using -ftracer stops buildworld at shutdown.c

From: jason henson (jason_at_ec.rr.com)
Date: 07/27/05

  • Next message: Jeremie Le Hen: "Re: ProPolice symbols in libc or libssp ?" 
    Date: Tue, 26 Jul 2005 18:46:55 -0400
To: hackers@freebsd.org

    Adding -ftracer to my make.conf causes problems with
    /src/sbin/shutdown/shutdown.c
    This is my make.conf settings
    CPUTYPE=athlon-xp
    CFLAGS=-O2 -pipe -ffast-math -funroll-loops -ftracer
    COPTFLAGES=-O2 -pipe -funroll-loops -ffast-math -ftracer

    Doing a little research I found these items:

    -ftracer
         Perform tail duplication to enlarge superblock size. This
    transformation simplifies the control flow of the function allowing
    other optimizations to do better job.

      Why do I get "warning: variable (or argument) `blah' might be
    clobbered by `longjmp' or `vfork'" and how do I fix it?

    When the C runtime executes a longjmp, the flow of control passes back
    to the last setjmp that was executed. In the process, the stack (which
    contains arguments and local variables) must be 'unwound', that is,
    returned to the exact same state that it was in when setjmp was executed.

    Unfortunately, due to optimizations such as storing the value of a
    variable in a register instead of on the stack, the compiler cannot
    always guarantee that the stack will be restored precisely to its
    original state. Thus some local variables and/or arguments can be
    'clobbered'.

    To avoid this warning, the variables and/or arguments should be
    prevented from being optimized by declaring them as volatile.

    So I sprinkled some volatiles around, but mostly got more errors that
    said "gcc volatile discards qualifiers from pointer target type" in the
    fprintf functions.

    The error went away after I removed -ftracer, but from what I have read
    it would be advantageous to change the variable pf to be stored on the
    stack and not in a register. That is if it is not stored on the stack.
      I did a search of all the listed include files in shutdown.c and found
    no matches for pf.

    Anyone one have any thoughts or insights on this? Am I completely off
    in left field?

    Jason

    ===> sbin/shutdown (all)
    cc -O2 -pipe -DNOPROFILE -ffast-math -ftracer -funroll-loops
    -march=athlon-xp -Wsystem-headers -Werror -Wall -Wno-format-y2k -W
    -Wno-unused-parameter -Wstrict-prototypes -Wmissing-prototypes
    -Wpointer-arith -Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch
    -Wshadow -Wcast-align -Wunused-parameter -Wchar-subscripts -Winline
    -Wnested-externs -Wredundant-decls -c /usr/src/sbin/shutdown/shutdown.c
    /usr/src/sbin/shutdown/shutdown.c: In function `timewarn':
    /usr/src/sbin/shutdown/shutdown.c:276: warning: variable 'pf' might be
    clobbered by `longjmp' or `vfork'
    *** Error code 1

    Stop in /usr/src/sbin/shutdown.
    *** Error code 1

    Stop in /usr/src/sbin.
    *** Error code 1

    Stop in /usr/src.
    *** Error code 1

    Stop in /usr/src.
    *** Error code 1

    Stop in /usr/src.
    _______________________________________________
    freebsd-hackers@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
    To unsubscribe, send any mail to "freebsd-hackers-unsubscribe@freebsd.org"

  • Next message: Jeremie Le Hen: "Re: ProPolice symbols in libc or libssp ?"

    Relevant Pages