Re: HELP! Install problem. How to free I/O locked by reserve=... in boot params?
On Wed, Mar 08, 2000 at 09:14:27PM +0100, Bjoern Brill wrote: Have you tried to insert the wd module with parameters (io=0x280 etc.)? I think it doesn't care if resources are reserved or not in this case. Hmmm, I used the insmod wd io=0x280, and it resulted in the mentioned error. What error? I couldn't find a verbatim error message in your previous postings. Sorry, I really didn't quote it strictly. It was: Device or resource busy Some ideas about what could be wrong: * I suppose the WD card connects to the ISA bus (would anybody want to produce PCI cards with this ancient chipset?). If so, you may have to tell the BIOS setup the resources the WD uses. They can be taken by other hardware or disabeled otherwise. The IRQ has to be marked as used by ISA card or the like, and you have to enable ISA shared memory or something like that at the correct address (0xcc000, size 16k in my case). Look for these options in the PnP BIOS Setup (or similar) section. You may have to disable auto config before you can change or even see any of this. * Check your /etc/conf.modules (or /etc/modules.conf - the name keeps changing) for conflicting wd options. * Try to use modprobe instead of insmod in case the wd module requires other modularized stuff that doesn't get autoloaded. Today everything started fine, without any problems. I don't know why, maybe someone has changed the BIOS setup, and the busy resource was really an interrupt? The failing machine is my friend's workstation which is used by many people, mainly in Window$, so everything may happen ;-). Anyway thanks a lot. -- Greetings Wojciech M. Zabolotny http://www.ise.pw.edu.pl/~wzab -- [EMAIL PROTECTED] http://www.gnupg.org Gnu Privacy Guard - protect your mail data with the FREE cryptographic system
Re: HELP! Install problem. How to free I/O locked by reserve=... in boot params?
Quoting Wojciech Zabolotny ([EMAIL PROTECTED]): Today everything started fine, without any problems. I don't know why, maybe someone has changed the BIOS setup, and the busy resource was really an interrupt? The failing machine is my friend's workstation which is used by many people, mainly in Window$, so everything may happen ;-). Anyway thanks a lot. When I first ran linux on a PC with DOS, the DEPCA DE100 card would not work unless powered off between boots. (I think it was going linux - DOS, but I can't be sure.) Cheers, -- Email: [EMAIL PROTECTED] Tel: +44 1908 653 739 Fax: +44 1908 655 151 Snail: David Wright, Earth Science Dept., Milton Keynes, England, MK7 6AA Disclaimer: These addresses are only for reaching me, and do not signify official stationery. Views expressed here are either my own or plagiarised.
Re: HELP! Install problem. How to free I/O locked by reserve=... in boot params?
On Fri, 3 Mar 2000 09:56:15 +0100, Wojciech Zabolotny wrote: Hi All, I have to install potato on a computer which locks with standard potato's installation kernel. This machine has a NIC (wd8003) located at 0x280. During the boot time one of the SCSI drivers (probably sim710) included in the standard kernel tries to detect the SCSI controller at the same address, locking the machine. The only solution is to use reserve=0x280,0x20 at the boot prompt. In this case kernel initializes correctly, however I can not insmod the NIC module later to complete the installation :-(. I've read the Boot-Prompt HOWTO, but there was nothing about such cases. I can not change the NIC's iobase. Is there any special boot parameter to disable the autodetection of particular address without blocking it for modules? Is there any way to free such reserved I/O space? Have you tried to insert the wd module with parameters (io=0x280 etc.)? I think it doesn't care if resources are reserved or not in this case. Regards, Bjoern Brill -- Bjorn Brill [EMAIL PROTECTED] Frankfurt am Main, Germany
Re: HELP! Install problem. How to free I/O locked by reserve=... in boot params?
On Wed, Mar 08, 2000 at 03:04:03AM +0100, Bjoern Brill wrote: Have you tried to insert the wd module with parameters (io=0x280 etc.)? I think it doesn't care if resources are reserved or not in this case. Hmmm, I used the insmod wd io=0x280, and it resulted in the mentioned error. However I've checked it in my home system. I've append the reserve=0x300,1 line to the lilo.conf (I have ne2000 at 0x300). In this case system starts correctly, then I switched off the eth0 interface and removed the ne module. After I tried: insmod ne, I got: #insmod ne ne.c: No PCI cards found. Use io=0xNNN value(s) for ISA cards. /lib/modules/2.0.36/net/ne.o: init_module: Device or resource busy However, when I typed insmod ne=0x300, I got: #insmod ne io=0x300 ne.c:v1.10 9/23/94 Donald Becker ([EMAIL PROTECTED]) NE*000 ethercard probe at 0x300: 00 c0 df 48 fa ef eth0: NE2000 found at 0x300, using IRQ 11. The 0x300 area is still reported in /proc/ioports as: 0300-0300 : reserved (even when the module is loaded!) So I can see a few possibilities: 1) The modules ne and wd differently handles the reserved I/O space 2) There is a difference in this subject between 2.0 and 2.2 kernels (I use 2.0.36 at home, and 2.2.14 in the failing machine) 3) This is not I/O space, what causes the resource unavailable error. Unfortunately the insmod gives very poor information about the reasons of failure (even the -v switch does not help). Thanks for suggestions. Tomorrow I'll try to investigate the problem more thoroughly. -- Wojciech Zabolotny http://www.ise.pw.edu.pl/~wzab http://www.freedos.org Free DOS for free people!
Re: HELP! Install problem. How to free I/O locked by reserve=... in boot params?
On Wed, 8 Mar 2000, Wojciech Zabolotny wrote: On Wed, Mar 08, 2000 at 03:04:03AM +0100, Bjoern Brill wrote: Have you tried to insert the wd module with parameters (io=0x280 etc.)? I think it doesn't care if resources are reserved or not in this case. Hmmm, I used the insmod wd io=0x280, and it resulted in the mentioned error. What error? I couldn't find a verbatim error message in your previous postings. However I've checked it in my home system. I've append the reserve=0x300,1 line to the lilo.conf (I have ne2000 at 0x300). In this case system starts correctly, then I switched off the eth0 interface and removed the ne module. After I tried: insmod ne, I got: #insmod ne ne.c: No PCI cards found. Use io=0xNNN value(s) for ISA cards. /lib/modules/2.0.36/net/ne.o: init_module: Device or resource busy However, when I typed insmod ne=0x300, I got: #insmod ne io=0x300 ne.c:v1.10 9/23/94 Donald Becker ([EMAIL PROTECTED]) NE*000 ethercard probe at 0x300: 00 c0 df 48 fa ef eth0: NE2000 found at 0x300, using IRQ 11. The 0x300 area is still reported in /proc/ioports as: 0300-0300 : reserved (even when the module is loaded!) So I can see a few possibilities: 1) The modules ne and wd differently handles the reserved I/O space 2) There is a difference in this subject between 2.0 and 2.2 kernels (I use 2.0.36 at home, and 2.2.14 in the failing machine) 3) This is not I/O space, what causes the resource unavailable error. Unfortunately the insmod gives very poor information about the reasons of failure (even the -v switch does not help). Thanks for suggestions. Tomorrow I'll try to investigate the problem more thoroughly. I have just upgraded a machine with an old WD NIC from 2.0.36 to 2.2.14 and tried to reproduce your problem but can't. Situation: the WD is at 0x300, IRQ 10, mem 0xcc000-0xc. The wd driver is modularized and gets autoloaded. The important lines in /etc/conf.modules are: alias eth0 wd options wdio=0x300 irq=10 mem=0xcc000 Now I boot through LILO and give a reserve=0x300,0x20 argument. The wd module loads just fine, using io=0x300 etc. And it really works - the machine imports /home and lots of other stuff via nfs so I should notice if it doesn't :) The io region is still reserved for the kernel (as you said), cat /proc/ioports yields 0300-031f : reserved So I'd say the reserve argument is not the problem. Some ideas about what could be wrong: * I suppose the WD card connects to the ISA bus (would anybody want to produce PCI cards with this ancient chipset?). If so, you may have to tell the BIOS setup the resources the WD uses. They can be taken by other hardware or disabeled otherwise. The IRQ has to be marked as used by ISA card or the like, and you have to enable ISA shared memory or something like that at the correct address (0xcc000, size 16k in my case). Look for these options in the PnP BIOS Setup (or similar) section. You may have to disable auto config before you can change or even see any of this. * Check your /etc/conf.modules (or /etc/modules.conf - the name keeps changing) for conflicting wd options. * Try to use modprobe instead of insmod in case the wd module requires other modularized stuff that doesn't get autoloaded. Good Luck, Bjoern -- Bjorn Brill [EMAIL PROTECTED] Frankfurt am Main, Germany
HELP! Install problem. How to free I/O locked by reserve=... in boot params?
Hi All, I have to install potato on a computer which locks with standard potato's installation kernel. This machine has a NIC (wd8003) located at 0x280. During the boot time one of the SCSI drivers (probably sim710) included in the standard kernel tries to detect the SCSI controller at the same address, locking the machine. The only solution is to use reserve=0x280,0x20 at the boot prompt. In this case kernel initializes correctly, however I can not insmod the NIC module later to complete the installation :-(. I've read the Boot-Prompt HOWTO, but there was nothing about such cases. I can not change the NIC's iobase. Is there any special boot parameter to disable the autodetection of particular address without blocking it for modules? Is there any way to free such reserved I/O space? -- TIA Wojciech M. Zabolotny http://www.ise.pw.edu.pl/~wzab -- [EMAIL PROTECTED] http://www.gnupg.org Gnu Privacy Guard - protect your mail data with the FREE cryptographic system
