the "pmap_page_protect" and "vm wire count" messages are indications that
you're triggering the work-around for a particular bug in the virtual-memory
subsystem. these messages are harmless. (I'm working on a real fix for
this bug, so the work-arounds will go away relatively soon.)
the bug is triggered by gdb in your case.
but after this point, it appears that something is corrupted in memory,
since subsequent php runs fail with the console messages indicating
that an illegal instruction was executed. hmm, maybe the work-around
is leaving a gdb breakpoint set in the real php page instead of in
a private copy. that would explain the symptoms. this is the first
I've seen of a problem like that, though.
a reboot should cure this symptom, but a less drastic measure that
might work would be to attempt to unmount the filesystem where the
php binary lives. even if the unmount fails (which it probably will
if it's eg. /usr), the corrupted data would be flushed from the cache.
is this problem reproducible? if so, I'll see if I can trigger it on
my machine this weekend.
-Chuck
On Sun, Aug 19, 2001 at 10:27:33PM -0700, Caffeinate The World wrote:
> # uname -a
> NetBSD ns01 1.5W NetBSD 1.5W (ALPHA-$Revision: 1.127.2.2 $) #5: Mon Jun
> 25 22:34:09 CDT 2001
> tom@ns01:/usr/src/sys/arch/alpha/compile/ns01 alpha
>
> php is from a snapshot on the snap.php.net
> php4-200108162235.tar.gz
>
> i was doing some scripting using cgi php on some very large data
> files.. php quit suddenly and said it ran out of allocated memory. i
> then tried running a simple command line "php --version" and it quit
> w/o further reports. then i ran gdb and ran php --version through gdb
> and it said that program quit with 0377.
>
> upon checking /var/log/messages
>
> Aug 20 00:05:26 ns01 /netbsd: pmap_page_protect: wired mapping for
> 677a000 not removed
> Aug 20 00:05:26 ns01 /netbsd: vm wire count 1
> Aug 20 00:05:26 ns01 /netbsd:
> Aug 20 00:05:26 ns01 /netbsd: CPU 0: fatal user trap:
> Aug 20 00:05:26 ns01 /netbsd:
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 trap entry = 0x3 (instruction
> fault)
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 a0 = 0x0
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 a1 = 0x120020260
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 a2 = 0x160214000
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 pc = 0x1602059c4
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 ra = 0x160204f64
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 curproc = 0xfffffc002747b710
> Aug 20 00:05:26 ns01 /netbsd: CPU 0 pid = 2377, comm = php
> Aug 20 00:05:26 ns01 /netbsd:
> Aug 20 00:05:30 ns01 /netbsd:
> Aug 20 00:05:30 ns01 /netbsd: CPU 0: fatal user trap:
> Aug 20 00:05:30 ns01 /netbsd:
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 trap entry = 0x3 (instruction
> fault)
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 a0 = 0x0
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 a1 = 0x120020260
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 a2 = 0x160214000
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 pc = 0x160204f68
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 ra = 0x160204f64
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 curproc = 0xfffffc002747b710
> Aug 20 00:05:30 ns01 /netbsd: CPU 0 pid = 2377, comm = php
> Aug 20 00:05:30 ns01 /netbsd:
>
> is this a netbsd issue or php? i've been running this php cgi for a few
> days now and nothing has changed with the system setup. now when ever
> php cgi is executed it will just quit quietly right away.
>
> __________________________________________________
> Do You Yahoo!?
> Make international calls for as low as $.04/minute with Yahoo! Messenger
> http://phonecard.yahoo.com/
--
PHP Development Mailing List <http://www.php.net/>
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
