Re: SPOOFED: Re: SVN rev 206755 breakage

From: Michael Butler <imb_at_protected-networks.net>
Date: Sun, 18 Apr 2010 09:37:46 -0400
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 04/18/10 01:57, Alexander Motin wrote:
> More important probably would be `pciconf -lvcb`.
> 
> Intel controllers after ICH6 change both ID and set of resources,
> depending on AHCI enabled in BIOS. There is separate set of IDs for
> controllers with AHCI enabled. As I can see, Linux handles ID 0x27c4 as
> non-AHCI SATA. If for some reason this ID could be used for both modes
> (I have doubts), we may try to set AHCI_Q_NOFORCE flag to make driver
> check PCI class/subclass, if it is correct there.
> 

atapci0_at_pci0:0:31:2:    class=0x010180 card=0xff101179 chip=0x27c48086
rev=0x02 hdr=0x00

    vendor     = 'Intel Corporation'


    device     = '82801GBM/GHM (ICH7-M Family) Serial ATA Storage
Controller'

    class      = mass storage


    subclass   = ATA


    bar   [10] = type I/O Port, range 32, base 0x1f0, size  8, enabled


    bar   [14] = type I/O Port, range 32, base 0x3f4, size  1, enabled


    bar   [18] = type I/O Port, range 32, base 0x170, size  8, enabled


    bar   [1c] = type I/O Port, range 32, base 0x374, size  1, enabled
    bar   [20] = type I/O Port, range 32, base 0x18b0, size 16, enabled
    cap 01[70] = powerspec 2  supports D0 D3  current D0

When AHCI is enabled, the device ID changes to 0x27c5.

In this case, the probe succeeds but, since MSE is not set, the
allocation of memory-IO space to BAR(5) is futile and the reset fails
since it addresses undecoded space. Thus the attach fails,

	imb
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (FreeBSD)

iEYEARECAAYFAkvLCyoACgkQQv9rrgRC1JK3UQCfXG1K3B7kOo35koBWdTohYt7/
qygAoM0kn0ZSYeD5P0Hu7kr3ci+otV3m
=sk9Y
-----END PGP SIGNATURE-----
Received on Sun Apr 18 2010 - 11:37:51 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:02 UTC