xenocara build failure
This is probably something stupid I'm doing, but I can't see it right this second. Trying to build xenocara from sources pulled from anon...@anoncvs3.usa.openbsd.org:/cvs as of about 60 minutes before sending this email message gives me cc -O2 -pipe -I/usr/xenocara/lib/freetype/include -I/usr/xenocara/lib/freetype/builds/unix -I/usr/xenocara/lib/freetype/src/lzw -DFT2_BUILD_LIBRARY -c /usr/xenocara/lib/freetype/src/type1/type1.c -o type1.o In file included from /usr/xenocara/lib/freetype/src/type1/type1.c:23: /usr/xenocara/lib/freetype/src/type1/t1load.c: In function 'parse_private': /usr/xenocara/lib/freetype/src/type1/t1load.c:1037: error: 'struct T1_Loader_' has no member named 'keywords_encountered' /usr/xenocara/lib/freetype/src/type1/t1load.c:1037: error: 'T1_PRIVATE' undeclared (first use in this function) /usr/xenocara/lib/freetype/src/type1/t1load.c:1037: error: (Each undeclared identifier is reported only once /usr/xenocara/lib/freetype/src/type1/t1load.c:1037: error: for each function it appears in.) In file included from /usr/xenocara/lib/freetype/src/type1/type1.c:23: /usr/xenocara/lib/freetype/src/type1/t1load.c: In function 'parse_dict': /usr/xenocara/lib/freetype/src/type1/t1load.c:1871: error: 'struct T1_Loader_' has no member named 'keywords_encountered' /usr/xenocara/lib/freetype/src/type1/t1load.c:1871: error: 'T1_PRIVATE' undeclared (first use in this function) /usr/xenocara/lib/freetype/src/type1/t1load.c:1872: error: 'struct T1_Loader_' has no member named 'keywords_encountered' /usr/xenocara/lib/freetype/src/type1/t1load.c:1873: error: 'T1_FONTDIR_AFTER_PRIVATE' undeclared (first use in this function) /usr/xenocara/lib/freetype/src/type1/t1load.c:1978: error: 'struct T1_Loader_' has no member named 'keywords_encountered' /usr/xenocara/lib/freetype/src/type1/t1load.c:1990: error: 'struct T1_Loader_' has no member named 'keywords_encountered' /usr/xenocara/lib/freetype/src/type1/t1load.c: In function 't1_init_loader': /usr/xenocara/lib/freetype/src/type1/t1load.c:2047: error: 'struct T1_Loader_' has no member named 'keywords_encountered' *** Error 1 in lib/freetype (:37 'type1.o': @cc -O2 -pipe -I/usr/xenocara/lib/freetype/include -I/usr/xenocara/lib/freetype/...) *** Error 1 in lib/freetype (Makefile:36 'build') *** Error 1 in lib (:48 'build') *** Error 1 in . (:48 'realbuild') *** Error 1 in /usr/xenocara (Makefile:35 'build') Any hints as to what I'm doing wrong?
Re: xenocara build failure
On May 13, 2013, at 9:47 PM, Scott McEachern wrote: >> >> *** Error 1 in lib/freetype (:37 'type1.o': @cc -O2 -pipe >> -I/usr/xenocara/lib/freetype/include -I/usr/xenocara/lib/freetype/...) >> *** Error 1 in lib/freetype (Makefile:36 'build') >> *** Error 1 in lib (:48 'build') >> *** Error 1 in . (:48 'realbuild') >> *** Error 1 in /usr/xenocara (Makefile:35 'build') >> >> Any hints as to what I'm doing wrong? >> > > I've seen this before. After you rebuild your system, reboot. (Yes, in > addition to after rebooting into the new kernel.) Bet your problem will be > solved. Tried that... still fails in the same place. I think I'll try grabbing source from a different CVS server and see if that makes a difference... grasping at straws.
Re: xenocara build failure
On May 13, 2013, at 11:54 PM, Maurice Janssen wrote: >> *** Error 1 in /usr/xenocara (Makefile:35 'build') >> >> Any hints as to what I'm doing wrong? > > Are you using make in parallel (-j) mode? If so, please try without -j I'm doing a simple make build (after a make bootstrap && make obj) Looking at the sources and the build output only leads to questions. 1) makefile issue? freetype fails to build but the makefile continues to the install phase where it finally fails trying to install the lib that didn't build. 2) t1load.c uses a structure element named keywords_encountered. The code that uses this element does not appear to be conditional. 3) No-where in the source can I find where anything with the name keywords_encountered is defined. There is a comment in Changelog.23: that says the field keywords_encountered was added to t1load.h back in 2006!!! Yet that field doesn't exist in my t1load.h and a cvs diff says my file is current. Confusion.
Re: xenocara build failure
> What is the revision of the file? Any sticky tags/dates? > > rev should be 1.1.1.2 You are on to something... cvs up -PAd should remove all tags, no? It didn't. When I moved t1load.h out of the way and re-updated I got a different version. Grumble... I think it's time to refresh all my sources from scratch now that I can't be sure that everything is up to date. Thanks, Marc
Re: xenocara build failure
On May 14, 2013, at 1:56 PM, Ted Unangst wrote: >> You are on to something... cvs up -PAd should remove all tags, no? It >> didn't. > > Yes, it should. No, it doesn't. There's a bug somewhere in cvs. *nod* Trashing the source and fetching from scratch showed a few things stuck at older tags. The nsd sources in src and several things in xenocara including freetype and Xfont. In any case my compile issue is resolved. Thanks to those who provided help.
Re: xenocara build failure
On May 14, 2013, at 2:07 PM, patrick keshishian wrote: > > and none of those files were locally modified? Do you have the output > of "cvs up -PAd" in regard to those specific files? There was no output. I have exactly one modified file in my xenocara tree. It's not in the lib subdir. I have a few more modifications in my src tree, but none in nsd which is the code that wasn't up to date. Otto Moerbeek wrote: > I believe others have seen problems wher the server was running > opencvs. No idea if this case is related to that. I tried updating from two different servers. anon...@anoncvs3.usa.openbsd.org:/cvs anon...@anoncvs.ca.openbsd.org:/cvs
Re: securing OpenBSD wireless network
Tor Houghton writes: > On Mon, Nov 19, 2007 at 07:59:17AM -0800, David Newman wrote: > > > > Well, if you want to prevent someone from accidentally connecting to your > > > network, yes. > > > > WEP keys can be captured is less than one minute: > > This fact is immaterial in context of my statement. Very true. The only time I consider turning on WEP is when I notice a neighbor is connecting to my net more often than not. Yes, by accident. Typically because their AP went down and needed to be reset and they hadn't noticed. Maybe one of these days I'll turn it on... // marc
Re: License Violation - ksh
"Pedro de Oliveira" writes: > Hello, > Someone on IRC just posted this link http://www.delilinux.de/oksh/ , seems > like someone ported OpenBSD ksh to Linux and licensed it under GPLv3. Isn't > this a license violation? See /usr/src/bin/ksh/LEGAL. People can do just about whatever they want with public domain code, including slapping a license on it. Anyone foolish (or lazy) enough to derive code from the licenses derivative deserves the license they get. // marc
Re: License Violation - ksh
Miod Vallat writes: > But some parts of OpenBSD's ksh are BSD-licensed files, which did not > come from pdksh initially. That's not what /usr/src/bin/ksh/LEGAL states, but I didn't look further. OK, looking I see that alloc.c and mknod.c have copyrights. > Relicensing these files under GPLv3 is only possible if their authors > gave permission. Very true. Were they relicensed, or tossed and re-written from scratch? // marc
Re: no 4.2-stable package updates??
"Unix Fan" writes: > This really does suck... While we as users appreciate developers hard work, A majority rely on -STABLE for updated and secure 3rd party software.. Really? You have statistics? I'd be curious to see how many run stable vs. old releases vs. current. > Why even have a -CURRENT ports tree?... -STABLE should be the tree > maintained for the 6 months between releases.. You do realize that ports maintenance is a volunteer effort and that volunteers get burned out. I'm speaking from experience in that I more-or-less managed the ports tree for several releases way back when when it was a lot easier. No packages (packages were just being born about the time I stopped) and there were less than 1/4 the number of ports you have today. You are asking for volunteers to double their work effort. I don't see that happening. If you want specific things done I suggest you come up with funding. I did that a few times, too, in the form of "when I can run foo on my machine I'll send you a check for $$$". // marc
Re: Usage of underscore in system accounts
Ingo Schwarze writes: > > and the early versions (say, OpenBSD 3.0) > > Hey, Nick! Stop making people feel so old, > winter is depressive time anyway...;) 3.0 is not old! I have one copy of 2.3 still in its shrink wrap left on my shelf :-) // marc
Re: ipv6 hostname.gif
> but when i try to translate them into configure files, i run into a problem > with the hostname.gif file: Not surprising as the hostname.xxx files do NOT use ifconfig syntax. This seems to fool lots of people. Do not look at the ifconfig man page. Look at the hostname.if man page where it states the form for Regular IPv6 network setup: addr_family [alias] addr prefixlen options > but this one does not. > inet6 alias 2001:470:1f04:16f::2 2001:470:1f04:16f::1 prefixlen 128 <- > this line fails Because it doesn't follow the form. The 4th param should be the prefixlen, not the literal value "prefixlen". You've tried to put two different hostname.if commands into one bogus entry. You want this: inet6 alias 2001:470:1f04:16f::2 128 dest 2001:470:1f04:16f::1 (assuming I didn't get the addresses backwords) Example: my hostname.gif0 looks like this: tunnel 208.201.244.208 208.201.234.221 inet6 alias server-gif0 64 inet6 alias 2001:05a8:0:1::0123 128 dest 2001:05a8:0:1::0122 ! route add -inet6 default ::1 ! route change -inet6 default -ifp gif0 where server-gif0 is an address from my /64 in /etc/hosts. // marc
Re: fvwm in base and repository with security issues?
Marco Peereboom writes: > > -Nix Fan. (Fvwm, IMHO is ugly..) > > Fluxbox is fugly, fvwm is awesome. > > this discussion is very useful! cwm (with the changes I've made :-) is looking nicer and nicer! I hope to see these changes in the tree, soon. // marc
Re: delete deleted data
"Brad Tilley" writes: > performed from the OpenBSD 4.2 install CD. I'll send it to the one > 'ISO Certified' company that agreed to examine it. If they cannot You keep throwing around the 'ISO Certified' tag as if it had some special meaning. Certified to what standard? It makes a difference. If they are certified to the 9001 standard, for example, all it means is that they have written procedures and they follow them. That's all it means. ISO 9001 certification is actually pretty easy to get. The companies that fail to get it are trying to hard. They come up with procedures that sound great but are impossible to follow. That's not what certification means. If I have a software company and write up a procedure that says "all code will be developed on a laptop while sitting in a Starbucks" and actually follow that procedure, then I can be an "ISO Certified" company. As for disk destruction... I don't know nor pretend to know what can and can not be recovered. Take a look at https://www.dss.mil/portal/ShowBinary/BEA%20Repository/new_dss_internet/isp/odaa/documents/clear_n_san_matrix_06282007_rev_11122007.pdf The DSS (Defense Security Service, part of the DoD) calls what you have done "clearing" the disk. It does not "sanitize" the disk. To sanitize you need to either degauss or destroy the disk. // marc
Re: OT Re: OpenBSD and ISDN TA
Diana Eichert writes: > On Wed, 9 Jan 2008, Stuart Henderson wrote: > > > run H.323 and you can experience much of that same pain again > > and more besides :-) > > (now we digress) > > give me X.25 any day, instead of this new fangled ISDN technology. Yeah, X.25 with a triple-X pad (X.3/X.28/X.29). a Yellow book version, none of that fancy new red or blue book stuff. It scares me that I remember such stuff. // marc
Re: OSS v4.0 released under BSD license
Deanna Phillips writes: > But for Linux binary emulation? No way. If you want that, run > Linux. What kind of people run Linux binaries on OpenBSD, > anyway? Don't give me that "I need Flash", since I spent months > of my life working on Gnash for OpenBSD just so you wouldn't > have to use the Adobe Linux binary.. and more months working on > PJSIP so that you wouldn't have to use Skype. Uhhh, railing against Linux binary emulation is fine, but don't use gnash as your argument. Gnash is not usable. It may play the run-of-the-mill youtube video, but using it for just about anything else does no more than leave a dump file on disk. // marc
Re: OSS v4.0 released under BSD license
Deanna Phillips writes: > ; Do something about it > ; Use another OS > ; Complain > > Which are you doing? None of the above. I ignore flash. My comment was only to point out that gnash is not the best example to show why Linux emulation isn't needed. Oh, I ignore Linux emulation, too. Neither is needed in my tiny little world. // marc
Re: photo/ image viewing software
Chris writes: > I am after a software that would allow me to view photos from my > digital camera which I usually mount in /mnt/camera. I tried from the Depends upon the camera. If your camera looks like a USB mass storage device then mount /dev/sdwhatever and use the display command from ImageMagick to look at images. If your camera looks like some special device then your stuck using one of the programs that you listed to get the images out of the camera. // marc
Re: Multi-Threaded SSH/SCP made by university of Puttsburgh
> Actually, this is the part of the discussion that interests me. Is > threading > a doomed hope on OpenBSD, a model of utilizing multiple cores which the > developers have zero interest supporting? Has the work on libc_r and the > like > been abandoned completely? libc_r is gone, replaced with libpthread libpthread is changed to fix specific reported problems, most recently (early Feb) a bug fix in dup2 by [EMAIL PROTECTED] librthread seems to be stagnating, waiting upon kernel work that no one is either interested in or, if interested, has the time to do. I'm not even sure there is consensus on what kernel changes are needed. For now libpthread (userland threads) are the most useful for most threaded applications. There are still bugs. libpthread based applications get no benefit from running on multi-core/processor machines. // marc
Re: xenocara source
> For the sake of new folks it may be wise to put a .cvsignore in > our /usr/src tree to prevent unintended cosequences of using the (also > suggested) prune switch on cvs (-P). -P will only remove EMPTY directories that cvs knows about. Putting xenocara (or in my case zenocara and ports) in /usr/src is pretty much a no-op when it comes to cvs. A "cd /usr/src; cvs up -Pd" displays "? xenocara". I can live with that. > When following anoncvs.html, if a new person goes to update > their /usr/src tree, they would thwack their /usr/src/xenocara tree. If the directory isn't empty it isn't thwacked If the directory isn't known by CVS it isn't thwacked. // marc
Re: minimac on openbsd
sonjaya writes: > Also default minimac is only 1 ethernet how to add another ethernet > can support in minimac and openbsd. Yes. My web/mail server is a mac mini. The only problem was finding a usb ethernet that worked. I had to try three of them before I found one supported. This one works fine: axe0 at uhub0 port 6 configuration 1 interface 0 \ "Cisco-Linksys USB200M v2" rev 2.00/0.01 addr 2 axe0: AX88772, address 00:16:b6:ef:6e:83 ukphy0 at axe0 phy 16: Generic IEEE 802.3u media interface, rev. 1: \ OUI 0x000ec6, model 0x0001 sonjaya also writes: > any other device sugesstion? Take a look at the back of the mini. What openings in the case do you see? For a second wired ethernet your only choice is USB as OpenBSD doesn't support firewire. I don't know if the built in wireless works as I've never tried it. It is recognized by the kernel: ath0 at pci2 dev 0 function 0 "Atheros AR5424" rev 0x01: apic 2 int 17 (irq 11) ath0: AR5424 10.3 phy 6.1 rf 10.2, WOR5_ETSIC, address 00:17:f2:4f:3f:75 Warning: the mini won't boot without a console attached. I believe you can make a dummy plug to fool it into thinking there is a console. Also, you need this in your rc.local or someplace similar: # Magic tweak to put a Mac mini in "server mode" meaning it will # reboot after a power failure. Requires machdep.allowaperture=1 # (or better) # if [ -x /usr/X11R6/bin/pcitweak ]; then /usr/X11R6/bin/pcitweak -w 00:1f:0 -b 0xa4 0x00 if [ "$(/usr/X11R6/bin/pcitweak -r 0:1f:0 -b 0xa4)" = "0x00" ]; then echo -n ' server-mode' fi fi // marc
Re: minimac on openbsd
Matthew Szudzik writes: > > doesn't support firewire. I don't know if the built in wireless works > > as I've never tried it. It is recognized by the kernel: > > > > ath0 at pci2 dev 0 function 0 "Atheros AR5424" rev 0x01: apic 2 int 17 ( irq 11) > > Atheros AR5424 wireless devices are not supported in OpenBSD. See > http://marc.info/?l=openbsd-misc&m=118896011519883 It may not work, don't know... but it is at least recognized by the kernel. $ ifconfig ath0 ath0: flags=8822 mtu 1500 lladdr 00:17:f2:4f:3f:75 groups: wlan media: IEEE802.11 autoselect status: no network ieee80211: nwid "" // marc
Re: Disable IPv6 on OpenBSD 4.0 - forking discussion to icmp echo request blockage
> servers with services running we want public. Why should I allow > someone to ping my dns server? If I'm having problems resolving a host address that is supposed to be handled by your server one of the first things I'll do is see if I have general connectivity to your server. I'll ping it. If there is no answer I'll most likely assume transient net errors and put the problem off until later. So what, you say. Well, if there are real DNS problems you won't be notified. Maybe you don't care. > If you need to see if the server is up telnet to port 53, a traceroute > will die at the hop above the firewall, I know which ip that is. I don't > care/need others to do so. If I can't ping I'll assume I can't telnet. A traceroute will confirn "net connectivity" issues. Eventually, assuming I need your DNS server to work correctly, I'll attempt to get in touch. >From my perspective the only thing your blocking ICMP has done is delay third party notification of DNS issues. To me (and I'll be the first to admit that this is nothing but opinion and I won't pretend that my opinion is any better than yours) I see more harm than good in blocking icmp. I like it when other people tell me I've screwed something up because I can find it and fix it faster. As for the person who wants to dispable ipv6... I think henning@ had the best solution: use pf. A rule such as block ipv6 drop quick all at the top of your ruleset should do the trick. // marc
Re: How can I view rule numbers under OpenBSD 4.0?
"C. L. Martinez" writes: > Oops .. I would say under OpenBSD 4.0 Not -w, but -v -v. I'm guessing the font you use makes -vv look like -w. // marc
Re: Block device required
"Greg Thomas" writes: > > > mount -t ext2fs /dev/rwd1c /mnt > > > > mount -t ext2fs /dev/wd1i /mnt > > How do you know it's i? FAQ 14.1 says: Devices without a disklabel: If a device does not currently have an OpenBSD disklabel on it but has another file system (for example, a disk with a pre-existing FAT32 file system), the OpenBSD kernel will "create" one in memory, and that can form the basis of a formal OpenBSD disklabel to be stored on disk. That created disklabel uses partition 'i'. Run disklabel(8) on the device (what you should have done in the first place) to verify this. // marc
Re: Help with v3.9 pf?
> bge1 - Internal Interface, 192.168.0.1 > sk0 - Management Interface, 192.168.0.36 These are on the same network. From your ifconfig: bge1inet 192.168.0.1 netmask 0x broadcast 192.168.255.255 sk0 inet 192.168.0.36 netmask 0xffc0 broadcast 192.168.0.63 I suspect that will cause all kinds of problems. 192.168.0.0/26 link#3 UC 00 - sk0 192.168/16 link#2 UC 00 - bge1 So which interface will packets destined for you gateway of 192.168.0.1 be sent? I *think* the narrow netmask wins so it goes to sk0. What will your pf ruleset do to those packets if seen on sk0? put sk0 on net 10 or narrow it's netmask to /32 (host route) and try again. // marc (just guessing)
Re: IPv6 and illegal prefixlen
> up giftunnel 212.182.166.172 64.71.128.81 > up inet6 2001:470:1F01:::1AE1 2001:470:1F01:::1AE0 prefixlen 128 > !route add -inet6 default 2001:470:1F01:::1AE0 Mine looks like this (and it works just fine) - hostname.gif0 - tunnel 208.201.244.208 208.201.234.221 inet6 alias 2001:05a8:0:1::0123 128 dest 2001:05a8:0:1::0122 ! route add -inet6 default ::1 ! route change -inet6 default -ifp gif0 - hostname.gif0 - With this setup route show also has the "route: illegal prefixlen" message. Ignore it. I don't think it has anything to do with your problem. // marc
Re: mail stats from procmail
> and when I use mailstats(8) (mailstats -C ~/.procmaillog) it says Wrong program. mailstats(8) is part of sendmail. The program you want is mailstat (no man page), installed in /usr/local/bin as part of the procmail package. But you knew that. > located" I also tried the "mailstat" command and it worked fine the > first time but from the second time onward it says "No mail arrived > since Dec 30 20:37" which is wrong because I received mails after that > time. There is also no manpage for "mailstat". Guess: you're looking at the wrong log file. Where does your .procmailrc put the log? Mine says: ... MAILDIR=$HOME/Mail LOGFILE=$MAILDIR/procmail.log ... which is why I tell mailstat to do this out of cron once a day: mailstat /home/marc/Mail/procmail.log // marc
Re: openbsd and APC UPS
> i would like to configure my UPS device for openbsd. But i am unsure > whether it is supported by any of openbsd port's ups programs. Is > there anyone that would allow to handle it under openbsd ? Yes and no. > uhidev1 at uhub3 port 1 configuration 1 interface 0 > uhidev1: American Power Conversion Back-UPS ES 600 FW:820.x1d.D USB FW:x1d, > rev 1.10/0.06, addr 2, iclass 3/0 The APC uhidev support sucks, so I've patched code that makes it work under ugen. ugen0: American Power Conversion Back-UPS RS 1500 FW:8.g8 .D USB FW:g8, rev 1.10/1.06, addr 2 I'm also using an old version of Apcupsd -- 3.10.17. At least I assume it is an old version. I got it and the usb patch from a friend. Since the code works with my unit I've seen no reason to see about changes or updates. Perhaps the need for this changed in the last couple of years. Don't know. If your interested in trying my version grab ftp://neko.snafu.org/pub/apcupsd-cvs.tar.gz The file "uhidev-patch" contains the needed kernel patch. The file "doit" is a shell script that will configure, make and install the daemon. Pay attention to the last 30 or so lines displayed when running the script. It will tell you what was installed and/or patched in /etc. // marc
Re: openbsd and APC UPS
> > >Also take a look at the OpenBSD section because you'll have > > >to recompile the kernel to make it work. > > That's incorrect, use config -e and you can disable uhid* without > changing to an unsupported kernel configuration. 1) The kernel config need not change. 2) disable uhid* would result in keyboards and mice not working. // marc
Re: openbsd and APC UPS
> uhidev-patch: > "don't match APC products (all of them are UPS's)" > > This isn't: > http://www.amazon.co.uk/APC-Keyboard-mouse-adapter-Type/dp/BC120J Yep. You're correct. However, that didn't exist when the patch was created. The patch comes from friend and dates back maybe 3 years. Yes, doing it with a quirk would be better, but at the time that wasn't an option. Since what I have works, I've never bothered changing. > yes. My previous post gave an alternative though (add a quirk). Sorry, perhaps I misread the post. I thought you were saying that the kernel must be left stock. Adding a quirk and re-compiling a kernel doesn't exactly do that. (Nor does my patch). Let me reiterate: what I have works, and has been working for 2 years or more. It is not pretty, nor is it the best way to go. It is acceptable as a "user beware" patch. Use it or not. // marc
Re: .forward for procmail
Exal de jesus Garcia Carrillo writes: > .forward on openbsd?, I have tryed with "|IFS=' ' && exec /usr/bin/procmail > -f- || exit 75 #exal" but doesn't work. /usr/bin/procmail? The port/package installs procmail in /usr/local/bin // marc
Re: acpi?
> This same issue however has stopped me from getting 4.0 to install on
> an intel core duo mac mini as well.
$ sysctl hw
hw.machine=i386
hw.model=Genuine Intel(R) CPU 1300 @ 1.66GHz ("GenuineIntel" 686-class)
hw.ncpu=2
hw.byteorder=1234
hw.physmem=1040629760
hw.usermem=1040621568
hw.pagesize=4096
hw.disknames=cd0,wd0
hw.diskcount=2
hw.cpuspeed=1667
hw.vendor=Apple Computer, Inc.
hw.product=Macmini1,1
hw.version=1.0
hw.serialno=G864806SW0C
hw.uuid=9cfe245e-d0c8-bd45-a79f-54ea5fbd3d97
I did this by 1) creating a bsd.rd with acpi support, and 2) blowing away
MacOS so it would boot from the CD without keyboard input. There have
since been changes made that make the keyboard work from the boot loader
(but not DDB or UKC) so all you need to do is enable acpi.
// marc
Re: raid and separate partitions
> So, question is - how is it possible? It looks like something I'm > missing. I don't know how, but it looks like all www date is written Order is important. It looks like you mounted /var/www before /var in which case /var/www is hidden. /etc/fstab must be in the order that you want filesystems mounted. An /etc/fstab on one of my machines... # devicemount point type optionsbackup fsck pass /dev/wd0a / ffsrw 1 1 /dev/wd0d /varffsrw,nodev,nosuid 1 2 /dev/wd0h /var/wwwffsrw,nodev,nosuid 1 2 ... // marc
Re: ifconfig commands for trunk0 to hostname.trunk0
> is this the way that it always is for configuring /etc/hostname.if > files? basically take the ifconfig command and put everything after > the interface name into the /etc/hostname.if file? Not quite. The format is sometimes different. There is a hostname.if(5) man page that should describe the differences. // marc
Re: LANDISK: USB 10/100 ethernet recommendation?
Diana Eichert writes: > Since I've never used a USB ethernet NIC, I usually live in the 1Gb / 10Gb > world, I thought I'd ask here for chipset recommendations. It took me three tries to get one that worked. The winner was made by linksys. The current version of the D-Link DUB-E100 does not work, nor does an off brand unit from Syba. > It appears axe(4) based USB device would be my best best. axe0 at uhub4 port 3 configuration 1 interface 0 axe0: Cisco-Linksys USB200M v2, rev 2.00/0.01, addr 2, AX88772, address 00:16:b6:ef:6e:83 ukphy0 at axe0 phy 16: Generic IEEE 802.3u media interface, rev. 1: OUI 0x000ec6, model 0x0001 // marc
Re: LANDISK: USB 10/100 ethernet recommendation?
Jonathan Gray writes: > Just about anything you can buy should work. The MosChip MCS7830 > ones (ie syba usb ethernet) being the only exception that comes to mind. The current rev of the D-link DUB-E100 does NOT work. At least it did't with code as of about the middle of last month. // marc
Re: Flash Player 9 on OpenBSD
> gnash looks very cool, but as of right now, it still can't play strongbad > email, nor google video. gnash (last tried was their cvs version as of about 1 month ago) would play maybe 30% of the flash I ran into on my amd64. It has a long way to go to become usable. If enabled I found that it played just about all of the "advertising" content and none of the content I actually wanted to see. // marc
Re: Flash Player 9 on OpenBSD
Frank Denis writes: > Well, I see two ways of having flash work with native apps: And these methods work on my hppa box? Or my Sparc64 box? Or on any non-i386/amd64 box? // marc
Re: Should fopen() succeed on a directory?
Adam writes: > Why does fopen()ing a directory for reading succeed instead of failing > with EISDIR? This has the possibly unexpected consequence of letting I believe it is so things like "grep -r regex *" work. // marc
Re: setting up a memory file system
Peter Matulis writes: > I am having difficulty finding documentation on how to set up a memory > file system from beginning to end. I keep reading about /tmp and swap > and docs that presume certain steps have been accomplished (disklabel). I suspect your difficulty is in expecting it to be more complex than it is. Short answer: add the appropriate line to your /etc/fstab. It really is that simple. My /tmp is an mfs. It is enabled with this line in /etc/fstab: swap /tmp mfsrw,async,noatime,nodev,noexec,-s=524288 0 0 The fstab(5) man page has the some of the nitty gritty. The rest is in the mount_mfs man page. You can ignore the vast majority of fine tuning options. If you want to populate your MFS with content look at the -P option of mount_mfs. // marc
Re: setting up a memory file system
> My main question is the device it uses. The man page has the > device /dev/sd0b. This needs to be set up somewhere. Still, I see > that people use "swap" in its place instead. Swap is the b partition in a generic kernel (and most other kernels, too). Use of "swap" in fstab just simplifies things. You don't have to know that the partition is sd0b or wd0b or some other device or some combination of devices in case swapon is used. // marc
Re: dmesg and fdisk do not match about usb external disk
frantisek holop writes: > so what's up with these dick measurements? I think you got that part just right :-) Expecthing cyl * head * sec/cyl to come up with the number of actual sectors on the disk is your problem. Modern disk don't have a fixed number of sec/track. They use Zone Bit Recording which uses a different number of sec/track depending upon the location of the track on the disk. The code tries to come up with an approximate CHS for historical reasons. It would probably be best if it just reported the number of sectors as that is the only important measure. // marc
amusing greylisting HELO/EHLO identification side-effect
I found it highly amusing that as a result of runnig the latest spamd in greylisting mode with this change Make spamd include the HELO/EHLO identification string sent by the connecting hosts in the tuple key when greylisting. catches a few more bogus hosts and will let us trap based on HELO later. drastically reduced my email -- because of this: GREY|192.43.244.163|<[EMAIL PROTECTED]>|<[EMAIL PROTECTED]>|1172616339|1172630739|1172630739|1|0 GREY|192.43.244.163|<[EMAIL PROTECTED]>|<[EMAIL PROTECTED]>|1172616284|1172630684|1172630684|1|0 GREY|192.43.244.163|<[EMAIL PROTECTED]>|<[EMAIL PROTECTED]>|1172616308|1172630708|1172630708|1|0 GREY|192.43.244.163|<[EMAIL PROTECTED]>|<[EMAIL PROTECTED]>|1172616479|1172630879|1172630879|1|0 etc. Oops. Somehow the white entry for the mailing list host was lost. // marc
Re: amusing greylisting HELO/EHLO identification side-effect
Bob Beck writes: > > etc. Oops. Somehow the white entry for the mailing list host was lost. > > No that indicates that you have messed up. you are showing Of course I messed up. Didn't mean to imply otherwise! > me output of an old spamdb command and claiming to be running the new spamd. See, I messed up again. The new spamd is running on one machine and I showed you the output of another (copy/paste error). > You better upgrade them both at the same time. I certainly didn't > lose entries when upgrading. Updating had little to do with the lost entries. It was my screw up during the update that did the damage. I just thought it amusing that my screw up affected little but the openbsd mailing lists :-) // marc
Re: forcing WD0/WD1 designation (soekris)
> i don't think you can modify attachments with config(8), but You can. Or you can build a custom kernel. My box finds my SATA drive before my ATA drive though I use the ATA drive as wd0. My kernel config has: wd1 at pciide? flags 0x wd0 at pciide? flags 0x wd* at pciide? flags 0x so they match in my desired order. Resulting dmesg is: ... wd1 at pciide0 channel 0 drive 0: wd1: 16-sector PIO, LBA48, 76319MB, 156301488 sectors wd1(pciide0:0:0): using BIOS timings, Ultra-DMA mode 6 wd0 at pciide1 channel 0 drive 0: wd0: 16-sector PIO, LBA48, 76319MB, 156301488 sectors wd0(pciide1:0:0): using PIO mode 4, Ultra-DMA mode 5 ... You can do the same with config -e. Lets see ukc> find wd 42 wd* at wdc*|pciide* channel -1 flags 0x0 ukc> add wd1 Clone Device (DevNo, 'q' or '?') ? 42 Insert before Device (DevNo, 'q' or '?') ? 42 42 wd1 at wdc*|pciide* channel -1 flags 0x0 ukc> add wd0 Clone Device (DevNo, 'q' or '?') ? 43 Insert before Device (DevNo, 'q' or '?') ? 43 43 wd0 at wdc*|pciide* channel -1 flags 0x0 ukc> find wd 42 wd1 at wdc*|pciide* channel -1 flags 0x0 43 wd0 at wdc*|pciide* channel -1 flags 0x0 44 wd* at wdc*|pciide* channel -1 flags 0x0 which will make a generic kernel work on my machine. // marc
Re: xenocara in /usr/src can cause problems ?
> I follow current and it seems to me strange that xenocara is under > /usr/src. I have my src tree in /usr/src and xenocara in > /usr/src/xenocara. So when I update my src tree with cvs it seems that Strangeness is a matter of personal choice. You do not have to put xenocara under /usr/src, it is just that /usr/src/xenocara is the default location. > I have problems : on one computer (src tree originally from CD) cvs > never finish and on an other (src tree from a cvs mirror) the command This is a local configuration issue on your machine. If you look in /usr/src/CVS/Entries you'll see a line that looks something like D/xenocara which is an instruction to the cvs process on your machine to access the xenocara directory when doing updates and other cvs operations. If you don't want xenocara automatically updated when doing a "cvs up" in /usr/src simply remove that line from the file.Instead of entering the xenocara directory and doing an update the cvs process will simply print "? xenocara" and otherwise skip the directory. You can then "cd /usr/src/xenocara; cvs up" when you wish to update the xenocara code. There may be a "proper" way to do what I've suggested using cvs, but I've always found it easier to edit the Entries file. // marc
Re: pf - drop or return - is stealth mode overrated?
"Kian Mohageri" writes: > I see no reason a host should receive any response at all when it is trying > to talk to a host that doesn't exist or a port that isn't actually listening. Traceroute. // marc
Re: Softupdates question
> Still curious how they would work on, say, /var/mysql or /var/postgresql, > but I can play with this on my own. > Has anyone already tried? Care to comment? FWIW I run softdep on ALL partitions except / and /var and have for many years. I exclude /var because on a crash I want the best chance items logged in /var/log/ to show up. // marc
Re: APC UPSD
"Jean-Daniel Beaubien" writes: > If you want to use an APC UPS you might need to compile nut from ports (or > download and compile the latest version). Or use apcuspd. Last time I got the source from sourceforge it just worked. ./configure --enable-usb make sudo make install neko[marc]# /etc/rc.apcupsd restart Stopping apcupsd power management Done. waiting for apcupsd to exit Starting apcupsd power management Done. neko[marc]# /etc/rc.apcupsd status APC : 001,037,0911 DATE : Mon May 21 13:40:56 PDT 2007 HOSTNAME : neko.snafu.org RELEASE : 3.14.0 VERSION : 3.14.0 (9 February 2007) openbsd UPSNAME : neko.snafu.org CABLE: USB Cable MODEL: Back-UPS RS 1500 UPSMODE : Stand Alone STARTTIME: Mon May 21 13:40:53 PDT 2007 STATUS : ONLINE LINEV: 118.0 Volts LOADPCT : 36.0 Percent Load Capacity BCHARGE : 100.0 Percent TIMELEFT : 91.5 Minutes MBATTCHG : 30 Percent MINTIMEL : 10 Minutes MAXTIME : 0 Seconds SENSE: Low LOTRANS : 097.0 Volts HITRANS : 132.0 Volts ALARMDEL : Always BATTV: 26.9 Volts LASTXFER : Low line voltage NUMXFERS : 0 TONBATT : 0 seconds CUMONBATT: 0 seconds XOFFBATT : N/A SELFTEST : NO STATFLAG : 0x0708 Status Flag MANDATE : 2004-09-25 SERIALNO : JB0439032522 BATTDATE : 2001-09-25 NOMINV : 120 NOMBATTV : 24.0 FIRMWARE : 8.g8 .D USB FW:g8 APCMODEL : Back-UPS RS 1500 END APC : Mon May 21 13:40:57 PDT 2007 // marc
Re: 4.1 install issue
"Bruce Bauer" writes: > You nailed it down! > After boot>-s, examining the dmesg shows the SATA drive is wd0 and the > former wd0 is wd1 and the former wd1 is wd2. Now if my thinking is > correct, all I should have to do is edit fstab to reflect the changed > drive positions and the system should be happy. If the loader is looking to wd0 for the kernel changing the fstab will do nothing. I have a custom kernel to set the sata as wd0 # IDE hard drives wd1 at pciide? flags 0x wd0 at pciide? flags 0x wd* at pciide? flags 0x but you can do the same thing with a GENERIC kernel by playing with config -e or boot -c neko[GENERIC]$ config -e bsd OpenBSD 4.1-current (GENERIC) #34: Sun May 27 15:11:12 PDT 2007 [EMAIL PROTECTED]:/usr/src/sys/arch/amd64/compile/GENERIC warning: no output file specified Enter 'help' for information ukc> find wd 42 wd* at wdc*|pciide* channel -1 flags 0x0 ukc> add wd1 Clone Device (DevNo, 'q' or '?') ? 42 Insert before Device (DevNo, 'q' or '?') ? 42 42 wd1 at wdc*|pciide* channel -1 flags 0x0 ukc> add wd0 Clone Device (DevNo, 'q' or '?') ? 43 Insert before Device (DevNo, 'q' or '?') ? 43 43 wd0 at wdc*|pciide* channel -1 flags 0x0 ukc> find wd 42 wd1 at wdc*|pciide* channel -1 flags 0x0 43 wd0 at wdc*|pciide* channel -1 flags 0x0 44 wd* at wdc*|pciide* channel -1 flags 0x0 ukc> // marc
Re: Rename multiple files at once
> for FILE in *jpg; do
> NEW=$(echo $FILE | sed -e 's/\.jpg$/_thumb.jpg/')
> mv "${FILE}" "${NEW}"
> done
There is no need for echo and sed. OpenBSD sh and ksh support
${var%suffix} which evaluates to the contents of var less the suffix.
for f in *.jpg; do mv $f ${f%.jpg}_thumb.jpg; done
does the trick in one line.
// marc
Re: following stable, extra file sets?
Maurice Janssen writes: > >Is there anyway to _not_ get these extra sets as part of > >following stable? I don't know that it hurts anything, but I have no > >use for them on the system and would like to keep it as minimalistic as > >possible. > > I'm sure it's possible to modify the tree in some way to prevent this, > but that's not supported and it may break other things (like cvs > updates). If you want to go unsupported and non-standard you can play with the makefiles. Games is easy: remove "games" from the list of SUBDIR in /usr/src/Makefile. misc is quite a bit harder as it contains the documentation for thing that you still want built and installed. > I guess the easiest way is to build a release on another system and > install only the file sets that you used during the initial > installation. Or, let it install then remove the unneeded files. The source contains a list of everything in a set and in the case of misc everything is machine independat. After a build you could do something like this (untested): cd / # remove the regular files cat /usr/src/distrib/sets/lists/misc/mi | while read f; do test -f $f && rm $f; done # remove the directories tail -r /usr/src/distrib/sets/lists/misc/mi | while read d; do test -d $d && rmdir $d; done // marc
sparcstation 5 for the taking in SF Bay Area
I've some old Sun SPARC equipment that must go. There are two conditions for the taking, though: 1) You must take all of it 2) I'm not going to ship it (I'm in the SF bay area) Even though I won't ship it, it is currently boxed for shipment in two boxes: #1: 25 x 25 x 21 inches, 57 lbs #2: 21 x 21 x 19 inches, maybe 30 lbs (forgot to weigh it) If you want it and you are not local you can schedule a pick-up with your favorite carrier on your dime. What I've got: * Sparcstation 5/110 with 96M ram, 8G disk, TGX (cgsix 1152x900) It's old enough that the idrom battery is dead, requiring a "boot disk" or "boot cdrom" command at the OK prompt to get it going * Solaris 1.1.2 AKA Sunos 4.1.4 CD * 611 enclosere with DDS2 tape drive * 411 enclosure with ST-150 tape drive * 411 enclosure with toshiba cd drive plus 2 CD carriers SCSI target switch not connected, drive coded for target 6 * No name disk enclosure with Seagate ST12400N 2G drive * Andataco disk enclosure with Seagate ST31200NH 1G drive * Various SCSI cables and terminators * Sun parallel <-> centronics cable for printer Front view: http://www.snafu.org/pics/misc/p-20061126-2056-2214.jpg Rear view: http://www.snafu.org/pics/misc/p-20061126-2056-2215.jpg (Pics snapped before it was boxed for shipping) No monitor. No keyboard. Nothing on the disks. The system works well enough to let me boot a bsd.rd from the CD and wipe all the disks. The DDS2 tape drive and the cables might be worth something. Taking the rest of the stuff is the price you'll have to pay to get it :-) // marc
Re: ifconfig problem - Botched upgrade?
> Most definitely, -A is normally the only way to see alias info. Just > wanted to let you know about the /32 mask for aliases. ???. Perhaps I missed some earlier context, but -A is not the only way to see alias info. $ ifconfig msk0 msk0: flags=8843 mtu 1500 lladdr 00:16:cb:a7:52:1c groups: egress media: Ethernet autoselect (100baseTX full-duplex,rxpause,txpause) status: active inet 208.201.244.208 netmask 0xff00 broadcast 208.201.244.255 inet6 fe80::216:cbff:fea7:521c%msk0 prefixlen 64 scopeid 0x1 inet 208.201.244.210 netmask 0x broadcast 208.201.244.210 inet6 2001:5a8:4:910:216:cbff:fea7:521c prefixlen 64 $ Note the second line from the bottom. It is an alias as are the inet6 entries. Note the command line. -A was not given. Per the man page: -A Causes full interface alias information for each interface to be displayed. Shows all aliases on all interfaces. Not necessary (and not valid) if you only want to look at one interface. // marc
Re: How do I configure Cyclades Z serial ports with OpenBSD?
> When my machine boots, I see: > > "Cyclades Cyclom-Z" rev 0x01 at pci1 dev 9 function 0 not configured > > So the OS/driver does see the card. You have that backwards. The OS does see the card, thus the message. The "not configured" part means the OS is not configured with a driver that supports that card. Not surprising as that support is not compiled into the GENERIC kernel. You'll have to create a custom kernel that doesn't comment out this line (from i386 generic -- I assume you're running an i386 box: you didn't say). #cz*at pci? # Cyclades-Z multi-port serial boards // marc
Re: using spamd to grey-TRAP *only*, with *no* grey-LIST delays, stutters, etc ?
snowcrash+openbsd writes: > > You've got the source. Why not read it and figure out the answer for > yourself? > > Source being available is true for just about everything, now, isnt't it? Well, I see you quoted from a private message I sent, and a selected quote at that. While noting that I said you could read the source to get your answers you neglected to mention that I read it for you and explained *exactly* what setting the delay to zero would do. > Signing off now -- just not worth the effort in here. Bye. // marc
Re: OBSD's perspective on SELinux
> Burroughs Computers essentially went out of business because their > computers refused to do illegal operations while IBM's computers > very happily did all sorts of illegal stuff. Way off topic here... Burroughs became part of Unisys and the architecture that "refused to do illegal operations" still exists in products sold today. Well, its emulated on intel hardware these days, so perhaps that doesn't count. I miss coding in algol :-) // marc
Re: The Name: UNIX
> Does OpenBSD = UNIX? Or, does OpenBSD = Unix? (or unix or unix-like or etc.) You are the first person I've seen to makes a distinction between unix, Unix, and UNIX and imply that the distinction has any meaning. // marc
Re: dmesglog request
> Some time ago I tried to submit a bug report through sendbug > but couldn't get it through. I didn't investigate much but IIRC > it was because the hostname of the box wasn't valid (missing MX records > or whatever) and the receiving smtp daemon discarded the message. Standard anti-spam action these days. The contents of /etc/myname had better be a full qualified doman name that matches the address of the box and the reverse DNS should point back to that name, too. If it doesn't then your mail will often be rejected. That's assuming you're not getting your address from dhcp, of course. If you get your address from your ISP via dhcp you'll have to find out what your ISP needs to set the hostname to something of your liking. Otherwise it is likely to be something like 206-45-94-138.static.mts.net (someone currently stuck in my greytrap database :-) // marc
Re: dmesglog request
Antti Harri writes: > > Standard anti-spam action these days. > [rest snipped] > > Maybe so, but I find it very strange for sendbug's server to have > such restrictions. I even didn't get any errors back. You find it strange that a site with a well known address and thus a magnet for spam would use spam filtering? I'd find it strange if it didn't. // marc
Re: Spamd - whitelisting round robin mail servers?
Jeff Simmons writes: > all out of date, and the link to the cvs list is broken. Anyone know of any > uptodate compilations? $ host -ttxt google.com google.com descriptive text "v=spf1 include:_netblocks.google.com ~all" $ host -ttxt _netblocks.google.com _netblocks.google.com descriptive text "v=spf1 ip4:216.239.32.0/19 ip4:64.233.160.0/19 ip4:66.249.80.0/20 ip4:72.14.192.0/18 ip4:209.85.128.0/17 ip4:66.102.0.0/20 ip4:74.125.0.0/16 ip4:64.18.0.0/20 ip4:207.126.144.0/20 ?all" That should cover google, no? // marc
Perl changes and majordomo
I notices that majordomo now gives this warning when running the digest command: $* is no longer supported at /usr/local/lib/majordomo/digest line 305. I assume it started when perl was updated to 5.10.0. As one who dislikes perl enough to have never learned it a clue as to what it means would be appreciated :-) // marc
Re: Perl changes and majordomo
On Nov 25, 2008, at 1:00 PM, Ingo Schwarze wrote: Then again, majordomo is long dead and has a bad license. Perhaps it is not worth so much effort, and the time might be better spent migrating to something else... mlmmj comes to mind... Thanks for that idea. Somehow I missed mlmmj last time I went looking for a majordomo replacement. It took some playing to make up for the dearth of documentation, but I've already converted one mailing list and see no reason why I shouldn't convert all and trash majordomo. // marc
Re: Firefox 2.0.0.12
Zbigniew Baniewski writes: > > ldconfig -SP /usr/bin /usr/sbin /usr/local/bin /usr/local/sbin > > /usr/X11R6/bin > > Just tried the sequence - can't see any difference, unfortunately. Not suprising as the firefox binary is not in any of the given paths. // marc
Re: minimac on openbsd
Hey Tas... I'm trying to bring a second mini on line and am having a problem that I don't recall how I worked around the first go around. Perhaps you can apply a clue bat! I hooked up a monitor and keyboard, installed a CD created from the latest snapshot, and was happy to see it boot OpenBSD as I powered it up while holding the C key. Much easier than I remember the first time when I had to build a custom i386 kernel. Alas, my joy was short lived: once the bsd.rd is running the USB keyboard isn't recognized. I'm sitting at the (I)nstall, (U)pgrade, or (S)hell? prompt and can't go any further. Do you remember the the necessary tweak to get bsd.rd to handle the USB keyboard? I don't remember this issue the last time I tried to bring one on line. Thanks, // marc
Re: minimac on openbsd
Otto Moerbeek writes: > On Thu, Apr 24, 2008 at 09:28:55PM -0700, Marco S Hyman wrote: > > Alas, my joy was short lived: once the bsd.rd is running the > > USB keyboard isn't recognized. I'm sitting at the (I)nstall, > > (U)pgrade, or (S)hell? prompt and can't go any further. Heh... didn't realize I copied misc. > On a macbook, not all USB ports are created equal. You have to use the > frontmost USB port for the external keyboard when installing. Perhaps > something similar is the case for the mini? A cookie for Otto. Moving the keyboard from the port closest to the edge to the port closest to the video connector (guess, from port 4 to port 1) did the trick. It was either that or the fact that this time I booted an amd64 CD instead of the i386 CD. My guess is that it was moving the keyboard that did the trick. Many, many thanks. Now on to some disk partitioning. // marc
Re: cwm keybindings misbehavior
> > CWM doesn't support Xinerama. > > It doesn't. It will. Xinerama is useful. Ditto EWMH. Ehh? I ran Xinerama. I ran CWM. It worked. CWM saw my dual screen as one big screen. For just about everything I wantede to do that was good enough. // marc
Re: Squid -> Privoxy -> Tor
Peter_APIIT writes: > I very disappointed with this forum because no one willing to help me. It is not a web forum, it is a mailing list. No one on the list has any obligation to help you. If you want support figure out your budget then call one of the commercial organizations listed at http://www.openbsd.org/support.html If you're not willing to pay for support and you're not willing to wait for some fellow user to read your message and maybe reply with something helpful spend your time reading the docs and figure it out yourself. Then, in the future when someone else asks the question you'll be able to answer. It is possible that you are the first on the mailing list to try to do whatever it is you want to do. // marc
Re: newsyslog default archived directory
"Dongsheng Song" writes: > OpenBSD execute /usr/bin/newsyslog every hour, where newsyslog > archived logs to ? Did you think to read the man page for newsyslog? It mentions a file called newsyslog.conf. Hmmm, a configuration file. I wonder what's in it. Oh, the man page also says SEE ALSO... newsyslog.conf(5). Read the man pages. Read the configuration file. You'll get your answer. // marc
Re: newsyslog default archived directory
I wrote... > Did you think to read the man page for newsyslog? It mentions > a file called newsyslog.conf. Hmmm, a configuration file. I > wonder what's in it. Oh, the man page also says SEE ALSO... > newsyslog.conf(5). O ops... Our man page does not have the SEE ALSO entry (it should, IMHO) but does list /etc/newsyslog.conf in the FILES section. I was on a different system when I checked. > Read the man pages. Read the configuration file. You'll get > your answer. That still stands. // marc
Re: inserting non-printable characters in sed?
Fred Snurd writes: > I have been looking at sed to insert non-printable characters into text > files. The sed(1) manpage states that I should be able to insert' > octal values by preceding the three character octal value with a backslash. The only place the word "octal" shows up in the sed man page is in the description of the OUTPUT of the [2addr]l command. It is not described as a meaningful input for the [2addr]s/re/replacement/flags command. // marc
Mac Mini (intel) status
What is the status of the Mac Mini? I saw some messages and patches regarding the mini flowing by a week or three ago but didn't own one at that time so wasn't paying attention. Are the patches in CVS? Thanks, // marc
Re: Mac Mini (intel) status
> Yes, works really fine on my mini. :-) Last time I installed a Thank you. The goal is to have the mini replace my dying sparc64 as a web server. Small, low power draw, quiet: I like that. // marc
Re: Mac Mini (intel) status
> Some recent CD: The 4.0 release CD and snapshots from Nov. 12, 2006 or > later should boot fine, a few snapshots between that didn't work. You > don't need a -current boot CD to install a -current (or snapshot) file > set. The CD I tried was a home grown snapshot of -current from October That may fall into your "a few snapshots between that didn't work" period. I'm building a new one now and will give it a try. I don't have a bluetooth keyboard to play with. I can, however, make a bsd.rd that has ACPI it that's what it takes. Time to play some more. > A 75 ohm resistor (100 ohm works also) as load between pin 2 and pin 7 > of a 15 pin highdensity D-SUB male connector (with DVI to VGA adapter) > or between pin C2 and C5 of a 29 pin DVI male connector as a dongle > for the DVI/VGA port (that's between analog green and analog ground, > 75 ohm is the standard load for that output pin). Just plug in the > dongle and the mini will boot happily also without a monitor. :-) Time to scrounge around for a connector. Maybe a trip to Fry's is in order. // marc
Re: Mac Mini (intel) status
> Ah, yes, booting from CD. Maybe I was really a little bit lucky with > that because it worked quite well here right from the beginning. ;-) I'm not being so lucky :-( > Boot Camp: No, it's not required, it works fine with a usual > OpenBSD-only configured internal harddisk, at least with > Boot-ROM-Version MM11.0055.B05 and Boot-ROM-Version MM11.0055.B08. Of macos says I'm running MM11.0055.B08. When I first fired the system up (just to verify that it was working OK) it downloaded new firmware and had be go throught the power cycle dance holding the power button to do the update so I expect the box came with MM11.0055.B05. > Some recent CD: The 4.0 release CD and snapshots from Nov. 12, 2006 or > later should boot fine, a few snapshots between that didn't work. You > don't need a -current boot CD to install a -current (or snapshot) file > set. Not working for me. I get this far: CD_ROM: 90 Loading /CDBOOT probing: pc0 com0 mem(699K 991M a20=on) disk: hd0+* cd0 boot> c and there it stays forever. I suspect the "c" following the boot prompt is left over from "hold c to boot from cd". The keyboard at this point is dead. Any ideas? I'd really like to get OpenBSD up on this beasty. I've tried several different home grown CDs plus the 11/29 snapshot CD from ftp.openbsd.org. // marc
Re: Mac Mini (intel) status
"Tasmanian Devil" writes: > > Well, at least theoretically, one could just replace the install script > > by one that does whatever you want it to, without asking any questions. > > Maybe that's easier than building an ACPI enabled bsd.rd? No, building an ACPI enabled bsd.rd was quite easy. It took more tries to build a working bsd for the running system -- just adding ACPI and MP to the generic kernel does NOT work with -current code -- than it did to get a working bsd.rd. However, the system dies under load. Took be about 4 tries to complete a build on the box. The lack of a keyboard under ddb doesn't make debugging easy. So I turned of ddb to get a crash but forgot to relocate /var/crash. Dumb, /var wasn't big enough. Then I got a crash, but had erased the bsd.gdb that went with it. Next time I build a new kernel I'll look at the dumps. // marc
Re: Mac Mini (intel) status
"Tasmanian Devil" writes: > Hmm... sounds like I shouldn't update my source tree (it's still from > Nov. 12, 2006). I had no crashes at all here so far. Are you running stock Nov. 12, 2006 or that code plus patches? As for creating a bsd.rd that works with the mini... The easiest way to do that is to follow the instructions in release(8) to generate a home grown release directory and burn that directory onto a CD. Before issuing the "make release" command in /usr/src/etc replace the file /sys/arch/i386/conf/RAMDISK_CD with a config file that will create a bsd.rd that the mini can use. The result is a custom release CD with a bsd.rd for the mini and containing all of the release sets. I've a config file that works, but I'm still tweaking it. When I'm finished I'll post it. The biggest issue is that option SMALL_KERNEL cant be used with the ACPI devices so, to make space, you have to remove many of the unused devices. I wound up removing just about everything that wasn't mentioned in the mini dmesg. // marc
Re: Mac Mini (intel) status
Miod Vallat writes:
> > I've a config file that works, but I'm still tweaking it. When I'm
> > finished I'll post it. The biggest issue is that option SMALL_KERNEL
> > cant be used with the ACPI devices so, to make space, you have to
> > remove many of the unused devices. I wound up removing just about
> > everything that wasn't mentioned in the mini dmesg.
>
> Do you mean the acpi code depends on functionality SMALL_KERNEL removes?
> If so, which one?
Yep. Let see. Don't remember. Ok, I'll just re-build with SMALL_KERNEL
and see what it was
There are two ACPI ramdisk issues
1) acpi code references sensor_add and sensor_task_unregister which is
in kern_sensors.c. kern_sensors.c is !small_kernel
2) /sys/dev/acpi/dsdt.c defines DST which interferes with option DST
in the config file. I changed DST -> DST1 in /sys/dev/acpi/dsdt.c
That same file also references tsleep but does not include
for the function prototype. Patch with both changes follows.
Index: dsdt.c
===
RCS file: /u8/cvso/src/sys/dev/acpi/dsdt.c,v
retrieving revision 1.67
diff -u -p -r1.67 dsdt.c
--- dsdt.c 29 Nov 2006 22:17:07 - 1.67
+++ dsdt.c 2 Dec 2006 20:47:34 -
@@ -19,6 +19,7 @@
#include
#include
#include
+#include
#include
@@ -802,7 +803,7 @@ int aml_parsenode(struct aml_scope *,str
#define LHS 0
#define RHS 1
-#define DST 2
+#define DST1 2
#define DST2 3
/* Allocate temporary storage in this scope */
@@ -2897,20 +2898,20 @@ aml_parsestring(struct aml_scope *scope,
tmpval = aml_alloctmp(scope, 4);
aml_parseterm(scope, &tmpval[LHS]);
aml_parseterm(scope, &tmpval[RHS]);
- aml_parsetarget(scope, &tmpval[DST], NULL);
+ aml_parsetarget(scope, &tmpval[DST1], NULL);
if (tmpval[LHS].type == AML_OBJTYPE_BUFFER &&
tmpval[RHS].type == AML_OBJTYPE_BUFFER) {
aml_resize(&tmpval[LHS],
tmpval[LHS].length+tmpval[RHS].length);
memcpy(&tmpval[LHS].v_buffer+tmpval[LHS].length,
tmpval[RHS].v_buffer, tmpval[RHS].length);
- aml_setvalue(scope, &tmpval[DST], &tmpval[LHS], 0);
+ aml_setvalue(scope, &tmpval[DST1], &tmpval[LHS], 0);
}
if (tmpval[LHS].type == AML_OBJTYPE_STRING &&
tmpval[RHS].type == AML_OBJTYPE_STRING) {
aml_resize(&tmpval[LHS],
tmpval[LHS].length+tmpval[RHS].length);
memcpy(&tmpval[LHS].v_string+tmpval[LHS].length,
tmpval[RHS].v_buffer, tmpval[RHS].length);
- aml_setvalue(scope, &tmpval[DST], &tmpval[LHS], 0);
+ aml_setvalue(scope, &tmpval[DST1], &tmpval[LHS], 0);
}
else {
aml_die("concat");
More on mini panics
gdb says:
(gdb) target kvm bsd.3.core
#0 0xd037e8dd in dumpsys () at /sys/arch/i386/i386/machdep.c:2650
2650if ((error = (*dump)(dumpdev, blkno,
(gdb) bt
#0 0xd037e8dd in dumpsys () at /sys/arch/i386/i386/machdep.c:2650
#1 0xd037e55c in boot (howto=256) at /sys/arch/i386/i386/machdep.c:2449
#2 0xd029da51 in panic (fmt=0xd045aa30 "trap type %d, code=%x, pc=%x")
at /sys/kern/subr_prf.c:221
#3 0xd038c3d8 in trap (frame=
{tf_fs = -676265984, tf_gs = -800194560, tf_es = 16, tf_ds = 16, tf_edi =
-778722000, tf_esi = -795422300, tf_ebp = -373113348, tf_ebx = 1, tf_edx = 0,
tf_ecx = 5591040, tf_eax = -372506624, tf_trapno = 6, tf_err = 0, tf_eip =
-801610747, tf_cs = 8, tf_eflags = 66119, tf_esp = 0, tf_ss = 3485696,
tf_vm86_es = -795422300, tf_vm86_ds = 3, tf_vm86_fs = 214484, tf_vm86_gs = 0})
at /sys/arch/i386/i386/trap.c:270
#4 0xd0200ed2 in calltrap ()
#5 0xd0386405 in pmap_page_remove_86 (pg=0xd12cc100)
at /sys/arch/i386/i386/pmap.c:2968
#6 0xd0353e9b in uvn_flush (uobj=0xd7aad4e0, start=0, stop=0, flags=20)
at pmap.h:414
#7 0xd0353a81 in uvn_detach (uobj=0xd7aad4e0) at /sys/uvm/uvm_vnode.c:404
#8 0xd03488f6 in uvm_unmap_detach (first_entry=0xd7af4d6c, flags=0)
at /sys/uvm/uvm_map.c:1613
#9 0xd034ac4d in uvmspace_free (vm=0xd7b3ece8) at /sys/uvm/uvm_map.c:3369
#10 0xd0345293 in uvm_exit (p=0xd7b1a9cc) at /sys/uvm/uvm_glue.c:296
#11 0xd0288c4a in reaper () at /sys/kern/kern_exit.c:433
#12 0xd02004ea in proc_trampoline ()
#13 0xd7c54168 in ?? ()
#14 0x00563000 in ?? ()
#15 0xd0555020 in ?? ()
#16 0x in ?? ()
core dump and matching bsd.gdb available.
// marc
panic in pmap_page_remove
I've got a reproducable crash on my Mac mini (intel). Sources are as of 30 Nov. No, I'm not running a generic kernel because a generic kernel doesn't run on the mini. And the keyboard doesn't work in ddb so getting more info is difficult at best. Here are photos of two different crashes. The first was while doing a make build, the second while building some ports. The crash and stack traces are pretty close in each case: panic panic trap --- trap (number 6) --- pmap_page_remove uvm_vnp_terminate uvn_attach uvm_unmap_detach uvmspace_free uvm_exit reaper http://www.snafu.org/crash/p-20061202-0035-2230.jpg http://www.snafu.org/crash/p-20061202-1454-2232.jpg Ideas? I'm trying to get a crash dump, but this last time the box hung at syncing disks... // marc
Re: problem with spamd-white table and greylisting
peter dunaskin writes: > > > My configuration: > >You forgot spamlogd > I did not, it starts automatically when spamd_grey is enabled in Oh. Sorry. I didn's see it in your ps grep. I see it with the same ps grep on my system. dumbcat[root]# ps aux | grep spamd _spamd 14450 0.0 0.1 9040 1376 ?? IsTue12AM1:53.41 spamd: (pf update) (spamd) _spamd 25485 0.0 0.8 8968 9592 ?? S Tue12AM0:50.54 /usr/libexec/spamd -g _spamd 10132 0.0 0.1 8968 1272 ?? I Tue12AM0:00.60 spamd: (/var/db/spamd update) (spamd) _spamd 14234 0.0 0.1 536 1216 ?? SsTue12AM0:09.53 /usr/libexec/spamlogd root 11990 0.0 0.0 1008 0 p1 R+ 7:36AM0:00.00 grep spamd (ksh) // marc
