From: Juan R. Sáenz-Diez (Yuri) (jrsdm++at++acm.org)
Date: 06/28/2000 14:34:57
Hi,
I happen to be experiencing the same problems and was just up to post a
similar message. Mine is an Onyx2 iR2, rest is the same, including the core.
I can't find the reason why my child dies, although it happens very
infrequently -- most likely when playing a demo
In my case, I don't see the "wrapping" message. I do not do anything special
about dlists. I have noticed that the stack traces back to pfConfig. Maybe
it's setting up something for the pipe's incremental stuff (clipmap) but
fails and takes a while to die? I can really fly for several minutes and
then -- crash.
Not that I have been helpful I'm afraid: just wanted to share the pain...
please help!
Regards,
Yuri.
---- Por favor, acuse recibo -- Please acknowledge receipt ---- Juan R. "Yuri" Sáenz-Diez Vis-Sim & VR Consultant +34 630 815 772 jrsdm++at++acm.org----- Original Message ----- From: Robert Doyle <doyle++at++overlord.nrl.navy.mil> To: <info-performer++at++sgi.com> Sent: Wednesday, June 28, 2000 8:38 PM Subject: Clip texturing problems.
> > > Machine: Onyx1-IR, 4 CPU, 2RM, Irix 6.5.7 > > Performers: > > I'm encountering clip texturing problems causing core dumps. The problem > occurs only in multiprocessor mode or when using display lists with > pfuTravSetDListMode(). If I run in single-processor mode or in > multiprocessor mode with display lists disabled, the problem does not > occur. > > In addition, the probability of a core dump increases with data base > size. Terrain with 90,000 triangles will core dump within minutes while > terrain with 20,000 triangles has not core dumped at all. All terrain > uses a 16K X 16K map. I can observe broad lines of incorrect texture > appearing when error messages (see below) begin to be printed. There > appear to be no problems with shared memory and performance is otherwise > very good. > > The clip texture is loaded using pfdLoadClipTexture() and the code is > similar to that found in > /usr/share/Performer/src/pguide/libpf/C/cliptex.c. > > The error messages are: > > PF Warning/Usage: pfMemory::unref() Attempt to unreference > memory > with 0 reference count. > PF Warning/Usage: pfMemory::unref() Attempt to unreference > memory > with 0 reference count. > PF Warning/Usage: pfMemory::ref() Reference count has > wrapped! > Clamping to 16777215 > PF Warning/Usage: pfMemory::unref() Attempt to unreference > memory > with 0 reference count. > PF Notice(12): Caught SIGCHLD. Exiting due to death of > child > with pid 25082. > > dbx shows: > > dbx version 7.2.1 Mar 13 1998 19:00:36 > Core from signal SIGBUS: Bus error > (dbx) where > > 0 pfTexLoad::pr_apply(void)(0x11, 0xffffffff, 0xf, 0x1, 0x4e460dc0, 0x1, > 0x2418cdc, 0x400) > ["/xlv61/performer/work/perf/lib/libpr/pfTexLoad.C":779, > 0x1e3b6d4] > 1 pfDispList::pr_caseDL_APPLYTLOAD(const DLElt*)(0x11, 0xffffffff, > 0xf, 0x1, > 0x4e460dc0, 0x1, 0x2418cdc, 0x400) > ["/xlv61/performer/work/perf/lib/libpr/pfDispList.C":1342, 0x1da2984] > 2 pfDispList::pr_drawFlat(void)(0x4f9c1d90, 0xffffffff, 0xf, 0x1, > 0x4f9c1ec8, > 0x1, 0x2418cdc, 0x400) > ["/xlv61/performer/work/perf/lib/libpr/pfDispList.C":3858, 0x1da0700] > 3 pfDispList::draw(void)(0x11, 0xffffffff, 0xf, 0x1, 0x4e460dc0, 0x0, > 0x2418cdc, 0x400) > ["/xlv61/performer/work/perf/lib/libpr/pfDispList.C":660, > 0x1d9fb10] > 4 pfPipe::pf_drawIncrementalState(void)(0x0, 0xffffffff, 0xf, 0x1, > 0x4e460dc0, 0x1, 0x2418cdc, 0x400) > ["/xlv61/performer/work/perf/lib/libpf/pfPipe.C":2353, 0x1ce9034] > 5 ::mpDraw(void)(0x11, 0xffffffff, 0xf, 0x0, 0x4e460dc0, 0x0, 0x0, > 0x400) > ["/xlv61/performer/work/perf/lib/libpf/pfProcess.C":6281, 0x1d56b08] > 6 ::pfConfig(0x0, 0x0, 0xf, 0x1, 0x0, 0x0, 0x0, 0x400) > ["/xlv61/performer/work/perf/lib/libpf/pfProcess.C":2851, 0x1d4af54] > 7 ::main(0x2, 0x7fff2f34, 0xf, 0x1, 0x4e460dc0, 0x1, 0x2418cdc, > 0x400) > ["/movies3/doyle/terrain-current/pgl-terrain/pgl.C":122, 0x1000a82c] > 8 __start() > ["/xlv55/kudzu-apr12/work/irix/lib/libc/libc_n32_M4/csu/crt1text.s":177, > 0x1000a428] > (dbx) q > > I'll be glad to provide more information if needed. > > Thanks, > Bob Doyle > > > -- > Robert J. Doyle, Jr. http://overlord.nrl.navy.mil > Code 5594 e-mail: doyle++at++overlord.nrl.navy.mil > Naval Research Laboratory phone: 202-767-8395 > Washington, DC 20375-5000 fax: 202-404-7402 > ----------------------------------------------------------------------- > List Archives, FAQ, FTP: http://www.sgi.com/software/performer/ > Submissions: info-performer++at++sgi.com > Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2b29 : Wed Jun 28 2000 - 14:39:20 PDT