linux-kernel - [PATCH 0/5] Additional radix tree patches for 4.10

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]

Date:   Tue, 13 Dec 2016 14:21:27 -0800
From:   Matthew Wilcox <mawilcox@...uxonhyperv.com>
To:     linux-kernel@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>
Cc:     Matthew Wilcox <mawilcox@...rosoft.com>, linux-mm@...ck.org,
        linux-fsdevel@...r.kernel.org,
        Konstantin Khlebnikov <koct9i@...il.com>,
        Ross Zwisler <ross.zwisler@...ux.intel.com>,
        "Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>,
        Tejun Heo <tj@...nel.org>
Subject: [PATCH 0/5] Additional radix tree patches for 4.10

From: Matthew Wilcox <mawilcox@...rosoft.com>

Hi Andrew,

One bug has popped up in testing which is reasonably hard to hit (fixed
by patch 2 in this series).  I needed to change the test suite to be
able to catch the bug in action.  The test suite was returning freed
memory to glibc's malloc, but we have to keep a cache of objects in
order to notice that we freed something non-zero and then neglected to
initialise it back to zero upon reallocation.

Please drop the existing "reimplement IDR and IDA using the radix tree"
patch from your queue and replace it with these five patches.  I'd like
to see the first four patches head to Linus along with the rest of the
radix tree patches, and we can continue with the plan of holding patch 5
until 4.11.

I took your suggestion of moving as much as possible of the IDA and
IDR functionality back to idr.c.  That prompted creation of some nicer
APIs, such as radix_tree_iter_delete() and radix_tree_iter_lookup().

There are also some good bugs fixed in this revision, such as:
 - returning the correct error value from ida_get_new_above() if we run
   out of memory,
 - handling attempts to allocate from a full IDA correctly.
 - Correctly handle deleting the only entry at root with
   radix_tree_iter_delete()
I wrote test cases for all these situations, so they shouldn't
regress again.

Matthew Wilcox (5):
  radix tree test suite: Cache recently freed objects
  radix-tree: Ensure counts are initialised
  radix tree test suite: Add new tag check
  radix tree test suite: Delete unused rcupdate.c
  Reimplement IDR and IDA using the radix tree

 include/linux/idr.h                     |  138 ++--
 include/linux/radix-tree.h              |   53 +-
 init/main.c                             |    3 +-
 lib/idr.c                               | 1144 ++++++-------------------------
 lib/radix-tree.c                        |  358 ++++++++--
 tools/include/linux/spinlock.h          |    4 +
 tools/testing/radix-tree/.gitignore     |    1 +
 tools/testing/radix-tree/Makefile       |   10 +-
 tools/testing/radix-tree/idr-test.c     |  200 ++++++
 tools/testing/radix-tree/linux.c        |   48 +-
 tools/testing/radix-tree/linux/export.h |    1 +
 tools/testing/radix-tree/linux/gfp.h    |    8 +-
 tools/testing/radix-tree/linux/idr.h    |    1 +
 tools/testing/radix-tree/linux/kernel.h |    2 +
 tools/testing/radix-tree/linux/slab.h   |    5 -
 tools/testing/radix-tree/main.c         |    6 +
 tools/testing/radix-tree/multiorder.c   |   45 +-
 tools/testing/radix-tree/rcupdate.c     |   86 ---
 tools/testing/radix-tree/tag_check.c    |    3 +
 tools/testing/radix-tree/test.h         |    2 +
 20 files changed, 914 insertions(+), 1204 deletions(-)
 create mode 100644 tools/include/linux/spinlock.h
 create mode 100644 tools/testing/radix-tree/idr-test.c
 create mode 100644 tools/testing/radix-tree/linux/idr.h
 delete mode 100644 tools/testing/radix-tree/rcupdate.c

-- 
2.10.2