[dm-crypt] cryptsetup --iter-time default should be configurable (and reported in --help)

[dm-crypt] cryptsetup --iter-time default should be configurable (and reported in --help)

[Daniel Kahn Gillmor]

[-- Attachment #1.1: Type: text/plain, Size: 635 bytes --]

Hi folks--

I just noticed that the default for cryptsetup --iter-time isn't visible
in the output of cryptsetup --help.

I went looking to change this, and saw that the default is neither
configurable nor easily extracted.

The attached patch should make the default for this parameter
configurable (e.g. ./configure --with-luks1-iter-time=1000), as well as
reporting the compiled-in default in the output of --help.

I'm not currently subscribed to the dm-crypt mailing list; please Cc
me on replies.

Thanks very much for all your work on dm-crypt -- it's a valuable
component in the operating systems i use!

Regards,

     --dkg


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1.2: make --iter-time default configurable and documented --]
[-- Type: text/x-diff, Size: 2283 bytes --]

From fe6a079e055dc7b4ba2e14531582d82411d3d7b8 Mon Sep 17 00:00:00 2001
From: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Date: Tue, 20 Nov 2012 13:43:28 -0500
Subject: [PATCH] make default LUKS PBKDF2 iteration time configurable

---
 configure.in     |    1 +
 src/cryptsetup.c |   10 ++++++----
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/configure.in b/configure.in
index 6b7a79b..c937ab3 100644
--- a/configure.in
+++ b/configure.in
@@ -346,6 +346,7 @@ CS_STR_WITH([luks1-hash],   [hash function for LUKS1 header], [sha1])
 CS_STR_WITH([luks1-cipher], [cipher for LUKS1], [aes])
 CS_STR_WITH([luks1-mode],   [cipher mode for LUKS1], [cbc-essiv:sha256])
 CS_NUM_WITH([luks1-keybits],[key length in bits for LUKS1], [256])
+CS_NUM_WITH([luks1-iter-time],[PBKDF2 iteration time for LUKS1 (in msec)], [1000])
 
 CS_STR_WITH([loopaes-cipher], [cipher for loop-AES mode], [aes])
 CS_NUM_WITH([loopaes-keybits],[key length in bits for loop-AES mode], [256])
diff --git a/src/cryptsetup.c b/src/cryptsetup.c
index 0cc7f07..eae14ee 100644
--- a/src/cryptsetup.c
+++ b/src/cryptsetup.c
@@ -44,7 +44,7 @@ static uint64_t opt_offset = 0;
 static uint64_t opt_skip = 0;
 static int opt_skip_valid = 0;
 static int opt_readonly = 0;
-static int opt_iteration_time = 1000;
+static int opt_iteration_time = DEFAULT_LUKS1_ITER_TIME;
 static int opt_version_mode = 0;
 static int opt_timeout = 0;
 static int opt_tries = 3;
@@ -1165,10 +1165,12 @@ static void help(poptContext popt_context,
 			 "<key file> optional key file for the new key for luksAddKey action\n"),
 			crypt_get_dir());
 
-		log_std(_("\nDefault compiled-in keyfile parameters:\n"
+		log_std(_("\nDefault compiled-in key and passphrase parameters:\n"
 			 "\tMaximum keyfile size: %dkB, "
-			 "Maximum interactive passphrase length %d (characters)\n"),
-			 DEFAULT_KEYFILE_SIZE_MAXKB, DEFAULT_PASSPHRASE_SIZE_MAX);
+			 "Maximum interactive passphrase length %d (characters)\n"
+			 "Default PBKDF2 iteration time for LUKS: %d (msec)\n"),
+			 DEFAULT_KEYFILE_SIZE_MAXKB, DEFAULT_PASSPHRASE_SIZE_MAX,
+			 DEFAULT_LUKS1_ITER_TIME);
 
 		log_std(_("\nDefault compiled-in device cipher parameters:\n"
 			 "\tloop-AES: %s, Key %d bits\n"
-- 
1.7.10.4


[-- Attachment #2: Type: application/pgp-signature, Size: 965 bytes --]

Re: [dm-crypt] cryptsetup --iter-time default should be configurable (and reported in --help)

[Milan Broz]

On 11/20/2012 08:02 PM, Daniel Kahn Gillmor wrote:
> I just noticed that the default for cryptsetup --iter-time isn't visible
> in the output of cryptsetup --help.
> 
> I went looking to change this, and saw that the default is neither
> configurable nor easily extracted.
> 
> The attached patch should make the default for this parameter
> configurable (e.g. ./configure --with-luks1-iter-time=1000), as well as
> reporting the compiled-in default in the output of --help.

Hi,

patch applied to git (just with small change s/msec/ms).

But I hope distro maintainers will not decrease this default without
reading section 5.9 in
http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions#5.9_Security_Aspects

Thanks,
Milan