mirror/netcdf-c
2
0
Fork 0
mirror of https://github.com/Unidata/netcdf-c.git synced 2024-11-27 07:30:33 +08:00
Code Issues Packages Projects Releases Wiki Activity
db98c57af6
netcdf-c/ncgen/genbin.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

563 lines
18 KiB
C
Raw Blame History

/*********************************************************************
* Copyright 2018, UCAR/Unidata
* See netcdf/COPYRIGHT file for copying and redistribution conditions.
* $Header: /upc/share/CVS/netcdf-3/ncgen/genbin.c,v 1.4 2010/05/27 21:34:17 dmh Exp $
*********************************************************************/
#include "includes.h"
#include <ctype.h> /* for isprint() */
#include "netcdf_aux.h"
#ifdef ENABLE_BINARY
#undef TRACE
/* Forward*/
static int genbin_defineattr(Symbol* asym);
static int genbin_definevardata(Symbol* vsym);
static int genbin_write(Generator*,Symbol*,Bytebuffer*,int,size_t*,size_t*);
static int genbin_writevar(Generator*,Symbol*,Bytebuffer*,int,size_t*,size_t*);
static int genbin_writeattr(Generator*,Symbol*,Bytebuffer*,int,size_t*,size_t*);
#ifdef USE_NETCDF4
static int genbin_deftype(Symbol* tsym);
static int genbin_definespecialattributes(Symbol* var);
#endif
/*
* Generate C code for creating netCDF from in-memory structure.
*/
void
genbin_netcdf(void)
{
int stat, ncid;
int idim, ivar, iatt;
int ndims, nvars, natts, ngatts;
const char* filename = rootgroup->file.filename;
#ifdef USE_NETCDF4
int ntyps, ngrps, igrp;
#endif
ndims = listlength(dimdefs);
nvars = listlength(vardefs);
natts = listlength(attdefs);
ngatts = listlength(gattdefs);
#ifdef USE_NETCDF4
ntyps = listlength(typdefs);
ngrps = listlength(grpdefs);
#endif /*USE_NETCDF4*/
/* Turn on logging */
#ifdef LOGGING
nc_set_log_level(ncloglevel);
#endif
/* create netCDF file, uses NC_CLOBBER mode */
cmode_modifier |= NC_CLOBBER;
#ifdef USE_NETCDF4
if(!usingclassic)
cmode_modifier |= NC_NETCDF4;
#endif
stat = nc_create(filename, cmode_modifier, &ncid);
CHECK_ERR(stat);
/* ncid created above is also root group*/
rootgroup->nc_id = ncid;
if (nofill_flag) {
stat = nc_set_fill(rootgroup->nc_id, NC_NOFILL, 0);
CHECK_ERR(stat);
}
#ifdef USE_NETCDF4
/* Define the group structure */
/* walking grdefs list will do a preorder walk of all defined groups*/
for(igrp=0;igrp<ngrps;igrp++) {
Symbol* gsym = (Symbol*)listget(grpdefs,igrp);
if(gsym == rootgroup) continue; /* ignore root group*/
stat = nc_def_grp(gsym->container->nc_id,gsym->name,&gsym->nc_id);
CHECK_ERR(stat);
}
#endif
#ifdef USE_NETCDF4
/* Define the types*/
if (ntyps > 0) {
int ityp;
for(ityp = 0; ityp < ntyps; ityp++) {
Symbol* tsym = (Symbol*)listget(typdefs,ityp);
genbin_deftype(tsym);
}
}
#endif
/* define dimensions from info in dims array */
if (ndims > 0) {
for(idim = 0; idim < ndims; idim++) {
Symbol* dsym = (Symbol*)listget(dimdefs,idim);
stat = nc_def_dim(dsym->container->nc_id,
dsym->name,
(dsym->dim.isunlimited?NC_UNLIMITED:dsym->dim.declsize),
&dsym->nc_id);
CHECK_ERR(stat);
}
}
/* define variables from info in vars array */
if (nvars > 0) {
for(ivar = 0; ivar < nvars; ivar++) {
Symbol* vsym = (Symbol*)listget(vardefs,ivar);
if (vsym->typ.dimset.ndims > 0) { /* a dimensioned variable */
/* construct a vector of dimension ids*/
int dimids[NC_MAX_VAR_DIMS];
for(idim=0;idim<vsym->typ.dimset.ndims;idim++)
dimids[idim] = vsym->typ.dimset.dimsyms[idim]->nc_id;
stat = nc_def_var(vsym->container->nc_id,
vsym->name,
vsym->typ.basetype->nc_id,
vsym->typ.dimset.ndims,
dimids,
&vsym->nc_id);
} else { /* a scalar */
stat = nc_def_var(vsym->container->nc_id,
vsym->name,
vsym->typ.basetype->nc_id,
vsym->typ.dimset.ndims,
NULL,
&vsym->nc_id);
}
CHECK_ERR(stat);
}
}
#ifdef USE_NETCDF4
/* define special variable properties */
if(nvars > 0) {
for(ivar = 0; ivar < nvars; ivar++) {
Symbol* var = (Symbol*)listget(vardefs,ivar);
genbin_definespecialattributes(var);
}
}
#endif /*USE_NETCDF4*/
/* define global attributes */
if(ngatts > 0) {
for(iatt = 0; iatt < ngatts; iatt++) {
Symbol* gasym = (Symbol*)listget(gattdefs,iatt);
genbin_defineattr(gasym);
}
}
/* define per-variable attributes */
if(natts > 0) {
for(iatt = 0; iatt < natts; iatt++) {
Symbol* asym = (Symbol*)listget(attdefs,iatt);
genbin_defineattr(asym);
}
}
if (nofill_flag) {
stat = nc_set_fill(rootgroup->nc_id, NC_NOFILL, 0);
CHECK_ERR(stat);
}
/* leave define mode */
stat = nc_enddef(rootgroup->nc_id);
CHECK_ERR(stat);
if(!header_only) {
/* Load values into those variables with defined data */
if(nvars > 0) {
for(ivar = 0; ivar < nvars; ivar++) {
Symbol* vsym = (Symbol*)listget(vardefs,ivar);
if(vsym->data != NULL) {
genbin_definevardata(vsym);
}
}
}
}
}
#ifdef USE_NETCDF4
#if 0
Turn off for now.
static void
genbin_defineglobalspecials(void)
{
int stat = NC_NOERR;
const char* format = NULL;
if(usingclassic) return;
if(!/*Main.*/format_attribute) return;
/* Watch out, this is a global Attribute */
format = kind_string(/*Main.*/format_flag);
stat = nc_put_att_text(rootgroup->nc_id,NC_GLOBAL,"_Format",strlen(format),format);
CHECK_ERR(stat);
}
#endif /*0*/
static int
genbin_definespecialattributes(Symbol* var)
{
int stat = NC_NOERR;
Specialdata* special = &var->var.special;
if(special->flags & _STORAGE_FLAG) {
if(special->_Storage == NC_CONTIGUOUS
|| special->_Storage == NC_COMPACT) {
stat = nc_def_var_chunking(var->container->nc_id, var->nc_id, special->_Storage, NULL);
} else { /* chunked */
if(special->nchunks == 0 || special->_ChunkSizes == NULL)
derror("NC_CHUNKED requested, but no chunksizes specified");
stat = nc_def_var_chunking(var->container->nc_id, var->nc_id, NC_CHUNKED, special->_ChunkSizes);
}
CHECK_ERR(stat);
}
if(special->flags & _FLETCHER32_FLAG) {
stat = nc_def_var_fletcher32(var->container->nc_id,
var->nc_id,
special->_Fletcher32);
CHECK_ERR(stat);
}
if(special->flags & (_DEFLATE_FLAG | _SHUFFLE_FLAG)) {
stat = nc_def_var_deflate(var->container->nc_id,
var->nc_id,
(special->_Shuffle == 1?1:0),
(special->_DeflateLevel >= 0?1:0),
(special->_DeflateLevel >= 0?special->_DeflateLevel
:0));
CHECK_ERR(stat);
}
if(special->flags & _ENDIAN_FLAG) {
stat = nc_def_var_endian(var->container->nc_id,
var->nc_id,
(special->_Endianness == NC_ENDIAN_LITTLE?
NC_ENDIAN_LITTLE
:NC_ENDIAN_BIG));
CHECK_ERR(stat);
}
if(special->flags & _NOFILL_FLAG) {
stat = nc_def_var_fill(var->container->nc_id,
var->nc_id,
(special->_Fill?NC_FILL:NC_NOFILL),
NULL);
CHECK_ERR(stat);
}
if(special->flags & _FILTER_FLAG) {
int k;
for(k=0;k<special->nfilters;k++) {
NC_H5_Filterspec* nfs = special->_Filters[k];
stat = nc_def_var_filter(var->container->nc_id,
var->nc_id,
nfs->filterid,
nfs->nparams,
nfs->params
);
}
CHECK_ERR(stat);
}
return stat;
}
#endif /*USE_NETCDF4*/
void
genbin_close(void)
{
int stat;
stat = nc_close(rootgroup->nc_id);
CHECK_ERR(stat);
}
#ifdef USE_NETCDF4
/*
Generate type definitions
*/
static int
genbin_deftype(Symbol* tsym)
{
unsigned long i;
int stat = 0;
ASSERT(tsym->objectclass == NC_TYPE);
switch (tsym->subclass) {
case NC_PRIM: break; /* these are already taken care of*/
case NC_OPAQUE:
stat = nc_def_opaque(tsym->container->nc_id,
tsym->typ.size,
tsym->name,
&tsym->nc_id);
CHECK_ERR(stat);
break;
case NC_ENUM:
{
Bytebuffer* datum;
stat = nc_def_enum(tsym->container->nc_id,
tsym->typ.basetype->nc_id,
tsym->name,
&tsym->nc_id);
CHECK_ERR(stat);
datum = bbNew();
for(i=0;i<listlength(tsym->subnodes);i++) {
Symbol* econst = (Symbol*)listget(tsym->subnodes,i);
ASSERT(econst->subclass == NC_ECONST);
generator_reset(bin_generator,NULL);
bbClear(datum);
generate_basetype(econst->typ.basetype,econst->typ.econst,datum,NULL,bin_generator);
stat = nc_insert_enum(tsym->container->nc_id,
tsym->nc_id,
econst->name,
bbContents(datum));
CHECK_ERR(stat);
}
bbFree(datum);
}
break;
case NC_VLEN:
stat = nc_def_vlen(tsym->container->nc_id,
tsym->name,
tsym->typ.basetype->nc_id,
&tsym->nc_id);
CHECK_ERR(stat);
break;
case NC_COMPOUND:
stat = nc_def_compound(tsym->container->nc_id,
tsym->typ.size,
tsym->name,
&tsym->nc_id);
CHECK_ERR(stat);
for(i=0;i<listlength(tsym->subnodes);i++) {
Symbol* efield = (Symbol*)listget(tsym->subnodes,i);
ASSERT(efield->subclass == NC_FIELD);
if(efield->typ.dimset.ndims == 0){
stat = nc_insert_compound(
tsym->container->nc_id,
tsym->nc_id,
efield->name,
efield->typ.offset,
efield->typ.basetype->nc_id);
} else {
int j;
int dimsizes[NC_MAX_VAR_DIMS]; /* int because inside compound */
/* Generate the field dimension constants*/
for(j=0;j<efield->typ.dimset.ndims;j++) {
unsigned int size = efield->typ.dimset.dimsyms[j]->dim.declsize;
dimsizes[j] = size;
}
stat = nc_insert_array_compound(
tsym->container->nc_id,
tsym->nc_id,
efield->name,
efield->typ.offset,
efield->typ.basetype->nc_id,
efield->typ.dimset.ndims,
dimsizes);
}
CHECK_ERR(stat);
}
break;
default: panic("definectype: unexpected type subclass");
}
return stat;
}
#endif /*USE_NETCDF4*/
static int
genbin_defineattr(Symbol* asym)
{
int stat = NC_NOERR;
Bytebuffer* databuf = bbNew();
generator_reset(bin_generator,NULL);
generate_attrdata(asym,bin_generator,(Writer)genbin_write,databuf);
stat = nc_reclaim_data(asym->container->nc_id,asym->typ.basetype->nc_id,bbContents(databuf),datalistlen(asym->data));
bbFree(databuf);
return stat;
}
/* Following is patterned after the walk functions in semantics.c */
static int
genbin_definevardata(Symbol* vsym)
{
int stat = NC_NOERR;
Bytebuffer* databuf = NULL;
if(vsym->data == NULL) goto done;
databuf = bbNew();
generator_reset(bin_generator,NULL);
generate_vardata(vsym,bin_generator,(Writer)genbin_write,databuf);
stat = nc_reclaim_data_all(vsym->container->nc_id,vsym->typ.basetype->nc_id,bbExtract(databuf),datalistlen(vsym->data));
done:
bbFree(databuf);
return stat;
}
static int
genbin_write(Generator* generator, Symbol* sym, Bytebuffer* memory,
int rank, size_t* start, size_t* count)
{
if(sym->objectclass == NC_ATT)
return genbin_writeattr(generator,sym,memory,rank,start,count);
else if(sym->objectclass == NC_VAR)
return genbin_writevar(generator,sym,memory,rank,start,count);
else
PANIC("illegal symbol for genbin_write");
return NC_EINVAL;
}
static int
genbin_writevar(Generator* generator, Symbol* vsym, Bytebuffer* memory,
int rank, size_t* start, size_t* count)
{
int stat = NC_NOERR;
int i;
char* data = bbContents(memory);
size_t nelems;
/* Compute total number of elements */
for(nelems=1,i=0;i<rank;i++) nelems *= count[i];
#ifdef GENDEBUG
{
int i;
fprintf(stderr,"startset = [");
for(i=0;i<rank;i++)
fprintf(stderr,"%s%lu",(i>0?", ":""),(unsigned long)start[i]);
fprintf(stderr,"] ");
fprintf(stderr,"countset = [");
for(i=0;i<rank;i++)
fprintf(stderr,"%s%lu",(i>0?", ":""),(unsigned long)count[i]);
fprintf(stderr,"]\n");
fflush(stderr);
}
#endif
if(rank == 0) {
size_t count[1] = {1};
stat = nc_put_var1(vsym->container->nc_id, vsym->nc_id, count, data);
} else {
stat = nc_put_vara(vsym->container->nc_id, vsym->nc_id, start, count, data);
}
CHECK_ERR(stat);
#if 0
/* Reclaim the data */
stat = nc_reclaim_data(vsym->container->nc_id, vsym->typ.basetype->nc_id, data, nelems);
CHECK_ERR(stat);
bbClear(memory); /* reclaim top-level memory */
#endif
return stat;
}
static int
genbin_writeattr(Generator* generator, Symbol* asym, Bytebuffer* databuf,
int rank, size_t* start, size_t* count)
{
int stat = NC_NOERR;
size_t len;
Datalist* list;
int varid, grpid, typid;
Symbol* basetype = asym->typ.basetype;
grpid = asym->container->nc_id;
varid = (asym->att.var == NULL?NC_GLOBAL : asym->att.var->nc_id);
typid = basetype->nc_id;
list = asym->data;
len = list->length;
/* Use the specialized put_att_XX routines if possible*/
if(isprim(typid)) {
switch (basetype->typ.typecode) {
case NC_BYTE: {
signed char* data = (signed char*)bbContents(databuf);
stat = nc_put_att_schar(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_CHAR: {
char* data = (char*)bbContents(databuf);
size_t slen = bbLength(databuf);
/* Revise length if slen == 0 */
if(slen == 0) {
bbAppend(databuf,'\0');
/* bbAppend frees the memory pointed to by char* data,
so re-assign. See Coverity issue: 1265731.*/
data = (char*)bbContents(databuf);
slen++;
}
stat = nc_put_att_text(grpid,varid,asym->name,slen,data);
CHECK_ERR(stat);
} break;
case NC_SHORT: {
short* data = (short*)bbContents(databuf);
stat = nc_put_att_short(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_INT: {
int* data = (int*)bbContents(databuf);
stat = nc_put_att_int(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_FLOAT: {
float* data = (float*)bbContents(databuf);
stat = nc_put_att_float(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_DOUBLE: {
double* data = (double*)bbContents(databuf);
stat = nc_put_att_double(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_STRING: {
const char** data;
data = (const char**)bbContents(databuf);
stat = nc_put_att_string(grpid,varid,asym->name,
bbLength(databuf)/sizeof(char*),
data);
} break;
case NC_UBYTE: {
unsigned char* data = (unsigned char*)bbContents(databuf);
stat = nc_put_att_uchar(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_USHORT: {
unsigned short* data = (unsigned short*)bbContents(databuf);
stat = nc_put_att_ushort(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_UINT: {
unsigned int* data = (unsigned int*)bbContents(databuf);
stat = nc_put_att_uint(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
case NC_INT64: {
long long* data = (long long*)bbContents(databuf);
stat = nc_put_att_longlong(grpid,varid,asym->name,typid,len,data);
CHECK_ERR2(stat,asym->lineno);
} break;
case NC_UINT64: {
unsigned long long* data = (unsigned long long*)bbContents(databuf);
stat = nc_put_att_ulonglong(grpid,varid,asym->name,typid,len,data);
CHECK_ERR(stat);
} break;
default: PANIC1("genbin_defineattr: unexpected basetype: %d",basetype->typ.typecode);
}
} else { /* use the generic put_attribute for user defined types*/
const char* data;
data = (const char*)bbContents(databuf);
stat = nc_put_att(grpid,varid,asym->name,typid,
len,(void*)data);
CHECK_ERR(stat);
#ifdef GENDEBUG
{
char out[4096];
memset(out,0x77,sizeof(out));
stat = nc_get_att(grpid,varid,asym->name,&out);
CHECK_ERR(stat);
}
#endif
}
return stat;
}
#endif /*ENABLE_BINARY*/
Reference in New Issue View Git Blame Copy Permalink