Date: Tue, 15 Dec 1998 00:47:30 +0100 From: Eivind Eklund <eivind@yes.no> To: Matt Dillon <dillon@FreeBSD.ORG>, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/sys/sys sem.h Message-ID: <19981215004730.U5444@follo.net> In-Reply-To: <199812142134.NAA18826@freefall.freebsd.org>; from Matt Dillon on Mon, Dec 14, 1998 at 01:34:56PM -0800 References: <199812142134.NAA18826@freefall.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Dec 14, 1998 at 01:34:56PM -0800, Matt Dillon wrote:
> dillon 1998/12/14 13:34:56 PST
>
> Modified files:
> sys/sys sem.h
> Log:
> Remove hardwired alignment, align to sizeof(long) which should be more
> portable (though the hardwired alignment of 16 would also have worked just
> fine and perhaps even resulted in better L1 cache operation).
This is a bad alignment. The correct restriction is something like
struct aligner {
char dummy;
union {
float f;
double d;
long int l;
void *p;
struct something *pstruct;
void (*func)();
} offsetter;
};
Align at offsetof(struct aligner, offsetter);
16 is better than sizeof(long), I think.
(I hope I got the set of types above correct - I think the pointer
set is right, at least).
Eivind.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19981215004730.U5444>