Re: [expert] Network card causes modem to disappear
-BEGIN PGP SIGNED MESSAGE- Fixed it! The clue was buried in the serial HOWTO. Not explicitly, mind you. But it suggested running "lspci -vv" in regard to seeing how the PCI cards were really configured. When I did that I saw that the modem card was using I/O 0xe000 instead of 0xec00 (which is what Windows says, and which worked fine in Linux without the network card installed). Anyway, putting that I/O address in the setserial call in rc.local did the trick. Can't say I fully understand why any of this was necessary, but at least I now have a functioning modem in Linux again. Doc Evans -BEGIN PGP SIGNATURE- Version: PGP 6.0.2 -- QDPGP 2.60 Comment: Key obtainable from servers: ID 0x6184B81D iQCVAwUBOQYnRv2CFbFhhLgdAQFnWwQAlXC5aJ/U+9XR6jHn8Sq79jQOHVyLbZOy ZSqdSl4/wMlxk3U0NgKyW+vUtfHWtbflHYan7ipi166Xvs2dcEk1UEw4y26TT42K /dlUIAZ7P2j2q2tJgwCNzk0u6vad7zMNWZUVzz3pfQ72tPWhL1llV2uNd28wYznA bm7rVf8TOzU= =IaUr -END PGP SIGNATURE- -- D.R. Evans N7DR / G4AMJ [EMAIL PROTECTED] "Palindor Chronicles" information and extracts: http://www.sff.net/people/N7DR/drevans.htp --
Re: [expert] Network card causes modem to disappear
Charles Curley wrote: > > On Mon, Apr 24, 2000 at 08:29:33AM -0600, D. R. Evans wrote: > -> -BEGIN PGP SIGNED MESSAGE- > -> > -> On 23 Apr 00, at 18:03, Charles Curley wrote: > -> > -> > -> > And in which case the card is out of spec. The PCI spec requires that > -> > software be able to assign a base address to the card in all address > -> > spaces. > -> > > -> > -> The card does appear to be in spec; I can indeed assign base addresses at > -> will. Also, the important thing to remember in all this is that the card > -> works perfectly under Windows with the network card also in place. > > OK, at least in this respect the card appears to be in spec. I recommend > that you not specify any base address if you can avoid it. This will help > the configuation software when you change your hardware around. > > The fact that the card works under Windows does not imply that it is in > spec. It is possible that Windows handles out of spec cards (by, perhaps, > being out of spec) where Linux (perhaps by rigourously complying with the > spec) does not. Also, chances are the Windows driver was written by the > hardware vendor, and could have code in it to compensate for the card's > being out of spec. > > I have not run across any instances of an out-of-spec card running under > Windows, but then I haven't looked for any. Most of my PCI work I did on > HP PA-RISC computers, and I had access to the engineers to verify spec > compliance (which was part of my job at the time). > > -> > -> Either I'm missing something in my trying to configure it under Linux or, > -> and I almost hate to say this) there's something slightly hosed in the PCI > -> PnP support on Linux. At this point I've tried every configuration trick I > -> can think of -- hence my call for help. > > Sorry, I'm out of ideas. > Sheesh, to make arguments on the basis of such data is the cause of jihads. let's see some output and some files like the configurations, and ~/kppprc, and maybe cat /proc/pci. We haven't begun to dig What network card and what modem? It is hard to suggest ideas and such when working in the dark Civileme
Re: [expert] Network card causes modem to disappear
On Mon, Apr 24, 2000 at 08:29:33AM -0600, D. R. Evans wrote: -> -BEGIN PGP SIGNED MESSAGE- -> -> On 23 Apr 00, at 18:03, Charles Curley wrote: -> -> -> > And in which case the card is out of spec. The PCI spec requires that -> > software be able to assign a base address to the card in all address -> > spaces. -> > -> -> The card does appear to be in spec; I can indeed assign base addresses at -> will. Also, the important thing to remember in all this is that the card -> works perfectly under Windows with the network card also in place. OK, at least in this respect the card appears to be in spec. I recommend that you not specify any base address if you can avoid it. This will help the configuation software when you change your hardware around. The fact that the card works under Windows does not imply that it is in spec. It is possible that Windows handles out of spec cards (by, perhaps, being out of spec) where Linux (perhaps by rigourously complying with the spec) does not. Also, chances are the Windows driver was written by the hardware vendor, and could have code in it to compensate for the card's being out of spec. I have not run across any instances of an out-of-spec card running under Windows, but then I haven't looked for any. Most of my PCI work I did on HP PA-RISC computers, and I had access to the engineers to verify spec compliance (which was part of my job at the time). -> -> Either I'm missing something in my trying to configure it under Linux or, -> and I almost hate to say this) there's something slightly hosed in the PCI -> PnP support on Linux. At this point I've tried every configuration trick I -> can think of -- hence my call for help. Sorry, I'm out of ideas. -- -- C^2 No windows were crashed in the making of this email. Looking for fine software and/or web pages? http://w3.trib.com/~ccurley
Re: [expert] Network card causes modem to disappear
-BEGIN PGP SIGNED MESSAGE- On 23 Apr 00, at 18:03, Charles Curley wrote: > And in which case the card is out of spec. The PCI spec requires that > software be able to assign a base address to the card in all address > spaces. > The card does appear to be in spec; I can indeed assign base addresses at will. Also, the important thing to remember in all this is that the card works perfectly under Windows with the network card also in place. Either I'm missing something in my trying to configure it under Linux or, and I almost hate to say this) there's something slightly hosed in the PCI PnP support on Linux. At this point I've tried every configuration trick I can think of -- hence my call for help. Doc Evans -BEGIN PGP SIGNATURE- Version: PGP 6.0.2 -- QDPGP 2.60 Comment: Key obtainable from servers: ID 0x6184B81D iQCVAwUBOQRaTP2CFbFhhLgdAQEfjgP7BJQg6vTe4UDyowaWm8GUQQ9ZmPkYvdoO +ocBHsexx9GEptM69v9n/lq/t2OqmcTFEq6cAq3vwx4dXpPPTeiD9C3LgBHWESp1 BRWCnUrLHZbsVZegCxObQwf7ZYhKlex+mbzhI0s5XEUr+4cmDup1Q2iNsdnNmub1 6lGVurebKPA= =Xacv -END PGP SIGNATURE- -- D.R. Evans N7DR / G4AMJ [EMAIL PROTECTED] "Palindor Chronicles" information and extracts: http://www.sff.net/people/N7DR/drevans.htp --
Re: [expert] Network card causes modem to disappear
On Wed, Apr 19, 2000 at 07:59:17AM +0200, Jean-Louis Debert wrote: -> "D. R. Evans" wrote: -> > So I'm wondering, is there any way to get a device-by-device listing of -> > what's occupying the various IRQs? (One hypothesis being that something -> > else was sitting on IRQ 5 and moved to IRQ 10 when the network card -> > appeared and occupied IRQ 5. I know that the PCI bus is not supposed to -> > allow two devices to occupy the same IRQ, but I'm grasping at straws.) -> -> -> You have that under /proc -> -> Note that PCI bus _IS_ supposed to allow to devices to occupy -> the same IRQ (there is arbitration _if needed_, i.e. if either -> of the devices indicates that it _CAN'T_ share a resource, -> which would probably be the case for either modem or ethernet, -> because of timing considerations). The PCI 2.x specs require that all devices daisy chain interrupt drivers and be able to share resources. This implies sufficient buffering to last a while between interrupt services, and encourages the use of bus mastering for delivering the data. If SCSI host adapters can share an interrupt, modems and ethernet for dang sure should be able to. -> -> Besides, did you consider that it might _NOT_ be the IRQ, -> but also a possible I/O (including memory mapped I/O) port -> or address, that could cause the conflict, and/or confuse -> the driver because the wrong device would answer ? -> The PCI arbitration is supposed to resolve this too, but -> maybe the devices _CAN'T_ use any other address, in which -> case you are screwed ... And in which case the card is out of spec. The PCI spec requires that software be able to assign a base address to the card in all address spaces. -> -> -> -- -> Jean-Louis Debert[EMAIL PROTECTED] -> 74 Annemasse France -> old Linux fan -- -- C^2 No windows were crashed in the making of this email. Looking for fine software and/or web pages? http://w3.trib.com/~ccurley
Re: [expert] Network card causes modem to disappear
-BEGIN PGP SIGNED MESSAGE- On 19 Apr 00, at 7:59, Jean-Louis Debert wrote: > "D. R. Evans" wrote: > > So I'm wondering, is there any way to get a device-by-device listing of > > what's occupying the various IRQs? (One hypothesis being that something > > else was sitting on IRQ 5 and moved to IRQ 10 when the network card > > appeared and occupied IRQ 5. I know that the PCI bus is not supposed to > > allow two devices to occupy the same IRQ, but I'm grasping at straws.) > > > You have that under /proc > Here is what I see in /proc/interrupts CPU0 0: 152398 XT-PIC timer 1:493 XT-PIC keyboard 2: 0 XT-PIC cascade 3: 0 XT-PIC es1371 8: 1 XT-PIC rtc 12: 11288 XT-PIC PS/2 Mouse 13: 1 XT-PIC fpu 14: 100956 XT-PIC ide0 15:251 XT-PIC ide1 NMI: 0 Interestingly, no IRQ 10 (for the modem) or IRQ 5 (for the network card). I imagine, though, that these are the interrupts that have actually occurred since booting. I hadn't brought up the network card, so no IRQ 5; and since I can't get to the modem I suppose I shouldn't be surprised (should I?) that IRQ 10 doesn't appear. FWIW, here is /proc/ioports: -001f : dma1 0020-003f : pic1 0040-005f : timer 0060-006f : keyboard 0070-007f : rtc 0080-008f : dma page reg 00a0-00bf : pic2 00c0-00df : dma2 00f0-00ff : fpu 0170-0177 : ide1 01f0-01f7 : ide0 0376-0376 : ide1 03c0-03df : vga+ 03f6-03f6 : ide0 03f8-03ff : serial(auto) d000-d007 : ide0 d008-d00f : ide1 d800-d83f : es1371 dc00-dc1f : eth0 ec00-ec07 : serial(set) The last entry comes from this in my rc.local file: # DRE setserial -v /dev/ttyS4 port 0xec00 uart 16550A irq 10 ^fourport - - What is the name of the serial driver? When I look at the loaded modules I see: ne2k-pci4256 0 (autoclean) (unused) 83906020 0 (autoclean) [ne2k-pci] lockd 33256 1 (autoclean) sunrpc 56612 1 (autoclean) [lockd] nls_iso8859-1 2052 2 (autoclean) nls_cp437 3580 2 (autoclean) vfat 11004 2 (autoclean) fat32640 2 (autoclean) [vfat] ide-scsi7584 1 supermount 14880 3 (autoclean) es1371 26308 0 soundcore 3524 4 [es1371] (this was after bringing up the eth0 interface, which is why it's listed). I expected to see something I could identify as a serial driver, but it's not obvious. One other thing that probably doesn't mean anything, at boot time I see a message that sits on my screen for about 1 millisecond, so I can't read it, but it contains the string "ttyS4". Looking in the log in dmesg or /var/tmp/messages doesn't help, because that particular message doesn't appear; maybe there's some other place I could look so I can get a chance to read the message? I don't think it's an error message, but it would be nice to be sure. > Note that PCI bus _IS_ supposed to allow to devices to occupy > the same IRQ (there is arbitration _if needed_, i.e. if either > of the devices indicates that it _CAN'T_ share a resource, > which would probably be the case for either modem or ethernet, > because of timing considerations). > I'd forgotten that, but of course you're right. (It's been a couple of years since I last looked at the PCI specs.) > Besides, did you consider that it might _NOT_ be the IRQ, > but also a possible I/O (including memory mapped I/O) port > or address, that could cause the conflict, and/or confuse > the driver because the wrong device would answer ? Yes. But the ioports look OK as well, per the list above. Doc Evans -BEGIN PGP SIGNATURE- Version: PGP 6.0.2 -- QDPGP 2.60 Comment: Key obtainable from servers: ID 0x6184B81D iQCVAwUBOP93Yf2CFbFhhLgdAQHw8AP+PnpU/D1WQWLd9uOXD3zueBi210gLvu2h KzjexG1GtrZIp5ZNfrvNDG1A7vZCPg0qTBnqeiG+UEJNrXX+j6ATNhyGjAKSm/uz t/3FMqyn+y7De3OnQfBExnAyRYqJ1wvCw7sVaM34q1TuUNbqVSsH4Oh0wyfA2bzJ IzoXN0za8qc= =2vOP -END PGP SIGNATURE- -- D.R. Evans N7DR / G4AMJ [EMAIL PROTECTED] "Palindor Chronicles" information and extracts: http://www.sff.net/people/N7DR/drevans.htp --
Re: [expert] Network card causes modem to disappear
"D. R. Evans" wrote: > So I'm wondering, is there any way to get a device-by-device listing of > what's occupying the various IRQs? (One hypothesis being that something > else was sitting on IRQ 5 and moved to IRQ 10 when the network card > appeared and occupied IRQ 5. I know that the PCI bus is not supposed to > allow two devices to occupy the same IRQ, but I'm grasping at straws.) You have that under /proc Note that PCI bus _IS_ supposed to allow to devices to occupy the same IRQ (there is arbitration _if needed_, i.e. if either of the devices indicates that it _CAN'T_ share a resource, which would probably be the case for either modem or ethernet, because of timing considerations). Besides, did you consider that it might _NOT_ be the IRQ, but also a possible I/O (including memory mapped I/O) port or address, that could cause the conflict, and/or confuse the driver because the wrong device would answer ? The PCI arbitration is supposed to resolve this too, but maybe the devices _CAN'T_ use any other address, in which case you are screwed ... -- Jean-Louis Debert[EMAIL PROTECTED] 74 Annemasse France old Linux fan
Re: [expert] Network card causes modem to disappear
Trying to think logically about this so that I can generate a model of what is going on. If I can use the modem with no network card in the box, then that seems a good test that at some level everything is OK. But if the modem goes away when I add the network card, then that seems to indicate that the presence of the network card somehow conflicts with the modem. What else do I know? Well, I know that the modem card is on IRQ 10 regardless of whether the network card is present. I also know that the network card is on IRQ 5. So I'm wondering, is there any way to get a device-by-device listing of what's occupying the various IRQs? (One hypothesis being that something else was sitting on IRQ 5 and moved to IRQ 10 when the network card appeared and occupied IRQ 5. I know that the PCI bus is not supposed to allow two devices to occupy the same IRQ, but I'm grasping at straws.) FWIW I have the following line in /etc/rc.d/rc.local: setserial /dev/ttyS4 port 0xec00 uart 16550A irq 10 ^fourport Does this look reasonable? If I use auto-irq instead of setting it explicitly to IRQ 10, setserial does report that 10 is reported back by the port, so explicitly setting it to 10 looks like it should be OK, although I suppose I could always experiment and do an auto-irq probe at boot time. You'll notice that I'm still not prepared (yet) to believe that this is a WinModem; I just can't fathom how it could ever have worked if it were a WinModem. And of course, the manufacturer says I have to go to Micron for support, and Micron says it supports only the OS that was on it the time at which the box was sold. They do confirm that this is not a WinModem -- if we believe them (actually, though, I've found Micron support to be very good in the past, that's the principal reason I bought from them again this time). Doc Evans -- D.R. Evans N7DR / G4AMJ [EMAIL PROTECTED] "Palindor Chronicles" information and extracts: http://www.sff.net/people/N7DR/drevans.htp --
Re: [expert] Network card causes modem to disappear
On Mon, 17 Apr 2000, you wrote: > > As I understand it, the modem gets configured to IRQ 10 during power-on by > the PCI gubbins, so there's no way to control that. > It's a PCI modem??? Do this...go look and make SURE you don't have to add any extra software for it to work under Windows. Have you tried dialing out from a plain, F8 boot? (i.e. at the boot prompt, hit your F8 key to NOT load *any* drivers.) I'm not sure how you'd do this with LILO. Maybe just have a generic Dos / Windows boot floppy with command.com on it and boot off that? Anyway, once you've gotten to a no-drivers-loaded boot, type "echo atdt[some-phone] >comX" (where X=com port of your alleged modem.) AFAIK, there is ALMOST no such thing as a PCI modem which is NOT a WinModem. I think you got ripped. > > And it really does work perfectly without the > network card in there; my > understanding is that a > WinModem wouldn't be accessible at all, since there > > wouldn't be a driver for it anywhere in the system. > > Yes? No? > That's the theory anyway... John
Re: [expert] Network card causes modem to disappear
-BEGIN PGP SIGNED MESSAGE- On 17 Apr 00, at 11:44, John Aldrich wrote: > On Mon, 17 Apr 2000, you wrote: > > The modem is on IRQ 10, and the network card is on IRQ5. The I/O > > addresses are also nicely different. > > > IRQ10 would seem to indicate a WinModem. Are you SURE it's > not a WinModem? If it is, you will ALWAYS have trouble > accessing it, even if it *appears* to be working. > John Yes, it's not a Winmodem -- although if anyone has a cast-iron way for me to tell for 100% sure, that would be nice. I specifically changed the order from the default (which is, of course, a Win-pseudo-modem) and they charged me for a real modem; the documentation (such as it is) also looks like it belongs to a real USRobotics/3COM modem. As I understand it, the modem gets configured to IRQ 10 during power-on by the PCI gubbins, so there's no way to control that. And it really does work perfectly without the network card in there; my understanding is that a WinModem wouldn't be accessible at all, since there wouldn't be a driver for it anywhere in the system. Yes? No? Doc Evans -BEGIN PGP SIGNATURE- Version: PGP 6.0.2 -- QDPGP 2.60 Comment: Key obtainable from servers: ID 0x6184B81D iQCVAwUBOPtiHv2CFbFhhLgdAQH1oAP/UrZK9Ukv2vG3tqEOJLBLH2fqEaPK3XTw l3n28WfxrSB0uhMls3ro0QdLsTAiIW9cDotifNqEizGYl7pBZWY0jymN8zmENQnJ 8XnlWiJjqr8WOY9VatKLbIHd3x+T4f8BVPgGmQrQFOIn59q1zZcyrloWSo8UMhtr tkhCiOvg3C4= =Di/A -END PGP SIGNATURE- -- D.R. Evans N7DR / G4AMJ [EMAIL PROTECTED] "Palindor Chronicles" information and extracts: http://www.sff.net/people/N7DR/drevans.htp --
Re: [expert] Network card causes modem to disappear
On Mon, 17 Apr 2000, you wrote: > The modem is on IRQ 10, and the network card is on IRQ5. The I/O addresses are also > nicely different. > IRQ10 would seem to indicate a WinModem. Are you SURE it's not a WinModem? If it is, you will ALWAYS have trouble accessing it, even if it *appears* to be working. John
[expert] Network card causes modem to disappear
I've tried to post this several times already over the course of the past few days, but I have seen nothing appear on the reflector. Haven't received any bounce messages either, so I don't know where the messages think they've been going. Anyway, apologies if you've seen this before. - A nasty problem 1. Running Mandrake 7.0 without any fancy modifications on a 733 MHz Micron box. 2. I can access the modem just fine (e.g. for PPP, etc.) The modem is on ttyS4 (COM5). [ Don't ask me, that's the way it came in WinDo$e, so I put it on ttyS4 for Linux. ] I have /dev/modem soft linked to /dev/ttyS4. 3. I add a Linksys Ether PCI II card (no other NICs in the box). 4. Suddenly I can no longer access the modem. At all. In any program. kppp, for example, says "searching for modem" and then "modem is busy". Even statserial doesn't work (complains about TIOCMGET not being supported, although of course it works fine when the network card is not in the box). 5. If I take the card out, I can access the modem again. 6. Even if I tell Kudzu not to configure the network card at boot time (i.e. the third option of the three that are presented), I STILL can't get to the modem if the network card is phsically inserted in a slot. 7. The really bad news is that everything works fine under WinDo$e. Actually, I suppose that's not really so bad, because it at least indicates that there's no fundamental hardware problem. So, any ideas as to how I might be able to have a network and a modem at the same time would be most welcome. Or even how to diagnose the problem. Oh, some extra info: The modem is on IRQ 10, and the network card is on IRQ5. The I/O addresses are also nicely different. And the network card appears to work OK (although I haven't "torture-tested" it, having been too busy trying to figure out why I don't have a modem any more). I have a separate serial port on COM1 (on the motherboard). It works fine regardless of whether the network card is in place. Doc Evans -- D.R. Evans N7DR / G4AMJ [EMAIL PROTECTED] "Palindor Chronicles" information and extracts: http://www.sff.net/people/N7DR/drevans.htp --