osmith
/
wireshark
forked from osmocom/wireshark
1
0
Fork 0
Code Pull Requests Releases Wiki Activity
wireshark/epan/oids.c

1413 lines
42 KiB
C
Raw Blame History

/* oids.c
* Object IDentifier Support
*
* (c) 2007, Luis E. Garcia Ontanon <luis@ontanon.org>
*
* Wireshark - Network traffic analyzer
* By Gerald Combs <gerald@wireshark.org>
* Copyright 1998 Gerald Combs
*
* SPDX-License-Identifier: GPL-2.0-or-later
*/
#include "config.h"
#include <glib.h>
#include <stdio.h>
#include <string.h>
#include <wsutil/report_message.h>
#include <epan/strutil.h>
#include <epan/wmem_scopes.h>
#include "uat.h"
#include "prefs.h"
#include "proto.h"
#include "packet.h"
#include "wsutil/filesystem.h"
#include "dissectors/packet-ber.h"
#include <wsutil/ws_assert.h>
#ifdef HAVE_LIBSMI
#include <smi.h>
static gboolean smi_init_done = FALSE;
static gboolean oids_init_done = FALSE;
static gboolean load_smi_modules = FALSE;
static gboolean suppress_smi_errors = FALSE;
#endif
#define D(level,args) do if (debuglevel >= level) { printf args; printf("\n"); fflush(stdout); } while(0)
#include "oids.h"
static int debuglevel = 0;
/*
* From SNMPv2-SMI and X.690
*
* Counter32 ::= [APPLICATION 1] IMPLICIT INTEGER (0..4294967295)
* Gauge32 ::= [APPLICATION 2] IMPLICIT INTEGER (0..4294967295)
* Unsigned32 ::= [APPLICATION 2] IMPLICIT INTEGER (0..4294967295) (alias of Gauge32)
* TimeTicks ::= [APPLICATION 3] IMPLICIT INTEGER (0..4294967295)
*
* If the BER encoding should not have the top bit set as to not become a negative number
* the BER encoding may take 5 octets to encode.
*/
#ifdef HAVE_LIBSMI
static const oid_value_type_t integer_type = { FT_INT32, BASE_DEC, BER_CLASS_UNI, BER_UNI_TAG_INTEGER, 1, 4, OID_KEY_TYPE_INTEGER, 1};
static const oid_value_type_t bytes_type = { FT_BYTES, BASE_SHOW_ASCII_PRINTABLE, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 0, -1, OID_KEY_TYPE_BYTES, 0};
static const oid_value_type_t oid_type = { FT_OID, BASE_NONE, BER_CLASS_UNI, BER_UNI_TAG_OID, 1, -1, OID_KEY_TYPE_OID, 0};
static const oid_value_type_t ipv4_type = { FT_IPv4, BASE_NONE, BER_CLASS_APP, 0, 4, 4, OID_KEY_TYPE_IPADDR, 4};
static const oid_value_type_t counter32_type = { FT_UINT64, BASE_DEC, BER_CLASS_APP, 1, 1, 5, OID_KEY_TYPE_INTEGER, 1};
static const oid_value_type_t unsigned32_type = { FT_UINT64, BASE_DEC, BER_CLASS_APP, 2, 1, 5, OID_KEY_TYPE_INTEGER, 1};
static const oid_value_type_t timeticks_type = { FT_UINT64, BASE_DEC, BER_CLASS_APP, 3, 1, 5, OID_KEY_TYPE_INTEGER, 1};
#if 0
static const oid_value_type_t opaque_type = { FT_BYTES, BASE_NONE, BER_CLASS_APP, 4, 1, 4, OID_KEY_TYPE_BYTES, 0};
#endif
static const oid_value_type_t nsap_type = { FT_BYTES, BASE_NONE, BER_CLASS_APP, 5, 0, -1, OID_KEY_TYPE_NSAP, 0};
static const oid_value_type_t counter64_type = { FT_UINT64, BASE_DEC, BER_CLASS_APP, 6, 1, 8, OID_KEY_TYPE_INTEGER, 1};
static const oid_value_type_t ipv6_type = { FT_IPv6, BASE_NONE, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 16, 16, OID_KEY_TYPE_BYTES, 16};
static const oid_value_type_t float_type = { FT_FLOAT, BASE_DEC, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 4, 4, OID_KEY_TYPE_WRONG, 0};
static const oid_value_type_t double_type = { FT_DOUBLE, BASE_DEC, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 8, 8, OID_KEY_TYPE_WRONG, 0};
static const oid_value_type_t ether_type = { FT_ETHER, BASE_NONE, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 6, 6, OID_KEY_TYPE_ETHER, 6};
static const oid_value_type_t string_type = { FT_STRING, BASE_NONE, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 0, -1, OID_KEY_TYPE_STRING, 0};
static const oid_value_type_t date_and_time_type = { FT_STRING, BASE_NONE, BER_CLASS_UNI, BER_UNI_TAG_OCTETSTRING, 8, 11, OID_KEY_TYPE_DATE_AND_TIME, 0};
#endif /* HAVE_LIBSMI */
static const oid_value_type_t unknown_type = { FT_BYTES, BASE_NONE, BER_CLASS_ANY, BER_TAG_ANY, 0, -1, OID_KEY_TYPE_WRONG, 0};
static oid_info_t oid_root = { 0, NULL, OID_KIND_UNKNOWN, NULL, &unknown_type, -2, NULL, NULL, NULL};
static void prepopulate_oids(void) {
if (!oid_root.children) {
char* debug_env = getenv("WIRESHARK_DEBUG_MIBS");
guint32 subid;
debuglevel = debug_env ? (int)strtoul(debug_env,NULL,10) : 0;
oid_root.children = wmem_tree_new(wmem_epan_scope());
/*
* make sure we got strings at least in the three root-children oids
* that way oid_resolved() will always have a string to print
*/
subid = 0; oid_add("itu-t",1,&subid);
subid = 1; oid_add("iso",1,&subid);
subid = 2; oid_add("joint-iso-itu-t",1,&subid);
}
}
static oid_info_t* add_oid(const char* name, oid_kind_t kind, const oid_value_type_t* type, oid_key_t* key, guint oid_len, guint32 *subids) {
guint i = 0;
oid_info_t* c = &oid_root;
prepopulate_oids();
oid_len--;
do {
oid_info_t* n = (oid_info_t *)wmem_tree_lookup32(c->children,subids[i]);
if(n) {
if (i == oid_len) {
if (n->name) {
if (!g_str_equal(n->name,name)) {
D(2,("Renaming Oid from: %s -> %s, this means the same oid is registered more than once",n->name,name));
}
wmem_free(wmem_epan_scope(), n->name);
}
n->name = wmem_strdup(wmem_epan_scope(), name);
if (! n->value_type) {
n->value_type = type;
}
return n;
}
} else {
n = wmem_new(wmem_epan_scope(), oid_info_t);
n->subid = subids[i];
n->kind = kind;
n->children = wmem_tree_new(wmem_epan_scope());
n->value_hfid = -2;
n->key = key;
n->parent = c;
n->bits = NULL;
wmem_tree_insert32(c->children,n->subid,n);
if (i == oid_len) {
n->name = wmem_strdup(wmem_epan_scope(), name);
n->value_type = type;
n->kind = kind;
return n;
} else {
n->name = NULL;
n->value_type = NULL;
n->kind = OID_KIND_UNKNOWN;
}
}
c = n;
} while(++i);
ws_assert_not_reached();
return NULL;
}
void oid_add(const char* name, guint oid_len, guint32 *subids) {
ws_assert(subids && *subids <= 2);
if (oid_len) {
gchar* sub = oid_subid2string(NULL, subids,oid_len);
D(3,("\tOid (from subids): %s %s ",name?name:"NULL", sub));
add_oid(name,OID_KIND_UNKNOWN,NULL,NULL,oid_len,subids);
wmem_free(NULL, sub);
} else {
D(1,("Failed to add Oid: %s (from subids)",name?name:"NULL"));
}
}
void oid_add_from_string(const char* name, const gchar *oid_str) {
guint32* subids;
guint oid_len = oid_string2subid(NULL, oid_str, &subids);
if (oid_len) {
gchar* sub = oid_subid2string(NULL, subids,oid_len);
D(3,("\tOid (from string): %s %s ",name?name:"NULL", sub));
add_oid(name,OID_KIND_UNKNOWN,NULL,NULL,oid_len,subids);
wmem_free(NULL, sub);
} else {
D(1,("Failed to add Oid: %s %s ",name?name:"NULL", oid_str?oid_str:NULL));
}
wmem_free(NULL, subids);
}
extern void oid_add_from_encoded(const char* name, const guint8 *oid, gint oid_len) {
guint32* subids = NULL;
guint subids_len = oid_encoded2subid(NULL, oid, oid_len, &subids);
if (subids_len) {
gchar* sub = oid_subid2string(NULL, subids,subids_len);
D(3,("\tOid (from encoded): %s %s ",name, sub));
add_oid(name,OID_KIND_UNKNOWN,NULL,NULL,subids_len,subids);
wmem_free(NULL, sub);
} else {
gchar* bytestr = bytes_to_str_punct(NULL, oid, oid_len, ':');
D(1,("Failed to add Oid: %s [%d]%s ",name?name:"NULL", oid_len, bytestr));
wmem_free(NULL, bytestr);
}
wmem_free(NULL, subids);
}
#ifdef HAVE_LIBSMI
/* de-allocate storage mallocated by libsmi */
/* */
/* XXX: libsmi provides access to smiFree as of libsmi v 0.4.8. */
/* On Windows: Wireshark 1.01 and later is built and distributed */
/* with libsmi 0.4.8 (or newer). */
/* On non-Windows systems, free() should be OK for libsmi */
/* versions older than 0.4.8. */
static void smi_free(void *ptr) {
#if (SMI_VERSION_MAJOR > 0) || (SMI_VERSION_MINOR > 4) || (SMI_VERSION_PATCHLEVEL >= 8)
smiFree(ptr);
#else
#ifdef _WIN32
#error Unsupported Windows libsmi version < 0.4.8
#endif
#define xx_free free /* hack so checkAPIs.pl doesn't complain */
xx_free(ptr);
#endif
}
typedef struct smi_module_t {
char* name;
} smi_module_t;
static smi_module_t* smi_paths = NULL;
static guint num_smi_paths = 0;
static uat_t* smi_paths_uat = NULL;
static smi_module_t* smi_modules = NULL;
static guint num_smi_modules = 0;
static uat_t* smi_modules_uat = NULL;
static GString* smi_errors;
UAT_DIRECTORYNAME_CB_DEF(smi_mod,name,smi_module_t)
static void smi_error_handler(char *path, int line, int severity, char *msg, char *tag) {
g_string_append_printf(smi_errors,"%s:%d %d %s %s\n",
path ? path : "-",
line, severity,
tag ? tag : "-",
msg ? msg : "");
}
static void* smi_mod_copy_cb(void* dest, const void* orig, size_t len _U_) {
const smi_module_t* m = (const smi_module_t*)orig;
smi_module_t* d = (smi_module_t*)dest;
d->name = g_strdup(m->name);
return d;
}
static void smi_mod_free_cb(void* p) {
smi_module_t* m = (smi_module_t*)p;
g_free(m->name);
}
static char* alnumerize(const char* name) {
char* s = g_strdup(name);
char* r = s;
char* w = r;
char c;
for (;(c = *r); r++) {
if (g_ascii_isalnum(c) || c == '_' || c == '-' || c == '.') {
*(w++) = c;
} else if (c == ':' && r[1] == ':') {
*(w++) = '.';
}
}
*w = '\0';
return s;
}
static const oid_value_type_t* get_typedata(SmiType* smiType) {
/*
* There has to be a better way to know if a given
* OCTETSTRING type is actually human readable text,
* an address of some type or some moe specific FT_
* Until that is found, this is the mappping between
* SNMP Types and our FT_s
*/
static const struct _type_mapping_t {
const char* name;
SmiBasetype base;
const oid_value_type_t* type;
} types[] = {
{"IpAddress", SMI_BASETYPE_UNKNOWN, &ipv4_type},
{"InetAddressIPv4",SMI_BASETYPE_UNKNOWN,&ipv4_type},
{"InetAddressIPv6",SMI_BASETYPE_UNKNOWN,&ipv6_type},
{"NetworkAddress",SMI_BASETYPE_UNKNOWN,&ipv4_type},
{"MacAddress",SMI_BASETYPE_UNKNOWN,&ether_type},
{"TimeTicks",SMI_BASETYPE_UNKNOWN,&timeticks_type},
{"Ipv6Address",SMI_BASETYPE_UNKNOWN,&ipv6_type},
{"TimeStamp",SMI_BASETYPE_UNKNOWN,&timeticks_type},
{"DisplayString",SMI_BASETYPE_UNKNOWN,&string_type},
{"SnmpAdminString",SMI_BASETYPE_UNKNOWN,&string_type},
{"DateAndTime",SMI_BASETYPE_UNKNOWN,&date_and_time_type},
{"Counter",SMI_BASETYPE_UNKNOWN,&counter32_type},
{"Counter32",SMI_BASETYPE_UNKNOWN,&counter32_type},
{"Unsigned32",SMI_BASETYPE_UNKNOWN,&unsigned32_type},
{"Gauge",SMI_BASETYPE_UNKNOWN,&unsigned32_type},
{"Gauge32",SMI_BASETYPE_UNKNOWN,&unsigned32_type},
{"NsapAddress",SMI_BASETYPE_UNKNOWN,&nsap_type},
{"i32",SMI_BASETYPE_INTEGER32,&integer_type},
{"octets",SMI_BASETYPE_OCTETSTRING,&bytes_type},
{"oid",SMI_BASETYPE_OBJECTIDENTIFIER,&oid_type},
{"u32",SMI_BASETYPE_UNSIGNED32,&unsigned32_type},
{"u64",SMI_BASETYPE_UNSIGNED64,&counter64_type},
{"f32",SMI_BASETYPE_FLOAT32,&float_type},
{"f64",SMI_BASETYPE_FLOAT64,&double_type},
{"f128",SMI_BASETYPE_FLOAT128,&bytes_type},
{"enum",SMI_BASETYPE_ENUM,&integer_type},
{"bits",SMI_BASETYPE_BITS,&bytes_type},
{"unk",SMI_BASETYPE_UNKNOWN,&unknown_type},
{NULL,SMI_BASETYPE_UNKNOWN,NULL} /* SMI_BASETYPE_UNKNOWN = 0 */
};
const struct _type_mapping_t* t;
SmiType* sT = smiType;
if (!smiType) return NULL;
do {
for (t = types; t->type ; t++ ) {
char* name = smiRenderType(sT, SMI_RENDER_NAME);
if (name && t->name && g_str_equal(name, t->name )) {
smi_free(name);
return t->type;
}
if (name) {
smi_free (name);
}
}
} while(( sT = smiGetParentType(sT) ));
for (t = types; t->type ; t++ ) {
if(smiType->basetype == t->base) {
return t->type;
}
}
return &unknown_type;
}
static guint get_non_implicit_size(SmiType* sT) {
SmiRange *sR;
guint size = 0xffffffff;
switch (sT->basetype) {
case SMI_BASETYPE_OCTETSTRING:
case SMI_BASETYPE_OBJECTIDENTIFIER:
break;
default:
return 0;
}
for ( ; sT; sT = smiGetParentType(sT) ) {
for (sR = smiGetFirstRange(sT); sR ; sR = smiGetNextRange(sR)) {
if (size == 0xffffffff) {
if (sR->minValue.value.unsigned32 == sR->maxValue.value.unsigned32) {
size = (guint32)sR->minValue.value.unsigned32;
} else {
return 0;
}
} else {
if (sR->minValue.value.unsigned32 != size || sR->maxValue.value.unsigned32 != size) {
return 0;
}
}
}
}
return size == 0xffffffff ? 0 : size;
}
static inline oid_kind_t smikind(SmiNode* sN, oid_key_t** key_p) {
*key_p = NULL;
switch(sN->nodekind) {
case SMI_NODEKIND_ROW: {
SmiElement* sE;
oid_key_t* kl = NULL; /* points to last element in the list of oid_key_t's */
const oid_value_type_t* typedata = NULL;
gboolean implied;
switch (sN->indexkind) {
case SMI_INDEX_INDEX:
break;
case SMI_INDEX_AUGMENT:
case SMI_INDEX_REORDER:
case SMI_INDEX_SPARSE:
case SMI_INDEX_EXPAND:
sN = smiGetRelatedNode(sN);
break;
case SMI_INDEX_UNKNOWN:
return OID_KIND_UNKNOWN;
};
implied = sN->implied;
for (sE = smiGetFirstElement(sN); sE; sE = smiGetNextElement(sE)) {
SmiNode* elNode = smiGetElementNode(sE) ;
SmiType* elType = smiGetNodeType(elNode);
oid_key_t* k;
guint non_implicit_size = 0;
char *oid1, *oid2;
if (elType) {
non_implicit_size = get_non_implicit_size(elType);
}
typedata = get_typedata(elType);
k = g_new(oid_key_t,1);
oid1 = smiRenderOID(sN->oidlen, sN->oid, SMI_RENDER_QUALIFIED);
oid2 = smiRenderOID(elNode->oidlen, elNode->oid, SMI_RENDER_NAME);
k->name = g_strconcat(oid1, ".", oid2, NULL);
smi_free (oid1);
smi_free (oid2);
k->hfid = -2;
k->ft_type = typedata ? typedata->ft_type : FT_BYTES;
k->display = typedata ? typedata->display : BASE_NONE;
k->next = NULL;
if (typedata) {
k->key_type = typedata->keytype;
k->num_subids = typedata->keysize;
} else {
if (elType) {
switch (elType->basetype) {
case SMI_BASETYPE_BITS:
case SMI_BASETYPE_OCTETSTRING: {
k->key_type = OID_KEY_TYPE_BYTES;
k->num_subids = non_implicit_size;
break;
}
case SMI_BASETYPE_ENUM:
case SMI_BASETYPE_OBJECTIDENTIFIER:
case SMI_BASETYPE_INTEGER32:
case SMI_BASETYPE_UNSIGNED32:
case SMI_BASETYPE_INTEGER64:
case SMI_BASETYPE_UNSIGNED64:
k->key_type = OID_KEY_TYPE_INTEGER;
k->num_subids = 1;
break;
default:
k->key_type = OID_KEY_TYPE_WRONG;
k->num_subids = 0;
break;
}
} else {
k->key_type = OID_KEY_TYPE_WRONG;
k->num_subids = 0;
}
}
if (!kl) {
/*
* The list is empty, so set the
* pointer to the head of the list
* to point to this entry.
*/
*key_p = k;
} else {
/*
* The list is non-empty, and kl
* points to its last element.
* Make the last element point to
* this entry as its successor.
*/
kl->next = k;
}
/*
* This entry is now the last entry in
* the list.
*/
kl = k;
}
if (implied && kl) {
switch (kl->key_type) {
case OID_KEY_TYPE_BYTES: kl->key_type = OID_KEY_TYPE_IMPLIED_BYTES; break;
case OID_KEY_TYPE_STRING: kl->key_type = OID_KEY_TYPE_IMPLIED_STRING; break;
case OID_KEY_TYPE_OID: kl->key_type = OID_KEY_TYPE_IMPLIED_OID; break;
default: break;
}
}
return OID_KIND_ROW;
}
case SMI_NODEKIND_NODE: return OID_KIND_NODE;
case SMI_NODEKIND_SCALAR: return OID_KIND_SCALAR;
case SMI_NODEKIND_TABLE: return OID_KIND_TABLE;
case SMI_NODEKIND_COLUMN: return OID_KIND_COLUMN;
case SMI_NODEKIND_NOTIFICATION: return OID_KIND_NOTIFICATION;
case SMI_NODEKIND_GROUP: return OID_KIND_GROUP;
case SMI_NODEKIND_COMPLIANCE: return OID_KIND_COMPLIANCE;
case SMI_NODEKIND_CAPABILITIES: return OID_KIND_CAPABILITIES;
default: return OID_KIND_UNKNOWN;
}
}
#define IS_ENUMABLE(ft) ( (ft == FT_UINT8) || (ft == FT_UINT16) || (ft == FT_UINT24) || (ft == FT_UINT32) \
|| (ft == FT_INT8) || (ft == FT_INT16) || (ft == FT_INT24) || (ft == FT_INT32) \
|| (ft == FT_UINT64) || (ft == FT_INT64) )
static void unregister_mibs(void) {
/* TODO: Unregister "MIBs" proto and clean up field array and subtree array.
* Wireshark does not support that yet. :-( */
/* smiExit(); */
}
static void restart_needed_warning(void) {
if (oids_init_done)
report_failure("Wireshark needs to be restarted for these changes to take effect");
}
static void register_mibs(void) {
SmiModule *smiModule;
SmiNode *smiNode;
guint i;
int proto_mibs = -1;
wmem_array_t* hfa;
GArray* etta;
gchar* path_str;
if (!load_smi_modules) {
D(1,("OID resolution not enabled"));
return;
}
/* TODO: Remove this workaround when unregistration of "MIBs" proto is solved.
* Wireshark does not support that yet. :-( */
if (oids_init_done) {
D(1,("Exiting register_mibs() to avoid double registration of MIBs proto."));
return;
}
hfa = wmem_array_new(wmem_epan_scope(), sizeof(hf_register_info));
etta = g_array_new(FALSE,TRUE,sizeof(gint*));
smiInit(NULL);
smi_init_done = TRUE;
smi_errors = g_string_new("");
smiSetErrorHandler(smi_error_handler);
path_str = oid_get_default_mib_path();
D(1,("SMI Path: '%s'",path_str));
smiSetPath(path_str);
for(i=0;i<num_smi_modules;i++) {
if (!smi_modules[i].name) continue;
if (smiIsLoaded(smi_modules[i].name)) {
continue;
} else {
char* mod_name = smiLoadModule(smi_modules[i].name);
if (mod_name)
D(2,("Loaded: '%s'[%u] as %s",smi_modules[i].name,i,mod_name ));
else
D(1,("Failed to load: '%s'[%u]",smi_modules[i].name,i));
}
}
if (smi_errors->len) {
if (!suppress_smi_errors) {
report_failure("The following errors were found while loading the MIBS:\n%s\n\n"
"The Current Path is: %s\n\nYou can avoid this error message "
"by removing the missing MIB modules at Edit -> Preferences"
" -> Name Resolution -> SMI (MIB and PIB) modules or by "
"installing them.\n" , smi_errors->str , path_str);
}
D(1,("Errors while loading:\n%s\n",smi_errors->str));
}
g_free(path_str);
g_string_free(smi_errors,TRUE);
for (smiModule = smiGetFirstModule();
smiModule;
smiModule = smiGetNextModule(smiModule)) {
D(3,("\tModule: %s", smiModule->name));
/* TODO: Check libsmi version at compile time and disable this
* workaround for libsmi versions where this problem is fixed.
* Currently there is no such version. :-(
*/
if (smiModule->conformance == 1) {
if (!suppress_smi_errors) {
report_failure("Stopped processing module %s due to "
"error(s) to prevent potential crash in libsmi.\n"
"Module's conformance level: %d.\n"
"See details at: https://bugs.debian.org/560325\n",
smiModule->name, smiModule->conformance);
}
continue;
}
for (smiNode = smiGetFirstNode(smiModule, SMI_NODEKIND_ANY);
smiNode;
smiNode = smiGetNextNode(smiNode, SMI_NODEKIND_ANY)) {
SmiType* smiType = smiGetNodeType(smiNode);
const oid_value_type_t* typedata = get_typedata(smiType);
oid_key_t* key;
oid_kind_t kind = smikind(smiNode,&key);
char *sub;
char *oid = smiRenderOID(smiNode->oidlen, smiNode->oid, SMI_RENDER_QUALIFIED);
oid_info_t* oid_data = add_oid(oid,
kind,
typedata,
key,
smiNode->oidlen,
smiNode->oid);
smi_free (oid);
sub = oid_subid2string(NULL, smiNode->oid, smiNode->oidlen);
D(4,("\t\tNode: kind=%d oid=%s name=%s ",
oid_data->kind, sub, oid_data->name));
wmem_free(NULL, sub);
if ( typedata && oid_data->value_hfid == -2 ) {
SmiNamedNumber* smiEnum;
hf_register_info hf;
char *name;
char *blurb;
name = g_strdup(oid_data->name);
blurb = smiRenderOID(smiNode->oidlen, smiNode->oid, SMI_RENDER_ALL);
/* Don't allow duplicate blurb/name */
if (strcmp(blurb, name) == 0) {
smi_free(blurb);
blurb = NULL;
}
hf.p_id = &(oid_data->value_hfid);
hf.hfinfo.name = name;
hf.hfinfo.abbrev = alnumerize(oid_data->name);
hf.hfinfo.type = typedata->ft_type;
hf.hfinfo.display = typedata->display;
hf.hfinfo.strings = NULL;
hf.hfinfo.bitmask = 0;
hf.hfinfo.blurb = blurb;
/* HFILL */
HFILL_INIT(hf);
oid_data->value_hfid = -1;
if ( IS_ENUMABLE(hf.hfinfo.type) && (smiEnum = smiGetFirstNamedNumber(smiType))) {
GArray* vals = g_array_new(TRUE,TRUE,sizeof(value_string));
for(;smiEnum; smiEnum = smiGetNextNamedNumber(smiEnum)) {
if (smiEnum->name) {
value_string val;
val.value = (guint32)smiEnum->value.value.integer32;
val.strptr = g_strdup(smiEnum->name);
g_array_append_val(vals,val);
}
}
hf.hfinfo.strings = g_array_free(vals, FALSE);
}
#if 0 /* packet-snmp does not handle bits yet */
} else if (smiType->basetype == SMI_BASETYPE_BITS && ( smiEnum = smiGetFirstNamedNumber(smiType) )) {
guint n = 0;
oid_bits_info_t* bits = g_new(oid_bits_info_t, 1);
gint* ettp = &(bits->ett);
bits->num = 0;
bits->ett = -1;
g_array_append_val(etta,ettp);
for(;smiEnum; smiEnum = smiGetNextNamedNumber(smiEnum), bits->num++);
bits->data = g_malloc(sizeof(struct _oid_bit_t)*bits->num);
for(smiEnum = smiGetFirstNamedNumber(smiType),n=0;
smiEnum;
smiEnum = smiGetNextNamedNumber(smiEnum),n++) {
guint mask = 1 << (smiEnum->value.value.integer32 % 8);
char* base = alnumerize(oid_data->name);
char* ext = alnumerize(smiEnum->name);
hf_register_info hf2 = { &(bits->data[n].hfid), { NULL, NULL, FT_UINT8, BASE_HEX, NULL, mask, NULL, HFILL }};
bits->data[n].hfid = -1;
bits->data[n].offset = smiEnum->value.value.integer32 / 8;
hf2.hfinfo.name = g_strconcat("%s:%s",oid_data->name, ":", smiEnum->name, NULL);
hf2.hfinfo.abbrev = g_strconcat(base, ".", ext, NULL);
g_free(base);
g_free(ext);
g_array_append_val(hfa,hf2);
}
#endif /* packet-snmp does not use this yet */
wmem_array_append_one(hfa,hf);
}
if ((key = oid_data->key)) {
for(; key; key = key->next) {
D(5,("\t\t\tIndex: name=%s subids=%u key_type=%d",
key->name, key->num_subids, key->key_type ));
if (key->hfid == -2) {
hf_register_info hf;
hf.p_id = &(key->hfid);
hf.hfinfo.name = key->name;
hf.hfinfo.abbrev = alnumerize(key->name);
hf.hfinfo.type = key->ft_type;
hf.hfinfo.display = key->display;
hf.hfinfo.strings = NULL;
hf.hfinfo.bitmask = 0;
hf.hfinfo.blurb = NULL;
/* HFILL */
HFILL_INIT(hf);
wmem_array_append_one(hfa,hf);
key->hfid = -1;
}
}
}
}
}
proto_mibs = proto_register_protocol("MIBs", "MIBS", "mibs");
proto_register_field_array(proto_mibs, (hf_register_info*)wmem_array_get_raw(hfa), wmem_array_get_count(hfa));
proto_register_subtree_array((gint**)(void*)etta->data, etta->len);
g_array_free(etta,TRUE);
oids_init_done = TRUE;
}
#endif
void oid_pref_init(module_t *nameres)
{
#ifdef HAVE_LIBSMI
static uat_field_t smi_fields[] = {
UAT_FLD_CSTRING(smi_mod,name,"Module name","The module's name"),
UAT_END_FIELDS
};
static uat_field_t smi_paths_fields[] = {
UAT_FLD_DIRECTORYNAME(smi_mod,name,"Directory path","The directory name"),
UAT_END_FIELDS
};
prefs_register_bool_preference(nameres, "load_smi_modules",
"Enable OID resolution",
"Resolve Object IDs to object names from the MIB and PIB"
" modules defined below."
" You must restart Wireshark for this change to take effect",
&load_smi_modules);
prefs_register_bool_preference(nameres, "suppress_smi_errors",
"Suppress SMI errors",
"While loading MIB or PIB modules errors may be detected,"
" which are reported. Some errors can be ignored."
" If unsure, set to false.",
&suppress_smi_errors);
smi_paths_uat = uat_new("SMI Paths",
sizeof(smi_module_t),
"smi_paths",
FALSE,
(void**)&smi_paths,
&num_smi_paths,
/* affects dissection of packets (as the MIBs and PIBs affect the
interpretation of e.g. SNMP variable bindings), but not set of
named fields
XXX - if named fields are generated from the MIBs and PIBs
for particular variable bindings, this *does* affect the set
of named fields! */
UAT_AFFECTS_DISSECTION,
"ChSNMPSMIPaths",
smi_mod_copy_cb,
NULL,
smi_mod_free_cb,
restart_needed_warning,
NULL,
smi_paths_fields);
prefs_register_uat_preference(nameres,
"smi_paths",
"SMI (MIB and PIB) paths",
"Search paths for SMI (MIB and PIB) modules. You must"
" restart Wireshark for these changes to take effect.",
smi_paths_uat);
smi_modules_uat = uat_new("SMI Modules",
sizeof(smi_module_t),
"smi_modules",
FALSE,
(void**)&smi_modules,
&num_smi_modules,
/* affects dissection of packets (as the MIBs and PIBs affect the
interpretation of e.g. SNMP variable bindings), but not set of
named fields
XXX - if named fields are generated from the MIBs and PIBs
for particular variable bindings, would this affect the set
of named fields? */
UAT_AFFECTS_DISSECTION,
"ChSNMPSMIModules",
smi_mod_copy_cb,
NULL,
smi_mod_free_cb,
restart_needed_warning,
NULL,
smi_fields);
prefs_register_uat_preference(nameres,
"smi_modules",
"SMI (MIB and PIB) modules",
"List of SMI (MIB and PIB) modules to load. You must"
" restart Wireshark for these changes to take effect.",
smi_modules_uat);
#else
prefs_register_static_text_preference(nameres, "load_smi_modules_static",
"Enable OID resolution: N/A",
"Support for OID resolution was not compiled into this version of Wireshark");
prefs_register_static_text_preference(nameres, "suppress_smi_errors_static",
"Suppress SMI errors: N/A",
"Support for OID resolution was not compiled into this version of Wireshark");
prefs_register_static_text_preference(nameres, "smi_module_path",
"SMI (MIB and PIB) modules and paths: N/A",
"Support for OID resolution was not compiled into this version of Wireshark");
#endif
}
void oids_init(void) {
prepopulate_oids();
#ifdef HAVE_LIBSMI
register_mibs();
#else
D(1,("libsmi disabled oid resolution not enabled"));
#endif
}
void oids_cleanup(void) {
#ifdef HAVE_LIBSMI
unregister_mibs();
#else
D(1,("libsmi disabled oid resolution not enabled"));
#endif
}
char* oid_subid2string(wmem_allocator_t *scope, guint32* subids, guint len) {
return rel_oid_subid2string(scope, subids, len, TRUE);
}
char* rel_oid_subid2string(wmem_allocator_t *scope, guint32* subids, guint len, gboolean is_absolute) {
wmem_strbuf_t *oid_str;
gsize oid_str_len;
if(!subids || len == 0)
return wmem_strdup(scope, "*** Empty OID ***");
oid_str = wmem_strbuf_new(scope, "");
if (!is_absolute)
wmem_strbuf_append_c(oid_str, '.');
do {
wmem_strbuf_append_printf(oid_str, "%u.",*subids++);
} while(--len);
/* Remove trailing "." (which is guaranteed to be there) */
oid_str_len = wmem_strbuf_get_len(oid_str);
wmem_strbuf_truncate(oid_str, oid_str_len - 1);
return wmem_strbuf_finalize(oid_str);
}
static guint check_num_oid(const char* str) {
const char* r = str;
char c = '.';
guint n = 0;
D(8,("check_num_oid: '%s'",str));
if (!r) return 0;
do {
D(9,("\tcheck_num_oid: '%c' %u",*r,n));
switch(*r) {
case '.': case '\0':
n++;
if (c == '.') return 0;
break;
case '1' : case '2' : case '3' : case '4' : case '5' :
case '6' : case '7' : case '8' : case '9' : case '0' :
continue;
default:
return 0;
}
} while((c = *r++));
return n;
}
guint oid_string2subid(wmem_allocator_t *scope, const char* str, guint32** subids_p) {
const char* r = str;
guint32* subids;
guint32* subids_overflow;
guint n = check_num_oid(str);
/*
* we cannot handle sub-ids greater than 32bytes
* keep a pilot subid of 64 bytes to check the limit
*/
guint64 subid = 0;
D(6,("oid_string2subid: str='%s'",str));
if (!n) {
*subids_p = NULL;
return 0;
}
*subids_p = subids = wmem_alloc0_array(scope, guint32, n);
subids_overflow = subids + n;
do switch(*r) {
case '.':
subid = 0;
subids++;
continue;
case '1' : case '2' : case '3' : case '4' : case '5' :
case '6' : case '7' : case '8' : case '9' : case '0' :
subid *= 10;
subid += *r - '0';
if( subids >= subids_overflow || subid > 0xffffffff) {
wmem_free(scope, *subids_p);
*subids_p=NULL;
return 0;
}
*(subids) *= 10;
*(subids) += *r - '0';
continue;
case '\0':
break;
default:
return 0;
} while(*r++);
return n;
}
guint oid_encoded2subid(wmem_allocator_t *scope, const guint8 *oid_bytes, gint oid_len, guint32** subids_p) {
return oid_encoded2subid_sub(scope, oid_bytes, oid_len, subids_p, TRUE);
}
guint oid_encoded2subid_sub(wmem_allocator_t *scope, const guint8 *oid_bytes, gint oid_len, guint32** subids_p,
gboolean is_first) {
gint i;
guint n = is_first ? 1 : 0;
guint32* subids;
guint32* subid_overflow;
/*
* we cannot handle sub-ids greater than 32bytes
* have the subid in 64 bytes to be able to check the limit
*/
guint64 subid = 0;
for (i=0; i<oid_len; i++) { if (! (oid_bytes[i] & 0x80 )) n++; }
*subids_p = subids = (guint32 *)wmem_alloc(scope, sizeof(guint32)*n);
subid_overflow = subids+n;
/* If n is 0 or 1 (depending on how it was initialized) then we found
* no bytes in the OID with first bit cleared, so initialize our one
* byte (if any) to zero and return. This *seems* to be the right thing
* to do in this situation, and at the very least it avoids
* uninitialized memory errors that would otherwise occur. */
if (is_first && n == 1) {
*subids = 0;
return n;
}
else if (!is_first && n == 0) {
return n;
}
for (i=0; i<oid_len; i++){
guint8 byte = oid_bytes[i];
subid <<= 7;
subid |= byte & 0x7F;
if (byte & 0x80) {
continue;
}
if (is_first) {
guint32 subid0 = 0;
if (subid >= 40) { subid0++; subid-=40; }
if (subid >= 40) { subid0++; subid-=40; }
*subids++ = subid0;
is_first = FALSE;
}
if( subids >= subid_overflow || subid > 0xffffffff) {
/* scope may be NULL in which case we must free our
* useless buffer before returning */
wmem_free(scope, *subids_p);
*subids_p = NULL;
return 0;
}
*subids++ = (guint32)subid;
subid = 0;
}
ws_assert(subids == subid_overflow);
return n;
}
oid_info_t* oid_get(guint len, guint32* subids, guint* matched, guint* left) {
oid_info_t* curr_oid = &oid_root;
guint i;
if(!(subids && *subids <= 2)) {
*matched = 0;
*left = len;
return curr_oid;
}
for( i=0; i < len; i++) {
oid_info_t* next_oid = (oid_info_t *)wmem_tree_lookup32(curr_oid->children,subids[i]);
if (next_oid) {
curr_oid = next_oid;
} else {
goto done;
}
}
done:
*matched = i;
*left = len - i;
return curr_oid;
}
oid_info_t* oid_get_from_encoded(wmem_allocator_t *scope, const guint8 *bytes, gint byteslen, guint32** subids_p, guint* matched_p, guint* left_p) {
guint subids_len = oid_encoded2subid(scope, bytes, byteslen, subids_p);
return oid_get(subids_len, *subids_p, matched_p, left_p);
}
oid_info_t* oid_get_from_string(wmem_allocator_t *scope, const gchar *oid_str, guint32** subids_p, guint* matched, guint* left) {
guint subids_len = oid_string2subid(scope, oid_str, subids_p);
return oid_get(subids_len, *subids_p, matched, left);
}
gchar *oid_resolved_from_encoded(wmem_allocator_t *scope, const guint8 *oid, gint oid_len) {
guint32 *subid_oid = NULL;
gchar * ret;
guint subid_oid_length = oid_encoded2subid(NULL, oid, oid_len, &subid_oid);
ret = oid_resolved(scope, subid_oid_length, subid_oid);
wmem_free(NULL, subid_oid);
return ret;
}
gchar *rel_oid_resolved_from_encoded(wmem_allocator_t *scope, const guint8 *oid, gint oid_len) {
guint32 *subid_oid = NULL;
gchar* ret;
guint subid_oid_length = oid_encoded2subid_sub(NULL, oid, oid_len, &subid_oid, FALSE);
ret = rel_oid_subid2string(scope, subid_oid, subid_oid_length, FALSE);
wmem_free(NULL, subid_oid);
return ret;
}
guint oid_subid2encoded(wmem_allocator_t *scope, guint subids_len, guint32* subids, guint8** bytes_p) {
guint bytelen = 0;
guint i;
guint32 subid;
guint8* b;
if ( !subids || subids_len <= 1) {
*bytes_p = NULL;
return 0;
}
for (subid=subids[0] * 40, i = 1; i<subids_len; i++, subid=0) {
subid += subids[i];
if (subid <= 0x0000007F) {
bytelen += 1;
} else if (subid <= 0x00003FFF ) {
bytelen += 2;
} else if (subid <= 0x001FFFFF ) {
bytelen += 3;
} else if (subid <= 0x0FFFFFFF ) {
bytelen += 4;
} else {
bytelen += 5;
}
}
*bytes_p = b = (guint8 *)wmem_alloc(scope, bytelen);
for (subid=subids[0] * 40, i = 1; i<subids_len; i++, subid=0) {
guint len;
subid += subids[i];
if ((subid <= 0x0000007F )) len = 1;
else if ((subid <= 0x00003FFF )) len = 2;
else if ((subid <= 0x001FFFFF )) len = 3;
else if ((subid <= 0x0FFFFFFF )) len = 4;
else len = 5;
switch(len) {
default: *bytes_p=NULL; return 0;
case 5: *(b++) = ((subid & 0xF0000000) >> 28) | 0x80;
/* FALL THROUGH */
case 4: *(b++) = ((subid & 0x0FE00000) >> 21) | 0x80;
/* FALL THROUGH */
case 3: *(b++) = ((subid & 0x001FC000) >> 14) | 0x80;
/* FALL THROUGH */
case 2: *(b++) = ((subid & 0x00003F80) >> 7) | 0x80;
/* FALL THROUGH */
case 1: *(b++) = subid & 0x0000007F ; break;
}
}
return bytelen;
}
gchar* oid_encoded2string(wmem_allocator_t *scope, const guint8* encoded, guint len) {
guint32* subids = NULL;
gchar* ret;
guint subids_len = oid_encoded2subid(NULL, encoded, len, &subids);
if (subids_len) {
ret = oid_subid2string(scope, subids,subids_len);
} else {
ret = wmem_strdup(scope, "");
}
wmem_free(NULL, subids);
return ret;
}
gchar* rel_oid_encoded2string(wmem_allocator_t *scope, const guint8* encoded, guint len) {
guint32* subids = NULL;
gchar* ret;
guint subids_len = oid_encoded2subid_sub(NULL, encoded, len, &subids, FALSE);
if (subids_len) {
ret = rel_oid_subid2string(scope, subids,subids_len, FALSE);
} else {
ret = wmem_strdup(scope, "");
}
wmem_free(NULL, subids);
return ret;
}
guint oid_string2encoded(wmem_allocator_t *scope, const char *oid_str, guint8 **bytes) {
guint32* subids;
guint32 subids_len;
guint byteslen;
if ( (subids_len = oid_string2subid(NULL, oid_str, &subids)) &&
(byteslen = oid_subid2encoded(scope, subids_len, subids, bytes)) ) {
wmem_free(NULL, subids);
return byteslen;
}
wmem_free(NULL, subids);
return 0;
}
gchar *oid_resolved_from_string(wmem_allocator_t *scope, const gchar *oid_str) {
guint32 *subid_oid;
guint subid_oid_length;
gchar *resolved;
subid_oid_length = oid_string2subid(NULL, oid_str, &subid_oid);
resolved = oid_resolved(scope, subid_oid_length, subid_oid);
wmem_free(NULL, subid_oid);
return resolved;
}
gchar *oid_resolved(wmem_allocator_t *scope, guint32 num_subids, guint32* subids) {
guint matched;
guint left;
oid_info_t* oid;
if(! (subids && *subids <= 2 ))
return wmem_strdup(scope, "*** Malformed OID ***");
oid = oid_get(num_subids, subids, &matched, &left);
while (! oid->name ) {
if (!(oid = oid->parent)) {
return oid_subid2string(scope, subids,num_subids);
}
left++;
matched--;
}
if (left) {
gchar *ret,
*str1 = oid_subid2string(NULL, subids,matched),
*str2 = oid_subid2string(NULL, &(subids[matched]),left);
ret = wmem_strconcat(scope, oid->name ? oid->name : str1, ".", str2, NULL);
wmem_free(NULL, str1);
wmem_free(NULL, str2);
return ret;
} else {
return oid->name ? wmem_strdup(scope, oid->name) : oid_subid2string(scope, subids,matched);
}
}
extern void oid_both(wmem_allocator_t *scope, guint oid_len, guint32 *subids, gchar** resolved_p, gchar** numeric_p) {
*resolved_p = oid_resolved(scope, oid_len,subids);
*numeric_p = oid_subid2string(scope, subids,oid_len);
}
extern void oid_both_from_encoded(wmem_allocator_t *scope, const guint8 *oid, gint oid_len, gchar** resolved_p, gchar** numeric_p) {
guint32* subids = NULL;
guint subids_len = oid_encoded2subid(NULL, oid, oid_len, &subids);
*resolved_p = oid_resolved(scope, subids_len,subids);
*numeric_p = oid_subid2string(scope, subids,subids_len);
wmem_free(NULL, subids);
}
void oid_both_from_string(wmem_allocator_t *scope, const gchar *oid_str, gchar** resolved_p, gchar** numeric_p) {
guint32 *subids;
guint subids_len;
subids_len = oid_string2subid(NULL, oid_str, &subids);
*resolved_p = oid_resolved(scope, subids_len,subids);
*numeric_p = oid_subid2string(scope, subids,subids_len);
wmem_free(NULL, subids);
}
/**
* Fetch the default OID path.
*/
extern gchar *
oid_get_default_mib_path(void) {
#ifdef HAVE_LIBSMI
GString* path_str;
char *path;
guint i;
path_str = g_string_new("");
if (!load_smi_modules) {
D(1,("OID resolution not enabled"));
return g_string_free(path_str, FALSE);
}
#ifdef _WIN32
path = get_datafile_path("snmp\\mibs");
g_string_append_printf(path_str, "%s;", path);
g_free (path);
path = get_persconffile_path("snmp\\mibs", FALSE);
g_string_append_printf(path_str, "%s", path);
g_free (path);
#else
g_string_append(path_str, "/usr/share/snmp/mibs");
if (!smi_init_done)
smiInit(NULL);
path = smiGetPath();
if (strlen(path) > 0 ) {
g_string_append(path_str, G_SEARCHPATH_SEPARATOR_S);
g_string_append_printf(path_str, "%s", path);
}
smi_free(path);
if (oids_init_done == FALSE)
{
#endif
for (i = 0; i < num_smi_paths; i++) {
if (!(smi_paths[i].name && *smi_paths[i].name))
continue;
g_string_append_printf(path_str, G_SEARCHPATH_SEPARATOR_S "%s", smi_paths[i].name);
}
#ifndef _WIN32
}
#endif
return g_string_free(path_str, FALSE);
#else /* HAVE_LIBSMI */
return g_strdup("");
#endif
}
#ifdef DEBUG_OIDS
char* oid_test_a2b(guint32 num_subids, guint32* subids) {
guint8* sub2enc = NULL;
guint8* str2enc = NULL;
guint32* enc2sub = NULL;
guint32* str2sub;
char* ret;
char* sub2str = oid_subid2string(NULL, subids, num_subids);
guint sub2enc_len = oid_subid2encoded(NULL, num_subids, subids,&sub2enc);
guint enc2sub_len = oid_encoded2subid(NULL, sub2enc, sub2enc_len, &enc2sub);
char* enc2str = oid_encoded2string(NULL, sub2enc, sub2enc_len);
guint str2enc_len = oid_string2encoded(NULL, sub2str,&str2enc);
guint str2sub_len = oid_string2subid(sub2str,&str2sub);
char* sub2enc_str = bytes_to_str_punct(NULL, sub2enc, sub2enc_len, ':');
char* enc2sub_str = enc2sub ? oid_subid2string(NULL, enc2sub,enc2sub_len) : wmem_strdup(NULL, "-");
char* str2enc_str = bytes_to_str_punct(NULL, str2enc, str2enc_len, ':');
char* str2sub_str = str2sub ? oid_subid2string(NULL, str2sub,str2sub_len) : wmem_strdup(NULL, "-");
ret = wmem_strdup_printf(NULL,
"oid_subid2string=%s \n"
"oid_subid2encoded=[%d]%s \n"
"oid_encoded2subid=%s \n "
"oid_encoded2string=%s \n"
"oid_string2encoded=[%d]%s \n"
"oid_string2subid=%s \n "
,sub2str
,sub2enc_len,sub2enc_str
,enc2sub_str
,enc2str
,str2enc_len,str2enc_str,
,str2sub_str
);
wmem_free(NULL, sub2enc_str);
wmem_free(NULL, enc2sub_str);
wmem_free(NULL, str2enc_str);
wmem_free(NULL, str2sub_str);
wmem_free(NULL, sub2str);
wmem_free(NULL, enc2sub);
wmem_free(NULL, sub2enc);
wmem_free(NULL, str2enc);
wmem_free(NULL, enc2str);
return ret;
}
void add_oid_debug_subtree(oid_info_t* oid_info, proto_tree *tree) {
static const char* oid_kinds[] = { "Unknown", "Node", "Scalar", "Table", "Row", "Column", "Notification", "Group", "Compliance", "Capabilities"};
static const char* key_types[] = {"OID_KEY_TYPE_WRONG","OID_KEY_TYPE_INTEGER",
"OID_KEY_TYPE_FIXED_STRING","OID_KEY_TYPE_FIXED_BYTES","OID_KEY_TYPE_STRING",
"OID_KEY_TYPE_BYTES","OID_KEY_TYPE_NSAP","OID_KEY_TYPE_OID","OID_KEY_TYPE_IPADDR"};
proto_item* pi = proto_tree_add_debug_text(tree,NULL,0,0,
"OidInfo: Name='%s' sub-id=%u kind=%s hfid=%d",
oid_info->name ? oid_info->name : "",
oid_info->subid,
oid_info->kind <= OID_KIND_CAPABILITIES ? oid_kinds[oid_info->kind] : "BROKEN",
oid_info->value_hfid);
proto_tree* pt = proto_item_add_subtree(pi,0);
oid_key_t* key;
for(key = oid_info->key; key; key = key->next) {
proto_tree_add_debug_text(pt,NULL,0,0,
"Key: name='%s' num_subids=%d type=%s",
key->name,
key->key_type <= OID_KEY_TYPE_IPADDR ? key_types[key->key_type] : "BROKEN"
);
};
if (oid_info->parent) {
pi = proto_tree_add_debug_text(pt,NULL,0,0,"Parent:");
pt = proto_item_add_subtree(pi,0);
add_oid_debug_subtree(oid_info->parent, pt);
}
}
#endif
/*
* Editor modelines
*
* Local Variables:
* c-basic-offset: 8
* tab-width: 8
* indent-tabs-mode: t
* End:
*
* ex: set shiftwidth=8 tabstop=8 noexpandtab:
* :indentSize=8:tabSize=8:noTabs=false:
*/
View Git Blame Copy Permalink