Performance issue with 5.4-RELEASE-p8 but not with 5.4-RELEASE

[Joseph Koshy]

On 10/24/05, Lukas Razik <lukas at razik.de> wrote:
>
> O.K.
>
> I've found the reason for the issue.
> After exchanging the newer
> src/sys/i386/include/cpufunc.h,v 1.142.6.1 2005/05/13 00:12:57
> src/sys/i386/i386/mp_machdep.c,v 1.235.2.6.2.3 2005/05/13 00:12:56
> from 5.4-RELEASE-p1 by the old
>
> src/sys/i386/include/cpufunc.h,v 1.142 2004/04/07 20:46:05
> src/sys/i386/i386/mp_machdep.c,v 1.235.2.6.2.2 2005/05/01 05:38:13
> from 5.4-RELEASE the issue disappeared.

>From the CVS commit log for this change at:

  http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/i386/i386/mp_machdep.c?only_with_tag=RELENG_5_4

  "Add a knob for disabling/enabling HTT, "machdep.hyperthreading_allowed".
   Default off due to information disclosure on multi-user systems."

Does turning HTT back on fix the regression?

--
FreeBSD Volunteer,     http://people.freebsd.org/~jkoshy