[dm-crypt] [ANNOUNCE] cryptsetup 1.1.1-rc1 (test release candidate)

[dm-crypt] [ANNOUNCE] cryptsetup 1.1.1-rc1 (test release candidate)

[Milan Broz]

Cryptsetup 1.1.1-rc1 Release Notes
==================================

The first cryptsetup 1.1.1-rc1 release candidate is available at

   http://code.google.com/p/cryptsetup/

Feedback and bug reports are welcomed.


Changes since version 1.1.0

Important changes:
~~~~~~~~~~~~~~~~~~

* Detects and use device-mapper udev support if available.

  This should allow synchronisation with udev rules and avoid races with udev.

  If package maintainer want to use old, direct libdevmapper device node creation,
  use configure option --disable-udev.

* Supports device topology detection for data alignment.

  If kernel provides device topology ioctl calls, the LUKS data area
  alignment is automatically set to optimal value.

  This means that stacked devices (like LUKS over MD/LVM)
  should use the most optimal data alignment.

  (You can still overwrite this calculation using --align-payload option.)

* Prefers some device paths in status display.
  (So status command will try to find top level device name, like /dev/sdb.)

* Fix package config file to use proper package version.

Other changes:
~~~~~~~~~~~~~~
* Fix luksOpen reading of passphrase on stdin (if "-" keyfile specified).
* Fix isLuks to initialise crypto backend (blkid instead is suggested anyway).
* Properly initialise crypto backend in header backup/restore commands.
* Do not verify unlocking passphrase in luksAddKey command.
* Allow no hash specification in plain device constructor - user can provide volume key directly.
* Try to use pkgconfig for device mapper library in configuration script.
* Add some compatibility checks and disable LUKS suspend/resume if not supported.
* Rearrange tests, "make check" now run all available test for package.
* Avoid class C++ keyword in library header.


Milan

Re: [dm-crypt] [ANNOUNCE] cryptsetup 1.1.1-rc1 (test release candidate)

[Milan Broz]

On 05/01/2010 07:13 PM, Milan Broz wrote:
> Cryptsetup 1.1.1-rc1 Release Notes
> ==================================
> 
> The first cryptsetup 1.1.1-rc1 release candidate is available at
> 
>    http://code.google.com/p/cryptsetup/
> 
> Feedback and bug reports are welcomed.

And one of the bug is that it no longer automatically loads dm_crypt module,
if not present.

I'll fix it and release new rc soon.

Milan

[dm-crypt] [ANNOUNCE] cryptsetup 1.1.1-rc2 (test release candidate)

[Milan Broz]

Cryptsetup 1.1.1-rc2 Release Notes
==================================
 
The second cryptsetup 1.1.1-rc2 release candidate is available at

   http://code.google.com/p/cryptsetup/

Feedback and bug reports are welcomed.


Changes since version 1.1.1-rc1
-------------------------------
The only important change is fix for automatic dm-crypt module loading regression.

Milan

Re: [dm-crypt] [ANNOUNCE] cryptsetup 1.1.1-rc1 (test release candidate)

[Thomas Bächler]

[-- Attachment #1: Type: text/plain, Size: 1162 bytes --]

Am 01.05.2010 19:13, schrieb Milan Broz:
> Important changes:
> ~~~~~~~~~~~~~~~~~~
> 
> * Detects and use device-mapper udev support if available.
> 
>   This should allow synchronisation with udev rules and avoid races with udev.
> 
>   If package maintainer want to use old, direct libdevmapper device node creation,
>   use configure option --disable-udev.

Awesome. This should eliminate some problems we've been having.

> * Supports device topology detection for data alignment.
> 
>   If kernel provides device topology ioctl calls, the LUKS data area
>   alignment is automatically set to optimal value.
> 
>   This means that stacked devices (like LUKS over MD/LVM)
>   should use the most optimal data alignment.
> 
>   (You can still overwrite this calculation using --align-payload option.)

Good news.

> * Rearrange tests, "make check" now run all available test for package.

At least some of these tests seem to require root privileges, so won't
work when building as user/in fakeroot.

Anyway, the package seems to work fine so far, but I only did a few
luksOpen calls, nothing more. Thanks for the great work.


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 262 bytes --]