Re: cvs commit: src/sys/kern init_main.c kern_thread.c

[ Available lists | Index of cvs-all | Month of Mar 2007 | Week of 28 Mar 2007 | Raw email | View thread | Wrap long lines | Reply | Tag ]
From
Attilio Rao <attilio@freebsd.org>
Date
28 Mar 2007 14:24:41
Subject
Re: cvs commit: src/sys/kern init_main.c kern_thread.c
Message-ID
3bbf2fe10703280724h3d59f3d8t159c2997d84cb950@mail.gmail.com


[ Hide this part ]
2007/3/27, Attilio Rao <attilio@freebsd.org>:
> 2007/3/27, John Baldwin <jhb@freebsd.org>:
> > jhb 2007-03-27 16:51:34 UTC
> >
> > FreeBSD src repository
> >
> > Modified files:
> > sys/kern init_main.c kern_thread.c
> > Log:
> > Align 'struct thread' on 16 byte boundaries so that the lower 4 bits are
> > always 0. Previously we aligned threads on a minimum of 8-byte boundaries.
> >
> > Note: This changes the uma zone to no longer cache align threads. We
> > really want the uma zone to do align threads to MAX(16, cache line size)
> > but there currently isn't a good way to express that to uma.
>
> We can workaround it using a function like:
> int uma_get_align(void)
> (which would be the counter part of uma_set_align).
> Its implementations is trivial.

I have a diff for that:
http://users.gufi.org/~rookie/works/patches/uma2.diff

I had no time to sync my branches with last HEAD changes, so I can't
fix uma_zcreate() call... just adding the uma_get_align() function.

Thanks,
Attilio


--
Peace can only be achieved by understanding - A. Einstein


Elapsed time: 0.187 seconds