64 bit Ubuntu cd on 32 bit system: more helpful error message

Bug #272087 reported by Wouter Stomp
28
This bug affects 4 people
Affects Status Importance Assigned to Milestone
gfxboot (Ubuntu)
Triaged
Low
Unassigned

Bug Description

When you boot the 64 bit ubuntu desktop cd on a 32 bit system, you get the following error message after you try to boot:

This kernel requires an x86-64 CPU, but only detected an i686 CPU.
Unable to boot - please use a kernel appropriate for your CPU.

Most people won't be able to make sense of this at all if they accidentally downloaded the wrong cd. There should be an understandable message telling you to download the 32 bit version of ubuntu.

Tags: iso-testing
tags: added: iso-testing
affects: ubuntu → casper (Ubuntu)
Changed in casper (Ubuntu):
importance: Undecided → Low
Changed in casper (Ubuntu):
status: Confirmed → Triaged
Revision history for this message
Stéphane Graber (stgraber) wrote :

Tentatively moving to gfxboot though the source of that message is likely the kernel, it's more likely that we can do something in gfxboot than in that part of the linux kernel.

affects: casper (Ubuntu) → gfxboot (Ubuntu)
Revision history for this message
Bryan Quigley (bryanquigley) wrote :

As an alternative, we could consider changing the wording in the kernel..:
arch/x86/boot/cpu.c: printf("This kernel requires an %s CPU, ",
arch/x86/boot/main.c: puts("Unable to boot - please use a kernel appropriate "
arch/x86/boot/cpu.c: cpu_name function, which only has x86_64 and i686 defined

SO:
 printf("This kernel requires an %s CPU, ", cpu_name(req_level));
 printf("but only detected an %s CPU.\n", cpu_name(cpu_level));

BECOMES:
 printf("This media requires an %s bit CPU, ", req_level);
 printf("but only detected an 32 bit CPU");

SO:
puts("Unable to boot - please use a kernel appropriate "
"for your CPU.\n");

BECOMES:
puts("Please acquire the 32 bit version of the media you are trying to use\n");

Thoughts on this strategy?

Revision history for this message
JohnWashington (ubuntu-johnwash) wrote :

Bryan, I can appreciate that you're trying to change the wording to make it more 'friendly', but at the same time you would be obscuring where the issue is detected. A user who is smart enough to download or otherwise obtain a CD, perhaps even burning it from an iso, is quite a smart user. I know a lot of people who stand no chance of doing such a thing. So surely these somewhat smart users are also smart enough to use Google and find out what the existing messages mean.

There's a danger of 'bike-shedding' here. Aren't there plenty of more important things for us to beg developers to fix?

Revision history for this message
Bryan Quigley (bryanquigley) wrote :

@JohnWashington

It was specifically mentioned in the discussion about going to "64 bit recommended" that we need a better error message. This is also for people who get an Ubuntu CD mailed to them. And even then, burning a CD is not nearly as technical as knowing what a kernel is...

Revision history for this message
JohnWashington (ubuntu-johnwash) wrote :

Bryan, you asked for thoughts. Don't shoot the messenger! :)

In part my objection is that the proposed messages have not just been dumbed down and convey less information, they're actually wrong. As you know, CDs contain streams of bits, they don't have any inherent "bitness". Media doesn't require a 64-bit cpu. You surely don't want naive users going to a shop and saying I want a pack of 32-bit blank CDs instead of these 64-bit ones.

Are the current messages a real problem? In other words, instead of some experienced users thinking they MAY be a problem, has this repeatedly come up as a show-stopper in actual usability testing?

Revision history for this message
Bryan Quigley (bryanquigley) wrote :

@JohnWashington

Apologies, if you thought I was shooting the messenger. I didn't mean it that way :/.

>Media doesn't require a 64-bit cpu. You surely don't want naive users going to a shop and saying I want a pack of 32-bit blank
I'm not sure that was implied by my statement. Not sure of any generic way to make this more clear.

> has this repeatedly come up as a show-stopper in actual usability testing?
I don't think it's come up as a show-stopper, and it's rated as "Low" which I think is correct. I think if we can come up with better language that could be accepted upstream, we could improve this for everyone. It's really just a language issue, no reason this should impact development time..

Revision history for this message
Bryan Quigley (bryanquigley) wrote :
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

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