[PATCH] Try again if SSL_write() fails to write

Murilo Opsfelder Araujo muriloo at linux.vnet.ibm.com
Fri Aug 3 14:22:55 EDT 2012


On 08/03/2012 11:54 AM, Woodhouse, David wrote:
> On Fri, 2012-08-03 at 09:41 -0300, Murilo Opsfelder Araujo wrote:
>> From: David Woodhouse<David.Woodhouse at intel.com>
>>
>> There are cases that SSL_write() returns SSL_ERROR_WANT_WRITE when, for
>> instance, the socket buffer is full and could not write.  This causes
>> openconnect to terminate and send BYE packet.
>>
>> This patch forces openconnect to try again until timeout is reached.
>>
>> Signed-off-by: David Woodhouse<David.Woodhouse at intel.com>
>
> You added that Signed-off-by: yourself, didn't you? I certainly hadn't
> granted it, before I went through and checked the rest of the
> surrounding code to make sure the patch was really correct.

Yes, I did.  I've used the patch you sent me via IRC and thought it was 
signed-off-by you.

> Please make sure you never make up a Signed-off-by for anyone else.

Sorry for the misunderstanding and thanks for the advice.

> I committed a variant of the patch this morning, after doing some of
> that extra checking.
>
> I think you found this using the EPEL packages? If you could file a bug
> in Red Hat bugzilla, that would be appreciated; it gives me the excuse I
> need for updating the package...

I've filed a bug:  https://bugzilla.redhat.com/show_bug.cgi?id=845636

Thanks (again) for the help and support.

-- 
Murilo




More information about the openconnect-devel mailing list