[pull] master from postgres:master

contrib/amcheck/verify_heapam.c

@@ -1942,7 +1942,7 @@ check_tuple(HeapCheckContext *ctx, bool *xmin_commit_status_ok,
 	if (RelationGetDescr(ctx->rel)->natts < ctx->natts)
 	{
 		report_corruption(ctx,
-						  psprintf("number of attributes %u exceeds maximum expected for table %u",
+						  psprintf("number of attributes %u exceeds maximum %u expected for table",
 								   ctx->natts,
 								   RelationGetDescr(ctx->rel)->natts));
 		return;

contrib/pg_buffercache/pg_buffercache_pages.c

@@ -320,7 +320,6 @@ pg_buffercache_numa_pages(PG_FUNCTION_ARGS)
 		uint64		os_page_count;
 		int			pages_per_buffer;
 		int			max_entries;
-		volatile uint64 touch pg_attribute_unused();
 		char	   *startptr,
 				   *endptr;
 
@@ -375,7 +374,7 @@ pg_buffercache_numa_pages(PG_FUNCTION_ARGS)
 
 			/* Only need to touch memory once per backend process lifetime */
 			if (firstNumaTouch)
-				pg_numa_touch_mem_if_required(touch, ptr);
+				pg_numa_touch_mem_if_required(ptr);
 		}
 
 		Assert(idx == os_page_count);

doc/src/sgml/amcheck.sgml

@@ -278,8 +278,8 @@ SET client_min_messages = DEBUG1;
          TOAST table.
         </para>
         <para>
-         This option is known to be slow.  Also, if the toast table or its
-         index is corrupt, checking it against toast values could conceivably
+         This option is known to be slow.  Also, if the TOAST table or its
+         index is corrupt, checking it against TOAST values could conceivably
          crash the server, although in many cases this would just produce an
          error.
         </para>

doc/src/sgml/bki.sgml

@@ -1042,7 +1042,7 @@ $ perl  rewrite_dat_with_prokind.pl  pg_proc.dat
     </listitem>
     <listitem>
      <para>
-      Define indexes and toast tables.
+      Define indexes and TOAST tables.
      </para>
     </listitem>
     <listitem>

doc/src/sgml/catalogs.sgml

@@ -1951,7 +1951,7 @@ SCRAM-SHA-256$<replaceable>&lt;iteration count&gt;</replaceable>:<replaceable>&l
       </para>
       <para>
        The OID of the data type that corresponds to this table's row type,
-       if any; zero for indexes, sequences, and toast tables, which have
+       if any; zero for indexes, sequences, and TOAST tables, which have
        no <structname>pg_type</structname> entry
       </para></entry>
      </row>

doc/src/sgml/installation.sgml

@@ -3847,17 +3847,13 @@ make: *** [postgres] Error 1
    <para>
     Both 32-bit and 64-bit builds are possible with the Microsoft Compiler suite.
     32-bit PostgreSQL builds are possible with
-    <productname>Visual Studio 2015</productname> to
+    <productname>Visual Studio 2019</productname> to
     <productname>Visual Studio 2022</productname>,
     as well as standalone Windows SDK releases 10 and above.
     64-bit PostgreSQL builds are supported with
     <productname>Microsoft Windows SDK</productname> version 10 and above or
-    <productname>Visual Studio 2015</productname> and above.
+    <productname>Visual Studio 2019</productname> and above.
     <!--
-        For 2015 requirements:
-        https://docs.microsoft.com/en-us/visualstudio/productinfo/vs2015-sysrequirements-vs
-        For 2017 requirements:
-        https://docs.microsoft.com/en-us/visualstudio/productinfo/vs2017-system-requirements-vs
         For 2019 requirements:
         https://docs.microsoft.com/en-us/visualstudio/releases/2019/system-requirements
         For 2022 requirements:

doc/src/sgml/logicaldecoding.sgml

@@ -1412,7 +1412,7 @@ commit_prepared_cb(...);  <-- commit of the prepared transaction
     currently used for decoded changes) is selected and streamed.  However, in
     some cases we still have to spill to disk even if streaming is enabled
     because we exceed the memory threshold but still have not decoded the
-    complete tuple e.g., only decoded toast table insert but not the main table
+    complete tuple e.g., only decoded TOAST table insert but not the main table
     insert.
    </para>
 

doc/src/sgml/ref/alter_table.sgml

@@ -852,7 +852,7 @@ WITH ( MODULUS <replaceable class="parameter">numeric_literal</replaceable>, REM
 
      <para>
       <literal>SHARE UPDATE EXCLUSIVE</literal> lock will be taken for
-      fillfactor, toast and autovacuum storage parameters, as well as the
+      fillfactor, TOAST and autovacuum storage parameters, as well as the
       planner parameter <varname>parallel_workers</varname>.
      </para>
     </listitem>

doc/src/sgml/ref/pg_amcheck.sgml

@@ -41,7 +41,7 @@ PostgreSQL documentation
   </para>
 
   <para>
-   Only ordinary and toast table relations, materialized views, sequences, and
+   Only ordinary and TOAST table relations, materialized views, sequences, and
    btree indexes are currently supported.  Other relation types are silently
    skipped.
   </para>
@@ -276,7 +276,7 @@ PostgreSQL documentation
      <term><option>--no-dependent-toast</option></term>
      <listitem>
       <para>
-       By default, if a table is checked, its toast table, if any, will also
+       By default, if a table is checked, its TOAST table, if any, will also
        be checked, even if it is not explicitly selected by an option
        such as <literal>--table</literal> or <literal>--relation</literal>.
        This option suppresses that behavior.
@@ -306,9 +306,9 @@ PostgreSQL documentation
      <term><option>--exclude-toast-pointers</option></term>
      <listitem>
       <para>
-       By default, whenever a toast pointer is encountered in a table,
+       By default, whenever a TOAST pointer is encountered in a table,
        a lookup is performed to ensure that it references apparently-valid
-       entries in the toast table. These checks can be quite slow, and this
+       entries in the TOAST table. These checks can be quite slow, and this
        option can be used to skip them.
       </para>
      </listitem>
@@ -368,9 +368,9 @@ PostgreSQL documentation
        End checking at the specified block number.  An error will occur if the
        table relation being checked has fewer than this number of blocks.
        This option does not apply to indexes, and is probably only useful when
-       checking a single table relation. If both a regular table and a toast
+       checking a single table relation. If both a regular table and a TOAST
        table are checked, this option will apply to both, but higher-numbered
-       toast blocks may still be accessed while validating toast pointers,
+       TOAST blocks may still be accessed while validating TOAST pointers,
        unless that is suppressed using
        <option>--exclude-toast-pointers</option>.
       </para>

doc/src/sgml/sepgsql.sgml

@@ -442,7 +442,7 @@ UPDATE t1 SET x = 2, y = func1(y) WHERE z = 100;
    <para>
     The default database privilege system allows database superusers to
     modify system catalogs using DML commands, and reference or modify
-    toast tables.  These operations are prohibited when
+    TOAST tables.  These operations are prohibited when
     <filename>sepgsql</filename> is enabled.
    </para>
   </sect3>

meson.build

@@ -279,6 +279,10 @@ elif host_system == 'windows'
     # define before including <time.h> for getting localtime_r() etc. on MinGW
     cppflags += '-D_POSIX_C_SOURCE'
   endif
+  if cc.get_id() == 'msvc'
+    # required for VA_ARGS_NARGS() in c.h; requires VS 2019
+    cppflags += '/Zc:preprocessor'
+  endif
 
   export_file_format = 'win'
   export_file_suffix = 'def'

src/backend/storage/ipc/shmem.c

@@ -679,12 +679,10 @@ pg_get_shmem_allocations_numa(PG_FUNCTION_ARGS)
 		 */
 		for (i = 0; i < shm_ent_page_count; i++)
 		{
-			volatile uint64 touch pg_attribute_unused();
-
 			page_ptrs[i] = startptr + (i * os_page_size);
 
 			if (firstNumaTouch)
-				pg_numa_touch_mem_if_required(touch, page_ptrs[i]);
+				pg_numa_touch_mem_if_required(page_ptrs[i]);
 
 			CHECK_FOR_INTERRUPTS();
 		}

src/bin/pg_amcheck/t/004_verify_heapam.pl

@@ -529,7 +529,7 @@ sub header
 		$tup->{t_infomask2} |= HEAP_NATTS_MASK;
 
 		push @expected,
-		  qr/${$header}number of attributes 2047 exceeds maximum expected for table 3/;
+		  qr/${$header}number of attributes 2047 exceeds maximum 3 expected for table/;
 	}
 	elsif ($offnum == 10)
 	{
@@ -552,7 +552,7 @@ sub header
 		$tup->{t_hoff} = 32;
 
 		push @expected,
-		  qr/${$header}number of attributes 67 exceeds maximum expected for table 3/;
+		  qr/${$header}number of attributes 67 exceeds maximum 3 expected for table/;
 	}
 	elsif ($offnum == 12)
 	{

src/include/c.h

@@ -376,25 +376,7 @@
  * pretty trivial: VA_ARGS_NARGS_() returns its 64th argument, and we set up
  * the call so that that is the appropriate one of the list of constants.
  * This idea is due to Laurent Deniau.
- *
- * MSVC has an implementation of __VA_ARGS__ that doesn't conform to the
- * standard unless you use the /Zc:preprocessor compiler flag, but that
- * isn't available before Visual Studio 2019.  For now, use a different
- * definition that also works on older compilers.
  */
-#ifdef _MSC_VER
-#define EXPAND(args) args
-#define VA_ARGS_NARGS(...) \
-	VA_ARGS_NARGS_ EXPAND((__VA_ARGS__, \
-				   63,62,61,60,                   \
-				   59,58,57,56,55,54,53,52,51,50, \
-				   49,48,47,46,45,44,43,42,41,40, \
-				   39,38,37,36,35,34,33,32,31,30, \
-				   29,28,27,26,25,24,23,22,21,20, \
-				   19,18,17,16,15,14,13,12,11,10, \
-				   9, 8, 7, 6, 5, 4, 3, 2, 1, 0))
-#else
-
 #define VA_ARGS_NARGS(...) \
 	VA_ARGS_NARGS_(__VA_ARGS__, \
 				   63,62,61,60,                   \
@@ -404,7 +386,6 @@
 				   29,28,27,26,25,24,23,22,21,20, \
 				   19,18,17,16,15,14,13,12,11,10, \
 				   9, 8, 7, 6, 5, 4, 3, 2, 1, 0)
-#endif
 
 #define VA_ARGS_NARGS_( \
 	_01,_02,_03,_04,_05,_06,_07,_08,_09,_10, \

src/include/catalog/pg_publication.h

@@ -146,7 +146,7 @@ extern Publication *GetPublicationByName(const char *pubname, bool missing_ok);
 extern List *GetRelationPublications(Oid relid);
 
 /*---------
- * Expected values for pub_partopt parameter of GetRelationPublications(),
+ * Expected values for pub_partopt parameter of GetPublicationRelations(),
  * which allows callers to specify which partitions of partitioned tables
  * mentioned in the publication they expect to see.
  *

src/include/port/pg_numa.h

@@ -24,12 +24,17 @@ extern PGDLLIMPORT int pg_numa_get_max_node(void);
  * This is required on Linux, before pg_numa_query_pages() as we
  * need to page-fault before move_pages(2) syscall returns valid results.
  */
-#define pg_numa_touch_mem_if_required(ro_volatile_var, ptr) \
-	ro_volatile_var = *(volatile uint64 *) ptr
+static inline void
+pg_numa_touch_mem_if_required(void *ptr)
+{
+	volatile uint64 touch pg_attribute_unused();
+
+	touch = *(volatile uint64 *) ptr;
+}
 
 #else
 
-#define pg_numa_touch_mem_if_required(ro_volatile_var, ptr) \
+#define pg_numa_touch_mem_if_required(ptr) \
 	do {} while(0)
 
 #endif

src/port/pg_numa.c

@@ -16,6 +16,7 @@
 #include "c.h"
 #include <unistd.h>
 
+#include "miscadmin.h"
 #include "port/pg_numa.h"
 
 /*
@@ -29,6 +30,19 @@
 #include <numa.h>
 #include <numaif.h>
 
+/*
+ * numa_move_pages() chunk size, has to be <= 16 to work around a kernel bug
+ * in do_pages_stat() (chunked by DO_PAGES_STAT_CHUNK_NR). By using the same
+ * chunk size, we make it work even on unfixed kernels.
+ *
+ * 64-bit system are not affected by the bug, and so use much larger chunks.
+ */
+#if SIZEOF_SIZE_T == 4
+#define NUMA_QUERY_CHUNK_SIZE 16
+#else
+#define NUMA_QUERY_CHUNK_SIZE 1024
+#endif
+
 /* libnuma requires initialization as per numa(3) on Linux */
 int
 pg_numa_init(void)
@@ -42,11 +56,48 @@ pg_numa_init(void)
  * We use move_pages(2) syscall here - instead of get_mempolicy(2) - as the
  * first one allows us to batch and query about many memory pages in one single
  * giant system call that is way faster.
+ *
+ * We call numa_move_pages() for smaller chunks of the whole array. The first
+ * reason is to work around a kernel bug, but also to allow interrupting the
+ * query between the calls (for many pointers processing the whole array can
+ * take a lot of time).
  */
 int
 pg_numa_query_pages(int pid, unsigned long count, void **pages, int *status)
 {
-	return numa_move_pages(pid, count, pages, NULL, status, 0);
+	unsigned long next = 0;
+	int			ret = 0;
+
+	/*
+	 * Chunk pointers passed to numa_move_pages to NUMA_QUERY_CHUNK_SIZE
+	 * items, to work around a kernel bug in do_pages_stat().
+	 */
+	while (next < count)
+	{
+		unsigned long count_chunk = Min(count - next,
+										NUMA_QUERY_CHUNK_SIZE);
+
+		CHECK_FOR_INTERRUPTS();
+
+		/*
+		 * Bail out if any of the chunks errors out (ret<0). We ignore (ret>0)
+		 * which is used to return number of nonmigrated pages, but we're not
+		 * migrating any pages here.
+		 */
+		ret = numa_move_pages(pid, count_chunk, &pages[next], NULL, &status[next], 0);
+		if (ret < 0)
+		{
+			/* plain error, return as is */
+			return ret;
+		}
+
+		next += count_chunk;
+	}
+
+	/* should have consumed the input array exactly */
+	Assert(next == count);
+
+	return 0;
 }
 
 int

src/tools/valgrind.supp

@@ -180,3 +180,17 @@
    Memcheck:Cond
    fun:PyObject_Realloc
 }
+
+# NUMA introspection requires touching memory first, and some of it may
+# be marked as noacess (e.g. unpinned buffers). So just ignore that.
+{
+   pg_numa_touch_mem_if_required
+   Memcheck:Addr4
+   fun:pg_numa_touch_mem_if_required
+}
+
+{
+   pg_numa_touch_mem_if_required
+   Memcheck:Addr8
+   fun:pg_numa_touch_mem_if_required
+}