Re: lockup after resume
On Wed, 28 Mar 2001, Georg-W. Koltermann wrote: I am experiencing a strange lockup with -current as of about a week ago: It will suspend and resume, but after the resume the console is dead and the system hangs after a short while. When I type on the console after a resume, nothing is shown, neither echo nor command output. If I break into DDB, the output suddenly appears, just above the DDB prompt. I can continue to UNIX, type another command, again nothing visible. Breaking into DDB again shows what I typed, and the output. After a few round-trips of this sort the system locks up solidly. If I'm in X (XFree86-4.0.2) after the resume, the system will respond for a few seconds and then lock up. All this happens with the GENERIC kernel as well as my cardbus kernel. Should I assume the console needs resetting after the resume? How could I try a reset? I can't find anything obvious in vidcontrol(1) or kbdcontrol(1). Assume that the i8254 needs reinitializing. The console driver just uses timeouts for screen updates. Timeouts depend on the i8254 generating interrupts. When you break into ddb, the screen gets updated directly. Bruce To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: Building aout ports in -current
On Wed, Mar 28, 2001 at 11:07:02PM -0600, Conrad Sabatier wrote: On 29-Mar-2001 Kris Kennaway wrote: On Wed, Mar 28, 2001 at 10:20:33PM -0600, Conrad Sabatier wrote: I've been notified that one of the ports I'm responsible for, xswallow, will no longer build under -current. Here's a snip from the build log: a.out support is entirely optional these days for ports, so it's up to you whether or not you care about fixing your port to build that way. Well, this is a plugin for Netscape, and (unless I'm mistaken) *must* be in aout format, or will ELF plugins work with Netscape??? Oops, you're right. Kris PGP signature
Re: Fun way to panic -current
Bruce Evans wrote: On Wed, 28 Mar 2001, John Baldwin wrote: On 28-Mar-01 Terry Lambert wrote: Run the 4.3 mountd on it. Boom! Kernel memory allocation way to large; unrecoverable! Does this really panic -current? It panics old versions of -current, and the -current mountd panics RELENG_4, but current versions of -current are supposed to check the parameters passwd by mountd (etc.) enough to avoid the panic. FYI, running releng_4 mountd on a releng_3 system also fails in a fairly spectacular fashion... occasionally panic'ing the system. Doug -- Perhaps the greatest damage the American system of education has done to its children is to teach them that their opinions are relevant simply because they are their opinions. Do YOU Yahoo!? To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
subscribe
subscribe To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
-current reliably panics during make release (current process = dd)
On a -current from this morning, I can reliably trigger a panic by "make release" (I believe the problem has existed for some time, though). The box can do make worlds without problem, but during make release, it stumples at the same place every time. This is a Dual P3 with all involved filesystems (/usr/src, /usr/obj and the chrootdir) nullfs-mounted from a SU vinum raid01 volume. monster# mount /dev/da0s1a on / (ufs, local, noatime, soft-updates) devfs on /dev (devfs, local) /dev/da3s1e on /usr (ufs, local, noatime, soft-updates) /dev/da0s1f on /var (ufs, local, noatime, soft-updates) procfs on /proc (procfs, local) /dev/vinum/vol/raid01 on /raid01 (ufs, NFS exported, local, noatime, soft-updates) /raid01/usr/src on /usr/src (nullfs, local, noatime) /raid01/usr/obj on /usr/obj (nullfs, local, noatime) /raid01/usr/ports on /usr/ports (nullfs, local, noatime) monster# The panic: Fatal trap 12: page fault while in kernel mode cpuid = 1; lapic.id = fault virtual address = 0x58 fault code = supervisor read, page not present instruction pointer = 0x8:0xc016749e stack pointer = 0x10:0xe009ddf4 frame pointer = 0x10:0xe009ddf4 code segment= base 0x0, limit 0xf, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags= interrupt enabled, resume, IOPL = 0 current process = 88329 (dd) kernel: type 12 trap, code=0 CPU1 stopping CPUs: 0x0001... stopped. Stopped at devsw+0x6: cmpl$0,0x58(%eax) db trace devsw(0,c41c3280,0,4,e0b8a36c) at devsw+0x6 vn_ioctl(c41c3280,4004667a,e009ded0,dfb3d740,dfb3d740) at vn_ioctl+0xb5 ioctl(dfb3d740,e009df80,805b5c0,bfbff868,bfbff880) at ioctl+0x20a syscall(2f,2f,2f,bfbff880,bfbff868) at syscall+0x405 syscall_with_err_pushed() at syscall_with_err_pushed+0x1b db I'll try to get a crashdump - 1GB (if anyone's interested?). Config attached. /Niels Chr. -- Niels Christian Bank-Pedersen, NCB1-RIPE. Network Manager, Tele Danmark NET, IP-section. "Hey, are any of you guys out there actually *using* RFC 2549?" machine i386 cpu I686_CPU ident MONSTER maxusers128 hints "/boot/device.hints" #makeoptionsDEBUG=-g options INET options FFS options SOFTUPDATES options NFS options CD9660 options DEVFS options PROCFS options COMPAT_43 options SCSI_DELAY=5000 options UCONSOLE options USERCONFIG options VISUAL_USERCONFIG options KTRACE options SYSVSHM options SYSVMSG options SYSVSEM options P1003_1B options _KPOSIX_PRIORITY_SCHEDULING options KBD_INSTALL_CDEV options SMP options APIC_IO options DDB device isa device pci device fdc device ata #device atadisk device atapicd #device atapifd #device atapist #optionsATA_STATIC_ID device sym device scbus device da #device sa #device cd device pass device atkbdc 1 device atkbd device psm device vga device splash device sc 1 device npx device sio device ppc device ppbus device lpt device miibus device fxp device md device random device loop device ether device pty device bpf
Re: malloc not recurseable?
* Michael Class [EMAIL PROTECTED] [010329 02:36] wrote: Hello, on a FBSD-5.0-current (as of yesterday) system I am having a question about the way malloc works. The manpage states that malloc cannot be recursed. Does this mean that our malloc is not thread safe? I checked the manpage on HPUX and there it explicitely states that malloc is thread-safe. The reason I am asking is that my X-Server is crashing randomly (not very often though, approx. once a day) with the following trace: (This is XFree86-4.0.3 from ports) #0 0x2820a9e8 in kill () from /usr/lib/libc.so.5 #1 0x2825bb3d in abort () from /usr/lib/libc.so.5 #2 0x2825a682 in isatty () from /usr/lib/libc.so.5 #3 0x2825a6b0 in isatty () from /usr/lib/libc.so.5 #4 0x2825b6a6 in malloc () from /usr/lib/libc.so.5 #5 0x80d19ff in Xalloc (amount=16) at utils.c:1225 #6 0x80cc30c in TimerSet (timer=0x0, flags=0, millis=50, func=0x8788ef0, arg=0x88acb00) at WaitFor.c:744 #7 0x87890fa in ?? () #8 0x878927d in ?? () #9 0x8788bf0 in ?? () #10 0x807da23 in xf86SigioReadInput (fd=7, closure=0x88acb00) at xf86Events.c:1039 #11 0x8093d48 in xf86SIGIO (sig=23) at sigio.c:99 #12 0xbfbfffac in ?? () #13 0x2825ad74 in isatty () from /usr/lib/libc.so.5 #14 0x2825afcd in isatty () from /usr/lib/libc.so.5 #15 0x2825b6f1 in malloc () from /usr/lib/libc.so.5 #16 0x80d19ff in Xalloc (amount=256) at utils.c:1225 #17 0x87578ac in ?? () #18 0x8757a98 in ?? () #19 0x8758f13 in ?? () #20 0x8759e26 in ?? () #21 0x80c8b6d in QueryFont (pFont=0x8c22a00, pReply=0xbfbfeba4, nProtoCCIStructs=256) at dixfonts.c:580 #22 0x80ad639 in ProcQueryFont (client=0x8b52d00) at dispatch.c:1388 #23 0x80ac045 in Dispatch () at dispatch.c:456 #24 0x80bc395 in main (argc=3, argv=0xbfbffdc0, envp=0xbfbffdd0) at main.c:439 #25 0x806b31d in _start () To me this looks like malloc is called in a signal-handler. But I am not sure, if this is the right interpretation. Any comments? There's a difference between being thread safe and signal safe. Malloc is not signal safe and the XFree coders need to remove the signal handler's use of malloc, basically rework the problem they are trying to solve. -- -Alfred Perlstein - [[EMAIL PROTECTED]|[EMAIL PROTECTED]] Represent yourself, show up at BABUG http://www.babug.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: NEWCARD broken in -current
On 28-Mar-01 Jesper Skriver wrote: On Wed, Mar 28, 2001 at 10:15:21PM +0200, Niels Chr. Bank-Pedersen wrote: On Wed, Mar 28, 2001 at 10:09:28PM +0200, Jesper Skriver wrote: cc -c -O -pipe -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -fformat-extensions -ansi -nostdinc -I- -I. -I/usr/src/sys -I/usr/src/sys/dev -I/usr/src/sys/../include -I/usr/src/sys/contrib/dev/acp ica/Subsystem/Include -D_KERNEL -include opt_global.h -elf -mpreferred-stack-boundary=2 /usr/src/sys/dev/pccbb/pccbb.c In file included from /usr/src/sys/dev/pccbb/pccbb.c:56: /usr/src/sys/sys/mutex.h:87: field `mtx_object' has incomplete type /usr/src/sys/dev/pccbb/pccbb.c: In function `pccbb_detach': /usr/src/sys/dev/pccbb/pccbb.c:533: warning: implicit declaration of function `MPASS' /usr/src/sys/dev/pccbb/pccbb.c:533: warning: implicit declaration of function `LOCK_LOG_LOCK' /usr/src/sys/dev/pccbb/pccbb.c:533: warning: implicit declaration of function `WITNESS_LOCK' /usr/src/sys/dev/pccbb/pccbb.c:539: warning: implicit declaration of function `WITNESS_UNLOCK' *** Error code 1 You'll need to #include sys/types.h and sys/lock.h in /usr/src/sys/dev/pccbb/pccbb.c Right, the below fixes it, should I commit ? Index: src/sys/dev/pccbb/pccbb.c === RCS file: /home/ncvs/src/sys/dev/pccbb/pccbb.c,v retrieving revision 1.12 diff -u -r1.12 pccbb.c --- src/sys/dev/pccbb/pccbb.c 2001/02/09 06:08:52 1.12 +++ src/sys/dev/pccbb/pccbb.c 2001/03/28 20:51:20 @@ -53,6 +53,8 @@ #include sys/kernel.h #include sys/kthread.h #include sys/malloc.h +#include sys/types.h +#include sys/lock.h #include sys/mutex.h #include sys/bus.h Please sort the includes or at least attempt to as that is style(9). You'll need to mvoe lock.h up before malloc.h. types.h is special, it belongs as the very first header unless sys/param.h is the first header (param.h includes types.h). -- John Baldwin [EMAIL PROTECTED] -- http://www.FreeBSD.org/~jhb/ PGP Key: http://www.baldwin.cx/~john/pgpkey.asc "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
i586 FP optimizations hosed.
Hi I have an SMP kernel with I586 and I686 support. If I boot it on a 686 it works. On a 586 it craps out with CPU1 stopping CPUs: 0x0001... Stopped. Stopped at i586_bzero_oops+0x1:jmp i586_bzero_oops late into the boot (during network/RPC init stuff). I thought the 586 FP stuff was disabled? M -- Mark Murray Warning: this .sig is umop ap!sdn To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: NEWCARD broken in -current
On 29-Mar-01 Jesper Skriver wrote: On Thu, Mar 29, 2001 at 01:13:09PM -0800, John Baldwin wrote: Index: src/sys/dev/pccbb/pccbb.c === RCS file: /home/ncvs/src/sys/dev/pccbb/pccbb.c,v retrieving revision 1.12 diff -u -r1.12 pccbb.c --- src/sys/dev/pccbb/pccbb.c 2001/02/09 06:08:52 1.12 +++ src/sys/dev/pccbb/pccbb.c 2001/03/28 20:51:20 @@ -53,6 +53,8 @@ #include sys/kernel.h #include sys/kthread.h #include sys/malloc.h +#include sys/types.h +#include sys/lock.h #include sys/mutex.h #include sys/bus.h Please sort the includes or at least attempt to as that is style(9). You'll need to mvoe lock.h up before malloc.h. types.h is special, it belongs as the very first header unless sys/param.h is the first header (param.h includes types.h). The above was committed, so would the below fix it right ? Index: src/sys/dev/pccbb/pccbb.c === RCS file: /home/ncvs/src/sys/dev/pccbb/pccbb.c,v retrieving revision 1.13 diff -u -r1.13 pccbb.c --- src/sys/dev/pccbb/pccbb.c 2001/03/29 10:23:45 1.13 +++ src/sys/dev/pccbb/pccbb.c 2001/03/29 22:06:51 @@ -52,9 +52,8 @@ #include sys/errno.h #include sys/kernel.h #include sys/kthread.h -#include sys/malloc.h -#include sys/types.h #include sys/lock.h +#include sys/malloc.h #include sys/mutex.h #include sys/bus.h Sure, as long as it compiles. :) -- John Baldwin [EMAIL PROTECTED] -- http://www.FreeBSD.org/~jhb/ PGP Key: http://www.baldwin.cx/~john/pgpkey.asc "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: make release broken in telnetd
Hi!
This was tricky. Due to the old bug in release/Makefile (it did
not pass -DRELEASE_CRUNCH when building list of object files for
crunched binary), ${OBJS} list for ppp was computed incorrectly,
and ppp/Makefile had a special glue to build empty object files:
: .if defined(RELEASE_CRUNCH)
: # We must create these objects because crunchgen will link them,
: # and we don't want any unused symbols to spoil the final link.
: CFLAGS+=-DNONAT -DNORADIUS -DNOI4B -DNOSUID
: OBJS+= chap_ms.o mppe.o id.o nat_cmd.o radius.o
: chap_ms.o mppe.o id.o nat_cmd.o radius.o:
: null_${.PREFIX}.c
: cc -c -o ${.TARGET} null_${.PREFIX}.c
: .endif
Recall that release/Makefile executes `make subclean all' against
the crunchgen(1) generated .mk file. Previously, `subclean' was
executed with the -DRELEASE_CRUNCH; this removed chap_ms.o, and
subsequent `make all' had a chance to build chap_ms.o from the
stub rule above:
: # make -n -DRELEASE_CRUNCH chap_ms.o
: null_chap_ms.c
: cc -c -o chap_ms.o null_chap_ms.c
Now that -DRELEASE_CRUNCH is moved to crunchgen(1) .conf files
(recall that I needed this so that ${OBJS} are computed correctly
for telnet/Makefile), ppp/Makefile got broken. `subclean' does
not cleans chap_ms.o, and subsequent `make all' considers it
up-to-date.
The attached patch should fix this. Please let me know...
Actually, I have just committed a fix to crunchgen(1) so that
it runs `make clean' with the ${BUILDOPTS}, to avoid possible
failures in the future. This fix alone should be enough to
fix the broken `make release', but please test with the attached
patch too.
: ru 2001/03/30 00:04:25 PST
:
: Modified files:
: usr.sbin/crunch/crunchgen crunchgen.c
: Log:
: `buildopts' may affect the selection of object files.
: Make sure we pass $(BUILDOPTS) to the `clean' target
: so that `make clean' works on the same set of object
: files. Otherwise, we may end up with an incorrectly
: built and up-to-date object file.
:
: Revision ChangesPath
: 1.26 +2 -2 src/usr.sbin/crunch/crunchgen/crunchgen.c
On Thu, Mar 29, 2001 at 07:10:57PM +0200, John Hay wrote:
Hi Ruslan,
Could you please try the attached patch and let me know?
I had to move -DRELEASE_CRUNCH to *_fixit.conf so that
${OBJS} are computed correctly for usr.bin/telnet.
I have tried it, but now it breaks in boot_crunch:
##
cc -O -pipe-DCRUNCHED_BINARY -c tunefs_stub.c
ld -dc -r -o tunefs.lo tunefs_stub.o /usr/obj//usr/src/sbin/tunefs/tunefs.o
crunchide -k _crunched_tunefs_stub tunefs.lo
cc -static -o boot_crunch boot_crunch.o sh.lo find.lo sed.lo test.lo rm.lo pwd.l
o 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 rtsol.lo pccardc.lo pcc
ardd.lo usbd.lo usbdevs.lo tunefs.lo -ll -ledit -lutil -lkvm -lmd -lcrypt -lftpi
o -lz -lnetgraph -ldialog -lncurses -lmytinfo -ldisk -lipx
ppp.lo: In function `MakeKey':
ppp.lo(.text+0xfe): undefined reference to `des_set_odd_parity'
ppp.lo: In function `DesEncrypt':
ppp.lo(.text+0x142): undefined reference to `des_set_key'
ppp.lo(.text+0x14f): undefined reference to `des_ecb_encrypt'
ppp.lo: In function `MPPEKeyChange':
ppp.lo(.text+0x7b6): undefined reference to `RC4_set_key'
ppp.lo(.text+0x7ca): undefined reference to `RC4'
ppp.lo: In function `MPPEOutput':
ppp.lo(.text+0x85b): undefined reference to `RC4_set_key'
ppp.lo(.text+0x898): undefined reference to `RC4'
ppp.lo(.text+0x8bd): undefined reference to `RC4'
ppp.lo: In function `MPPEInput':
ppp.lo(.text+0x9f7): undefined reference to `RC4_set_key'
ppp.lo(.text+0xa18): undefined reference to `RC4'
ppp.lo(.text+0xa4e): undefined reference to `RC4'
*** Error code 1
Stop in /usr/src/release/boot_crunch.
*** Error code 1
...
###
--
Ruslan Ermilov Oracle Developer/DBA,
[EMAIL PROTECTED] Sunbay Software AG,
[EMAIL PROTECTED] FreeBSD committer,
+380.652.512.251Simferopol, Ukraine
http://www.FreeBSD.org The Power To Serve
http://www.oracle.com Enabling The Information Age
Index: Makefile
===
RCS file: /home/ncvs/src/usr.sbin/ppp/Makefile,v
retrieving revision 1.84
diff -u -p -r1.84 Makefile
--- Makefile2001/03/26 14:41:22 1.84
+++ Makefile2001/03/30 07:16:55
@@ -92,13 +92,7 @@ DPADD+= ${LIBNETGRAPH}
.endif
.if defined(RELEASE_CRUNCH)
-# We must create these objects because crunchgen will link them,
-# and we don't want any unused symbols to spoil the final link.
CFLAGS+=-DNONAT -DNORADIUS -DNOI4B -DNOSUID
-OBJS+= chap_ms.o mppe.o id.o nat_cmd.o radius.o
-chap_ms.o mppe.o id.o nat_cmd.o radius.o:
- null_${.PREFIX}.c
- cc -c -o ${.TARGET} null_${.PREFIX}.c
.endif
.include bsd.prog.mk
