Re: [PATCH 3/6] libsemanage: never call memcpy with a NULL value

[William Roberts <bill.c.roberts@xxxxxxxxx>]

On Feb 27, 2017 2:16 PM, "William Roberts" <bill.c.roberts@xxxxxxxxx> wrote:

On Feb 27, 2017 12:42, "Nicolas Iooss" <nicolas.iooss@xxxxxxx> wrote:

clang's static analyzer reports "Argument with 'nonnull' attribute
passed null" in append_str(), because argument t may be NULL but is used
in a call to memcpy().

Make append_str() do nothing when called with t=NULL.

Signed-off-by: Nicolas Iooss <nicolas.iooss@xxxxxxx>
---
 libsemanage/src/semanage_store.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/libsemanage/src/semanage_store.c b/libsemanage/src/semanage_store.c
index f468faba4b64..47ec93185e06 100644
--- a/libsemanage/src/semanage_store.c
+++ b/libsemanage/src/semanage_store.c
@@ -1194,8 +1194,14 @@ static char *append(char *s, char c)
 static char *append_str(char *s, const char *t)
 {
        size_t s_len = (s == NULL ? 0 : strlen(s));
-       size_t t_len = (t == NULL ? 0 : strlen(t));
-       char *new_s = realloc(s, s_len + t_len + 1);
+       size_t t_len;
+       char *new_s;
+
+       if (t == NULL) {
+               return s;
+       }
+       t_len = strlen(t);
+       new_s = realloc(s, s_len + t_len + 1);

Overflow possibility here?

I guess since s and t lengths come from strlen() and the architectures we worry about running code on, overflowing would be pretty impossible here.

        if (new_s == NULL) {
                return NULL;
        }
--
2.11.1

_______________________________________________
Selinux mailing list
Selinux@xxxxxxxxxxxxx
To unsubscribe, send email to Selinux-leave@xxxxxxxxxxxxx.
To get help, send an email containing "help" to Selinux-request@xxxxxxxxxxxxx.

_______________________________________________
Selinux mailing list
Selinux@xxxxxxxxxxxxx
To unsubscribe, send email to Selinux-leave@xxxxxxxxxxxxx.
To get help, send an email containing "help" to Selinux-request@xxxxxxxxxxxxx.