Re: PLEASE TEST: IPI deadlock avoidance patch

Willem Jan Withagen wrote:

>
>>>
>>> Signal 6 is SIGABRT, which is usually intentional. You'd have to 
>>> check the
>>> output for a specific process that abended.  I'd also have to scan the
>>> make code for any abort() calls.
>>>  
>>>
>> I have not given it much attention yet, since I'm bussy doing other 
>> things right now.
>> But on my dual AMD64 box compiling the current 5.3 Beta generated 
>> also 'Abort Trap' while doing a
>> buildworld -j 32. Compiling without -j worked fine, and now I'm at 
>> 5.3 B1, so I can test again.
>> It ws running a 5.3B1 kernel.
>>
>> I'll let you know if the problem persists
>
>
> The above was with last mondays kernel.
> This mornings 5.3B1 when running 'make -j 32 buildworld', did not 
> generate any errors

I'm sorry to report that I was to fast with reporting good news.....

make -j 64 buildworld did generate quite few errors on the second run.....

This is AMD-64 on a dual opteron system:
    config, dmesg, uname => 
htttp://withagen.dyndns.org/FreeBSD/cores/opteron.tyan.106.{config,dmesg,uname}

Things are like:
===> sbin/devfs
cc -pipe -g  -Wsystem-headers -Wall -Wno-format-y2k -W 
-Wstrict-prototypes -Wmis
sing-prototypes -Wpointer-arith -Wreturn-type -Wcast-qual 
-Wwrite-strings -Wswit
ch -Wshadow -Wcast-align -c /home2/src/sbin/devfs/rule.c
cc -pipe -g  -Wsystem-headers -Wall -Wno-format-y2k -W 
-Wstrict-prototypes -Wmis
sing-prototypes -Wpointer-arith -Wreturn-type -Wcast-qual 
-Wwrite-strings -Wswit
ch -Wshadow -Wcast-align -c /home2/src/sbin/devfs/devfs.c
gzip -cn /home2/src/sbin/devfs/devfs.8 > devfs.8.gz
*** Signal 6
cc -pipe -g    -o chat chat.o
*** Error code 134
===> usr.bin/checknr
cc -pipe -g  -Wsystem-headers -Wall -Wno-format-y2k -W 
-Wstrict-prototypes -Wmis
sing-prototypes -Wpointer-arith -Wreturn-type -Wcast-qual 
-Wwrite-strings -Wswit
ch -Wshadow -Wcast-align  -o devfs devfs.o rule.o
cc -pipe -g   -c /home2/src/usr.bin/checknr/checknr.c
gzip -cn /home2/src/usr.bin/checknr/checknr.1 > checknr.1.gz
===> sbin/dhclient
*** Error code 134
===> sbin/dhclient/common
===> bin/expr
*** Error code 134

Not shure who writes these Error code 134?? If it is make, then it 
should perhaps in the '-j X' version have the executed command (and 
working directory) appended. Gives a better indication where things go 
haywire.
But in the current output there is no telling who aborted what.

--WjW