Re: [PATCH 0/30] W1: w1 core fixes, ds2490 updates, strong pullup

[David Fries <david@fries.net>]

On Tue, Jul 29, 2008 at 04:13:56PM -0700, Andrew Morton wrote:
> On Mon, 28 Jul 2008 21:04:33 -0500
> David Fries <david@fries.net> wrote:
> 
> > What follows is a long list of fixes and enhancements to the one wire
> > system, and even some documentation.
> > 
> > I no longer have any deadlocks, a thread was eliminated (along with
> > its one second wakeup interval), the cpu and time overhead are much
> > reduced for one wire accesses.  The time for the ds2490 to read a
> > temperature sensor went from 3.91 seconds (.002s user, 3.001s system)
> > to 0.860 seconds (0.004s user, 0.004s system).  I also added support
> > for the strong pullup to provide more current when requested.
> 
> This is all dreadfully late for 2.6.27, but it does seem to be rather
> important, so let's aim for 2.6.27.

Arguments for sooner: fixes some bad bugs, lower risk as it is
isolated to the w1 driver.

Arguments for later, the bugs aren't new, the first version of the
patch was sent in March and I have yet to get a response from anyone
using the ds1wm master (in some ARM handhelds for battery readings),
maybe if it gets in the merge window someone with the hardware will
actually try it before a kernel release.

I'm fine with either.

> The w1 code seems to have rather a lot of comments which start with /**
> but which aren't kerneldoc comments.  But /** is exclusively used to
> signify the start of a kerneldoc comment.  Please let's not invent new
> commenting styles like this.  Documentation/CodingStyle is there to
> help.

I did a quick look, some should be api documentation and updated for
kerneldoc, some clearly aren't.  I'll let Evgeniy Polyakov address
those.

> Please be aware that this:
> 
> Mime-Version: 1.0
> Content-Type: multipart/signed; micalg=pgp-sha1;
> 	protocol="application/pgp-signature"; boundary="OgqxwSJOaUobr8KG"
> 
> is rather receiver-hostile.  My MUA (at least) (sylpheed) manages to
> make a complete mess when saving-to-file, so I needed to go through all
> the patches and do various manual steps to fix this up.  I may still have
> some "=066"s in the changelogs.

I would have resubmitted it if you had asked, I'm using mutt, which
isn't exactly new or unknown, I guess the incompatibility is why
encryption and signing e-mail hasn't taken taken off as it could have.
There are some problems left in the changelog.

How about one tar.gz?

> WARNING: consider using strict_strtol in preference to simple_strtol

> Please use checkpatch.

I did use checkpatch.pl, simple_strtol was the only warning (I ignored
it to be consistent, the fix follows).  strict_strtol is safe here as
it only reads one integer from sysfs.  The patch didn't change,
checkpatch.pl did, I've updated to that as well.


The next set of patches has these minor updates, and checkpatch.pl
returns no errors or warnings.

0005-W1-feature-enable-hardware-strong-pullup.txt
Switch to strict_strtol, code style fixups.

0017-W1-w1_io.c-reset-comments-and-msleep.txt
Fix changelog long lines.

0020-W1-ds2490.c-add-support-for-strong-pullup.txt
Whitespace code style fixups.

0030-W1-ds2490.c-optimize-ds_set_pullup.txt
Whitespace code style fixups.


-- 
David Fries <david@fries.net>
http://fries.net/~david/ (PGP encryption key available)