<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.24.4">
</HEAD>
<BODY>
<BLOCKQUOTE TYPE=CITE>
<BR>
<BLOCKQUOTE TYPE=CITE>
-------- Message transféré --------<BR>
<B>De</B>: Dan Williams <<A HREF="mailto:Dan%20Williams%20%3cdcbw@redhat.com%3e">dcbw@redhat.com</A>><BR>
<B>Reply-to</B>: Development discussions related to Fedora <fedora-devel-list@redhat.com><BR>
<B>À</B>: Development discussions related to Fedora <<A HREF="mailto:Development%20discussions%20related%20to%20Fedora%20%3cfedora-devel-list@redhat.com%3e">fedora-devel-list@redhat.com</A>><BR>
<B>Sujet</B>: Re: built-in USB drivers<BR>
<B>Date</B>: Mon, 09 Feb 2009 10:43:56 -0500<BR>
<BR>
<PRE>
On Sun, 2009-02-08 at 16:06 +0100, Philippe Carriere wrote:
> Hi,
>
> since Fedora 10, USB drivers (more precisely, ehci-hcd for which a
> patch is required to be snd-ubs-us112l - alsa-1.0.18 - compatible) are
> "built-in" in the official kernel package.
> For previous releases (F9), USB drivers always were modules, with the
> advantage that providing a patched module (using kmod rpmfusion
> framework) just required, after installing the kmod rpms, to
> disable/enable the module.
>
> I do not know the motivation of Fedora developers for choosing
> built-in rather than module in the pre-compilation configuration: is
> it indeed unavoidable ? Or does it exist a fairly simple way (I found
> nothing like this in documentations) to disable a built-in driver so
> as to replace it by a module, avoiding lengthy kernel compilation for
> just a few modified lines ?
Building modules in speeds up boot times. Obviously you only want to
build in modules that almost all of the machines of a certain
architecture are likely to need. The USB host controller drivers are
probably the best candidate for this.
</PRE>
</BLOCKQUOTE>
</BLOCKQUOTE>
<BR>
I'm probably a too old man to understand that (a few seconds of boot) time saving is preferable than (keeping some degrees of) freedom. Anyway, almost all is not all and undoubtfully not all hardware-to-be.<BR>
<BR>
<BLOCKQUOTE TYPE=CITE>
<BLOCKQUOTE TYPE=CITE>
<PRE>
My suggestion would be to encourage the upstream ALSA maintainers to
accept the patch necessary for your hardware, and then that patch will
find its way into the Fedora kernel when the ALSA folks push updates to
Linus.
Dan
</PRE>
</BLOCKQUOTE>
</BLOCKQUOTE>
<BR>
Well, ehci-hcd is not Alsa maintainers dependent and as can be seen in the attached emails, we do our best effort for having the patch integrated in the next kernel release. However, such integration always requires some times while we are also able, presently (ans since about 8 months), to propose a patched module for interested users, a fairly better way than a full compiled kernel. And such a situation will be probably encountered for other hardwares in a near future.<BR>
<BR>
Phil.<BR>
<BR>
<TABLE CELLSPACING="0" CELLPADDING="0" WIDTH="100%">
<TR>
<TD>
<HR NOSHADE>
<BR>
<BR>
philippe-f.carriere@wanadoo.fr
<DIV ALIGN=center></DIV><BR>
<BR>
<BR>
<DIV ALIGN=center></DIV><BR>
<BR>
<BR>
<DIV ALIGN=center></DIV><BR>
<BR>
<BR>
<BR>
</TD>
</TR>
</TABLE>
</BODY>
</HTML>