https://fedorahosted.org/sssd/ticket/1869
Currently the private data passed to the PAM request is a structure allocated on the client context. But in the odd case where the back end would be stopped or stuck until the idle timeout hits, the DP callback.
This patch introduces a new structure allocated on responder context, whose only purpose is to live as long as the request is active.
The attached patch is mostly a workaround, I think the correct fix would be to start using a tevent_req-styled request structure similar to the one we use in the NSS responder, but we've seen users hitting this problem even on RHEL5 with sssd-1-5 recently, so I would like to fix the problem in some way. If this patch is accepted, I'll file a ticket to convert the PAM responder to a tevent_req interface for better manageability.
On 04/05/2013 05:12 PM, Jakub Hrozek wrote:
https://fedorahosted.org/sssd/ticket/1869
Currently the private data passed to the PAM request is a structure allocated on the client context. But in the odd case where the back end would be stopped or stuck until the idle timeout hits, the DP callback.
This patch introduces a new structure allocated on responder context, whose only purpose is to live as long as the request is active.
The attached patch is mostly a workaround, I think the correct fix would be to start using a tevent_req-styled request structure similar to the one we use in the NSS responder, but we've seen users hitting this problem even on RHEL5 with sssd-1-5 recently, so I would like to fix the problem in some way. If this patch is accepted, I'll file a ticket to convert the PAM responder to a tevent_req interface for better manageability.
Works fine. Ack.
On Mon, Apr 08, 2013 at 02:31:37PM +0200, Pavel Březina wrote:
On 04/05/2013 05:12 PM, Jakub Hrozek wrote:
https://fedorahosted.org/sssd/ticket/1869
Currently the private data passed to the PAM request is a structure allocated on the client context. But in the odd case where the back end would be stopped or stuck until the idle timeout hits, the DP callback.
This patch introduces a new structure allocated on responder context, whose only purpose is to live as long as the request is active.
The attached patch is mostly a workaround, I think the correct fix would be to start using a tevent_req-styled request structure similar to the one we use in the NSS responder, but we've seen users hitting this problem even on RHEL5 with sssd-1-5 recently, so I would like to fix the problem in some way. If this patch is accepted, I'll file a ticket to convert the PAM responder to a tevent_req interface for better manageability.
Works fine. Ack.
Pushed to master.
On Mon, Apr 08, 2013 at 11:59:39PM +0200, Jakub Hrozek wrote:
On Mon, Apr 08, 2013 at 02:31:37PM +0200, Pavel Březina wrote:
On 04/05/2013 05:12 PM, Jakub Hrozek wrote:
https://fedorahosted.org/sssd/ticket/1869
Currently the private data passed to the PAM request is a structure allocated on the client context. But in the odd case where the back end would be stopped or stuck until the idle timeout hits, the DP callback.
This patch introduces a new structure allocated on responder context, whose only purpose is to live as long as the request is active.
The attached patch is mostly a workaround, I think the correct fix would be to start using a tevent_req-styled request structure similar to the one we use in the NSS responder, but we've seen users hitting this problem even on RHEL5 with sssd-1-5 recently, so I would like to fix the problem in some way. If this patch is accepted, I'll file a ticket to convert the PAM responder to a tevent_req interface for better manageability.
Works fine. Ack.
Pushed to master.
Also pushed to sssd-1-9
On Fri, Apr 05, 2013 at 05:12:45PM +0200, Jakub Hrozek wrote:
If this patch is accepted, I'll file a ticket to convert the PAM responder to a tevent_req interface for better manageability.
This ticket is: https://fedorahosted.org/sssd/ticket/1878
sssd-devel@lists.fedorahosted.org
-
Jakub Hrozek -
Pavel Březina