Re: 3c905B IRQ=12
In a message dated 6/16/03 05:33:26 AM Eastern Daylight Time,
iisnt2002@yahoo.com writes:
>
> Please have a look And Let me known How can I Fix that
> problem.
>
> /lib/modules/2.2.20/net/3c509.o: init_module: Device
> or resource busy
> Hint: insmod errors can be caused by incorrect module
> parameters, including inva
> lid IO or IRQ parameters
> /lib/modules/2.2.20/net/3c509.o: insmod
> /lib/modules/2.2.20/net/3c509.o failed
> /lib/modules/2.2.20/net/3c509.o: insmod 3c509 failed
>
Debian Linux(s) user(installer toget it going) problems are not much that
there's actaully anything wrong at all, with anything, but that Linux installers,
of OS or devices or code packages or just packages that run on Linux are most
often, of a single nature. They are are not so much an OS issue, but that as
carefully slow as Debian Linux release/adopti0on elements are, Linux changes
are ongoing in virtually every facet of the gem. Modules are a prime example,
numbers.numbers.last-number+1more are different code than
number+/-priors.numbers+/-priors/numbers-last-or-prior;
they constitute cascades of upgrades/fixes. Cascades of cascades. Focus is
the point, on which branch has which focus needed on which of its branches
branch choice, that branch's branches choice, down the many forks in each of the
many roads of the trip, to which leaf is to be sought, despite the leaf's own
site on one one of the group of twigs it budded with of the branch it draws
water from.
Thus, as configuration value datum, in tables, in files, kept somewhere, are
different. That only one or few elements in a/some table(s,
related ones;
often needed "matches"
for comparison match
to choose which row each table-of-that-part/thing
of two or more tables,
to use row's after-the-match values,
one more or more
)
The point in you problem example is, like the problem almost always occurring
in addition of either hardware of software, code run, as in it becoming
manifested as outer code when it is being run, calls functions of other code
packages (libraries; packages of functions that more than one portion of code or
expecially packages of such other code (when that runs in turn) that are run
under other code packages, can have the reliability of all using the same one of
that thing/function. Thus, such code of functions (and defs and include calls
to libraries sub to the one at hand) in a version of a library, can so often
provide at least one incompatible table column value used by other function's
functions called in outer code, as parameters or range of values that are
permitted, to pass some "test at the top" comparison, which faithfully trigger
error message from then running inner code, citing element of outer code it choked
on.
> /lib/modules/2.2.20/net/3c509.o: init_module: Device
> or resource busy
Is sent out as:
/lib/modules/2.2.20/net/3c509.o: init_module: Device or resource
busy
the particular library, a module library file of code of
the library directory,
thus directory is a collection of libraries
modules; modules are a sub set collection of
modules-used-libraries
(cited at top, module code, code of one or a type-of things,
and influence replys produced,
especially error crys about input they can't match
to their inner shared upward defs-followed
(type,range) code line cited value
or table/array-of-all-those-things row
version 2.2.20 (kernal uses modules now,
to change what loaded runtime kernal actually consists of,
like when devices or types of them used,
and especially to result in
much smaller/faster kernal of no code
of things not to be used (many thousands)
so collection is of that kernal's version#
net; networking devices sub set of modules
as there are now so damn many of such things
and now so many of them in individual files of
subsets of themselves,
like mfgrs. older vs. newer version cascades, of
mfgrs. stuff/things,
drivers, by their own version differences
of code I/O element(s) consequences
3c509whatever; 3COM' older, world's greatest number of
NICs placed, so it got great number of customer service
(PC makers customers, each order BIG) modifications
(pre-standards needs notice)
so big a seller LOTS of other brands ended up compatible
(BY DOS usage)
So many 3C509s were out there, quite early a Linux boot saw it
and hooked it right up, saying so in cascade of boot messages
Regarding
Device or resource busy
is module's actual error message, cascaded UP thru preceeding
elements of error messageS, as it enters and is combine prefixed
by outer calling code cascade (UP/OUT)
> Hint: insmod errors can be caused by incorrect module
> parameters, including inva
> lid IO or IRQ parameters
> /lib/modules/2.2.20/net/3c509.o: insmod
> /lib/modules/2.2.20/net/3c509.o failed
> /lib/modules/2.2.20/net/3c509.o: insmod 3c509 failed
"hint:" is module writers report of author's getting banged on often
by PC elements code bangers mistakes (Corp. home brew support
contract eternal additional income "fix"?"es")
Note: such authoring is probably part of module code that is
still within the module
The rest of the lines are from module-effectively-is-driver:
if can't anything-at-all-imperfect, any module block
causes outer block return to it
for comparion of return=inner_returns[returning_block[it's_return]]
begins(match(error-x,error-y,error-z))
then:
try test block X test results
if-then ___> try test block X.X results
same, pretty much, just again, next branch
trying to be helpful, that's Linux people out there
then out-msg = error-msgs-text-array(one2sendout2OS)
So, in newer net module elements file's content, 3C509, and many more things
get set aside of net module library function set's tables etc. So modern
Fortune-all-of-em kernal of great impression and results, can simply be that.
You can use Debian site package manager pages to track which pkg. component
you got for network stuff. Then web trace by searches for "3c509", which you
can even do from home page, but that can provide more branches to spend time
going down'em, to end up finding the lib/module/net file 3c509 is in,
sorted by reverse date, and then which net module (sorted reverse date too)
has 3c509 still in it, and download that version of net module that you need
instead, search for the modules dir your crying one is in, make a dir in that
dir named ".orig" to copy!(NOT delete/risk-loss-of-later-need-desperate-ism)
your file of the version of the module to .orig (non symboilc is better to keep
file's date same, as that can be revealing in problems. Link (on symbolic!,
true!) is same file, upper link file deletion makes no change to OS filesystem's
dir table having two sub table (filename,inode) entries for same inode, just
one less for same inode, which is what it uses as it's filename thing.
Thus, you can replace the net module file (executable very desired!, run
"type" followed by filename) with prior one that has 3c509 in it. Bang, you done,
for newer kernals(versions here too) that pick modules of net/any/all stuff
they actually will find and use. Older kernals need to be rebuilt (different
dir of source code of kernal, w/make file you must read first in case some
lines need to be commented in or out or comment switched, or vars set to what you
actually got there, probably very few & quite often none; if kernal build
"make" execute fails, then here is cause; make run takes several runs, ea. with
word following make command(file, executable, followed by param word(s) ), till
final just make, to get it to actually look around for what's in the box here
and now, and clean up lots of large pieces files left over, man on make covers
this clearly, but along with so damn many other options too (using man, NEVER
go down all those many branches in everything, it is protocoled to be
complete, so it's always way beyond the concern of current info search, but man's
beauty part is; what it doesn't mention!, is doen't get into or handle in any
way!, bc. it knows nothing! about anything! but it's! stuff, of which it knows
all!, and the slash key let's you type in a topics main word, to leap straight
to that word(s-all-that-one), and you can fast look in section it took you to
w/another /search, but it's damn concise by protocol too, as I just mentioned).
But this email forem is quite likely to get you a response to copy/paste thru
google & Debian to paragraphs telling you all the above, in summary. I just
have a habit of responding with the truth, the whole truth, and nothing but
all that huge tree of truth details. But you and so many others can just save
this as
"HowToSaveYourOwnUNIX/LINUX_ASS-WhenNobobyElseWillAndYouGottaDoItTheOldWay", and
"HowToGetit-Of-Unix-Linux-Code-Builders-Agreements-On-Building-It-_My_Part_A-Way-We-Can-Each-Do-JUST-Part-Of-THAT-Party-So-All-Of-It-Does-Actually
-Work-With-The-Other-Guys-Partys".
Sorta.
Regards,
Jim Cunningham
P.S.
The Bell Labs Pure (SUPER PHDS) Research guidelines of the Unix guys standard
procedures to meet demands of Gov. TOP Scientist examiner's progress reports,
demanded that particular organizational methodology, and thus evaluations
thru to delivery of grant granter's goals and expectations, to keep getting the
rest of the grant. Those organized demands were Gov. learned in wartime, World
War 2, and used to win the war, and of wartime expectation levels. Failure,
and it's a hearing you really can get a death sentence at, this is wartime.
Of course, the GOV learned that from Bell Labs and it's equals, at crank up of
the war effort, anyway. But after the war, AT&T at BELL realized GOV HAD
finally learned how to manage those guys that were too smart to even discuss the
work with, Organize'em. Not the work anymore, make that the problem of those
workers. Failure is a review which actually can get AT&T's loss of the money
they use to pay you, and you do get a job-death sentence.
P.P.S.
Thers's still a whole lot of 3C509s out there, by pass down, I got 3 at a
"Trade Center" for 3$ ea. several years ago, and there isn't a reason in this
world to replace them. They work just as good as the brand new clone ones, the
ethernet protocol has not changed, that you can get for a full $5 each at,
claimded to be wholesaler, quite big suppliers.
Reply to: