Hi everybody, I've made another attempt, but still no luck. Here are the details...thanks in advance for any suggestions! :-) I stuck my 537EPG modem into a Soekris net5501 (http://www.soekris.com/net5501.htm). This little box runs stock Debian with a stripped down 2.6.24.7 kernel. It has one 3.3V 32-bit PCI slot (3.3V only, no 5V). If you take a look at http://en.wikipedia.org/wiki/File:PCI_Keying.png, my PCI slot has only one key in it, closest to the metal bracket. The 537EPG modem actually has 2 keys in it, indicating it *should* be universal (3.3V or 5V) 32-bit. The same thing happens as with my other machine -- kermit segfaults and dmesg/console shows a stack trace / kernel oops with si3052Init at the top. A command like "echo ATZ > /dev/modem" does the same. minicom simply doesn't work at all (not sure why it doesn't generate this error)...it just acts like the modem is not responsive. I've tried disabling sound in /etc/init.d/537_boot, this did not seem to help. I tried pci=routeirq, pollirq, and also no special options when booting the kernel -- all resulted in the same behavior. One difference I noted is that I don't get the "ACPI: PCI interrupt for device 0000:01:00.0 disabled" in dmesg, as this platform does not support ACPI (acpi_pci_irq_enable also does not show up in the stack trace). Here is the lspci output on the Soekris box: === soekris:~/Modem# lspci -vvv -d 8086:1080 00:0e.0 Modem: Intel Corporation FA82537EP 56K V.92 Data/Fax Modem PCI (rev 04) (prog-if 00 [Generic]) Subsystem: Dell Unknown device 1000 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- Interrupt: pin A routed to IRQ 10 Region 0: Memory at a0005000 (32-bit, non-prefetchable) [size=4K] Region 1: I/O ports at e500 [size=256] Capabilities: [80] Power Management version 2 Flags: PMEClk- DSI- D1- D2- AuxCurrent=55mA PME(D0+,D1-,D2-,D3hot+,D3cold+) Status: D0 PME-Enable- DSel=0 DScale=0 PME- === dmesg output when loading the Intel537 module: === sound disabled Intel 537EP card found 537ep:softcore_init_struct: driver serial already allocated device. 537ep:softcore_init_struct: Unregistering serial driver. === Stack trace / kernel oops (only difference between this trace and the last one I posted is that this one does not have acpi_pci_irq_enable as the Soekris net5501 does not support ACPI): === BUG: unable to handle kernel NULL pointer dereference at virtual address 00000006 printing eip: d0c278b9 *pde = 00000000 Oops: 0000 [#1] Modules linked in: Intel537 Pid: 1589, comm: kermit Not tainted (2.6.24.7-kraba5501 #3) EIP: 0060:[<d0c278b9>] EFLAGS: 00010283 CPU: 0 EIP is at read8+0x19/0x20 [Intel537] EAX: 00000000 EBX: 00000000 ECX: d0fb20a8 EDX: 00000006 ESI: d0ff8540 EDI: 0f000001 EBP: cdb31e00 ESP: cdb31e00 DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068 Process kermit (pid: 1589, ti=cdb30000 task=cfc2b570 task.ti=cdb30000) Stack: cdb31e30 d0c25df0 d0ff8540 00000006 cdb31e30 d0c26197 cdb31e90 00000000 d0ffaaa0 d0fb2098 cfe351a0 0f000001 cdb31e80 d0c24de3 d0ff8540 cfeee00c cfeee00c cdb31e68 cff38000 cdb31e60 cff30000 cfcb9c00 cfcb9c00 0f000001 Call Trace: [<d0c25df0>] si3052Init+0x40/0x370 [Intel537] [<d0c26197>] si3052Res+0x37/0x90 [Intel537] [<d0c24de3>] afe_init+0x123/0x1a0 [Intel537] [<d0c231af>] modem_init+0x13f/0x1e0 [Intel537] [<d0c1faf5>] open+0xa5/0x180 [Intel537] [<d0c226fe>] softserial_open+0x12e/0x140 [Intel537] [<c025218b>] tty_open+0x177/0x244 [<c0153d7d>] chrdev_open+0xe0/0xf6 [<c0153c9d>] chrdev_open+0x0/0xf6 [<c0150932>] __dentry_open+0xc1/0x16d [<c0150a58>] nameidata_to_filp+0x24/0x33 [<c0150a9e>] do_filp_open+0x37/0x3e [<c015080b>] get_unused_fd_flags+0x49/0xa8 [<c0150ae1>] do_sys_open+0x3c/0x6d [<c0150b4b>] sys_open+0x1c/0x1e [<c0103baa>] syscall_call+0x7/0xb ======================= Code: fc b8 01 00 00 00 89 ec 5d c3 90 8d b4 26 00 00 00 00 55 89 e5 8b 55 0c 31 c0 83 fa 64 77 10 8b 45 08 8b 40 0c 8b 80 8c 00 00 00 <0f> b6 04 02 5d c3 90 55 89 e5 8b 55 08 8b 45 0c 8b 52 0c 8b 8a EIP: [<d0c278b9>] read8+0x19/0x20 [Intel537] SS:ESP 0068:cdb31e00 ---[ end trace 6963d572264669f2 ]--- === lsmod: === soekris:~/Modem# lsmod Module Size Used by Intel537 4316516 1 === -Chris On Mar 8, 2010, at 3:29 PM, Philippe Vouters wrote: > > Chris, > > See http://www.intel.com/design/modems/products/537ep.htm for the > specifications of your Intel's JA82537EP chip. Again it claims PCI 2.3 > compliant. > -- > sip:Vouters.Philippe@xxxxxxxxx > IM: philvout (Yahoo!) > http://vouters.dyndns.org/ > Philippe Vouters (Fontainebleau/France) > > > > -------- Message initial -------- > De: Chris Tracy <chris.t.tracy@xxxxxxxxx> > À: Philippe Vouters <Philippe.Vouters@xxxxxxxxxxx> > Cc: Bjorn Wielens <Uniacke1@xxxxxxxx>, Antonio Olivares > <olivares14031@xxxxxxxxx>, discuss@xxxxxxxxxxxxx, Marvin Stodolsky > <marvin.stodolsky@xxxxxxxxx> > Sujet: Re: Intel 537EP - driver loads but cannot use modem > Date: Mon, 8 Mar 2010 15:05:32 -0500 > > I just took a look at the actual board. There are only 2 chips: > > Si3018-KS (16-pin, http://www.datasheetsite.com/datasheet/SI3018-KS) > 0505CGCAAQ > > JA82537EP (64-pin) > 0529DGEA30 > INTEL '02 > > I guess the Si3018 and Si3052 chips are used in pairs in this application -- http://www.silabs.com/Marcom%20Documents/Resources/Si3052_PB.pdf. The Intel JA82537EP chip must contain all of the Si3052 functionality...? The Si3018 looks to just be the interface to the physical phone line while the Si3052 interfaces to the PCI bus. > > Cheers, > -Chris > > > On Mar 8, 2010, at 12:00 PM, Philippe Vouters wrote: > >> >> Chris, >> >> After objdump'ing intel-536-537/coredrv/537core.lib on my Linux, it >> ought to appear the code is all there ready to work correctly for your >> board. The only matter which I am concerned with is if your PCI bus >> fulfills the hardware limitation of the chip I mentioned in both my >> mails to you AND my Intel-Readme. If you look carefully at your board, >> it should contain such a component. >> -- >> sip:Vouters.Philippe@xxxxxxxxx >> IM: philvout (Yahoo!) >> http://vouters.dyndns.org/ >> Philippe Vouters (Fontainebleau/France) >> >> >> >> -------- Message initial -------- >> De: Chris Tracy <chris.t.tracy@xxxxxxxxx> >> À: Bjorn Wielens <Uniacke1@xxxxxxxx> >> Cc: Philippe Vouters <Philippe.Vouters@xxxxxxxxxxx>, Antonio Olivares >> <olivares14031@xxxxxxxxx>, discuss@xxxxxxxxxxxxx, Marvin Stodolsky >> <marvin.stodolsky@xxxxxxxxx> >> Sujet: Re: Intel 537EP - driver loads but cannot use modem >> Date: Mon, 8 Mar 2010 11:38:00 -0500 >> >> Bjorn, >> >> Good question -- the board in question that I am dealing with does indeed say "537EPG" on it, not "537EP". >> >> -Chris >> >> >> On Mar 7, 2010, at 3:30 PM, Bjorn Wielens wrote: >> >>> I wonder if this is the same issue I ran in to last year when I picked >>> up such a modem... >>> Also identified as a 537, but the board itself was from a Dell system, >>> and branded "537EPG". >>> >>> Chris, can you examine the back of your board? Does it say "537EP" or >>> "537EPG"? >>> >>> If there are distinguishing marks, it will at least be easier to >>> determine whether this is a hardware or a software limitation. >>> >>> Bjorn. >>> >>> >>> On 03/07/2010 04:11 PM, Philippe Vouters wrote: >>>> >>>> Antonio, >>>> >>>>> From intel-536-537/makefile_537, PCIDEV=8086:1080 well corresponds to a >>>> 537EP modem absolutely handled by the >>>> http://vouters.dyndns.org/Intel/intel-536EP-537EP_2010_01_06.tar.bz2 >>>> What brings an uncertainty between 536EP and 537SP is PCIDEV=8086:1040 >>>> >>>> Now Chris is facing a problem with his 537EP modem in modem_init (see >>>> kernel stack trace). >>>> >>>> There is not much I can do to help as the code involved resides in >>>> 537_core.lib which has always been Intel's proprietary code and shipped >>>> in binary format. This is all the problem with Winmodems. >>>> >>>> Regards, >>> >> >> >> >> > > > >
