latchset/luksmeta

luksmeta's Issues

Tool failed to recognize the state of a LUKS1 volume

OS: Fedora; Component: luksmeta.x86_64 9-5.fc31 @fedora

Hey. Tool failed to recognize the state of a LUKS1 volume. Surprisingly a tool dedicated to LUKS v2 volumes succeeded in that task.

# lsblk -o name,fstype,type,mode,mountpoint
NAME                                          FSTYPE      TYPE  MODE       MOUNTPOINT
[...]
└─sda6                                        crypto_LUKS part  brw-rw---- 
  └─luks-acf8855f-302b-4ea5-90f6-12a256deb1fa LVM2_member crypt brw-rw---- 
# cryptsetup luksDump /dev/sda6
[...]
Version:       	1
[...]
Key Slot 0: ENABLED
[...]
# luksmeta show -d /dev/sda6
Device is not initialized (/dev/sda6)

man page - luksmeta exit code

The current man page lists the textual error messages and refers to "sysexit.h" for the actual error code.
Usability of the information would be higher if the numeric error code was included in the man page instead of asking the user to look at a source header file.

0 /* successful termination /
64 / base value for error messages /
64 / command line usage error /
65 / data format error /
66 / cannot open input /
67 / addressee unknown /
68 / host name unknown /
69 / service unavailable /
70 / internal software error /
71 / system error (e.g., can't fork) /
72 / critical OS file missing /
73 / can't create (user) output file /
74 / input/output error /
75 / temp failure; user is invited to retry /
76 / remote error in protocol /
77 / permission denied /
78 / configuration error /
/ maximum listed value */

https://unix.stackexchange.com/questions/110348/how-do-i-get-the-list-of-exit-codes-and-or-return-codes-and-meaning-for-a-comm/110350

Fix manpage section

Severity: minor

Hello,

Debian "lintian" checker reported a mismatch in the provided luksmeta.8 manpage.
Possible fix:

--- a/luksmeta.8
+++ b/luksmeta.8
@@ -1,7 +1,7 @@
." generated with Ronn/v0.7.3
." http://github.com/rtomayko/ronn/tree/0.7.3
.
-.TH "LUKSMETA" "1" "June 2017" "" ""
+.TH "LUKSMETA" "8" "June 2017" "" ""
.
.SH "NAME"
\fBluksmeta\fR - Utility for storing metadata in a LUKSv1 header

typos in the manpage

Severity: minor

Hello,

Debian's lintian check reported two typos in the manpage:

-\fBEX_NOINPUT\fR : An error occured while reading from standard input.
+\fBEX_NOINPUT\fR : An error occurred while reading from standard input.

-Destroy all data (including LUKSMeta data) in the LUKSv1 header gap and initalize the gap for LUKSMeta storage:
+Destroy all data (including LUKSMeta data) in the LUKSv1 header gap and initialize the gap for LUKSMeta storage:

[patch] test-luksmeta breakage "Not compatible PBKDF options."

In Debian, luksmeta started to fail the test suite recently, error message is "Not compatible PBKDF options." Turns out creating a container breaks some constraints when running as non-root. Workaround is to force the v1 format using "--type luks1".

Full patch is at https://sources.debian.org/patches/luksmeta/9-3/fix.test-luksmeta.patch/
See also Debian bug report at https://bugs.debian.org/923513

Assertion "test_layout" failing with recent versions of cryptsetup

When trying to build LUKSMeta with cryptsetup 2.0.5 or 2.0.6 on Arch Linux, the compilation works fine, but the test suite fails with errors of the form

Assertion `test_layout((range_t[]) { ... })' failed.

Click to see the full test-suite.log

==================================
   luksmeta 9: ./test-suite.log
==================================

# TOTAL: 7
# PASS:  2
# SKIP:  0
# XFAIL: 0
# FAIL:  5
# XPASS: 0
# ERROR: 0

.. contents:: :depth: 2

FAIL: test-lm-assumptions
=========================

00000000:00001024 (!= 0)
00001024:04194304 (== 0)
unexpected nonzero: 3072
lt-test-lm-assumptions: test-lm-assumptions.c:32: main: Assertion `test_layout((range_t[]) { { 0, 1024 }, END(1024), })' failed.
FAIL test-lm-assumptions (exit status: 134)

FAIL: test-lm-init
==================

00000000:00001024 (!= 0)
00001024:01052672 (== 0)
unexpected nonzero: 3072
lt-test-lm-init: test-lm-init.c:63: main: Assertion `test_layout((range_t[]) { { 0, 1024 }, { 1024, offset - 1024, true }, { offset, 4096 }, END(offset + 4096), })' failed.
FAIL test-lm-init (exit status: 134)

FAIL: test-lm-one
=================

00000000:00001024 (!= 0)
00001024:01052672 (== 0)
unexpected nonzero: 3072
lt-test-lm-one: test-lm-one.c:56: main: Assertion `test_layout((range_t[]) { { 0, 1024 }, { 1024, offset - 1024, true }, { offset, 4096 }, { offset + 4096, 4096 }, END(offset + 8192), })' failed.
FAIL test-lm-one (exit status: 134)

FAIL: test-lm-two
=================

00000000:00001024 (!= 0)
00001024:01052672 (== 0)
unexpected nonzero: 3072
lt-test-lm-two: test-lm-two.c:60: main: Assertion `test_layout((range_t[]) { { 0, 1024 }, { 1024, offset - 1024, true }, { offset, 4096 }, { offset + 4096, 4096 }, END(offset + 8192), })' failed.
FAIL test-lm-two (exit status: 134)

FAIL: test-lm-big
=================

00000000:00001024 (!= 0)
00001024:01052672 (== 0)
unexpected nonzero: 3072
lt-test-lm-big: test-lm-big.c:119: main: Assertion `test_layout((range_t[]) { { 0, 1024 }, { 1024, offset - 1024, true }, { offset, 4096 }, { offset + 4096, 4096 }, { offset + 8192, 4096 }, END(offset + 12288), })' failed.
FAIL test-lm-big (exit status: 134)

 
Downgrading to crypsetup 2.0.4 solves the problem, so the change seems to have been introduced somewhere between version 2.0.4 and 2.0.5. I did not dig deeper, but the upstream commit c2bce3e9 (Wipe full header areas (including unused) during LUKS format.) looks promising.