Added two client requests: VALGRIND_COUNT_ERRORS and VALGRIND_COUNT_LEAKS.
The first returns the number of errors found so far, and is a core request.
The second returns the number of bytes found
reachable/dubious/leaked/suppressed by all leak checks so far, for Memcheck and
Addrcheck.
Both are useful for using Valgrind in regression test suites where multiple
tests are present in a single file -- one can run Valgrind with no output
(using --logfile-fd=-1) and use the requests after each test to determine if
any errors happened.
Had to rename and make public vg_n_errs_found --> VG_(n_errs_found) to do so.
Nb: leak errors are not counted as errors for the purposes of
VALGRIND_COUNT_ERRORS. This was decided as the best thing to do after
discussion with Olly Betts, who original suggested these changes.
Pulled out common client request code shared between Memcheck and Addrcheck.
Added a regression test for this.
Added some documentation too.
git-svn-id: svn://svn.valgrind.org/valgrind/trunk@1533 a5019735-40e9-0310-863c-91ae7b9d1cf9
diff --git a/coregrind/vg_errcontext.c b/coregrind/vg_errcontext.c
index b04be54..f36426e 100644
--- a/coregrind/vg_errcontext.c
+++ b/coregrind/vg_errcontext.c
@@ -46,7 +46,7 @@
Supp* VG_(get_suppressions) ( void ) { return vg_suppressions; }
/* Running count of unsuppressed errors detected. */
-static UInt vg_n_errs_found = 0;
+UInt VG_(n_errs_found) = 0;
/* Running count of suppressed errors detected. */
static UInt vg_n_errs_suppressed = 0;
@@ -292,7 +292,7 @@
pointless to continue the Valgrind run after this point. */
if (VG_(clo_error_limit)
&& (vg_n_errs_shown >= M_VG_COLLECT_NO_ERRORS_AFTER_SHOWN
- || vg_n_errs_found >= M_VG_COLLECT_NO_ERRORS_AFTER_FOUND)) {
+ || VG_(n_errs_found) >= M_VG_COLLECT_NO_ERRORS_AFTER_FOUND)) {
if (!stopping_message) {
VG_(message)(Vg_UserMsg, "");
@@ -354,7 +354,7 @@
p->supp->count++;
vg_n_errs_suppressed++;
} else {
- vg_n_errs_found++;
+ VG_(n_errs_found)++;
}
/* Move p to the front of the list so that future searches
@@ -405,7 +405,7 @@
p->supp = is_suppressible_error(&err);
vg_errors = p;
if (p->supp == NULL) {
- vg_n_errs_found++;
+ VG_(n_errs_found)++;
if (!is_first_shown_context)
VG_(message)(Vg_UserMsg, "");
pp_Error(p, False);
@@ -422,10 +422,12 @@
errors that the skin want to report immediately, eg. because they're
guaranteed to only happen once. This avoids all the recording and
comparing stuff. But they can be suppressed; returns True if it is
- suppressed. Bool `print_error' dictates whether to print the error. */
+ suppressed. Bool `print_error' dictates whether to print the error.
+ Bool `count_error' dictates whether to count the error in VG_(n_errs_found)
+*/
Bool VG_(unique_error) ( ThreadState* tst, ErrorKind ekind, Addr a, Char* s,
void* extra, ExeContext* where, Bool print_error,
- Bool allow_GDB_attach )
+ Bool allow_GDB_attach, Bool count_error )
{
Error err;
Addr m_eip, m_esp, m_ebp;
@@ -443,7 +445,8 @@
(void)SK_(update_extra)(&err);
if (NULL == is_suppressible_error(&err)) {
- vg_n_errs_found++;
+ if (count_error)
+ VG_(n_errs_found)++;
if (print_error) {
if (!is_first_shown_context)
@@ -502,7 +505,7 @@
VG_(message)(Vg_UserMsg,
"ERROR SUMMARY: "
"%d errors from %d contexts (suppressed: %d from %d)",
- vg_n_errs_found, n_err_contexts,
+ VG_(n_errs_found), n_err_contexts,
vg_n_errs_suppressed, n_supp_contexts );
if (VG_(clo_verbosity) <= 1)
@@ -551,7 +554,7 @@
VG_(message)(Vg_UserMsg,
"IN SUMMARY: "
"%d errors from %d contexts (suppressed: %d from %d)",
- vg_n_errs_found, n_err_contexts,
+ VG_(n_errs_found), n_err_contexts,
vg_n_errs_suppressed,
n_supp_contexts );
VG_(message)(Vg_UserMsg, "");