Re: asrock, problem with nic after windows-boot

iforone wrote:
> (E-Mail Removed) wrote:
> > hello,
> > i got here an asrock board K8NF4G-SATA2 with an realtek phy rtl8201cl
> > networkchip and a nforce chipset. i work with a debian like kanotix
> > system kernel 2.16.20
> >
> > under linux all works fine. but i got the following phenomenon. when i
> > boot windows xp and after this i boot linux, then is there no network
> > available, the system got no ip address from the dhcp server, seems as
> > the cable is disconnected. when i at this point go to the bios and
> > disable onboard lan, make power off, switch on and enable the lan in
> > the bios again, then the nic works correct under linux. but when i boot
> > windows later, the same problem after this und linux.
> > any ideas how i can fix this ? any ideas whats the problem ?
> >
> > greetings Randolf Balasus
>
> All I can say is I have (almost?) the exact same thing occur all the
> time now, with my dual boot Debian (sarge 3.1r1, 2.6.8-2 kernel) and
> win98. It's a fairly common recent happening now for me, but didn't
> used to occur often - and I've found that the quickest way to obtain a
> dhcp address (as doled out from my router), is to just ...yep -->
> Reboot winblows.
>
> If you open the Run box (while booted into XPoop) and type CMD, then
> type in 'IPCONFIG /all' you likley will see a 169.xxx.xxx.xxx address
> given to your NIC - this is some sort of twisted Multicast IP addy
> (which is what mine defaults to when this occurs). I've tried win98's
> version of IPCONFIG /release and /renew to no avail - so Win=Reboot,
> and Linux=BeRoot ;-)
>
> WAG; it may have something to do with the fact that I have smbd and
> nmbd daemons running while in Linux (and perhaps others associated with
> File Server setup) , which then wreak havoc on the router's DHCP server
> capabilities. When I first encountered this problem, I did try setting
> up a purely Static IP addy on win98, but eventually gave up after
> awhile (In Debian, it's a cakewalk to do), as my stupid 4 port router,
> has so many different settings, and I just fell shy of doing it
> correctly, and ended up removing ...and then reinstalling all the
> necessary networking components from win98, but even left out a few I'm
> not using (that are placed there as default during setup). All to no
> avail. Oh, I also made the mistake years ago of setting up my router to
> clone my NIC's MAC addy (which I think my ISP (which also doles out my
> public addy using DHCP, but I always seem to obtain the 'same' addy
> with each lease) doesn't like/want me to alter)... Rebooting into
> winblows (twice) fixes the problem (atleast temporarily).

By the way - I really wouldn't think of this as a "hardware" problem,
and so I also added c.o.l.networking (just now with this post, the 3rd
post so far in this thread) to expand the pool of available help -
perhaps using the 'followup-to' header would be more appropro, I'm not
sure.

On 9 Jun 2006, in the Usenet newsgroup comp.os.linux.hardware, in article
<(E-Mail Removed). com>, iforone wrote:

>iforone wrote:
>> (E-Mail Removed) wrote:
>>> i got here an asrock board K8NF4G-SATA2 with an realtek phy rtl8201cl
>>> networkchip and a nforce chipset. i work with a debian like kanotix
>>> system kernel 2.16.20

That's more likely 2.6.16.20, which was released four days ago.

>>> under linux all works fine. but i got the following phenomenon. when i
>>> boot windows xp and after this i boot linux, then is there no network
>>> available, the system got no ip address from the dhcp server, seems as
>>> the cable is disconnected. when i at this point go to the bios and
>>> disable onboard lan, make power off, switch on and enable the lan in
>>> the bios again, then the nic works correct under linux. but when i boot
>>> windows later, the same problem after this und linux.
>>> any ideas how i can fix this ? any ideas whats the problem ?

I'm guessing here, but this sounds like the problem that Donald Becker
reported about the RTL8139[BC] some time ago. What driver are you using?
Do you see anything relating to the card in /var/log/messages - see where
the kernel is looking for the card at boot time.

Lessee...

===========================
>So once again I booted up the machine in Windows (98 SE) and then
>rebooted into Linux again and ... the hardware adress was once again 0.
>Shutting down the power and the booting up in Linux again gives the card
>the right address again.

This can be solved by using the 'pci-scan' module and updated driver
from
http://www.scyld.com/network/updates.html

You have a RTL8139B or RTL8139C with PCI power management.
Windows leaves PCI power management capable devices in D3-warm power
state. Neither the BIOS nor Linux knows about PCI power management.
The 'pci-scan' module, combined with the updated driver set, knows how
to restore the device to full power "D0" state before using it.
===========================

Does the card work if you just power off, and then power back on - without
kicking the BIOS?

>> If you open the Run box (while booted into XPoop) and type CMD, then
>> type in 'IPCONFIG /all' you likley will see a 169.xxx.xxx.xxx address
>> given to your NIC - this is some sort of twisted Multicast IP addy
>> (which is what mine defaults to when this occurs).

See RFC3330 - it's actually a random address in the 169.254.x.x range,
called ZeroConf. Microsoft adopted this concept (originally from Apple)
so that when the MCSE so fscks up the DHCP server that even windoze can't
get an address, this service will reach between it's legs and grab some
address. This will let the systems work locally, but is otherwise unusable
for networking.

In this case, windoze is able to see the card, but isn't able to get it to
talk on the wire. Rather than provide a clue that something is wrong,
windoze just fakes it and hopes that this works well enough for you.

>> I've tried win98's version of IPCONFIG /release and /renew to no avail
>> - so Win=Reboot, and Linux=BeRoot ;-)

The problem is that windoze has screwed up the configuration of the card,
and only a power-on reset (or the front panel reset button if you still
have one of those) will grab the card's attention.

>> WAG; it may have something to do with the fact that I have smbd and
>> nmbd daemons running while in Linux (and perhaps others associated with
>> File Server setup) , which then wreak havoc on the router's DHCP server
>> capabilities.

Nope - nothing to do with it.

>By the way - I really wouldn't think of this as a "hardware" problem,
>and so I also added c.o.l.networking (just now with this post, the 3rd
>post so far in this thread) to expand the pool of available help

comp.os.linux.hardware Hardware compatibility with the Linux operating system.
comp.os.linux.networking Networking and communications under Linux.

Well... I haven't seen posts from Donald Becker in several years, but he
tended to post mainly in c.o.l.networking.

>perhaps using the 'followup-to' header would be more appropro, I'm not
>sure.

Normally it is, and as you quoted the entire thread without trimming
(normally a no-no), every one there should be able to pick it up. I did
set the followup-to, so the O/P should look in that group for further
responses.

Old guy

Moe Trin wrote:
> I'm guessing here, but this sounds like the problem that Donald Becker
> reported about the RTL8139[BC] some time ago. What driver are you using?
> Do you see anything relating to the card in /var/log/messages - see where
> the kernel is looking for the card at boot time.

Hi;
I know you know I'm not the OP, and for the sake of clarification, my
issue seems to be the almost exact *opposite* of the OP's; In my win98
+ Debian Sarge 3.1r1 (2.6.8-2 kernel) dual-boot setup, it's when I boot
up into *win98* after being in Debian for a while (anywhere from 1 day
to however long) that Microcrud's 'ZeroConf' crap occurs in win98 for
me. My (quickest) workaround has been to just reboot....ala the MS way
:-/

Some information that may be relevant;
* WakeOnLan connection from NIC to Mobo is 'connected', though I don't
use it.
* My NIC (Netgear) info;

~$ sudo lspci -vv
[...]
0000:00:0e.0 Ethernet controller: National Semiconductor Corporation
DP83815 (MacPhyter) Ethernet Controller
Subsystem: Netgear: Unknown device f312
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR+ FastB2B-
Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
Latency: 90 (2750ns min, 13000ns max)
Interrupt: pin A routed to IRQ 11
Region 0: I/O ports at 1400 [size=256]
Region 1: Memory at f4000000 (32-bit, non-prefetchable)
[size=4K]
Capabilities: [40] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=320mA
PME(D0+,D1+,D2+,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME+

* ACPI is active (via 'acpi=force' in GRUB Kernel command line) on this
circa 1999 Intel 440BX (PentiumII) mobo;

~$ dmesg | grep -i ACPI
BIOS-e820: 000000000bffdc00 - 000000000bfffc00 (ACPI data)
BIOS-e820: 000000000bfffc00 - 000000000c000000 (ACPI NVS)
ACPI: RSDP (v000 PTLTD ) @
0x000f6a80
ACPI: RSDT (v001 PTLTD RSDT 0x00000001 PTL 0x01000000) @
0x0bffde96
ACPI: FADT (v001 GATEWA TABOR II 0x19991104 PTL 0x000f4240) @
0x0bfffb8c
ACPI: DSDT (v001 GATEWA TABOR II 0x00000001 MSFT 0x01000004) @
0x00000000
ACPI: PM-Timer IO Port: 0x8008
Kernel command line: root=/dev/hdc1 ro acpi=force
ACPI: Subsystem revision 20040326
ACPI: IRQ9 SCI: Level Trigger.
ACPI: Interpreter enabled
ACPI: Using PIC for interrupt routing
ACPI: PCI Root Bridge [PCI0] (00:00)
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Link [LNKA] (IRQs 3 4 5 7 9 10 11 12) *0, disabled.
ACPI: PCI Interrupt Link [LNKB] (IRQs 3 4 5 7 9 10 *11 12)
ACPI: PCI Interrupt Link [LNKC] (IRQs 3 4 5 7 9 10 11 12) *0, disabled.
ACPI: PCI Interrupt Link [LNKD] (IRQs 3 4 5 7 *9 10 11 12)
ACPI: Power Resource [PFAN] (on)
PCI: Using ACPI for IRQ routing
ACPI: PCI Interrupt Link [LNKD] enabled at IRQ 9
ACPI: PCI interrupt 0000:00:07.2[D] -> GSI 9 (level, low) -> IRQ 9
ACPI: PCI Interrupt Link [LNKB] enabled at IRQ 11
ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt 0000:00:0e.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt 0000:00:10.0[A] -> GSI 9 (level, low) -> IRQ 9
ACPI: PCI interrupt 0000:01:00.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: (supports S0 S1 S5)
ACPI: PCI interrupt 0000:00:0e.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt 0000:00:0c.0[A] -> GSI 11 (level, low) -> IRQ 11
ACPI: PCI interrupt 0000:00:07.2[D] -> GSI 9 (level, low) -> IRQ 9
pciehp: acpi_pciehprm:\_SB_.PCI0 evaluate _BBN fail=0x5
pciehp: acpi_pciehprm:get_device PCI ROOT HID fail=0x5
shpchp: acpi_shpchprm:\_SB_.PCI0 evaluate _BBN fail=0x5
shpchp: acpi_shpchprm:get_device PCI ROOT HID fail=0x5

* These last 4 'fail' entries above only occur using the 2.6 kernel,
and while I don't have a 2.4 setup installed right now, I recall these
errors specifically *not* being there without using 'hald' and whatever
udev elements(vs devfs) the 2.6.8-2 kernel uses.

* Speaking of the 2.4 kernel; when Debian was on this system using the
2.4 kernel, this issue never occured (IIRC). I've reinstalled Debian
quite a few times, for various reasons, over the last year or
so...sometimes using 2.4, sometimes 2.6, sometimes using the
[+]FileServer install along with [+]Desktop, but most times not. Either
way, this is the 1st install of Debian using a separate (physical) HD
drive (on a very LOUD small 9.1 GB Western Digital)...and in this case
- both FileServer + Desktop were chosen during install time.

~$ dmesg | grep -i hd*
[...]
hda: HDS728080PLAT20, ATA DISK drive # win98 on 80GB Hitachi
hdb: LITE-ON LTR-32123S, ATAPI CD/DVD-ROM drive
hdc: WDC AC29100D, ATA DISK drive # Debian Sarge on 9.1GB WD
hdd: _NEC DVD_RW ND-3550A, ATAPI CD/DVD-ROM drive

Also note;
I've uninstalled the AppleTalk Services(?) a couple months back --
During Boot, the system would take a l-o-n-g time and get hung up
(30-60seconds) trying to locate/find/initialize the AppleTalk stuff
that gets installed when choosing 'FileServer' - drats, I don't have
the exact dmesg/messages from that time (maybe a few months ago). I
think I tracked down the issue and used Synaptic to uninstall, though I
don't believe I "purged".

....Aha!!
It was called 'Netatalk', 'atalkd', and 'afpd'...
w00t! -- I just figured out how to use gzip to view docs that have root
as owner :-) (KDE, Konquerer complains - permission denied, for good
reason, I know)

$ cd /var/log
$ ls -al | less
-rw-r----- 1 root adm 67234 2006-06-10 06:17 syslog
-rw-r----- 1 root adm 24864 2006-06-08 06:25 syslog.0
-rw-r----- 1 root adm 14621 2006-05-26 06:25 syslog.1.gz
-rw-r----- 1 root adm 616 2006-05-07 06:25 syslog.2.gz
-rw-r----- 1 root adm 13858 2006-05-06 06:25 syslog.3.gz
-rw-r----- 1 root adm 14734 2006-05-03 06:25 syslog.4.gz
-rw-r----- 1 root adm 10989 2006-05-01 06:25 syslog.5.gz
(note the Dates)

$ sudo gzip -dc syslog.5.gz | less
[...]
May 1 05:44:12 localhost kernel: eth0: DSPCFG accepted after 0 usec.
May 1 05:44:12 localhost kernel: eth0: link up.
May 1 05:44:12 localhost kernel: eth0: Setting full-duplex based on
negotiated link capability.
May 1 05:44:12 localhost kernel: NET: Registered protocol family 17
May 1 05:44:15 localhost lpd[3078]: restarted
May 1 05:44:15 localhost kernel: NET: Registered protocol family 5
May 1 05:44:15 localhost atalkd[3097]: restart (2.0.2)
May 1 05:44:16 localhost atalkd[3097]: zip_getnetinfo for eth0
May 1 05:44:36 localhost last message repeated 2 times
May 1 05:44:46 localhost atalkd[3097]: config for no router
May 1 05:44:47 localhost atalkd[3097]: ready 0/0/0
May 1 05:44:59 localhost papd[3105]: restart (2.0.2)
May 1 05:44:59 localhost papd[3105]: CUPS support enabled (1.1)
May 1 05:45:00 localhost afpd[3107]: Registering CNID module [last]
May 1 05:45:00 localhost afpd[3107]: Registering CNID module [cdb]
May 1 05:45:00 localhost afpd[3107]: Registering CNID module [dbd]
May 1 05:45:00 localhost afpd[3107]: Loading ConfigFile
May 1 05:45:00 localhost afpd[3107]: Finished parsing Config File
May 1 05:45:01 localhost kernel: NET: Registered protocol family 10
May 1 05:45:01 localhost kernel: Disabled Privacy Extensions on device
c02cc960(lo)
May 1 05:45:01 localhost kernel: IPv6 over IPv4 tunneling driver
May 1 05:45:02 localhost rpc.statd[3149]: Version 1.0.6 Starting
May 1 05:45:02 localhost rpc.statd[3149]: statd running as root. chown
/var/lib/nfs/sm to choose di
fferent user
May 1 05:45:02 localhost /usr/sbin/cron[3154]: (CRON) INFO (pidfile fd
= 3)
May 1 05:45:02 localhost /usr/sbin/cron[3155]: (CRON) STARTUP (fork
ok)
May 1 05:45:02 localhost /usr/sbin/cron[3155]: (CRON) INFO (Running
@reboot jobs)


> Lessee...
>
> =========================================
[..snipped...most of this older post from Donald Becker, archived by
Moe...]
> This can be solved by using the 'pci-scan' module and updated driver
> from
> http://www.scyld.com/network/updates.html
[...snipped...]
> ==========================================

So...
because my problem is the opposite, I wouldn't think this would be
useful to me(?) since it's winblows that has this 169.254.x.x addy
*after* being booted into Debian.

> >> you likley will see a 169.xxx.xxx.xxx address
> >> given to your NIC - this is some sort of twisted Multicast IP addy
> >> (which is what mine defaults to when this occurs).
>
> See RFC3330 - it's actually a random address in the 169.254.x.x range,
> called ZeroConf. Microsoft adopted this concept (originally from Apple)
> [...]

Thanks for the refresher - I couldn't recall the specifics offhand -
http://en.wikipedia.org/wiki/Zeroconf has some info...and your
abundance of networking experience helps too.

> The problem is that windoze has screwed up the configuration of the card,
> and only a power-on reset (or the front panel reset button if you still
> have one of those) will grab the card's attention.

I wonder how much the Client for MS Networking (win98) munges the
settings.

> >> WAG; it may have something to do with the fact that I have smbd and
> >> nmbd daemons running while in Linux (and perhaps others associated with
> >> File Server setup) , which then wreak havoc on the router's DHCP server
> >> capabilities.
>
> Nope - nothing to do with it.

Okee Dokee - I'll easily accept that from someone as experienced as
you, no prob :-)

[Aside]
> >perhaps using the 'followup-to' header would be more appropro, I'm not
> >sure.
>
> Normally it is, and as you quoted the entire thread without trimming
> (normally a no-no),

Understood ...In this case - is about the only time that I can think
of offhand, that I'd quote the entire (albeit small) thread...and only
if I catch it early on, as I have in this case.

> every one there should be able to pick it up. I did
> set the followup-to, so the O/P should look in that group for further
> responses.

Noted - and (hopefully) retained :-)
[/Aside]

Thanks for any/all input about these problems, and I hope I haven't
hijacked this thread...if I have - just disregard this post - this
minor issue is not a huge deal for me, considering all the other issues
I'm having with my installation (too intricate - especially for this
thread).

Thanks again Moe Trin
Regards

On 10 Jun 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<(E-Mail Removed) .com>, iforone wrote:

>I know you know I'm not the OP, and for the sake of clarification, my
>issue seems to be the almost exact *opposite* of the OP's; In my win98
>+ Debian Sarge 3.1r1 (2.6.8-2 kernel) dual-boot setup, it's when I boot
>up into *win98* after being in Debian for a while (anywhere from 1 day
>to however long) that Microcrud's 'ZeroConf' crap occurs in win98 for
>me. My (quickest) workaround has been to just reboot....ala the MS way
>:-/

OK - let's clarify this - when you reboot to get it to work, it this a
power-cycle (or front panel reset button), or just the three-finger-salute?

The rational is that a power-cycle or front panel reset switch does a full
hardware reset - there actually is a wire on the motherboard going to all
ISA and/or PCI sockets and the "important" chips on the motherboard (such
as the CPU) labeled /RESET. Dragging that "low" for <mumble> clock cycles
resets all hardware to a known state. The three-finger-salute on the
other hand, just generates a signal to the operating system ALONE to start
running some piece of code (in windoze, this used to cause it to restart
the software - in Linux, look in /etc/inittab to see what your system is
configured to do), but it does exactly nothing to the hardware. Also, if
the operating system is out in la-la land somewhere and not reading the
keyboard, the three-finger-salute isn't even seen, never mind acted upon.

The ZeroConf address would come up in windoze if it thinks it loaded the
device driver OK (meaning only that no error messages were received while
loading the driver), and then when it tried to contact a DHCP server, it
got no error indication from the hardware, but also was not able to contact
any DHCP server.

In your case, this is _somewhat_ similar to the O/P (although you seem to
have a different NIC), in that the one operating system is leaving the
hardware in a mode that the other operating system doesn't recognize, or
doesn't know how to correct.

>Some information that may be relevant;
>* WakeOnLan connection from NIC to Mobo is 'connected', though I don't
>use it.

I don't _think_ it's relevant - if it were, I'd expect either no, or
constant booting problems

> 0000:00:0e.0 Ethernet controller: National Semiconductor Corporation
> DP83815 (MacPhyter) Ethernet Controller
> Subsystem: Netgear: Unknown device f312

http://pciids.sourceforge.net

100b National Semiconductor Corporation
0020 DP83815 (MacPhyter) Ethernet Controller
103c 0024 Pavilion ze4400 builtin Network
1385 f311 FA311 / FA312 (FA311 with WoL HW)

OK, I think. My notes suggest that's a 'natsemi' driver.

>* ACPI is active (via 'acpi=force' in GRUB Kernel command line) on this
>circa 1999 Intel 440BX (PentiumII) mobo;

Can't help there.

>I've uninstalled the AppleTalk Services(?) a couple months back --
>During Boot, the system would take a l-o-n-g time and get hung up
>(30-60seconds) trying to locate/find/initialize the AppleTalk stuff
>that gets installed when choosing 'FileServer'

That's a disadvantage of some of these overly helpful installation
programs. If I wanted support for Apple (or Novell, or Samba, or...), I'd
have told you. Stop being so helpful!! Hack! Smash! Grrr...

>w00t! -- I just figured out how to use gzip to view docs that have root
>as owner :-) (KDE, Konquerer complains - permission denied, for good
>reason, I know)

Security - it's not entirely unknown for all kinds of "sensitive" information
to accidentally get into the logs. We've had more than one user who attempted
to log in as his password, and then enter his username when prompted for the
secret word (invariably, some real wiz from Mahogany Row). Sigh...

>$ sudo gzip -dc syslog.5.gz | less

That will do the trick.

>So...
>because my problem is the opposite, I wouldn't think this would be
>useful to me(?) since it's winblows that has this 169.254.x.x addy
>*after* being booted into Debian.

It could still be something similar - point is, how do you induce the
failure mode. I have only one system that dual boots (Slackware and
Red Hat), and I can not get '/sbin/shutdown -r now' to cleanly restart
the system. I must use '/sbin/shutdown -h now' and then poke the reset
button after the system reports "System Halted" - which is a lie anyway
as I can still ping the "halted" box (though I can't connect to any
other network services). If I don't do this (reset), the system gets
lost during the subsequent reboot, and wedges solid.

>I wonder how much the Client for MS Networking (win98) munges the
>settings.

I dunno - the snippet from Becker was responding to someone with that
problem in win98.

>Thanks for any/all input about these problems, and I hope I haven't
>hijacked this thread...

QUICK, SOMEONE CALL THE USENET POLICE!! HIJACKING IN PROGRESS!!!

>if I have - just disregard this post - this minor issue is not a huge
>deal for me, considering all the other issues I'm having with my
>installation (too intricate - especially for this thread).

Not a problem. You changed the subject line, so if the O/P responds to
my original response or your response over in c.o.l.h, the subject line
will tell things apart - not that it really matters, as the problems are
related.

Old guy

Moe Trin wrote:
> OK - let's clarify this - when you reboot to get it to work, it this a
> power-cycle (or front panel reset button), or just the three-finger-salute?


It doesn't seem to matter whether or not ctrl+alt+del was used - or
logout/shutdown, from the Debian(KDE) menu, or if it's from a 'cold'
boot - no difference either way :-(

> The rational is that a power-cycle or front panel reset switch does a full
> hardware reset - there actually is a wire on the motherboard going to all
> ISA and/or PCI sockets and the "important" chips on the motherboard (such
> as the CPU) labeled /RESET.

I like to think of myself as more of a 'hardware' guy than 'software'
guy, and you seem to have quite a good understanding of hardware
(especially older systems) - which leads me to my question about a
"RESET" button; which I thought was mostly/only specifically related to
AT systems (of which I have an old P166 that's in a BabyAT case).
Everything after - I thought was _ATX_.

Basically, my understanding is: The AT systems had/have Full power(120v
AC) running to the actual On/Off switch (which caused many a zapper for
some - and allows for a RESET front panel button [and "turbo" mode
:-]), and the Mobo could (potentially and literally) blowup in your
face, if one forgot to discharge the capacitors, and/or unplug the
power -- which is why ATX was introduced. (I'm sure there are other
reasons too, that I can't recall ATM).

> Dragging that "low" for <mumble> clock cycles
> resets all hardware to a known state. The three-finger-salute on the
> other hand, just generates a signal to the operating system ALONE to start
> running some piece of code

I have always understood this to mean a *soft* or *warm* boot (as
opposed to a *cold* boot, - since also the RAM doesn't necessarily get
discharged (or checked) upon a restart ('shutdown -r now' , or through
the GUI).

> (in windoze, this used to cause it to restart
> the software - in Linux, look in /etc/inittab to see what your system is
> configured to do), but it does exactly nothing to the hardware.

Understood and agreed ;-) - it's a Keyboard assigned Interrupt signal
(AFAICT)

> Also, if
> the operating system is out in la-la land somewhere and not reading the
> keyboard, the three-finger-salute isn't even seen, never mind acted upon.

which is why when a system *hangs*, this ctrl+alt+del keycode usually
does squat anyway.

> The ZeroConf address would come up in windoze if it thinks it loaded the
> device driver OK (meaning only that no error messages were received while
> loading the driver),

I get _no_ errors when chainloading Winblows (win98) through GRUB - but
it 'does' take a little time after my autoexec.bat file loads - it sits
at the text/boot screen showing C:\DOSKey /insert (I can see all
onscreen bootup messages for any OS, even though I may have a "buggy"
BIOS)...and all my silly BIOS's 'quick/fast/silent boot' options are
set to off.

Another piece of possible relevant info;
I've totally disabled NetBIOS in win98 (those nasty 135-139 ports),
something akin to this <http://www.grc.com/su-rebinding9x.htm>

> In your case, this is _somewhat_ similar to the O/P (although you seem to
> have a different NIC), in that the one operating system is leaving the
> hardware in a mode that the other operating system doesn't recognize, or
> doesn't know how to correct.

I think so too - precisely.

> >* WakeOnLan connection from NIC to Mobo is 'connected', though I don't
> >use it.
>
> I don't _think_ it's relevant - if it were, I'd expect either no, or
> constant booting problems

hmm...I'll have to also delve deeper into that as an issue - I say so;
b/c one Debian installation (awhile back, and on the /same/ HD as
win98) freaked out on me while the machine was totally Off , except for
the WOL - the NIC light always stays On (awaiting instructions) even
after a Clean and Full power shutdown.

What eventually happened was;
I was shutdown completely (except for that WOL I guess) and after a day
of being away, I booted up (into GRUB, then chose Debian) and could not
get passed a certain point in the text bootup messages (I can't recall
/exactly/ where it hung, but everything I tried seemed to make it
worse) . Finally I couldn't even use the Debian install, and used only
98 for awhile, while I was trying to sort through the Debian issues and
tryingto recover. It's a very long process/story that ended me back
into re-installing Both OSes...(Details left out, for brevity in this
thread).

> 100b National Semiconductor Corporation
> 0020 DP83815 (MacPhyter) Ethernet Controller
> 103c 0024 Pavilion ze4400 builtin Network
> 1385 f311 FA311 / FA312 (FA311 with WoL HW)
>
> OK, I think. My notes suggest that's a 'natsemi' driver.

Right you are;
~$ dmesg | grep -i natsemi
natsemi dp8381x driver, version 1.07+LK1.0.17, Sep 27, 2002
http://www.scyld.com/network/natsemi.html
natsemi 0000:00:0e.0: EEPROM did not reload in 20000 usec.
natsemi eth0: NatSemi DP8381[56] at 0xcc82b000 (0000:00:0e.0),
00:09:5b:8d:10:f5, IRQ 11, port TP.

I still have trouble using commands (and it's options) such as
'modprobe' and 'depmod' to find out more useful info about which
modules are loaded, being used, their dependencies, etc.

FWIW..I have a plain vanilla Kernel installed using initrd(?) and the
cramfs - I only say this , in case some funky module parsing commands
are being told to me

~$ dmesg | grep -i init
checking if image is initramfs...it isn't (ungzip failed); looks like
an initrd
Freeing initrd memory: 4216k freed
Initializing Cryptographic API
RAMDISK driver initialized: 16 RAM disks of 8192K size 1024 blocksize

~$ dmesg | grep -i cram
RAMDISK: cramfs filesystem found at block 0
VFS: Mounted root (cramfs filesystem) readonly.

IOW - there's nothing in /usr/src/linux - there's not even a /linux
subDIR in there.

~$ cat /boot/grub/menu.lst
[...]
title Debian GNU/Linux, kernel 2.6.8-2-386
root (hd1,0)
kernel /boot/vmlinuz-2.6.8-2-386 root=/dev/hdc1 ro acpi=force
initrd /boot/initrd.img-2.6.8-2-386
savedefault
boot
[...]

~$ cat /etc/fstab
[...]
proc /proc proc defaults 0 0
/dev/hdc1 / ext3 defaults,errors=remount-ro 0
1
[...]

> >* ACPI is active (via 'acpi=force' in GRUB Kernel command line) on this

> Can't help there.

Okee dokey ;-)

> >I've uninstalled the AppleTalk Services(?) a couple months back --

> That's a disadvantage of some of these overly helpful installation
> programs. If I wanted support for Apple (or Novell, or Samba, or...), I'd
> have told you. Stop being so helpful!! Hack! Smash! Grrr...

I hear that - ....you sure as heck said it! -- even though to be fair,
I must confess that I'd most certainly munge an install if I had use
'expert26' mode to choose each and every item, dependency and set every
parameter - I mean (just on the area of Fonts alone) there was a slew
of questions relating to using a certain Font Server, whether or not I
wanted Pongo(?), (Pango?) to handle this or not...something like that -
and that was just the Fonts section of the Install.

My newness to *nix systems has me a bit all over the place - and
uncertain about certain commands, their options, (and mild scripting)
to help find (only the relevant) info buried deep within log files and
for debugging - heck, I haven't even yet figured out how to use the
'Find' command properly yet :-/ ...and 'mount' and 'umount' have me
pulling hair on occasion.

But it's posts from people, such as yourself and others, that when they
post a 'command' - I try to follow along and learn (and go as far as
looking up the 'man' pages , and issuing the command myself, depending
upon what the thread's subject is about) - and that's why I always post
the 'command' + the 'output' - so that others (newbs perhaps) can gain
from it hopefully.

Some things I know well, and other things I know nothing about - that's
how I like(need to) learn, ...so I 'retain' it, and if I use it
constantly - it sticks with me most times. Plus there are many other
areas (other than PCs/Computers) of my life that require my focus and
attention.

I'm not the young man I once was, and while I'm not elderly either -
memory retention (or more correctly, the lack thereof) is just one of
those things that creeps up on you with age.

> We've had more than one user who attempted
> to log in as his password, and then enter his username when prompted for the
> secret word (invariably, some real wiz from Mahogany Row). Sigh...

ouch...
As long as I can actually 'get' to the info, I'm very much ok with
tight security, and actually am trying to learn more about it, use it,
enforce it correctly. I am *king* of my domain (though I won't logon as
root, unless necessary to perform a task, and I have sudo
(/etc/sudoers) setup to cooperate nicely. I guess it's mostly Groups,
and User permissions and such I need to really learn more about - and
not so much the 9-10 digit attributes POSIX stuff, but the Group
'number' assignments and such - nevermind that for now - sorry
[/rambling on]

I need to learn how to tell those darn serfs to do what I want them to
do ;-)

> >$ sudo gzip -dc syslog.5.gz | less

> That will do the trick.

thank you kindly

> >So...
> >because my problem is the opposite, I wouldn't think this would be
> >useful to me(?) since it's winblows that has this 169.254.x.x addy
> >*after* being booted into Debian.
>
> It could still be something similar - point is, how do you induce the
> failure mode.

It doesn't seem to matter whether it's a warm boot (reboot) or a cold
boot

> I have only one system that dual boots (Slackware and
> Red Hat), and I can not get '/sbin/shutdown -r now' to cleanly restart
> the system. I must use '/sbin/shutdown -h now' and then poke the reset
> button after the system reports "System Halted"

This is extremely close to the way my system behaves when using a
Knoppix 3.6 LiveCD - it never entirely/fully shuts down, no matter what
I do..I have to usually drop to root and type 'shutdown -r now' to
reboot into win98 or Debian and then shutdown. All the Hard Disks and
Opticals get their power cut and most everything else appears off, but
I can still hear the system running (PSU fan), and other items are not
powered off in this state.

> - which is a lie anyway
> as I can still ping the "halted" box (though I can't connect to any
> other network services). If I don't do this (reset), the system gets
> lost during the subsequent reboot, and wedges solid.

Exactly - ...BTW is that an AT system or ATX ?

> >I wonder how much the Client for MS Networking (win98) munges the
> >settings.
>
> I dunno - the snippet from Becker was responding to someone with that
> problem in win98.

My NetBIOS settings (the lack of) or protocol (kind of, though not
really a protocol) in combo with my WOL/WOR options have made this
system real funky, I think. Heck, it may be a hardware issue after all
(well, atleast maybe the way PME settings, ACPI, WOL, and the OS
software and BIOS interact).

> QUICK, [snip]

haha ;-)

> Not a problem. You changed the subject line, so if the O/P responds to
> my original response or your response over in c.o.l.h, the subject line
> will tell things apart - not that it really matters, as the problems are
> related.

Got it - and thanks, for both the help -and- the conversation.

Regards

iforone wrote:
> Moe Trin wrote:
> > OK - let's clarify this - when you reboot to get it to work, it this a
> > power-cycle (or front panel reset button), or just the three-finger-salute?
>
>
> It doesn't seem to matter whether or not ctrl+alt+del was used - or
> logout/shutdown, from the Debian(KDE) menu, or if it's from a 'cold'
> boot - no difference either way :-(

[ Correction ]
Whoops - I thought you were asking about while when switching OSes -
but anyway, to answer your above question(s) - I _think_ I (in win98,
where my problem exists, after using Debian) usually just -->Start Menu
| Restart.
[ /Correction ]

I will now go and try again, and see if the 3finger salute makes a diff
or not - I'll report back

OK I'm back... - and yes it seems the 3 finger salutation does the
trick too...i.e.; the same thing occurs as when using Start | Restart
from within win98 (same as rebooting) -- details at the bottom (after
my "mount" dilemma and resolution)


[ 'mount' dilemma ]
Besides I need to reboot anyway, since I can't seem to 'umount' hda1
properly,
To mount my hda (win98 only + plus and old dilapidated and corrupted
Debian install)
I did;
~$ sudo mkdir /mnt/tmp
~$ sudo mount -t vfat o,uid=username,ro /dev/hda1 /mnt/tmp # this
failed
~$ sudo mount -t vfat -o,uid=username,ro /dev/hda1 /mnt/tmp # this
worked

Like an idiot, I mounted hda1, when I needed to mount the whole hda
(all 3 win98 partitions)....but I *should* be able to just easily
umount it and remount the whole hda - but n-o-o-o :-( woe is me...

....I've tried many (different) iterations of umount and mount;

~$ sudo umount /dev/hda1 /mnt/tmp
umount: /mnt/tmp: device is busy
umount: /mnt/tmp: device is busy
umount: /mnt/tmp: device is busy
umount: /mnt/tmp: device is busy

~$ sudo umount -r ro /dev/hda /mnt/tmp
umount: ro: not found
umount: /dev/hda: not mounted
umount: /dev/hda1 busy - remounted read-only

~$ cat /etc/mtab
/dev/hdc1 / ext3 rw,errors=remount-ro 0 0
proc /proc proc rw 0 0
sysfs /sys sysfs rw 0 0
devpts /dev/pts devpts rw,gid=5,mode=620 0 0
tmpfs /dev tmpfs rw,size=10M,mode=0755 0 0
/dev/hda1 /mnt/tmp vfat ro 0 0 # <--this is the win98 mount table entry

so I tried these;

~$ sudo mount -o ro /dev/hda /mnt/tmp
mount: you must specify the filesystem type

~$ sudo mount -t vfat -o ro /dev/hda /mnt/tmp
mount: /dev/hda already mounted or /mnt/tmp busy
mount: according to mtab, /dev/hda1 is already mounted on /mnt/tmp

Big Note; hda is NOT listed in my /etc/fstab which likely plays a role
in this unmounting ability too - but I can recall this umounting pita
also occuring when there were entries in /etc/fstab for hda - and even
when the entries were made to load, but not mount upon boot;
(something like this)
# <file system> <mount point> <type> <options> <dump> <pass>
/dev/hda1 /mnt/hda1 win95 noauto,nouserid,noexec 0 0

I even tried remounting read-only, which is what I wanted anyway. Using
'mount -r' does indeed remount read-only (see my last 2 mount commands
above in this post). I always get device is busy (no worry, for another
time and place ;-))

Oh - I solved the mount issue it seems...It was stupid KDE - once I
closed it out, right before shutdown/reboot, I successfully umounted
hda1 - no more entry in /etc/mtab. Though I still have a issues with a
bad superblock (or unrecognized filesystem) on that hda.
~$ sudo mount -t vfat -o ro /dev/hda /mnt/tmp
mount: wrong fs type, bad option, bad superblock on /dev/hda,
missing codepage or other error
In some cases useful info is found in syslog - try
dmesg | tail or so

....yeah ...yeah I know - but hda has my working win98 system (and a
really munged Debian install leftover, I think)

~$ dmesg | tail
FAT: invalid media value (0xb9)
VFS: Can't find a valid FAT filesystem on dev hda.
FAT: invalid media value (0xb9)
VFS: Can't find a valid FAT filesystem on dev hda.
FAT: bogus number of reserved sectors
VFS: Can't find a valid FAT filesystem on dev hda2.
FAT: bogus number of reserved sectors
VFS: Can't find a valid FAT filesystem on dev hda3.
FAT: invalid media value (0xb9)
VFS: Can't find a valid FAT filesystem on dev hda.

Well -- that's finally done and solved...what a pita :-/

[ /mount dilemma ]

Back on Topic;
Details of win98 IPCONFIG (actually Winipcfg) after closing out Debian
and booting into win98
=====================================
/////////////////////////////////////
Windows 98 IP Configuration

Host Name . . . . . . . . . : FUCKYOUBILLYRAT <--no joke
DNS Servers . . . . . . . . :
Node Type . . . . . . . . . : Broadcast
NetBIOS Scope ID. . . . . . :
IP Routing Enabled. . . . . : No
WINS Proxy Enabled. . . . . : No
NetBIOS Resolution Uses DNS : No

Ethernet adapter :

Description . . . . . . . . : NETGEAR FA31X PCI Adapter
Physical Address. . . . . . : 00-00-00-00-00-00 <-- Note No MAC addy
DHCP Enabled. . . . . . . . : Yes
IP Address. . . . . . . . . : 169.254.196.62 <--here's the ratbag's
idea
Subnet Mask . . . . . . . . : 255.255.0.0
Default Gateway . . . . . . :
DHCP Server . . . . . . . . : 255.255.255.255
Primary WINS Server . . . . :
Secondary WINS Server . . . :
Lease Obtained. . . . . . . : 06 10 06 5:27:00 PM
Lease Expires . . . . . . . :


==== Below is a Working config - After the 3 finger salutation =======
(or just a Start Menu | Restart)

Windows 98 IP Configuration

Host Name . . . . . . . . . : FUCKYOUBILLYRAT
DNS Servers . . . . . . . . : xxx.206.251.15
xxx.206.251.16
xxx.206.251.79
^^^^^
xxx's used in place of legitmate working 1st octet
Node Type . . . . . . . . . : Broadcast
NetBIOS Scope ID. . . . . . :
IP Routing Enabled. . . . . : No
WINS Proxy Enabled. . . . . : No
NetBIOS Resolution Uses DNS : No

Ethernet adapter :

Description . . . . . . . . : NETGEAR FA31X PCI Adapter
Physical Address. . . . . . : 00-0x-xx-xx-xx-xx <--Real MAC addy x-ed
out
DHCP Enabled. . . . . . . . : Yes
IP Address. . . . . . . . . : 192.168.0.2 <--#PC addy DHCP via Router
Subnet Mask . . . . . . . . : 255.255.255.0
Default Gateway . . . . . . : 192.168.0.1 <--#Router addy
DHCP Server . . . . . . . . : 192.168.0.1 <--#Router addy
Primary WINS Server . . . . :
Secondary WINS Server . . . :
Lease Obtained. . . . . . . : 06 10 06 5:30:46 PM
Lease Expires . . . . . . . : 06 13 06 5:30:46 PM

/////////////////////////////////////
=====================================

more Drats - how'd my last post get s-o-o-o... darn l-o-n-g :-$ ????
and now this one too :-(

Regards

On 11 Jun 2006 15:52:57 -0700, iforone wrote:
>
> Like an idiot, I mounted hda1, when I needed to mount the whole hda
> (all 3 win98 partitions)....but I *should* be able to just easily
> umount it and remount the whole hda - but n-o-o-o :-( woe is me...

Then try this

mkdir /p1
mkdir /p2
mkdir /p3

mount -t auto /dev/hda1 /p1
mount -t auto /dev/hda2 /p2
mount -t auto /dev/hda3 /p3

ls /p1
ls /p2
ls /p3

Bit Twister wrote:
> On 11 Jun 2006 15:52:57 -0700, iforone wrote:
> >
> > Like an idiot, I mounted hda1, when I needed to mount the whole hda
> > (all 3 win98 partitions)....but I *should* be able to just easily
> > umount it and remount the whole hda - but n-o-o-o :-( woe is me...
>
> Then try this
>
> mkdir /p1
> mkdir /p2
> mkdir /p3
>
> mount -t auto /dev/hda1 /p1
> mount -t auto /dev/hda2 /p2
> mount -t auto /dev/hda3 /p3
>
> ls /p1
> ls /p2
> ls /p3

Very Nice :-) Thanks a ton!

I did;
~$ sudo mkdir /mnt/p1
~$ sudo mkdir /mnt/p2
~$ sudo mkdir /mnt/p3

then;
$ sudo mount -t auto /dev/hda1 /mnt/p1
$ sudo mount -t auto /dev/hda5 /mnt/p2
$ sudo mount -t auto /dev/hda4 /mnt/p3


Then to verify - I did;
~$ ls -al /mnt/p1
~$ ls -al /mnt/p2
~$ ls -al /mnt/p3

and viewed the partitions within Konquerer -- YES

I think I'm wrong about KDE being the umount culprit though;
I wrote earlier;
> Oh - I solved the mount issue it seems...It was stupid KDE - once I
> closed it out, right before shutdown/reboot, I successfully umounted
> hda1 - no more entry in /etc/mtab.

It seems I need to completely Logout of X (using Kmenu, Logout) and
then choose "Console Login" from kdm's menu, which enters one into text
mode only (runlevel2(?) or is it 1 in Debian(?)) -- anyway - from
there, I Login as 'root', supply passwd, then a 'umount' does the
trick;

# umount /dev/hda1

So ....do I understand that one mounts "partitions' only ?? One at a
time? On separate mount points? I understand that one mounts "file
systems*, not devices, as written in the 'man mount' pages -- nut do
you think because the Superblock is a bit munged, and/or b/c hda
contains multiple file systems - that I need to mount 'partitions'
only? and cannot mount the whole HD in one shot?

~$ sudo fdisk -l /dev/hda

Disk /dev/hda: 82.3 GB, 82348277760 bytes
255 heads, 63 sectors/track, 10011 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System
/dev/hda1 * 1 1275 10241406 c W95 FAT32 (LBA)
/dev/hda2 1276 3188 15366172+ f W95 Ext'd (LBA)
/dev/hda3 3189 5620 19535040 83 Linux
/dev/hda4 5621 10011 35270707+ b W95 FAT32 <--
See Notations below
/dev/hda5 1276 3188 15366141 b W95 FAT32

[Notations]
This hda4 is a copy (and a Primary partition, as most know) of a win98
partition done while Debian installation - using it's "Copy partition"
feature - and note that I canNOT ScanDisk that partition (Scandskw.exe)
while in win98, nor DOS's version(Scandisk.exe) - it won't do it, it
complains of lack of RAM, and something else weird. Here's the thing
though, I can easily access (read / write) that partition's data (hda4)
while in booted into either Debian, -or- win98. I have that part
mounted right now, in Debian (as you can tell from my above 'mount'
commands - thanks to your advise :-))

Regards

On 11 Jun 2006 16:36:06 -0700, iforone wrote:
>
> So ....do I understand that one mounts "partitions' only ?? One at a
> time? On separate mount points?

That is how I misunderstand it

> and cannot mount the whole HD in one shot?

That is how I misunderstand it.

But now you can unmount them with
umount /mnt/p*

Bit Twister wrote:
> On 11 Jun 2006 16:36:06 -0700, iforone wrote:
> >
> > So ....do I understand that one mounts "partitions' only ?? One at a
> > time? On separate mount points?
>
> That is how I misunderstand it

Ahhhh ;-)

> > and cannot mount the whole HD in one shot?
>
> That is how I misunderstand it.
>
> But now you can unmount them with
> umount /mnt/p*

Thanks again...some notes for those following along;
I notice that the way I mounted made each partition RW, but I wanted RO
~$ mount
/dev/hdc1 on / type ext3 (rw,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
tmpfs on /dev type tmpfs (rw,size=10M,mode=0755)
/dev/hda1 on /mnt/p1 type vfat (rw)
/dev/hda5 on /mnt/p2 type vfat (rw)
/dev/hda4 on /mnt/p3 type vfat (rw)

So - here we go again ;-) But this time I managed to do it...

I was trying to Remount 'ro' using the 'mount' command with it's
options - but it seems that this simple command does the trick;
~$ sudo umount -r /mnt/p1
umount: /dev/hda1 busy - remounted read-only - but /etc/mtab still said
RW

and for some reason, right after that I decided to try this
$ sudo umount -r /dev/hda1 /mnt/p1
umount: /dev/hda1 busy - remounted read-only
umount: /dev/hda1 busy - remounted read-only
because 'cat /etc/mtab' was still reporting RW - I guess the /dev/hd*
_and_ the mount point
need to be given in the command?

Ahhh yes;
~$ cat /etc/mtab
[...]
/dev/hda1 /mnt/p1 vfat ro 0 0 <--Note the RO after issuing above
command
/dev/hda5 /mnt/p2 vfat rw 0 0
/dev/hda4 /mnt/p3 vfat rw 0 0

So I remounted the the other 2;
~$ sudo umount -r /dev/hda5 /mnt/p2
and
~$ sudo umount -r /dev/hda4 /mnt/p3

and voila;
~$ cat /etc/mtab
[...]
/dev/hda1 /mnt/p1 vfat ro 0 0
/dev/hda5 /mnt/p2 vfat ro 0 0
/dev/hda4 /mnt/p3 vfat ro 0 0

I could have (and can still) made the mount points more logically
ordered;
IOW - make hda4 = /mnt/p4, hda5 = /mnt/p5 , etc...

Do you know of a way to use 'mount' to remount RO, rather than using
'umount -r' (which is fine by me, but trying various 'mount' (with
various options), and it not working, has me wanting to know, what I'm
doing wrong).
Thanks again

Regards