CURRENT Intel ICH5 (82801EB) errors

[Pyun YongHyeon]

On Thu, Nov 17, 2005 at 12:36:25PM +0300, Andrew Bliznak wrote:
 > On 11/17/05, Pyun YongHyeon <pyunyh at gmail.com> wrote:
 > > On Thu, Nov 17, 2005 at 10:49:20AM +0300, Andrew Bliznak wrote:
 > >  > On 11/17/05, Pyun YongHyeon <pyunyh at gmail.com> wrote:
 > >  > > On Wed, Nov 16, 2005 at 01:53:25PM +0300, Andrew Bliznak wrote:
 > >  > >  > Just for information
 > >  > >  >
 > >  > >  > FreeBSD beef.xxxxxx.com 7.0-CURRENT FreeBSD 7.0-CURRENT #0: Tue Nov 15
 > >  > >  > 13:03:12 EET 2005
 > >  > >  > andrew at pyvo.xxxxxx.com:/usr/home/andrew/C/obj/usr/home/andrew/C/src/sys/GENERIC
 > >  > >  >  i386
 > >  > >  >
 > >  > >  > from dmesg:
 > >  > >  > ---
 > >  > >  > Nov 15 17:02:31 beef kernel: pcm0: <Intel ICH5 (82801EB)> port
 > >  > >  > 0xd000-0xd0ff,0xe100-0xe13f mem
 > >  > >  > 0xf8201000-0xf82011ff,0xf8202000-0xf82020ff irq 17 at device 31.5 on
 > >  > >  > pci0
 > >  > >  > Nov 15 17:02:31 beef kernel: pcm0: unable to initialize the card
 > >  > >  > Nov 15 17:02:31 beef kernel: device_attach: pcm0 attach returned 6
 > >  > >
 > >  > > How about attached patch?
 > >  >
 > >  > No luck, rebuild module and reload
 > >  > ---
 > >  > Nov 17 09:37:50 beef kernel: pcm0: <Intel ICH5 (82801EB)> port
 > >  > 0xd000-0xd0ff,0xe100-0xe13f mem
 > >  > 0xf8201000-0xf82011ff,0xf8202000-0xf82020ff irq 17 at device 31.5 on
 > >  > pci0
 > >  > Nov 17 09:37:50 beef kernel: pcm0: primary codec not ready!pcm0:
 > >  > unable to initialize the card
 > >  > Nov 17 09:37:50 beef kernel: device_attach: pcm0 attach returned 6
 > >  > ---
 > >  >
 > >  > Will try late with reboot, thanks anyway.
 > >  >
 > >
 > > Hmm, then try this one. It just ignores PCR bit.
 > 
 > Rebuild, reload. This time much better, and I hear sound from speakers
 > ( debug.bootverbose 0->1)
 > ---

Great.
In order to commit this change to tree I need more feedbacks from
ICH[4-7] users as the patch changes fundamental operation procedures.
(i.e. ignoring PCR bit while accessing codec.)
It's also possible to introduce a new tunable to ignore PCR bit
but I'd like to avoid adding additional one as possible as I can.

-- 
Regards,
Pyun YongHyeon