named-checkconf core dump (bind-9.3.1rc1)
Stefan Puiu
stefanpuiu at itcnetworks.ro
Tue Feb 22 09:37:07 UTC 2005
Thanks. Now, I'm not sure if these lines are supposed to go in the app
or in the library (rbtdb.c), but this is probably up to ISC.
I've managed to elucidate the mistery of the difference in line numbers
between my stack trace and John's - it looks like gcc associates the
first line number in a multi-line C expression to the generated code
when writing debugging information in the binary, while the Sun cc uses
the last line - that's why sometimes the line numbers were skewed by one
or two.
JINMEI Tatuya / 神明達哉 wrote:
>Attached below. (this is probably needless to say but) please note
>that it's not an official patch, and use it at your own risk.
>
> JINMEI, Tatuya
> Communication Platform Lab.
> Corporate R&D Center, Toshiba Corp.
> jinmei at isl.rdc.toshiba.co.jp
>
>--- named-checkconf.c.orig Mon Mar 8 18:04:14 2004
>+++ named-checkconf.c Tue Feb 22 17:55:47 2005
>@@ -25,6 +25,8 @@
>
> #include <isc/commandline.h>
> #include <isc/dir.h>
>+#include <isc/entropy.h>
>+#include <isc/hash.h>
> #include <isc/log.h>
> #include <isc/mem.h>
> #include <isc/result.h>
>@@ -35,6 +37,7 @@
>
> #include <bind9/check.h>
>
>+#include <dns/fixedname.h>
> #include <dns/log.h>
> #include <dns/result.h>
>
>@@ -203,6 +206,7 @@
> isc_mem_t *mctx = NULL;
> isc_result_t result;
> int exit_status = 0;
>+ isc_entropy_t *ectx = NULL;
> isc_boolean_t load_zones = ISC_FALSE;
>
> while ((c = isc_commandline_parse(argc, argv, "djt:vz")) != EOF) {
>@@ -252,6 +256,10 @@
>
> RUNTIME_CHECK(setup_logging(mctx, &logc) == ISC_R_SUCCESS);
>
>+ RUNTIME_CHECK(isc_entropy_create(mctx, &ectx) == ISC_R_SUCCESS);
>+ RUNTIME_CHECK(isc_hash_create(mctx, ectx, DNS_NAME_MAXWIRE)
>+ == ISC_R_SUCCESS);
>+
> dns_result_register();
>
> RUNTIME_CHECK(cfg_parser_create(mctx, logc, &parser) == ISC_R_SUCCESS);
>@@ -279,6 +287,9 @@
> cfg_parser_destroy(&parser);
>
> isc_log_destroy(&logc);
>+
>+ isc_hash_destroy();
>+ isc_entropy_detach(&ectx);
>
> isc_mem_destroy(&mctx);
>
>
>
>
More information about the bind-users
mailing list