dm-crypt: Fix parsing of extended IV arguments.
PATCH 1/1 dm-crypt: Fix parsing of extended IV arguments.
The dm-crypt cipher specification in a mapping table is defined as
cipher[:keycount]-chainmode-ivmode[:ivopts] or with the new crypt API format capi:cipher_api_spec-ivmode[:ivopts].
For ESSIV, the parameter includes hash specification, for example aes-cbc-essiv:sha256.
The implementation expected that additional IV option never includes another dash '-' character
Unfortunately, with SHA3, we have now names like sha3-256, so the mapping table parser fails:
dmsetup create test --table "0 8 crypt aes-cbc-essiv:sha3-256 9c1185a5c5e9fc54612808977ee8f5b9e 0 /dev/sdb 0"
or (new format)
dmsetup create test --table "0 8 crypt capi:cbc(aes)-essiv:sha3-256 9c1185a5c5e9fc54612808977ee8f5b9e 0 /dev/sdb 0"
device-mapper: crypt: Ignoring unexpected additional cipher options
device-mapper: table: 253:0: crypt: Error creating IV
device-mapper: ioctl: error adding target to table
This patch fixes the dm-crypt constructor to ignore additional dash in IV options and also removes bogus warning (that is ignored anyway).
[This patch should go into stable tree as well.]
--- drivers/md/dm-crypt.c | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-)
drivers/md/dm-crypt.c +17 -8
@@ -2414,9 +2414,21 @@ static int crypt_ctr_cipher_new(struct dm_target *ti, char *cipher_in, char *key* capi:cipher_api_spec-iv:ivopts */ tmp = &cipher_in[strlen ("capi:" )]; - cipher_api =strsep (&tmp,"-" ); - *ivmode =strsep (&tmp,":" ); - *ivopts = tmp; + +/* Separate IV options if present, it can contain another '-' in hash name */ + *ivopts =strrchr (tmp,':' ); +if (*ivopts) { + **ivopts ='\0' ; + (*ivopts)++; + } +/* Parse IV mode */ + *ivmode =strrchr (tmp,'-' ); +if (*ivmode) { + **ivmode ='\0' ; + (*ivmode)++; + } +/* The rest is crypto API spec */ + cipher_api = tmp;if (*ivmode && !strcmp (*ivmode,"lmk" )) cc->tfms_count =64 ;
@@ -2486,11 +2498,8 @@ static int crypt_ctr_cipher_old(struct dm_target *ti, char *cipher_in, char *key goto bad_mem; chainmode =strsep (&tmp,"-" ); - *ivopts =strsep (&tmp,"-" ); - *ivmode =strsep (&*ivopts,":" ); - -if (tmp) -DMWARN ("Ignoring unexpected additional cipher options" ); + *ivmode =strsep (&tmp,":" ); + *ivopts = tmp;/* * For compatibility with the original dm-crypt mapping format, if
Signature · 1 line
2.20.1
Quoted 14 lines · Milan Broz
The dm-crypt cipher specification in a mapping table is defined as
cipher[:keycount]-chainmode-ivmode[:ivopts] or with the new crypt API format capi:cipher_api_spec-ivmode[:ivopts].For ESSIV, the parameter includes hash specification, for example aes-cbc-essiv:sha256.
The implementation expected that additional IV option never includes another dash '-' character
Unfortunately, with SHA3, we have now names like sha3-256, so the mapping table parser fails:
dmsetup create test --table "0 8 crypt aes-cbc-essiv:sha3-256
9c1185a5c5e9fc54612808977ee8f5b9e0/dev/sdb0"
or (new format)
dmsetup create test --table "0 8 crypt capi:cbc(aes)-essiv:sha3-2569c1185a5c5e9fc54612808977ee8f5b9e0/dev/sdb0"
Quoted 8 lines
device-mapper: crypt: Ignoring unexpected additional cipher options
device-mapper: table: 253:0: crypt: Error creating IV
device-mapper: ioctl: error adding target to tableThis patch fixes the dm-crypt constructor to ignore additional dash in IV options and also removes bogus warning (that is ignored anyway).
[This patch should go into stable tree as well.]
Rather than this it'd be useful to just be more explicit, e.g.:
Once I know which commit exposed us to this problem I can take care of getting this fix staged for 5.0-rcX inclussion.
Thanks,
Mike
Quoted 9 lines · Mike Snitzer
Quoted 4 lines · Milan Broz
This patch fixes the dm-crypt constructor to ignore additional dash in IV options and also removes bogus warning (that is ignored anyway).
[This patch should go into stable tree as well.]
Rather than this it'd be useful to just be more explicit, e.g.:
There is no single reference commit.
The new capi: was introduced in
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/md/dm-crypt.c?id=33d2f09fcb357fd1861c4959d1d3505492bf91f8
but I think the second part applies even to older kernel, but not sure it makes sense to fix it there (I guess not)...
So perhaps the best is to apply it to stable kernel >= 4.12
Milan
Quoted 11 lines · Milan Broz
Quoted 9 lines · Mike Snitzer
Quoted 4 lines · Milan Broz
This patch fixes the dm-crypt constructor to ignore additional dash in IV options and also removes bogus warning (that is ignored anyway).
[This patch should go into stable tree as well.]
Rather than this it'd be useful to just be more explicit, e.g.:
Quoted 8 lines
There is no single reference commit.
The new capi: was introduced in
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/md/dm-crypt.c?id=33d2f09fcb357fd1861c4959d1d3505492bf91f8but I think the second part applies even to older kernel, but not sure it makes sense to fix it there (I guess not)...
So perhaps the best is to apply it to stable kernel >= 4.12
OK, now staged, I tweaked your header though because (like your email reply) it was lacking proper wrapping.
Thanks,
Mike