Array 5 liveCD - report (X11, ide issues; suspend success)

[Matt Zimmerman]

On Sun, Feb 20, 2005 at 02:53:49PM -0800, Zack Weinberg wrote:

> The first problem I was hoping to solve, is that since upgrading to kernel
> 2.6 and ACPI (which has otherwise been quite nice, e.g. battery life is
> improved) I have not been able to use suspend-to-RAM.  This used to work
> perfectly with 2.4 and APM.
>
> I'm pleased to report that
> following the directions at http://www.ubuntulinux.org/wiki/HoaryPM
> also works perfectly, from a text console or from an X session.  (A
> minor improvement would be, from an X session, to present the
> xscreensaver unlock dialogue immediately on resume - I had to guess
> that the reason I was seeing a black screen was because the
> screensaver was active.)

These two statements seem to be in conflict; the instructions on HoaryPM
include setting up suspend-to-RAM.  Does this mean that suspend-to-RAM works
after you follow the instructions?

> While I'm complaining about X issues, this annoyance isn't fixed in
> Ubuntu either: If I am in an X session, and I close the lid, when I
> reopen it the backlight does not turn back on.  I must press Fn+F8
> (the magic direct-to-BIOS sequence to switch between the LCD panel and
> the external VGA connector).  I then get a garbled screen.  Switching
> to a text console and back corrects the display.  David Dawes gave me
> to understand this was fixed in the XFree86 4.4 "nv" driver, so I was
> really hoping it would be fixed in Xorg whosis as well...

Please report a bug (Package: xserver-xorg).

> A few other comments: 
> 
>  - The LiveCD boot sequence should notice when no network adapter
>    has a good link-layer connection, and not bother running the DHCP
>    client.

It already does that; if it doesn't work for your hardware (with the latest
daily live CD), then either the driver lacks the capability to detect
linkbeat, or there is a bug.

>  - The initrd loads every single IDE driver with -k, and then tries to
>    remove all the unnecessary ones later.  Since they were all loaded
>    with -k, this fails.  (Debian's initrd has the exact same bug.
>    Y'all at least seem to have made it not spew error messages while
>    failing to remove them.)  It should either not use -k so that
>    removal works, or better, it should figure out from PCI data which
>    drivers are relevant, and only load those.

This works fine for me (the modules are removed as should be expected).

If it isn't working for you, it certainly isn't because of the -k option.
modprobe -k <foo> is exactly equivalent to modprobe <foo> with
module-init-tools.

We intend to make this smarter for Hoary, but this is the sort of change
that needs to be made _way_ ahead of feature freeze.

>  - Sound does not work, except that mysteriously it starts working
>    after a suspend/resume cycle.  (AC'97 sound chip, IIRC.)  In Debian
>    it doesn't work at all.

Please file a bug with the relevant dmesg output.

>  - I tried to use "su" and was prompted for the root password, which I
>    did not know.  "sudo -s" worked, since I was successful in guessing
>    that the password for the "ubuntu" user was "ubuntu".  I don't see
>    why either of these users has a password at all on the livecd,
>    though.

The fact that you were prompted for a password was a bug which was fixed
shortly after Array 5.

Additionaly, there have since been further refinements, and there are no
longer any passwords set on the live CD at all.

>  - The only documentation on the livecd is for the installer, which is
>    not particularly useful.  A README at top level, including all the
>    stuff that's in the F1-F10 boot help screens, plus other useful
>    info (such as the root password ;-) would be really nice.

There is no password, and I'm not sure that I understand the value of
duplicating the information available in the help screens.  Is there
something in particular that you feel needs to be documented, which is
specific to the live CD?

-- 
 - mdz