PCMCIA USB card problems
I've got an 4 Port PCMCIA connector with a VIA chipset. The manual claims it's an OHCI/EHCI device, but FreeBSD identifies it as an VIA 83C572 UHCI device. After the identification there is a nonesaying errormessage and the system crashs. Compiling the kernel without UHCI I get the following output when I put the card into the slot: cardbus0: Expecting link target, got 0xff cardbus0: Resource not specified in CIS: id=14, size=100 cardbus0: Resource not specified in CIS: id=20, size=20 cardbus0: serial bus, USB at device 0.0 (no driver attached) cardbus0: Expecting link target, got 0xff cardbus0: Resource not specified in CIS: id=14, size=100 cardbus0: Resource not specified in CIS: id=20, size=20 cardbus0: serial bus, USB at device 0.1 (no driver attached) cardbus0: Expecting link target, got 0xff cardbus0: Resource not specified in CIS: id=10, size=100 cardbus0: Resource not specified in CIS: id=14, size=100 ehci1: VIA VT6202 USB 2.0 controller mem 0xc0208000-0xc02080ff,0xc0209000-0xc02090ff irq 11 at device 0.2 on cardbus0 ehci1: [GIANT-LOCKED] usb1: EHCI version 1.0 usb1: wrong number of companions (2 != 0) usb1: unrecoverable error, controller halted usb1: blocking intrs 0x10 usb1: run timeout ehci1: USB init failed err=13 device_attach: ehci1 attach returned 5 The system doesn't crash, but obviously it doesn't work either. Anyway that output looks more useful, maybe someone here can give me a hint, or is the chipset simply not supportet (the VIA 83C572 is in the supportet hardware list, but I don't know weather that realy is the chipset or just what FreeBSD thinks it is). signature.asc Description: OpenPGP digital signature
nforce2, onboard sound, digital out?
My brother has an Asus A7N8X Deluxe (nforce2) with onboard sound. I installed 6-stable on the system, but I cannot figure out how to activate the digital output. Has anyone ever managed to do this? signature.asc Description: OpenPGP digital signature
6-stable doesn't recognize detatchment of CD-Rom
I don't know into which list this belongs, so I'm posting it to mobile and stable. I've got a Thinkpad-R40, a first generation Centrino system. When I detach the CD-Rom drive nothing is shown on dmesg. The same when I reenter it. No dmesg output. When I try to mount a cd after detaching and reentering the drive the system freezes. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
how to reach a developer
Because of the devfs discussion I had a look into sys/fs/devfs/devfs_rules.c . Somewhere in the sourcecode he/she's asking a question and I would like to answer it. There just is no email address provided. The question is as follows: /* * XXX: Does it matter whether we do * * foreach(dk in ds) * foreach(de in dm) * apply(dk to de) * * as opposed to * * foreach(de in dm) * foreach(dk in ds) * apply(dk to de) * * The end result is obviously the same, but does the order * matter? */ Supposing that there is only read access to dk and only de is manipulated by the operation apply(dk to de); I have to say that the second example is a lot more efficient, because the first one causes lots of unnecessary cache to memory writebacks. Which would be gracefuly avoided by the second one. - Dominic ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
devfs doesn't set access rights
On my FreeBSD 6-stable (the last build is less then 24hours ago) my devfs doesn't apply permissions set in /etc/devfs.conf when I attach new devices. I have to call: /etc/rc.d/devfs restart manually for the settings to be applied. This is rather uncomfortable in some cases (especially with my PDA). I don't know weather this is the normal behaviour, anyway if it is, shouldn't devfs be restarted automatically as soon as external devices (usb, firewire, ir) are attached? ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: devfs doesn't set access rights
It's a simple and working solution and I think it simply should be made default for ALL attach events. To me it makes more sense than having 2 different systems for the same thing. Anyway, now I'm going to set up a devfs.rules (and learn a new different syntax for doing exactly the same thing). This looks like something that could have been in the unix haters handbook. Brooks Davis wrote: This is normal. devfs.conf is for boot only, you need devfs.rules for runtime. Unfortunatly, the documentation of this fact and the docuementation of devfs.rules sucks. -- Brooks Marwan Burelle wrote: One possibility is to call /etc/rc.d/devfs when a device is attached in /etc/devd.conf (take a look at devd(8) and devd.conf(5)) For exemple, I have something like this in my /etc/devd.conf : attach 100 { device-name umass[0-9]+ ; action /bin/sleep 3; /etc/rc.d/devfs restart ; }; attach : what kind of event 100 : as usual, for ordering rules device-name ... : the device concerned action ... : what to do, here I use sleep, because action is called when the attach event arrives, not when the device node is created. There maybe a better way, but it works© ;) ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: devfs doesn't set access rights
Well, here are my conclusions. 1) The manpage for devfs.rules should mention that you have to set devfs_system_ruleset=yourruleset in /etc/rc.conf . I didn't see it anywhere. 2) I have transferred all the rules in /etc/devfs.conf to /etc/devfs.rules, because I don't like to have 2 solutions for one task and because of the race condition mentioned in this list. 3) There is one kind of thing I cannot do in /etc/devfs.rules, creating links. So I am still doing this in /etc/devfs.conf (links don't cause race conditions anyway, I hope). One would think that link ttyU0 pilot would simply be translated to ln -s /dev/ttyU0 /dev/pilot which would simply create the link and it would work as soon as a ttyU0 device (my PDA) is there. But instead it checks weather the device exists (of course it doesn't since I didn't press the hotsync button during boot) and omits the creation of the link. So either the behaviour of /etc/rc.d/devfs has to be changed or it has to be dumped entirely and /sbin/devfs has to implement a way to create links (the solution I would prefer). - Dominic ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: devfs doesn't set access rights
Here is a small patch for /etc/rc.d/devfs which makes it behave the way I suggested in my last mail. --- devfs.old Tue Dec 13 20:58:52 2005 +++ devfs Tue Dec 13 21:00:57 2005 @@ -43,8 +43,8 @@ cd /dev while read action device parameter; do case ${action} in - l*) if [ -c ${device} -a ! -e ${parameter} ]; then - ln -fs ${device} ${parameter} + l*) if [ ! -e ${parameter} ]; then + ln -s ${device} ${parameter} fi ;; o*) if [ -c ${device} ]; then ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: devfs doesn't set access rights
I agree that a more flexible approach is generally better, but since it's not necessary I don't want to create an extra entry in my /etc/usbd.conf (one more file to worry about during mergemaster). Thus I think the best solution would be to dump devfs.conf and /etc/rc.d/devfs entirely (or leave it optional for backwards compatibility) and give devfs the ability to create links. I have a dislike for inconsistent solutions and I think most people here share this dislike. It's just what happens when software grows. Sooner or later you get to the point where you should throw away 80% of what's there and turn the rest into something that does what needs to be done. Johny Mattsson wrote: Note that while for you it might be sufficient to have a hard coded link from /dev/pilot to /dev/ttyU0, that is not necessarily the case if you have multiple USB serial gadgets. The way I handled it is by using an entry in /etc/usbd.conf: # # Fixup permissions for Palm Handheld ucom devices # (set here and not in devfs.rules since we only want to apply # the change to Palm com ports, not all USB com ports) device PalmHandheld devname ucom[0-9]+ vendor 0x0830 attach /bin/chmod 0666 /dev/${DEVNAME} ; /bin/ln -s /dev/${DEVNAME} /dev/pilot detach rm -f /dev/pilot Having said that, it would be great to be able to create links from devfs.rules - I'd certainly use it for my /dev/cdrom entry! Cheers, /Johny ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
buildworld fails
Gcc segfaults for me during a buildworld. In my make.conf: CPUTYPE?= pentium-m CFLAGS= -O2 -pipe My last successful build was: FreeBSD 6.0-STABLE (TPR40-6) #0: Sat Nov 26 23:13:30 CET 2005 # make update was run right before building. Here's a tail of the buildworld log. building profiled panel library ranlib libpanel_p.a cat /usr/src/lib/libpanel/../../contrib/ncurses/man/panel.3x panel.3 gzip -cn panel.3 panel.3.gz === lib/libpcap (all) cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c grammar.c -o grammar.po cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c /usr/src/lib/libpcap/../../contrib/libpcap/pcap-bpf.c -o pcap-bpf.po cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c /usr/src/lib/libpcap/../../contrib/libpcap/pcap.c -o pcap.po cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c /usr/src/lib/libpcap/../../contrib/libpcap/inet.c -o inet.po cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c /usr/src/lib/libpcap/../../contrib/libpcap/fad-getad.c -o fad-getad.po cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c /usr/src/lib/libpcap/../../contrib/libpcap/gencode.c -o gencode.po cc -pg -O2 -pipe -march=pentium-m -DHAVE_CONFIG_H -Dyylval=pcapyylval -I/usr/src/lib/libpcap -I. -D_U_=__attribute__((unused)) -DHAVE_SNPRINTF -DHAVE_VSNPRINTF -DINET6 -I/usr/src/lib/libpcap/../../contrib/libpcap -c /usr/src/lib/libpcap/../../contrib/libpcap/optimize.c -o optimize.po /usr/src/lib/libpcap/../../contrib/libpcap/optimize.c: In function `convert_code_r': /usr/src/lib/libpcap/../../contrib/libpcap/optimize.c:2213: internal compiler error: Segmentation fault: 11 Please submit a full bug report, with preprocessed source if appropriate. See URL:http://gcc.gnu.org/bugs.html for instructions. *** Error code 1 Stop in /usr/src/lib/libpcap. *** Error code 1 Stop in /usr/src/lib. *** Error code 1 Stop in /usr/src. *** Error code 1 Stop in /usr/src. *** Error code 1 Stop in /usr/src. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: buildworld fails
The error also appeared without optimizations. So far compiling with optimizations has never caused any trouble for me. Now it is working without problems. I suspect that it was a cooling problem of my RAM. Adi Pircalabu wrote: On Tue, 06 Dec 2005 15:48:30 +0100 [LoN]Kamikaze [EMAIL PROTECTED] wrote: Gcc segfaults for me during a buildworld. In my make.conf: CPUTYPE?= pentium-m CFLAGS= -O2 -pipe Hi, you probably know it's not recommended to use optimization flags when building the base system. Please comment CFLAGS in /etc/make.conf and retry with the same source tree. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]