diff mbox series

Correct access attribute on memfrob (bug 28475)

Message ID alpine.DEB.2.22.394.2110191724140.10354@digraph.polyomino.org.uk
State Superseded
Headers show
Series Correct access attribute on memfrob (bug 28475) | expand

Checks

Context Check Description
dj/TryBot-apply_patch success Patch applied to master at the time it was sent
dj/TryBot-32bit success Build for i686

Commit Message

Joseph Myers Oct. 19, 2021, 5:24 p.m. UTC
As noted in bug 28475, the access attribute on memfrob in <string.h>
is incorrect: the function both reads and writes the memory pointed to
by its argument, so it needs to use __read_write__, not
__write_only__.  This incorrect attribute results in a build failure
for accessing uninitialized memory for s390x-linux-gnu-O3 with
build-many-glibcs.py using GCC mainline.

Correct the attribute.  Fixing this shows up that some calls to
memfrob in elf/ tests are reading uninitialized memory; I'm not
entirely sure of the purpose of those calls, but guess they are about
ensuring that the stack space is indeed allocated at that point in the
function, and so it matters that they are calling a function whose
semantics are unknown to the compiler.  Thus, add a memset call before
the memfrob call in those tests to avoid the use of uninitialized
memory.

Tested for x86_64, and with build-many-glibcs.py (GCC mainline) for
s390x-linux-gnu-O3.

Comments

Florian Weimer Oct. 19, 2021, 7:48 p.m. UTC | #1
* Joseph Myers:

> As noted in bug 28475, the access attribute on memfrob in <string.h>
> is incorrect: the function both reads and writes the memory pointed to
> by its argument, so it needs to use __read_write__, not
> __write_only__.  This incorrect attribute results in a build failure
> for accessing uninitialized memory for s390x-linux-gnu-O3 with
> build-many-glibcs.py using GCC mainline.

> Correct the attribute.

That part looks okay.

> Fixing this shows up that some calls to memfrob in elf/ tests are
> reading uninitialized memory; I'm not entirely sure of the purpose of
> those calls, but guess they are about ensuring that the stack space is
> indeed allocated at that point in the function, and so it matters that
> they are calling a function whose semantics are unknown to the
> compiler.  Thus, add a memset call before the memfrob call in those
> tests to avoid the use of uninitialized memory.

Using explicit_bzero would be more idomatic, I think.

Florian
diff mbox series

Patch

diff --git a/elf/tst-execstack-needed.c b/elf/tst-execstack-needed.c
index 8b794a3d47..a42597a4c1 100644
--- a/elf/tst-execstack-needed.c
+++ b/elf/tst-execstack-needed.c
@@ -26,6 +26,7 @@  static void
 deeper (void (*f) (void))
 {
   char stack[1100 * 1024];
+  memset (stack, 123, sizeof stack);
   memfrob (stack, sizeof stack);
   (*f) ();
   memfrob (stack, sizeof stack);
diff --git a/elf/tst-execstack-prog.c b/elf/tst-execstack-prog.c
index 8663153372..40d4d00858 100644
--- a/elf/tst-execstack-prog.c
+++ b/elf/tst-execstack-prog.c
@@ -25,6 +25,7 @@  static void
 deeper (void (*f) (void))
 {
   char stack[1100 * 1024];
+  memset (stack, 123, sizeof stack);
   memfrob (stack, sizeof stack);
   (*f) ();
   memfrob (stack, sizeof stack);
diff --git a/elf/tst-execstack.c b/elf/tst-execstack.c
index 114f341d76..f9a6509be0 100644
--- a/elf/tst-execstack.c
+++ b/elf/tst-execstack.c
@@ -227,6 +227,7 @@  static void
 deeper (void (*f) (void))
 {
   char stack[1100 * 1024];
+  memset (stack, 123, sizeof stack);
   memfrob (stack, sizeof stack);
   (*f) ();
   memfrob (stack, sizeof stack);
diff --git a/string/string.h b/string/string.h
index 04e1b7067d..201d7e31c8 100644
--- a/string/string.h
+++ b/string/string.h
@@ -495,7 +495,7 @@  extern char *strfry (char *__string) __THROW __nonnull ((1));
 
 /* Frobnicate N bytes of S.  */
 extern void *memfrob (void *__s, size_t __n) __THROW __nonnull ((1))
-    __attr_access ((__write_only__, 1, 2));
+    __attr_access ((__read_write__, 1, 2));
 
 # ifndef basename
 /* Return the file name within directory of FILENAME.  We don't