Re: ThinkPad R31 - Lucent windmodem woes....
Date: Mon, 25 Nov 2002 10:23:23 -0800 From: Wes Peters [EMAIL PROTECTED] Sender: [EMAIL PROTECTED] P.A.Osborne wrote: On Mon, Nov 25, 2002 at 07:54:32AM -0800, Wes Peters wrote: cual0 (or sio0) is almost certainly NOT your winmodem. I suspect it is an onboard serial or IR port, as is sio1. More reading of the ltmdm documentation might reveal how to discover what sio device node is associated with the winmodem when the module attaches. You'll then need to use the /dev/MAKEDEV script to create /dev entries for that sio node if you don't already have them. Thats the problem exactly! If I *knew* what device it really mapped to then I would be laughing, finding that out seems to be nigh on impossible, hence my mail asking if anyone had been round these hoops and succeeded. The last time I tried ltmdm (quite some time ago) it seemed to take the first non-allocated sio, which would be sio2 in your case. It's at least a good starting point. Or you could do like the rest of us and buy a pccard modem that doesn't suck. ;^) That is something that I don't want to do until I am certain that the winmodem is not an option. If you do get it working you're not likely to be impressed by the through- put or the load imposed on your system. I certainly wasn't. I believe IBM is now using the ICH3 chipset which includes the Agre (nee Lucent) AC'97 WinModem. This is a different beast from the older Lucent WinModems and, unless something has recently changed, it is not supported under FreeBSD with the ltmdm port. FWIW, under Windows, the new modem is vastly superior to my older internal modem on my 600E. I get better data rates and more reliable connections, but it does load the CPU quite a bit. Still, as little as I use a modem, it would be fine if there was an available driver. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: [EMAIL PROTECTED] Phone: +1 510 486-8634 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: running many programs on multiprocessor system
Stephen Montgomery-Smith wrote: [ ... ] In the mean time, can anyone tell me or point to me documentation telling me any kind of rules that assembler functions called from C programs have to obey? (And maybe vice versa.) Whether a given chunk of code was orignally written in assembler or in C doesn't matter, since C programs usually get translated into assembly by the compiler. Regardless of the origin, the code needs to follow the calling conventions of the platform, as well as the interfaces defined by the C header files; see a KR, or man elf and the references from there: Hewlett Packard, Elf-64 Object File Format. Santa Cruz Operation, System V Application Binary Interface. Unix System Laboratories, Object Files, Executable and Linking Format (ELF). In practice, compiling a small program in C and taking a look at the corresponding assembly via cc -S will probably be fairly helpful as well. -Chuck To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: IPsec/gif VPN tunnel packets on wrong NIC in ipfw?
On Mon, Nov 25, 2002 at 05:46:47PM +0100, Eric Masson wrote: In my case, the lan joined by the vpn use rfc1918 adresses, and if I want the vpn traffic to flow correctly, I must invalidate incoming rfc1918 address checking on the external firewall interface. I don't think it increases security ;) So Is there any fix floating around or is this definitely the right behaviour ? I see /usr/src/sys/netinet/ip_input.c changed last night when I ran cvsup: * @(#)ip_input.c 8.2 (Berkeley) 1/4/94 * $FreeBSD: src/sys/netinet/ip_input.c,v 1.130.2.44 2002/11/25 05:23:00 silby Exp $ cvs log says: revision 1.130.2.44 date: 2002/11/25 05:23:00; author: silby; state: Exp; lines: +20 -2 MFC rev 1.217 Add a sysctl to control the generation of source quench packets, and set it to 0 by default. Partially obtained from:NetBSD Suggested by: David Gilbert revision 1.130.2.43 date: 2002/11/21 01:27:31; author: luigi; state: Exp; lines: +1 -0 MFC: obey to fw_one_pass in bridge and layer 2 firewalling (the latter only affects ipfw2 users). Move fw_one_pass from ip_fw[2].c to ip_input.c to avoid depending on IPFIREWALL. I haven't tried it yet but the above sounds like its addressing the problem I have had with formerly tunneled packets being run thru IPFW after emerging from the tunnel. -- David Kelly N4HHE, [EMAIL PROTECTED] = The human mind ordinarily operates at only ten percent of its capacity -- the rest is overhead for the operating system. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
MCLGET Page Fault Trap
Hi!
We got this strange page fault after 18 days of uptime, where this box
has done the same task all over these days. It might be broken RAM,
but I thought I might share this with you.
Alex
root@cherub ...src/sys/CHERUB # gdb -k
GNU gdb 4.18 (FreeBSD)
Copyright 1998 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you
are
welcome to change it and/or distribute copies of it under certain
conditions.
Type show copying to see the conditions.
There is absolutely no warranty for GDB. Type show warranty for
details.
This GDB was configured as i386-unknown-freebsd.
(kgdb) exec-file /usr/crash/kernel.9
(kgdb) symbol-file kernel.debug
Reading symbols from kernel.debug...Deprecated bfd_read called at
/usr/src/gnu/usr.bin/binutils/gdb/../../../../contrib/gdb/gdb/dbxread.c
line 2627 in elfstab_build_psymtabs
Deprecated bfd_read called at
/usr/src/gnu/usr.bin/binutils/gdb/../../../../contrib/gdb/gdb/dbxread.c
line 933 in fill_symbuf
done.
(kgdb) core-file /usr/crash/vmcore.9
IdlePTD at phsyical address 0x002ce000
initial pcb at physical address 0x0024c4c0
panicstr: page fault
panic messages:
---
Fatal trap 12: page fault while in kernel mode
fault virtual address = 0xc0902743
fault code = supervisor write, page not present
instruction pointer = 0x8:0xc0166a3b
stack pointer = 0x10:0xc022d5e4
frame pointer = 0x10:0xc022d5f8
code segment= base 0x0, limit 0xf, type 0x1b
= DPL 0, pres 1, def32 1, gran 1
processor eflags= interrupt enabled, resume, IOPL = 0
current process = Idle
interrupt mask = net
trap number = 12
panic: page fault
syncing disks...
Fatal trap 12: page fault while in kernel mode
fault virtual address = 0x30
fault code = supervisor read, page not present
instruction pointer = 0x8:0xc01b2a40
stack pointer = 0x10:0xc022d408
frame pointer = 0x10:0xc022d410
code segment= base 0x0, limit 0xf, type 0x1b
= DPL 0, pres 1, def32 1, gran 1
processor eflags= interrupt enabled, resume, IOPL = 0
current process = Idle
interrupt mask = net bio cam
trap number = 12
panic: page fault
Uptime: 18d0h40m10s
dumping to dev #ad/0x20021, offset 262304
dump ata2: resetting devices .. done
127 126 125 124 123 122 121 120 119 118 117 116 115 114 113 112 111 110
109 108 107 106 105 104 103 102 101 100 99 98 97 96 95 94 93 92 91 90 89
88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65
64 63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41
40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17
16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
---
#0 dumpsys () at /usr/src/sys/kern/kern_shutdown.c:487
487 if (dumping++) {
(kgdb) up
#1 0xc014aed7 in boot (howto=260) at
/usr/src/sys/kern/kern_shutdown.c:316
316 dumpsys();
(kgdb) up
#2 0xc014b2fc in poweroff_wait (junk=0xc0224dcc, howto=-1071494929)
at /usr/src/sys/kern/kern_shutdown.c:595
595 boot(bootopt);
(kgdb) up
#3 0xc01f44b6 in trap_fatal (frame=0xc022d3c8, eva=48)
at /usr/src/sys/i386/i386/trap.c:974
974 panic(%s, trap_msg[type]);
(kgdb) up
#4 0xc01f4189 in trap_pfault (frame=0xc022d3c8, usermode=0, eva=48)
at /usr/src/sys/i386/i386/trap.c:867
867 trap_fatal(frame, eva);
(kgdb) up
#5 0xc01f3d73 in trap (frame={tf_fs = 16, tf_es = 16, tf_ds = 16,
tf_edi = -1071244832, tf_esi = 0, tf_ebp = -1071459312,
tf_isp = -1071459340, tf_ebx = -1071391268, tf_edx = 6865984,
tf_ecx = 2, tf_eax = 0, tf_trapno = 12, tf_err = 0,
tf_eip = -1071961536, tf_cs = 8, tf_eflags = 66050, tf_esp = 0,
tf_ss = 0}) at /usr/src/sys/i386/i386/trap.c:466
466 (void) trap_pfault(frame, FALSE, eva);
(kgdb) up
#6 0xc01b2a40 in acquire_lock (lk=0xc023dddc)
at /usr/src/sys/ufs/ffs/ffs_softdep.c:266
266 lk-lkt_held = CURPROC-p_pid;
(kgdb) up
#7 0xc01b7062 in softdep_fsync_mountdev (vp=0xc86a0b40)
at /usr/src/sys/ufs/ffs/ffs_softdep.c:4024
4024ACQUIRE_LOCK(lk);
(kgdb) up
#8 0xc01bb292 in ffs_fsync (ap=0xc022d484)
at /usr/src/sys/ufs/ffs/ffs_vnops.c:134
134 softdep_fsync_mountdev(vp);
(kgdb) up
#9 0xc01b9f23 in ffs_sync (mp=0xc0c01200, waitfor=2, cred=0xc0741600,
p=0xc02619e0) at vnode_if.h:558
558 rc = VCALL(vp, VOFFSET(vop_fsync), a);
(kgdb) up
#10 0xc017adff in sync (p=0xc02619e0, uap=0x0)
at /usr/src/sys/kern/vfs_syscalls.c:576
576 VFS_SYNC(mp, MNT_NOWAIT,
(kgdb) up
#11 0xc014ac72 in boot (howto=256) at
/usr/src/sys/kern/kern_shutdown.c:235
235 sync(proc0, NULL);
(kgdb) up
#12 0xc014b2fc in poweroff_wait (junk=0xc0224dcc, howto=-1071494929)
at
Re: ThinkPad R31 - Lucent windmodem woes....
On Mon, Nov 25, 2002 at 10:23:23AM -0800, Wes Peters wrote: If I *knew* what device it really mapped to then I would be laughing, finding that out seems to be nigh on impossible, hence my mail asking if anyone had been round these hoops and succeeded. The last time I tried ltmdm (quite some time ago) it seemed to take the first non-allocated sio, which would be sio2 in your case. It's at least a good starting point. Ok I will give that a go in a bit, thanks. That is something that I don't want to do until I am certain that the winmodem is not an option. If you do get it working you're not likely to be impressed by the through- put or the load imposed on your system. I certainly wasn't. Thats kind of what I am expecting mainly die to the bad press (with good reason) in the past. However the modem runs find under WinXP with no noticeable major increase in load on the CPU - hence I dont't want to buy a real modem unless I have to. Paul To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: ThinkPad R31 - Lucent windmodem woes....
On Mon, Nov 25, 2002 at 11:04:20AM -0800, Kevin Oberman wrote: I believe IBM is now using the ICH3 chipset which includes the Agre (nee Lucent) AC'97 WinModem. This is a different beast from the older Lucent WinModems and, unless something has recently changed, it is not supported under FreeBSD with the ltmdm port. Yes thats the beast, in which case: rats! FWIW, under Windows, the new modem is vastly superior to my older internal modem on my 600E. I get better data rates and more reliable connections, but it does load the CPU quite a bit. Still, as little as I use a modem, it would be fine if there was an available driver. Thats my thought exactly - at work the laptop is on the net any way - it would just be nice to be able to dial up on it using FreeBSD than currently being forced to use Windows ( as I am at the moment to ssh to work). Ho hum Paul To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: ThinkPad R31 - Lucent windmodem woes....
Sender: [EMAIL PROTECTED] Date: Mon, 25 Nov 2002 13:21:15 -0800 From: Wes Peters [EMAIL PROTECTED] I believe IBM is now using the ICH3 chipset which includes the Agre (nee Lucent) AC'97 WinModem. This is a different beast from the older Lucent WinModems and, unless something has recently changed, it is not supported under FreeBSD with the ltmdm port. FWIW, under Windows, the new modem is vastly superior to my older internal modem on my 600E. I get better data rates and more reliable connections, but it does load the CPU quite a bit. Still, as little as I use a modem, it would be fine if there was an available driver. Thanks for the update. It might be worth checking to see if the new chipset is supported under Linux. Porting the changes for the chipset to FreeBSD may be easier than porting the entire driver. It is supported under Linux with a binary-only driver supplied by Agre. Of course, this driver will not run under FreeBSD, so it will be necessary to write shims similar to those in the ltmdm port. I believe Watanabe-san was either working on these or he was posting information from others working on them on his web page. Because this modem is included in the very popular ICH sets, I am hopeful that it won't take too much longer, but I don't have a clue about Linux kernel modules, so I am not in a position to contribute much to this. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: [EMAIL PROTECTED] Phone: +1 510 486-8634 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: ThinkPad R31 - Lucent windmodem woes....
To rudely follow up my own posts, I have been unable to reach Watanabe-san's web page. At least the English language page seems to have moved or gone away. Also the message I had seen that I thought implied he was working on it really was relating to an issue of a different flavor of Lucent WinModem. At this time I am unsure if anyone is working on this seriously. Larry Rosenman was trying to track something down on this, but I have not seen any recent messages. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: [EMAIL PROTECTED] Phone: +1 510 486-8634 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: ThinkPad R31 - Lucent windmodem woes....
On Tue, 2002-11-26 at 02:24, Wes Peters wrote: luthien# ppp Working in interactive mode Using interface: tun0 ppp ON luthien dial isp Warning: deflink: /dev/cual0: Bad file descriptor ppp ON luthien However the file is there: luthien# ls -la /dev/cual0 crw-rw 1 uucp dialer 228, 128 Nov 25 10:29 /dev/cual0 cual0 (or sio0) is almost certainly NOT your winmodem. I suspect it is an onboard serial or IR port, as is sio1. More reading of the ltmdm documentation might reveal how to discover what sio device node is associated with the winmodem when the module attaches. You'll then need to use the /dev/MAKEDEV script to create /dev entries for that sio node if you don't already have them. Actually, cual0 IS the ltmdm device node name.. (Along with ttyl0) Of course it doesn't work if the driver doesn't attach. If Paul runs dmesg | grep ltmdm he should see a line like - ltmdm0: Lucent Winmodem port 0xe800-0xe8ff,0xecb8-0xecbf mem 0xf8ffec00-0xf8ffecff irq 11 at device 8.0 on pci8 ltmdm0: type Virtual 16550A If not it wasn't detected. In that case it either isn't a lucent winmodem, or, if it is the PCI ID isn't known. BTW while on general principle I don't like win modems etc, these ones DO seem to work quite well. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 9A8C 569F 685A D928 5140 AE4B 319B 41F4 5D17 FDD5 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Loading if_xl from the loader
I installed FreeBSD on a work colleges computer and found something odd.. I have the following in /boot/loader.conf # -- sysinstall generated deltas -- # userconfig_script_load=YES agp_load=YES snd_pdm_load=YES snd_ich_load=YES if_xl_load=YES miibus_load=YES hw.ata.atapi_dma=1 But when I check dmesg.. CPU: Pentium III/Pentium III Xeon/Celeron (1002.28-MHz 686-class CPU) Origin = GenuineIntel Id = 0x68a Stepping = 10 Features=0x383f9ffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR,SSE real memory = 133103616 (129984K bytes) avail memory = 126459904 (123496K bytes) Preloaded elf kernel kernel at 0xc030b000. Preloaded elf module miibus.ko at 0xc030b0a8. Preloaded elf module snd_ich.ko at 0xc030b148. Preloaded elf module snd_pcm.ko at 0xc030b1e8. Preloaded elf module agp.ko at 0xc030b288. Pentium Pro MTRR support enabled ??? I saw the loader load the module too :) If I do kldload if_xl after it's booted the xl device is probed fine - xl0: 3Com 3c905-TX Fast Etherlink XL port 0xc800-0xc83f irq 10 at device 4.0 on pci1 xl0: Ethernet address: 00:60:08:ae:15:f7 miibus0: MII bus on xl0 nsphy0: DP83840 10/100 media interface on miibus0 nsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto This is some Packard Bell prebuilt thing if it matters. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 9A8C 569F 685A D928 5140 AE4B 319B 41F4 5D17 FDD5 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Loading if_xl from the loader
I installed FreeBSD on a work colleges computer and found something odd.. I have the following in /boot/loader.conf # -- sysinstall generated deltas -- # userconfig_script_load=YES agp_load=YES snd_pdm_load=YES snd_ich_load=YES if_xl_load=YES miibus_load=YES hw.ata.atapi_dma=1 But when I check dmesg.. CPU: Pentium III/Pentium III Xeon/Celeron (1002.28-MHz 686-class CPU) Origin = GenuineIntel Id = 0x68a Stepping = 10 Features=0x383f9ffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV ,PAT,PSE36,MMX,FXSR,SSE real memory = 133103616 (129984K bytes) avail memory = 126459904 (123496K bytes) Preloaded elf kernel kernel at 0xc030b000. Preloaded elf module miibus.ko at 0xc030b0a8. Preloaded elf module snd_ich.ko at 0xc030b148. Preloaded elf module snd_pcm.ko at 0xc030b1e8. Preloaded elf module agp.ko at 0xc030b288. Pentium Pro MTRR support enabled ??? I saw the loader load the module too :) speculation Could if_xl depend on miibus and thus not be loaded because miibus hasn't been loaded yet? Try putting `miibus_load=YES' before `if_xl_load=YES' /speculation If I do kldload if_xl after it's booted the xl device is probed fine - xl0: 3Com 3c905-TX Fast Etherlink XL port 0xc800-0xc83f irq 10 at device 4. 0 on pci1 xl0: Ethernet address: 00:60:08:ae:15:f7 miibus0: MII bus on xl0 nsphy0: DP83840 10/100 media interface on miibus0 nsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto This is some Packard Bell prebuilt thing if it matters. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 9A8C 569F 685A D928 5140 AE4B 319B 41F4 5D17 FDD5 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Loading if_xl from the loader
On Tue, 2002-11-26 at 10:07, Mike Hogsett wrote: Could if_xl depend on miibus and thus not be loaded because miibus hasn't been loaded yet? Try putting `miibus_load=YES' before `if_xl_load=YES' I just tried that, no change :( FWIW the loader does know that if_xl needs miibus and loads that first. Hmm, I just rebooted it again and found that if I interrupt the loader after it's loaded all the modules (including if_xl :) then do 'lsmod' if_xl doesn't show up! -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 9A8C 569F 685A D928 5140 AE4B 319B 41F4 5D17 FDD5 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Loading if_xl from the loader
On Tue, 2002-11-26 at 10:14, Daniel O'Connor wrote: Hmm, I just rebooted it again and found that if I interrupt the loader after it's loaded all the modules (including if_xl :) then do 'lsmod' if_xl doesn't show up! Ooh, even better I find that if I load if_xl first (either by making it the first *_load=YES command, or manually) it is listed and detects fine.. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 9A8C 569F 685A D928 5140 AE4B 319B 41F4 5D17 FDD5 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Loading if_xl from the loader
On Tue, 26 Nov 2002, Daniel O'Connor wrote: Date: Tue, 26 Nov 2002 10:14:23 +1030 From: Daniel O'Connor [EMAIL PROTECTED] To: Mike Hogsett [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: Loading if_xl from the loader On Tue, 2002-11-26 at 10:07, Mike Hogsett wrote: Could if_xl depend on miibus and thus not be loaded because miibus hasn't been loaded yet? Try putting `miibus_load=YES' before `if_xl_load=YES' I just tried that, no change :( FWIW the loader does know that if_xl needs miibus and loads that first. Hmm, I just rebooted it again and found that if I interrupt the loader after it's loaded all the modules (including if_xl :) then do 'lsmod' if_xl doesn't show up! Daniel, What happens if you take miibus_load=YES out of loader.conf? Is it still loaded as a dependancy by the loader when it trys to load if_xl? Just an idea.. Jarrod Sayers Information Strategy and Technology Services University of South Australia Phone: +61 8 8302 4809 There are only 10 types of people in the world: Those who understand binary, and those who don't To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
