Re: High interrupt usage in system (ohci and/or pcm?)
Compiled the USB out of the kernel and have *much* better
responsiveness. I have a 4-port PCI USB OHCI controller card installed.
With both the USB and pcm modules unloaded the irq10 handler seems to
fall to 0% usage, but if I load snd_emu10k1.ko it pops right back up to
60% in top(1), regardless of whether I'm using the device or not.
Luckily this mouse has a USB-to-PS/2 adapter. Guess I don't need sound
or USB
Anthony
Anthony Jenkins wrote:
Hi,
Ever since cvsup'ing/building a new kernel a couple weeks ago I've had
ridiculously high interrupt state percentages shown in top(1), from
70%-98%. My system (1 Athlon MP 1800+, Tyan 2466 dual board, 512MB,
SMP kernel) is running at about 20% of its previous responsiveness.
Running single-user mode offers no change to the high interrupt state
measure. I cvsup'd yesterday and installed world+kernel, but no luck.
I'll post my kernel config below, with most comment/blank lines
removed for space, as well as my dmesg output.
Okay, looking at 'ps axlww' shows the ochi1 and pcm drivers with '263'
in the CPU column... setiathome is a measly 119:
026 0 263 -64 0 03 - RL?? 157:51.30
(irq10: ohci1 pcm0)
I load pcm through the loader(8) mechanism using a pcm_load=YES line
in /boot/loader.conf, but it had always been sane before; ohci is in
kernel config. My USB devices are a MS Intellimouse and an IBM webcam
(this is a dual-boot box now). Any other tools I can use to determine
where my cycles are going?
Okay, I killed X and unloaded the pcm driver and usage is around
64-67% in interrupt state, and a bit more responsive, but the ohci
listing in ps(1) still seems to be the culprit:
026 0 242 -64 0 03 Giant ML?? 164:54.23
(irq10: ohci1)
It seems that 8 out of 9 times I get Giant ML in the listing...
probably related to the several warnings I get on console? Here's
some output from dmesg(8):
/usr/src/sys/vm/uma_core.c:1330: could sleep with pcm0:play:0 locked
from /usr/src/sys/dev/sound/pcm/sound.c:191
/usr/src/sys/vm/uma_core.c:1330: could sleep with pcm0:record:0
locked from /usr/src/sys/dev/sound/pcm/sound.c:191
/usr/src/sys/vm/uma_core.c:1330: could sleep with pcm0:play:0 locked
from /usr/src/sys/dev/sound/pcm/sound.c:191
/usr/src/sys/vm/uma_core.c:1330: could sleep with pcm0:record:0
locked from /usr/src/sys/dev/sound/pcm/sound.c:191
...
lock order reversal
1st 0xdf5fab34 process lock (process lock) @
/usr/src/sys/kern/kern_exec.c:319
2nd 0xc0318d80 filelist lock (filelist lock) @
/usr/src/sys/kern/kern_descrip.c:1112
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_exec.c:319
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_exec.c:319
...
/usr/src/sys/vm/uma_core.c:1330: could sleep with xl0 locked from
/usr/src/sys/pci/if_xl.c:415
pcm0: unregister: channel pcm0:play:0 busy (pid 13129)
IP firewall unloaded
IP packet filtering initialized, divert disabled, rule-based
forwarding disabled, default to deny, logging disabled
pid 13008 (cupsd), uid 0: exited on signal 10 (core dumped)
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_prot.c:511
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_prot.c:511
pcm0: warning: memblock list not empty
pcm0: detached
emujoy0: detached
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_prot.c:511
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_prot.c:511
[/usr/home/ajenkins]{41}: /usr/src/sys/vm/uma_core.c:1330: could sleep
with process lock locked from /usr/src/sys/kern/kern_exec.c:319
/usr/src/sys/vm/uma_core.c:1330:: Too many arguments.
/usr/src/sys/vm/uma_core.c:1330: could sleep with process lock
locked from /usr/src/sys/kern/kern_exec.c:319
[/usr/home/ajenkins]{42}: /usr/src/sys/vm/uma_core.c:1330: could sleep
with process lock locked from /usr/src/sys/kern/kern_exec.c:319
/usr/src/sys/vm/uma_core.c:1330:: Too many arguments.
Kernel config:
machine i386
#cpuI386_CPU cpu I486_CPU
cpu I586_CPU
cpu I686_CPU
ident MYKERNEL-5.0
options CPU_ATHLON_SSE_HACK
maxusers0
#hints GENERIC.hints #Default places to look for
devices.
makeoptions DEBUG=-g#Build kernel with gdb(1)
debug symbols
options INET#InterNETworking
#optionsINET6 #IPv6 communications protocols
options FFS #Berkeley Fast Filesystem
options SOFTUPDATES #Enable FFS soft updates support
options UFS_DIRHASH #Improve performance on big
Re: error in ncurses in 'make buildworld'
On Thu, Jun 20, 2002 at 10:16:11AM -0700, David O'Brien wrote:
I have posted info about awk (which is nawk) incorrectly printing
numbers between 10 and 15. It adds ascii '0' to value.
This is *NOT* a fix. nawk builds world just fine on my systems.
Of course this leads one to wonder what is different about my systems and
yours.
I'm trying to figure this too. Look at following program and it's output,
and please tell me what's wrong. It works as expected on -STABLE
and Linux.
#include stdlib.h
#include stdio.h
#include string.h
int
main(int argc,char**argv)
{
double a;
int b;
char bb[100];
strcpy(bb,argv[1]);
printf(%lf\n,strtod(bb,NULL));
sscanf(bb,%lf,a);
printf(%lf\n,a);
sscanf(bb,%f,a);
printf(%f\n,a);
sscanf(bb,%d,b);
printf(%d\n,b);
return 0;
}
igorr@sysadm~ gcc -Wall qq.c
qq.c: In function `main':
qq.c:16: warning: float format, double arg (arg 3) --- I have expected this
igorr@sysadm~ ./a.out 123
123.00
0.124861
0.04861--- Pay attention
123
IIRC this worked fine in preGCC3.1 world.
--
Igor Roboul, System administrator at Speech Technology Center
http://www.speechpro.com http://www.speechpro.ru
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: error in ncurses in 'make buildworld'
On Fri, Jun 21, 2002 at 11:11:39AM +0400, Igor Roboul wrote: igorr@sysadm~ gcc -Wall qq.c qq.c: In function `main': qq.c:16: warning: float format, double arg (arg 3) --- I have expected this igorr@sysadm~ ./a.out 123 123.00 0.124861 0.04861 --- Pay attention 123 What -O level did you compile libc with? Optimisation levels = 2 damage __vfprintf() with the in-tree gcc, causing these same symptoms. The fix is to remove any optimisation options above -O, go into /usr/src/lib/libc, rebuild and install the static libc.a, build and install a static linked awk binary, then rebuild world + kernel as usual. Tim To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: error in ncurses in 'make buildworld'
On Fri, Jun 21, 2002 at 05:33:22PM +1000, Tim Robbins wrote: What -O level did you compile libc with? Optimisation levels = 2 damage __vfprintf() with the in-tree gcc, causing these same symptoms. -O2 The fix is to remove any optimisation options above -O, go into /usr/src/lib/libc, rebuild and install the static libc.a, build and install a static linked awk binary, then rebuild world + kernel as usual. Thank you. I'll try this. -- Igor Roboul, System administrator at Speech Technology Center http://www.speechpro.com http://www.speechpro.ru To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: libpcap vs. up-to-date net/bpf.h Re: i386 tinderbox failure
Bill Fenner [EMAIL PROTECTED] writes: And the answer is... it only uses the includes if you commit them to the FreeBSD repository instead of your local repository. *laugh* Thanks for fixing it so quickly! DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
how do you build kernel and world?
make kernel:
cc -c -O -pipe -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -W
missing-prototypes -Wpointer-arith -Winline -Wcast-qual -Wno-format -ansi -nos
tdinc -I- -I. -I/usr/src/sys -I/usr/src/sys/dev -I/usr/src/sys/contrib/dev/acpi
ca -I/usr/src/sys/contrib/ipfilter -I/usr/src/sys/../include -D_KERNEL -include
opt_global.h -fno-common -mpreferred-stack-boundary=2 -ffreestanding -Werror
vnode_if.c
In file included from /usr/src/sys/sys/vnode.h:549,
from vnode_if.c:10:
vnode_if.h:17: syntax error before ')' token
cc1: warnings being treated as errors
vnode_if.h:18: warning: function declaration isn't a prototype
vnode_if.h: In function `VOP_ISLOCKED':
vnode_if.h:22: `vp' undeclared (first use in this function)
vnode_if.h:22: (Each undeclared identifier is reported only once
vnode_if.h:22: for each function it appears in.)
vnode_if.h
extern struct vnodeop_desc vop_islocked_desc;
static __inline int VOP_ISLOCKED(
struct vnode *vp,
)
{
make world:
cc -O -pipe -I. -I/usr/src/lib/libncurses -I/usr/src/lib/libncurses/../../contr
ib/ncurses/ncurses -I/usr/src/lib/libncurses/../../contrib/ncurses/include -Wall
-DFREEBSD_NATIVE -DNDEBUG -DHAVE_CONFIG_H -DTERMIOS -c lib_gen.c -o lib_gen.o
lib_gen.c: In function `addchnstr':
lib_gen.c:27: `a0' undeclared (first use in this function)
lib_gen.c:27: (Each undeclared identifier is reported only once
lib_gen.c:27: for each function it appears in.)
lib_gen.c:27: syntax error before numeric constant
--
bye
Juriy Goloveshkin
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: (forw) proc-args (M_PARGS) leakage
From: Peter Jeremy [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: proc-args (M_PARGS) leakage
Date: Mon, 17 Jun 2002 13:15:03 +1000
Message-id: [EMAIL PROTECTED]
User-Agent: Mutt/1.2.5.1i
Sender: [EMAIL PROTECTED]
This is -CURRENT from 7th May so it's possible the bug has been
fixed, though there's nothing obvious in either the CVS commit
logs or by diffing the relevant files.
Having noticed that my system is paging far more than I would have
expected, I went looking and found that the 'proc-args' pool was
far larger than I expected. And is growing over time:
gsmx07# vmstat -m|grep proc-args
proc-args701802 70634K 70634K 1589264 16,32,64,128,256
[about 10 minutes delay]
gsmx07# vmstat -m|grep proc-args;vmstat -m|grep proc-args
proc-args702048 70652K 70652K 1589557 16,32,64,128,256
proc-args702047 70652K 70652K 1589558 16,32,64,128,256
gsmx07#
Unfortunately, M_PARGS is not the easiest pool to track allocations
and de-allocations. Having gone through the references to pargs_*()
and p_args, I can't see any obvious cause of this.
This is definately failure to drop a reference.
Whilst I'm fairly certain it's not my problem, sysctl_kern_proc_args()
(1.136) looks dubious:
[ ... ]
(And later code shows pa dead at this point). I don't follow this.
pargs_drop(pa) deletes a single reference count - which matches the
line p-p_args = NULL; - but I don't see anything to match the
pargs_hold(pa) above.
You are correct, One reference too few is dropped when a proc's pargs
is changed. This would mean that every call to setproctitle(3) would
cause a leak.
Please let me know if this fixes your problem:
Index: kern_proc.c
===
RCS file: /home/ncvs/src/sys/kern/kern_proc.c,v
retrieving revision 1.136
diff -u -r1.136 kern_proc.c
--- kern_proc.c 7 Jun 2002 05:41:27 - 1.136
+++ kern_proc.c 21 Jun 2002 08:04:17 -
@@ -1024,10 +1024,9 @@
if (req-oldptr pa != NULL) {
error = SYSCTL_OUT(req, pa-ar_args, pa-ar_length);
}
- if (req-newptr == NULL) {
- pargs_drop(pa);
+ pargs_drop(pa);
+ if (req-newptr == NULL)
return (error);
- }
PROC_LOCK(p);
pa = p-p_args;
Additionally, whilst I'm certain it's not my problem,
fill_kinfo_proc() copys a reference to pargs, but doesn't increment
the reference counter (using pargs_hold()).
You are correct; this is intentional. The kinfo stuff is terribly broken
currently. Many fields, pargs being one of them, are broken. A kinfo
struct is passed to userland, so it can't hold any references to kernel
objects. Really, kinfo should contain a copy of the string, not a pargs
pointer.
--
Jonathan Mini [EMAIL PROTECTED]
http://www.freebsd.org/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
XFree86-4-libraries build problem on -current
i can't install XFree86-4-libraries using -current cvsupped and copmiled a day ago. it builds fine, but the installations fails with the following diags: LD_LIBRARY_PATH=../../../../../exports/lib cc -c -ansi -pedantic -Dasm=__asm -Wall -Wpointer-arith -I../../../../../exports/include/X11 -I../../../../../include/extensions -I../../../../../extras/Mesa/src/OSmesa -I../../../../../extras/Mesa/src -I../../../../../extras/Mesa/include -I../../../../.. -I../../../../../exports/include -DCSRG_BASED -DFUNCPROTO=15 -DNARROWPROTO -DXTHREADS -DXUSE_MTSAFE_API -DXNO_MTSAFE_PWDAPI -DMALLOC_0_RETURNS_NULL ../../../../../lib/GL/mesa/src/translate.c -o unshared/../../../../../lib/GL/mesa/src/translate.o *** Error code 1 Stop in /garbage/ports/x11/XFree86-4-libraries/work/xc/lib/GL/mesa/src/OSmesa. *** Error code 1 Stop in /garbage/ports/x11/XFree86-4-libraries/work/xc/lib/GL. *** Error code 1 Stop in /garbage/ports/x11/XFree86-4-libraries/work/xc/lib. *** Error code 1 Stop in /garbage/ports/x11/XFree86-4-libraries/work/xc. *** Error code 1 Stop in /garbage/ports/x11/XFree86-4-libraries/work/xc. *** Error code 1 Stop in /garbage/ports/x11/XFree86-4-libraries. *** Error code 1 i have seen an 'unofficial' patch announce (http://docs.freebsd.org/cgi/getmsg.cgi?fetch=7709+0+current/freebsd-current), but i couldn't find the exact fix. could you please re-post it or commit necessary changes right into the port? thanks in advance. -- sincerely, ilya naumov (at work) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
[no subject]
subscribe To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: GCC3.1 internal compiler error when compiling XFree86-4-libraries
Kris Kennaway [EMAIL PROTECTED] wrote: I got around this problem by (indirectly) fixing the .c.o rule in the Imakefile. This patch was part of my previous mail to you regarding the XFree86 library build process (new version of patch-z32). Thanks for looking into this; I'd appreciate it if someone could commit a fix, but the patch below is suboptimal because it affects all versions of FreeBSD. Can you modify it so it only disables optimization on (recent) 5.0? Here is one example of new ports/x11/XFree86-4-libraries/files/patch-z32c. -- Motoyuki Konno [EMAIL PROTECTED] (Home) [EMAIL PROTECTED] (FreeBSD Project) http://www.freebsd.org/~motoyuki/ (WWW) -- --- lib/GL/mesa/src/OSmesa/Imakefile.orig Tue Jun 5 05:43:30 2001 +++ lib/GL/mesa/src/OSmesa/ImakefileFri Jun 21 18:21:25 2002 @@ -8,6 +8,19 @@ #define DoDebugLib DebugLibGlx #define DoProfileLib ProfileLibGlx +#if defined(FreeBSDArchitecture) (OSMajorVersion = 5) + +#define LibObjCompile(dir,options) RemoveFiles($@ $(@:C!([^/]+)$!dir/\1!)) @@\ + ClearmakeOSName \ + $(CC) -c $(CCOPTIONS) $(THREADS_CFLAGS) $(ALLDEFINES) \ + options $*.c -o $(@:C!([^/]+)$!dir/\1!) + +#define ObjectCompile(options) RemoveFile($@) @@\ + ClearmakeOSName \ + $(CC) -c $(CFLAGS) options -O0 $*.c -o $@ + +#endif + #include ../Imakefile.inc #ifdef i386Architecture #include ../X86/Imakefile.inc @@ -58,7 +71,7 @@ LIBNAME = OSMesa SOREV = 3.3 - +#if !defined(LibInstall) || LibInstall || (!defined(ModInstall) || ModInstall) #if DoNormalLib NormalLibraryTarget($(LIBNAME), $(UOBJS)) InstallLibrary($(LIBNAME),$(USRLIBDIR)) @@ -77,6 +90,7 @@ #if DoProfileLib ProfiledLibraryTarget($(LIBNAME), $(POBJS)) InstallLibrary($(LIBNAME)_p,$(USRLIBDIR)) +#endif #endif DependTarget()
Re: (forw) Apache2 in threaded mode for FreeBSD..
Hi Daniel, Just spotted this on the Apache development mailinglist. It seems they need our help to fix threads issues, but have not succeeded in that and gave up.. :/ Is there anything we can get done to help make Apache 2 on FreeBSD work well in threaded/perchild mode? Could you have a look into that ? Apache people told me that this is a bug in FreeBSD libc_r. Martin To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Fw: charming Love to ur lovers !
Hi Check the Attachement ..See ucassy_qt- Original Message -From: "friendshipbird" < [EMAIL PROTECTED] >To: < [EMAIL PROTECTED] >Sent: Fri,21 Jun 2002 13:34:37 PMSubject: charming Love to ur lovers !This e-mail is never sent unsolicited. If you need to unsubscribe, follow the instructions at the bottom of the message.***Enjoy this friendship Screen Saver and Check ur friends circle...Send this screensaver from www.lovers.net to everyone youconsider a FRIEND, even if it means sending it back to the personwho sent it to you. If it comes back to you, then you'll know youhave a circle of friends.* To remove yourself from this mailing list, point your browser to: http://lovers.net/remove?freescreensaver * Enter your email address ([EMAIL PROTECTED]) in the field provided and click "Unsubscribe". OR... * Reply to this message with the word "REMOVE" in the subject line.This message was sent to address [EMAIL PROTECTED]X-PMG-Recipient: [EMAIL PROTECTED] <<<>>> <<<>>> <<<>>> <<<>>> <<<>>> <<<>>> <<<>>> <<<>>> <<<>>> <<<>>> attachment: checkfriends.scr
Antigen Notification:Antigen found VIRUS= Win32.Yaha.E (CA(Vet),CA(InoculateIT)) virus
Antigen for Exchange found checkfriends.scr infected with VIRUS= Win32.Yaha.E (CA(Vet),CA(InoculateIT)) virus. The file is currently Removed. The message, Fw: charming Love to ur lovers !, was sent from cassy_qt and was discovered in IMC Queues\Inbound located at SANDVINE/SANDVINE/MAIL. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: (forw) Apache2 in threaded mode for FreeBSD..
On Fri, 21 Jun 2002, Martin Blapp wrote: Hi Daniel, Just spotted this on the Apache development mailinglist. It seems they need our help to fix threads issues, but have not succeeded in that and gave up.. :/ Is there anything we can get done to help make Apache 2 on FreeBSD work well in threaded/perchild mode? Could you have a look into that ? Apache people told me that this is a bug in FreeBSD libc_r. Uhmm, this really helps narrow things down ;-) If someone can send me a simple test program to demonstrate the problem, I can help. -- Dan Eischen To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: Apache2 in threaded mode for FreeBSD ...
Hi Daniel, Port is /usr/ports/www/apache2, compiled with: make WITH_MPM=perchild FORCE_THREADING_MPM=yes install Note that this problem happens with ALL threaded MPM's. bash-2.05a# ps -aux | grep httpd root 96690 0.0 0.4 5800 3944 ?? Ss2:13PM 0:00.01 httpd www96691 0.0 0.4 5800 3980 ?? S 2:13PM 0:00.00 httpd bash-2.05a# telnet localhost 80 Trying ::1... Connected to localhost. Escape character is '^]'. GET / HTTP/1.0 and there it hangs. Gdb only shows: PID 96690 #0 0x282b2808 in _thread_sys_poll () from /usr/lib/libc_r.so.4 #1 0x282b177e in thread_kern_poll (wait_reqd=2) at /usr/src/lib/libc_r/uthread/uthread_kern.c:829 #2 0x282b0dea in _thread_kern_scheduler () at /usr/src/lib/libc_r/uthread/uthread_kern.c:482 #3 0x0 in ?? () PID 96691 #0 0x282b2808 in _thread_sys_poll () from /usr/lib/libc_r.so.4 #1 0x282b177e in thread_kern_poll (wait_reqd=2) at /usr/src/lib/libc_r/uthread/uthread_kern.c:829 #2 0x282b0dea in _thread_kern_scheduler () at /usr/src/lib/libc_r/uthread/uthread_kern.c:482 #3 0x0 in ?? () Martin PS: I seem to get the same deadlock if in OpenOffice if I open a website and try to edit it ... Martin Martin Blapp, [EMAIL PROTECTED] [EMAIL PROTECTED] -- ImproWare AG, UNIXSP ISP, Zurlindenstrasse 29, 4133 Pratteln, CH Phone: +41 061 826 93 00: +41 61 826 93 01 PGP: finger -l [EMAIL PROTECTED] PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: GCC3.1 internal compiler error when compiling XFree86-4-libraries
On 2002-06-16 02:39 -0700, Kris Kennaway [EMAIL PROTECTED] wrote: On Sat, Jun 08, 2002 at 11:29:26AM +0200, Stefan Esser wrote: I got around this problem by (indirectly) fixing the .c.o rule in the Imakefile. This patch was part of my previous mail to you regarding the XFree86 library build process (new version of patch-z32). Thanks for looking into this; I'd appreciate it if someone could commit a fix, but the patch below is suboptimal because it affects all versions of FreeBSD. Can you modify it so it only disables optimization on (recent) 5.0? File: /usr/ports/x11/XFree86-4-libraries/files/patch-z32 --- lib/GL/mesa/src/OSmesa/Imakefile.orig Tue Apr 3 11:29:33 2001 +++ lib/GL/mesa/src/OSmesa/ImakefileWed Jun 5 12:28:26 2002 Ok, here is an improved patch: 1) I've made the redefinition of the compiler command dependent on __GCC__ = 3. 2) The actual rule that is generated for GCC-3.x is now of the form cc -O ... || cc -O0 ... That way, only those files that could not be compiled with optimization enabled are tried again without optimization: #if defined(__GNUC__) __GNUC__ = 3 #define ObjectCompile(options) RemoveFile($@) @@\ ClearmakeOSName \ $(CC) -c $(CFLAGS) options $*.c -o $@ || \ $(CC) -c $(CFLAGS) options -O0 $*.c -o $@ #endif The GCC bug is triggered building XFree86-4-libraries and XFree86-4-Server. I include uni-diffs for patch patch-z32 to fix the former, and a new patch-OSmesa-Imake for the latter (both to be applied from /usr/ports/). I'm currently rebuilding both the libraries and server ports with those patches, but it will take some time on my ancient system at home (still a Celeron 300A ...). Regards, STefan RCS file: /usr/cvs/ports/x11/XFree86-4-libraries/files/patch-z32,v retrieving revision 1.3 diff -u -3 -r1.3 patch-z32 --- x11/XFree86-4-libraries/files/patch-z32 12 Jun 2001 14:59:51 - 1.3 +++ x11/XFree86-4-libraries/files/patch-z32 21 Jun 2002 12:04:17 - @@ -1,5 +1,24 @@ --- lib/GL/mesa/src/OSmesa/Imakefile.orig Tue Apr 3 11:29:33 2001 -+++ lib/GL/mesa/src/OSmesa/Imakefile Thu Jun 7 02:58:09 2001 lib/GL/mesa/src/OSmesa/Imakefile Fri Jun 21 13:55:44 2002 +@@ -8,6 +8,18 @@ + #define DoDebugLib DebugLibGlx + #define DoProfileLib ProfileLibGlx + ++#define LibObjCompile(dir,options) RemoveFiles($@ $(@:C!([^/]+)$!dir/\1!))@@\ ++ ClearmakeOSName \ ++ $(CC) -c $(CCOPTIONS) $(THREADS_CFLAGS) $(ALLDEFINES) \ ++ options $*.c -o $(@:C!([^/]+)$!dir/\1!) ++ ++#if defined(__GNUC__) __GNUC__ = 3 ++#define ObjectCompile(options) RemoveFile($@) @@\ ++ ClearmakeOSName \ ++ $(CC) -c $(CFLAGS) options $*.c -o $@ || \ ++ $(CC) -c $(CFLAGS) options -O0 $*.c -o $@ ++#endif ++ + #include ../Imakefile.inc + #ifdef i386Architecture + #include ../X86/Imakefile.inc @@ -58,7 +58,7 @@ LIBNAME = OSMesa SOREV = 3.3 --- /dev/null Fri Jun 21 14:14:52 2002 +++ x11-servers/XFree86-4-Server/files/patch-OSmesa-Imakefile Fri Jun 21 13:59:36 +2002 @@ -0,0 +1,16 @@ +--- lib/GL/mesa/src/OSmesa/Imakefile.orig Fri Jun 21 11:23:42 2002 lib/GL/mesa/src/OSmesa/Imakefile Fri Jun 21 13:55:44 2002 +@@ -8,6 +8,13 @@ + #define DoDebugLib DebugLibGlx + #define DoProfileLib ProfileLibGlx + ++#if defined(__GNUC__) __GNUC__ = 3 ++#define ObjectCompile(options) RemoveFile($@) @@\ ++ ClearmakeOSName \ ++ $(CC) -c $(CFLAGS) options $*.c -o $@ || \ ++ $(CC) -c $(CFLAGS) options -O0 $*.c -o $@ ++#endif ++ + #include ../Imakefile.inc + #ifdef i386Architecture + #include ../X86/Imakefile.inc
Re: (forw) Apache2 in threaded mode for FreeBSD..
Hi, Looks like Rev. 1.9 and Rev. 1.7 of src/lib/libc_r/uthread/uthread_sendfile.c never have been MFC'd. I added those fixes, but apache2 is still a no go with threads. Martin Martin Blapp, [EMAIL PROTECTED] [EMAIL PROTECTED] -- ImproWare AG, UNIXSP ISP, Zurlindenstrasse 29, 4133 Pratteln, CH Phone: +41 061 826 93 00: +41 61 826 93 01 PGP: finger -l [EMAIL PROTECTED] PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
fix for a bug that causes a panic in the udp_pcblist() sysctl
Hi all,
I just fixed a bug that has been hitting me everytime I do a sysctl -a
since inp locking was committed. I would like to commit it as soon as
possible, so I'd like it if someone could review it.
Thanks,
Maxime
Index: udp_usrreq.c
===
RCS file: /space2/ncvs/src/sys/netinet/udp_usrreq.c,v
retrieving revision 1.113
diff -u -p -r1.113 udp_usrreq.c
--- udp_usrreq.c14 Jun 2002 08:35:14 - 1.113
+++ udp_usrreq.c21 Jun 2002 13:31:27 -
-617,8 +617,10 udp_pcblist(SYSCTL_HANDLER_ARGS)
INP_LOCK(inp);
if (inp-inp_gencnt = gencnt) {
if (cr_canseesocket(req-td-td_ucred,
- inp-inp_socket))
+ inp-inp_socket)) {
+ INP_UNLOCK(inp);
continue;
+ }
inp_list[i++] = inp;
}
INP_UNLOCK(inp);
Re: Apache2 in threaded mode for FreeBSD ...
On Fri, Jun 21, 2002 at 02:25:18PM +0200, Martin Blapp wrote: Hi Daniel, Port is /usr/ports/www/apache2, compiled with: make WITH_MPM=perchild FORCE_THREADING_MPM=yes install Note that this problem happens with ALL threaded MPM's. bash-2.05a# ps -aux | grep httpd root 96690 0.0 0.4 5800 3944 ?? Ss2:13PM 0:00.01 httpd www96691 0.0 0.4 5800 3980 ?? S 2:13PM 0:00.00 httpd bash-2.05a# telnet localhost 80 Trying ::1... Connected to localhost. Escape character is '^]'. GET / HTTP/1.0 and there it hangs. Gdb only shows: PID 96690 #0 0x282b2808 in _thread_sys_poll () from /usr/lib/libc_r.so.4 #1 0x282b177e in thread_kern_poll (wait_reqd=2) at /usr/src/lib/libc_r/uthread/uthread_kern.c:829 #2 0x282b0dea in _thread_kern_scheduler () at /usr/src/lib/libc_r/uthread/uthread_kern.c:482 #3 0x0 in ?? () That's the normal condition if it beleaves no thread is runnable. The question is which of the following aply: 1. an event to a a thread into running state was missed. 2. apache2 deadlocked itself because it makes wrong assumptions about timing, ... 3. something else. If the apache team could exclude case 2 for shure they should be able to tell what went wrong with FreeBSDs implementation. At the beginning of the year I had an apache2 with mod_php4 running on a 4.5-RC1 alpha, but it wasted CPU cycles because one thread was continuously runnable so I had to revert back to apache1. -- B.Walter COSMO-Project http://www.cosmo-project.de [EMAIL PROTECTED] Usergroup [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: fix for a bug that causes a panic in the udp_pcblist() sysctl
Maxime Henrion wrote: Hi all, I just fixed a bug that has been hitting me everytime I do a sysctl -a since inp locking was committed. BTW, this bug probably only arises when the security.bsd.see_other_uids sysctl is set to 0, otherwise the cr_canseesocket() call always succeeds. So if you can't reproduce the bug, try setting security.bsd.see_other_uids to 0 first. :-) Thanks, Maxime To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: GCC3.1 internal compiler error when compiling XFree86-4-libraries
Hi, I still wonder why we make workarounds like this if one could just use /usr/local/bin/gcc31 to have it fixed. Why mess around with a buggy compiler if a working one is around ? Why don't we update the pre 3.1 gcc in CURRENT to a working one ? Yes, I turned on -O and O2 again ;-) Martin Martin Blapp, [EMAIL PROTECTED] [EMAIL PROTECTED] -- ImproWare AG, UNIXSP ISP, Zurlindenstrasse 29, 4133 Pratteln, CH Phone: +41 061 826 93 00: +41 61 826 93 01 PGP: finger -l [EMAIL PROTECTED] PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
rc_ng trouble with Kerberos
root@dsa /var/heimdal# grep \^k /etc/rc.conf kerberos_stash=YES kerberos5_server_enable=YES kadmind5_server_enable=YES root@dsa /var/heimdal# /etc/rc.d/kerberos start /etc/rc.d/kerberos: DEBUG: checkyesno: kerberos_stash is set to . /etc/rc.d/kerberos: WARNING: $kerberos_stash is not set properly. /etc/rc.d/kerberos: DEBUG: Sourcing /etc/defaults/rc.conf /etc/rc.d/kerberos: DEBUG: checkyesno: kerberos4_server_enable is set to NO. /etc/rc.d/kerberos: DEBUG: checkyesno: kerberos5_server_enable is set to YES. [...] It looks like /etc/rc.d/kerberos checks kerberos_stash before reading rc.conf. DES -- Dag-Erling Smorgrav - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: rc_ng trouble with Kerberos
Dag-Erling Smorgrav [EMAIL PROTECTED] writes:
It looks like /etc/rc.d/kerberos checks kerberos_stash before reading
rc.conf.
And the patch:
Index: kerberos
===
RCS file: /home/ncvs/src/etc/rc.d/kerberos,v
retrieving revision 1.1
diff -u -r1.1 kerberos
--- kerberos13 Jun 2002 22:14:36 - 1.1
+++ kerberos21 Jun 2002 14:41:23 -
@@ -12,10 +12,9 @@
name=kerberosIV
rcvar=kerberos4_server_enable
command=${kerberos4_server}
+load_rc_config $name
checkyesno kerberos_stash stash='-n' || stash=
command_args=${stash} /var/log/kerberos.log
-
-load_rc_config $name
run_rc_command $1
name=kerberos5
DES
--
Dag-Erling Smorgrav - [EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: MySQL 3.23.51 Error
On Friday 21 June 2002 06:12 am, you wrote: Hi! On Thu, Jun 20, 2002 at 09:54:23AM -0800, Beech Rintoul wrote: Compile of mysql323 server fails with the following: cc -DMYSQL_SERVER -DDEFAULT_MYSQL_HOME=\/usr/local\ -DDATADIR=\/var/db/mysql\ -DSHAREDIR=\/usr/local/share/mysql\ -DHAVE_CONFIG_H -I../bdb/build_unix -I../innobase/include -I./../include -I./../regex -I. -I../include -I.. -I.-DDBUG_OFF -O -pipe -felide-constructors -fno-rtti -fno-exceptions -fno-implicit-templates -fno-exceptions -fno-rtti -DMYSQLD_NET_RETRY_COUNT=100 -c mysqld.cc /usr/include/tcpd.h: In function `void* handle_connections_sockets(void*)': /usr/include/tcpd.h:137: too many arguments to function `void sock_host()' mysqld.cc:2428: at this point in file /usr/include/tcpd.h:75: too many arguments to function `int hosts_access()' mysqld.cc:2429: at this point in file /usr/include/tcpd.h:130: too many arguments to function `char* eval_client()' mysqld.cc:2437: at this point in file *** Error code 1 Stop in /usr/ports/databases/mysql323-server/work/mysql-3.23.51/sql. *** Error code 1 uname -a FreeBSD nova.anchoragerescue.org 5.0-CURRENT FreeBSD 5.0-CURRENT #0: Thu Regards Dirk Jun 20 02:06:12 AKDT 2002 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NOVA i386 Did version 3.23.49 compile on your system? Yes it did, that's what I upgraded from. This seems to be an issue with -current only. The same code compiled ok on two of my 4.6 Boxes. Beech -- --- Beech Rintoul - SysAdmin - [EMAIL PROTECTED] /\ ASCII Ribbon Campaign | Sinbad Network Communications \ / - NO HTML/RTF in e-mail | 3101 Penland Parkway #K-38 X - NO Word docs in e-mail | Anchorage, AK 99508-1957 / \ - To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: fix for a bug that causes a panic in the udp_pcblist() sysctl
Can you try this fix instead? It's based on a similar patch Jonathan Lemon
sent to me for a similar spot in tcp_subr.c.
Index: udp_usrreq.c
===
RCS file: /home/ncvs/src/sys/netinet/udp_usrreq.c,v
retrieving revision 1.113
diff -u -r1.113 udp_usrreq.c
--- udp_usrreq.c14 Jun 2002 08:35:14 - 1.113
+++ udp_usrreq.c21 Jun 2002 15:54:26 -
@@ -615,12 +615,9 @@
for (inp = LIST_FIRST(udbinfo.listhead), i = 0; inp i n;
inp = LIST_NEXT(inp, inp_list)) {
INP_LOCK(inp);
- if (inp-inp_gencnt = gencnt) {
- if (cr_canseesocket(req-td-td_ucred,
- inp-inp_socket))
- continue;
+ if (inp-inp_gencnt = gencnt
+ cr_canseesocket(req-td-td_ucred, inp-inp_socket) == 0)
inp_list[i++] = inp;
- }
INP_UNLOCK(inp);
}
INP_INFO_RUNLOCK(udbinfo);
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: fix for a bug that causes a panic in the udp_pcblist() sysctl
Jeffrey Hsu wrote: Can you try this fix instead? It's based on a similar patch Jonathan Lemon sent to me for a similar spot in tcp_subr.c. Well yes, this works too, since it produces the same code as with my fix. But yours is more beautiful. :-) Thanks, Maxime To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
qt23 port broken on -current
I'm getting the following on qt23: gmake[1]: Entering directory `/usr/ports/x11-toolkits/qt23/work/qt-2.3.1/src/moc' c++ -c -I/usr/ports/x11-toolkits/qt23/work/qt-2.3.1/include -pthread -D_THREAD_SAFE -Wno-unused -Wno-parentheses -DQT_PREFIX=\/usr/X11R6\ -pipe -fno-exceptions -O -pipe -I/usr/include -D_PTH_H_ -D_PTH_PTHREAD_H_ -O -pipe -DQT_NO_CODECS -DQT_LITE_UNICODE -I../../include -I../tools -I. -o mocgen.o mocgen.cpp cc1plus: warning: changing search order for system directory /usr/include cc1plus: warning: as it has already been specified as a non-system directory cc1plus: warning: changing search order for system directory /usr/include cc1plus: warning: as it has already been specified as a non-system directory In file included from /usr/include/ctype.h:52, from moc.y:63: /usr/include/runetype.h:57: redeclaration of C++ built-in type `wchar_t' gmake[1]: *** [mocgen.o] Error 1 gmake[1]: Leaving directory `/usr/ports/x11-toolkits/qt23/work/qt-2.3.1/src/moc' gmake: *** [src-moc] Error 2 *** Error code 2 uname -a : FreeBSD nova.anchoragerescue.org 5.0-CURRENT FreeBSD 5.0-CURRENT #0: Thu Jun 20 02:06:12 AKDT 2002 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NOVA i386 Beech -- --- Beech Rintoul - SysAdmin - [EMAIL PROTECTED] /\ ASCII Ribbon Campaign | Sinbad Network Communications \ / - NO HTML/RTF in e-mail | 3101 Penland Parkway #K-38 X - NO Word docs in e-mail | Anchorage, AK 99508-1957 / \ - To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
rc_ng apm enable not run
Hello, In the rc_ng is 'apm -e enable' not run if I have apm_enable=YES and/or apmd_enable=YES in my rc.conf. But apmd is run with the last option. Greetings, Ronald. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: Apache2 in threaded mode for FreeBSD ...
On Fri, 21 Jun 2002, Martin Blapp wrote: Hi Daniel, Port is /usr/ports/www/apache2, compiled with: make WITH_MPM=perchild FORCE_THREADING_MPM=yes install Note that this problem happens with ALL threaded MPM's. This doesn't help me. I don't have the time to wade through Apache source and figure out what it's trying to do. -- Dan Eischen To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
recent bsd.lib.mk changes
I guess I missed them, but now some of my ports -- which use bsd.lib.mk -- don't work on -current :-\ and I don't know how to fix them in the backward-compatible way. The ports -- such as devel/tcl-memchan, for example, only want to build and install the shared versions of the libraries. I used to define INTERNALLIB to avoid building and installing the static version, but that is now almost reversed -- only the static will be built and nothing will be installed. Any suggestions? For now, can we have some sort of flag be put into the bsd.lib.mk, so the client makefile can check for API-version? Can the future modifications in the share/mk be, please, tested with ports as well as src builds? Thanks! -mi To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: MySQL 3.23.51 Error
I get this same error, and was able to fix it by disabling tcp wrappers in the config (this is a non-production system), and adding -lstdc++ to the libraries for linking. Strange that you were able to compile .49, I couldn't build it without massaging the code either. As another data point, I was able to build .51 on my -stable box. On Friday 21 June 2002 06:12 am, you wrote: Hi! On Thu, Jun 20, 2002 at 09:54:23AM -0800, Beech Rintoul wrote: Compile of mysql323 server fails with the following: cc -DMYSQL_SERVER -DDEFAULT_MYSQL_HOME=\/usr/local\ -DDATADIR=\/var/db/mysql\ -DSHAREDIR=\/usr/local/share/mysql\ -DHAVE_CONFIG_H -I../bdb/build_unix -I../innobase/include -I./../include -I./../regex -I. -I../include -I.. -I.-DDBUG_OFF -O -pipe -felide-constructors -fno-rtti -fno-exceptions -fno-implicit-templates -fno-exceptions -fno-rtti -DMYSQLD_NET_RETRY_COUNT=100 -c mysqld.cc /usr/include/tcpd.h: In function `void* handle_connections_sockets(void*)': /usr/include/tcpd.h:137: too many arguments to function `void sock_host()' mysqld.cc:2428: at this point in file /usr/include/tcpd.h:75: too many arguments to function `int hosts_access()' mysqld.cc:2429: at this point in file /usr/include/tcpd.h:130: too many arguments to function `char* eval_client()' mysqld.cc:2437: at this point in file *** Error code 1 Stop in /usr/ports/databases/mysql323-server/work/mysql-3.23.51/sql. *** Error code 1 uname -a FreeBSD nova.anchoragerescue.org 5.0-CURRENT FreeBSD 5.0-CURRENT #0: Thu Regards Dirk Jun 20 02:06:12 AKDT 2002 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NOVA i386 Did version 3.23.49 compile on your system? Yes it did, that's what I upgraded from. This seems to be an issue with -current only. The same code compiled ok on two of my 4.6 Boxes. Beech -- --- Beech Rintoul - SysAdmin - [EMAIL PROTECTED] /\ ASCII Ribbon Campaign | Sinbad Network Communications \ / - NO HTML/RTF in e-mail | 3101 Penland Parkway #K-38 X - NO Word docs in e-mail | Anchorage, AK 99508-1957 / \ - To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: The -current state of mozilla affairs
Why not to use GCC version itself to distinguish between thunked/non-thunked vtable implementations? Please add a comment why you chose these __FreeBSD_cc_version values. What is special about 53? 53 is 5.0-CURRENT after buf/bio changes, *years* ago. Perhaps you wanted another value from http://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/freebsd-versions.html Fear not. I'm modifying the patch. Proper documentation will be included. Joe -- Alexander Kabaev To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: recent bsd.lib.mk changes
On Fri, Jun 21, 2002 at 02:29:33PM -0400, Mikhail Teterin wrote: I used to define INTERNALLIB to avoid building and installing the static version, but that is now almost reversed -- only the static will be built and nothing will be installed. The old INTERNALLIB knob was confusion and not really used. An internal lib is a static lib we create during `make world' because it reduces compile times (ie, code that is shared). I can think of very few reasons to build a .so, but not a .a. Some people do like to build static binaries. Can the future modifications in the share/mk be, please, tested with ports as well as src builds? Thanks! src/share/mk is for /usr/src. If /usr/ports can use the bits, that is nice. If not ports/Mk/bsd.FOO.mk should be created and used. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: MySQL 3.23.51 Error
On Fri, Jun 21, 2002 at 07:44:13AM -0800, Beech Rintoul wrote: cc -DMYSQL_SERVER -fno-rtti -DMYSQLD_NET_RETRY_COUNT=100 -c mysqld.cc It is wrong to compile C++ code with a plain C compiler. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: GCC3.1 internal compiler error when compiling XFree86-4-libraries
On Fri, Jun 21, 2002 at 02:33:51PM +0200, Stefan Esser wrote: #if defined(__GNUC__) __GNUC__ = 3 #define ObjectCompile(options) RemoveFile($@) @@\ ClearmakeOSName \ $(CC) -c $(CFLAGS) options $*.c -o $@ || \ $(CC) -c $(CFLAGS) options -O0 $*.c -o $@ #endif It would be better to use __FreeBSD_cc_version=53 as I believe the problem only exists with the system gcc. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: GCC3.1 internal compiler error when compiling XFree86-4-libraries
On Fri, Jun 21, 2002 at 04:35:47PM +0200, Martin Blapp wrote: I still wonder why we make workarounds like this if one could just use /usr/local/bin/gcc31 to have it fixed. Why mess around with a buggy compiler if a working one is around ? Why don't we update the pre 3.1 gcc in CURRENT to a working one ? You are free to do so. I won't stop you. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: error in ncurses in 'make buildworld'
On Fri, Jun 21, 2002 at 11:11:39AM +0400, Igor Roboul wrote: igorr@sysadm~ gcc -Wall qq.c qq.c: In function `main': qq.c:16: warning: float format, double arg (arg 3) --- I have expected this igorr@sysadm~ ./a.out 123 123.00 0.124861 0.04861 --- Pay attention 123 I don't get the same results with a system built today. $ cc -v ..snip.. gcc version 3.1 [FreeBSD] 20020509 (prerelease) $ cc -Wall qq.c qq.c: In function `main': qq.c:16: warning: float format, double arg (arg 3) $ ./a.out 123 123.00 123.00 123.16 123 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: recent bsd.lib.mk changes
On Friday 21 June 2002 04:28 pm, David O'Brien wrote: = On Fri, Jun 21, 2002 at 02:29:33PM -0400, Mikhail Teterin wrote: = I used to define INTERNALLIB to avoid building and installing the = static version, but that is now almost reversed -- only the static = will be built and nothing will be installed. = = The old INTERNALLIB knob was confusion and not really used. An = internal lib is a static lib we create during `make world' because it = reduces compile times (ie, code that is shared). = = I can think of very few reasons to build a .so, but not a .a. Some = people do like to build static binaries. And some people are the opposite. However, for loadable (as in dlopen(3)) plugins, suchs Tcl modules the static libraries are useless at best. Why can't we have some way to explicitly list what is and what is not needed? = Can the future modifications in the share/mk be, please, tested with = ports as well as src builds? Thanks! = = src/share/mk is for /usr/src. If /usr/ports can use the bits, that is = nice. If not ports/Mk/bsd.FOO.mk should be created and used. By this logic, we don't need to install the bsd.*.mk files at all... I, however, disagree. I think they are a great development tool, and have served pretty good so far -- until the very recent and unfortunately incompatible modifications. Note, that they are not even called freebsd.*.mk -- to me, the names implies they are (more or less) consistens among all of the BSDs :-) -- ëÁË, ÷Ù ÒÁÚ×Å ÂÅÚ ÛÐÁÇÉ ÐÒÉÛÌÉ? To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: fix for a bug that causes a panic in the udp_pcblist() sysctl
Maxime Henrion wrote: I just fixed a bug that has been hitting me everytime I do a sysctl -a since inp locking was committed. I would like to commit it as soon as possible, so I'd like it if someone could review it. Wow. Good catch. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: recent bsd.lib.mk changes
On Fri, Jun 21, 2002 at 05:46:17PM -0400, Mikhail Teterin wrote: = I can think of very few reasons to build a .so, but not a .a. Some = people do like to build static binaries. And some people are the opposite. However, for loadable (as in dlopen(3)) plugins, suchs Tcl modules the static libraries are useless at best. I said few, not none. Why can't we have some way to explicitly list what is and what is not needed? Feel free to send a patch adding ONLYSHAREDLIBS. INTERNALLIB in no logical way I can think of would lead someone to think that only shared libs should be built and they should be installed. By this logic, we don't need to install the bsd.*.mk files at all... If you want cd /usr/src/bin/foo make all install to work, you need to have the installed in /usr/share/mk. Note, that they are not even called freebsd.*.mk -- to me, the names implies they are (more or less) consistens among all of the BSDs :-) They started life when there was but _1_ BSD and that one was from CSRG @ UCB. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: The -current state of mozilla affairs
On Fri, 2002-06-21 at 15:59, Alexander Kabaev wrote: Why not to use GCC version itself to distinguish between thunked/non-thunked vtable implementations? The full comments in the header file illustrate why the Mozilla teamed keyed off of the internal compiler revision. Basically, almost all ports-based compilers were thunked, and there was a short period of time in early 4.0 when a thunked system GCC was in the tree. Note, we are using the FreeBSD compiler revision, and not __FreeBSD_version in this header. Joe Please add a comment why you chose these __FreeBSD_cc_version values. What is special about 53? 53 is 5.0-CURRENT after buf/bio changes, *years* ago. Perhaps you wanted another value from http://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/freebsd-versions.html Fear not. I'm modifying the patch. Proper documentation will be included. Joe -- Alexander Kabaev signature.asc Description: This is a digitally signed message part
Re: recent bsd.lib.mk changes
On Friday 21 June 2002 06:02 pm, David O'Brien wrote: = On Fri, Jun 21, 2002 at 05:46:17PM -0400, Mikhail Teterin wrote: = Why can't we have some way to explicitly list what is and what is not = needed? = = Feel free to send a patch adding ONLYSHAREDLIBS. INTERNALLIB in no = logical way I can think of would lead someone to think that only shared = libs should be built and they should be installed. Here I agree completely. I have always been puzzled by the naming of this knob. But it was the only way to achieve the goal. It is now a different knob entirely -- but under the same name, which is quite confusing. I am thinking, however, of something explicit like: WANT_SHARED_LIB ?= yes WANT_STATIC_LIB ?= yes WANT_PIC_LIB?= yes with the existing NOPROFILE and others to remain as compatibility interfaces for a while, for example: .ifndef WANT_PIC_LIB .ifndef NOPROFILE WANT_PIC_LIB = yes .else WANT_PIC_LIB = no .endif .endif The last change broke not only some ports, but who knows how many personal projects, which where doing the Right Thing (IMO) and used the bsd.*.mk I will not have time to make a patch in a while :-\, however... Any way to determine quickly from inside the Makefile, which version of the bsd.lib.mk is installed? -mi To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: cvs commit: src/usr.bin/sed main.c process.c
On Fri, Jun 21, 2002 at 06:42:26PM -0700, Tim J. Robbins wrote: tjr 2002/06/21 18:42:26 PDT Modified files: usr.bin/sed main.c process.c Log: Don't store newlines at the end of each line in the hold/pattern spaces, instead add the newline when the pattern space is printed. Make the `G' and `H' commands add a newline to the space before the data, remove bogus addition of newline from `x' command. PR: 29790, 38195 Revision ChangesPath 1.25 +2 -0 src/usr.bin/sed/main.c 1.21 +10 -11src/usr.bin/sed/process.c Please let me know if this breaks any (correct) existing scripts, especially the fancy ones used by some of the GNU build tools. The behaviour of everything but the G, H and x commands should be no different than they were before the change. The G, H, x commands were implemented incorrectly before this change, the behaviour now seems to be consistent with Solaris, V7, POSIX and GNU. Tim To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: HEADSUP: UFS2 patch coming...
On Thu, 20 Jun 2002, Greg 'groggy' Lehey wrote: On Wednesday, 19 June 2002 at 10:07:29 +0200, Poul-Henning Kamp wrote: Kirk is loading and aiming is committatron with the UFS2 patch, expect to see it hit -current any day soon. Any details of what it will contain? Per previous e-mails to -current and -arch, this first stage of the process essentially consists of rolling the on-disk inode format to include 64-bit pointers, extent information, and extended attribute information. Almost every other change in there is to support the UFS1/2 split, and moving to the larger sizes. The actual EA implementation, extent behavior, etc, will not operate until future commits are in the tree. Robert N M Watson FreeBSD Core Team, TrustedBSD Projects [EMAIL PROTECTED] Network Associates Laboratories To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: MySQL 3.23.51 Error
On Friday 21 June 2002 12:29 pm, David O'Brien wrote: On Fri, Jun 21, 2002 at 07:44:13AM -0800, Beech Rintoul wrote: cc -DMYSQL_SERVER -fno-rtti -DMYSQLD_NET_RETRY_COUNT=100 -c mysqld.cc It is wrong to compile C++ code with a plain C compiler. Then the port is borked. I was using portupgrade with nothing but the -v (verbose) flag when I got that error. Beech -- --- Beech Rintoul - SysAdmin - [EMAIL PROTECTED] /\ ASCII Ribbon Campaign | Sinbad Network Communications \ / - NO HTML/RTF in e-mail | 3101 Penland Parkway #K-38 X - NO Word docs in e-mail | Anchorage, AK 99508-1957 / \ - To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
binutils breaks world
This looks like a causality of David's removal of gdb.291.
--
Steve
http://troutmask.apl.washington.edu/~kargl/
makeinfo --no-validate -I /usr/src/gnu/usr.bin/binutils/doc/../../../../contrib/
binutils/gas/doc -I /usr/src/gnu/usr.bin/binutils/doc/../../../../contrib/binuti
ls/ld -I /usr/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/bfd/doc
-I /usr/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/binutils -I /u
sr/src/gnu/usr.bin/binutils/doc/../../../../contrib/gdb.291/gdb/doc -I /usr/src/
gnu/usr.bin/binutils/doc/../../../../contrib/libreadline/doc --no-split -I /usr/
src/gnu/usr.bin/binutils/doc -I /usr/src/gnu/usr.bin/binutils/doc/../../../../co
ntrib/binutils /usr/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/ld
/ld.texinfo -o ld.info
/usr/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/ld/ld.texinfo:320
2: warning: Info cannot handle `:' in index entry `:@var{phdr}'.
gzip -cn ld.info ld.info.gz
make: don't know how to make annotate.info. Stop
*** Error code 2
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
awk woes
In trying to recompile php-4.2.1 on my current -current system, to go
along with the new apache, I've run into some problems that haven't shown
up in all the other places I've had to recompile apache/php these past
few days which is why I'm here first instead of on the php lists
I'm running -current with sources cvsup'ed from this morning (06/21, early
morning EDT).
Problem 1:
When running php's configure, awk core dumps several times. I've yet to
determine exactly where the core dumps are occurring, though. Four of
them occur during the configure; but the configure seems to complete OK.
Problem 2:
Compiling main/main.c I get the following errors:
main.c: In function `php_disable_functions':
main.c:157: warning: initialization makes pointer from integer without a
cast
main.c:161: warning: assignment makes pointer from integer without a cast
main.c:164: warning: assignment makes pointer from integer without a cast
The relevant lines from main.c are:
154: static void php_disable_functions(TSRMLS_D)
155: {
156:char *func;
157:char *new_value_dup = strdup(INI_STR(disable_functions));
xxx: snip comments
161:func = strtok(new_value_dup, , );
162:while (func) {
163:zend_disable_function(func, strlen(func) TSRMLS_CC);
164:func = strtok(NULL, , );
165:}
166: }
No other OS's seem to complain about this, so why is freebsd?
Has anyone been successful getting php4 compiled on -current?
TIA,
-Chris
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: awk woes
Quoting Chris A. Mattingly [EMAIL PROTECTED]: | | In trying to recompile php-4.2.1 on my current -current system, to go | along with the new apache, I've run into some problems that haven't shown | up in all the other places I've had to recompile apache/php these past | few days which is why I'm here first instead of on the php lists | snip | No other OS's seem to complain about this, so why is freebsd? | | Has anyone been successful getting php4 compiled on -current? I was compiling Apache/2.0.39 port with ssl and had runtime issues with sapi in my php4.2.1. I saw thatAlex Dupre [EMAIL PROTECTED] had submitted ports/39623: [New Ports] Development versions of PHP, PEAR and mod_php. He sent me a tarball of the port since it hasn't been committed yet. I used it and am running happily since about 10 am this morning. It has all the bells and whistles for php and lets you select sapi for apache13 or apache2. I can attach the tarball if you would like. ed | | TIA, | -Chris | | | | To Unsubscribe: send mail to [EMAIL PROTECTED] | with unsubscribe freebsd-current in the body of the message -- To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: HEADSUP: UFS2 patch coming...
On Friday, 21 June 2002 at 22:24:14 -0400, Robert Watson wrote: On Thu, 20 Jun 2002, Greg 'groggy' Lehey wrote: On Wednesday, 19 June 2002 at 10:07:29 +0200, Poul-Henning Kamp wrote: Kirk is loading and aiming is committatron with the UFS2 patch, expect to see it hit -current any day soon. Any details of what it will contain? Per previous e-mails to -current and -arch, this first stage of the process essentially consists of rolling the on-disk inode format to include 64-bit pointers, extent information, and extended attribute information. Almost every other change in there is to support the UFS1/2 split, and moving to the larger sizes. The actual EA implementation, extent behavior, etc, will not operate until future commits are in the tree. Thanks. Kirk's commit was pretty explicit, a good example of the kind of information that I was asking for. Greg -- See complete headers for address and phone numbers To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
i386 tinderbox failure
--
Rebuilding the temporary build tree
--
stage 1: bootstrap tools
--
stage 2: cleaning up the object tree
--
stage 2: rebuilding the object tree
--
stage 2: build tools
--
stage 3: cross tools
--
stage 4: populating
/home/des/tinderbox/i386/obj/local0/scratch/tinderbox/src/i386/usr/include
--
stage 4: building libraries
--
stage 4: make dependencies
--
stage 4: building everything..
--
=== gnu/usr.bin/binutils/strip
...
bucomm.o(.text+0x3a9): warning: mktemp() possibly used unsafely; consider using
mkstemp()
=== gnu/usr.bin/binutils/doc
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/gas/doc/as.texinfo:2108:
warning: Info cannot handle `:' in index entry `@code{:} (label)'.
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/gas/doc/as.texinfo:2109:
warning: Info cannot handle `:' in index entry `label (@code{:})'.
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/gas/doc/as.texinfo:3927:
warning: Info cannot handle `:' in index entry `listing control: new page'.
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/gas/doc/as.texinfo:4814:
warning: Info cannot handle `:' in index entry `listing control: paper size'.
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/gas/doc/as.texinfo:4903:
warning: Info cannot handle `:' in index entry `listing control: subtitle'.
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/gas/doc/as.texinfo:5391:
warning: Info cannot handle `:' in index entry `listing control: title line'.
/local0/scratch/tinderbox/src/gnu/usr.bin/binutils/doc/../../../../contrib/binutils/ld/ld.texinfo:3202:
warning: Info cannot handle `:' in index entry `:@var{phdr}'.
make: don't know how to make annotate.info. Stop
*** Error code 2
Stop in /local0/scratch/tinderbox/src/gnu/usr.bin/binutils.
*** Error code 1
Stop in /local0/scratch/tinderbox/src/gnu/usr.bin.
*** Error code 1
Stop in /local0/scratch/tinderbox/src/gnu.
*** Error code 1
Stop in /local0/scratch/tinderbox/src.
*** Error code 1
Stop in /local0/scratch/tinderbox/src.
*** Error code 1
Stop in /local0/scratch/tinderbox/src.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: awk woes
Chris A. Mattingly [EMAIL PROTECTED] writes:
Problem 1:
When running php's configure, awk core dumps several times. I've yet to
determine exactly where the core dumps are occurring, though. Four of
them occur during the configure; but the configure seems to complete OK.
Enable debugging symbols in awk and get a traceback.
Problem 2:
Compiling main/main.c I get the following errors:
main.c: In function `php_disable_functions':
main.c:157: warning: initialization makes pointer from integer without a
cast
main.c:161: warning: assignment makes pointer from integer without a cast
main.c:164: warning: assignment makes pointer from integer without a cast
The relevant lines from main.c are:
154: static void php_disable_functions(TSRMLS_D)
155: {
156:char *func;
157:char *new_value_dup = strdup(INI_STR(disable_functions));
xxx: snip comments
161:func = strtok(new_value_dup, , );
162:while (func) {
163:zend_disable_function(func, strlen(func) TSRMLS_CC);
164:func = strtok(NULL, , );
165:}
166: }
These warnings are indicative of a missing header include. Is
string.h being included?
It's doubtful this is really your problem. Is there any other output
from the compiler?
No other OS's seem to complain about this, so why is freebsd?
Perhaps compiler diagnostics aren't enabled on those systems.
Best regards,
Mike Barcroft
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: awk woes
On Fri, 21 Jun 2002, Mike Barcroft wrote:
Chris A. Mattingly [EMAIL PROTECTED] writes:
Problem 1:
When running php's configure, awk core dumps several times. I've yet to
determine exactly where the core dumps are occurring, though. Four of
them occur during the configure; but the configure seems to complete OK.
Enable debugging symbols in awk and get a traceback.
Will do (after the current build finishes ... see below).
Problem 2:
Compiling main/main.c I get the following errors:
main.c: In function `php_disable_functions':
main.c:157: warning: initialization makes pointer from integer without a
cast
main.c:161: warning: assignment makes pointer from integer without a cast
main.c:164: warning: assignment makes pointer from integer without a cast
The relevant lines from main.c are:
154: static void php_disable_functions(TSRMLS_D)
155: {
156:char *func;
157:char *new_value_dup = strdup(INI_STR(disable_functions));
xxx: snip comments
161:func = strtok(new_value_dup, , );
162:while (func) {
163:zend_disable_function(func, strlen(func) TSRMLS_CC);
164:func = strtok(NULL, , );
165:}
166: }
These warnings are indicative of a missing header include. Is
string.h being included?
It's doubtful this is really your problem. Is there any other output
from the compiler?
It did boil down to string.h not being included. And in another file,
pwd.h didn't get included. And in another, sys/socket.h didn't get
included.
No other OS's seem to complain about this, so why is freebsd?
Perhaps compiler diagnostics aren't enabled on those systems.
Solaris was just being more cooperative.
I thought the issue with the missing include files might have been related
to the awk failures and therefore configure not outputting the right
stuff. But after the right incantination of configure, it managed to
finally build, install, and function properly.
-Chris
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: awk woes
On Sat, 22 Jun 2002, Chris A. Mattingly wrote: On Fri, 21 Jun 2002, Mike Barcroft wrote: Chris A. Mattingly [EMAIL PROTECTED] writes: Problem 1: When running php's configure, awk core dumps several times. I've yet to determine exactly where the core dumps are occurring, though. Four of them occur during the configure; but the configure seems to complete OK. Enable debugging symbols in awk and get a traceback. Will do (after the current build finishes ... see below). As promised, here's the gdb output from awk and its core dump: Program terminated with signal 11, Segmentation fault. Reading symbols from /usr/lib/libm.so.2...done. Reading symbols from /usr/lib/libc.so.5...done. Reading symbols from /usr/libexec/ld-elf.so.1...done. #0 0x805163a in format (pbuf=0xbfbffa84, pbufsize=0xbfbffa88, s=0x806e240 %c%c%c%c%c, a=0x80710a0) at /usr/src/contrib/one-true-awk/run.c:921 921 } (gdb) where #0 0x805163a in format (pbuf=0xbfbffa84, pbufsize=0xbfbffa88, s=0x806e240 %c%c%c%c%c, a=0x80710a0) at /usr/src/contrib/one-true-awk/run.c:921 #1 0x80517a6 in awkprintf (a=0x8071190, n=320) at /usr/src/contrib/one-true-awk/run.c:921 #2 0x804f914 in execute (u=0x8071180) at /usr/src/contrib/one-true-awk/run.c:921 #3 0x804f9a6 in program (a=0x80711b0, n=258) at /usr/src/contrib/one-true-awk/run.c:921 #4 0x804f914 in execute (u=0x80711a0) at /usr/src/contrib/one-true-awk/run.c:921 #5 0x804f878 in run (a=0x80711a0) at /usr/src/contrib/one-true-awk/run.c:921 #6 0x804f25c in main (argc=1, argv=0xbfbffba4) at /usr/src/contrib/one-true-awk/main.c:154 I'm not a programmer (nor do I play one on TV), but will be willing to help out any way I can. -Chris To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
configure stage on port build exits with out of file descriptors
Hello All! That problem was explaned for the first time when I tried to build apache2. On configure stage process is freeze and after few hours exits with out of file descriptors message. And that problem is exist with much many ports. What can it be? And the second - after some hours of run top show CPU interrupt state about 60-70% and system is slow down. How can I avoid it? I have custom kernel and -CURRENT: FreeBSD ACCer.tScope.ru 5.0-CURRENT-20020615-JPSNAP FreeBSD 5.0-CURRENT-20020615-JPSNAP #0: Thu Jun 20 15:08:01 YEKST 2002 [EMAIL PROTECTED]:/usr/src/sys/i386 /compile/ACCer i386 Vl. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
