streams module no longer compiles
David: If I read things correctly you relocated the svr4 pieces from sys to sys/compat in todays -CURRENT. It appears that the streams module depends on several include files which you moved. As a result, the streams module no longer compiles. It fails in make depend with the errors mkdep -f .depend -a -nostdinc -D_KERNEL -DKLD_MODULE -I- -I. -I@ -I@/../include -I/usr/include /usr/src/sys/modules/streams/../../dev/streams/streams.c /usr/src/sys/modules/streams/../../dev/streams/streams.c:57: svr4/svr4_types.h: No such file or directory /usr/src/sys/modules/streams/../../dev/streams/streams.c:58: svr4/svr4_util.h: No such file or directory /usr/src/sys/modules/streams/../../dev/streams/streams.c:59: svr4/svr4_signal.h: No such file or directory /usr/src/sys/modules/streams/../../dev/streams/streams.c:60: svr4/svr4_ioctl.h: No such file or directory /usr/src/sys/modules/streams/../../dev/streams/streams.c:61: svr4/svr4_stropts.h: No such file or directory /usr/src/sys/modules/streams/../../dev/streams/streams.c:62: svr4/svr4_socket.h: No such file or directory mkdep: compile failed *** Error code 1 changing CFLAGS as follows - CFLAGS+=-O +CFLAGS+=-O -I{$CURDIR}/../../sys/compat seems to fix the problem I am not sure whether this is correct form, there may be a better way. Regards, George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Linux ABI no longer supports staroffice
Marcel: Up until this weekend, I was able to use the staroffice52 port with little problem (I had installed it earlier without benefit of the port and it worked fine.) I did a 5.0-current kernel rebuild on Thursday with sources current on that day and things were fine. When I rebuilt my kernel yesterday afternoon with sources from Saturday morning, the port stopped working. I get the following error messages I18N: X Window System doesn't support locale "C" _X11TransSocketOpen: socket() failed for local _X11TransSocketOpenCOTSClient: Unable to open socket for local _X11TransOpen: transport open failed for local/dinolt1.bingdrive:0 setup.bin: cannot open display ":0.0" Please check your "DISPLAY" environment variable, as well as the permissions to access that display (See "man X" resp. "man xhost" for details) I am running the install as root. The Window Manager is up and running, other programs including linux-netscape have no trouble running in the same environment. This is not a clue for me, but it may mean something to others. (By the way, these errors did not appear prior to the changes.) I was wondering whether the setrlimit changes had something to do with this. This seems to be the major change to the linux code in the last few days. Apparently you were the one to make those changes, so I am writing to you (Marcel) Anyway, thought someone should know. George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Failure build libkvm
Quick report: I saw a failure building libkvm while trying to do a build world using sources cvsuped aound 1.00 A.M. Pacific Daylight Time (U.S.) Sunday, Aug 6. The error messages I saw were cc -O -pipe -DLIBC_SCCS -I/usr/src/lib/libkvm/../../sys -I/usr/obj/usr/src/i386/usr/include -c /usr/src/lib/libkvm/kvm_proc.c -o kvm_proc.o In file included from /usr/src/lib/libkvm/kvm_proc.c:57: /usr/src/lib/libkvm/../../sys/sys/tty.h:84: field `t_rsel' has incomplete type /usr/src/lib/libkvm/../../sys/sys/tty.h:85: field `t_wsel' has incomplete type *** Error code 1 Maybe someone has an idea of what has happened recently. Unfortunately, (actually, for obscure reasons it is fortunately) I have to work on Sunday so I won't be able to investigate further. George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Even with version 1.54 of mfs_vnops.c and a freshly built worldand booted kernel (with ``options MFS''),I am still seeing these. What could I be missing?
Mahtew Jacob wrote: Hmm. They went away for me- I thought, but no, you're right. They're still there. This is what I actually use to get rid of them: Index: mfs_vnops.c === RCS file: /home/ncvs/src/sys/ufs/mfs/mfs_vnops.c,v retrieving revision 1.53 diff -u -r1.53 mfs_vnops.c --- mfs_vnops.c 2000/05/05 09:59:07 1.53 +++ mfs_vnops.c 2000/07/25 21:47:09 @@ -80,6 +80,7 @@ { vop_reclaim_desc,(vop_t *) mfs_reclaim }, { vop_strategy_desc, (vop_t *) mfs_strategy }, { vop_unlock_desc, (vop_t *) vop_defaultop }, + { vop_getwritemount_desc, (vop_t *) vop_stdgetwritemount }, { NULL, NULL } }; static struct vnodeopv_desc mfs_vnodeop_opv_desc = I have applied the patch and rebuilt my kernel. This seems to have fixed the problem for me. I will keep monitoring the situation. Maybe this can get committed to -current? Thanks, George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Even with version 1.54 of mfs_vnops.c and a freshly built world andbooted kernel (with ``options MFS''), I am still seeing these. What could I bemissing?
Thanks, -- Jos Backus [EMAIL PROTECTED] WebTV Networks, Inc., Mountain View, CARe: Still seeing /kernel: mfs_badop[vop_getwritemount] = 45 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Jos Backus wrote: Even with version 1.54 of mfs_vnops.c and a freshly built world and booted kernel (with ``options MFS''), I am still seeing these. What could I be missing? Thanks, -- Jos Backus [EMAIL PROTECTED] WebTV Networks, Inc., Mountain View, CA I can confirm that on my system with world built from sources cvsuped on Sunday and a kernel just built from sources cvsuped today, I see these messages with an mfs mounted /tmp. Regards, George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
(noperiph:ahc0:0:-1:-1): ... error
I am trying to run a recent (as of today) and am seeing the following error when I try to boot:: (noperiph:ahc0:0:-1:-1): SCSI bus reset delivered. 0 SCBs aborted. panic: Bogus resid sgptr value 0xbd68609 (I copied this from the console after the boot failure, there may be minor mistakes.) This started happening when I started compiling kernels built from sources cvsuped around Jul 18. I am not sure what is causing these messages. The "noperiph" message appears to come from xpt_print_path in /usr/src/sys/cam/cam_xpt.c while the panic seems to be written by ahc_calc_residual in /usr/src/sys/dev/aic7xxx/aic7xxx.c. From a quick look at the code, the problem is not directly in the code pointed to by the messages. I have an Adaptec 2940UW. A much older kernel reports it as Adaptec 2940 Ultra SCSI adapter with aic7880 Wide Channel A, SCSI Id=7, 16/255 SCBs. The Bios on the board is version 2.20.0 I have 4 drives and a UMAX scanner connected to the bus. More details available if needed. I am hopeful that someone will recognize the problem. Regards, George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Bootstrapping perl ...
Mark: Got through the buildworld with a recent cvsup. I am now running with this "world". Many thanks. I am not sure who is supposed to manage the fact that /usr/ports/Mk/bsd.ports.mk doesn't yet know about perl-5.006. One appears to need to set PERL_VER and PERL_VERSION. I just added an ."if {OSVERSION} = 56" , but this clearly isn't enough for people who have yet to upgrade to the newest "current'. Does this mean that the OSVERSION needs to be incremented to indicate the perl upgrade? Isn't life wonderful. Solutions to problems always lead to new problems. Regards, George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Bootstrapping perl ...
I have found that the following small diff to /usr/src/Makefile.inc1 allows the buildworld to proceed: *** Makefile.inc1~ Mon Jun 26 23:00:54 2000 --- Makefile.inc1 Mon Jun 26 23:27:05 2000 *** *** 123,129 WORLDTMP= ${OBJTREE}${.CURDIR}/${BUILD_ARCH} # /usr/games added for fortune which depend on strfile STRICTTMPPATH= ${WORLDTMP}/usr/sbin:${WORLDTMP}/usr/bin:${WORLDTMP}/usr/games ! TMPPATH= ${STRICTTMPPATH}:${PATH} # # Building a world goes through the following stages --- 123,129 WORLDTMP= ${OBJTREE}${.CURDIR}/${BUILD_ARCH} # /usr/games added for fortune which depend on strfile STRICTTMPPATH= ${WORLDTMP}/usr/sbin:${WORLDTMP}/usr/bin:${WORLDTMP}/usr/games ! TMPPATH= ${STRICTTMPPATH}:${OBJTREE}${.CURDIR}/gnu/usr.bin/perl/miniperl:${PATH} # # Building a world goes through the following stages This is probably only a temporary fix and is not the right approach, but it does illustrate that the problem is that the miniperl from /usr/bin/miniperl was being called by the /usr/src/gnu/usr.bin/perl/Makefile.inc and not the one built during the buildworld. Since "miniperl" is referenced in the perl/Makefile.inc either one has to "hardcode" the version of miniperl, i.e. include the path of the version of miniperl for each reference in that makefile, or one should install miniperl someplace in ${WORLDTMP} which is included in the path. The former appears to be more in the spirit of "build-tools" since these "tools' do not install their targets. They assume that the targets operate in place. The later (installing miniperl inside of ${WORLDTMP) would require an update to the build-tools target in perl/miniperl. Maybe none of the above is the correct approach, but these are some ideas. Regards, George Dinolt Note: My buildworld has not completed as I write this but it has gotten through the bootstrap and makedepends part of the perl build and I need to get to bed. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: linux broken
Michael: I suspect that you may have options COMPAT_LINUX defined in your kernel config file. Re your problem with Linux not working. I had the same problem. It seems to be related to the changes in the generation of the file linux_assym.h. A new method for generating these kind of symbols was introducted into the kernel build system today. The changes were not promulgated to the generation of linux_assym.h. The following change to /usr/src/modules/linux/Makefile will allow one to at least build a workable linux module. *** Makefile~ Sat May 27 10:04:09 2000 --- MakefileSat Jun 3 00:11:05 2000 *** *** 17,23 CLEANFILES= linux_assym.h linux_genassym.o linux_assym.h:linux_genassym.o ! genassym ${.ALLSRC} ${.TARGET} linux_locore.o: linux_locore.s linux_assym.h ${CC} -c -x assembler-with-cpp -DLOCORE ${CFLAGS} \ --- 17,23 CLEANFILES= linux_assym.h linux_genassym.o linux_assym.h:linux_genassym.o ! sh /usr/src/sys/kern/genassym.sh ${.ALLSRC} ${.TARGET} linux_locore.o: linux_locore.s linux_assym.h ${CC} -c -x assembler-with-cpp -DLOCORE ${CFLAGS} \ If you don't keep your srcs in /usr/src, you will have to modify the fixes appropriately I am sure that there is a much cleaner way of fixing this, but I will leave it to the gurus to work that out. I think a similar fix to the kernel Makefile would allow one to build a working kernel with COMPAT_LINUX built in. There may be other modules or pieces of the kernel which may be affected by the new symbol generation facility. I have not been bitten by the problem yet nor have I looked for them, so I can't speak first hand of the issues. Hope this helps. Regards, George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Problem building a snapshot for current, no more more
Andreas Klemm wrote: -current of June 01: For me it was in stage 4 at: ln -sf libm.so.2 /usr/obj/usr/src/i386/usr/lib/libm.so cd /usr/src/lib/libncurses; make depend; make all; make install ./make_keys /usr/src/lib/libncurses/../../contrib/ncurses/ncurses/tinfo/keys.lis t init_keytry.h sh /usr/src/lib/libncurses/../../contrib/ncurses/ncurses/base/MKlib_gen.sh "cc - E -I. -I/usr/src/lib/libncurses -I/usr/src/lib/libncurses/../../contrib/ncurses/ ncurses -I/usr/src/lib/libncurses/../../contrib/ncurses/include" "awk" curses .h | fgrep undef nomacros.h *** Error code 1 Stop in /usr/src/lib/libncurses. *** Error code 1 Stop in /usr/src. *** Error code 1 -- Andreas Klemm http://people.FreeBSD.ORG/~andreas http://www.freebsd.org/~fsmp/SMP/SMP.html powered by Symmetric MultiProcessor FreeBSD New APSFILTER 542 and songs from our band - http://people.freebsd.org/~andreas Actually, once I followed the directions that "John W. DeBoskey" [EMAIL PROTECTED] sent me in a private messages things worked fine. I had no troubles with curses or anything else. My cvs src tree was from late May 31. The instructions he sent me were - Hi, I've already fixed this. Unfortunately, fixit.flp is now larger than 1.4Meg. The following patch will get you past the problem: Index: fixit_crunch.conf === RCS file: /home/ncvs/src/release/fixit_crunch.conf,v retrieving revision 1.22 diff -u -r1.22 fixit_crunch.conf --- fixit_crunch.conf 2000/05/31 15:26:03 1.22 +++ fixit_crunch.conf 2000/06/01 03:20:17 @@ -28,8 +28,7 @@ # /usr/bin stuff -progs ftp less telnet vi grep -ln less more +progs ftp telnet vi grep ln vi view ln vi ex However, you'll have no pager on the fixit floppy. You should be able to run 'chroot /snap/release /mk' to continue your build (substituting your own release build area in the above). Good Luck, John --- I am currently trying this with a cvs tree current as of about 12:05 A.M (PDT) on June 2. These include changes listed above. I will let you know how things go. Regards, George Dinolt. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Problem building a snapshot for current, no more more
In attempting to build a snapshot for current, I came across the following problem in the release.4 target of the Makefile in /usr/src/release. ... cc -static -o boot_crunch boot_crunch.o sh.lo find.lo sed.lo test.lo rm.lo pwd.lo ppp.lo sysinstall.lo newfs.lo minigzip.lo cpio.lo fsck.lo ifconfig.lo route.lo slattach.lo mount_nfs.lo dhclient.lo arp.lo hostname.lo pccardc.lo pccardd.lo usbd.lo usbdevs.lo -ll -ledit -lutil -lkvm -lmd -lcrypt -lftpio -lz -lnetgraph -ldialog -lncurses -lmytinfo -L/usr/src/release/libdisk/obj -ldisk -lipx dhclient.lo: In function `script_init': dhclient.lo(.text+0x416b): warning: mktemp() possibly used unsafely; consider using mkstemp() strip boot_crunch crunchgen: /usr/src/release/fixit_crunch.conf: more: warning: could not find source directory crunchgen: /usr/src/release/fixit_crunch.conf: more: warning: could not find any .o files crunchgen: /usr/src/release/fixit_crunch.conf: more: error: no objpaths specified or calculated crunchgen: /usr/src/release/fixit_crunch.conf: more: ignoring program because of errors Run "make -f fixit_crunch.mk objs exe" to build crunched binary. *** Error code 1 Stop in /usr/src/release. Note that the crunchgen could not find the sources for "more." Does this have something to do with the replacement of "more" by "less" recently? I am not sure what the correct fix is. Someone with more understanding of what is going on might want to address this. For example, should one replace "more" by "less" (and "lesskey" and "lessecho") ? I have not tried this and tested whether things would still fit on the floppy. George Dinolt To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Failed compile of ext2_alloc.c
Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Robert: I compiled recent (cvsup around 9:00 A.M. PDT) sources for the 5.0 kernel with and without options FFS_EXTATTR and obtained the following error message in both cases. cc -c -O -pipe -march=k6 -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -fformat-extensions -ansi -nostdinc -I- -I. -I../.. -I../../../include -D_KERNEL -include opt_global.h -elf -mpreferred-stack-boundary=2 ../../gnu/ext2fs/ext2_alloc.c In file included from ../../gnu/ext2fs/ext2_alloc.c:54: ../../ufs/ufs/ufsmount.h:90: field `um_extattr' has incomplete type *** Error code 1 I have not investigated. I think this may have to do with your additions of extended attributes. Hope this helps. George Dinolt [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Failed compile of ext2_alloc.c
Robert: Re the supplied patches. I applied them and the kernel compiled, loaded and ran. There seem to be some problems with the modules I use for my ethernet cards so I was not able to give it a "thorough" testing. I don;t think the module issues relate to the file system. I seemed to have hit an inconsistency between the modules (fxp and dc) and kernel as part of the CVS updates I have been doing. .I hope to investigate more later. Regards, George Dinolt [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: newpcm (or newpnp?)
Chris Giordano wrote: Previous to newpnp, I forced the modem (sio2) to irq 15 and the soundcard to irq 5 via pnp commands in my /boot/kernel.conf. I did this because I had found by experimentation that this was the only combination of the supposedly available PnP configurations which enabled both devices to work. I can report similar problems with my US Robotics internal 56K Fax/Modem and my SoundBlaster 16 isa card. On Freebsd 3.4 I "hard wire" the cards using kernel.conf file. On -Current I found that the only way I can get the cards to work is to configure my IWILL XA100P motherboard so that irq 3 and 5 are set up for "legacy" cards and are not automagically assigned. (I disable the 2nd serial port as well so that irq 3 is available.) When I do this reconfiguration of the motherboard, the modem and sound card work together with few problems. I can play sound and do a compilation without problems. I do have 192MB of memory add a 450Mhz processor so there is plenty of spare "horsepower" for essentially a "single user" system. I do occasionally get a flutter on startup and can reliably see a cutoff of sound at the end of reading a wave or midi file. But for most things it seems to work fine. The "say" program from the rsynth port provides a good demonstration of the "cutoff" phenomena. George Dinolt ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: -current buildworld dies, retch.
Otay, please tell me how to fix: === usr.sbin/ifmcstat cc -O -pipe -DINET6 -I/usr/obj/usr/src/i386/usr/include -c /usr/src/usr.sbin/ifmcstat/ifmcstat.c gzip -cn /usr/src/usr.sbin/ifmcstat/ifmcstat.8 ifmcstat.8.gz /usr/src/usr.sbin/ifmcstat/ifmcstat.c: In function `main': /usr/src/usr.sbin/ifmcstat/ifmcstat.c:109: storage size of `arpcom' isn't known It appears that in the change from "KERNEL" to "_KERNEL" the usage was not updated in ifmcstat.c. The following patch appears to fix the problem, (I just completed a make world) and ifmcstat seems to work. -- diff -c ifmcstat-old/ifmcstat.c ifmcstat/ifmcstat.c *** ifmcstat-old/ifmcstat.c Wed Dec 29 19:26:58 1999 --- ifmcstat/ifmcstat.c Wed Dec 29 21:59:06 1999 *** *** 48,58 #include netinet/in.h #ifndef __NetBSD__ # ifdef __FreeBSD__ ! # define KERNEL # endif # include netinet/if_ether.h # ifdef __FreeBSD__ ! # undef KERNEL # endif #else # include net/if_ether.h --- 48,58 #include netinet/in.h #ifndef __NetBSD__ # ifdef __FreeBSD__ ! # define _KERNEL # endif # include netinet/if_ether.h # ifdef __FreeBSD__ ! # undef _KERNEL # endif #else # include net/if_ether.h To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Netscape core dump, happily :-)
I am running FreeBSD 4.0. I have been seeing the empty core files from netscape 4.61 and (as of last night) 4.7 for at least a week. I use fvwm2 and afterstep-devel. The behavior seems to be independant of the window manager. I rebuilt all my ports (including X and new X aout files) with the new kernel and world as of Saturday, but this has had no effect. Mostly this behavior is anoying, although occasionally I will see a crash when I try to read mail. Things seem a little stabler with netscape in linux emulation mode, but I haven't kept track closely enough to tell. -- George ([EMAIL PROTECTED]) To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message