[20189] in Kerberos_V5_Development
Re: Assertion failure due to repeated K5_KEY_GSS_SPNEGO_STATUS
daemon@ATHENA.MIT.EDU (Adam Dabrowski)
Fri Oct 16 04:34:24 2020
To: Greg Hudson <ghudson@mit.edu>, <krbdev@mit.edu>
From: Adam Dabrowski <adabrowski@nomachine.com>
Message-ID: <097fe0aa-c7e1-6cd4-08e6-626707f4ae5f@nomachine.com>
Date: Fri, 16 Oct 2020 10:34:03 +0200
MIME-Version: 1.0
In-Reply-To: <b632f154-b76c-1c3d-779f-ed263e010ee3@mit.edu>
Content-Language: en-GB
Content-Type: text/plain; charset="utf-8"
Errors-To: krbdev-bounces@mit.edu
Content-Transfer-Encoding: 8bit
W dniu 16.10.2020 o 00:55, Greg Hudson pisze:
> On 10/15/20 5:49 PM, Adam Dabrowski wrote:
>> 1. Libgssapi_krb.so.2.2 is loaded dynamically with dlopen.
>> 2. Call to gss_init_sec_context - library initialization, first
>> registration of K5_KEY_GSS_SPNEGO_STATUS.
>> 3. libgssapi_krb.so.2.2 is unloaded, however libkrb5support stays in the
>> process memory.
>> 4. Steps 1., 2..
>> 5. Abort due to failed assertion 'destructors_set[keynum] == 0' in
>> k5_key_register.
> If you edit gss_spnegoint_lib_fini() and add:
>
> k5_key_delete(K5_KEY_GSS_SPNEGO_STATUS);
>
> does the problem go away? I think this was neglected in commit
> d160bc733a3dbeb6d84f4e175234ff18738d9f66 .
Yes, as I mentioned in previous mail, I've already tried that and it
prevents crashes.
I just wasn't sure if empty gss_spnegoint_lib_fini is simple oversight
or has some rationale
behind it.
_______________________________________________
krbdev mailing list krbdev@mit.edu
https://mailman.mit.edu/mailman/listinfo/krbdev
