Date: Fri, 21 Jul 1995 09:08:47 +1000 From: Bruce Evans <bde@zeta.org.au> To: bde@zeta.org.au, terry@cs.weber.edu Cc: bugs@freebsd.org Subject: Re: possible ffs_vget() race condition Message-ID: <199507202308.JAA10420@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
[Cc list trimmed] >> The shift method has the advantage of giving the correct result :-). >> You meant ((~(u_quad_t)0) >> 1). This depends on u_quad_t being >> larger than int. >I don't understand why this would depend on quad being larger than int; >because the naked 1 is of type int? I should have said larger or the same size as an int. ~(foo_t)0 is usually wrong if foo_t is shorter than an int (it is the same as ~0 because of the broken ANSI value-preserving promotion rules), so using it gives a bad example for beginners to follow. ((foo_t)(-1)) always works although you may need further casts (e.g., on the i386, (u_short)(-1) is (int)0xffff, so shifting right 1 works right but shifting left 16 and then shifting right gives surprising results. The type of the shift count doesn't affect the type of the result. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199507202308.JAA10420>