[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: 2.4.18 pcmcia troubles



On Wed, Aug 06, 2003 at 05:23:56AM +0000, Andrew McGuinness wrote:
> ScruLoose wrote:

> >Hm.  I don't *think* that worked, but I'm not entirely sure.
> >The i82092 module loads, but subsequent attempts to modprobe
> >pcnet_cs result in the familiar error:
> 
> Hmmm. I would have expected loading the module to fail.  Was there any 
> output at all on "modprobe i82092".  Did anything appear in 
> /var/log/syslog?

Yep, "modprobe i82092" produced this output:

Linux Kernel Card Services 3.1.22
  options: [pci] [cardbus] [pm]
Warning: loading /lib/modules/2.4.18-1-586tsc/kernel/drivers/pcmcia/i82092.o will taint the kernel: no licence

The first two lines of that show up in syslog too, but the warning
doesn't.

i82092 shows up in lsmod as being "unused"...

> >Does this look like the module doesn't support my hardware, or is there
> >some sort of vital step I'm leaving out in between there?
> >
> Not sure.

My situation right now stands with pcmcia_core and i82092 listed in
lsmod; modprobe-ing pcnet_cs fails after complaining about "ds: no
socket drivers loaded!"; and modprobe-ing ds itself fails with that
same complaint.

And a note: if I boot into the old 2.2.20 kernel with "skip i82365" in
discover's config, then it boots up with no pcmcia modules loaded
except pcmcia_core.  
I can then modprobe i82365, and then I can modprobe pcnet_cs, and it
succeeds, and the link light comes on on my dongle.
So I *think* I can conclude from this that the behaviour under 2.4
indicates a "module doesn't support the hardware" kind of problem...
does that sound reasonable?

> >Now, when you say "refers to it"...  Is this information I should know
> >where to look for myself?
> >
> Not really; I was digging around in kernel sources.

Um.  I'll hazard a guess that it's a bit early for me to get into that.
;-)

> >That does lead me to another question:  would it be any significant
> >security risk (or instability risk) to use a pre-release kernel on a
> >firewall?
> >
> I wouldn't think it's a security risk.  There is some instability risk; 
> in my experience the pre kernels are usually stable enough, but there's 
> no guarantee.  If something breaks in a released stable kernel, various 
> people will be very embarassed, but in a pre-release they'll just look 
> all innocent and say "but it *was* a test kernel" :-)

Fair enough. That sounds a lot like what I would have guessed, but it's
nice to have a second opinion.

> In any case, 2.4.22rc1 has *just* appeared on the mirrors. (rc=release 
> candidate).  You can interpret that as "it's stable enough now, let's 
> go", or that a final release is days not weeks away and you might as 
> well wait, depending on taste...

Hehe!  A week ago I would never have dreamed I'd be sitting here right
now  trying to decide my personal 'taste' on compiling a prerelease 
kernel...
I hope you can forgive a little bit of newbie excitement on this.  ;-)
I think maybe I will take a crack at 2.4.22rc1 (although knowing me
the release version will be out when I actually get around to it)

Sweet timing either way, though.
-- 
,-------------------------------------------------------------------------.
>   -ScruLoose-   |       What makes a person so poisonous righteous      <
>  Please do not  |  That they'd think less of anyone who just disagreed? <
> reply off-list. |                     - Moxy Fruvous                    <
`-------------------------------------------------------------------------'

Attachment: pgp8bPvjOhNcJ.pgp
Description: PGP signature


Reply to: