Boot problem custom Kernel 5.7
Hello, for around 10 years I build a custom cd bootable version of OpenBSD. The project was started with OpenBSD 3.4 and is updated with each OpenBSD release. Basically I take the GERNERIC kernel configuration and change config bsd swap generic to config bsd root on rd0a And I also add: option RAMDISK_HOOKS option MINIROOTSIZE=4480 pseudo-device rd 1 The rest are changes to the ramdisk configuration. I added mount_mfs to the ramdisk and some scripts to boot from cd. Starting with OpenBSD 5.7 I cannot build a bootable kernel and I don't know why. The kernel build itself shows no errors, but when I boot from CD I get: Booting from DVD/CD... CD-ROM: E0 Loading /CDBOOT probing: pc0 com0 mem[639K 698; a20=on] disk: hd0+ cd0 OpenBSD/amd64 CDBOOT 3.23 boot cannot open cd0a:/etc/random.seed: No such file or directory booting cd0a:/bsd.gz: 5798084+2048752+2532360+0+589824=0xa782d8 entry point at 0x1000160 [7205c766, 3404, 24448b12, 1080a304] == At this point the system reboots. No further messages. Booting with -c results also in a reboot. Then I took the RAMDISK_CD configuration and added option MFS. This time the kernel boots with the customised ramdisk, but the RAMDISK_CD lacks some features like the pf firewall. Adding more features to the RAMDISK_CD configuration I cannot compile the kernel anymore. So the question is what can I do to find out the cause of the problem or maybe someone has already a clue what the problem is. Thanks, Oliver
Re: caron no longer working in xterm
On May 13 18:20:23, h...@stare.cz wrote: On Apr 22 07:03:22, s...@spacehopper.org wrote: On 2015-04-21, Jan Stary h...@stare.cz wrote: On Mar 31 19:24:00, h...@stare.cz wrote: Since about three snapshots ago, I cannot use the caron (háček) in xterm to diacritize my Czech (latin2) writings. This used to work. No really, this used to work. Now it stopped working on the machines I upgraded. Writing a non-capitalized caron'd letter in vim (in an xterm) and then capitalizing it with ~ does produce the correct capitalized caron'd version of the letter. But typing it straight like in [caron]S still does not work. Jan Are you able to backtrack by date in cvs to identify what caused it? Was it the xkeyboard-config update? Probably not. The diff in /usr/xenocara/data/xkeyboard-config between 6 months ago and now (below) doesn't seem to contain anythong related to the cz keyboard. The diff to the previous state goes back way longer before I started seeing this. What else could have caused this? Thanks Jan Index: man/xkeyboard-config.7 === RCS file: /cvs/xenocara/data/xkeyboard-config/man/xkeyboard-config.7,v retrieving revision 1.9 retrieving revision 1.7 diff -u -p -u -p -r1.9 -r1.7 --- man/xkeyboard-config.730 Mar 2015 17:11:29 - 1.9 +++ man/xkeyboard-config.717 Sep 2014 16:11:50 - 1.7 @@ -1,5 +1,5 @@ .\ WARNING: this man page is autogenerated. Do not edit or you will lose all your changes. -.TH XKEYBOARD-CONFIG 7 xkeyboard-config 2.14 X Version 11 +.TH XKEYBOARD-CONFIG 7 xkeyboard-config 2.12 X Version 11 .SH NAME xkeyboard-config \- XKB data description files .SH DESCRIPTION @@ -126,7 +126,7 @@ logidinovoedgeLogitech diNovo Edge Keyb mx1998 Memorex MX1998 mx2500 Memorex MX2500 EZ-Access Keyboard mx2750 Memorex MX2750 -microsoft4000Microsoft Natural Ergonomic Keyboard 4000 +microsoft4000Microsoft Natural Wireless Ergonomic Keyboard 4000 microsoft7000Microsoft Natural Wireless Ergonomic Keyboard 7000 microsoftinetMicrosoft Internet Keyboard microsoftpro Microsoft Natural Keyboard Pro / Microsoft Internet Keyboard Pro @@ -244,7 +244,6 @@ ara(digits) Arabic (digits) ara(qwerty) Arabic (qwerty) ara(qwerty_digits) Arabic (qwerty/digits) ara(buckwalter) Arabic (Buckwalter) -ara(mac) Arabic (Macintosh) _ al Albanian @@ -539,7 +538,6 @@ _ is Icelandic is(Sundeadkeys) Icelandic (Sun dead keys) is(nodeadkeys) Icelandic (eliminate dead keys) -is(mac_legacy) Icelandic (Macintosh, legacy) is(mac) Icelandic (Macintosh) is(dvorak) Icelandic (Dvorak) @@ -691,8 +689,6 @@ ru(mac) Russian (Macintosh) ru(srp) Serbian (Russia) ru(bak) Bashkirian ru(chm) Mari -ru(phonetic_azerty) Russian (phonetic azerty) -ru(phonetic_fr) Russian (phonetic French) _ rs Serbian @@ -1005,7 +1001,6 @@ ctrl:menu_rctrl Menu as Right Ctrl ctrl:ctrl_ralt Right Alt as Right Ctrl ctrl:swap_lalt_lctl Swap Left Alt key with Left Ctrl key ctrl:swap_lwin_lctl Swap Left Win key with Left Ctrl key -ctrl:swap_rwin_rctl Swap Right Win key with Right Ctrl key ctrl:swap_lalt_lctl_lwin Left Alt as Ctrl, Left Ctrl as Win, Left Win as Alt .TE @@ -1161,7 +1156,7 @@ lB l. Option Description numpad:pcDefault numeric keypad keys numpad:mac Numeric keypad keys always enter digits (as in Mac OS) -numpad:microsoft NumLock on: digits, Shift switches to arrow keys, Numlock off: always arrow keys (as in MS Windows) +numpad:microsoft Shift with numeric keypad keys works as in MS Windows numpad:shift3Shift does not cancel Num Lock, chooses 3rd level instead srvrkeys:noneSpecial keys (Ctrl+Alt+key) handled in a server apple:alupckeys Apple Aluminium Keyboard: emulate PC keys (Print, Scroll Lock, Pause, Num Lock) @@ -1253,22 +1248,6 @@ Option Description japan:kana_lock Kana Lock key is locking japan:nicola_f_bsNICOLA-F style Backspace japan:hztg_escapeMake Zenkaku Hankaku an additional ESC - -.TE - - -.SS -Korean Hangul/Hanja keys -.BR -.TS -left,box; -lB lB -___ -lB l. -Option Description -korean:hw_keys Hardware Hangul/Hanja keys -korean:ralt_rctrlRight Alt as Hangul, right Ctrl as Hanja -korean:rctrl_raltRight Ctrl as Hangul, right Alt as Hanja .TE Index: pc/Makefile === RCS file: /cvs/xenocara/data/xkeyboard-config/pc/Makefile,v retrieving revision 1.5 retrieving revision 1.3 diff -u -p -u -p -r1.5 -r1.3 --- pc/Makefile 30 Mar 2015 17:11:29 - 1.5 +++ pc/Makefile 17 Sep 2014 16:11:50 - 1.3 @@ -1,7 +1,7 @@ -#$OpenBSD: Makefile,v 1.5 2015/03/30 17:11:29
Re: Boot problem custom Kernel 5.7
entry point at 0x1000160 [7205c766, 3404, 24448b12, 1080a304] == At this point the system reboots. No further messages. Your kernel is probably too large. A limitation in the bootblocks.
Re: Boot problem custom Kernel 5.7
entry point at 0x1000160 [7205c766, 3404, 24448b12, 1080a304] == At this point the system reboots. No further messages. Your kernel is probably too large. A limitation in the bootblocks. Do you mean the kernel with the ramdisk or without? What would be the best way to workaround this limit? Decrease ramdisk size? Remove driver/options from the kernel? Those are your changes, outside the OpenBSD tree. You are on your own.
Re: Blob-free OpenBSD kernel needed
Hi On 2015-06-06, Артур Истомин wrote: Your rant is cogent. But if so, why OpenBSD does not supply microcode updates from Intel/AMD? There are tons of security fixes. I just wonder: Is there really such microcode available that is open source? -- Greetings Elias
Re: Correction to the AnonCVS documentation
Please ignore/delete this. The SMTP server I was using was initially blocked, and thanks to the list owner, it is now fixed. As a result, this message may have been backlocked and sent. It is essentially a duplicate of another message, which has received reasonable replies. On Mon, 8 Jun 2015 04:04:57 +1000 bytevolc...@safe-mail.net wrote: Hello all, I noticed on http://www.openbsd.org/faq/faq5.html#BldGetSrc that there is information about preloading the tree, but does not mention that getting to -current requires -rHEAD at least the first time using 'cvs update' after pre-loading the tree with the source files from the last release. I discovered (the semi-hard way) that to get to -current, you need -rHEAD to be specified in the command line. This is after pre-loading the tree, because the (src/sys/xenocara).tar.gz files all have a check out of OPENBSD_5_7 (or whatever -stable release), frozen to that point at the time of releasing. Running cvs up simply gets me to the latest -stable. My .cvsrc (between the hyphen lines): --- # $OpenBSD: dot.cvsrc,v 1.1 2013/04/01 16:55:26 espie Exp $ # cvs -q -danon...@anoncvs.au.openbsd.org:/cvs diff -up update -Pd checkout -P --- It seems the (src/sys/xenocara).tar.gz files are not in the snapshots directory of any of the FTP mirrors I've seen. I am willing to write a patch for this, but I want to ensure: 1. I am not missing anything hidden somewhere in another FAQ. 2. There are no mirrors which have the (src/sys/xenocara).tar.gz files in their /pub/OpenBSD/snapshots directory. 3. That there is no other CVS command magic that I have missed, that may have been covered by the FAQ. 4. That there are no other reasons not to document this information. I will appreciate input from users and devs alike.
Correction to the AnonCVS documentation
Hello all, I noticed on http://www.openbsd.org/faq/faq5.html#BldGetSrc that there is information about preloading the tree, but does not mention that getting to -current requires -rHEAD at least the first time using 'cvs update' after pre-loading the tree with the source files from the last release. I discovered (the semi-hard way) that to get to -current, you need -rHEAD to be specified in the command line. This is after pre-loading the tree, because the (src/sys/xenocara).tar.gz files all have a check out of OPENBSD_5_7 (or whatever -stable release), frozen to that point at the time of releasing. Running cvs up simply gets me to the latest -stable. My .cvsrc (between the hyphen lines): --- # $OpenBSD: dot.cvsrc,v 1.1 2013/04/01 16:55:26 espie Exp $ # cvs -q -danon...@anoncvs.au.openbsd.org:/cvs diff -up update -Pd checkout -P --- It seems the (src/sys/xenocara).tar.gz files are not in the snapshots directory of any of the FTP mirrors I've seen. I am willing to write a patch for this, but I want to ensure: 1. I am not missing anything hidden somewhere in another FAQ. 2. There are no mirrors which have the (src/sys/xenocara).tar.gz files in their /pub/OpenBSD/snapshots directory. 3. That there is no other CVS command magic that I have missed, that may have been covered by the FAQ. 4. That there are no other reasons not to document this information. I will appreciate input from users and devs alike.
Re: Boot problem custom Kernel 5.7
On 2015-06-09 13:34, Oliver wrote: On Tue, 09 Jun 2015, Theo de Raadt wrote: entry point at 0x1000160 [7205c766, 3404, 24448b12, 1080a304] == At this point the system reboots. No further messages. Your kernel is probably too large. A limitation in the bootblocks. Do you mean the kernel with the ramdisk or without? What would be the best way to workaround this limit? Decrease ramdisk size? Remove driver/options from the kernel? Back when I was building live CDs/DVDs, I used GENERIC or GENERIC.MP kernels with cd0a as the root filesystem, and MFS mounted /etc, /var, /root, /tmp, /home, and /dev. Read-only filesystems on media were all CD9660. I ran into cdboot(8) issues with very large /usr/local filesystems (e.g. KDE and Gnome). My simple circumvention was to vnconfig(8) /usr/local from an .iso image file after boot completed.
Re: Boot problem custom Kernel 5.7
On Tue, 09 Jun 2015, Theo de Raadt wrote: entry point at 0x1000160 [7205c766, 3404, 24448b12, 1080a304] == At this point the system reboots. No further messages. Your kernel is probably too large. A limitation in the bootblocks. Do you mean the kernel with the ramdisk or without? What would be the best way to workaround this limit? Decrease ramdisk size? Remove driver/options from the kernel?
Re: Blob-free OpenBSD kernel needed
2015-06-09 18:48 GMT+02:00 Elias Diem li...@webconect.ch: I just wonder: Is there really such microcode available that is open source? No.
Re: caron no longer working in xterm
On Tue, Jun 09, 2015 at 04:43:45PM BST, Jan Stary wrote: Writing a non-capitalized caron'd letter in vim (in an xterm) and then capitalizing it with ~ does produce the correct capitalized caron'd version of the letter. But typing it straight like in [caron]S still does not work. Hi Jan, Has anything changed in terms of default(?) keyboard shortcuts? Maybe it's not xterm but your DE or WM that's changed? This wouldn't be a first[0] ;^) [0] https://medium.com/medium-eng/the-curious-case-of-disappearing-polish-s-fa398313d4df Cheers, Raf
Error while building userland
Hello misc@, I am building -current for the first time and I have an error while building userland. More specifically, when rebuilding the symbolic link, make spew this error: *** Parse error in /usr/src/gnu/usr.bin/binutils: Malformed conditional (${BINUTILS_VERSION} == binutils-2.17) (Makefile.bsd-wrapper:13) *** Parse error: Need an operator in 'binutils-2.17' (Makefile.bsd-wrapper:13) *** Error 1 in gnu/usr.bin (bsd.subdir.mk:48 'obj') *** Error 1 in gnu (bsd.subdir.mk:48 'obj') *** Error 1 in /usr/src (bsd.subdir.mk:48 'obj') I followed as closely as I can instruction from section 5 of the FAQ. I installed a snapshot, build a GENERIC kernel. cvs is up to date. I tried from both snapshot and freshly compiled kernel for good mesure but both give the same result. As I said, it's my first time building a kernel. I was using release or snapshot until now. So it might be an error from my part. I don't know what to do at this point so any pointer is welcome. -- Cordialement, Coues Ludovic +336 148 743 42
Re: OpenBSD 5.7 CD 2
On Mon, Jun 1, 2015 at 10:13 AM, OpenBSD Store Misc m...@openbsdstore.com wrote: CD 2 has been dispatched for replacement on all orders. arrived for me today on the east coast of the USA (New York City area). -ken
Re: AnonCVS and -rHEAD
On 06/07/15 22:57, bytevolc...@safe-mail.net wrote: On Sun, 07 Jun 2015 22:27:05 -0400 Nick Holland n...@holland-consulting.net wrote: ... Where did you get these source files? Just checked the ones on the mirrors, they do not have such a tag on them. I'll check the CDs..um...not tonight. ... I am quite certain the ones I got were from the ftp mirror. I've had them for almost a month, but I recall getting them from ftp.iinet.net.au. Now that you mention it, this is actually the first time I've encountered such a thing while updating my copy of the source tree to -current; I've never had to do this with tarballs of earlier releases. yeah. You might want to look into this. I've checked files from my mirror, ftp.iinet.net.au, and the CD. None of them have this 5.7 tag you claim is there. Now, IF you had a machine you ran on an older release, checked out a -stable version of the tree, then tried to upgrade to -current and overwrote the tree with the new source tar.gz file -- the old Tag files would still be there, and not altered or removed by the newer tar.gz files. Nick.
Web2py running on openbsd 5.7 with new httpd server
Hi, I would like to know if anyone have web2py running on OpenBSD 5.7 using new httpd server. I've started web2py with fcgihandler and tried a simple configuration with no luck (probably wrong, that's why I am asking here): # cat /etc/httpd.conf # $OpenBSD: httpd.conf,v 1.14 2015/02/04 08:39:35 florian Exp $ # # Macros # ext_addr=XX.XX.XX.XX # # Global Options # prefork 2 # # Servers # # A minimal default server server default { listen on $ext_addr port 80 location /* { fastcgi socket \ /tmp/fcgi.sock } root /var/www/htdocs } # Include MIME types instead of the built-in ones types { include /usr/share/misc/mime.types } Any help would be very appreciated. Thanks!! Regards, Alvaro