[Pkg-cryptsetup-devel] Bug#430158: Bug#430158: closed by Jonas Meurer <jonas at freesources.org> (closing because of inactivity)

[Helmut Grohne]

Hi,

> Ok, now I got the problem. You're correct, that the tries option doesn't
> work for the cryptsetup version in debian/stable (2:1.0.4+svn26-1).
> That's due to a mistake in upstream development. This has been fixed in
> the cryptsetup package version 2:1.0.4+svn26-2, but unfortunately this
> didn't make it into etch. Please see bugs #414326 and #412064 for
> further information.

I was indeed trying using an etch system, sorry about that. I though
this was ok, since /usr/share/initramfs-tools/scrips/local-top/cryptroot
were roughly the same.

> As you can see, this bug has been fixed for a long time, so I'm closing
> the bugreport again. Unfortunately, this will never make it into etch,
> which is due to debians stable release policy. If you insist on this
> feature, I could prepare a backport of the current cryptsetup package
> version 2:1.0.6~pre1+svn46-1 for backports.org. What do you think about
> that?

I normally don't use etch, so could you be so kind to explain me how you
fixed it? I really don't see this:

My unstable system has 2:1.0.6~pre1+svn45-1, so this should have the fix
included. The script for activating cryptsetup is the one mentioned
above.

On line 173 I can see:
	cryptcreate="/sbin/cryptsetup luksOpen $cryptsource $crypttarget"

$cryptsource is tested for -e, so this will not include --tries and
$crypttarget is similarly tested for -e /dev/mapper/$crypttarget, so
this command does not include a --tries.

Later in line 198 (I don't run a keyscript not bootsplash):
	$cryptcreate < /dev/console > /dev/console 2>&1

The command is just verbatim executed, no --tries.

The whole script doesn't even contain -T nor --tries, so how do the
number of tries get into that script? I really don't see that, sorry.

Helmut