Re: armv7 targeting via poudriere-devel (armv7 via aarch64): net/dante build ends up with ./conftest stuck in write loop

On 2022-Jun-10, at 12:36, Mark Millard <marklmi@yahoo.com> wrote:

> The aarch64 context is Cortex-A72 based: a 8 GiByte RPi4B.
> The build is part of a "bulk -a -c" test that I've
> been doing.
> 
> top reports the builder activity looks like:
> 
>  PID   JID USERNAME    PRI NICE     SIZE       RES STATE    C   TIME     CPU COMMAND
> 21812    11 root        103    0   4740Ki    1200Ki CPU3     3 113:33  99.93% ./conftest
> 19726    11 root         52    0   5964Ki    1948Ki wait     0   0:01   0.00% /bin/sh ./configure --with-socks-conf=/usr/local/etc/socks.conf --with-sockd-conf=/usr/local/etc/sockd.conf --prefix=/usr
> 19327    11 root         52    0   5096Ki    1432Ki wait     1   0:00   0.00% /bin/sh -e -c (cd /wrkdirs/usr/ports/net/dante/work/dante-1.4.3 &&  _LATE_CONFIGURE_ARGS="" ;  if [ -z "" ] &&  ./configu
> 19217    11 root         52    0   6516Ki    2216Ki wait     0   0:00   0.00% /usr/bin/make -C /usr/ports/net/dante configure
> 
> # truss -p 21812 | more
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> write(9,"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,1024) = 1024 (0x400)
> . . .
> 
> # truss -c -p 21812
> ^Csyscall                     seconds   calls  errors
> write                   2.771261517  206935       0
>                      ------------- ------- -------
>                        2.771261517  206935       0
> 
> 
> (I checked on a separate system and the problem repeated there.)
> 
> I'm going to kill process 21812.

The log file gives some more context about what was stuck.
It notes "Terminated" for the process kill that I did:

. . .
checking for getpass... yes
checking for system V getpwnam... no
checking read/send-side pipe system... Terminated
unknown
configure: WARNING: unable to determine PIPEBUFFER type
checking for errorless select behavior with high timeouts... yes
checking for errno symbol E2BIG... OK
. . .

===
Mark Millard
marklmi at yahoo.com


===
Mark Millard
marklmi at yahoo.com