netcdf-c/libdispatch/doffsets.c
Dennis Heimbigner 8b9253fef2 Fix various problem around VLEN's
re: https://github.com/Unidata/netcdf-c/issues/541
re: https://github.com/Unidata/netcdf-c/issues/1208
re: https://github.com/Unidata/netcdf-c/issues/2078
re: https://github.com/Unidata/netcdf-c/issues/2041
re: https://github.com/Unidata/netcdf-c/issues/2143

For a long time, there have been known problems with the
management of complex types containing VLENs.  This also
involves the string type because it is stored as a VLEN of
chars.

This PR (mostly) fixes this problem. But note that it adds new
functions to netcdf.h (see below) and this may require bumping
the .so number.  These new functions can be removed, if desired,
in favor of functions in netcdf_aux.h, but netcdf.h seems the
better place for them because they are intended as alternatives
to the nc_free_vlen and nc_free_string functions already in
netcdf.h.

The term complex type refers to any type that directly or
transitively references a VLEN type. So an array of VLENS, a
compound with a VLEN field, and so on.

In order to properly handle instances of these complex types, it
is necessary to have function that can recursively walk
instances of such types to perform various actions on them.  The
term "deep" is also used to mean recursive.

At the moment, the two operations needed by the netcdf library are:
* free'ing an instance of the complex type
* copying an instance of the complex type.

The current library does only shallow free and shallow copy of
complex types. This means that only the top level is properly
free'd or copied, but deep internal blocks in the instance are
not touched.

Note that the term "vector" will be used to mean a contiguous (in
memory) sequence of instances of some type. Given an array with,
say, dimensions 2 X 3 X 4, this will be stored in memory as a
vector of length 2*3*4=24 instances.

The use cases are primarily these.

## nc_get_vars
Suppose one is reading a vector of instances using nc_get_vars
(or nc_get_vara or nc_get_var, etc.).  These functions will
return the vector in the top-level memory provided.  All
interior blocks (form nested VLEN or strings) will have been
dynamically allocated.

After using this vector of instances, it is necessary to free
(aka reclaim) the dynamically allocated memory, otherwise a
memory leak occurs.  So, the recursive reclaim function is used
to walk the returned instance vector and do a deep reclaim of
the data.

Currently functions are defined in netcdf.h that are supposed to
handle this: nc_free_vlen(), nc_free_vlens(), and
nc_free_string().  Unfortunately, these functions only do a
shallow free, so deeply nested instances are not properly
handled by them.

Note that internally, the provided data is immediately written so
there is no need to copy it. But the caller may need to reclaim the
data it passed into the function.

## nc_put_att
Suppose one is writing a vector of instances as the data of an attribute
using, say, nc_put_att.

Internally, the incoming attribute data must be copied and stored
so that changes/reclamation of the input data will not affect
the attribute.

Again, the code inside the netcdf library does only shallow copying
rather than deep copy. As a result, one sees effects such as described
in Github Issue https://github.com/Unidata/netcdf-c/issues/2143.

Also, after defining the attribute, it may be necessary for the user
to free the data that was provided as input to nc_put_att().

## nc_get_att
Suppose one is reading a vector of instances as the data of an attribute
using, say, nc_get_att.

Internally, the existing attribute data must be copied and returned
to the caller, and the caller is responsible for reclaiming
the returned data.

Again, the code inside the netcdf library does only shallow copying
rather than deep copy. So this can lead to memory leaks and errors
because the deep data is shared between the library and the user.

# Solution

The solution is to build properly recursive reclaim and copy
functions and use those as needed.
These recursive functions are defined in libdispatch/dinstance.c
and their signatures are defined in include/netcdf.h.
For back compatibility, corresponding "ncaux_XXX" functions
are defined in include/netcdf_aux.h.
````
int nc_reclaim_data(int ncid, nc_type xtypeid, void* memory, size_t count);
int nc_reclaim_data_all(int ncid, nc_type xtypeid, void* memory, size_t count);
int nc_copy_data(int ncid, nc_type xtypeid, const void* memory, size_t count, void* copy);
int nc_copy_data_all(int ncid, nc_type xtypeid, const void* memory, size_t count, void** copyp);
````
There are two variants. The first two, nc_reclaim_data() and
nc_copy_data(), assume the top-level vector is managed by the
caller. For reclaim, this is so the user can use, for example, a
statically allocated vector. For copy, it assumes the user
provides the space into which the copy is stored.

The second two, nc_reclaim_data_all() and
nc_copy_data_all(), allows the functions to manage the
top-level.  So for nc_reclaim_data_all, the top level is
assumed to be dynamically allocated and will be free'd by
nc_reclaim_data_all().  The nc_copy_data_all() function
will allocate the top level and return a pointer to it to the
user. The user can later pass that pointer to
nc_reclaim_data_all() to reclaim the instance(s).

# Internal Changes
The netcdf-c library internals are changed to use the proper
reclaim and copy functions.  It turns out that the places where
these functions are needed is quite pervasive in the netcdf-c
library code.  Using these functions also allows some
simplification of the code since the stdata and vldata fields of
NC_ATT_INFO are no longer needed.  Currently this is commented
out using the SEPDATA \#define macro.  When any bugs are largely
fixed, all this code will be removed.

# Known Bugs

1. There is still one known failure that has not been solved.
   All the failures revolve around some variant of this .cdl file.
   The proximate cause of failure is the use of a VLEN FillValue.
````
        netcdf x {
        types:
          float(*) row_of_floats ;
        dimensions:
          m = 5 ;
        variables:
          row_of_floats ragged_array(m) ;
              row_of_floats ragged_array:_FillValue = {-999} ;
        data:
          ragged_array = {10, 11, 12, 13, 14}, {20, 21, 22, 23}, {30, 31, 32},
                         {40, 41}, _ ;
        }
````
When a solution is found, I will either add it to this PR or post a new PR.

# Related Changes

* Mark nc_free_vlen(s) as deprecated in favor of ncaux_reclaim_data.
* Remove the --enable-unfixed-memory-leaks option.
* Remove the NC_VLENS_NOTEST code that suppresses some vlen tests.
* Document this change in docs/internal.md
* Disable the tst_vlen_data test in ncdump/tst_nccopy4.sh.
* Mark types as fixed size or not (transitively) to optimize the reclaim
  and copy functions.

# Misc. Changes

* Make Doxygen process libdispatch/daux.c
* Make sure the NC_ATT_INFO_T.container field is set.
2022-01-08 18:30:00 -07:00

349 lines
12 KiB
C

/*********************************************************************
* Copyright 2018, UCAR/Unidata
* See netcdf/COPYRIGHT file for copying and redistribution conditions.
* $Header: /upc/share/CVS/netcdf-3/ncgen/offsets.c,v 1.1 2009/09/25 18:22:40 dmh Exp $
*********************************************************************/
/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* Copyright by The HDF Group. *
* Copyright by the Board of Trustees of the University of Illinois. *
* All rights reserved. *
* *
* This file is part of HDF5. The full HDF5 copyright notice, including *
* terms governing use, modification, and redistribution, is contained in *
* the files COPYING and Copyright.html. COPYING can be found at the root *
* of the source code distribution tree; Copyright.html can be found at the *
* root level of an installed copy of the electronic HDF5 document set and *
* is linked from the top-level documents page. It can also be found at *
* http://hdfgroup.org/HDF5/doc/Copyright.html. If you do not have *
* access to either file, you may request a copy from help@hdfgroup.org. *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
/*
This code is a variantion of the H5detect.c code from HDF5.
Author: D. Heimbigner 10/7/2008
*/
#include "config.h"
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <assert.h>
#include "nclog.h"
#ifdef OFFSETTEST
static void* emalloc(size_t);
typedef int nc_type;
typedef struct nc_vlen_t {
size_t len;
void* p;
} nc_vlen_t;
#define NC_NAT 0 /* NAT = 'Not A Type' (c.f. NaN) */
#define NC_BYTE 1 /* signed 1 byte integer */
#define NC_CHAR 2 /* ISO/ASCII character */
#define NC_SHORT 3 /* signed 2 byte integer */
#define NC_INT 4 /* signed 4 byte integer */
#define NC_FLOAT 5 /* single precision floating point number */
#define NC_DOUBLE 6 /* double precision floating point number */
#define NC_UBYTE 7 /* unsigned 1 byte int */
#define NC_USHORT 8 /* unsigned 2-byte int */
#define NC_UINT 9 /* unsigned 4-byte int */
#define NC_INT64 10 /* signed 8-byte int */
#define NC_UINT64 11 /* unsigned 8-byte int */
#define NC_STRING 12 /* string */
#define NC_STRING 12 /* string */
#define NC_VLEN 13
#define NC_OPAQUE 14
#define NC_ENUM 15
#define NC_COMPOUND 16
#endif
#include "netcdf.h"
#include "ncoffsets.h"
/*
The heart of this is the following macro,
which computes the offset of a field x
when preceded by a char field.
The assumptions appear to be as follows:
1. the offset produced in this situation indicates
the alignment for x relative in such a way that it
depends only on the types that precede it in the struct.
2. the compiler does not reorder fields.
3. arrays are tightly packed.
4. nested structs are alignd according to their first member
(this actually follows from C language requirement that
a struct can legally be cast to an instance of its first member).
Given the alignments for the various common primitive types,
it is assumed that one can use them anywhere to construct
the layout of a struct of such types.
It seems to work for HDF5 for a wide variety of machines.
Note that technically, this is compiler dependent, but in practice
all compilers seem to mimic the gcc rules.
*/
#define COMP_ALIGNMENT(DST,TYPE) {\
struct {char f1; TYPE x;} tmp; \
DST.type_name = #TYPE ; \
DST.alignment = (size_t)((char*)(&(tmp.x)) - (char*)(&tmp));}
#if 0
char* ctypenames[NCTYPES] = {
(char*)NULL,
"char","unsigned char",
"short","unsigned short",
"int","unsigned int",
"long long","unsigned long long",
"float","double",
"void*","nc_vlen_t"
};
#endif
static NCtypealignvec vec[NC_NCTYPES];
static NCtypealignset set;
static int NC_alignments_computed = 0;
/* Argument is a netcdf type class, except compound|ENUM */
int
NC_class_alignment(int ncclass, size_t* alignp)
{
int stat = NC_NOERR;
NCalignment* align = NULL;
int index = 0;
if(!NC_alignments_computed)
NC_compute_alignments();
switch (ncclass) {
case NC_BYTE: index = NC_UCHARINDEX; break;
case NC_CHAR: index = NC_CHARINDEX; break;
case NC_SHORT: index = NC_SHORTINDEX; break;
case NC_INT: index = NC_INTINDEX; break;
case NC_FLOAT: index = NC_FLOATINDEX; break;
case NC_DOUBLE: index = NC_DOUBLEINDEX; break;
case NC_UBYTE: index = NC_UCHARINDEX; break;
case NC_USHORT: index = NC_USHORTINDEX; break;
case NC_UINT: index = NC_UINTINDEX; break;
case NC_INT64: index = NC_LONGLONGINDEX; break;
case NC_UINT64: index = NC_ULONGLONGINDEX; break;
case NC_STRING: index = NC_PTRINDEX; break;
/* Here class matters */
case NC_VLEN: index = NC_NCVLENINDEX; break;
case NC_OPAQUE: index = NC_UCHARINDEX; break;
case NC_ENUM: /* fall thru */
case NC_COMPOUND: /* fall thru */
default:
nclog(NCLOGERR,"nc_class_alignment: class code %d cannot be aligned",ncclass);
goto done;
}
align = &vec[index];
if(alignp) *alignp = align->alignment;
done:
return stat;
}
void
NC_compute_alignments(void)
{
if(NC_alignments_computed) return;
/* Compute the alignments for all the common C data types*/
/* First for the struct*/
/* initialize*/
memset((void*)&set,0,sizeof(set));
memset((void*)vec,0,sizeof(vec));
COMP_ALIGNMENT(set.charalign,char);
COMP_ALIGNMENT(set.ucharalign,unsigned char);
COMP_ALIGNMENT(set.shortalign,short);
COMP_ALIGNMENT(set.ushortalign,unsigned short);
COMP_ALIGNMENT(set.intalign,int);
COMP_ALIGNMENT(set.uintalign,unsigned int);
COMP_ALIGNMENT(set.longlongalign,long long);
COMP_ALIGNMENT(set.ulonglongalign,unsigned long long);
COMP_ALIGNMENT(set.floatalign,float);
COMP_ALIGNMENT(set.doublealign,double);
COMP_ALIGNMENT(set.ptralign,void*);
COMP_ALIGNMENT(set.ncvlenalign,nc_vlen_t);
/* Then the vector*/
COMP_ALIGNMENT(vec[NC_CHARINDEX],char);
COMP_ALIGNMENT(vec[NC_UCHARINDEX],unsigned char);
COMP_ALIGNMENT(vec[NC_SHORTINDEX],short);
COMP_ALIGNMENT(vec[NC_USHORTINDEX],unsigned short);
COMP_ALIGNMENT(vec[NC_INTINDEX],int);
COMP_ALIGNMENT(vec[NC_UINTINDEX],unsigned int);
COMP_ALIGNMENT(vec[NC_LONGLONGINDEX],long long);
COMP_ALIGNMENT(vec[NC_ULONGLONGINDEX],unsigned long long);
COMP_ALIGNMENT(vec[NC_FLOATINDEX],float);
COMP_ALIGNMENT(vec[NC_DOUBLEINDEX],double);
COMP_ALIGNMENT(vec[NC_PTRINDEX],void*);
COMP_ALIGNMENT(vec[NC_NCVLENINDEX],nc_vlen_t);
NC_alignments_computed = 1;
}
#ifdef OFFSETTEST
/* Compute the alignment of TYPE when it is preceded
by a field of type TYPE1
*/
#define COMP_ALIGNMENT1(DST,TYPE1,TYPE) {\
struct {TYPE1 f1; TYPE x;} tmp; \
DST.type_name = #TYPE ; \
DST.alignment = (size_t)((char*)(&(tmp.x)) - (char*)(&tmp));}
/* Compute the alignment of TYPE when it is preceded
by a field of type TYPE1 and a field of type TYPE2
*/
#define COMP_ALIGNMENT2(DST,TYPE1,TYPE2,TYPE) {\
struct {TYPE1 f1, TYPE2 f2; TYPE x;} tmp; \
DST.type_name = #TYPE ; \
DST.alignment = (size_t)((char*)(&(tmp.x)) - (char*)(&tmp));}
/* Compute the alignment of TYPE when it is preceded
by a field of type TYPE1 and a field of type TYPE2
*/
#define COMP_SIZE0(DST,TYPE1,TYPE2) {\
struct {TYPE1 c; TYPE2 x;} tmp; \
DST = sizeof(tmp); }
static char*
padname(char* name)
{
#define MAX 20
if(name == NULL) name = "null";
int len = strlen(name);
if(len > MAX) len = MAX;
char* s = (char*)emalloc(MAX+1);
memset(s,' ',MAX);
s[MAX+1] = '\0';
strncpy(s,name,len);
return s;
}
static void
verify(NCtypealignvec* vec)
{
int i,j;
NCtypealignvec* vec16;
NCtypealignvec* vec32;
int* sizes8;
int* sizes16;
int* sizes32;
vec16 = (NCtypealignvec*)emalloc(sizeof(NCtypealignvec)*NCTYPES);
vec32 = (NCtypealignvec*)emalloc(sizeof(NCtypealignvec)*NCTYPES);
sizes8 = (int*)emalloc(sizeof(int)*NCTYPES);
sizes16 = (int*)emalloc(sizeof(int)*NCTYPES);
sizes32 = (int*)emalloc(sizeof(int)*NCTYPES);
COMP_SIZE0(sizes8[1],char,char);
COMP_SIZE0(sizes8[2],unsigned char,char);
COMP_SIZE0(sizes8[3],short,char);
COMP_SIZE0(sizes8[4],unsigned short,char);
COMP_SIZE0(sizes8[5],int,char);
COMP_SIZE0(sizes8[6],unsigned int,char);
COMP_SIZE0(sizes8[7],long long,char);
COMP_SIZE0(sizes8[8],unsigned long long,char);
COMP_SIZE0(sizes8[9],float,char);
COMP_SIZE0(sizes8[10],double,char) ;
COMP_SIZE0(sizes8[11],void*,char);
COMP_SIZE0(sizes8[12],nc_vlen_t,char);
COMP_SIZE0(sizes16[1],char,short);
COMP_SIZE0(sizes16[2],unsigned char,short);
COMP_SIZE0(sizes16[3],short,short);
COMP_SIZE0(sizes16[4],unsigned short,short);
COMP_SIZE0(sizes16[5],int,short);
COMP_SIZE0(sizes16[6],unsigned int,short);
COMP_SIZE0(sizes16[7],long long,short);
COMP_SIZE0(sizes16[8],unsigned long long,short);
COMP_SIZE0(sizes16[9],float,short);
COMP_SIZE0(sizes16[10],double,short) ;
COMP_SIZE0(sizes16[11],void*,short);
COMP_SIZE0(sizes16[12],nc_vlen_t*,short);
COMP_SIZE0(sizes32[1],char,int);
COMP_SIZE0(sizes32[2],unsigned char,int);
COMP_SIZE0(sizes32[3],short,int);
COMP_SIZE0(sizes32[4],unsigned short,int);
COMP_SIZE0(sizes32[5],int,int);
COMP_SIZE0(sizes32[6],unsigned int,int);
COMP_SIZE0(sizes32[7],long long,int);
COMP_SIZE0(sizes32[8],unsigned long long,int);
COMP_SIZE0(sizes32[9],float,int);
COMP_SIZE0(sizes32[10],double,int) ;
COMP_SIZE0(sizes32[11],void*,int);
COMP_SIZE0(sizes32[12],nc_vlen_t*,int);
COMP_ALIGNMENT1(vec16[1],char,short);
COMP_ALIGNMENT1(vec16[2],unsigned char,short);
COMP_ALIGNMENT1(vec16[3],short,short);
COMP_ALIGNMENT1(vec16[4],unsigned short,short);
COMP_ALIGNMENT1(vec16[5],int,short);
COMP_ALIGNMENT1(vec16[6],unsigned int,short);
COMP_ALIGNMENT1(vec32[7],long long,short);
COMP_ALIGNMENT1(vec32[8],unsigned long long,short);
COMP_ALIGNMENT1(vec16[9],float,short);
COMP_ALIGNMENT1(vec16[10],double,short);
COMP_ALIGNMENT1(vec16[11],void*,short);
COMP_ALIGNMENT1(vec16[12],nc_vlen_t*,short);
COMP_ALIGNMENT1(vec32[1],char,short);
COMP_ALIGNMENT1(vec32[2],unsigned char,short);
COMP_ALIGNMENT1(vec32[3],char,short);
COMP_ALIGNMENT1(vec32[4],unsigned short,short);
COMP_ALIGNMENT1(vec32[5],int,int);
COMP_ALIGNMENT1(vec32[6],unsigned int,int);
COMP_ALIGNMENT1(vec32[7],long long,int);
COMP_ALIGNMENT1(vec32[8],unsigned long long,int);
COMP_ALIGNMENT1(vec32[9],float,int);
COMP_ALIGNMENT1(vec32[10],double,int);
COMP_ALIGNMENT1(vec32[11],void*,int);
COMP_ALIGNMENT1(vec32[12],nc_vlen_t*,int);
for(i=0;i<NCTYPES;i++) {
printf("%s: size=%2d alignment=%2d\n",
padname(vec[i].type_name),sizes8[i],vec[i].alignment);
}
for(i=0;i<NCTYPES;i++) {
printf("short vs %s: size=%2d alignment=%2d\n",
padname(vec[i].type_name),sizes16[i],vec16[i].alignment);
}
for(i=0;i<NCTYPES;i++) {
printf("int vs %s: size=%2d alignment=%2d\n",
padname(vec[i].type_name),sizes32[i],vec32[i].alignment);
}
}
void *
emalloc(size_t bytes) {
size_t *memory;
memory = malloc(bytes);
if(memory == 0) {
printf("malloc failed\n");
exit(2);
}
return memory;
}
int
main(int argc, char** argv)
{
int i;
compute_alignments();
verify(vec);
/*
for(i=0;i<NCTYPES;i++) {
printf("%s:\talignment=%d\n",vec[i].type_name,vec[i].alignment);
}
*/
exit(0);
}
#endif /*OFFSETTEST*/