Package "postgresql-16"

  postgresql-16 (16.14-0ubuntu0.24.04.1) noble-security; urgency=medium

  * New upstream version (LP: #2152636).

    + A dump/restore is not required for those running 16.X.

    + However, if you are upgrading from a version earlier than 16.10, see
      those release notes as well please.

    + Prevent unbounded recursion while processing startup packets

      A malicious client could crash the connected backend by alternating
      rejected SSL and GSS encryption requests indefinitely.

      The PostgreSQL Project thanks Calif.io (in collaboration with Claude and
      Anthropic Research) for reporting this problem. (CVE-2026-6479)

    + Fix assorted integer overflows in memory-allocation calculations

      Various places were incautious about the possibility of integer overflow
      in calculations of how much memory to allocate. Overflow would lead to
      allocating a too-small buffer which the caller would then write past the
      end of. This would at least trigger server crashes, and probably could
      be exploited for arbitrary code execution. In many but by no means all
      cases, the hazard exists only in 32-bit builds. (CVE-2026-6473)

    + Properly quote object names in logical replication origin checks

      ALTER SUBSCRIPTION ... REFRESH PUBLICATION interpolated schema and
      relation names into SQL commands without quoting them, allowing
      execution of arbitrary SQL on the publisher. (CVE-2026-6638)

    + Reject over-length options in ts_headline()

      The StartSel, StopSel and FragmentDelimiter strings must not exceed 32Kb
      in length, but this was not checked for. An over-length value would
      typically crash the server. (CVE-2026-6473)

    + Guard against malicious time zone names in timeofday() and pg_strftime()

      A crafted time zone setting could pass % sequences to snprintf(),
      potentially causing crashes or disclosure of server memory. Another path
      to similar results was to overflow the limited-size output buffer used
      by pg_strftime(). (CVE-2026-6474)

    + When creating a multirange type, ensure the user has CREATE privilege on
      the schema specified for the multirange type

      The multirange type can be put into a different schema than its parent
      range type, but we neglected to apply the required privilege check when
      doing so. (CVE-2026-6472)

    + Use timing-safe string comparisons in authentication code

    + Use timingsafe_bcmp() instead of memcpy() or strcmp() when checking
      passwords, hashes, etc. It is not known whether the data dependency of
      those functions is usefully exploitable in any of these places, but in
      the interests of safety, replace them. (CVE-2026-6478)

    + Mark PQfn() as unsafe, and avoid using it within libpq

      For a non-integral result type, PQfn() is not passed the size of the
      output buffer, so it cannot check that the data returned by the server
      will fit. A malicious server could therefore overwrite client memory.
      This is unfixable without an API change, so mark the function as
      deprecated. Internally to libpq, use a variant version that can apply
      the missing check. (CVE-2026-6477)

    + Prevent path traversal in pg_basebackup and pg_rewind

      These applications failed to validate output file paths read from their
      input, so that a malicious source could overwrite any file writable by
      these applications. Constrain where data can be written by rejecting
      paths that are absolute or contain parent-directory references.
      (CVE-2026-6475)

    + Guard against field overflow within contrib/intarray's query_int type
      and contrib/ltree's ltxtquery type

      Parsing of these query structures did not check for overflow of 16-bit
      fields, so that construction of an invalid query tree was possible. This
      can crash the server when executing the query. (CVE-2026-6473)

    + Guard against overly long values of contrib/ltree's lquery type

      Values with more than 64K items caused internal overflows, potentially
      resulting in stack smashes or wrong answers. (CVE-2026-6473)

    + Prevent SQL injection and buffer overruns in contrib/spi

      check_foreign_key() was insufficiently careful about quoting key values,
      and also used fixed-length buffers for constructing queries. While this
      module is only meant as example code, it still shouldn't contain such
      dangerous errors. (CVE-2026-6637)

    + Details about these and many further changes can be found at:
      https://www.postgresql.org/docs/16/release-16-14.html.

  * d/postgresql-16.NEWS: update NEWS file.

 -- Athos Ribeiro <email address hidden> Fri, 15 May 2026 09:22:56 -0300

  postgresql-16 (16.13-0ubuntu0.24.04.1) noble-security; urgency=medium

  * New upstream version (LP: #2127668).

    + This release encompasses changes from upstream's 16.12 and 16.13
      releases. The former contains fixes for 4 CVEs (among other things), and
      the latter was a hotfix for regressions introcuced in 16.12.

    + A dump/restore is not required for those running 16.X.

    + However, if you are upgrading from a version earlier than 16.10, see
      those release notes as well please.

    + Guard against unexpected dimensions of oidvector/int2vector (Tom Lane)

      These data types are expected to be 1-dimensional arrays containing no
      nulls, but there are cast pathways that permit violating those
      expectations. Add checks to some functions that were depending on those
      expectations without verifying them, and could misbehave in consequence.
      (CVE-2026-2003)

    + Harden selectivity estimators against being attached to operators that
      accept unexpected data types (Tom Lane)

      contrib/intarray contained a selectivity estimation function that could
      be abused for arbitrary code execution, because it did not check that
      its input was of the expected data type. Third-party extensions should
      check for similar hazards and add defenses using the technique intarray
      now uses. Since such extension fixes will take time, we now require
      superuser privilege to attach a non-built-in selectivity estimator to an
      operator. (CVE-2026-2004)

    + Fix buffer overrun in contrib/pgcrypto's PGP decryption functions
      (Michael Paquier)

      Decrypting a crafted message with an overlength session key caused a
      buffer overrun, with consequences as bad as arbitrary code execution.
      (CVE-2026-2005)

    + Fix inadequate validation of multibyte character lengths (Thomas Munro,
      Noah Misch)

      Assorted bugs allowed an attacker able to issue crafted SQL to overrun
      string buffers, with consequences as bad as arbitrary code execution.
      After these fixes, applications may observe “invalid byte sequence for
      encoding” errors when string functions process invalid text that has
      been stored in the database. (CVE-2026-2006)

    + Details about these and many further changes can be found at:
      https://www.postgresql.org/docs/16/release-16-12.html and
      https://www.postgresql.org/docs/16/release-16-13.html.

  * d/postgresql-16.NEWS: Update.

 -- Athos Ribeiro <email address hidden> Wed, 25 Feb 2026 11:12:24 -0300

  postgresql-16 (16.11-0ubuntu0.24.04.1) noble-security; urgency=medium

  * New upstream version (LP: #2127667).

    + A dump/restore is not required for those running 16.X.

    + However, if you are upgrading from a version earlier than 16.10, see
      those release notes as well please.

    + Check for CREATE privileges on the schema in CREATE STATISTICS (Jelte
      Fennema-Nio)

      This omission allowed table owners to create statistics in any schema,
      potentially leading to unexpected naming conflicts. (CVE-2025-12817)

    + Avoid integer overflow in allocation-size calculations within libpq
      (Jacob Champion)

      Several places in libpq were not sufficiently careful about computing
      the required size of a memory allocation. Sufficiently large inputs
      could cause integer overflow, resulting in an undersized buffer, which
      would then lead to writing past the end of the buffer. (CVE-2025-12818)

    + Details about these and many further changes can be found at:
      https://www.postgresql.org/docs/16/release-16-11.html.

  * d/postgresql-16.NEWS: Update NEWS file.

 -- Athos Ribeiro <email address hidden> Mon, 24 Nov 2025 09:58:56 -0300

  postgresql-16 (16.10-0ubuntu0.24.04.1) noble-security; urgency=medium

  * New upstream version (LP: #2112531).

    + A dump/restore is not required for those running 16.X.

    + However, if you have any BRIN numeric_minmax_multi_ops indexes, it is
      advisable to reindex them after updating. Keep reading for further
      information.

    + Also, if you are upgrading from a version earlier than 16.9, see those
      release notes as well please.

    + Tighten security checks in planner estimation functions (Dean Rasheed)

      The fix for CVE-2017-7484, plus followup fixes, intended to prevent
      leaky functions from being applied to statistics data for columns that
      the calling user does not have permission to read. Two gaps in that
      protection have been found. One gap applies to partitioning and
      inheritance hierarchies where RLS policies on the tables should restrict
      access to statistics data, but did not.

      The other gap applies to cases where the query accesses a table via a
      view, and the view owner has permissions to read the underlying table
      but the calling user does not have permissions on the view. The view
      owner's permissions satisfied the security checks, and the leaky
      function would get applied to the underlying table's statistics before
      we check the calling user's permissions on the view. This has been fixed
      by making security checks on views occur at the start of planning. That
      might cause permissions failures to occur earlier than before.
      (CVE-2025-8713)

    + Prevent pg_dump scripts from being used to attack the user running the
      restore (Nathan Bossart)

      Since dump/restore operations typically involve running SQL commands as
      superuser, the target database installation must trust the source
      server. However, it does not follow that the operating system user who
      executes psql to perform the restore should have to trust the source
      server. The risk here is that an attacker who has gained superuser-level
      control over the source server might be able to cause it to emit text
      that would be interpreted as psql meta-commands. That would provide
      shell-level access to the restoring user's own account, independently of
      access to the target database.

      To provide a positive guarantee that this can't happen, extend psql with
      a \restrict command that prevents execution of further meta-commands,
      and teach pg_dump to issue that before any data coming from the source
      server. (CVE-2025-8714)

    + Convert newlines to spaces in names included in comments in pg_dump
      output (Noah Misch)

      Object names containing newlines offered the ability to inject arbitrary
      SQL commands into the output script. (Without the preceding fix,
      injection of psql meta-commands would also be possible this way.)
      CVE-2012-0868 fixed this class of problem at the time, but later work
      reintroduced several cases. (CVE-2025-8715)

    + Fix incorrect distance calculation in BRIN numeric_minmax_multi_ops
      support function (Peter Eisentraut, Tom Lane)

      The results were sometimes wrong on 64-bit platforms, and wildly wrong
      on 32-bit platforms. This did not produce obvious failures because the
      logic is only used to choose how to merge values into ranges; at worst
      the index would become inefficient and bloated. Nonetheless it's
      recommended to reindex any BRIN indexes that use the
      numeric_minmax_multi_ops operator class.

    + Details about these and many further changes can be found at:
      https://www.postgresql.org/docs/16/release-16-10.html.

  * d/postgresql-16.NEWS: Update.

  * d/p/hurd-iovec: drop patch applied upstream.

 -- Athos Ribeiro <email address hidden> Mon, 25 Aug 2025 17:36:09 -0300