Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, Nov 18, 2013 at 6:44 AM, Marek Polacek pola...@redhat.com wrote: On Mon, Nov 18, 2013 at 02:51:41PM +0100, Jakub Jelinek wrote: On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote: --- gcc/config/bootstrap-ubsan.mk.mp2013-11-12 13:46:13.345182065 +0100 +++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ Hopefully with my pending patch you can remove the -lpthread -ldl again, but ok for now. You shouldn't use -ldl directly. Not all OSes have libdl. You should extract the libdl check from gcc/configure.ac and set LIBDL instead by changing gcc/Makefile.in PLUGINLIBS = @pluginlibs@ to LIBDL = @libdl@ PLUGINLIBS = @pluginlibs@ $(LIBD) Then you can use POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread $(LIBDL) \ -- H.J.
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Fri, Nov 29, 2013 at 11:22 AM, H.J. Lu hjl.to...@gmail.com wrote: On Mon, Nov 18, 2013 at 6:44 AM, Marek Polacek pola...@redhat.com wrote: On Mon, Nov 18, 2013 at 02:51:41PM +0100, Jakub Jelinek wrote: On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote: --- gcc/config/bootstrap-ubsan.mk.mp2013-11-12 13:46:13.345182065 +0100 +++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ Hopefully with my pending patch you can remove the -lpthread -ldl again, but ok for now. You shouldn't use -ldl directly. Not all OSes have libdl. You should extract the libdl check from gcc/configure.ac and set LIBDL instead by changing gcc/Makefile.in PLUGINLIBS = @pluginlibs@ to LIBDL = @libdl@ PLUGINLIBS = @pluginlibs@ $(LIBD) Then you can use POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread $(LIBDL) \ Something like this. Only tested with normal build. -- H.J. --- diff --git a/config/bootstrap-ubsan.mk b/config/bootstrap-ubsan.mk index 0cd8b17..c298cd1 100644 --- a/config/bootstrap-ubsan.mk +++ b/config/bootstrap-ubsan.mk @@ -2,6 +2,7 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \ + $(LIBDL) \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs diff --git a/gcc/Makefile.in b/gcc/Makefile.in index 4d683a0..cb64241 100644 --- a/gcc/Makefile.in +++ b/gcc/Makefile.in @@ -340,12 +340,15 @@ CLOOGINC = @CLOOGINC@ # Set to 'yes' if the LTO front end is enabled. enable_lto = @enable_lto@ +# Library for dlopen +LIBDL=@libdl@ + # Compiler and flags needed for plugin support PLUGINCC = @CXX@ PLUGINCFLAGS = @CXXFLAGS@ # Libs and linker options needed for plugin support -PLUGINLIBS = @pluginlibs@ +PLUGINLIBS = @pluginlibs@ $(LIBDL) enable_plugin = @enable_plugin@ diff --git a/gcc/configure.ac b/gcc/configure.ac index 91a22d5..80cd248 100644 --- a/gcc/configure.ac +++ b/gcc/configure.ac @@ -5404,14 +5404,6 @@ if test x$enable_plugin = xyes; then AC_MSG_RESULT([unable to check]) fi - # Check -ldl - saved_LIBS=$LIBS - AC_SEARCH_LIBS([dlopen], [dl]) - if test x$ac_cv_search_dlopen = x-ldl; then -pluginlibs=$pluginlibs -ldl - fi - LIBS=$saved_LIBS - # Check that we can build shared objects with -fPIC -shared saved_LDFLAGS=$LDFLAGS saved_CFLAGS=$CFLAGS @@ -5454,6 +5446,16 @@ if test x$enable_plugin = xyes; then AC_DEFINE(ENABLE_PLUGIN, 1, [Define to enable plugin support.]) fi +# Check -ldl +libdl= +saved_LIBS=$LIBS +AC_SEARCH_LIBS([dlopen], [dl]) +if test x$ac_cv_search_dlopen = x-ldl; then + libdl=-ldl +fi +LIBS=$saved_LIBS +AC_SUBST(libdl) + # Enable --enable-host-shared AC_ARG_ENABLE(host-shared,
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Fri, Nov 29, 2013 at 11:22:00AM -0800, H.J. Lu wrote: On Mon, Nov 18, 2013 at 6:44 AM, Marek Polacek pola...@redhat.com wrote: On Mon, Nov 18, 2013 at 02:51:41PM +0100, Jakub Jelinek wrote: On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote: --- gcc/config/bootstrap-ubsan.mk.mp2013-11-12 13:46:13.345182065 +0100 +++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ Hopefully with my pending patch you can remove the -lpthread -ldl again, but ok for now. You shouldn't use -ldl directly. Not all OSes have libdl. You should extract the libdl check from gcc/configure.ac and set LIBDL instead by changing gcc/Makefile.in -static-libubsan should add all the libraries needed of libubsan.a by now, so -lpthread -ldl should be just removed from POSTSTAGE1_LDFLAGS. Jakub
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Fri, Nov 29, 2013 at 08:32:34PM +0100, Jakub Jelinek wrote: On Fri, Nov 29, 2013 at 11:22:00AM -0800, H.J. Lu wrote: On Mon, Nov 18, 2013 at 6:44 AM, Marek Polacek pola...@redhat.com wrote: On Mon, Nov 18, 2013 at 02:51:41PM +0100, Jakub Jelinek wrote: On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote: --- gcc/config/bootstrap-ubsan.mk.mp2013-11-12 13:46:13.345182065 +0100 +++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ Hopefully with my pending patch you can remove the -lpthread -ldl again, but ok for now. You shouldn't use -ldl directly. Not all OSes have libdl. You should extract the libdl check from gcc/configure.ac and set LIBDL instead by changing gcc/Makefile.in -static-libubsan should add all the libraries needed of libubsan.a by now, so -lpthread -ldl should be just removed from POSTSTAGE1_LDFLAGS. So ok to install this one? 2013-11-29 Marek Polacek pola...@redhat.com * bootstrap-ubsan.mk (POSTSTAGE1_LDFLAGS): Remove -lpthread -ldl. --- gcc/bootstrap-ubsan.mk.mp3 2013-11-29 20:50:04.788238860 +0100 +++ gcc/bootstrap-ubsan.mk 2013-11-29 20:50:25.870322185 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs Marek
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Fri, Nov 29, 2013 at 08:55:26PM +0100, Marek Polacek wrote: 2013-11-29 Marek Polacek pola...@redhat.com * bootstrap-ubsan.mk (POSTSTAGE1_LDFLAGS): Remove -lpthread -ldl. --- gcc/bootstrap-ubsan.mk.mp32013-11-29 20:50:04.788238860 +0100 +++ gcc/bootstrap-ubsan.mk2013-11-29 20:50:25.870322185 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs Please add -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ too, so that it is able to find libsanitizer.spec. Ok with that change. Jakub
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Fri, Nov 29, 2013 at 08:57:23PM +0100, Jakub Jelinek wrote: On Fri, Nov 29, 2013 at 08:55:26PM +0100, Marek Polacek wrote: 2013-11-29 Marek Polacek pola...@redhat.com * bootstrap-ubsan.mk (POSTSTAGE1_LDFLAGS): Remove -lpthread -ldl. --- gcc/bootstrap-ubsan.mk.mp3 2013-11-29 20:50:04.788238860 +0100 +++ gcc/bootstrap-ubsan.mk 2013-11-29 20:50:25.870322185 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs Please add -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ too, so that it is able to find libsanitizer.spec. Ok with that change. Thanks, will apply the following then. 2013-11-29 Marek Polacek pola...@redhat.com * bootstrap-ubsan.mk (POSTSTAGE1_LDFLAGS): Remove -lpthread -ldl. Add -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/. --- gcc/bootstrap-ubsan.mk.mp3 2013-11-29 20:50:04.788238860 +0100 +++ gcc/bootstrap-ubsan.mk 2013-11-29 20:58:23.322131822 +0100 @@ -2,6 +2,7 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan \ + -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs Marek
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Fri, Nov 29, 2013 at 12:02 PM, Marek Polacek pola...@redhat.com wrote: On Fri, Nov 29, 2013 at 08:57:23PM +0100, Jakub Jelinek wrote: On Fri, Nov 29, 2013 at 08:55:26PM +0100, Marek Polacek wrote: 2013-11-29 Marek Polacek pola...@redhat.com * bootstrap-ubsan.mk (POSTSTAGE1_LDFLAGS): Remove -lpthread -ldl. --- gcc/bootstrap-ubsan.mk.mp3 2013-11-29 20:50:04.788238860 +0100 +++ gcc/bootstrap-ubsan.mk 2013-11-29 20:50:25.870322185 +0100 @@ -2,6 +2,6 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs Please add -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ too, so that it is able to find libsanitizer.spec. Ok with that change. Thanks, will apply the following then. 2013-11-29 Marek Polacek pola...@redhat.com * bootstrap-ubsan.mk (POSTSTAGE1_LDFLAGS): Remove -lpthread -ldl. Add -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/. --- gcc/bootstrap-ubsan.mk.mp3 2013-11-29 20:50:04.788238860 +0100 +++ gcc/bootstrap-ubsan.mk 2013-11-29 20:58:23.322131822 +0100 @@ -2,6 +2,7 @@ STAGE2_CFLAGS += -fsanitize=undefined STAGE3_CFLAGS += -fsanitize=undefined -POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \ +POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan \ + -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/ \ -B$$r/prev-$(TARGET_SUBDIR)/libsanitizer/ubsan/.libs Marek I pushed it to binutils-gdb. -- H.J.
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote:
--- gcc/config/bootstrap-ubsan.mk.mp 2013-11-12 13:46:13.345182065 +0100
+++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100
@@ -2,6 +2,6 @@
STAGE2_CFLAGS += -fsanitize=undefined
STAGE3_CFLAGS += -fsanitize=undefined
-POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \
+POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl \
Hopefully with my pending patch you can remove the -lpthread -ldl again, but
ok for now.
+ for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (gsi))
+ {
+ gimple stmt = gsi_stmt (gsi);
+
+ if (gimple_code (stmt) != GIMPLE_CALL)
if (is_gimple_call (stmt))
Ok with those changes.
Jakub
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, Nov 18, 2013 at 02:51:41PM +0100, Jakub Jelinek wrote:
On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote:
--- gcc/config/bootstrap-ubsan.mk.mp2013-11-12 13:46:13.345182065
+0100
+++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100
@@ -2,6 +2,6 @@
STAGE2_CFLAGS += -fsanitize=undefined
STAGE3_CFLAGS += -fsanitize=undefined
-POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \
+POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl
\
Hopefully with my pending patch you can remove the -lpthread -ldl again, but
ok for now.
+ for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (gsi))
+ {
+ gimple stmt = gsi_stmt (gsi);
+
+ if (gimple_code (stmt) != GIMPLE_CALL)
if (is_gimple_call (stmt))
Ok with those changes.
Oh, one more thing, please update gcc/doc/, the -fsanitize= description is
far from up to date there.
Jakub
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, Nov 18, 2013 at 02:51:41PM +0100, Jakub Jelinek wrote:
On Wed, Nov 13, 2013 at 12:13:48AM +0100, Marek Polacek wrote:
--- gcc/config/bootstrap-ubsan.mk.mp2013-11-12 13:46:13.345182065
+0100
+++ gcc/config/bootstrap-ubsan.mk 2013-11-12 13:46:49.812314016 +0100
@@ -2,6 +2,6 @@
STAGE2_CFLAGS += -fsanitize=undefined
STAGE3_CFLAGS += -fsanitize=undefined
-POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread \
+POSTSTAGE1_LDFLAGS += -fsanitize=undefined -static-libubsan -lpthread -ldl
\
Hopefully with my pending patch you can remove the -lpthread -ldl again, but
ok for now.
Cool.
+ for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (gsi))
+ {
+ gimple stmt = gsi_stmt (gsi);
+
+ if (gimple_code (stmt) != GIMPLE_CALL)
if (is_gimple_call (stmt))
Fixed.
Ok with those changes.
Thanks. Also I'll have to add some headers after the gimple.h reorg,
but that is an obvious change.
Marek
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, Nov 18, 2013 at 02:52:34PM +0100, Jakub Jelinek wrote:
Oh, one more thing, please update gcc/doc/, the -fsanitize= description is
far from up to date there.
Ok, the following (incremental) hopefully improves the docs. Joseph, would
you mind having a look at this? Thanks,
2013-11-18 Marek Polacek pola...@redhat.com
* doc/invoke.texi: Extend -fsanitize=undefined documentation.
--- gcc/doc/invoke.texi.mp3 2013-11-18 15:57:47.104103101 +0100
+++ gcc/doc/invoke.texi 2013-11-18 17:08:51.305594441 +0100
@@ -5260,9 +5260,45 @@ data race bugs.
See @uref{http://code.google.com/p/data-race-test/wiki/ThreadSanitizer} for
more details.
@item -fsanitize=undefined
-Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector
+Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
Various computations will be instrumented to detect undefined behavior
-at runtime, e.g.@: division by zero or various overflows.
+at runtime. Current suboptions are:
+
+@itemize @bullet
+
+@item @option{-fsanitize=shift}
+
+This option enables checking that the result of a shift operation is
+not undefined. Note that what exactly is considered undefined differs
+slightly between C and C++, as well as between ANSI C and C99, etc.
+
+@item @option{-fsanitize=integer-divide-by-zero}
+
+Detect integer division by zero as well as @code{INT_MIN / -1} division.
+Note that the latter is only made undefined from C99 onwards.
+
+@item @option{-fsanitize=unreachable}
+
+With this option, the compiler will turn the @code{__builtin_unreachable}
+call into a diagnostics message call instead. When reaching the
+@code{__builtin_unreachable} call, the behavior is undefined.
+
+@item @option{-fsanitize=vla-bound}
+
+This option instructs the compiler to check that the size of a variable
+length array is positive. This option does not have any effect in
+@option{-std=c++1y} mode, as the standard requires the exception be thrown
+instead.
+
+@item @option{-fsanitize=null}
+
+This option enables pointer checking. Particularly, the application
+built with this option turned on will issue an error message when it
+tries to dereference a NULL pointer, or if a reference (possibly an
+rvalue reference) is bound to a NULL pointer.
+
+@end itemize
+
While @option{-ftrapv} causes traps for signed overflows to be emitted,
@option{-fsanitize=undefined} gives a diagnostic message.
This currently works only for the C family of languages.
Marek
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, 18 Nov 2013, Marek Polacek wrote:
+@item @option{-fsanitize=shift}
+
+This option enables checking that the result of a shift operation is
+not undefined. Note that what exactly is considered undefined differs
+slightly between C and C++, as well as between ANSI C and C99, etc.
We generally refer to ISO C90, not ANSI C.
+Detect integer division by zero as well as @code{INT_MIN / -1} division.
+Note that the latter is only made undefined from C99 onwards.
INT_MIN / -1 is unambiguously undefined in C90 - it's a signed arithmetic
overflow (result not within the range of its type). It's INT_MIN % -1
where there's more ambiguity, but I consider the wording changes in C11 as
a defect correction that should be applied back to C90. (A comment on
what the semantics should be, not on whether the documentation accurately
reflects the code.)
--
Joseph S. Myers
jos...@codesourcery.com
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, Nov 18, 2013 at 04:58:36PM +, Joseph S. Myers wrote:
On Mon, 18 Nov 2013, Marek Polacek wrote:
+@item @option{-fsanitize=shift}
+
+This option enables checking that the result of a shift operation is
+not undefined. Note that what exactly is considered undefined differs
+slightly between C and C++, as well as between ANSI C and C99, etc.
We generally refer to ISO C90, not ANSI C.
Fixed.
+Detect integer division by zero as well as @code{INT_MIN / -1} division.
+Note that the latter is only made undefined from C99 onwards.
INT_MIN / -1 is unambiguously undefined in C90 - it's a signed arithmetic
overflow (result not within the range of its type). It's INT_MIN % -1
where there's more ambiguity, but I consider the wording changes in C11 as
a defect correction that should be applied back to C90. (A comment on
what the semantics should be, not on whether the documentation accurately
reflects the code.)
I removed that sentence to not confuse readers. (We issue runtime
error for INT_MIN % -1 for all c90, c99, c11 modes.) Thanks.
Ok now?
2013-11-18 Marek Polacek pola...@redhat.com
* doc/invoke.texi: Extend -fsanitize=undefined documentation.
--- gcc/doc/invoke.texi.mp3 2013-11-18 15:57:47.104103101 +0100
+++ gcc/doc/invoke.texi 2013-11-18 18:55:00.178009402 +0100
@@ -5260,9 +5260,44 @@ data race bugs.
See @uref{http://code.google.com/p/data-race-test/wiki/ThreadSanitizer} for
more details.
@item -fsanitize=undefined
-Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector
+Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
Various computations will be instrumented to detect undefined behavior
-at runtime, e.g.@: division by zero or various overflows.
+at runtime. Current suboptions are:
+
+@itemize @bullet
+
+@item @option{-fsanitize=shift}
+
+This option enables checking that the result of a shift operation is
+not undefined. Note that what exactly is considered undefined differs
+slightly between C and C++, as well as between ISO C90 and C99, etc.
+
+@item @option{-fsanitize=integer-divide-by-zero}
+
+Detect integer division by zero as well as @code{INT_MIN / -1} division.
+
+@item @option{-fsanitize=unreachable}
+
+With this option, the compiler will turn the @code{__builtin_unreachable}
+call into a diagnostics message call instead. When reaching the
+@code{__builtin_unreachable} call, the behavior is undefined.
+
+@item @option{-fsanitize=vla-bound}
+
+This option instructs the compiler to check that the size of a variable
+length array is positive. This option does not have any effect in
+@option{-std=c++1y} mode, as the standard requires the exception be thrown
+instead.
+
+@item @option{-fsanitize=null}
+
+This option enables pointer checking. Particularly, the application
+built with this option turned on will issue an error message when it
+tries to dereference a NULL pointer, or if a reference (possibly an
+rvalue reference) is bound to a NULL pointer.
+
+@end itemize
+
While @option{-ftrapv} causes traps for signed overflows to be emitted,
@option{-fsanitize=undefined} gives a diagnostic message.
This currently works only for the C family of languages.
Marek
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Mon, 18 Nov 2013, Marek Polacek wrote: 2013-11-18 Marek Polacek pola...@redhat.com * doc/invoke.texi: Extend -fsanitize=undefined documentation. OK. -- Joseph S. Myers jos...@codesourcery.com
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On Wed, Nov 13, 2013 at 06:45:06AM +0100, Markus Trippelsdorf wrote: On 2013.11.13 at 00:13 +0100, Marek Polacek wrote: 2) bootstrap-ubsan almost passes, but the bootstrap fails when building all-fixincludes. The problem here is that libiberty.a is built with -fsanitize=undefined, but fixincludes, when linking, don't link libubsan in. My attemps to tweak FIXINC_CFLAGS/LDFLAGS/BOOT_LDFLAGS and whatnot weren't successfull. I'm using the following patch locally as a part to enable slim-lto-bootstrap. Maybe it helps in your case, too? Unfortunately, doesn't seem to help :(. Thanks anyway. Marek
Re: [PATCH] Implement -fsanitize=null + new sanopt pass
On 2013.11.13 at 00:13 +0100, Marek Polacek wrote: 2) bootstrap-ubsan almost passes, but the bootstrap fails when building all-fixincludes. The problem here is that libiberty.a is built with -fsanitize=undefined, but fixincludes, when linking, don't link libubsan in. My attemps to tweak FIXINC_CFLAGS/LDFLAGS/BOOT_LDFLAGS and whatnot weren't successfull. I'm using the following patch locally as a part to enable slim-lto-bootstrap. Maybe it helps in your case, too? diff --git a/Makefile.in b/Makefile.in index f9e8e0d5cb79..5db913fa0b8d 100644 --- a/Makefile.in +++ b/Makefile.in @@ -2930,6 +2930,7 @@ configure-build-fixincludes: test ! -f $(BUILD_SUBDIR)/fixincludes/Makefile || exit 0; \ $(SHELL) $(srcdir)/mkinstalldirs $(BUILD_SUBDIR)/fixincludes ; \ $(BUILD_EXPORTS) \ + CFLAGS=$(STAGE_CFLAGS); export CFLAGS; \ echo Configuring in $(BUILD_SUBDIR)/fixincludes; \ cd $(BUILD_SUBDIR)/fixincludes || exit 1; \ case $(srcdir) in \ @@ -2965,6 +2966,7 @@ all-build-fixincludes: configure-build-fixincludes $(BUILD_EXPORTS) \ (cd $(BUILD_SUBDIR)/fixincludes \ $(MAKE) $(BASE_FLAGS_TO_PASS) $(EXTRA_BUILD_FLAGS) \ + CFLAGS=$(STAGE_CFLAGS) \ $(TARGET-build-fixincludes)) @endif build-fixincludes @@ -7813,6 +7815,7 @@ configure-fixincludes: test ! -f $(HOST_SUBDIR)/fixincludes/Makefile || exit 0; \ $(SHELL) $(srcdir)/mkinstalldirs $(HOST_SUBDIR)/fixincludes ; \ $(HOST_EXPORTS) \ + CFLAGS=$(STAGE_CFLAGS); export CFLAGS; \ echo Configuring in $(HOST_SUBDIR)/fixincludes; \ cd $(HOST_SUBDIR)/fixincludes || exit 1; \ case $(srcdir) in \ @@ -7847,6 +7850,7 @@ all-fixincludes: configure-fixincludes $(HOST_EXPORTS) \ (cd $(HOST_SUBDIR)/fixincludes \ $(MAKE) $(BASE_FLAGS_TO_PASS) $(EXTRA_HOST_FLAGS) $(STAGE1_FLAGS_TO_PASS) \ + CFLAGS=$(STAGE_CFLAGS) \ $(TARGET-fixincludes)) @endif fixincludes -- Markus
