linux-kernel - [PATCH v2 00/17] Support for read-write lock deadlock detection

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20190516080015.16033-1-duyuyang@gmail.com>
Date:   Thu, 16 May 2019 15:59:58 +0800
From:   Yuyang Du <duyuyang@...il.com>
To:     peterz@...radead.org, will.deacon@....com, mingo@...nel.org
Cc:     bvanassche@....org, ming.lei@...hat.com, frederic@...nel.org,
        tglx@...utronix.de, boqun.feng@...il.com, paulmck@...ux.ibm.com,
        linux-kernel@...r.kernel.org, Yuyang Du <duyuyang@...il.com>
Subject: [PATCH v2 00/17] Support for read-write lock deadlock detection

Hi Peter and Ingo,

Historically, the read-write lock (recursive-read locks included) is not
well supported in lockdep. This patchset attempts to solve this problem
sound and complete.

The bulk of the algorithm is in patch #10, which is actually not complex at
all. Hopefully, it simply works.

Now that we have read-write locks suppported, we have all the 262 cases
passed, though I have to flip some cases which, I think, are wrong.

P.S. To Boqun, I haven't got time to read your patchset except that I did
carefully read your design doc and learnt from it a lot. It is helpful.
Please give this patchset at least a look.

Change from v1:

 - As Peter pointed out the assumption that a lock class cannot ever be read
   and recursive read does not hold. The value of LOCK_TYPE_RECURSIVE is
   changed to 3 from 2, which maintains that if type1 and type2 are exclusive
   then type1 & type2 == 0. And since it is legal for a task can have a read
   and then a recursive read lock, we allow this in deadlock check.

 - Some typo fixes and rewording in comments or changelog.


Thanks,
Yuyang

--

Yuyang Du (17):
  locking/lockdep: Add lock type enum to explicitly specify read or
    write locks
  locking/lockdep: Add read-write type for dependency
  locking/lockdep: Add helper functions to operate on the searched path
  locking/lockdep: Update direct dependency's read-write type if it
    exists
  locking/lockdep: Rename deadlock check functions
  locking/lockdep: Adjust BFS algorithm to support multiple matches
  locking/lockdep: Introduce mark_lock_unaccessed()
  locking/lockdep: Introduce chain_hlocks_type for held lock's
    read-write type
  locking/lockdep: Hash held lock's read-write type into chain key
  locking/lockdep: Support read-write lock's deadlock detection
  locking/lockdep: Adjust lockdep selftest cases
  locking/lockdep: Remove useless lock type assignment
  locking/lockdep: Add nest lock type
  locking/lockdep: Support recursive read locks
  locking/lockdep: Adjust selftest case for recursive read lock
  locking/lockdep: Add more lockdep selftest cases
  locking/lockdep: Remove irq-safe to irq-unsafe read check

 include/linux/lockdep.h            |   57 +-
 include/linux/rcupdate.h           |    2 +-
 kernel/locking/lockdep.c           |  464 +++++++++++----
 kernel/locking/lockdep_internals.h |    4 +
 lib/locking-selftest.c             | 1099 +++++++++++++++++++++++++++++++++++-
 5 files changed, 1489 insertions(+), 137 deletions(-)

-- 
1.8.3.1