Fwd: Is disabling ctrl-alt-backspace really such a good idea? - no.

[John Moser]

On Thu, Feb 12, 2009 at 5:16 PM, Thomas Jaeger <thjaeger at gmail.com> wrote:
> This is not a healthy discussion.  We have people claiming that they
> can't live without C-A-B, yet they're unable to come up with any
> *concrete* situations where they need it.  I don't doubt that these
> issues exist, but my guess is that in most of those cases, C-A-B is the
> wrong way to go about it.  To make any progress here, we need more data,
> it's as simple as that.  Otherwise it's impossible to make any sort of
> generalizations as to why these situations happen.  For example, if it
> turned out that most of those lock-ups are due to grabs gone awry, it
> wouldn't be top difficult to implement and advertise a Release-All-Grabs
> key, that would take care of the issue without the need for killing X.
>

I have a solution for you:  X is still alive enough to catch C-A-B,
have it request a core dump from the kernel.  A vmem dump and
vregister dump would also be helpful but we can't reliably enter video
memory or play with video registers without triggering a
hardware-state issue and locking on a request to said hardware.