When doing make MALLOC_PRODUCTION=yes buildworld, a compiler warning will be issued which halts compilation. The warning is issued because line 463 of head/sbin/fsck_ffs/fsutil.c uses "%ld" and "%lld", for variables of type time_t, which I think are of the type int in this case when compiling for armv6. Fix: Typecasting like this solved the problem: printf("%21s:%8ld %2ld.%ld%% %8lld msec %2lld.%lld%%\n", buftype[i], (long)readcnt[i], (long) readcnt[i] * 100 / diskreads, (long) (readcnt[i] * 1000 / diskreads) % 10, (long long) msec, (long long) msec * 100 / totalmsec, (long long) (msec * 1000 / totalmsec) % 10); How-To-Repeat: Following this guide to step 13 should do the trick: http://ogris.de/howtos/freebsd-raspberry.html
On Mon, 2013-02-25 at 18:32 +0000, Josef Larsson wrote: > >Number: 176424 > >Category: arm > >Synopsis: Compiler warning, TARGET_ARCH=armv6, make MALLOC_PRODUCTION=yes buildworld > >Confidential: no > >Severity: non-critical > >Priority: low > >Responsible: freebsd-arm > >State: open > >Quarter: > >Keywords: > >Date-Required: > >Class: sw-bug > >Submitter-Id: current-users > >Arrival-Date: Mon Feb 25 18:40:00 UTC 2013 > >Closed-Date: > >Last-Modified: > >Originator: Josef Larsson > >Release: FreeBSD server 9.1-RELEASE > >Organization: > >Environment: > FreeBSD server 9.1-RELEASE FreeBSD 9.1-RELEASE #0 r243825: Tue Dec 4 09:23:10 UTC 2012 root@farrell.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC amd64 > >Description: > When doing make MALLOC_PRODUCTION=yes buildworld, a compiler warning will be issued which halts compilation. The warning is issued because line 463 of head/sbin/fsck_ffs/fsutil.c uses "%ld" and "%lld", for variables of type time_t, which I think are of the type int in this case when compiling for armv6. > >How-To-Repeat: > Following this guide to step 13 should do the trick: > http://ogris.de/howtos/freebsd-raspberry.html > >Fix: > Typecasting like this solved the problem: > > printf("%21s:%8ld %2ld.%ld%% %8lld msec %2lld.%lld%%\n", buftype[i], (long)readcnt[i], (long) readcnt[i] * 100 / diskreads, (long) (readcnt[i] * 1000 / diskreads) % 10, (long long) msec, (long long) msec * 100 / totalmsec, (long long) (msec * 1000 / totalmsec) % 10); FYI, this was fixed with r247269. Also, there's no need to use MALLOC_PRODUCTION anymore, a change about a month ago fixed the performance and memory-hogging problems. -- Ian
State Changed From-To: open->closed Closing this PR. The issue was fixed in r247269. Thanks!