Re: dhcp, cable modems...
On Thu, Mar 01, 2001 at 06:27:22PM -0800, Jason Victor wrote: > I'm trying to configure my cable modem to work with > Linux. I'm using Optimum Online. I have a couple of > questions. First of all, is my ethernet card working > if i type dmesg and see something about "eth0: Xircom > CardBus." That is the NIC i have. second, i've been > trying to connect with dhcpcd, and can't figure it > out. i set up dhcpcd to work on eth0, and type ifup > eth0, and it says "ignoring unknown interface eth0." > i'm stuck. by the way, it sits at the starting dhcp: > thing for a long time. Run ifconfig -a to see if eth0 exists or not. If it does, you just need to get the right stuff to use it. If it doesn't, then you need to config your pcmcia and modules stuff to set it up. You might want to try doing things by hand, i.e. running pump -i eth0 (dhcpcd is obsolete; It's been removed from Debian because it is irreparably full of bugs and security holes. It still works, though, and it has a similar command line option. Note that pump will log everything it does to syslog. Take a look if anything goes wrong.) You say that dhcpcd hangs for a long time. Try running tcpdump while you run dhcpcd. (This may not work too well, since the interface will be down when you start, unless you assign a bogus address to it. Also, dhcpcd will take down the interface tcpdump is sniffing on...) See if you're getting any response. If pump (or dhcpcd) works, then your problem is in configuring the high level tools like ifup. I personally never bother with them on my desktop machines, and rewrite /etc/init.d/networking to run ifconfig, route, and ipchains the way I like. (I haven't finished hacking the pcnet-cs driver to support my Surecom EP-427X for my laptop :(. However, if you just want something simple that works, stick with ifup I guess. > i am using storm linux (basically debian 2.2) is there > any way to use the storm administration system to do > this? there is an option to set up eth0 with dhcp. i > need 2 give it a hostname... where do i get that? DHCP can pass a hostname to the DHCP server. [EMAIL PROTECTED] in Ontario uses this: they give you a host name to use when you sign up, and you have to supply that for their DHCP server to answer. Other providers, like Eastlink in Nova Scotia, don't require this, and just check the ethernet MAC address. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: dhcp, cable modems...
On Thu, Mar 01, 2001 at 11:40:51PM -0500, MaD dUCK wrote: > also sprach xsdg (on Thu, 01 Mar 2001 11:29:05PM -0500): > > > (dhcpcd is obsolete; It's been removed from Debian because it is > > do what? i missed something. oh, and dhcpcd is not the dhcp daemon, > but the dhcp client daemon, Yes. > and it's so much better than pump... I don't find it better than pump. Last I checked, it does dumb stuff like giving up after a while and leaving the interface down. If my cable modem provider is doing an upgrade so there is a temporary outage, dhcpcd would (after an hour or so) give up and bring down the network connection, leaving me disconnected even after my cable modem came back online. I don't know how pump would handle the situation, since I've stopped using dhcp except to obtain an IP config initially. My ISP doesn't capriciously change my IP address, so I statically config my interfaces. (Sure, this breaks things when the IP does actually change, but I want that, because in that case I need to change the DNS A record for llama.nslug.ns.ca so my email will come in, so I want large breakage right away that will catch my attention :) See http://lists.debian.org/debian-powerpc-0102/msg00097.html and http://lists.debian.org/debian-powerpc-0102/msg00126.html to find out what the current state of DHCP in Debian is. As you can see, there is a dhcp-client package. It might float your boat better than pump, if you don't like pump. It's interesting to note that even though apt-cache search dhcpcd only finds pump, (i.e. dhcpcd is not part of woody anymore), you can still download the dhcpcd package through packages.debian.org/dhcpcd, and it's on the mirrors. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: dhcp, cable modems...
On Fri, Mar 02, 2001 at 07:32:02AM -0500, Michael P. Soulier wrote: > On Fri, Mar 02, 2001 at 01:38:36AM -0400, Peter Cordes wrote: > > > It's interesting to note that even though apt-cache search dhcpcd only > > finds pump, (i.e. dhcpcd is not part of woody anymore), you can still > > download the dhcpcd package through packages.debian.org/dhcpcd, and it's on > > the mirrors. > > I've never tried pump, only dhcpcd, so can you answer two questions for > me? > > 1. Does pump permit you to specific a client ID? It lets you specify a host name, with -h. It doesn't say anything about using anything other than the ethernet MAC address as the client ID, though. > 2. Does pump give you a hook for running a script whenever the IP changes? Yup. from pump(8)'s description of the config file: script executable-filename When events occur in negotiation with the server, calls the given executable or script. Scripts are called when a lease is granted, when a renewal is negotiated, and when the interface is brought down and the address released. The scripts are called with two or three arguments, depending on the con- dition, as documented in the table above. > > If so, then I'll probably switch, if it's not maintained in woody anymore. You might have to switch to dhcp-client, which is dhcpcd renamed, AFAIK. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Can anyone get their hands on a Surecom EP-427X?
Hi,
Can anyone borrow a Surecom EP-427X fast ethernet PC card to try a quick
experiment on for me? I'm trying to get Linux to support this NIC (it's an
ne2000 clone, so it should require only a bit of hacking...) I want to see
whether the problem I'm seeing is my laptop's fault, or what.
I've never used any other pc cards with my laptop (an old used AST Ascentia
810N, which has a Cirrus Logic PD6720 pc card controller.). I don't know if
my laptop's pc card hardware even works totally right, since I got it second
hand. I was able to use a DOS boot disk to run the enabler and packet
driver from the disk that came with the NIC, and run a packet driver test
program to ping another computer on my LAN with the card, so obviously
it is possible to make it work.
This card is a 10/100baseT 16bit pc card NIC, from Surecom. Note that this
is not the same as the EP-427 or EP-427T. They are 10baseT only, and are
already supposed to be supported by linux.
I've got one of these cards, and the driver floppy includes a "linux"
directory, with a patched copy of pcmcia-cs.c, and 8390.c, and instructions
to use this with a 2.0.30 kernel! I've figured out some stuff about how it
works, and figured out that the change Surecom made was adding:
{ /* SURECOM EP-427X */ 0x01c8, 0x00, 0x00, 0x21, 0 }
to the hw_info array. Using the memory-cs driver, and dd /dev/mem0a, as
described in the PCMCIA-HOWTO, the 0x01c8 offset does indeed point to the
card's ethernet MAC address.
The problem is that pcnet_cs.o says:
pcnet_cs: GetFirstTuple: No more items, and refuses to load. Another weird
thing that happened is that the linux info from surecom says the right
config for the card is:
card "SURECOM EP-427X 10/100M PCMCIA Adapter"
version "PCMCIA", "100BASE"
bind "pcnet_cs"
However, when I run /etc/init.d/pcmcia start, the drivers read the card's
CIS info as: "PCMCIA", "1". I changed the version line in the config file
to match that, so it would try to load pcnet_cs. I think my laptop is
having problems reading from the CIS on the card, since the tuples are
stored there, and so is the version string stuff, and my laptop had trouble
with both. The other possibility is that the card has some lame CIS that
linux doesn't like :( By having somebody else test it, I hope to find out
which it is.
It would be great if someone could stick one of these NICs in their linux
laptop and tell me what it says for version, ("PCMCIA", "100BASE"), or
("PCMCIA", "1"). Also likely to be helpful would be setting up config to
load memory_cs instead of pcnet_cs, and running
dd if=/dev/mem0a count=20 | od -Ax -t x1 > ep427x.dump
and sending me the dump, along with what the MAC address if you have it.
(This is described in the PCMCIA-HOWTO.)
If you want to try to get the card actually working, patch pcnet_cs.c like
this:
--- old-pcnet_cs.c Wed Nov 22 19:05:45 2000
+++ pcnet_cs.c Fri Feb 23 18:54:21 2001
@@ -199,6 +199,8 @@
{ /* Socket EA */ 0x4000, 0x00, 0xc0, 0x1b,
DELAY_OUTPUT | HAS_MISC_REG | USE_BIG_BUF },
{ /* SuperSocket RE450T */ 0x0110, 0x00, 0xe0, 0x98, 0 },
+{ /* SURECOM EP-427X */ 0x01c8, 0x00, 0x00, 0x21, 0 },
{ /* Volktek NPL-402CT */ 0x0060, 0x00, 0x40, 0x05, 0 },
{ /* NEC PC-9801N-J12 */ 0x0ff0, 0x00, 0x00, 0x4c, 0 },
(the line numbers are probably messed up, so patch(1) probably won't apply
this cleanly. Just add the line with the + in front of it by using a text
editor.)
Recompile it and see if it works.
Thanks,
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: XF86Setup&XF86Config + gnome doesn't start
On Sat, Mar 03, 2001 at 12:25:40PM -0800, [EMAIL PROTECTED] wrote: > I typed the command and XF86Config was found at /usr/X11R6/lib/X11/ but > when I try to open it with emacs it writes: "file exists, but cannot be > read." Maybe a broken symlink? Maybe it was created by root without read for other permission, and you are running as an ordinary user? BTW, the locate command is good for finding files. I'll just tell you where to find XF86Config, though. If it exists at all, it'll be in /etc/X11. That's where all the X config files live. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: XF86Setup&XF86Config + gnome doesn't start
On Sat, Mar 03, 2001 at 02:48:25PM -0800, [EMAIL PROTECTED] wrote: > On Sat, 03 March 2001, Peter Cordes wrote: > > > XF86Config doesn't exist in /etc, therefore I asume that it isn't >installed properly or not at all. But if I have installed hole packages in >first cd disk, what is then missing?? You're checking /etc/X11/XF86Config, not /etc/XF86Config, right? /etc/X11/XF86Config is _the_ location for it. If that's what you meant, (or you meant the it wasn't in any directory rooted in /etc), then yeah, it's not there at all. Time to run the configurator: xf86cfg (for X4.0), or the old standby, xf86config. If you have problems, take a look at the man pages for these programs. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: cron & anacron
On Sun, Mar 04, 2001 at 11:25:43AM -0500, Dan Christensen wrote: > [EMAIL PROTECTED] (Bostjan Muller) writes: > On my system, the default big cron jobs are automatically ignored by > cron if anacron is install. Here is part of my /etc/crontab: > > 25 6* * * roottest -e /usr/sbin/anacron || run-parts --report > /etc/cron.daily > 47 6* * 7 roottest -e /usr/sbin/anacron || run-parts --report > /etc/cron.weekly > 52 61 * * roottest -e /usr/sbin/anacron || run-parts --report > /etc/cron.monthly > > So don't remove cron. My system is like that too. You're wrong, though; cron doesn't do anything if anacron is installed. _nothing_, not even run anacron. /usr/sbin/anacron runs as a daemon and does the run-parts stuff. If you check the package dependency, anacron only recomends, not depends on, cron. Go ahead and remove cron. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: cron & anacron
On Sun, Mar 04, 2001 at 07:55:51PM -0500, Dan Christensen wrote:
> Peter Cordes <[EMAIL PROTECTED]> writes:
>
> > On Sun, Mar 04, 2001 at 11:25:43AM -0500, Dan Christensen wrote:
> > > [EMAIL PROTECTED] (Bostjan Muller) writes:
> > > On my system, the default big cron jobs are automatically ignored by
> > > cron if anacron is install. Here is part of my /etc/crontab:
> > >
> > > 25 6* * * roottest -e /usr/sbin/anacron || run-parts --report
> > > /etc/cron.daily
> > > 47 6* * 7 roottest -e /usr/sbin/anacron || run-parts --report
> > > /etc/cron.weekly
> > > 52 61 * * roottest -e /usr/sbin/anacron || run-parts --report
> > > /etc/cron.monthly
> > >
> > > So don't remove cron.
> >
> > My system is like that too. You're wrong, though; cron doesn't do
> > anything if anacron is installed. _nothing_, not even run anacron.
> > ...
> > Go ahead and remove cron.
>
> On my system there are a lot of files in /etc/cron.d which cron *does*
> process, even though I have anacron installed. So I think my advice
> is sound, unless Bostjan has an installation with nothing in that
> directory.
Err, yes, you're right. I should have gotten more sleep before I wrote my
last message :(. I've got an old 486 laptop that I run really stripped
down. (I'm a hacker and I know what I'm doing, so it works for me:).
I like the idea of running cron/anacron stuff whenever AC power is present,
I'll have to try that, esp. if anacron does it automatically.
On my desktop, exim and logcheck have stuff in /etc/cron.d. Everything
else drops stuff in cron.{hourly,daily,...}.
grep 'cron\.d/' /pub/debian/Contents-i386 comes up with 35 files that want
to be in /etc/cron.d in all of Woody. Shouldn't be too much work to deal
with, esp since most of them didn't look like the kind of package you would
typically have on a laptop (e.g. new servers).
> My only point was that removing cron wouldn't accomplish anything that
> Bostjan wanted, and would have the potential of removing functionality.
Removing cron would make it stop reading files every so often, which causes
disk writes unless you have the filesystem mounted with the noatime
option. (I do this on my laptop, and haven't had problems with it.)
Fortunately, cron itself is not at all CPU intensive, only the cron jobs it
runs are, so you wouldn't save a whole lot.
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: WHERE are the docs?
On Mon, Mar 05, 2001 at 09:45:00AM -0500, William F. Dowling wrote: > Sometimes docs don't get installed when you think they should, so > using 'find ...' to find them is fruitless. locate(1) is good. It can search the whole system quickly. > Suggestion for Debian > maintainers: at install time let me specify "I want all relevant docs > by default". On one system I have (not my notebook) I managed > (newbily surfing through dselect) to have installed gcc, but no gcc > info pages. It took me a while to realize I needed the gcc-doc > package. Why not have foo-docless packages for those who really don't > want docs, and let the package foo contain the docs? Because that wastes space and bandwidth on the mirrors. Packages with separate doc-packages should recommends: foo-doc or suggests: foo-docs. If they don't recommend or suggest the package with the docs, then you should complain or file a bug report. :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: WHERE are the docs?
On Mon, Mar 05, 2001 at 04:47:12PM -0500, xsdg wrote: > On Mon, Mar 05, 2001 at 02:14:58PM -0400, Peter Cordes wrote: > > locate(1) is good. It can search the whole system quickly. > locate doesn't actually search the system. When a user runs updatedb(1?), >a database is created of all files accessible to the user running the >command. The locate command searches through the database. On debian >systems, updatedb is run in cron.daily. Yes, that's _why_ it can _effectively_ search the whole system quickly. find takes a long time because it has to traverse the directory tree. Locate just has to read its file where it keeps a list of every filename on the system. This is a good thing, because sequential reads are a lot faster than seeks and reads to scattered locations. I don't know if findutils runs updatedb in its postinst script, but if not, then a newly installed system won't have locate working for the first day. After that, the only downside is that it won't find files that were installed since last night. If a user recently installed a package, they probably looked at its contents. Still, things aren't perfect, but locate is a lot better than nothing. Another useful program is glimpse. I haven't used it, but it's supposed to make a (quickly) searchable database of the contents of all text files. It also only runs periodically. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: sound blaster configuration??
On Mon, Mar 05, 2001 at 04:23:26PM -0800, [EMAIL PROTECTED] wrote: > I have tried to find information in /usr/src/linux.. but I don't have any directory like ../src/linux/ .. thanks anyway.. That would be the kernel source tree. Grab it from www..kernel.org/pub/linux/kernel/v2.2/linux-2.2.18.tar.bz2, or install the kernel-source package. If you aren't going to recompile the kernel to suit your tastes, there might be a kernel-doc package which just has linux/Documentation, but not the rest of the source. If not, then just get the whole tree. A lot of really good docs on getting the kernel to do stuff is in linux/Documentation. > > > > > We first need more information: > > > > Is this Sound Blaster 16 PCI or ISA? > > Is it PnP or not? > > Which Kernel do you use? > > > > pnpdump -c > /etc/isapnp.conf && isapnp /etc/isapnp.conf did not give >any results, didn't find my sound card.. So I think it isn't a ISA.. I am >using kernel 2.2.. If it's PCI, lspci will find it (independently of whether any drivers are loaded for it. PCI is good, because there is a standard for scanning the bus and finding config info). (If you don't have lspci, you look at the obsolete /proc/pci. lspci interprets the binary data in the new /proc/bus/pci interface.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Friendliest Laptop??
On Mon, Mar 05, 2001 at 08:39:19PM -0800, Heather wrote: > Vendors right off the top of my head: Ascentia (sp?), Imperial, Kachinatech > (but the model they sell is far slower than you want), Dell if you twist their > arm, ARM Computer. I'm sure there are a few more, maybe I'll bother to > look at the hardware howto and narrow it down sometime. Apple makes laptops that run debian-powerpc quite nicely, but are pretty pricey :( Good battery life and no noisy power-sucking fans though. Ascentia was a line of laptops made by AST. AST is out of business now. Good thing, too, according to this tremendously pissed off customer: http://www.netppl.fi/~findians/paper.html I don't know if all AST laptops suck as much as this guy says they do, but my AST Ascentia 810N is, um, ok. It has a nice trackball, not a little mouse nipple in the middle of the keyboard, and it runs cool enough to be called a laptop (even for people wearing shorts :). The biggest problem is that I can't upgrade the RAM past 20MB :( I got it for only 300 $ last year, so I'd say it's worth it. Anyway, you [Yannick] might want to consider a used laptop unless you have a good reason for needing an 800MHz Pentium III. You can get a pretty good machine for a whole lot less than you'd pay for a new one. My next laptop will probably be a powerpc or an AMD K6-2 (or some other K6-* processor). If you ever need to seriously crunch anything, just ssh to a powerful desktop box. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Friendliest Laptop??
On Mon, Mar 05, 2001 at 09:45:00PM -0800, Heather wrote: > > On Mon, Mar 05, 2001 at 08:39:19PM -0800, Heather wrote: > > > Vendors right off the top of my head: Ascentia (sp?), Imperial, > > > Kachinatech > > > (but the model they sell is far slower than you want), Dell if you twist > > > their arm, ARM Computer. I'm sure there are a few more, maybe I'll bother > > > to look at the hardware howto and narrow it down sometime. > > > > Apple makes laptops that run debian-powerpc quite nicely, but are pretty > > pricey :( Good battery life and no noisy power-sucking fans though. > > > > Ascentia was a line of laptops made by AST. AST is out of business now. > > Good thing, too, according to this tremendously pissed off customer: > > http://www.netppl.fi/~findians/paper.html > > The Ascentias are still listed in LAPTOP magazine's chart this last month. > Unless someone else took up the product line, this company failure was recent? Hmm, I guess they must still be around. I thought I read that somewhere, but I guess I must have mis-remembered. Maybe it was just one product line that they stopped making. Also, the fact that AST sends you off to http://ast.com/Quick_Support_Guide.htm for support gave the impression that they weren't up to much anymore. Anyway, I guess they're still around, since the page says: NOTE: Prior to February 1999, all AST brand products were manufactured and sold by AST Research, Inc., which is now providing support under the business name ARI Service (www.ari-service.com). After January 1999, all AST brand products were manufactured and sold by AST Computers, LLC, which provides support under the business name AST (www.ast.com). Maybe they just made a big change (and hopefully got rid of the crap described http://www.netppl.fi/~findians/paper.html!). Well, I don't know what's really going on, so I'll stop rambling now. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: How do I finish an install of X?
On Tue, Mar 06, 2001 at 02:28:12PM +, Rod Young wrote: > > I appreciate the updatedated KDE info. Thought right now I don't have a > network card for it. > What is the difference between dselect and apt-get? They are two different front-ends for package management. apt lets you do stuff like apt-get update, and apt-get dist-upgrade, or apt-get install mozilla. dselect uses ncurses to give you a menu of packages to select from. dselect is good if you want to poke around and see what there is. (apt-cache search is good for this too.) apt-get is faster for installing something when you know what you want. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: How do I finish an install of X?
On Tue, Mar 06, 2001 at 02:29:55PM -0800, Heather wrote: > console-apt, usually typed 'capt' and previously called 'deity', uses the apt > mechanisms but also provides a curses front end. It's only about a billion > times more fun to use than dselect... > > apt-get install console-apt > apt-get update > capt > > ...enjoy! Thanks dude. I need to get a slower internet connection so I can play captris longer while capt does its download... ;-) It's been too long since you used apt-get; I think you meant apt-get update; apt-get install console-apt; cap; (update before install :) Anyway, now that I've got it installed on my IA32 box (it's got broken deps on powerpc :(, it doesn't seem to ever show anything about recommended or suggested packages, though. I like how dselect lets you see the raw package info if you want to. I read the man page and the online help but didn't see anything about it. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: How do I finish an install of X?
On Wed, Mar 07, 2001 at 11:20:03AM +1100, Brendan J Simon wrote: > Peter Cordes wrote: > > > Anyway, now that I've got it installed on my IA32 box (it's got broken deps > > on powerpc :(, it doesn't seem to ever show anything about recommended or > > suggested packages, though. I like how dselect lets you see the raw package > > info if you want to. I read the man page and the online help but didn't see > > anything about it. > > Have you tried "apt-cache show "? > eg. "apt-cache show apt" or "apt-cache showpkg apt" showpkg? Ahh, so there is an easy way to see the reverse deps of a package! Why didn't I find this before :/ Anyway, I had broken dependencies because I hadn't done a dist-upgrade since I did an update, and apparently a new version of apt that provides: libapt-pkg-libc6.2-3-2-3.1 had been uploaded. That is needed for aptitude, but console-apt depends on libapt-pkg2.7, which is not available in unstable anymore. (I've got more than one computer, so I can afford to run unstable on some of them :). The version of apt in woody provide libapt-pkg2.7, so I was able to install it on my woody machine. It's running sid that's causing the problem, not powerpc. (My most recent problems have been due to stuff not being available in powerpc, so my first assumption was that it was powerpc that was the problem.) Aptitude works like a charm on bigfoot (my huge ppc tower:). It will do the actual installs now. Just press "g" to carry out the installs/upgrades/removals/purges selected. It puts up a list of what will be done for comfirmation, like console-apt, but you can make changes on that list directly, instead of having to remember a list of packages you want to change something with. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: french and czech keyboard together
On Wed, Mar 07, 2001 at 02:52:48PM -0800, Heather wrote: > > someone knows how to get working together czech and french keyboard > > without the need of resetting LC_CTYPE every time you want to setxkmap to > > different language (btw: fr is iso-1, cz is iso-2 and without appropriate > > LC_CTYPE you can't type in special accented characters > [...] > There's at least one toy for toggling keyboards within X, but for the console > mode I'm not so sure. What about setkeys? You might need to use consolechars to load a font that will display the characters you want to type, though. Anyway, all the goodies are in the console-* packages. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: maximum hard drive limitations?
On Fri, Mar 09, 2001 at 02:18:38AM +1100, Drew Parsons wrote: > This question isn't strictly Linux, but I hope you'll be kind :) > > I have a Toshiba 490CDT model, which comes with a 3.8GB harddrive, and the > Toshiba docs seem to be saying that the maximum drive that can put in the > computer is 6GB. You might have problems booting off >6GB, if the limit is the BIOS. Usually you can't boot off >7GB (except with LBA32). This isn't a problem. Just make /boot early on the drive. AFAIK, laptop IDE is the same as regular IDE, so there should be no reason why you can't address the same number of sectors as you can there (once linux is booted and controlling the IDE controller directly). However, all I've done is to say that I can't think of any good reason why there should be a 6GB limit. Borrow somebody's >6GB drive, and see if it works. (i.e. boot off it, or (safer for the drive's data) boot from a rescue floppy and dd if=/dev/hda of=/dev/zero bs=1024k, and see how many blocks it transfers. (1024k = 1MB, so the block count will be MB.) Do this, rather than df or cat /proc/partitions, because this makes sure you can actually read out that far on the disk. Running badblocks might be a good idea too. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: 2.4/sid on toshiba satellite
On Thu, Mar 08, 2001 at 11:56:33AM -0700, Joel Dudley wrote: > Hello all, > I have 2.2.17/sid running pretty well on my Toshiba Satellite Pro 4300 > laptop minus sound and modem support (yamaha pci sound, lucent winmodem). I > tried to install the 2.4.2 kernel because I know that it supports my sound > card. However, after I load the kernal nothing works because the kernel > detects several devices on IRQ 11. In the BIOS IRQ 11 is used for the PCI > bus and I am not able to change that (crappy BIOS :-( ) why would 2.2.17 > work fine with the PCI BUS and 2.4.2 freak out and not find my devices? It > actually finds the yamaha sound card but complains about IRQ use. Thanks. Try using setpci. I don't remember exactly what it can do, but it might let you reconfig a PCI device to use a different IRQ. It's in the pciutils package, IIRC. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Losing interrupts
On Thu, Mar 08, 2001 at 11:53:05PM +0100, Mikael Cardell wrote: > After I upgraded my Thinkpad 570 to Linux 2.4.0 I started noticing > that it complained about losing interrupts and that the HDD wouldn't > spin up after suspend one time in 10 or so. It's very irritating and > means I have to reboot to get the HDD spinning again. If I revert to a > 2.2.x kernel, this never happens. > > I'm guessing it has something to do with the APM support and I've > tried different configurations, like allowing or disallowing > interrupts during calls to the APM BIOS, but to no avail. I will check > again with a 2.4.2 kernel as soon as I get back on a fatter pipe (I'm > at home trying to fight the flu right now), and I hope that will fix > things. > > Has anyone else had any similar problems? Yes. I've seen this on my AST 810N. I don't use my laptop a very much (I will more once I get the NIC working and get stuff transfered to a new HD...). When I have used it, I've had problems with this sometimes. IIRC, having apmd _not_ running let it spin down the driver without problems. I used hdparm -S96 /dev/hda to set the spin-down time on my drive. I don't remember what I figured out about it, and whether the kernel will spin down the drive when it's ready or what. You said it only happens one time in 10 for you. I had it happening every time, or never, depending on how I set stuff up, IIRC. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Suspend to disk partition on ThinkPad T20
On Tue, Mar 13, 2001 at 04:22:18PM +0100, Russell Coker wrote: > Making a DOS partition is a problem as I don't have Windows installed and it > seems that there is no DOS software to setup such a partition. Who needs DOS? yeti:~$ grep mkdosfs /pub/debian/Contents-i386 sbin/mkdosfs otherosfs/dosfstools usr/share/doc/dosfstools/README.mkdosfs otherosfs/dosfstools usr/share/man/man8/mkdosfs.8.gz otherosfs/dosfstools usr/share/man/pl/man8/mkdosfs.8.gzdoc/manpages-pl happy hacking :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Badiane: make-kpkg PCMCIA conflict
On Tue, Mar 13, 2001 at 09:44:47AM -0800, Francois BOTTIN wrote: > > --- Badiane Ka <[EMAIL PROTECTED]> wrote: > > I have created a deb package with make-kpkg and upon > > executing dpkg -i I get a message saying that there is > > conflict between my custom 2.2.18pre1 image and > > pcmcia. How do I compile, install and get a kernel to > > run under debian without running into a bunch of > > module problems. I would like to be able to upgrade > > my system without having all of these problems. If > > there is literature about this problem please point me > > in that direction and I will follow. > > > > Badiane > > > I don't know if there is any documentation about that, but what I do is: > - remove the kernel and pcmcia packages that are in use (yes, there's a > warning saying that it's dangerous to remove a running kernel's > package...) > - install the new kernel package, and then the pcmcia one. > > Simply, do not try rebooting between the two stages... You might want to copy your vmlinuz in case of accidents. It's a bad idea to ever have no kernels on your machine. At least make sure you have a boot disk. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
debian-laptop@lists.debian.org
On Tue, Mar 13, 2001 at 06:02:50PM +0100, Andreas Mohr wrote: > On Tue, Mar 13, 2001 at 05:46:31PM +0100, Dieter Faulbaum wrote: > > > > Hello, > > > > I try to install a 2.2-system on such a laptop (a 486DX2 with 8MB > > memory) but get thesh error messages after loading the root.bin > > disc (the last lines several times): > > > > Freeing unused kernel memory: 148K freed > > VM: do_try_to_free_pages failed for kswapd > > > > > > Can anyone tell me what can be wrong? I use the rescue.bin disc from > > safe. Are 8MB not enough for this version? > I'm pretty damn sure that this is only a debug message left over. > I had this, too, with certain 2.2.x kernels. > > AFAICT this is only informative, not critical. > > But OTOH ICBW :-) Translation for non-acronym-speakers: AFAICT = as far as I can tell OTOH = on the other hand ICBW = I could be wrong. I think it's a problem if you see "do_try_to_free_pages failed". The kernel may even have killed off its own kswapd when that happened. Try installing with the compact or the idepci images. They have less stuff compiled in, so they're more likely to not run out of memory before you can set up a swap partition. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Laptop Recommendations
On Fri, Mar 16, 2001 at 04:44:03PM -, Todd Kokoszka wrote: > Hi, > > I'm interested in buying an old laptop and installing Debian on it. What are > people's recommendations? Get something with enough RAM for what you want. Esp. older laptops sometimes use custom RAM that only works with that model, so upgrades are expensive. Try to get one with at lithium battery, since they're better. Nickel-metal-hydride (NiMH) batteries are ok. Don't get nickel-cadmium, since they don't hold as much charge, and suffer from the "memory effect", where they lose effective capacity if you don't discharge them all the way. (you can reverse this a bit, but it's not good.) In general, laptop upgrades cost a lot more than for desktop machines. Don't buy a really cheap laptop and plan to upgrade. (unless it was really, really cheap, and not too bad.) Hard drives are an exception. Laptop ones cost more, but they are compatible. Some laptops with slow-by-modern-standards processors have CPUs which were the fastest for their time, and use a lot of power. I've heard tales of P200MMX laptops that would run for half an hour on a full charge. If you're going to be near an outlet most of the time, you can get away with buying a power-hungry laptop, though. Another thing to watch for is that older laptops might not be PCI-based, in which case they probably won't take 32-bit (cardbus) PC cards, and you won't be able to tweak the hard drive performance with hdparm very well (probably). If this matters to you, then get a PCI laptop. (I think most pentium laptops would use PCI). -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: quality of postscript display
On Fri, Mar 16, 2001 at 04:40:32PM -0500, Steven K Thompson wrote: > I have Debian 2.2r2 on a Vaio F340 laptop. When I view a > postscript file with gv, the type is broken and hard to > read on the display. This is true for example when I try > to view files in the Debian documentation that are provided > in postscript form, such as debian-guide, and is also true > when I have created the postscript file myself with dvips. > In contrast, the type is very clear > for example when viewing a latex dvi file with xdvi. > > Would my problem have to do with the fonts used in the > postscript file or something quite different? > > I would appreciate any explanation and advice anyone can > provide on this. dvips-produced postscript files always look terrible on screen. Read the HTML version. If you want to produce screen-viewable stuff with LaTeX, use pdflatex to make a PDF. That way, you'll get scalable fonts and things will work well. I don't know all the details of how fonts work and stuff, but I do know that what you see is normal. If you just want to view on screen, then stick with xdvi. (BTW, some people on my LUG mailing list were trying to produce a PDF version of the debian-guide, for the benefit of people with windows computers. pstopdf ends up using bad fonts (unless you configure it a certain way, which I don't remember), so it's not even readable when viewed with xpdf. pdflatex produced a PDF where the text was offset half off the page to the right. If you have the spare time to figure out what's going on, that would be great. Debian should make a PDF version of the debian-guide available, at least one we figure out how to make one that looks right and uses scalable fonts.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Switching keys around
On Sun, Mar 18, 2001 at 10:13:02AM -0800, Heather wrote:
> > On Sat, 17 Mar 2001, Joanne Hunter wrote:
> > > 2) How can I do the same under the Linux console? (Is it possible?)
> >
> > Yes, it's possible. I don't know how, though, and think it involves
> > editing the keymap files and loading them. Read the documentation on the
> > `console-tools' (or `kbd', if you have that installed instead) packages,
> > they /may/ help.
> >
> > Daniel
>
> There is a directory where alternative keymaps are kept (for example, Dvorak).
> You can copy the one for qwerty and add your own things, or change them. Of
> course the normal keymaps are .gz so you need to ungzip them first ... but
> after that they are plaintext and it made good sense to me, at least. Plus,
> in my experience, the new one wouldn't work until I gzipped it again.
You have to gzip again because /etc/init.d/keymaps-lct.sh specifies the .gz
extension, instead of just giving the base name, and letting loadkeys add
the .gz if there isn't an un-gzipped one.
I spend enough time on the console that I've spent some time figuring out
the programs in console-tools, so I'll pass along what I've figured out:
There's two ways go about making changes to the console keymap. One way is
to edit /etc/console-tools/default.kmap. This has the disadvantage of not
recording what you changed, and requiring more hacking if an upgrade to
console-tools-data has a new keymap (this hasn't happened happened to me
since I've been running debian, though...). To swap any keys, use showkey
to find out their numbers, then edit default.kmap and swap the numbers in
the "keycode" lines. e.g. to swap the "6" and "7" keys that are above the
alphabetic keys, find the lines:
keycode 7 = six asciicircum
control keycode 7 = Control_asciicircum
alt keycode 7 = Meta_six
shift alt keycode 7 = Meta_asciicircum
keycode 8 = sevenampersandbraceleft
Control_underscore
alt keycode 8 = Meta_seven
shift alt keycode 8 = Meta_ampersand
control alt keycode 8 = Meta_Control_underscore
and change them to:
keycode 7 = sevenampersandbraceleft
Control_underscore
alt keycode 7 = Meta_seven
shift alt keycode 7 = Meta_ampersand
control alt keycode 7 = Meta_Control_underscore
keycode 8 = six asciicircum
control keycode 8 = Control_asciicircum
alt keycode 8 = Meta_six
shift alt keycode 8 = Meta_asciicircum
There will only be a single line for simpler keys like "Insert", and "Alt".
Instead of changing /etc/console-tools/default.kmap, you can override parts
of it. One of the directories loadkeys searches by default is
/usr/local/share/keymaps, so that's a good place to put stuff. I have two
files there:
/usr/local/share/keymaps/ctrl_nocaps.kmap:
keycode 58 = Control
keycode 29 = Control
keycode 97 = Control
/usr/local/share/keymaps/windowskeys.kmap:
keycode 125 = Alt
keycode 126 = Alt
keycode 127 = Alt
(this is on an x86 desktop machine).
Then, change /etc/init.d/keymaps-lct.sh to load a base keymap ("us" in my
case) , and the "patch" files you want. Mine (with cruft removed) is:
#!/bin/sh
case "$1" in
start | restart | force-reload | reload)
/bin/loadkeys us ctrl_nocaps windowskeys
;;
stop)
;;
*)
echo "Usage: $0 {start|stop|restart|reload|force-reload}"
exit 1
;;
esac
the .kmap extension is added by default, and gzipped files are handled
automatically, by loadkeys.
(BTW, ctrl_nocaps makes caps lock be control. This makes a standard PC
keyboard have control in the Right place, which is where the wrong-headed PC
kbd designers usually place the much-less-useful caps-lock key. I named it
after the Xkbd option that does that for XF86, IIRC. windowskeys assigns
all three of the suckers to be Alt.)
On my old computers without windoze keys, I make the right alt key be alt
instead of altgr, so I can hold it with my thumb while I use the arrow keys
to switch consoles. This lets me switch consoles with one hand. If you
are a boring english-speaker like me, you probably won't ever use altgr for
anything...
Hope this helps.
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Display errors on Sony Viao
On Mon, Mar 19, 2001 at 04:47:00PM -, jimmy sandhar wrote: > Hi All, > While trying to run the startx my system errors as follows: > X: exec of /etc/X11/X failed > xinit: Connection refused ( errno 111 ): unable to connect to X server > xinit: No such process (errno 3): Server error > Please help Install the appropriate X server package. If you're running potato, then install xserver-mach64, or whichever one is right for your video hardware. If you're running woody or sid, (and thus have XFree86 4.0), install xserver-xfree86. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Today's Challange
On Fri, Mar 23, 2001 at 08:46:31AM -0800, [EMAIL PROTECTED] wrote: > G'Day ! > > I need to install Debian on several computers. However the twist is that > the hardware on these computers is such that the only way to access the > drives is by pulling the drive and making it the slave drive on a working > Linux desktop computer. (ie no floppies, no CD, no NIC, no Modem on > computers being upgraded. They are currently NT). > > I have read the docs included on the Debian CD and they hint it is > possible but do not offer specifics. > > Any suggestions ? Unless you have debian systems with known-good configurations, you will not be happy if you need to rearrange your drives every time you want to install a package you forgot you wanted. Get some NICs, or at least set up ppp over null modem (this will be ok for fine tuning, but pretty slow for major installs/updates, unless you leave it overnight). As for the mechanics of it, put the drive in the working linux box, then partition it, mkfs/mkswap it, mount it on /target, mount other partitions in the appropriate spots under /target, untar base-2.2.tgz in /target, configure lilo, and, umm, do some other stuff that I probably forgot here, like put a kernel+System.map in /target/boot. For convenience, you might as well just leave the drive jumpered as slave, if your BIOS can boot it that way. Otherwise, you might have problems with lilo if it was set up to boot from hdb, but the drive that used to be hdb was now hda. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Today's Challange
On Fri, Mar 23, 2001 at 07:22:27PM -0800, Heather wrote: > And if you want to boot from floppy or not at all -- perhaps as a security > measure? -- then make sure a floppy is in, and use boot=/dev/fd0 You can put this in lilo.conf to boot a floppy from lilo: other=/dev/fd0 label=floppy unsafe This lets you boot off the HD, so you can get into linux fast without having the BIOS grind your FD. If you want to boot off a floppy, you can do so without going into the BIOS setup. (The "unsafe" option makes it possible to run /sbin/lilo without a floppy in the drive.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: compaq internet keys
On Mon, Mar 26, 2001 at 12:47:59AM -0500, Brian Stults wrote: > Does anyone know how to remap the "internet zone" keys on a Compaq > 1800? I tried xkeycaps, but it doesn't recognize the key when pressed. > xev also does not notice the key being pressed. The X server can't send events if the kernel isn't passing keycodes to it, or if it doesn't know what events to map the keycodes to. Use showkey(1) to find out what the key codes are, then feed something like keycode 200 = Hyper_L or keycode 220 = XF86LightBulb (I just fired up xkeycaps to see what weird stuff you could do with extra keys, and under Vendor keysyms, XF86LightBulb really exists... :) BTW, some/all of this may not work at all, since I just pulled it out of my ass :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: compaq internet keys
On Mon, Mar 26, 2001 at 10:50:14AM -0800, Alan Chen wrote: > BTW, I think you might need to be in a non-X console for showkey to > work. Yes, of course. > I have similiar keys on my HP laptop, but I haven't gotten to > work yet. Talking to the author of the keyboard howto, if showkey > doesn't respond to keypresses for your "zone" keys, you might need to > play with some kernel code to see if the kernel is even receiving data > for those buttons via the kb interface. You can see if pressing them makes anything at all happen, by looking at the interrupt counters in /proc/stat. Run watch -n1 grep intr /proc/stat and press the keys. If any interrupt counters other than the timer (irq 0), ethernet, and disk (and anything else that generates interrupts by itself) go up, then you've got something. The first number after intr is the total of all interrupts, the numbers after are the counts for irq0, irq1, ... BTW, this will work even if there is no driver using that IRQ, which is necessary for /proc/interrupts to show it. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Help - PCMCIA card disables serial port!
On Mon, Mar 26, 2001 at 12:54:37PM -0800, Michael Dickey wrote: > Inserting my PCMCIA modem/ethernet card disables my serial port. I am sure > that it is not an irq problem, as they both use different irqs when > functioning Can you get at the serial ports that "disappear" by using setserial to assign the right port and IRQ to ttyS1, while ttyS0 is the pcmcia modem? > Before pcmcia card: > $ cat /proc/tty/driver/serial > serinfo:1.0 driver:4.27 > 0: uart:16550A port:3F8 irq:4 baud:9600 tx:0 rx:0 > 1: uart:16550A port:480 irq:3 tx:0 rx:0 CTS|DSR|CD|RI > 2: uart:unknown port:3E8 irq:4 > 3: uart:unknown port:2E8 irq:3 > > After inserting card: > $ cat /proc/tty/driver/serial > serinfo:1.0 driver:4.27 > 0: uart:16550A port:A80 irq:3 baud:9600 tx:0 rx:0 CTS|DSR > 1: uart:unknown port:2F8 irq:3 > 2: uart:unknown port:3E8 irq:4 > 3: uart:unknown port:2E8 irq:3 > ... -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: compaq internet keys
On Mon, Mar 26, 2001 at 01:35:33PM -0800, Alan Chen wrote: > Well, there was no joy on /proc/stat. None of the interrupts > responded to "zone" keypresses, no interrupts on the "rocker" switch > on my keypad either. I'm thinking that something needs to be enabled > first. Back to plan A, look at kernel code. :) Feel free to look at the kernel code, but I suspect it won't get you anywhere. The keyboard uses interrupts, not polled IO, so if your actions don't generate interrupts, then no part of the kernel is getting notified of them. I guess it's possible that these keys just set some kind of state that the kernel has to poll to see keypresses, but that would be really stupid. It would work so much better just to send a keycode. OTOH, some of your keys might just send stuff to some power-management hardware, but if you're talking about "internet" keys, then that's probably not it. Since the keys don't seem to do anything normally, I would guess that the extra keys have to be enabled by the kernel somehow. The kernel can set some stuff on the keyboard (e.g. the numlock/etc. lights), so I guess it's possible that the kernel has to set the hardware to use these keys before they do anything. (still seems lame, I don't know why they would design the hardware like that.) Happy hacking, -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: slink, potato and what else
On Tue, Mar 27, 2001 at 03:11:39PM +, Rod Young wrote: > What sid woody and the other versions? http://www.debian.org/releases/ -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Badiane: CPU speed on my TP600X
On Tue, Mar 27, 2001 at 08:18:33AM -0800, Badiane Ka wrote: > My CPU speed on my laptop shows up as 132MHz instead > of 500MHz. Does anyone know how I can reset it? Does it have speedstep or something? You can probably change an option in your BIOS to use full speed mode. Boot up in such a way as to have the CPU running at full speed when it checks the bogomips. You can get big problems if the kernel things bogomips is too low (since it will wait too short a time when CPU is running fast), than if it is too high (kernel will wait too long when running at low CPU speed.) Waiting too long is almost always safer than not waiting long enough, for typical IO devices, so you should make sure the kernel sees the high speed when it's calibrating. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Badiane: CPU speed on my TP600X
On Tue, Mar 27, 2001 at 11:42:09AM -0500, Alan Shutko wrote: > Badiane Ka <[EMAIL PROTECTED]> writes: > > > My CPU speed on my laptop shows up as 132MHz instead > > of 500MHz. Does anyone know how I can reset it? > > [...] > You can just reboot on AC power and /proc/cpuinfo will show the right > value, but it's just a cosmetic problem. It's _not_ cosmetic. The kernel doesn't just measure bogomips to make you feel good because you have a lot of them. (The CPU MHz value is cosmetic, but it gets measured along with bogomips). It measures so it will know how many times through the loop it has to go to wait for e.g. 10 microseconds. If your hardware requires certain IO accesses to be separated by at least x amount of time, you'll have problems. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Dell Latitude C800 and NIC card.
On Wed, Mar 28, 2001 at 04:07:14PM +0100, Steve Dobson wrote: > I have held back going the kernel 2.4.x, and one of the reasons is for > device support. If you getting it going under 2.4.x please let me know. If you leave out pcmcia when you compile the kernel, you can use the standalone pcmcia-cs kernel modules and user space stuff, so you get all the stability and hardware support that you had before. I think pcmcia-cs.sourceforge.net recommends doing this, unless you want to hack with the kernel drivers. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: ailing gpm - tp380ED
On Thu, Mar 29, 2001 at 10:49:36AM -0800, Jeff Coppock wrote: > I just checked my version: > > gpm-Linux 1.17.8, $Date: 1999/01/03 21:02:51 $ > > jc > > On Thu, Mar 29, 2001 at 09:49:56AM -0800, [EMAIL PROTECTED] wrote: > > > > G'Day ! > > > > Perhaps I was using an old GPM library A couple of years ago back > > when I upgraded to XF86 3.x, GPM and X conflicted. I always had to > > kill the GPM demaon, before starting X. > > > > I guess this was fixed. > > > > cheers, > > Jim Parker GPM never conflicted with X, except maybe before it supported the repeater option. The problem has always been that the PS/2 device driver only allows a single process to open the device and read events, IIRC. The serial driver is different. gpm and X can both open /dev/mouse simultaneously if it is linked to ttyS0. (I don't know whether they both get a copy of the mouse data, or whether gpm gets some and X gets some. If the latter is the case, it wouldn't work anyway. Oh well.) I've always used gpm in repeater mode with my serial mice, since I've got one with 3 buttons that gets reset to microsoft protocol unless I hold down the left button when something opens the device. (The control lines on the serial port are toggled on device open, or something.) Since I want it to stay in MouseSystems protocol mode for 3 button support, I don't want the X server openning the device when I start it. (I sometimes like to bounce between console and X.) Anyway, my point is that GPM has been working like this for over 2 years, I think. It's all about configuring it right, not some recent feature. (Unless it's even smarter now...) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: ailing gpm - tp380ED
On Thu, Mar 29, 2001 at 03:59:17PM -0800, [EMAIL PROTECTED] wrote: > G'Day ! > > Beginning to think its hardware (or maybe not). First had the problem in > '94 on a 486 w/ VESA Bus and maybe 8M RAM, and serial mouse. Back then I > modified my startx script to kill GPM before starting the X server. > > Old habits never die ;-) It won't be hardware. The hardware doesn't decide whether the kernel copies mouse movement info to multiple processes that have the device open (unless the driver's behaviour changes depending on minor differences in hardware or something!). By far the most important factor is what kernel version you use, since this problem/non-problem is dependent on exactly how the driver behaves when X opens /dev/psaux while gpm has /dev/psaux open. (or vice versa). Someone said that restarting gpm while X was running was not good, even though they weren't using gpm in repeater mode. This sounds somewhat familiar, actually. I think I've seen something like it on ps/2 mouse systems. One possible driver behaviour which would give this result would be for the driver to pass mouse data to whichever program most recently openned the device. I don't have time to test this now, but I can't let you get away with blaming innocent hardware for stuff it didn't do! ;-> -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Horrible font on Toshiba 740CDT
On Fri, Mar 30, 2001 at 01:45:33PM +0100, Conry, Lance wrote: > Hi List > > I've just installed potato on my Toshiba 740CDT. When I open a console in > X, the fonts are horrible. Much like what it looks like in windows when you > have the incorrect screen resolution and or colour depth. I've not had this > problem before, but then again, this is the first laptop install I've done. Try installing some of the other fonts packages. I don't know what ones to recommend, but that's probably your problem. > > I'm running the SVGA X server, 1024x768 60hz 16. > > The spec says the machine will run at 1024x768 at 60Hz, with 64k colours. > I've also enabled 640x480 and 800x600, but 1024x768 is the default. > Interestingly, when I press the +/- I can only cycle through two > modes, which suggests that they aren't all available. I'm wondering if > 1024x768 isn't working. Is there any way to find what the current > resolution is? Read the stdout from the X server (which is probably logged to a file if you're using xdm or equivalent). It will say what modes it let you use, and why it tossed the modes it isn't letting you use. (If you used a setup tool to write your XF86Config, find the Modes line in the Display subsection you are using (depth=16), and list all three modes if they aren't there.) xvidtune will actually show what's going on, if you want to be sure, instead of correlating the screen sizes with the list of modes the server is using. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: installing w/udma66
On Sat, Mar 31, 2001 at 04:00:15PM -0500, Dave Linsalata wrote: > Hey guys, > > I was wondering - if I have an udma66 HD, can I use the udma66 installation > disk set to install Debian with the HD on the ata/66 channel? Or do I still > have to switch it to the ata/33 channel to install it? I was able to install on a desktop Athlon with an AOpen AK72 motherboard (VIA KX133 chipset, 82c686a south bridge), and the hard drive on the UDMA66 controller. (drive supports udma66 also). Things worked fine as long as I didn't enable DMA. (well, if you consider 5MB/s disk IO fine...) I'm working on getting udma66 running with 2.4.3 + Andre Hendrick's ide patch, but I guess I'm doing something wrong because it's still using PIO. Still, I was able to get linux installed without moving any drives around. This will work as long as it is possible to use your controller in non-udma66 modes. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: HP Pavillion n5270
On Wed, Apr 04, 2001 at 12:36:24PM -0500, John R. Sheets wrote: > I've only tried playing a couple MP3's so far, mainly because I > haven't figured out how to change the sound volume yet. Does anyone > know how to do that? Do you need mixer support compiled into the > kernel (and if so, which kernel option?), or are there userspace > volume apps? apt-get install aumix If aumix doesn't do it, then you're probably out of luck. However, you could try alsa, in case it has better support. http://alsa-project.org/ > Another issue I have with the maestro3 driver is that depmod -a keeps > complaining about unresolved symbols. I suppose I may have compiled > it against the wrong kernel source tree or something. However, since > it loads and runs, I haven't been motivated enough to dig any deeper. The unresolved symbols must be in a module you haven't tried to load yet. When you try, insmod will refuse to load it. (And no, you can't override it or something, for good reason. The only way it could "override" would be to load broken code into the kernel. This is not the same as trying to load a module from a different kernel version.) 2.4.x on powerpc has a lot of problems with some parts of the tree. Maybe some of these are present on IA32 as well. Stuff like the enhanced RTC driver (used to? don't remember if it's working yet) would break the compile if you said Y in the config, or not load if you compiled it as a module. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: How to paste with two buttons?
On Wed, Apr 04, 2001 at 10:43:21PM +0200, Oliver Ob wrote: > xsdg schrieb: > > > > On Mon, Apr 02, 2001 at 08:15:46PM -0500, Charles Blair wrote: > > >My compaq machine has two things for mouse buttons (touchpad). > > > How do I paste? > > press SHIFT+INS > > --xsdg > > > i would be more delighted if the ctrl+v combination worked. If you actually want that behaviour (which breaks a bunch of progs that use ^v for other things), you can set it up for xterms by binding insert() to ^V using X resources. (This would only change text programs that run in the xterm.) Or, you could use xkeycaps to set up your keyboard to send an Insert keysym when you press control-v. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: installing w/udma66
On Thu, Apr 05, 2001 at 08:22:19AM +0200, Andreas Tille wrote: > On Sat, Mar 31, 2001 at 04:00:15PM -0500, Dave Linsalata wrote: > > I was wondering - if I have an udma66 HD, can I use the udma66 installation > > disk set to install Debian with the HD on the ata/66 channel? Or do I still > > have to switch it to the ata/33 channel to install it? > Possibly this Problem is related: > I'm using kernel 2.4.2 and get the following messages: > > PCI: Using IRQ router PIIX [8086/7110] at 00:07.0 > got res[1000:1fff] for resource 0 of Ricoh Co Ltd RL5c476 II > got res[10001000:10001fff] for resource 0 of Ricoh Co Ltd RL5c476 II (#2) > Limiting direct PCI/PCI transfers. > > ... > ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx > ^^ > PIIX4: IDE controller on PCI bus 00 dev 39 > PIIX4: chipset revision 1 > PIIX4: not 100% native mode: will probe irqs later > ide0: BM-DMA at 0x1040-0x1047, BIOS settings: hda:DMA, hdb:pio > ide1: BM-DMA at 0x1048-0x104f, BIOS settings: hdc:pio, hdd:pio > hda: HITACHI_DK23BA-20, ATA DISK drive > ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 > > > What is limited here I don't know. Maybe it's turning off transfers that go directly from one PCI device to another, instead of to the CPU or RAM (I said maybe. I don't know what it's doing, or why.) Read the kernel code that prints out that message. > and why is 33MHz assumed if I have UDMA 66MHz? That's innocent. Keep reading: it says Assuming 33MHz ... for PIO modes. i.e. it doesn't get used except for PIO transfers. The driver needs the bus timing to know how long it has to wait before stuffing more bits down the pipe. PIO modes are the ones where the CPU does all the copying itself. UDMA modes have the IDE controller write to memory directly, so the CPU doesn't need to worry about the timing. There might be a file in /proc that will give you the info on stuff. On my VIA KX133 chipset machine, the via82cxxx code creates /proc/ide/via with status info. (This is with Andre Hendrick's IDE patch.) > These kernel messages are new since kernel 2.4.2 (which are reported > by logcheck. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: HP Pavillion n5270
On Thu, Apr 05, 2001 at 07:26:46PM -0500, John R. Sheets wrote: > On Apr 04, 2001, Peter Cordes <[EMAIL PROTECTED]> wrote: > > > Another issue I have with the maestro3 driver is that depmod -a keeps > > > complaining about unresolved symbols. I suppose I may have compiled > > > it against the wrong kernel source tree or something. However, since > > > it loads and runs, I haven't been motivated enough to dig any deeper. > > > > The unresolved symbols must be in a module you haven't tried to load yet. > > When you try, insmod will refuse to load it. (I should have read your message more carefully, I didn't notice that you said the module with the broken syms did actually load.) > > The maestro3 module does successfully load, though. Does that make sense? > > <[EMAIL PROTECTED] ~># lsmod | grep maestro > maestro3 25232 1 > ac97_codec 7152 0 [maestro3] > soundcore 2384 2 [maestro3] > <[EMAIL PROTECTED] ~># depmod -a > depmod: *** Unresolved symbols in /lib/modules/2.2.18/misc/maestro3.o > AFAIK, that doesn't make sense. Maybe I wrongly assumed that depmod would check against the currently running kernel. Could depmod be checking symbols against a wrong System.map or something? (I think there are options to do so, but I didn't think that was default.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: How to paste with two buttons?
On Sat, Apr 07, 2001 at 04:48:23PM -0400, xsdg wrote: > On Sat, Apr 07, 2001 at 12:49:43PM +0200, Mikael Hedin wrote: > > xsdg <[EMAIL PROTECTED]> writes: > > > > How do I paste? > > > press SHIFT+INS > > > > Where do you learn all this magic? Is there some doc that teaches > > about these keyboard combos, in X or console? > IRC hearsay :o) There probably is a doc that contains this and other > magical key combos, but I have no idea what it is/how to access it. > the xterm(1) manpage documents the key/event bindings, including the cool stuff like control-rightclick = font select menu. The one line about the SHIFT+INS behaviour is: Shift Insert:insert-selection(PRIMARY, CUT_BUFFER0) \n\ so it's not surprising you heard about this from xterm hackers instead of seeing it yourself. Unless I knew what that line was supposed to do, I don't think I'd be able to figure it out without trying it or reading a whole bunch more docs, but reading the bindings list can give you an idea of what kind of stuff xterm has. use xkeycaps to find out what modifiers/keysyms your keys set/send. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Thinkpad video and digital camera issues
On Sun, Apr 08, 2001 at 10:06:54AM -0700, Heather wrote: > It does however seem logical if the PCMCIA interface card isn't being > spotted, that its further smartmedia item wouldn't be spotted either :( You could build hardware that worked as described. The interface card could stay inert and not signal anything the the host until a smartmedia card was inserted. That shuffling smartmedia cards would send regular PCMCIA events to the host, so they wouldn't need to invent a further level of events for smartmedia cards. No idea if that's how they do it or not, just a theory. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Query on CF cards..
On Mon, Apr 09, 2001 at 11:17:24AM +, anup neelanath wrote: > Hi, > Let me intoduce myself first. I am Anup from India. I am looking in the > net > for buying a digital camera - one with compact flash card. I got ur mail > address from one of the replies u had given for a query on compact flash > card. I hope u will be able to answer my query.. This is a mailing list for the Debian GNU/Linux operating system. (see www.debian.org). Your question doesn't depend on windoze software, so I know the answer, but this isn't really the right mailing list for it. > I have a desktop PC - windows, and no USB port. The cameras now adays > come > with USB cables. I saw that the CF card is IDE compatible. Does it mean that > I can directly connect the CF card to the IDE interface (as if connecting a > HD), or do I need to have some interface to connect it to the IDE cable. > You can't connect it to an IDE cable. You need a PCMCIA adapter. > I understand that if I buy a PCMCIA adaptor for CF card I can use it in > the > PCMCIA slot of a laptop and in that case also will I see the CF card as a > Hard disk? Yup, my friend's camera works that way. When you put the card in a pcmcia slot, it shows up as an IDE device with a normal FAT filesystem. It was /dev/hde, IIRC. (Actually, it might have been a RAM card with a battery, not a flash card.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: reiserfs?
On Mon, Apr 09, 2001 at 06:55:36PM +0100, Alexander Clouter wrote: > Anyone have any ideas about how a laptop's battery time is lowered by the > tree-balancing. This is not a problem for mains-computers however laptops > anything that uses CPU cycles (although they may normally be idle ones) > uses juice. It wouldn't be very fast if it used so much CPU that it consumed a noticeable amount more electric power. Rebalancing only has to happen when something else is writing, and usually rebalancing doesn't take much time at all. Every file system has some CPU overhead for management. Reiserfs doesn't do a lot of copying around blocks on disk, since that would be slow. Rearranging a bit of memory is really quite fast. Don't worry about the CPU load from reiserfs. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Different X settings for internal and external display
On Thu, Apr 12, 2001 at 12:37:52PM +0300, Torsten Reuss wrote: > Hi everybody, > > I am running potato on a Dell Latitude with an ATI M1 grapics chip and > XFree86 3.3.6 and the mach64 X-Server. X correctly activates either the > internal or the external display, depending on whether I am docked or not. > > So fine so far, the only problem I have is that 60Hz on the external Monitor > are pretty annoying and 75Hz on the internal LCD cause quite some flickering > an making me afraid that the display will start to burn. Unfortunately I > have not yet managed to enable X to automatically detect what monitor I am > using and use appropriate settings. If anybody else did and can send me > his/her XF86Config I'll be happy for that. Links to documentation on these > issues are also welcome... > You could do it manually, and make two XF86Config files. One for the monitor, one for the LCD. Set up startx to pass the appropriate -config option to X, and voila :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: dselect Trouble
On Sat, Apr 14, 2001 at 08:08:41PM +0200, Andreas Tscharner wrote: > Hello World, > > I have installed Debian 2.2rc0 on my new computer, and wanted to make an > upgrade. I use apt through dselect. In my sources.list (in /etc/apt/) I > inserted the german official mirror, made an update and entered "Install". > After having downloaded all packages, it makes a package scan, I there I > run into the following error message: > > 88% [Scanning packages]Template parse error near "" at > /usr/lib/perl5/Debian/DebConf/Template.pm line 102, chunk 2. > E: Sub-process /usr/sbin/dpkg-preconfigure --apt returned with an error > code (29) > .. > .. > > How can I fix this error? Any ideas? Are you upgrading to unstable or testing? I remember seeing something like this in testing a few weeks ago. Once you get all the packages installed, it should go away... You could turn of preconfig by commenting the line in /etc/apt/apt.conf. (It's not needed, it just lets you run the config scripts all at once, instead of sitting around while it gets around to asking you stuff every now and then...) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Kernel panic. No init found
On Sat, Apr 14, 2001 at 09:17:03PM +0200, Vincent Laisney wrote: > Hi debs! > > I have installed Debian Potato 2.2.19pre9 on my laptop. Everything was well, > but I had no APM support. I have tried to recompile the Kernel with > > make menuconfig > make-kpkg kernel_image > dpkg -i ../.deb > > But wenn I have rebooted I have got a message > "Kernel panic no init found. You can give at boot the option init= ..." What does it say before this? > > The worst is that I get this message for my old (and correct) Kernel, wenn I > boot with on the vmlinuz.old > > Yes, I have read in the Archives of this List that the solution is to add > apm=on on the boot line or add the line append "apm=on" in lilo.conf. It's a > pity to read this too late... lilo append lines are equivalent to putting the command line options in yourself. Just boot linux apm=on if you want. > Can somebody help me and give me the data to put after "init=" ? > Why can I not boot from my old kernel? Most likely the problem is the it's getting the wrong root directory. specify root= whatever it should be. It's not too likely the /sbin/init disappeared, unless something really bad happened. init=/bin/bash is another thing that can work, but no init= line will help until the kernel find the correct root filesystem. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Kernel 2.4.2's compilation on Toshiba 2800-100
On Sun, Apr 15, 2001 at 08:11:36AM +0200, Thierry leurent wrote: > Hello, > > I have a Toshiba 2800-100 with the debian potato. > I compile the kernel 2.4.2 and when i reboot the system, i receive the > message "Kernel decompression OK" and it freeze > > Do you have a solution ??? What config options did you use? Did you choose any different options from what you used in other kernel's you've compiled for that machine? -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Kernel 2.4.2's compilation on Toshiba 2800-100
On Sun, Apr 15, 2001 at 09:24:45PM +0200, Thierry leurent wrote: > On Sun, 15 Apr 2001 04:15:16 -0300 > Peter Cordes <[EMAIL PROTECTED]> wrote: > > > On Sun, Apr 15, 2001 at 08:11:36AM +0200, Thierry leurent wrote: > > > Hello, > > > > > > I have a Toshiba 2800-100 with the debian potato. > > > I compile the kernel 2.4.2 and when i reboot the system, i receive the > > > message "Kernel decompression OK" and it freeze > > > > > > Do you have a solution ??? > > > > What config options did you use? Did you choose any different options > > from > > what you used in other kernel's you've compiled for that machine? > > I take the original config of the potato, i unselect modules for non yamaha's > soundcards and choose PIII CPU in a first time after i use pentium MMX CPU. > It's possible that come from the version of glibc, gcc or from the Frame > buffer module ?? Try again without any framebuffer modules. You can just use VGA text mode on the console with x86 hardware. If you pick the wrong framebuffer, the kernel might boot but it won't be able to show anything on the screen. The GCC in potato shouldn't have any problems compiling the kernel. glibc makes zero difference. The kernel doesn't touch glibc. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: putting sound on irq 5
On Thu, Apr 19, 2001 at 01:10:32PM +, p wrote: > laptop debs, > > regarding my sony picturebook, how do i put the sound > controller on irq 5. currently, it's on irq 9 and will > not happily coexist with usb. Try messing with stuff in the bios setup. If that doesn't work, try using setpci to mess with them. (Be prepared for a crash if you mess up with setpci. I haven't used it on my system, so I don't know what kind of stuff to expect.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: hdd defrag
On Fri, Apr 20, 2001 at 11:14:35AM -0700, Jeff Coppock wrote: > I was wondering if it's necessary to defragment ext2fs or reiserfs? I do > this on a regular basis with my Windoze machines. > > What utilities could be used for this, if it's recommended? ext2 is smart, and does a good job of not fragmenting files. search the web, lots of good info is out there. Reiserfs is also smart like that. (BTW, it's the allocation algorithms in particular that are smart. They put new files where they will have room to grow without fragmenting.) If you have an almost-full filesystem, there won't be room to avoid fragmentation, so you should avoid that. ext2 is most efficient when less than 90% full, or so. (I don't know what a good cutoff value should be, but 90% is probably reasonable.) There is an ext2defrag or something around, so you can use that if you want to. Most people never defrag their FSes, and have no problems with them. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Help: Screwed up Partition + - Won't boot
On Fri, Apr 27, 2001 at 12:37:25AM -0400, [EMAIL PROTECTED] wrote: > - F > I can't imagine that you can't do a straight(startup/install) boot from the > CD, so going on from that assumption: > Boot frm CD into install > exit to shell > fsck primary linux part(hda2?) - if that's usable go on, if not your > DOA-SOL > make mount point ( I use /tmp/root or the like) > mount primary linux > From here on its more free form: > -, save what data you absolutely need to floppy > - might try undoing last step : or > free up some space the normal way (deleting stuff you can get along > without for a short while) especially stuff in /var/cache. dpkg --purge could help here, on big unneeded packages. Or, you could rm -rf some of /usr/share/doc, and reinstall the packages later. Unless you have filesystem damage, or you screwed up your partition table, it should be possible to restore everything to the way it should be. (I skimmed some of the original message, so I might have missed something.) As long as apt still works, you shouldn't have problems. debsums can help you track down packages with screwed up files. (using the contents of /var/lib/dpkg/info/*.list with some shell scripting can get you a list of packages with missing files.) BTW, after booting off a CD or floppy, chroot(8) is your friend. As long as a shell binary and libc are working, you're set. Another handy command is openvt(1). > maybe try copying the modules from the CD to the /target/lib/modules > dir for the 2.2.17 > Hope this helps. > Ed You can usually boot without loading any modules. Just get the kernel into /boot, edit /etc/lilo.conf, are run it. Reinstall the kernel modules binary package if you want to keep using 2.2.17, or just compile 2.4.3 (or 2.2.19) yourself. (I think there are packages of 2.2.19, if you want to go that route). happy hacking, -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: 4.0.2 trouble on the laptop.
On Fri, Apr 27, 2001 at 01:57:23PM -0400, Eugene Cheipesh wrote: > Hello, > > I'm having a strange problem on my laptop. it seems that the whole > screen is offset by about 3 mm upwards and at the bottom I have a black > and white line. Besides that everything works great! And clues would be > greatly appreciated, my searches have unveiled nothing of use. Try xvidtune. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: hdd crashed?
On Sat, Apr 28, 2001 at 09:50:44AM +0200, Joachim Schiele wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > On Friday 27 April 2001 08:47, you wrote: > some people here told me to start up in single mode > on lilo boot promt when you start up and lilo appears simply type single > then i have scaned my hdd for errors (physical ones i think) with that: > e2fsck -c /dev/hda3 ;-) this did the job i think > > at last the errors with don't appear anymore ;-) If it just failed randomly for no apparent reason the first time, you should expect it to happen again. Now would be a great time to upgrade your hard drive to a large, new, reliable one :). At the very least, start making regular backups if you weren't already doing so. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: ppp- and dns-problem
On Sat, Apr 28, 2001 at 10:17:21PM +0200, Vincent Laisney wrote: > Hi, > I have a problem for the configuration of ppp (pppd) with a combi-card > modem+ethernet. I use either ethenet or a ppp-connection with an ISP. > I have tried wvdialconfig for the configuration and i could not resolve > the > domaines names. I have also tried pppconfig (where it is possible to > specify > a static-dns-server. But it doesn't work. With > netstat -rn > I get no default line. > I have tried to set up directly a route with > > route add default gw 62.104.196.134 > > and I get the following error > > SIOCADDRT: Network is unreachable BTW, SIOCADDRT == Socket IO Control: Add Route. (io control = ioctl()) This error happens if the gateway you are trying to add to the routing table wasn't already reachable through some other entry in the routing table. i.e. on an ethernet, if you have eth0 with an IP address of 10.0.0.15, (netmask 255.0.0.0 (/24)), then your routing table before you add the default route will be: Destination Gateway Genmask Flags Metric RefUse Iface 10.0.0.00.0.0.0 255.0.0.0 U 0 00 eth0 If I ran route add default gw 192.168.0.1, it won't work, because the kernel doesn't know how to reach that host. That makes it pretty hard to start sending all our packets through it, doesn't it :) After running route add default gw 10.0.0.1, it will be: 10.0.0.00.0.0.0 255.0.0.0 U 0 00 eth0 0.0.0.0 10.0.0.10.0.0.0 UG0 00 eth0 If you're using ppp, you should just tell pppd to set a default route. It has an option to do this, called defaultroute. Just add that to your /etc/ppp/options, or /etc/ppp/peers/filename. If you want to manually add a default route with ppp, the gateway you specify _must_ be the ppp peer, i.e. the IP that ifconfig ppp0 shows as the Point-to-point IP. (If you already added a non-default route with the ppp peer as the gateway, you still have to use the ppp peer as the gateway. There is no way to ask a host that you can't reach directly to forward a packet for you, since there is no way to tell the first hop router who to forward to, other than the final destination. Think about it. Or don't if this isn't what you were trying to do, since it's not easy to understand unless you know networking :) Also, why were you talking about DNS in the start of your message, then switched to talking about routing. The route command has nothing at all to do with DNS. It will use DNS if you give host names instead of IPs, but that's all. The routing configuration affects whether DNS will work or not, because the servers you specify in /etc/resolv.conf must be reachable, but that's all there is in terms of interaction between routing and DNS. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE P.S. My last sentence is only true under the assumption that IP addresses are unique. More generally, you could route packets so they got to one 10.0.0.1 or a different 10.0.0.1. However, if that's the case, it shouldn't be, so you should change it :)
Re: WEP
On Fri, May 11, 2001 at 06:24:40AM -0500, tom wrote: > Can someone point me to something that will tell me how to set up > an encrypted WEP using only Linux? > > I am not even able to find much explaining the network ID > (SESS_ID). > I got something working with ADHOC and no security, but I'd like > to try for a named network and some security. > > Any suggestions? Suggestion: don't trust WEP. It's another layer of obscurity, but it's designed wrong and not at all secure. Some past articles on /. have links to papers by people who've broken WEP. I wouldn't trust it to keep my spare change secure. That said, it doesn't hurt to use it as long as you still treat it as an insecure network, and use SSH or IPSec for everything. It's another layer of stuff to deter a casual attacker. Just don't get a false sense of security. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: stupid kernel question
On Fri, May 11, 2001 at 09:44:17AM -0700, Sean 'Shaleh' Perry wrote: > > On 11-May-2001 [EMAIL PROTECTED] wrote: > > G'Day ! > > > > After I have install the kernel sources from a Debian package, can I use > > the patch files from kernel.org to bring it up to whatever release I want > > ? > > > > For example if I have installed a 2.2.17 .deb package , can I use the > > patches to bring it up to 2.2.19, or is it better to use a new .deb > > package ? > > > > the debian kernel MAY be subtly different from the released kernel.org kernel > so patches may not apply cleanly. I have also found that by using patches to > upgrade I sometimes missed things or had cruft left behind. No one likes the > 20mb downloads, but it is how I do it. It's annoying that there doesn't seem to be an rsync mirror of the kernel tree, since one could use that after applying patches to make sure everything was in sync... The PowerPC kernel has a couple of rsync trees available, and it's really handy. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: stupid kernel question
On Fri, May 11, 2001 at 01:09:25PM -0500, Craig T. Milling wrote: > > On Fri, 11 May 2001, Peter Cordes wrote: > > > > > It's annoying that there doesn't seem to be an rsync mirror of the kernel > > tree, since one could use that after applying patches to make sure > > everything was in sync... > > > > for clarification, www.kernel.org lists an rsync tree > > rsync://rsync.kernel.org/pub/ > > are you lamenting the lack of rsync://kernel.mirror.(com|edu|org|net)/ ? No, I already knew about that. I meant "tree" as in the source tree for a single kernel version, unpacked, not a tree of kernel versions. They let you rsync a mirror of kernel.org's collection of tarballs. What I want is a source tree I can rsync my untarred kernel source against. This would be useful after doing the following: tar xjf kernel.tar.bz2 patch < foo.patch patch < bar.patch patch < baz.patch # mess with some stuff, and come back weeks later so you don't remember # exactly what patches you've applied, so you can't patch -R. rsync -avz --delete rsync://rsync.kernel.org/pub/linux-2.4.4 linux # now your source tree is exactly what you'd have if you'd downloaded a # 2.4.4 tarball and unpacked it into a clean directory, but you didn't have to # take all that bandwidth to do it. PowerPC users will know what I'm talking about, because the PPC developers make their kernel trees available. See http://penguinppc.org/dev/kernel.shtml for how this works. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: SpeedStep / Geyserville lockups?
On Mon, May 14, 2001 at 02:28:09PM -0700, Karsten M. Self wrote: > on Mon, May 14, 2001 at 07:08:40PM +0100, Alexander Clouter ([EMAIL > PROTECTED]) wrote: > > On Sun, 13 May 2001, Karsten M. Self wrote: > > > > > > Is anyone else out there seeing lockups under 2.2.x (2.2.18 here) with a > > > PIII SpeedStep (aka Geyserville) chip? > > > > > > I've had lockups over the past week I suspect are due to CPU step-speed > > > changes resulting from power source switching. > > > > > isn't there an option in the bios to banish this 'feature'? I remember > > there being reference to this in a couple of places. Once disabled I > > guess you could solely use 'CPU Idle' calls to cool your cpu, which I > > assume has the same *effect*. > > There is. > > I've set CPU speed to "maximum" (forget the precise opts, believe they're > max, min, auto, reverse). System's been up for 19 hours w/o problems, > previously I was lucky to get an hour. > > Note also that SpeedStep is probably more useful under an OS which > doesn't have preemptive multitasking. My understanding of Win9X/ME was > that it ran at full tilt unless specifically idled, this information > coming from VMWare experiences. NT/2K and GNU/Linux have an idle cycle > which actually idles the CPU. > > What's the CPU Idle call you're referringn to? That's the APM idle, enable it in the BIOS config so the kernel will tell the BIOS when it is idle. The ordinary idle you're talking about is when the kernel runs the halt instruction, which does indeed keep the CPU cool (and save a lot of power if your CPU takes most of the power in the system) but the BIOS will never know to go into sleep mode. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: SpeedStep / Geyserville lockups?
On Tue, May 15, 2001 at 01:01:29AM -0700, Heather wrote: > > > > > Is anyone else out there seeing lockups under 2.2.x (2.2.18 here) > > > > > with a > > > > > PIII SpeedStep (aka Geyserville) chip? > > > > > > > > > > I've had lockups over the past week I suspect are due to CPU > > > > > step-speed > > > > > changes resulting from power source switching. > > I believe that Speedstep machines are much more stable if they are always > (re)booted while attached to wall power. Suspends and resumes should behave > much better, but *those* depend on whether your APM BIOS is crappy :( Yes, that's what the guy who maintains APM in the kernel told me after his talk at OLS last year. I told him about my laptop that lets you switch between 33 and 66 MHz with a keypress, and he said to always boot up at high speed because waiting too long is a lot safer than not waiting long enough when it comes to the micro-delays that are timed with bogomips. > I personally suspect that some things may not be all that well behaved "at > the wrong speed" too but I have no direct experience with that. I would be surprised unless you make your clock speed slower by more like an order of magnitude. (depending on the drivers you are using, of course). Most of the time, the kernel will wait only the minimum delay specified by the hardware. If there is a maximum delay, it will be much longer than the minimum, so waiting longer won't hurt. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: SpeedStep / Geyserville lockups?
On Wed, May 16, 2001 at 08:44:12PM +1200, Andrew McMillan wrote: > > I personally suspect that some things may not be all that well behaved "at > > the wrong speed" too but I have no direct experience with that. > > That's really interesting. I haven't had any lock-ups on this system at all > (I > finally got an IBM A21p). I have seen bogomips ratings of between 187 and > 1690. > > Perhaps the BIOS manages to sort this out for me? No, linux doesn't use the BIOS after it boots up. > > Suspend / resume have also worked flawlessly, apart from the (apparently) > standard > sound driver bugs, requiring ALSA to be restarted on resume. You're just lucky. Some drivers must not need accurate delay timing, and you must not be using any that do. Give your hardware a hug :) It's always a good idea to boot up a full speed, so detected bogomips = max bogomips, but if you've _never_ seen any lockups or data errors/corruption, you probably don't need to change now. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: thinkpad ps/2 mouse under X with gpm
On Wed, May 16, 2001 at 04:32:10PM +0100, Vivek Dasmohapatra wrote: > On Wed, 16 May 2001, Wenzhuo Zhang wrote: > > > > Im running debian 2.2 on a IBM Thinkpad i1200. The kernel detects a ps/2 > > > mouse port. I installed gpm, wich works on the console. > [cut] > > > I tried several setting in XF86Setup and xf86config to get the mouse > > > running under X. The only thing that works is > [cut] > > Try > > Protocol"PS/2" > > Device "/dev/psaux" > > No. Only one thing can grab the psaux device at a time, Yes. > so at best this > will be flaky. No. >You may need to pass extra options to gpm, or repeat it in > a different protocol. Recent gpm versions (or maybe it's done this all along) will release the mouse when you're not in X. I've got a couple systems with PS/2 mice, and gpm/X don't have any problems interacting. Both X and gpm are reading from /dev/psaux. If you don't believe this, then use strace on GPM, and observe that it close(2)s the mouse device when you switch back to X, and reopens it when you switch to a console. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Lost mail
On Wed, May 16, 2001 at 01:14:51PM +0100, Dave Swegen wrote: > Last week I received some mail regarding debian/linux on thinkpads. > However, since my machine went belly up over the weekend I have lost > these peoples email addresses. Therefore, if you did mail me regarding > thinkpads, and haven't received a reply, please mail me again, and I will > try and help you out. > If it was on the list, check the archives at http://lists.debian.org/ -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: thinkpad ps/2 mouse under X with gpm
On Wed, May 16, 2001 at 05:17:45PM +0100, Vivek Dasmohapatra wrote: > On Wed, 16 May 2001, Peter Cordes wrote: > > > Recent gpm versions (or maybe it's done this all along) will release the > > Judging from observer behaviour, I would guess that it has not done this > all along. > > > mouse when you're not in X. I've got a couple systems with PS/2 mice, and > > gpm/X don't have any problems interacting. Both X and gpm are reading from > > /dev/psaux. > > Fine. Is this the case in potato? Or was it introduced in the woody > package? Somebody with a potato system try this out. strace gpm and see what it does, just to be sure. > > If you don't believe this, then use strace on GPM, and observe that it > > close(2)s the mouse device when you switch back to X, and reopens it when > > I have no reason to disbelieve you, good :) > but until such time as all psaux > grabbing software is this considerate, it's still a good thing to know, > yes? Right. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: my dell runs
On Thu, May 17, 2001 at 06:52:58AM -0700, Ignasi Palou-Rivera wrote: > --- Matthias Weiss <[EMAIL PROTECTED]> wrote: > > The simple solution was: I had no window manager installed 8-() > > Further I had to disable gpm mouse service otherwise I couldn't use > > the mouse under X. > > Same thing happened to me last week when I installed Potato/Woody on a > TP 755 Cs. I'm sure there must be good reasons, but it's strange that > no window manager is required with the core X packages. Not really. You could run a box as an X terminal, and run the window manager on a remote machine, especially if you use something like XDMCP. (X -indirect hostname). Of course, I don't think having lwm installed (62kB installed size) would kill anyone :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: lilo display
On Thu, May 17, 2001 at 01:40:26PM -0400, Steven K Thompson wrote: > I did not get any response when I asked this query before. It > would be great if someone knows how to get a more straightforward > display of the available operating systems at bootup, especially > if it is someone besides myself who turns the machine on. > > > When I installed Debian 2.2r2 on my laptopI configured Lilo to give a choice > of > three operating systems: > 1. Debian, 2. Redhat, and 3. Windows/Dos. Debian is the default. > > When I press the tab key at the boot prompt the display of the options is > rather > chaotic > in appearance, like this: > > > debian 1 2 redhat > 3 dos > > > Is there a way I can change this into a simple list of the three options, with > the numbers and > labels in the same order for each? I would appreciate any advice here. > First, get rid of all the comments you don't need in lilo.config. There's a man page for it, so you don't need the docs for stuff next to the items, unless you usually find that you don't remember what they do without checking. Anyway, get yourself an up-to-date version of lilo (potato's lilo might not have boot-menu), and use this: # bootloader options boot=/dev/hda lba32 install=/boot/boot-menu.b # instead of boot.b compact # turn this off if it won't boot with it. # UI options delay=20 prompt timeout=100 # kernel options append="apm=on" root=/dev/hda4 read-only # images default=debian #default=redhat image=/vmlinuz label=debian root=/dev/hda4 image=/redhat/redboot/vmlinuz-2.2.16-22 label=redhat root=/dev/hda5 other=/dev/hda1 label=dos # a couple fun ones I like to have other=/usr/lib/hwtools/memtest86.bin label=memtest86 image=/usr/lib/hwtools/memmxtest.bin label=memmxtest other=/dev/fd0 label=floppy unsafe # don't try to read floppy when installing bootloader A lot shorter and easier to read, eh? :) (BTW, if /sbin/lilo complains, _be_ _careful_. Don't set your computer on fire with this, etc. :) If you don't install a lilo with boot-menu.b, then you're probably best off following the advice of the other people who've suggested things. (GRUB, or using a boot-message). -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Xircom and Potatoe
On Thu, May 17, 2001 at 07:21:14PM -0300, Edson Akio Maruyama wrote: > At 16:38 14/05/01 +0100, you wrote: > >Hi > > > >Hopefully someone can put me out of my misery with this card FYI, common usage of the expression "put something out of its misery" means to end the misery by ending the things life! E.g. shoot a horse with a broken leg, or throw out a half-working computer. I don't think that's what you had in mind, but the packet of death will be arriving shortly ;-) > ifconfig eth0 111.222.333.444 up netmask 255.255.255.0 > > route add default gw 111.222.333.1 dev eth0 333 and 444 don't fit in an 8 bit integer, so they are impossible. The private nets like 10.0.0.0 are good for examples, since people using them instead of the actual numbers won't break the internet :) Today's lesson has been brought to you by Mr. Pedant. :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Inspiron 8000 & Progeny install woes
On Fri, May 18, 2001 at 05:28:45PM +0100, Vivek Dasmohapatra wrote: > On Fri, 18 May 2001, Jim Richardson wrote: > > > Which video card would you recommend for this laptop? (I am about to buy > > one) the GeForce or ATi? > > iirc, nvidia require you to use a binary only driver, which means you'd be > pretty much on your own if you had kernel related problems, as nobody > outside of nvidia would have the info needed to help you. I could be > wrong though. There is a Free driver in XF86 called "nv" that supports GeForce chips, but it doesn't do 3D accel at all AFAIK. (If it does do any 3d stuff, it's very likely a lot slower than the closed-source drivers from nvidia.) The binary-only drivers work well (They have the same codebase as the windoze drivers, so the OpenGL implementation is actual certified as OpenGL, officially.) , but really suck from the point of view of running a Free system, since there's a kernel module as well as an X server module and OpenGL libraries. The kernel mod comes with some wrapper code to interface with the rest of the kernel, so you can run whatever kernel you want and recompile the module as needed. I've got a GF2 MX in my desktop machine, and it kicks ass, but requires all that proprietary software :( -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: install on Toshiba 430CDT laptop
On Fri, May 18, 2001 at 05:52:29PM -0300, Edson Akio Maruyama wrote: > At 12:57 16/05/01 -0700, you wrote: > >I have a Toshiba 430CDT laptop, and I installed the > >base system of Potato on the machine, which seemed to > > [diet quote] > > i?ve gotta had the net adapter "manually" configured... through the > following commands... > > ifconfig eth0 111.222.333.444 up netmask 255.255.255.0 > route add default gw 111.222.333.1 dev eth0 > > where, those funny numbers... 111.222 are NOT real IPs... and then > you should place in your real numbers... :) right peter? :) yeah :) Note that 111.222.333.444 is one IP address, not IPs (plural). (Well, it would be an IP addr if 333 and 444 were less than 256 :) > hey, ALL.. need some help on this... i could make the net adapter work, > BUT... i?ve gotta do this set up when i log on... how can i do this > automatically? tried to save some "bash" file as executable... but it > didn?t quite work... i dont know why.. but the "graphical" setup doesn?t > quite work either... and it keeps telling me i?ve got no NIC... : Ok, what you need to do is find out what your IP address is, and configure your network card for that. You can put something in the pcmcia scripts to configure the network when you insert the card. To find out your IP, either ask your admin, or run a DHCP client like pump. (I think people were saying they were having problems sending/receiving traffic after running pump, but that it still found an address for them.) After running pump, run ifconfig eth0, and note the information there. You need: the IP address, the netmask, and the broadcast address. Run route. Note the address of the default gateway. Once you have this, put ifconfig eth0 111.222.333.444 up netmask 255.255.255.0 broadcast 111.222.333.255 route add default gw 111.222.333.1 somewhere where it will run when you insert the network card. (not sure where that might be ...) (BTW, I think someone said that running ifconfig eth0 up after running pump did something. That might be totally bogus, though... If it works, it's a nicer solution, since it will work whatever IP you want to set. (Not as nice as fixing the broken kernel driver, though.)) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: New ... please help
On Fri, May 18, 2001 at 10:12:20AM -0600, Razvan Roncov wrote: > hello... > I just installed Debian on my laptop ... but I have 3 "little" (I hope they > are little) > problems: > 1. My network card was not installed Surecom (pcmcia slot) Is that a Surecom EP-427X? If so, do you have time to help me get Linux to support it? I've got an EP-427X myself, but I'm not sure if its my (old) laptop that's the problem, or the driver. (The driver disk that comes with the card includes source that is a small change to a really old version of pcmcia-cs's ne2k driver, for use with a 2.0 kernel. I think I've figured out what Surecom added, but it doesn't seem to work on my laptop :( > 2. My resolution for XServer all though I set it to 800x600 ... it still > displays a biger resolution >therefore I cant see the whole desktop Edit /etc/X11/XF86Config and change the order of the Modes line, in the Screen section that applies to your video card. > 3. does Debian support USB mouse ... if yes ... how could i configure it ... ? > I would appreciate any advice I can get .... Don't know about that one... -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: X keymaps
On Sat, May 19, 2001 at 03:29:03PM +1000, Martin wrote: > Hi guys > > i have debian up and running on my portege 7020ct but the keymap is really > screwed up... > > ie 'y' and 'z' swapped over, half the symbol keys produce different symbols, > backspace or other funky control characters... > > is there any good resources on 'Xresources' fixes for things like this? a > place with a how-to or example configs would be fantastic... You can pick a different XKB keymap by editing the section in your /etc/X11/XF86Config. XkbModel and XkbLayout are the ones to look for. (This applies to XFree86 3.3.6; X 4.x uses XF86Config-4 in Debian.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: kitchens (was your mail)
On Wed, May 23, 2001 at 03:29:47PM -0700, Heather wrote: > > The keyboard of Alexander Clouter <[EMAIL PROTECTED]> wrote: > > >> [something about tacky kitchens] > > > > > > I'm not sure if I can see a .deb package for it.. > > > > Sorry, we working on the task-kichen package. > > > > Maybe u should try fridge-0.1.deb and fridge-contents-0.1-be(er)ta.deb > > first. > > > > CC. --- a bit silly today. ;) > > Found in the console-apt grocery store: > > ale-clone > blender > cucipop > cup > filters-nonfree > several types of fortune cookies > gifsicle > numerous varieties of java > lgrind > muffin > mush > plugger > quite a bit of squid > a wip > more shells than you can shake a stick at. > > Now I've got a debtop range, you've got someone working on fridge, all > we need is a sink ;P Got that: It's called emacs. (On win32 it even uses a kitchen sink icon.) When we serve the meal, we can use xserver[-xfree86] to hold the plates :) I don't know what to make of sawfish or bluefish. At least we don't have green eggs and ham. Fortunately, bsdgames includes ordinary fish :) And don't forget to say grace before digging in, as console-apt is now renamed deity. :-> -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: kitchen
On Thu, May 24, 2001 at 01:27:03PM -0700, Heather wrote: > > On Thu, 24 May 2001, Daniel E Baumann wrote: > > > > [cut wine package listings] > > > > > However, I do not recommend using non-free programs :). > > > > Huh? Wine is in main. Not non-free. > > > > BTW, you'll also need the libwine package to install wine. > > I believe that he means, wine would usually be most useful to run some > -very- non-free programs. > > Nonetheless, I have a *huge* mswin software collection, much of it shareware > and freeware, and a few non MS licensed apps that could be useful, but I'm > completely chicken about Unstable for the moment. Problems with unstable are mostly in dependencies and pre/postinst scripts, and other minor packaging errors. Of course, if software you use a lot has a version in unstable that is actually an alpha or beta version, but a stable version is in woody, then you might not want to run all unstable. What I do is run woody, but with the unstable repositories in my sources.list, and APT::Default-Release "testing"; in my apt.conf. (see apt_preferences(5), etc.) This way, apt-get install package gets the package from woody, unless it only exists in unstable. apt-get install package/unstable gets the unstable version. apt-get -t unstable lets apt upgrade the dependencies to their unstable version if necessary. I've found that running an unstable libc is not actually as bad as it sounds. It's a really important package, so Ben makes sure he doesn't break it :) Of course, I wouldn't be running unstable packages if I didn't think I could hack my way out of any problems, so it's not for novices, but having a few packages from unstable is not a big deal, even if they are important ones. If all my packages were the unstable versions, I would probably see a lot more bugs caused by packaging errors than I do currently. Running the unstable versions of packages you are actively interested in is a good way to contribute to debian, and takes a lot less effort than hoping that lots of people will test everything by running unstable across the board. (Notice that I managed to argue that my position is the Right Thing to do. Err, I didn't quite mean to do that, but whatever. :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: kitchen
On Fri, May 25, 2001 at 12:22:28PM -0700, Adam Shand wrote: > > [apt_preferences(5) stuff] > this is really cool. guess it's been a while since i read the apt man > page! It's not mentioned in the apt man page, and there is no reference to apt_preference in them :(. I only found it because I was looking at the BTS page for apt for another thing, and I saw a bug about apt_prefereces not being referenced. So, I thought, "Oooh secret documentation, cool :)". > i've wanted this for so long ... I hadn't really thought about it, since I'm relatively new to Debian (about a year), and I've been busy with school during term. I was pretty happy to discover that feature, though :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Flaky HW: keyboard & power (Compal/TuxTops 20U)
On Sat, May 26, 2001 at 05:02:27PM -0700, Karsten M. Self wrote: > It's that trusty TuxTops box again. > > Two things I've noticed in the past few days: > > - Slight adjustments of the power cord cause the box to switch in and > out of battery / AC modes. This is accompanied by a high-frequency > squealing sound, and logs APM events to system logs (which I monitor > via xconsole on xrootconsole. > > - The built-in keyboard appears to have cut out. I can plug in a > Happy Hacker keyboard into the side port and enter commands. On > rebooting, the system didn't respond to keypresses. > > I'm inclined to pop the box open and look for any visibly loose > connectors. Not sure what the proscribed surgical techniques are, the > accompanying manual doesn't address disassembly. > > Thoughts? Check for traces lifting off the PCB near the solder joint where the connector is attached. A pair of speakers had the problem with the AC power connector, where moving the connector would make or break the connection. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: PCMCIA Modem Driver?
On Thu, May 31, 2001 at 03:57:01PM +0100, Daniel Faulkner wrote: > Hi, > Recently a friend asked me 2 install linux on his laptop and > most of it works well. > (touchpad mouse was suprisingly easy) > Sound, looks hard but from the sound of things has been done > before and something I haven't had much of a go at yet. > PCMCIA bus/controller thing I think is OK > But the actual cards are the problem. There is a DVD card (not > looked for a driver yet) > And a modem (which is causing the problems) > > The modem is: > PCMCIA modem, 3COM Megahertz 56k Global WinModem PC Card Model > 3CCM656 It says WinModem right on it, which means that the DSP functions are handled by the CPU, not by dedicated hardware. You _WILL_ have problems, since most manufacturers don't tell anyone how to get the CPU to do this, they just release windoze drivers. > Laptop is: > Gateway Solo 2500 > OS is: > Debian Linux, Potato (Disks say 2.2.r3 on front) Kernal 2.2.18 > (I think if I remember right) > > So far what I've seen makes it look like a new modem is needed > for certain but as linux is continualy updated I would like to > know if there is a way to get the modem to work in linux. Try www.linmodems.org, though. They've got something going, but I don't know if it's useable or supports your card. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: power button dell inspiron 4000
On Sat, Jun 09, 2001 at 12:06:24PM +1000, Tim Connors wrote: > I've given up on this project for a while, but here's some more info: > Flashed with the latest BIOS, but pretty much everything is the same in > the setup menu (gee, I like being able to change bios options when the > machine is already running an OS!). I booted winbloze, and found out it > doesn't even disable the power button (good chance to run scandisk, > anyway :) > > So I think I will ask Dell for some help and/or for them to add yet > another thing to the BIOS. > > Incidentally - when the machine hard-crashes you need to press the power > button for 4 seconds to turn it off. So somewhere, it is asking the OS for > permission to turn off, and doesn't get through to it when it is crashed. No, it isn't asking the OS. My computer has a BIOS option for making the "power button" switch be either suspend or soft-off. When it's set to be suspend, it generates an APM event when you push it. In this mode, you can also hold it down for 4 seconds to turn the computer off, accomplishing the same thing that pressing it at all does in soft-off mode. > I'm slightly confuzzled by this Try reading the bios manual, if you have a copy of it. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Raunchiest SEX VIDEOS and NO CCARD required !!!
On Wed, Jun 13, 2001 at 05:18:17AM -0700, Dhruva Reddy wrote: > Let's try that again... > > I second that. My local LUG does this, and I have > never seen spam on their list. nslug.ns.ca does that, but I think the admins have it set up so they see stuff sent from non-subscribers, and can let it onto the list if it's not spam. When I first saw this spam, I thought "wrong kind of lap-top action"... :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: moving to a new laptop
On Tue, Jun 19, 2001 at 07:03:17PM +0200, Philipp Bliedung wrote: > Hi > > I'm going to move to a new laptop pretty soon. My question: > Can I simply "move" all the software to the new one? (Well, except for > the XConfig and other hardware stuff like sound, 3D, etc...) > My biggest concern is that I have to install all the packages again that > are on my laptop rigth now. What about programs that I compiled myself - > do I have to compile them again? You can move whole filesystems with dump | netcat, netcat -l | restore. If you want to just move a subdir, use GNU partimage, or use tar | netcat. 100baseT is your friend :) > I have a lot of packages from woody and sid and I really don't feel like > getting them again (I only have potato stuff on CD) - I really want to > avoid that, especially because it took me a long time to configure > everything. Everything in /usr except /usr/local should be restorable give the info in /var. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: moving to a new laptop
On Thu, Jun 21, 2001 at 10:11:56AM -0700, Kelsey Jordahl wrote: > >>>>> Jeff Coppock writes: > > >The folks in my IS dept. at work use a program called Ghost > > (I think its from Norton) to copy an entire partition, like an > > image, from one HDD to another. They use it for HDD upgrades > > all the time. I even moves the Partition table. So uprading > > from a 6GB HDD to a 12GB HDD, you'd end up with a 6GB partition > > on the new drive. Then you can go in and add more partitions > > with the extra space or expand the one partition, or whatever. > > >I don't know what other products like this that are avaiable, > > or what cost might be associated with it, but maybe there's > > something like it in Debian or for Linux in general. > > I haven't used partimage, but plan to try it the next time I have this > kind of problem to solve: partimage is great. It does exactly what Jeff says Ghost does, and does it for FAT and ext2. (maybe others?). If you're just doing ext2, then you might as well use dump, unless you really like ncurses GUIs like partimage has... dump more or less does what partimage does, with fewer features, but is easier to redirect i/o to/from. Both of them write only the actual data, so a partimage or dump of a 10GB FS with 500MB of data on it will only take up about 500MB, not the whole 10GB. I have personally used partimage while shuffling around a FAT partition, and it worked great. Absolutely no problems. partimage was solid, which is what you want with something as important as all your data! -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: LCD sync rates
On Thu, Jun 28, 2001 at 10:59:28AM -0700, Heather wrote: > > and the slightest variation can result in a black screen when your start > > the X server. > > I recommend having your net connection live and be ssh'd in from another box. > That keeps you a visible text session. > > You can sometimes use a different GUI utility (e.g. SVGAlib app, or vga_reset > or something) to reinit the video and keep working without having to reboot > to yank its chain. CTRL+ALT+backspace or CTRL+ALT+F1 still work even when the display doesn't want to show you anything. If the X server locks up, you can always use alt+sysrq+u, alt+sysrq+b to umount/remount-readonly all your partitions, then reboot. This is handled very early on in the kernel, so it doesn't stop working no matter what user space software does, unless the kernel itself actually locks up, or the keyboard physically stops working. Of course, rebooting takes time, so if ctrl+alt+backspace doesn't work then logging in remotely with ssh or a null modem before messing with stuff will save time. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: LCD sync rates
On Thu, Jun 28, 2001 at 08:21:15PM -0700, Heather wrote: > > On Thu, Jun 28, 2001 at 10:59:28AM -0700, Heather wrote: > > > > and the slightest variation can result in a black screen when your > > > > start > > > > the X server. > > > > > > I recommend having your net connection live and be ssh'd in from another > > > box. > > > That keeps you a visible text session. > > > > > > You can sometimes use a different GUI utility (e.g. SVGAlib app, or > > > vga_reset > > > or something) to reinit the video and keep working without having to > > > reboot > > > to yank its chain. > > > > CTRL+ALT+backspace or CTRL+ALT+F1 still work even when the display > > doesn't want to show you anything. > > Not necessarily. The X server is also responsible for input focus - one might > argue that's its primary job - and if it's *really* unhappy, it won't get > around > to your useful keystroke. Hmm, good point. That makes sense, given my past experience with X... > Too busy dealing with a crying vidcard. Maybe next > week sometime. > > Meanwhile your monitor is squealing at you :( :( :( The X server doesn't know that the monitor isn't syncing, since the vid card doesn't act any differently whether the monitor handles the video output or not. (I think it can detect whether or not a monitor is connected, though.) I don't know if this is the case for most LCD displays in laptops, or if the video controller will stop working if the vid mode is out of range. That wouldn't make much sense though, because that would make it hard to get back to a working range... > > > If the X server locks up, you can > > always use alt+sysrq+u, alt+sysrq+b > > Magic sysrq's might only work at a console prompt. cf above, your keyboard > may be inoperable. No, they always work, since it's all in the kernel, before it goes through the "keyboard mode" stuff that kbdmode messes with. That being the case, it doesn't matter if X is running and has your keyboard in raw mode. You obviously don't see the output from your keystrokes, but they show up in the kernel log messages, even alt+sysrq+?. Err, one other thing that can bring down alt+sysrq: The system could deadlock with interrupts disabled, in which case you lose, and nothing will ever get through, not ethernet, not mouse, not keyboard, not serial. I don't think X can disable interrupts, though, so as long as the kernel isn't buggy, you're fine (he says optimistically... :) Anyway, we both agree that logging in remotely is the way to go. I was just pointing out that, especially for this problem, you can get away with not doing that if you can deal with the system when it's having problems. (For the display-won't-sync problem, nothing is going to happen that overwrites the kernel in memory or whatever, so it's unlikely that the system will hang and require a reboot.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Debian into an existing Loopback file system
On Thu, Aug 02, 2001 at 10:20:16AM +0200, A. Demarteau (linux rules!) wrote: > on my laptop I use a loopback-file for the /usr partitions. > This is however, very slow. > I think you would be better off using the umsdos filesystem (which hapily > co-exsists with windows on the same disk-partition without troubles). I don't think loopback adds that much overhead. I would stay away from umsdos. I think using an initrd would let you boot with the root fs on a loopback device, which is good because kernel-image-2.4.7, etc. packages use an initrd anyway. Loadlin knows how to load an initrd for a kernel, as does lilo. I would use chroot to get the install done. There are distros that are designed for people who want to give linux a try without repartitioning. I tried one a couple years ago, and it had a windoze-based installer that created the image file, and set up loadlin with an initrd thing. If you get stuck, just look at how they do it, or just use one of those and morph it into Debian once you've installed :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: Debian into an existing Loopback file system
On Thu, Aug 02, 2001 at 10:11:20PM -0700, Curt Howland wrote: > > There are distros that are designed for people who want to give linux > > a try without repartitioning. I tried one a couple years ago, and it > > had a windoze-based installer that created the image file, and set up > > loadlin with an initrd thing. If you get stuck, just look at how they > > do it, or just use one of those and morph it into Debian once you've > > installed :) > > This is exactly what I'm talking about. Dragon Linux in a > one-gig loop-back "image file". > > And I would like to "morph it into Debian", but I'm not a > programmer. BTW, you'd want to be a sysadmin for this job, not a programmer. I guess you're not one of those either :( Lucky for you, you can just ask us hackers on the list :) > I don't know all the ins and outs, thus my > original statement that I know the best way to change > distributions is to wipe out the old one and install the > new one. Well, I would like to keep the present launch > process. That shouldn't be a problem. You should be able to keep using the kernel that's currently installed by copying the modules for it into the new Debian /lib/modules, and /boot/System.map-* to. (the kernel image itself is probably on the FAT volume, not in the image file. upgrading the kernel could be trickier. I don't know if the initrd that the kernel you're using would continue to work with Debian kernels. Solve that hurdle when you come to it. > > Debian web page and docs have no "neat loop-back image file" > type of install described, just second partition style. > > Has anyone done this, changing from an non-Debian to a > Debian "distribution" without actually erasing anything? Our LUG web server migrated from redhat to debian this way. They had Debian in a chroot, but I don't know exactly what they did after that. Probably just moved aside /usr, and moved /chroot/usr to /, the same for the other tier-1 directories. Make sure you don't break shared libraries while you move /lib... Booting off a floppy is the safe way to do this. Hmm... a challenge. This might do it: # /chroot is the debian installation mkdir /old mv /[^c]* /cdrom /old LD_LIBRARY_PATH=/chroot/lib /chroot/bin/mv /chroot/* / run lilo, whatever. (you probably can run lilo out of an image file (if lilo can find the disk address of the files it needs through a loopback device), but you'd have to rerun whenever you defrag and the image gets moved. As for setting up the chroot env. in the first place, install potato by untarring base2.2.tgz inside its own directory. chroot in and run /sbin/unconfigured.sh. (copy stuff from your existing /etc as necessary. Probably you should cp -a /etc /chroot/old-etc, so you will have everything in case you want it.) Woody would be be a little trickier, because it doesn't use a base tarball, so the install options aren't as flexible :( I would recomment installing potato, and apt-get dist-upgrade from there after editting /etc/apt/sources.list (and apt.conf). -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE
Re: dhcp, cable modems...
On Thu, Mar 01, 2001 at 06:27:22PM -0800, Jason Victor wrote: > I'm trying to configure my cable modem to work with > Linux. I'm using Optimum Online. I have a couple of > questions. First of all, is my ethernet card working > if i type dmesg and see something about "eth0: Xircom > CardBus." That is the NIC i have. second, i've been > trying to connect with dhcpcd, and can't figure it > out. i set up dhcpcd to work on eth0, and type ifup > eth0, and it says "ignoring unknown interface eth0." > i'm stuck. by the way, it sits at the starting dhcp: > thing for a long time. Run ifconfig -a to see if eth0 exists or not. If it does, you just need to get the right stuff to use it. If it doesn't, then you need to config your pcmcia and modules stuff to set it up. You might want to try doing things by hand, i.e. running pump -i eth0 (dhcpcd is obsolete; It's been removed from Debian because it is irreparably full of bugs and security holes. It still works, though, and it has a similar command line option. Note that pump will log everything it does to syslog. Take a look if anything goes wrong.) You say that dhcpcd hangs for a long time. Try running tcpdump while you run dhcpcd. (This may not work too well, since the interface will be down when you start, unless you assign a bogus address to it. Also, dhcpcd will take down the interface tcpdump is sniffing on...) See if you're getting any response. If pump (or dhcpcd) works, then your problem is in configuring the high level tools like ifup. I personally never bother with them on my desktop machines, and rewrite /etc/init.d/networking to run ifconfig, route, and ipchains the way I like. (I haven't finished hacking the pcnet-cs driver to support my Surecom EP-427X for my laptop :(. However, if you just want something simple that works, stick with ifup I guess. > i am using storm linux (basically debian 2.2) is there > any way to use the storm administration system to do > this? there is an option to set up eth0 with dhcp. i > need 2 give it a hostname... where do i get that? DHCP can pass a hostname to the DHCP server. Rogers@Home in Ontario uses this: they give you a host name to use when you sign up, and you have to supply that for their DHCP server to answer. Other providers, like Eastlink in Nova Scotia, don't require this, and just check the ethernet MAC address. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: dhcp, cable modems...
On Thu, Mar 01, 2001 at 11:40:51PM -0500, MaD dUCK wrote: > also sprach xsdg (on Thu, 01 Mar 2001 11:29:05PM -0500): > > > (dhcpcd is obsolete; It's been removed from Debian because it is > > do what? i missed something. oh, and dhcpcd is not the dhcp daemon, > but the dhcp client daemon, Yes. > and it's so much better than pump... I don't find it better than pump. Last I checked, it does dumb stuff like giving up after a while and leaving the interface down. If my cable modem provider is doing an upgrade so there is a temporary outage, dhcpcd would (after an hour or so) give up and bring down the network connection, leaving me disconnected even after my cable modem came back online. I don't know how pump would handle the situation, since I've stopped using dhcp except to obtain an IP config initially. My ISP doesn't capriciously change my IP address, so I statically config my interfaces. (Sure, this breaks things when the IP does actually change, but I want that, because in that case I need to change the DNS A record for llama.nslug.ns.ca so my email will come in, so I want large breakage right away that will catch my attention :) See http://lists.debian.org/debian-powerpc-0102/msg00097.html and http://lists.debian.org/debian-powerpc-0102/msg00126.html to find out what the current state of DHCP in Debian is. As you can see, there is a dhcp-client package. It might float your boat better than pump, if you don't like pump. It's interesting to note that even though apt-cache search dhcpcd only finds pump, (i.e. dhcpcd is not part of woody anymore), you can still download the dhcpcd package through packages.debian.org/dhcpcd, and it's on the mirrors. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: dhcp, cable modems...
On Fri, Mar 02, 2001 at 07:32:02AM -0500, Michael P. Soulier wrote: > On Fri, Mar 02, 2001 at 01:38:36AM -0400, Peter Cordes wrote: > > > It's interesting to note that even though apt-cache search dhcpcd only > > finds pump, (i.e. dhcpcd is not part of woody anymore), you can still > > download the dhcpcd package through packages.debian.org/dhcpcd, and it's on > > the mirrors. > > I've never tried pump, only dhcpcd, so can you answer two questions for > me? > > 1. Does pump permit you to specific a client ID? It lets you specify a host name, with -h. It doesn't say anything about using anything other than the ethernet MAC address as the client ID, though. > 2. Does pump give you a hook for running a script whenever the IP changes? Yup. from pump(8)'s description of the config file: script executable-filename When events occur in negotiation with the server, calls the given executable or script. Scripts are called when a lease is granted, when a renewal is negotiated, and when the interface is brought down and the address released. The scripts are called with two or three arguments, depending on the con- dition, as documented in the table above. > > If so, then I'll probably switch, if it's not maintained in woody anymore. You might have to switch to dhcp-client, which is dhcpcd renamed, AFAIK. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Can anyone get their hands on a Surecom EP-427X?
Hi,
Can anyone borrow a Surecom EP-427X fast ethernet PC card to try a quick
experiment on for me? I'm trying to get Linux to support this NIC (it's an
ne2000 clone, so it should require only a bit of hacking...) I want to see
whether the problem I'm seeing is my laptop's fault, or what.
I've never used any other pc cards with my laptop (an old used AST Ascentia
810N, which has a Cirrus Logic PD6720 pc card controller.). I don't know if
my laptop's pc card hardware even works totally right, since I got it second
hand. I was able to use a DOS boot disk to run the enabler and packet
driver from the disk that came with the NIC, and run a packet driver test
program to ping another computer on my LAN with the card, so obviously
it is possible to make it work.
This card is a 10/100baseT 16bit pc card NIC, from Surecom. Note that this
is not the same as the EP-427 or EP-427T. They are 10baseT only, and are
already supposed to be supported by linux.
I've got one of these cards, and the driver floppy includes a "linux"
directory, with a patched copy of pcmcia-cs.c, and 8390.c, and instructions
to use this with a 2.0.30 kernel! I've figured out some stuff about how it
works, and figured out that the change Surecom made was adding:
{ /* SURECOM EP-427X */ 0x01c8, 0x00, 0x00, 0x21, 0 }
to the hw_info array. Using the memory-cs driver, and dd /dev/mem0a, as
described in the PCMCIA-HOWTO, the 0x01c8 offset does indeed point to the
card's ethernet MAC address.
The problem is that pcnet_cs.o says:
pcnet_cs: GetFirstTuple: No more items, and refuses to load. Another weird
thing that happened is that the linux info from surecom says the right
config for the card is:
card "SURECOM EP-427X 10/100M PCMCIA Adapter"
version "PCMCIA", "100BASE"
bind "pcnet_cs"
However, when I run /etc/init.d/pcmcia start, the drivers read the card's
CIS info as: "PCMCIA", "1". I changed the version line in the config file
to match that, so it would try to load pcnet_cs. I think my laptop is
having problems reading from the CIS on the card, since the tuples are
stored there, and so is the version string stuff, and my laptop had trouble
with both. The other possibility is that the card has some lame CIS that
linux doesn't like :( By having somebody else test it, I hope to find out
which it is.
It would be great if someone could stick one of these NICs in their linux
laptop and tell me what it says for version, ("PCMCIA", "100BASE"), or
("PCMCIA", "1"). Also likely to be helpful would be setting up config to
load memory_cs instead of pcnet_cs, and running
dd if=/dev/mem0a count=20 | od -Ax -t x1 > ep427x.dump
and sending me the dump, along with what the MAC address if you have it.
(This is described in the PCMCIA-HOWTO.)
If you want to try to get the card actually working, patch pcnet_cs.c like
this:
--- old-pcnet_cs.c Wed Nov 22 19:05:45 2000
+++ pcnet_cs.c Fri Feb 23 18:54:21 2001
@@ -199,6 +199,8 @@
{ /* Socket EA */ 0x4000, 0x00, 0xc0, 0x1b,
DELAY_OUTPUT | HAS_MISC_REG | USE_BIG_BUF },
{ /* SuperSocket RE450T */ 0x0110, 0x00, 0xe0, 0x98, 0 },
+{ /* SURECOM EP-427X */ 0x01c8, 0x00, 0x00, 0x21, 0 },
{ /* Volktek NPL-402CT */ 0x0060, 0x00, 0x40, 0x05, 0 },
{ /* NEC PC-9801N-J12 */ 0x0ff0, 0x00, 0x00, 0x4c, 0 },
(the line numbers are probably messed up, so patch(1) probably won't apply
this cleanly. Just add the line with the + in front of it by using a text
editor.)
Recompile it and see if it works.
Thanks,
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BCE
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: XF86Setup&XF86Config + gnome doesn't start
On Sat, Mar 03, 2001 at 12:25:40PM -0800, [EMAIL PROTECTED] wrote: > I typed the command and XF86Config was found at /usr/X11R6/lib/X11/ but > when I try to open it with emacs it writes: "file exists, but cannot be > read." Maybe a broken symlink? Maybe it was created by root without read for other permission, and you are running as an ordinary user? BTW, the locate command is good for finding files. I'll just tell you where to find XF86Config, though. If it exists at all, it'll be in /etc/X11. That's where all the X config files live. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: XF86Setup&XF86Config + gnome doesn't start
On Sat, Mar 03, 2001 at 02:48:25PM -0800, [EMAIL PROTECTED] wrote: > On Sat, 03 March 2001, Peter Cordes wrote: > > > XF86Config doesn't exist in /etc, therefore I asume that it isn't >installed properly or not at all. But if I have installed hole packages in >first cd disk, what is then missing?? You're checking /etc/X11/XF86Config, not /etc/XF86Config, right? /etc/X11/XF86Config is _the_ location for it. If that's what you meant, (or you meant the it wasn't in any directory rooted in /etc), then yeah, it's not there at all. Time to run the configurator: xf86cfg (for X4.0), or the old standby, xf86config. If you have problems, take a look at the man pages for these programs. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: cron & anacron
On Sun, Mar 04, 2001 at 11:25:43AM -0500, Dan Christensen wrote: > [EMAIL PROTECTED] (Bostjan Muller) writes: > On my system, the default big cron jobs are automatically ignored by > cron if anacron is install. Here is part of my /etc/crontab: > > 25 6* * * roottest -e /usr/sbin/anacron || run-parts --report >/etc/cron.daily > 47 6* * 7 roottest -e /usr/sbin/anacron || run-parts --report >/etc/cron.weekly > 52 61 * * roottest -e /usr/sbin/anacron || run-parts --report >/etc/cron.monthly > > So don't remove cron. My system is like that too. You're wrong, though; cron doesn't do anything if anacron is installed. _nothing_, not even run anacron. /usr/sbin/anacron runs as a daemon and does the run-parts stuff. If you check the package dependency, anacron only recomends, not depends on, cron. Go ahead and remove cron. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: cron & anacron
On Sun, Mar 04, 2001 at 07:55:51PM -0500, Dan Christensen wrote:
> Peter Cordes <[EMAIL PROTECTED]> writes:
>
> > On Sun, Mar 04, 2001 at 11:25:43AM -0500, Dan Christensen wrote:
> > > [EMAIL PROTECTED] (Bostjan Muller) writes:
> > > On my system, the default big cron jobs are automatically ignored by
> > > cron if anacron is install. Here is part of my /etc/crontab:
> > >
> > > 25 6* * * roottest -e /usr/sbin/anacron || run-parts --report
>/etc/cron.daily
> > > 47 6* * 7 roottest -e /usr/sbin/anacron || run-parts --report
>/etc/cron.weekly
> > > 52 61 * * roottest -e /usr/sbin/anacron || run-parts --report
>/etc/cron.monthly
> > >
> > > So don't remove cron.
> >
> > My system is like that too. You're wrong, though; cron doesn't do
> > anything if anacron is installed. _nothing_, not even run anacron.
> > ...
> > Go ahead and remove cron.
>
> On my system there are a lot of files in /etc/cron.d which cron *does*
> process, even though I have anacron installed. So I think my advice
> is sound, unless Bostjan has an installation with nothing in that
> directory.
Err, yes, you're right. I should have gotten more sleep before I wrote my
last message :(. I've got an old 486 laptop that I run really stripped
down. (I'm a hacker and I know what I'm doing, so it works for me:).
I like the idea of running cron/anacron stuff whenever AC power is present,
I'll have to try that, esp. if anacron does it automatically.
On my desktop, exim and logcheck have stuff in /etc/cron.d. Everything
else drops stuff in cron.{hourly,daily,...}.
grep 'cron\.d/' /pub/debian/Contents-i386 comes up with 35 files that want
to be in /etc/cron.d in all of Woody. Shouldn't be too much work to deal
with, esp since most of them didn't look like the kind of package you would
typically have on a laptop (e.g. new servers).
> My only point was that removing cron wouldn't accomplish anything that
> Bostjan wanted, and would have the potential of removing functionality.
Removing cron would make it stop reading files every so often, which causes
disk writes unless you have the filesystem mounted with the noatime
option. (I do this on my laptop, and haven't had problems with it.)
Fortunately, cron itself is not at all CPU intensive, only the cron jobs it
runs are, so you wouldn't save a whole lot.
--
#define X(x,y) x##y
Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca)
"The gods confound the man who first found out how to distinguish the hours!
Confound him, too, who in this place set up a sundial, to cut and hack
my day so wretchedly into small pieces!" -- Plautus, 200 BCE
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: WHERE are the docs?
On Mon, Mar 05, 2001 at 09:45:00AM -0500, William F. Dowling wrote: > Sometimes docs don't get installed when you think they should, so > using 'find ...' to find them is fruitless. locate(1) is good. It can search the whole system quickly. > Suggestion for Debian > maintainers: at install time let me specify "I want all relevant docs > by default". On one system I have (not my notebook) I managed > (newbily surfing through dselect) to have installed gcc, but no gcc > info pages. It took me a while to realize I needed the gcc-doc > package. Why not have foo-docless packages for those who really don't > want docs, and let the package foo contain the docs? Because that wastes space and bandwidth on the mirrors. Packages with separate doc-packages should recommends: foo-doc or suggests: foo-docs. If they don't recommend or suggest the package with the docs, then you should complain or file a bug report. :) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: WHERE are the docs?
On Mon, Mar 05, 2001 at 04:47:12PM -0500, xsdg wrote: > On Mon, Mar 05, 2001 at 02:14:58PM -0400, Peter Cordes wrote: > > locate(1) is good. It can search the whole system quickly. > locate doesn't actually search the system. When a user runs updatedb(1?), >a database is created of all files accessible to the user running the >command. The locate command searches through the database. On debian >systems, updatedb is run in cron.daily. Yes, that's _why_ it can _effectively_ search the whole system quickly. find takes a long time because it has to traverse the directory tree. Locate just has to read its file where it keeps a list of every filename on the system. This is a good thing, because sequential reads are a lot faster than seeks and reads to scattered locations. I don't know if findutils runs updatedb in its postinst script, but if not, then a newly installed system won't have locate working for the first day. After that, the only downside is that it won't find files that were installed since last night. If a user recently installed a package, they probably looked at its contents. Still, things aren't perfect, but locate is a lot better than nothing. Another useful program is glimpse. I haven't used it, but it's supposed to make a (quickly) searchable database of the contents of all text files. It also only runs periodically. -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: sound blaster configuration??
On Mon, Mar 05, 2001 at 04:23:26PM -0800, [EMAIL PROTECTED] wrote: > I have tried to find information in /usr/src/linux.. but I don't have any directory like ../src/linux/ .. thanks anyway.. That would be the kernel source tree. Grab it from www..kernel.org/pub/linux/kernel/v2.2/linux-2.2.18.tar.bz2, or install the kernel-source package. If you aren't going to recompile the kernel to suit your tastes, there might be a kernel-doc package which just has linux/Documentation, but not the rest of the source. If not, then just get the whole tree. A lot of really good docs on getting the kernel to do stuff is in linux/Documentation. > > > > > We first need more information: > > > > Is this Sound Blaster 16 PCI or ISA? > > Is it PnP or not? > > Which Kernel do you use? > > > > pnpdump -c > /etc/isapnp.conf && isapnp /etc/isapnp.conf did not give >any results, didn't find my sound card.. So I think it isn't a ISA.. I am >using kernel 2.2.. If it's PCI, lspci will find it (independently of whether any drivers are loaded for it. PCI is good, because there is a standard for scanning the bus and finding config info). (If you don't have lspci, you look at the obsolete /proc/pci. lspci interprets the binary data in the new /proc/bus/pci interface.) -- #define X(x,y) x##y Peter Cordes ; e-mail: X([EMAIL PROTECTED] , ns.ca) "The gods confound the man who first found out how to distinguish the hours! Confound him, too, who in this place set up a sundial, to cut and hack my day so wretchedly into small pieces!" -- Plautus, 200 BCE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
