Please build ata_piix into kernel (not as module)

Bug #311730 reported by Martin Pitt
16
This bug affects 1 person
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Fix Released
Wishlist
Tim Gardner

Bug Description

2.8.28-4 started building most common modules statically into the kernel, according to https://blueprints.launchpad.net/ubuntu/+spec/jaunty-boot-performance-kernel . This didn't include ata_piix, though, which is used in relatively current Dell notebooks (Latitude D430 here):

  00:1f.1 IDE interface: Intel Corporation 82801G (ICH7 Family) IDE Controller (rev 01)

Booting without an initramfs does not work on this machine due to this. I booted into initramfs, and removed ata_generic, pata_acpi, and ata_piix; modprobing the latter detects the internal hard drive, modprobing the first two does not give any kernel messages.

So should this be built in as well, or is this controller supposed to use a different driver?

Thank you!

Revision history for this message
Martin Pitt (pitti) wrote :
Revision history for this message
Martin Pitt (pitti) wrote :
Changed in linux:
importance: Undecided → Wishlist
Changed in linux:
status: New → Triaged
Revision history for this message
Tim Gardner (timg-tpi) wrote :

@Martin - I've been working towards compiling in all of the devices that most PCs require in order to mount a root file system without going through the initramfs step. I started a discussion thread on the kernel-team mailing list:

https://lists.ubuntu.com/archives/kernel-team/2009-January/003987.html

Changed in linux:
assignee: nobody → timg-tpi
status: Triaged → In Progress
Revision history for this message
Martin Pitt (pitti) wrote :

Hello Tim,

I tested your kernel in http://people.ubuntu.com/~rtg/2.6.28-4.6-build-in-ata, and indeed with it I can boot without an initramfs, which cuts down my boot speed (grub press 'b' to gdm login appearing) from 61 to 43 seconds.

Caveats:
 - specifying root device with root=UUID=DEADBEEF does not work. I remember Scott saying that the kernel itself would be able to handle UUIDs for ext3, but apparently it doesn't? Going back from UUIDs to hardcoded device names would be quite a regression if we'd do it by default.

 - usplash does not start (no big deal, that can be fixed easily with an init script, I can do that)

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package linux - 2.6.28-4.7

---------------
linux (2.6.28-4.7) jaunty; urgency=low

  [ Tim Gardner ]

  * Enable CONFIG_ATH5K=m for i386/amd64
    - LP: #306719
  * Build all i386/amd64 AGP/DRM components as modules.
    - LP: #312721
  * git commands are now installed outside the default $PATH
    Use 'git CMD' instead of 'git-CMD'.
  * Build in most PATA/SATA drivers. This should allow most i386/amd64 systems to boot
    without an initramfs, though some support work is still required in initramfs-tools
    and grub.
    - LP: #311730

 -- Tim Gardner <email address hidden> Fri, 02 Jan 2009 07:33:09 -0700

Changed in linux:
status: In Progress → Fix Released
Revision history for this message
Andres Järv (andresjarv) wrote :

Not loading the initrd also sped up the first part of the boot sequence for me noticably. Just got an error about the hardware clock but everything seems to be working just fine.

Revision history for this message
Bogdan Gribincea (bogdan-gribincea) wrote :

Since pata_acpi and ata_generic were built statically into the kernel my PATA controller stopped working. I need a way to disable pata_acpi and use ata_generic instead.

Please look at bug #315006 for detailed info

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.