Re: [PATCH] Call BUILT_IN_ASAN_HANDLE_NO_RETURN before BUILT_IN_UNWIND_RESUME (PR sanitizer/81021).

[Jeff Law]

On 06/13/2017 02:09 AM, Martin Liška wrote:
> Hi.
> 
> For a function that does not handle an expection (and calls 
> BUILT_IN_UNWIND_RESUME),
> we need to emit call to BUILT_IN_ASAN_HANDLE_NO_RETURN. That will clean up 
> stack
> which can possibly contain poisoned shadow memory that will not be cleaned-up
> in function prologue.
> 
> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
> 
> Ready to be installed?
> Martin
> 
> gcc/testsuite/ChangeLog:
> 
> 2017-06-12  Martin Liska  
> 
>   PR sanitizer/81021
>   * g++.dg/asan/pr81021.C: New test.
> 
> gcc/ChangeLog:
> 
> 2017-06-12  Martin Liska  
> 
>   PR sanitizer/81021
>   * tree-eh.c (lower_resx): Call BUILT_IN_ASAN_HANDLE_NO_RETURN
>   before BUILT_IN_UNWIND_RESUME when ASAN is used.
OK.
Jeff

Re: [PATCH] Call BUILT_IN_ASAN_HANDLE_NO_RETURN before BUILT_IN_UNWIND_RESUME (PR sanitizer/81021).

[Martin Liška]

PING^2

On 06/20/2017 02:15 PM, Martin Liška wrote:
> PING^1
> 
> On 06/13/2017 10:09 AM, Martin Liška wrote:
>> Hi.
>>
>> For a function that does not handle an expection (and calls 
>> BUILT_IN_UNWIND_RESUME),
>> we need to emit call to BUILT_IN_ASAN_HANDLE_NO_RETURN. That will clean up 
>> stack
>> which can possibly contain poisoned shadow memory that will not be cleaned-up
>> in function prologue.
>>
>> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>>
>> Ready to be installed?
>> Martin
>>
>> gcc/testsuite/ChangeLog:
>>
>> 2017-06-12  Martin Liska  
>>
>>  PR sanitizer/81021
>>  * g++.dg/asan/pr81021.C: New test.
>>
>> gcc/ChangeLog:
>>
>> 2017-06-12  Martin Liska  
>>
>>  PR sanitizer/81021
>>  * tree-eh.c (lower_resx): Call BUILT_IN_ASAN_HANDLE_NO_RETURN
>>  before BUILT_IN_UNWIND_RESUME when ASAN is used.
>> ---
>>  gcc/testsuite/g++.dg/asan/pr81021.C | 33 +
>>  gcc/tree-eh.c   | 14 ++
>>  2 files changed, 47 insertions(+)
>>  create mode 100644 gcc/testsuite/g++.dg/asan/pr81021.C
>>
>>
> 

Re: [PATCH] Call BUILT_IN_ASAN_HANDLE_NO_RETURN before BUILT_IN_UNWIND_RESUME (PR sanitizer/81021).

[Martin Liška]

PING^1

On 06/13/2017 10:09 AM, Martin Liška wrote:
> Hi.
> 
> For a function that does not handle an expection (and calls 
> BUILT_IN_UNWIND_RESUME),
> we need to emit call to BUILT_IN_ASAN_HANDLE_NO_RETURN. That will clean up 
> stack
> which can possibly contain poisoned shadow memory that will not be cleaned-up
> in function prologue.
> 
> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
> 
> Ready to be installed?
> Martin
> 
> gcc/testsuite/ChangeLog:
> 
> 2017-06-12  Martin Liska  
> 
>   PR sanitizer/81021
>   * g++.dg/asan/pr81021.C: New test.
> 
> gcc/ChangeLog:
> 
> 2017-06-12  Martin Liska  
> 
>   PR sanitizer/81021
>   * tree-eh.c (lower_resx): Call BUILT_IN_ASAN_HANDLE_NO_RETURN
>   before BUILT_IN_UNWIND_RESUME when ASAN is used.
> ---
>  gcc/testsuite/g++.dg/asan/pr81021.C | 33 +
>  gcc/tree-eh.c   | 14 ++
>  2 files changed, 47 insertions(+)
>  create mode 100644 gcc/testsuite/g++.dg/asan/pr81021.C
> 
> 

[PATCH] Call BUILT_IN_ASAN_HANDLE_NO_RETURN before BUILT_IN_UNWIND_RESUME (PR sanitizer/81021).

[Martin Liška]

Hi.

For a function that does not handle an expection (and calls 
BUILT_IN_UNWIND_RESUME),
we need to emit call to BUILT_IN_ASAN_HANDLE_NO_RETURN. That will clean up stack
which can possibly contain poisoned shadow memory that will not be cleaned-up
in function prologue.

Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.

Ready to be installed?
Martin

gcc/testsuite/ChangeLog:

2017-06-12  Martin Liska  

PR sanitizer/81021
* g++.dg/asan/pr81021.C: New test.

gcc/ChangeLog:

2017-06-12  Martin Liska  

PR sanitizer/81021
* tree-eh.c (lower_resx): Call BUILT_IN_ASAN_HANDLE_NO_RETURN
before BUILT_IN_UNWIND_RESUME when ASAN is used.
---
 gcc/testsuite/g++.dg/asan/pr81021.C | 33 +
 gcc/tree-eh.c   | 14 ++
 2 files changed, 47 insertions(+)
 create mode 100644 gcc/testsuite/g++.dg/asan/pr81021.C


diff --git a/gcc/testsuite/g++.dg/asan/pr81021.C b/gcc/testsuite/g++.dg/asan/pr81021.C
new file mode 100644
index 000..daa0525c273
--- /dev/null
+++ b/gcc/testsuite/g++.dg/asan/pr81021.C
@@ -0,0 +1,33 @@
+// { dg-do run }
+
+#include 
+
+struct ConfigFile {
+ConfigFile(std::string filename, std::string delimiter) { throw "error"; }
+ConfigFile(std::string filename) {}
+};
+
+struct Configuration {
+ConfigFile _configFile;
+
+Configuration(const std::string , const char *baseName) 
+: _configFile(root + baseName, "=") { }
+Configuration(const std::string , const char *a, const char *b) 
+: _configFile(root + a + b) { }
+};
+
+
+void test() {
+std::string root("etc");
+try {
+Configuration config(root, "notthere");
+}
+catch (...) {
+// exception is thrown, caught here and ignored...
+}
+Configuration config(root, "a", "b"); // ASAN error during constructor here
+}
+
+int main(int argc, const char *argv[]) {
+test();
+}
diff --git a/gcc/tree-eh.c b/gcc/tree-eh.c
index fc016d795b7..fdd348c52e9 100644
--- a/gcc/tree-eh.c
+++ b/gcc/tree-eh.c
@@ -3304,6 +3304,20 @@ lower_resx (basic_block bb, gresx *stmt,
 	  gimple_call_set_lhs (x, var);
 	  gsi_insert_before (, x, GSI_SAME_STMT);
 
+	  /* When exception handling is delegated to a caller function, we
+	 have to guarantee that shadow memory variables living on stack
+	 will be cleaner before control is given to a parent function.  */
+	  if ((flag_sanitize & SANITIZE_ADDRESS) != 0
+	  && !lookup_attribute ("no_sanitize_address",
+DECL_ATTRIBUTES (current_function_decl)))
+	{
+	  tree decl
+		= builtin_decl_implicit (BUILT_IN_ASAN_HANDLE_NO_RETURN);
+	  gimple *g = gimple_build_call (decl, 0);
+	  gimple_set_location (g, gimple_location (stmt));
+	  gsi_insert_before (, g, GSI_SAME_STMT);
+	}
+
 	  fn = builtin_decl_implicit (BUILT_IN_UNWIND_RESUME);
 	  x = gimple_build_call (fn, 1, var);
 	  gsi_insert_before (, x, GSI_SAME_STMT);