[Freeipa-devel] [PATCHES] 366-372 Additional Coverity fixes

[Jan Cholasta]

Dne 11.11.2014 v 11:40 Alexander Bokovoy napsal(a):
> On Tue, 11 Nov 2014, Jan Cholasta wrote:
>>> From 82d7d37ca310af015018ebb2da2f9a72c4dabcaa Mon Sep 17 00:00:00 2001
>> From: Jan Cholasta <jcholast at redhat.com>
>> Date: Mon, 10 Nov 2014 18:10:27 +0000
>> Subject: [PATCH 4/7] Fix unchecked return value in ipa-kdb
>>
>> https://fedorahosted.org/freeipa/ticket/4713
>> ---
>> daemons/ipa-kdb/ipa_kdb_mspac.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/daemons/ipa-kdb/ipa_kdb_mspac.c
>> b/daemons/ipa-kdb/ipa_kdb_mspac.c
>> index c8f6c76..debcd1b 100644
>> --- a/daemons/ipa-kdb/ipa_kdb_mspac.c
>> +++ b/daemons/ipa-kdb/ipa_kdb_mspac.c
>> @@ -2071,6 +2071,9 @@ krb5_error_code ipadb_sign_authdata(krb5_context
>> context,
>>                             ipactx->kdc_hostname,
>> strlen(ipactx->kdc_hostname),
>>                             NULL, NULL, &result) == 0) {
>>                 kerr = ipadb_reinit_mspac(ipactx, true);
>> +                if (kerr != 0 && kerr != ENOENT) {
>> +                    goto done;
>> +                }
>>             }
>>         }
>>
> I'm not sure we should drop the sign_authdata request here. If we were
> able to re-initialize our view of trusted domains, we simply cannot
> re-sign incoming PAC but this is handled in ipadb_verify_pac() and
> ipadb_sign_pac() and if the former returns NULL value for PAC, we exit
> with a return code of 0 while this change will fail a cross-realm TGT
> request unconditionally.
>

OK, what would be a proper fix? Just ignore the return value of 
ipadb_reinit_mspac here?

-- 
Jan Cholasta