cores and 4GB of RAM. The heart of the problem is perhaps the LSILogic SAS 1068-IT (v6.6.0.2) controller (there's no RAID hardware). By default, the installer loads the following modules: Fusion MPT SCSI Host Driver Fusion MPT SAS Host Driver But there are other drivers, which seemed to me more appropriate: LSI Logic MegaRAID LSI Logic MegaRAID SAS LSI Logic Management Module The problem is, when I load the LSI Logic MegaRAID SAS module, it is rejected, saying: Module not supported by Novell, setting U Taint flag. Interestingly, with the delivered SuSE Enterprise Linux CD-ROM, I have no problems installing, but I also have only one month of updates. Has anybody already had that problem? Phil
![](https://seccdn.libravatar.org/avatar/696255f62c3cefe5d517f913058b70bc.jpg?s=120&d=mm&r=g)
Hi Philip, Questions & comments follow... On Friday 13 October 2006 09:11, Philip Mötteli wrote:
The problem is, that the installer just doesn't see the hard-drives.
I presume you're discussing the standard 10.1 GUI installer and not the ncurses (text) based 'manual' installation system? I'd say your problem is the GUI installer not correctly identifying the controller at the time it probes your hardware. This prevents it from loading the correct module during the first phase of the installation. It might be useful for you to know that the 10.1 ncurses based manual installation system provides a convenient way to both 'view loaded' and 'manually load' supported modules for various categories like 'network cards', 'filesystems', etc. If you know which module the Enterprise installer is loading, you can manually load the same one *before* proceeding with the 10.1 (manual) installation. <snip>
The problem is, when I load the LSI Logic MegaRAID SAS module, it is rejected, saying:
Module not supported by Novell, setting U Taint flag.
If you follow up with 'lsmod' you might be surprised to find the module actually *has* been loaded. My understanding is that the taint flag is a legal/technical warning indicating the module isn't supported by Novell (i.e. no installation support) because it is proprietary, distributed in binary form, hence not GPL. If I'm wrong on this point, I'm sure I'll be straightened out here in short order! ;-) hth & regards, Carl
![](https://seccdn.libravatar.org/avatar/5e70f769092f3372f14b4f2df58a17e1.jpg?s=120&d=mm&r=g)
<snip>
The problem is, when I load the LSI Logic MegaRAID SAS module, it is rejected, saying:
Module not supported by Novell, setting U Taint flag.
If you follow up with 'lsmod' you might be surprised to find the module actually *has* been loaded. My understanding is that the taint flag is a legal/technical warning indicating the module isn't supported by Novell (i.e. no installation support) because it is proprietary, distributed in binary form, hence not GPL. If I'm wrong on this point, I'm sure I'll be straightened out here in short order! ;-) Actually, as was very recently explained to me, the U taint flag means you won't get any support from Novell if you wind up with kernel
On 13/10/06 07:59, Carl Hartung wrote: problems, but that the kernel dudes *might* help. At least it isn't a T flag -- then your only recourse is the module author. The horse is already dead, I won't beat it much longer -- but it sure is discomforting to know that, sooner or later (probably sooner), one is going to have to consult an "approved" hardware database before making any hardware purchases for use with Linux -- imagine the heartbreak when somewhere, some kid with a penguin fetish gets a brand-new state-of-the-art system just for graduating from high school, only to find out that he can only get support for <the OS from Redmond>, all because the motherboard has some essential features that have no open-source drivers at all.
![](https://seccdn.libravatar.org/avatar/7574aaee71d8971a36f4283a7cad6b2c.jpg?s=120&d=mm&r=g)
* Darryl Gregorash <raven@accesscomm.ca> [10-14-06 08:22]:
The horse is already dead, I won't beat it much longer -- but it sure is discomforting to know that, sooner or later (probably sooner), one is going to have to consult an "approved" hardware database before making any hardware purchases for use with Linux -- imagine the heartbreak when somewhere, some kid with a penguin fetish gets a brand-new state-of-the-art system just for graduating from high school, only to find out that he can only get support for <the OS from Redmond>, all because the motherboard has some essential features that have no open-source drivers at all.
Good Morning. Good to see you awake. Pray tell, how does your statement differ from linux history. THE single biggest detriment to linux is and has been fender support, ie: drivers, and is better today than yesterday. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2
![](https://seccdn.libravatar.org/avatar/5e70f769092f3372f14b4f2df58a17e1.jpg?s=120&d=mm&r=g)
On 14/10/06 06:28, Patrick Shanahan wrote:
* Darryl Gregorash <raven@accesscomm.ca> [10-14-06 08:22]:
The horse is already dead, I won't beat it much longer -- but it sure is discomforting to know that, sooner or later (probably sooner), one is going to have to consult an "approved" hardware database before making any hardware purchases for use with Linux -- imagine the heartbreak when somewhere, some kid with a penguin fetish gets a brand-new state-of-the-art system just for graduating from high school, only to find out that he can only get support for <the OS from Redmond>, all because the motherboard has some essential features that have no open-source drivers at all.
Good Morning. Good to see you awake. Pray tell, how does your statement differ from linux history. THE single biggest detriment to linux is and has been fender support, ie: drivers, and is better today than yesterday.
Good morning to you, too, Patrick. I don't know about the Linux history, I didn't follow this stuff at all much before now. I didn't begin to bother with all this crap, until suddenly I began to get a U taint flag with the driver for the SBAwe32 card I had installed. It was nice to learn that Novell doesn't support only about the largest manufacturer of soundcards on the planet -- I first thought it was because the card is so old, Novell was trying to tell me it is time to replace it. Well, it was time for that, it only does 2.0 sound, but that is hardly relevant to the whole issue of tainting the kernel. Yes, the situation is much improved from yesterday, as I have learned -- and soon it will be worse than yesterday, because no one is going to force <insert any hardware manufacturer's name here> to release his source code, simply because some whining little brats have got it into their heads that they are going to try. If Linus et al wish to play the same game as Billy, it would help if Linus had an installed base approaching at least one-quarter what Billy has got. Well, he doesn't, and never will at this rate, because people expect support for their OS from the people who wrote the OS(*). The more probable response from most manufacturers will be "well, if this is your attitude, we really do not need to release *any* Linux drivers at all, because you folks are less than 5% of our customer base anyway". How much might things be different, if Linus and company try a few carrots instead -- "this problem has been reported to us, it is apparently with your driver, we got all this information out of the person who reported it to us, it looks like it might be a problem with <whatever>, if we can assist further, please contact <whoever>." Instead, we get wifi drivers for (I forget the brand) that aren't even included in the distribution, and anyway the cards aren't even recognized at install time, so anyone with one of those and nothing else has to fire up <the OS from Redmond> to get on the Internet to find them, *if* he even knows how to properly phrase his search pattern so that Google will return more useful than useless results (assuming he even knows how to use Google properly in the first place). Scratch one from the Linux installed base, this guy isn't coming back. (*) Which, in the mindset of the vast majority of the populace, for Linux distros equates to the people whose name is on the box -- most people don't know, and don't care, about all the workings of the inner sanctum of the kernel or of KDE. If they buy SuSELinux, they expect Novell to support them, because Novell's name is on the box.
![](https://seccdn.libravatar.org/avatar/7574aaee71d8971a36f4283a7cad6b2c.jpg?s=120&d=mm&r=g)
* Darryl Gregorash <raven@accesscomm.ca> [10-14-06 09:07]:
I don't know about the Linux history, I didn't follow this stuff at all much before now.
:^)
I didn't begin to bother with all this crap, until suddenly I began to get a U taint flag with the driver for the SBAwe32 card I had installed. It was nice to learn that Novell doesn't support only about the largest manufacturer of soundcards on the planet -- I first thought it was because the card is so old, Novell was trying to tell me it is time to replace it. Well, it was time for that, it only does 2.0 sound, but that is hardly relevant to the whole issue of tainting the kernel.
It is not that Novell doesn't support the card, and where is this comming from? It's that the hardware manufacturer doesn't provide drivers/support for linux where it does for windoz. Why is it that you surmise it Novell's responsibility to provide drivers/support that is not available from the hardware manufacturer?
Yes, the situation is much improved from yesterday, as I have learned -- and soon it will be worse than yesterday, because no one is going to force <insert any hardware manufacturer's name here> to release his source code, simply because some whining little brats have got it into their heads that they are going to try.
there, there... attitude not correct...
If Linus et al wish to play the same game as Billy,
You have completely lost track. This is *not* the plan, stan.
it would help if Linus had an installed base approaching at least one-quarter what Billy has got. Well, he doesn't, and never will at this rate, because people expect support for their OS from the people who wrote the OS(*). The more probable response from most manufacturers will be "well, if this is your attitude, we really do not need to release *any* Linux drivers at all, because you folks are less than 5% of our customer base anyway".
You are just ranting. remainder deleted. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2
![](https://seccdn.libravatar.org/avatar/5e70f769092f3372f14b4f2df58a17e1.jpg?s=120&d=mm&r=g)
On 14/10/06 08:32, Patrick Shanahan wrote:
* Darryl Gregorash <raven@accesscomm.ca> [10-14-06 09:07]:
I don't know about the Linux history, I didn't follow this stuff at all much before now.
:^)
I didn't begin to bother with all this crap, until suddenly I began to get a U taint flag with the driver for the SBAwe32 card I had installed. It was nice to learn that Novell doesn't support only about the largest manufacturer of soundcards on the planet -- I first thought it was because the card is so old, Novell was trying to tell me it is time to replace it. Well, it was time for that, it only does 2.0 sound, but that is hardly relevant to the whole issue of tainting the kernel.
It is not that Novell doesn't support the card, and where is this comming from? It's that the hardware manufacturer doesn't provide drivers/support for linux where it does for windoz. Why is it that you surmise it Novell's responsibility to provide drivers/support that is not available from the hardware manufacturer?
Well, someone had to write the driver that caused the kernel to be marked as tainted -- if it wasn't Novell (and no, I do not think it is their responsibility), and it was not the kernel writers (it certainly isn't *their* responsibility), then surely it was the manufacturer, yes? Support? Egad, it would suffice that the Linux community could do something more than say "find it yourself, it isn't our problem" -- because it certainly becomes a problem when that user dumps Linux, and moreover turns his 10 best friends away from Linux with a quick "you can't get support for anything" -- perception is all, and it doesn't matter what the truth actually is. So the rest of your reply becomes a total non sequitur -- and I stand by my "whining little brats" comment.
![](https://seccdn.libravatar.org/avatar/7574aaee71d8971a36f4283a7cad6b2c.jpg?s=120&d=mm&r=g)
* Darryl Gregorash <raven@accesscomm.ca> [10-14-06 13:22]:
Well, someone had to write the driver that caused the kernel to be marked as tainted -- if it wasn't Novell (and no, I do not think it is their responsibility), and it was not the kernel writers (it certainly isn't *their* responsibility), then surely it was the manufacturer, yes? Support? Egad, it would suffice that the Linux community could do something more than say "find it yourself, it isn't our problem" -- because it certainly becomes a problem when that user dumps Linux, and moreover turns his 10 best friends away from Linux with a quick "you can't get support for anything" -- perception is all, and it doesn't matter what the truth actually is.
So the rest of your reply becomes a total non sequitur -- and I stand by my "whining little brats" comment.
No, linux is a *community* project. You need to become involved and, instead of laying blame, contact the hardware vendors and manufacturers and advise that you will not be buying and/or using their products since they fail to provide the necessary drivers for you to utilize your operating system of choice. The vendors and manufacturers make their money from sales. If they do not know that they are loosing sales and the number of lost sales, they will do nothing. so don't be the "whinning little brats", at least not to the people who *are* supporting you. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2
![](https://seccdn.libravatar.org/avatar/11b4b3cf016b1d6a62454324eaaacc59.jpg?s=120&d=mm&r=g)
On Saturday 14 October 2006 13:23, Patrick Shanahan wrote:
The vendors and manufacturers make their money from sales. If they do not know that they are loosing sales and the number of lost sales, they will do nothing.
This is what matters. In hardware market vendors don't want to loose any percent of market share, so they will support even if that means 1% more. Linux community mistake is not to make list of contact information for major vendors. Not the common customer support contact info, as they don't count number of requests, but somebody that is looking where to expand market and boost sales. -- Regards, Rajko M.
![](https://seccdn.libravatar.org/avatar/72ee3b9e0735cf98a1e936a90fc087ed.jpg?s=120&d=mm&r=g)
On Sat, 2006-10-14 at 10:32 -0400, Patrick Shanahan wrote:
* Darryl Gregorash <raven@accesscomm.ca> [10-14-06 09:07]:
I don't know about the Linux history, I didn't follow this stuff at all much before now.
:^)
I didn't begin to bother with all this crap, until suddenly I began to get a U taint flag with the driver for the SBAwe32 card I had installed. It was nice to learn that Novell doesn't support only about the largest manufacturer of soundcards on the planet -- I first thought it was because the card is so old, Novell was trying to tell me it is time to replace it. Well, it was time for that, it only does 2.0 sound, but that is hardly relevant to the whole issue of tainting the kernel.
It is not that Novell doesn't support the card,
Actually it is. The "Unsupported" taint flag is something SUSE started and SUSE/Novell continues to use. It means there is no support for that particular driver. Please note that other Creative cards are supported, e.g.
/sbin//modinfo snd-emu10k1|grep supported supported: yes
(snd-emu10k1 is the driver for Soundblaster Live, Audigy etc. More modern cards)
![](https://seccdn.libravatar.org/avatar/7574aaee71d8971a36f4283a7cad6b2c.jpg?s=120&d=mm&r=g)
* Anders Johansson <andjoh@rydsbo.net> [10-14-06 15:38]:
Actually it is. The "Unsupported" taint flag is something SUSE started and SUSE/Novell continues to use. It means there is no support for that particular driver.
but *unsupported* means one of two things, closed source driver or no driver....
Please note that other Creative cards are supported, e.g.
yes, I have one on mb, multi-channel and it works well. -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org HOG # US1244711 Photo Album: http://wahoo.no-ip.org/gallery2
![](https://seccdn.libravatar.org/avatar/72ee3b9e0735cf98a1e936a90fc087ed.jpg?s=120&d=mm&r=g)
On Sat, 2006-10-14 at 17:18 -0400, Patrick Shanahan wrote:
* Anders Johansson <andjoh@rydsbo.net> [10-14-06 15:38]:
Actually it is. The "Unsupported" taint flag is something SUSE started and SUSE/Novell continues to use. It means there is no support for that particular driver.
but *unsupported* means one of two things, closed source driver or no driver....
Not at all. "Unsupported" is something you get from a driver, you wouldn't see "U" if there was no driver at all. "Unsupported" could come from a non-GPL driver as well, but then it would be accompanied by a "G" for non-GPL.
Please note that other Creative cards are supported, e.g.
yes, I have one on mb, multi-channel and it works well.
"Supported" doesn't mean "it works", it means "we'll do our best to fix it if it doesn't". There are more drivers that work but that aren't supported
![](https://seccdn.libravatar.org/avatar/696255f62c3cefe5d517f913058b70bc.jpg?s=120&d=mm&r=g)
On Saturday 14 October 2006 08:20, Darryl Gregorash wrote:
Actually, as was very recently explained to me, the U taint flag means you won't get any support from Novell if you wind up with kernel problems, but that the kernel dudes *might* help. At least it isn't a T flag -- then your only recourse is the module author.
Thanks for the clarification, Darryl!
The horse is already dead, I won't beat it much longer -- but it sure is discomforting to know that, sooner or later (probably sooner), one is going to have to consult an "approved" hardware database before making any hardware purchases for use with Linux
I agree :-( And I note your choice of "approved" vs. "certified". <sigh!>
![](https://seccdn.libravatar.org/avatar/d055a2899adeaeb6967fdde6994b0079.jpg?s=120&d=mm&r=g)
Philip Mötteli a écrit :
Hi,
I try to install SuSE Linux 10.1 on a Dell 2950. The problem is, that the installer just doesn't see the hard-drives. This is a system with two dual-cores and 4GB of RAM. The heart of the problem is perhaps the LSILogic SAS 1068-IT (v6.6.0.2) controller (there's no RAID hardware). By default, the installer loads the following modules:
Fusion MPT SCSI Host Driver Fusion MPT SAS Host Driver
But there are other drivers, which seemed to me more appropriate:
LSI Logic MegaRAID LSI Logic MegaRAID SAS LSI Logic Management Module
The problem is, when I load the LSI Logic MegaRAID SAS module, it is rejected, saying:
Module not supported by Novell, setting U Taint flag.
Interestingly, with the delivered SuSE Enterprise Linux CD-ROM, I have no problems installing, but I also have only one month of updates.
Has anybody already had that problem? Phil
--Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
I've such a problem each time I build a new computer with the latest hardware. In year 2001 Linux was unable to manage my LSIlogic Ultra SCSI card, and the problem was solved with the next release. In year 2006 with SuSE 10.0 I'd to manually choose the necessary drivers for my chipset and Sata controllers and with 10.1 everything was fine. So you can expect a solution with 10.2 release. Most manufacturers don't support Linux and often it is necessary to do reverse engineering in order to define the drivers. Michel.
participants (7)
-
Anders Johansson
-
Carl Hartung
-
Catimimi
-
Darryl Gregorash
-
Patrick Shanahan
-
Philip Mötteli
-
Rajko M