[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: [LMDB] Unexpected behavior for MDB_MULTIPLE with zero-length page

To: Bryan Matsuo <bryan.matsuo@gmail.com>
Subject: Re: [LMDB] Unexpected behavior for MDB_MULTIPLE with zero-length page
From: Hallvard Breien Furuseth <h.b.furuseth@usit.uio.no>
Date: Sun, 12 Feb 2017 08:56:04 +0100
Cc: "OpenLDAP, Technical" <openldap-technical@openldap.org>
In-reply-to: <CAJ_FXgpBz_jXkepvG2GTRO8WmYcH-tX5-b_JSDU3d8jiiO86bA@mail.gmail.com>
References: <CAJ_FXgpBz_jXkepvG2GTRO8WmYcH-tX5-b_JSDU3d8jiiO86bA@mail.gmail.com>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.7.0

On 12/02/17 05:19, Bryan Matsuo wrote:

I noticed that, for in a corner case of the mdb_cursor_put function when
MDB_MULTIPLE is provided, if the number of items is specified to be zero
this may lead to garbage data being written or a segfault. My intuition
would be that this is essentially a noop, but I also would have thought
EINVAL would be an acceptable error code


EINVAL.  If the key does not exist, you're asking LMDB to create a key
with no data.  That is not supported.  I'll file an ITS, with a few
other corner cases I noticed as well.

References:
- [LMDB] Unexpected behavior for MDB_MULTIPLE with zero-length page
  - From: Bryan Matsuo <bryan.matsuo@gmail.com>

Prev by Date: [LMDB] Unexpected behavior for MDB_MULTIPLE with zero-length page
Next by Date: Valgrind detecting memory leak on 2-way multi master setup on RHEL 6 - openldap 2.4.44
Index(es):
- Chronological
- Thread