Re: modem/isapnp trouble SOLVED
I solved this by deactivating my script that was attempting to set the
isapnp manually. The modem did end up at ttyS0 (vs the old ttyS2), so
I adjusted /etc/ppp/peers/provider accordingly.
Interestingly, my soundcard (isapnp Soundblaster) also seems to have
ended up using new resources; at least the resources shown by
/proc/isapnp no longer match those in /etc/alsa/modutils/1.0. But it
still works! I think I saw some messages fly by that my settings were
being ignored.
I was never clear on whether specifying IO, IRQ, etc was necessary or
a good idea for ALSA, since it claimed to be PnP. I think at one
point it was necessary, but it seems it no longer is.
The other tidbit I uncovered is that /proc/interrupts only shows
interrupts in use. So if a driver is inactive, the interrupt may be
"reserved" but not show up on /proc/interrupts. Apparently the driver
needs to be in active use, not just loaded (but perhaps this is
because it gets unloaded if not in use?).
On Tue, Aug 10, 2004 at 01:20:04PM -0700, Ross Boylan wrote:
> I am unable to activate my dial-up modem since I rebooted yesterday,
> August 9. I have changed neither my hardware, BIOS, kernel, or boot
> procedure since my last successful boot on July 26. I have updated my
> software reflecting changes in testing since then.
>
> I have tried several things to diagnose and solve the problem, so far
> without result. If anyone could offer me any guidance, I'd be be very
> grateful.
>
> On the same machine, the modem works under Windows 98 but Win2k can't
> connect to it. It used to work under all systems.
>
> Running on Intel PC style hardware, Athlon CPU, Gigabyte 7IXE4 Rev 1.0
> motherboard, AMIBIOS F9d, USRobotics internal ISA PNP modem. The
> first COM port is disabled in BIOS. There is a general shortage of
> interrupts. Soundblaster card also ISAPNP. I use devfs and evms. I
> think I have my BIOS set for non-PNP OS.
>
> The immediate symptom of the problem was that /dev/ttyS2 disappeared.
> On boot serial driver 5.05c shows
> July 26 now
> ttyS01 ttyS01 0x02f8 IRQ 3
> ttyS02 ttyS00 0x03e8 IRQ 5
> i.e., ttyS02 seems to have become ttyS00. However, changing
> /etc/ppp/peers/provider to use /dev/ttyS0 instead of /dev/ttyS2 does
> not help:
>
> Aug 10 10:42:44 wheat pppd[3869]: pppd 2.4.2 started by ross, uid 1000
> Aug 10 10:42:45 wheat chat[3870]: Can't get terminal parameters:
> Input/output error
> Aug 10 10:42:45 wheat pppd[3869]: Connect script failed
> Aug 10 10:42:46 wheat pppd[3869]: Exit.
>
> My main suspicion centers on trouble with isapnp. Other candidates
> include devfs, changes in the system initialization procedure from
> software upgrades, a crash induced change in the hardware, and random
> drift in the pnp settings.
>
> I'll take these in turn.
>
> ISAPNP
>
>
> On closer inspection, it seemed that the soundcard had grabbed the IRQ
> $ cat /proc/interrupts
> CPU0
> 0: 458829 XT-PIC timer
> 1: 3634 XT-PIC keyboard
> 2: 0 XT-PIC cascade
> 3: 0 XT-PIC serial
> 5: 99 XT-PIC SoundBlaster
> 8: 1 XT-PIC rtc
> 11: 390 XT-PIC usb-ohci, eth0
> 12: 19545 XT-PIC PS/2 Mouse
> 14: 48819 XT-PIC ide0
> 15: 371190 XT-PIC ide1
> NMI: 0
> LOC: 458835
> ERR: 96
> I tried disabling the soundcard via isapnp. Doing this before setting
> the modem, and then enabling the soundcard, had no effect. Disabling
> it entirely caused system initialization to hang at "enabling ALSA
> settings". Fortunately, single user mode still started.
>
> However, the information on the IRQ of the modem is inconsistent:
> # cat /proc/tty/driver/serial
> serinfo:1.0 driver:5.05c revision:2001-07-08
> 0: uart:16550A port:3F8 irq:4 tx:0 rx:0 CTS|DSR|CD|RI
> 1: uart:16550A port:2F8 irq:3 baud:2400 tx:0 rx:0 RTS
>
> That's IRQ 4, despite the fact that I have a startup script, executing
> after the standard S15isapnp (at S25), that cat's the following to
> /proc/isapnp:
> card 0 USR3070
> dev 0 USR3070
> irq 0 5
> port 0 0x03e8
> activate
> card 0 CTL00e4
> dev 0 CTL0045
> activate
> dev 0 CTL7002
> deactivate
> dev 0 CTL0022
> activate
>
> Note the irq 0 5, i.e., set the modem to use IRQ 5.
>
> Further, cat /proc/isapnp shows the modem is at IRQ 5, along with the
> soundcard (I have edited out some of the supplementary info):
> Card 1 'CTL00e4:Creative SB AWE64 PnP' PnP version 1.0 Product version
> 1.0
> Logical device 0 'CTL0045:Audio'
> Device is active
> Active port 0x220,0x330,0x388
> Active IRQ 5 [0x2]
> Active DMA 1,5
> Resources 0
> Priority preferred
> Port 0x220-0x220, align 0x0, size 0x10, 16-bit address decoding
> Port 0x330-0x330, align 0x0, size 0x2, 16-bit address decoding
> Port 0x388-0x3f8, align 0x0, size 0x4, 16-bit address decoding
> IRQ 5 High-Edge
> DMA 1 8-bit byte-count compatible
> DMA 5 16-bit word-count compatible
> Logical device 1 'CTL7002:Game'
> Compatible device PNPb02f
> Device is not active
> Logical device 2 'CTL0022:WaveTable'
> Device is active
> Active port 0x620,0xa20,0xe20
> Resources 0
> Priority preferred
> Port 0x620-0x620, align 0x0, size 0x4, 16-bit address decoding
> Port 0xa20-0xa20, align 0x0, size 0x4, 16-bit address decoding
> Port 0xe20-0xe20, align 0x0, size 0x4, 16-bit address decoding
> Card 2 'USR3070:U.S. Robotics 56K Voice INT' PnP version 1.0
> Logical device 0 'USR3070:Unknown'
> Device is active
> Active port 0x3e8
> Active IRQ 5 [0x3]
> Resources 0
> Priority preferred
> Port 0x2f8-0x2f8, align 0x0, size 0x8, 10-bit address decoding
> IRQ 3 High-Edge
> Alternate resources 0:1
> Priority acceptable
> Port 0x3f8-0x3f8, align 0x0, size 0x8, 10-bit address decoding
> IRQ 4 High-Edge
> Alternate resources 0:2
> Priority acceptable
> Port 0x3e8-0x3e8, align 0x0, size 0x8, 10-bit address decoding
> IRQ 3,4,5,7 High-Edge
> Alternate resources 0:3
> Priority acceptable
> Port 0x2e8-0x2e8, align 0x0, size 0x8, 10-bit address decoding
> IRQ 3,4,5,7 High-Edge
>
> I'm not sure what it means that /proc/isapnp and
> /proc/tty/driver/serial contradict each other on the IRQ of the modem.
>
>
> DEVFS
> Since the pnp settings come after the devfsd activation in the boot
> sequence, I'm also not sure how the two are supposed to communicate.
> This leads to my second suspect, devfs. At the initial run of
> /etc/rcS.d/S01devfsd the isapnp scripts haven't run, so I wouldn't
> expect the modem to be properly configured. The kernel is compiled
> with devfs activated at boot time. I have installed
> ii devfsd 1.3.25-16 Daemon for the device filesystem
> However, that was there before, and everything worked. I think it's
> supposed to detect new devices; perhaps it's not.
>
> SOFTWARE CHANGES
> There have recently been a fairly large set of upgrades. In
> particular, initscripts, which was at 2.85-20 and sysvinit, which was
> at 2.85-22, were upgraded. I don't see anything in the changelogs
> that would suggest changes that would affect me, however.
>
> Here are some other possibly relevant updates since the last
> successful boot:
> Preparing to replace base-config 2.35 (using
> .../base-config_2.38_all.deb) ...
> Preparing to replace base-config 2.38 (using
> .../base-config_2.40_all.deb) ...
> Preparing to replace base-files 3.0.16 (using
> .../base-files_3.1_i386.deb) ...
> Preparing to replace coreutils 5.0.91-2 (using
> .../coreutils_5.2.1-2_i386.deb) ...
> Preparing to replace initscripts 2.85-22 (using
> .../initscripts_2.86-1_i386.deb) ...
> Preparing to replace libdevmapper1.00 2:1.00.17-5 (using
> .../libdevmapper1.00_2%3a1.00.19-2_i386.deb) ...
> Preparing to replace ppp 2.4.2+20040202-3 (using
> .../ppp_2.4.2+20040428-2_i386.deb) ...
> Preparing to replace syslinux 2.04-2 (using
> .../syslinux_2.10-1_i386.deb) ...
> Preparing to replace sysv-rc 2.85-22 (using .../sysv-rc_2.86-1_all.deb)
> ...
> Preparing to replace sysvinit 2.85-22 (using
> .../sysvinit_2.86-1_i386.deb) ...
>
>
> CRASH
> I rebooted because the system was badly hung, unresponsive to
> keyboard or mouse. Perhaps that saved, or failed to save, some magic
> data.
>
> PNP DRIFT
> My imperfect understanding of PNP is that it is not entirely
> deterministic, so that the same configuration may not come up twice,
> even if you don't change anything. So perhaps I'm just randomly in a
> bad state. This is one reason I explicitly set things via isapnp, but
> as you can see above my script does not set all elements of all
> devices. And perhaps some of my commands are being ignore as
> inconsistent with the state of other devices (e.g., the soundcard has
> ended up on IRQ 5 for whatever reason, so my request to assign the
> modem to that IRQ doesn't take).
>
> HARDWARE FAILURE?
>
Reply to: