From: Michael Buesch <mb@bu3sch.de>
To: Joakim Tjernlund <joakim.tjernlund@transmode.se>
Cc: "scottwood@freescale.com" <scottwood@freescale.com>,
Alessandro Rubini <rubini@gnudd.com>,
"bgat@billgatliff.com" <bgat@billgatliff.com>,
"linuxppc-dev@lists.ozlabs.org" <linuxppc-dev@lists.ozlabs.org>,
Jonathan Haws <Jonathan.Haws@sdl.usu.edu>
Subject: Re: Accessing flash directly from User Space [SOLVED]
Date: Sat, 31 Oct 2009 21:35:31 +0100 [thread overview]
Message-ID: <200910312135.34196.mb@bu3sch.de> (raw)
In-Reply-To: <OF7FB0F99B.2962ECBA-ONC1257660.006EDDB5-C1257660.006F2808@transmode.se>
On Saturday 31 October 2009 21:14:07 Joakim Tjernlund wrote:
> Michael Buesch <mb@bu3sch.de> wrote on 31/10/2009 17:42:54:
> >
> > On Saturday 31 October 2009 14:26:48 Joakim Tjernlund wrote:
> > > >
> > > > > On Friday 30 October 2009 16:08:55 Alessandro Rubini wrote:
> > > > > > > asm("eieio; sync");
> > > > > >
> > > > > > Hmm...
> > > > > > : : : "memory"
> > > > > >
> > > > > > And, doesn't ";" start a comment in assembly? (no, not on powerpc
> > > > > it seems)
> > > > >
> > > > > Yes, I think the barrier is wrong.
> > > > > Please try with
> > > > >
> > > > > #define mb() __asm__ __volatile__("eieio\n sync\n" : : :
> > > > > "memory")
> > > >
> > > > That definition worked great. I must have missed the : : : "memory" bit when
> > > > I was digging through code.
> > > >
> > > > Thanks, that gives me about a 2x speedup over the msync() calls.
> > >
> > > Exactly when should you use the barrier? At every access,
> > > every read or when changing from write to read?
> >
> > Well, it depends on the device you are accessing. I'll give you a small pseudo example.
> >
> > mmio[0] = address;
> > mmio[1] = data;
> > mb();
> > mmio[3] |= 0x01; /* This triggers an operation -> address=data */
> > /* probably also need an mb() here, if the following code
> > * depends on the operation to be triggered. */
>
> So anything that depends on the previous accesses needs a mb()
No, not really. I would always put an mb() where the comment is.
Imagine you have two instances of the above (probably in a loop):
mmio[0] = address;
mmio[1] = data;
mb();
mmio[3] |= 0x01; /* This triggers an operation -> address=data */
mb();
mmio[0] = address;
mmio[1] = data;
mb();
mmio[3] |= 0x01; /* This triggers an operation -> address=data */
mb();
> hmm, the mmio[0] and mmio[1] are written in order I hope?
We do not care in this example, as the write to [3] does trigger
the device operation. We do only care that [0] and [1] are set
when [3] is written. We do not care in what order [0] and [1] are written.
This is just an artificial example, but I think you get my point.
My point just being is that an mb() does only belong where it matters.
And what matters is defined by your device specifications.
--
Greetings, Michael.
next prev parent reply other threads:[~2009-10-31 20:36 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-10-27 19:59 Accessing flash directly from User Space Jonathan Haws
2009-10-27 20:13 ` Bill Gatliff
2009-10-27 20:15 ` Jonathan Haws
2009-10-27 20:30 ` Bill Gatliff
2009-10-27 20:31 ` Jonathan Haws
2009-10-27 22:24 ` Jonathan Haws
2009-10-27 22:32 ` Scott Wood
2009-10-27 22:35 ` Jonathan Haws
2009-10-27 22:42 ` Scott Wood
2009-10-27 22:52 ` Jonathan Haws
2009-10-28 9:56 ` Kenneth Johansson
2009-10-28 14:45 ` Jonathan Haws
2009-10-29 21:36 ` Joerg Albert
2009-10-29 21:39 ` Jonathan Haws
2009-10-29 16:33 ` Scott Wood
2009-10-29 16:48 ` Jonathan Haws
[not found] ` <BB99A6BA28709744BF22A68E6D7EB51F0330D368AA@midas.usurf.usu.edu >
2009-10-27 22:42 ` Alessandro Rubini
2009-10-28 11:24 ` Josh Boyer
2009-10-28 14:38 ` Jonathan Haws
2009-10-28 17:44 ` Accessing flash directly from User Space [SOLVED] Jonathan Haws
2009-10-29 9:00 ` Joakim Tjernlund
2009-10-29 9:15 ` Joakim Tjernlund
2009-10-29 17:01 ` Jonathan Haws
2009-10-29 11:08 ` Kenneth Johansson
2009-10-29 17:02 ` Jonathan Haws
2009-10-29 23:30 ` Scott Wood
2009-10-30 14:50 ` Jonathan Haws
2009-10-30 14:56 ` Michael Buesch
2009-10-30 14:57 ` Jonathan Haws
[not found] ` <BB99A6BA28709744BF22A68E6D7EB51F0330E23124@midas.usurf.usu.edu >
2009-10-30 15:08 ` Alessandro Rubini
2009-10-30 15:24 ` Michael Buesch
2009-10-30 15:33 ` Jonathan Haws
2009-10-31 13:26 ` Joakim Tjernlund
2009-10-31 16:42 ` Michael Buesch
2009-10-31 20:14 ` Joakim Tjernlund
2009-10-31 20:35 ` Michael Buesch [this message]
2009-10-31 22:31 ` Joakim Tjernlund
2009-11-01 10:07 ` Segher Boessenkool
2009-10-30 15:48 ` Micha Nelissen
2009-10-30 16:46 ` Jonathan Haws
2009-10-30 17:49 ` Gabriel Paubert
2009-10-30 15:57 ` Scott Wood
[not found] ` <BB99A6BA28709744BF22A68E6D7EB51F0330D3688B@midas.usurf.usu.edu >
2009-10-27 22:32 ` Accessing flash directly from User Space Alessandro Rubini
2009-10-27 22:33 ` Jonathan Haws
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200910312135.34196.mb@bu3sch.de \
--to=mb@bu3sch.de \
--cc=Jonathan.Haws@sdl.usu.edu \
--cc=bgat@billgatliff.com \
--cc=joakim.tjernlund@transmode.se \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=rubini@gnudd.com \
--cc=scottwood@freescale.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).