DropRelFileNodeBuffers, DropDatabaseBuffers, FlushRelationBuffers, and
FlushDatabaseBuffers have to scan the whole shared_buffers pool because
we have no index structure that would find the target buffers any more
efficiently than that. This gets expensive with large NBuffers. We can
shave some cycles from these loops by prechecking to see if the current
buffer is interesting before we acquire the buffer header lock.
Ordinarily such a test would be unsafe, but in these cases it should be
safe because we are already assuming that the caller holds a lock that
prevents any new target pages from being loaded into the buffer pool
concurrently. Therefore, no buffer tag should be changing to a value of
interest, only away from a value of interest. So a false negative match
is impossible, while a false positive is safe because we'll recheck after
acquiring the buffer lock. Initial testing says that this speeds these
loops by a factor of 2X to 3X on common Intel hardware.
Patch for DropRelFileNodeBuffers by Jeff Janes (based on an idea of
Heikki's); extended to the remaining sequential scans by Tom Lane
WALSender now woken up after each background flush by WALwriter, avoiding
multi-second replication delay for an all-async commit workload.
Replication delay reduced from 7s with default settings to 200ms and often
much less, allowing significantly reduced data loss at failover.
Andres Freund and Simon Riggs
In lazy_scan_heap, we could issue bogus warnings about incorrect
information in the visibility map, because we checked the visibility
map bit before locking the heap page, creating a race condition. Fix
by rechecking the visibility map bit before we complain. Rejigger
some related logic so that we rely on the possibly-outdated
all_visible_according_to_vm value as little as possible.
In heap_multi_insert, it's not safe to clear the visibility map bit
before beginning the critical section. The visibility map is not
crash-safe unless we treat clearing the bit as a critical operation.
Specifically, if the transaction were to error out after we set the
bit and before entering the critical section, we could end up writing
the heap page to disk (with the bit cleared) and crashing before the
visibility map page made it to disk. That would be bad. heap_insert
has this correct, but somehow the order of operations got rearranged
when heap_multi_insert was added.
Also, add some more comments to visibilitymap_test, lazy_scan_heap,
and IndexOnlyNext, expounding on concurrency issues.
Per extensive code review by Andres Freund, and further review by Tom
Lane, who also made the original report about the bogus warnings.
The original coding misbehaved if "char" is signed, and also made the
extremely poor decision to print control characters literally when trying
to complain about them. Report and patch by Shigeru Hanada.
In passing, also fix core dump risk in report_parse_error() should the
parse state be something other than what it expects.
It failed to check for error return from xsltApplyStylesheet(), as reported
by Peter Gagarinov. (So far as I can tell, libxslt provides no convenient
way to get a useful error message in failure cases. There might be some
inconvenient way, but considering that this code is deprecated it's hard to
get enthusiastic about putting lots of work into it. So I just made it say
"failed to apply stylesheet", in line with the existing error checks.)
While looking at the code I also noticed that the string returned by
xsltSaveResultToString was never freed, resulting in a session-lifespan
memory leak.
Back-patch to all supported versions.
This is currently only cosmetic, since all the call sites just curl up
and die in event of a failure return. It might be important for some
future use-case, though, and in any case it quiets warnings from the
clang static analyzer (as reported by Anna Zaks).
Josh Kupershmidt
When we allowed read-only transactions to skip assigning XIDs
we introduced the possibility that a fully deleted btree page
could be reused. This broke the index link sequence which could
then lead to indexscans silently returning fewer rows than would
have been correct. The actual incidence of silent errors from
this is thought to be very low because of the exact workload
required and locking pre-conditions. Fix is to remove pages only
if index page opaque->btpo.xact precedes RecentGlobalXmin.
Noah Misch, reviewed by Simon Riggs
Re-implements similar functionality in 9.1 and previously which
was removed during split of checkpointer and bgwriter.
Requested/spotted by Magnus Hagander
We allow non-superusers to create procedural languages (with restrictions)
and range datatypes. Previously, the automatically-created support
functions for these objects ended up owned by the creating user. This
represents a rather considerable security hazard, because the owning user
might be able to alter a support function's definition in such a way as to
crash the server, inject trojan-horse SQL code, or even execute arbitrary
C code directly. It appears that right now the only actually exploitable
problem is the infinite-recursion bug fixed in the previous patch for
CVE-2012-2655. However, it's not hard to imagine that future additions of
more ALTER FUNCTION capability might unintentionally open up new hazards.
To forestall future problems, cause these support functions to be owned by
the bootstrap superuser, not the user creating the parent object.
It's not very sensible to set such attributes on a handler function;
but if one were to do so, fmgr.c went into infinite recursion because
it would call fmgr_security_definer instead of the handler function proper.
There is no way for fmgr_security_definer to know that it ought to call the
handler and not the original function referenced by the FmgrInfo's fn_oid,
so it tries to do the latter, causing the whole process to start over
again.
Ordinarily such misconfiguration of a procedural language's handler could
be written off as superuser error. However, because we allow non-superuser
database owners to create procedural languages and the handler for such a
language becomes owned by the database owner, it is possible for a database
owner to crash the backend, which ideally shouldn't be possible without
superuser privileges. In 9.2 and up we will adjust things so that the
handler functions are always owned by superusers, but in existing branches
this is a minor security fix.
Problem noted by Noah Misch (after several of us had failed to detect
it :-(). This is CVE-2012-2655.
We used to only allow offsets less than +/-13 hours, then it was +/14,
then it was +/-15. That's still not good enough though, as per today's bug
report from Patric Bechtel. This time I actually looked through the Olson
timezone database to find the largest offsets used anywhere. The winners
are Asia/Manila, at -15:56:00 until 1844, and America/Metlakatla, at
+15:13:42 until 1867. So we'd better allow offsets less than +/-16 hours.
Given the history, we are way overdue to have some greppable #define
symbols controlling this, so make some ... and also remove an obsolete
comment that didn't get fixed the last time.
Back-patch to all supported branches.
First, the previous code failed to account for the fact that, during Hot
Standby operation, the startup process takes AccessExclusiveLocks on
relations without setting MyDatabaseId. This resulted in fast path
strong lock counts failing to be incremented with the startup process
took locks, which in turn allowed conflicting lock requests to succeed
when they should not have. Report by Erik Rijkers, diagnosis by Heikki
Linnakangas.
Second, LockReleaseAll() failed to honor the allLocks and lockmethodid
restrictions with respect to fast-path locks. It's not clear to me
whether this produces any user-visible breakage at the moment, but it's
certainly wrong. Rearrange order of operations in LockReleaseAll to fix.
Noted by Tom Lane.
Overly tight coding caused the password transformation loop to stop
examining input once it had processed a byte equal to 0x80. Thus, if the
given password string contained such a byte (which is possible though not
highly likely in UTF8, and perhaps also in other non-ASCII encodings), all
subsequent characters would not contribute to the hash, making the password
much weaker than it appears on the surface.
This would only affect cases where applications used DES crypt() to encode
passwords before storing them in the database. If a weak password has been
created in this fashion, the hash will stop matching after this update has
been applied, so it will be easy to tell if any passwords were unexpectedly
weak. Changing to a different password would be a good idea in such a case.
(Since DES has been considered inadequately secure for some time, changing
to a different encryption algorithm can also be recommended.)
This code, and the bug, are shared with at least PHP, FreeBSD, and OpenBSD.
Since the other projects have already published their fixes, there is no
point in trying to keep this commit private.
This bug has been assigned CVE-2012-2143, and credit for its discovery goes
to Rubin Xu and Joseph Bonneau.
We used to mimic the way a stack is constructed when descending the tree
during normal GiST inserts, but that was quite complicated during a buffered
build. It was also wrong: in GiST, the left-to-right relationships on
different levels might not match each other, so that when you know the
parent of a child page, you won't necessarily find the parent of the page to
the right of the child page by following the rightlinks at the parent level.
This sometimes led to "could not re-find parent" errors while building a
GiST index.
We now use a simple hash table to track the parent of every internal page.
Whenever a page is split, and downlinks are moved from one page to another,
we update the hash table accordingly. This is also better for performance
than the old method, as we never need to move right to re-find the parent
page, which could take a significant amount of time for buffers that were
created much earlier in the index build.
There were two bugs here: We forgot to call gistFreeBuildBuffers() function
at the end of build, and we passed interXact == true to BufFileCreateTemp,
so the file wasn't automatically cleaned up at end-of-transaction either.
The initial implementation of pg_dump's --section option supposed that the
existing --schema-only and --data-only options could be made equivalent to
--section settings. This is wrong, though, due to dubious but long since
set-in-stone decisions about where to dump SEQUENCE SET items, as seen in
bug report from Martin Pitt. (And I'm not totally convinced there weren't
other bugs, either.) Undo that coupling and instead drive --section
filtering off current-section state tracked as we scan through the TOC
list to call _tocEntryRequired().
To make sure those decisions don't shift around and hopefully save a few
cycles, run _tocEntryRequired() only once per TOC entry and save the result
in a new TOC field. This required minor rejiggering of ACL handling but
also allows a far cleaner implementation of inhibit_data_for_failed_table.
Also, to ensure that pg_dump and pg_restore have the same behavior with
respect to the --section switches, add _tocEntryRequired() filtering to
WriteToc() and WriteDataChunks(), rather than trying to implement section
filtering in an entirely orthogonal way in dumpDumpableObject(). This
required adjusting the handling of the special ENCODING and STDSTRINGS
items, but they were pretty weird before anyway.
Minor other code review for the patch, too.
The result of (maintenance_work_mem * 1024) / BLCKSZ doesn't fit in a signed
32-bit integer, if maintenance_work_mem >= 2GB. Use double instead. And
while we're at it, write the calculations in an easier to understand form,
with the intermediary steps written out and commented.
AbortOutOfAnyTransaction failed to do anything if the state it saw on
entry corresponded to failing partway through StartTransaction. I fixed
AbortCurrentTransaction to cope with that case way back in commit
60b2444cc3, but evidently overlooked that
AbortOutOfAnyTransaction should do likewise.
Back-patch to all supported branches. It's not clear that this omission
has any more-than-cosmetic consequences, but it's also not clear that it
doesn't, so back-patching seems the least risky choice.
This patch fixes three places (which AFAICT is all of them) where runtime
was O(N^2) in the number of TOC entries, by using an index array to replace
linear searches of the TOC list. This performance issue is a bit less bad
than those recently fixed, because it depends on the number of items dumped
not the number in the source database, so the problem can be dodged by
doing partial dumps.
The previous coding already had an instance of one of the two index arrays
needed, but it was only calculated in parallel-restore cases; now we need
it all the time. I also chose to move the arrays into the ArchiveHandle
data structure, to make this code a bit more ready for the day that we
try to sling multiple ArchiveHandles around in pg_dump or pg_restore.
Since we still need some server-side work before pg_dump can really cope
nicely with tens of thousands of tables, there's probably little point in
back-patching.
Drop special handling of host component with slashes to mean
Unix-domain socket. Specify it as separate parameter or using
percent-encoding now.
Allow omitting username, password, and port even if the corresponding
designators are present in URI.
Handle percent-encoding in query parameter keywords.
Alex Shulgin
some documentation improvements by myself
Set E081 Basic Privileges to supported, since by the letter of it, we
support it, even though not all possible forms of USAGE privileges are
implemented.
Write the file to a temporary name and then rename() it into the
permanent name, to ensure it can't end up half-written and corrupt
in case of a crash during shutdown.
Unlink the file after it has been read so it's removed from the data
directory and not included in base backups going to replication slaves.
The only interesting-for-performance case wherein we force heapscan here
is when we're rebuilding the relcache init file, and the only such case
that is likely to be examining a catalog big enough to be syncscanned is
RelationBuildTupleDesc. But the early-exit optimization in that code gets
broken if we start the scan at a random place within the catalog, so that
allowing syncscan is actually a big deoptimization if pg_attribute is large
(at least for the normal case where the rows for core system catalogs have
never been changed since initdb). Hence, prevent syncscan here. Per my
testing pursuant to complaints from Jeff Frost and Greg Sabino Mullane,
though neither of them seem to have actually hit this specific problem.
Back-patch to 8.3, where syncscan was introduced.
Previously, casts to name could generate invalidly-encoded results.
Also, make these functions match namein() more exactly, by consistently
using palloc0() instead of ad-hoc zeroing code.
Back-patch to all supported branches.
Karl Schnaitter and Tom Lane
The previous coding presented a significant bottleneck when dumping
databases containing many thousands of schemas, since the total time
spent searching would increase roughly as O(N^2) in the number of objects.
Noted by Jeff Janes, though I rewrote his proposed patch to use the
existing findObjectByOid infrastructure.
Since this is a longstanding performance bug, backpatch to all supported
versions.
When backing up from a standby server, the backup process
will not automatically switch xlog segment. So we must
accept a partially transferred xlog file in this case, but
rename it into position anyway.
In passing, merge the two callbacks for segment end and
stop stream into a single callback, since their implementations
were close to identical, and rename this callback to
reflect that it stops streaming rather than continues it.
Patch by Magnus Hagander, review by Fujii Masao
