Leonard Li wrote:
> >This is slightly offtopic, but due lack of a dbx-related mailinglist and
> >because it happens while debugging ksh93 I am dropping it here for now:
> >
> >While trying to debug ksh93 with "check -leaks" and/or "check -memuse"
> >dbx (version "Sun Dbx Debugger 7.4 117845-01 2005/03/05") simply crashes
> >like this:
> >-- snip --
> >% env - SHELL=$SHELL TERM=$TERM HOME=$HOME EDITOR=$EDITOR
> >/opt/onbld/bin/bldenv ./opensolaris.sh
> ># build the OS/Net B37 tree where ksh93 has been integrated
> >% cd /home/test001/ksh93/on_build1/test1/usr/src/lib/libshell
> ># run dbx like this:
> >% (LD_LIBRARY_PATH=$ROOT/lib dbx ../../cmd/ksh/i386/ksh)
> >(dbx) check -memuse
> >(dbx) run -o emacs
> >run -o emacs
> >Running: ksh -o emacs
> >(process id 19500)
> >Reading rtcapihook.so
> >Reading libdl.so.1
> >Reading rtcaudit.so
> >Reading libmapmalloc.so.1
> >Reading libgen.so.1
> >Reading rtcboot.so
> >Reading librtc.so
> >RTC: Enabling Error Checking...
> >dbx: can't find a system call entry point -- program not linked with
> >libc?
> >
> >dbx: internal error: signal SIGSEGV (no mapping at the fault address)
[snip]
> >... and again no success... ;-(
> >
> >Does anyone have a suggestion what I can do at this point to get "dbx"'s
> >"check -leaks / -memuse" functionality working on i386 ?
> 
> RTC is depended on libc.so for various reason. In this case, it tries to
> locate an special instruction (lcall, sysenter or syscall) but fail to
> find one. A workaround is to include functions like close(),
> sigpending() or setuid() in your customized libc and make sure one of
> these instructions is used in those functions.

Uhm... the problem is that this is no custom libc - it's just the libc
version in my B37 workspace which is almost identical to the one on the
CD - except the detail that I compiled it from source (using Sun Studio
10+recommended compiler patches).
I even copied /lib/libc.so.1 (after unmounting /libc/libc.so.1 to get
rid of the hardware-optimizsed version) to the proto/root_i386/
workspace with the same result - which means the dbx problem occurs even
when both libraries are identical - at which point I assume that dbx
somehow does not like to use libraries referenced via LD_LIBRARY_PATH
... ;-(

... and I need help with this catch22 - I cannot successfully hack&test
any RTC patches for libast without getting this part running...
heeellllppp... ;-(

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) roland.mainz at nrubsig.org
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 7950090
 (;O/ \/ \O;)

Reply via email to