utf8s_to_utf16s could return -EINVAL(-22), then uniname[i] =
uniname[-22] = '\0' will corrupt other local variable.
uniname is a local variable having different address everytime
when call nls_cstring_to_uniname. so uniname[-22] could corrupt
some different addresses.
and also when filename reaches at maximum length, need to leave
last one space in buffer for '\0'
Change-Id: I44bbdd3249150ad5617c7cd13207cfa37b102a77
Signed-off-by: Shengzhe Zhao <a18689@motorola.com>
Signed-off-by: Shiyong Li <a22381@motorola.com>
Reviewed-on: https://gerrit.mot.com/
1102179
SLTApproved: Slta Waiver
SME-Granted: SME Approvals Granted
Tested-by: Jira Key
Reviewed-by: Igor Kovalenko <igork@motorola.com>
Submit-Approved: Jira Key
Reviewed-on: https://gerrit.mot.com/
1276880
Reviewed-by: Hua Tan <tanhua1@motorola.com>
if (nls == NULL) {
len = utf16s_to_utf8s(uniname, MAX_NAME_LENGTH, UTF16_HOST_ENDIAN, p_cstring, MAX_NAME_LENGTH);
+ if (len >= MAX_NAME_LENGTH * MAX_CHARSET_SIZE)
+ len--;
+
p_cstring[len] = 0;
return;
}
#endif
for (j = 0; j < i; j++)
SET16_A(upname + j * 2, nls_upper(sb, uniname[j]));
- uniname[i] = '\0';
+
+ if (i >= 0 && i < MAX_NAME_LENGTH)
+ uniname[i] = '\0';
+ else
+ lossy = TRUE;
}
else {
i = j = 0;