SEO

I had to go back to Slack 12.0, since my USB disk became unusable. I got a
never-ending stream of messages to the effect that a new address had been
assigned to a new device, presumably my USB disk. The addresses ran
sequentially up to 127, then back at 1.
It is clearly a 12.1 issue, now I am back at 12.0, and everything is OK.
Does anyone know more about this phenomenon?

On Sat, 10 May 2008 17:03:34 +0200, jjg wrote:

> I had to go back to Slack 12.0, since my USB disk became unusable. I got
> a never-ending stream of messages to the effect that a new address had
> been assigned to a new device, presumably my USB disk. The addresses ran
> sequentially up to 127, then back at 1. It is clearly a 12.1 issue, now
> I am back at 12.0, and everything is OK. Does anyone know more about
> this phenomenon?

are you by any chance using KDE and relying on some
form of automounting and/or udev+KDE magic mounting?

did you by some hazard upgraded your 12.0 to 12.1 and
not read the few files that were floating around with
their names in CAPITAL chars and somewhat incentive
'drink me/eat me' naming scheme?

if so that's not "clearly a 12.1 issue" :-)
if not your case, please develop the missing part
in your post, that one when you describe how you
installed 12.1 and the rest of the WWWWWH questions,
oh dear...

loki harfagr wrote:

> On Sat, 10 May 2008 17:03:34 +0200, jjg wrote:
>
>> I had to go back to Slack 12.0, since my USB disk became unusable. I got
>> a never-ending stream of messages to the effect that a new address had
>> been assigned to a new device, presumably my USB disk. The addresses ran
>> sequentially up to 127, then back at 1. It is clearly a 12.1 issue, now
>> I am back at 12.0, and everything is OK. Does anyone know more about
>> this phenomenon?
>
> are you by any chance using KDE and relying on some
> form of automounting and/or udev+KDE magic mounting?

I _am_ using KDE, but no automounting.

>
> did you by some hazard upgraded your 12.0 to 12.1 and

No, I did _not_ upgrade. I never do. I just did a fresh install of 12.1,
wiping out 12.0 (except my /home volume). However, I had a backup of /etc
on the USB drive (my next step would have been to recover the essentials
from it), and I could not get at it.

> not read the few files that were floating around with
> their names in CAPITAL chars and somewhat incentive
> 'drink me/eat me' naming scheme?

Well, I saw a few README files, but since I did not plan an upgrade, I saw
no point in reading them more than cursorily. I did not notice anything
that was not already present in 12.0. Was I wrong? Which one(s) do you
mean?

>
> if so that's not "clearly a 12.1 issue" :-)

Well, all I saw was that my USB drive did not do anything sensible with the
new kernel, until I had reinstated the old one. I also saw that quite a few
drivers were not present, or had the wrong format, so I could not load them
afterwards.

> if not your case, please develop the missing part
> in your post, that one when you describe how you
> installed 12.1 and the rest of the WWWWWH questions,
> oh dear...

Well, perhaps you could be a bit more specific...

jjg <jjge@xs4all.nl> wrote:
> I had to go back to Slack 12.0, since my USB disk became unusable. I got a
> never-ending stream of messages to the effect that a new address had been
> assigned to a new device, presumably my USB disk. The addresses ran
> sequentially up to 127, then back at 1.
> It is clearly a 12.1 issue, now I am back at 12.0, and everything is OK.
> Does anyone know more about this phenomenon?

jjg,

I've seen the same thing happen in another version of Slackware. I
think it was 10.2. I boot Slackware using a homemade grub boot CD-R
(kernel on the CD-R). I do this so I can carry my Linux system around
with me on a hard drive in a USB enclosure.

I end up booting from a variety of machines, since I sometimes try to
get some work done at my wife's place of work while I'm passing the
time (we car pool). I ask her "What machine can I boot from?", and she
directs me to an office cubicle that she thinks will be unoccupied for
a while.

In doing this I've found some machines that will sometimes show the
error you described. I was able to get around the error by using a
different USB connection on the machine. This really surprised me, but
it worked. That is not much of a solution, but it might work for you
in the short run.

Another suggestion is that you should try connecting your USB drive to
a completely different machine. This is a lot of work I know if you
have only data on your USB drive and not the whole OS. (You would have
to install Slackware 12.1 on the other machine too--or put Slackware
12.1 on the USB drive and use a boot disk like what I do.) My
experience showed that trying different machines can make a difference
with this problem.

BTW, I've only found a few machines that give me the error you
described.

-Joe

On Sat, 10 May 2008 21:05:24 +0200, jjg wrote:

> loki harfagr wrote:
>
>> On Sat, 10 May 2008 17:03:34 +0200, jjg wrote:
>>
>>> I had to go back to Slack 12.0, since my USB disk became unusable. I
>>> got a never-ending stream of messages to the effect that a new address
>>> had been assigned to a new device, presumably my USB disk. The
>>> addresses ran sequentially up to 127, then back at 1. It is clearly a
>>> 12.1 issue, now I am back at 12.0, and everything is OK. Does anyone
>>> know more about this phenomenon?
>>
>> are you by any chance using KDE and relying on some form of
>> automounting and/or udev+KDE magic mounting?
>
> I _am_ using KDE, but no automounting.

OK, I asked just to be sure there wasn't a supplemental
layer that could have had side effects.

>> did you by some hazard upgraded your 12.0 to 12.1 and
>
> No, I did _not_ upgrade. I never do. I just did a fresh install of 12.1,
> wiping out 12.0 (except my /home volume). However, I had a backup of
> /etc on the USB drive (my next step would have been to recover the
> essentials from it), and I could not get at it.

OK, if /tmp and swap also were fresh I don't imagine /home files
would have side effects with the problem you describe.
(I suppose there isn't any stuff like calls to rescan-scsi-bus or
akin in some /home files)

>> not read the few files that were floating around with their names in
>> CAPITAL chars and somewhat incentive 'drink me/eat me' naming scheme?
>
> Well, I saw a few README files, but since I did not plan an upgrade, I
> saw no point in reading them more than cursorily. I did not notice
> anything that was not already present in 12.0. Was I wrong? Which one(s)
> do you mean?

I only had a faint idea somehow your initrd was b0rken, there were so
many turmoils between udev kernel and hotplug between their versions
off 12.0 to their current line that most anything could happen, as
I use to follow -current and also test a few stuff off the main slack
I frequently had bullets in my toes especially regarding USB stuff
like wacom tablets, multicard readers thumbdisks and harddisks,
curiously I hadn't have the slightest problem with audio hardware
like it used to be the case in the previous years )

>> if so that's not "clearly a 12.1 issue" :-)
>
> Well, all I saw was that my USB drive did not do anything sensible with
> the new kernel, until I had reinstated the old one. I also saw that
> quite a few drivers were not present, or had the wrong format, so I
> could not load them afterwards.

Now that's very strange and it certainly *should not* happen on a
fresh installation, problem is it brings me back to the idea there's
a problem in your kernel/initrd/parms.
Any chance you could test again and note the 'missing drivers' and
precise the 'wrong format'? (what do you mean by 'wrong format?
some msg like "module is xxx-smp and kernel is not-smp" ?
if so that's a problem when chosing kernel and making initrd
and lilo)

>
>> if not your case, please develop the missing part in your post, that
>> one when you describe how you installed 12.1 and the rest of the WWWWWH
>> questions, oh dear...
>
> Well, perhaps you could be a bit more specific...

specific on what? You're the one who can se your screen, from here
all I can do is to try having an idea out of the blue (though I'm
not RM and have not a fast BSOD launching supercomputer at hand ;-)

anyway, your problem looks strange, maybe it's something like
Joseph Rosevear said, but as long as you can't test again and give
information we'll never know :-)

Joseph Rosevear wrote:
>
>jjg <jjge@xs4all.nl> wrote:
>
>> I had to go back to Slack 12.0, since my USB disk became unusable. I got a
>> never-ending stream of messages to the effect that a new address had been
>> assigned to a new device, presumably my USB disk. The addresses ran
>> sequentially up to 127, then back at 1.
>
>I've seen the same thing happen in another version of Slackware. I
>think it was 10.2. I boot Slackware using a homemade grub boot CD-R
>(kernel on the CD-R). I do this so I can carry my Linux system around
>with me on a hard drive in a USB enclosure.
>
>I end up booting from a variety of machines, since I sometimes try to
>get some work done at my wife's place of work while I'm passing the
>time (we car pool). I ask her "What machine can I boot from?", and she
>directs me to an office cubicle that she thinks will be unoccupied for
>a while.
>
>In doing this I've found some machines that will sometimes show the
>error you described. I was able to get around the error by using a
>different USB connection on the machine. This really surprised me, but
>it worked. That is not much of a solution, but it might work for you
>in the short run.

>I've only found a few machines that give me the error you
>described.

I would like to see your method written up and added to the
Revised Slackware Book at http://www.slackbook.org/. It looks
like a really useful thing to be able to do.

BTW, would it work to make the USB disk bootable, or do you
have to boot from a CD?

--
Sam I Am says:
I do NOT like Green Eggs and Spam!!

On Sat, 10 May 2008 23:33:47 +0000, Joseph Rosevear wrote:

> jjg <jjge@xs4all.nl> wrote:
>> I had to go back to Slack 12.0, since my USB disk became unusable. I
>> got a never-ending stream of messages to the effect that a new address
>> had been assigned to a new device, presumably my USB disk. The
>> addresses ran sequentially up to 127, then back at 1. It is clearly a
>> 12.1 issue, now I am back at 12.0, and everything is OK. Does anyone
>> know more about this phenomenon?
....
> In doing this I've found some machines that will sometimes show the
> error you described. I was able to get around the error by using a
> different USB connection on the machine. This really surprised me, but
> it worked. That is not much of a solution, but it might work for you in
> the short run.

On this point specifically I suspect that's related to the
actual fact that not all USB connectors are equal, on my
different machines I have, thru USB, a multicard reader
and a WD external hard disk that will refuse to work properly
on some connectors with a prompt in /var/log/mess. about
the plug "not being a full hi speed" though they're on the
motherboard cinched to the same USB main as the next tried
connector that'll work.

<posted & mailed>

loki harfagr wrote:
[..]
> OK, if /tmp and swap also were fresh I don't imagine /home files
> would have side effects with the problem you describe.
> (I suppose there isn't any stuff like calls to rescan-scsi-bus or
> akin in some /home files)

No, certainly not. (my wife and I are the only users, so I should know)

>> Well, all I saw was that my USB drive did not do anything sensible with
>> the new kernel, until I had reinstated the old one. I also saw that
>> quite a few drivers were not present, or had the wrong format, so I
>> could not load them afterwards.
>
> Now that's very strange and it certainly *should not* happen on a
> fresh installation, problem is it brings me back to the idea there's
> a problem in your kernel/initrd/parms.

Well, at least I did not supply any params on bootup. I just press RET when
asked.

> Any chance you could test again and note the 'missing drivers' and
> precise the 'wrong format'? (what do you mean by 'wrong format?
> some msg like "module is xxx-smp and kernel is not-smp" ?
> if so that's a problem when chosing kernel and making initrd
> and lilo)
>

Just from memory... I think it was the usbhid.ko driver for the USB mouse
(but that did not work under 12.0 either). As far as the USB disk is
concerned, I just checked rc.modules, and I tried to modprobe a few modules
mentioned (and commented out) in it, and most of them were not present. I
would assume that modules listed in rc.modules would normally exist, but
perhaps I am wrong?

And of course I could build my own kernel and modules, and maybe I will, but
the problem is that I am temporarily without computer when I am rebuilding,
and I don't want to stay offline too long.

And therefore, I will have to take a break before I decide to try
again... :-(

> Joseph Rosevear said, but as long as you can't test again and give
> information we'll never know :-)

PS just as a sideline: I do not assume that my USB drive is 100% OK. In
fact, I have quite strong suspicions. Under 12.0 it occasionally "hangs",
and I have a subtree that cannot be removed. So something may be wrong. But
back under 12.0 it became alive and usable again, so for the time being I
assume it is something in 12.1, even though it might be triggered by a
problem in my drive.

loki harfagr wrote:

> On Sat, 10 May 2008 23:33:47 +0000, Joseph Rosevear wrote:
>
>> jjg <jjge@xs4all.nl> wrote:
>>> I had to go back to Slack 12.0, since my USB disk became unusable. I
>>> got a never-ending stream of messages to the effect that a new address
>>> had been assigned to a new device, presumably my USB disk. The
>>> addresses ran sequentially up to 127, then back at 1. It is clearly a
>>> 12.1 issue, now I am back at 12.0, and everything is OK. Does anyone
>>> know more about this phenomenon?
> ...
>> In doing this I've found some machines that will sometimes show the
>> error you described. I was able to get around the error by using a
>> different USB connection on the machine. This really surprised me, but
>> it worked. That is not much of a solution, but it might work for you in
>> the short run.
>
> On this point specifically I suspect that's related to the
> actual fact that not all USB connectors are equal, on my
> different machines I have, thru USB, a multicard reader
> and a WD external hard disk that will refuse to work properly
> on some connectors with a prompt in /var/log/mess. about
> the plug "not being a full hi speed" though they're on the
> motherboard cinched to the same USB main as the next tried
> connector that'll work.

I have heard this as well, and, although I find it difficult to believe
(well, hardware design is so straightforward, isn't it?), I have tried it
occasionally. But in this particular case, the USB port does not seem to
matter.

[...]
jjg wrote:
> Just from memory... I think it was the usbhid.ko driver for the USB mouse
> (but that did not work under 12.0 either).

Under 12.0 I get:

modprobe usbhid
FATAL: Error inserting usbhid
(/lib/modules/2.6.21.5-smp/kernel/drivers/usb/input/usbhid.ko): Invalid
module format

and under 12.1 I remember having seen the same...