kern/118842: [apic] Boot failure with apic enabled (FreeBSD 7)

David Naylor naylor.b.david at gmail.com
Thu Jul 17 21:30:04 UTC 2008


The following reply was made to PR kern/118842; it has been noted by GNATS.

From: David Naylor <naylor.b.david at gmail.com>
To: Luoqi Chen <Luoqi.Chen at brion.com>
Cc: "bug-followup at FreeBSD.org" <bug-followup at freebsd.org>,
 "dragonsa at highveldmail.co.za" <dragonsa at highveldmail.co.za>,
 "jhb at freebsd.org" <jhb at freebsd.org>
Subject: Re: kern/118842: [apic] Boot failure with apic enabled (FreeBSD 7)
Date: Thu, 17 Jul 2008 22:46:04 +0200

 --nextPart2810632.UEoOsYPp01
 Content-Type: text/plain;
   charset="iso-8859-1"
 Content-Transfer-Encoding: quoted-printable
 Content-Disposition: inline
 
 IT WORKS!!!
 
 I applied the below patch on FreeBSD-Current (cvsup today) and it works. =20
 
 I am still getting lots of "ata0: stat0=3D0x80 err=3D0x80 lsb=3D0x80 msb=3D=
 0x80"=20
 messages and the primary console does not change to grey when switching fro=
 m=20
 kernel messages to userland (does not happen when apic disable). =20
 
 Both hard drives are detected and all the CPU's are detected and working. =
 =20
 
 There was a problem previously with ad8 disconnecting and reconnecting.  It=
 =20
 had caused the computer to slow down and even crash once.  So far it does n=
 ot=20
 appear to be a problem...
 
 Thank you :-)
 
 David
 
 On Thursday 17 July 2008 11:44:00 Luoqi Chen wrote:
 > I've identified the cause: the SATA device in nForce chipset has the
 > HyperTransport MSI mapping capability, and we unconditionally enable the
 > MSI->HT mapping even though we don't use MSI in our ATA driver, this like=
 ly
 > confused the device.
 >
 > John, it is unclear to me why we enable the MSI->HT mapping at all given
 > that we always allocate MSI addresses from local APIC. To utilize HT
 > interrupt mechanism, shouldn't we be programming MSI with addresses from
 > the HT MSI mapping capability register (the device's or that of a parent
 > bridge)?
 >
 > For those who are having the same problem, please try commenting out the
 > MSI mapping enabling code and see if it works for you,
 >
 > Index: pci.c
 > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
 > RCS file: /home/ncvs/src/sys/dev/pci/pci.c,v
 > retrieving revision 1.355.2.2
 > diff -u -r1.355.2.2 pci.c
 > --- pci.c       12 Feb 2008 20:20:56 -0000      1.355.2.2
 > +++ pci.c       17 Jul 2008 09:36:12 -0000
 > @@ -564,9 +564,11 @@
 >                                                     (long long)addr);
 >                                 }
 >
 > +#if 0
 >                                 /* Enable MSI -> HT mapping. */
 >                                 val |=3D PCIM_HTCMD_MSI_ENABLE;
 >                                 WREG(ptr + PCIR_HT_COMMAND, val, 2);
 > +#endif
 >                                 break;
 >                         }
 >                         break;
 >
 > Thanks
 > -lq
 
 --nextPart2810632.UEoOsYPp01
 Content-Type: application/pgp-signature; name=signature.asc 
 Content-Description: This is a digitally signed message part.
 
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.4 (FreeBSD)
 
 iD8DBQBIf6+RUaaFgP9pFrIRAvu3AJ9Yno10lfwHzG0PYONnaHjIZyGZ9wCghuTp
 1dpaz7mAE2Gm7/oKXv4A++Q=
 =Rju4
 -----END PGP SIGNATURE-----
 
 --nextPart2810632.UEoOsYPp01--


More information about the freebsd-bugs mailing list