Created attachment 26098 [details]
output of lspci -nn

Hello,

beginning with commit
"agp/intel-agp: Clear entire GTT on startup"
i.e.

5877960869333e42ebeb733e8d9d5630ff96d350 in 2.6.32.3 (2.6.32.y long term stable),

fc61901373987ad61851ed001fe971f3ee8d96a3 in 2.6.32-rc6 (mainline),

both USB and Sound are broken under some not fully understood circumstances on Intel's D945GCLF2 Mini-ITX-Board (Atom 330 CPU, BIOS version LF94510J.86A.0099.2008.0731.3030).

For USB the bug is triggered at least if support for uhci-hcd, ehcid-hcd and usb-storage is in-kernel.  Symptoms are messages like

    ehci_hcd 0000:00:1d.7: port 3 reset error -110
    hub 1-0:1.0: hub_port_status failed (err = -32)
    hub 1-0:1.0: port_wait_reset: err = -32
    hub 1-0:1.0: port 3 not enabled, trying reset again...
    hub 1-0:1.0: Cannot enable port 3.  Maybe the USB cable is bad

and an attached USB stick is unreadable.

For sound, there are a lot of different symptoms, for example messages like

    hda-intel: azx_get_response timeout, switching to polling mode: last cmd=0x200f0000
    hda-intel: Codec #2 probe error; disabling it...
    hda_intel: azx_get_response timeout, switching to single_cmd mode: last cmd=0x200f0000
    hda-intel: no codecs initialized

or, depending on module parameters, a _lot_ of messages spit on the terminal requiring a hard reboot. Sound is not usable in any situation.


Reverting the respective commits restored the expected behaviour, tested for 2.6.32.11 and 2.6.34-rc5, configs are attached.

This commit has been found guilty for other trouble too, see #14484.