|
Home > Archive > Debian Developers > August 2004 > Suppressing loading of a kernel module
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Suppressing loading of a kernel module
|
|
| Douglas Bates 2004-08-15, 5:52 pm |
| I am running Debian testing with the 2.6.7-1-k7 kernel on a computer
with the Nvidia Nforce 2 chipset. I would like to use the optical
S/PDIF output to drive a 5.1 speaker system.
As I understand it the proprietary Nvidia sound driver module (nvsound)
provides better support for 6 channel audio than does the i810_audio
driver in the kernel sources. I have downloaded and installed the
NFORCE-Linux-x86-1.0-0283-pkg1.run file from the Nvidia download site.
That appears to have been successful and I have an nvsound module for
this kernel. However, on reboot the i810_audio module continues to be
loaded before the nvsound module and it takes precedence. I have been
unable to discover how to suppress the loading of the i810_audio module.
It is not clear to me what is causing it to be loaded. I can see that
some of the entries in /lib/modules/2.6.7-1-k7/modules.alias and
/lib/modules/2.6.7-1-k7/modules.pcimap have the same signature for
i810_audio and nvsound and I have commented out those aliases for
i810_audio but that module still gets loaded.
Please cc: me on replies. I have suspended my debian-devel subscription
while I am at a location with only dial-up access.
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| James Michael Greenhalgh 2004-08-15, 5:52 pm |
| On Sun, 2004-08-15 at 17:11, Douglas Bates wrote:
> As I understand it the proprietary Nvidia sound driver module (nvsound)
> provides better support for 6 channel audio than does the i810_audio
> driver in the kernel sources. I have downloaded and installed the
> NFORCE-Linux-x86-1.0-0283-pkg1.run file from the Nvidia download site.
> That appears to have been successful and I have an nvsound module for
> this kernel. However, on reboot the i810_audio module continues to be
> loaded before the nvsound module and it takes precedence. I have been
> unable to discover how to suppress the loading of the i810_audio module.
Could this be hotplug? You may wish to add the i810_audio to the
/etc/hotplug/blacklist
HTH,
James
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| MiguelGea 2004-08-15, 5:52 pm |
| On dg, 2004-08-15 at 23:11, Douglas Bates wrote:
> I am running Debian testing with the 2.6.7-1-k7 kernel on a computer
> with the Nvidia Nforce 2 chipset. I would like to use the optical
> S/PDIF output to drive a 5.1 speaker system.
>
> As I understand it the proprietary Nvidia sound driver module (nvsound)
> provides better support for 6 channel audio than does the i810_audio
> driver in the kernel sources. I have downloaded and installed the
> NFORCE-Linux-x86-1.0-0283-pkg1.run file from the Nvidia download site.
> That appears to have been successful and I have an nvsound module for
> this kernel. However, on reboot the i810_audio module continues to be
> loaded before the nvsound module and it takes precedence. I have been
> unable to discover how to suppress the loading of the i810_audio module.
> It is not clear to me what is causing it to be loaded. I can see that
> some of the entries in /lib/modules/2.6.7-1-k7/modules.alias and
> /lib/modules/2.6.7-1-k7/modules.pcimap have the same signature for
> i810_audio and nvsound and I have commented out those aliases for
> i810_audio but that module still gets loaded.
>
Add the module name (i810_audio) to /etc/hotplug/blacklist
> Please cc: me on replies. I have suspended my debian-devel subscription
> while I am at a location with only dial-up access.
>
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Andrew Pollock 2004-08-15, 8:48 pm |
| On Sun, Aug 15, 2004 at 05:49:38PM -0400, James Michael Greenhalgh wrote:
>
> Could this be hotplug? You may wish to add the i810_audio to the
> /etc/hotplug/blacklist
>
I've been meaning to ask, at what point did hotplug start overlapping into
discover's space?
I too found that #236423, whilst originally I believed to be caused by
discover, later seemed to be caused by hotplug. (For all I know it may have
been caused by hotplug all along).
It seems to me that hotplug seems to be much more indiscriminate than
discover when it comes to loading modules.
regards
Andrew
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Douglas Bates 2004-08-16, 7:59 am |
| MiguelGea wrote:[vbcol=seagreen]
> On dg, 2004-08-15 at 23:11, Douglas Bates wrote:
>
>
> Add the module name (i810_audio) to /etc/hotplug/blacklist
>
Thanks to you and James for the suggestions regarding the hotplug
blacklist. It turns out that it was discover that was picking up the
i810_audio module and I needed to add
skip i810_audio
to /etc/discover.conf. Then I found out that to use alsa I should have
added lines to /etc/discover.conf to skip all the OSS drivers, including
the nvsound driver. So now I finally have the system configured with
the snd_intel8x0 driver.
An interesting way to spend a Sunday.
--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
| |
| Marco d'Itri 2004-08-16, 8:50 pm |
| On Aug 16, Andrew Pollock <apollock@debian.org> wrote:
> I've been meaning to ask, at what point did hotplug start overlapping into
> discover's space?
A few months ago. Actually hotplug was always supposed to behave like
this, but it was broken on debian systems because pciutils lacked the
pcimodules program.
> I too found that #236423, whilst originally I believed to be caused by
> discover, later seemed to be caused by hotplug. (For all I know it may have
> been caused by hotplug all along).
Anyway, #236423 is not a bug in hotplug or discover: it's not their
fault if a driver spews the kernel log with crap after being loaded.
I'm closing it because I remember that it's a duplicate of another bug
opened against a kernel package (a patch was provided, but I do not know
the current status of the bug).
BTW, please remember to ALWAYS Cc reassigned bugs to the appropriate
maintainers, because they are not otherwise notified so I did not notice
this bug before.
> It seems to me that hotplug seems to be much more indiscriminate than
> discover when it comes to loading modules.
It seems to me that our binary kernel packages contain drivers which are
not suitable for general use. They should either not be built by default
or blacklisted by the kernel package.
--
ciao, |
Marco | [7541 pr/9WdNkiFUyg]
|
|
|
|
|