linux-security-module.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v6 0/2] KEYS: Read keys to internal buffer & then copy to userspace
@ 2020-03-20 19:19 Waiman Long
  2020-03-20 19:19 ` [PATCH v6 1/2] KEYS: Don't write out to userspace while holding key semaphore Waiman Long
  2020-03-20 19:19 ` [PATCH v6 2/2] KEYS: Avoid false positive ENOMEM error on key read Waiman Long
  0 siblings, 2 replies; 6+ messages in thread
From: Waiman Long @ 2020-03-20 19:19 UTC (permalink / raw)
  To: David Howells, Jarkko Sakkinen, James Morris, Serge E. Hallyn,
	Mimi Zohar, David S. Miller, Jakub Kicinski
  Cc: keyrings, linux-kernel, linux-security-module, linux-integrity,
	netdev, linux-afs, Sumit Garg, Jerry Snitselaar, Roberto Sassu,
	Eric Biggers, Chris von Recklinghausen, Waiman Long

v6:
 - Make some variable name changes and revise comments as suggested by
   Jarkko. No functional change from v5.

v5:
 - Merge v4 patches 2 and 3 into 1 to avoid sparse warning. Merge some of 
   commit logs into patch 1 as well. There is no further change.

v4:
 - Remove the __user annotation from big_key_read() and user_read() in
   patch 1.
 - Add a new patch 2 to remove __user annotation from rxrpc_read().
 - Add a new patch 3 to remove __user annotation from dns_resolver_read().
 - Merge the original patches 2 and 3 into a single patch 4 and refactor
   it as suggested by Jarkko and Eric.

v3:
 - Reorganize the keyctl_read_key() code to make it more readable as
   suggested by Jarkko Sakkinen.
 - Add patch 3 to use kvmalloc() for safer large buffer allocation as
   suggested by David Howells.

v2:
 - Handle NULL buffer and buflen properly in patch 1.
 - Fix a bug in big_key.c.
 - Add patch 2 to handle arbitrary large user-supplied buflen.

The current security key read methods are called with the key semaphore
held.  The methods then copy out the key data to userspace which is
subjected to page fault and may acquire the mmap semaphore. That can
result in circular lock dependency and hence a chance to get into
deadlock.

To avoid such a deadlock, an internal buffer is now allocated for getting
out the necessary data first. After releasing the key semaphore, the
key data are then copied out to userspace sidestepping the circular
lock dependency.

The keyutils test suite was run and the test passed with these patchset
applied without any falure.

Waiman Long (2):
  KEYS: Don't write out to userspace while holding key semaphore
  KEYS: Avoid false positive ENOMEM error on key read

 include/keys/big_key-type.h               |  2 +-
 include/keys/user-type.h                  |  3 +-
 include/linux/key-type.h                  |  2 +-
 net/dns_resolver/dns_key.c                |  2 +-
 net/rxrpc/key.c                           | 27 +++-----
 security/keys/big_key.c                   | 11 ++-
 security/keys/encrypted-keys/encrypted.c  |  7 +-
 security/keys/internal.h                  | 12 ++++
 security/keys/keyctl.c                    | 84 ++++++++++++++++++++---
 security/keys/keyring.c                   |  6 +-
 security/keys/request_key_auth.c          |  7 +-
 security/keys/trusted-keys/trusted_tpm1.c | 14 +---
 security/keys/user_defined.c              |  5 +-
 13 files changed, 114 insertions(+), 68 deletions(-)

-- 
2.18.1


^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2020-03-21  1:51 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-03-20 19:19 [PATCH v6 0/2] KEYS: Read keys to internal buffer & then copy to userspace Waiman Long
2020-03-20 19:19 ` [PATCH v6 1/2] KEYS: Don't write out to userspace while holding key semaphore Waiman Long
2020-03-20 19:19 ` [PATCH v6 2/2] KEYS: Avoid false positive ENOMEM error on key read Waiman Long
2020-03-20 22:19   ` Jarkko Sakkinen
2020-03-20 22:21     ` Jarkko Sakkinen
2020-03-21  1:51     ` Waiman Long

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).