[ofa-general] Re: [PATCHv2 1/2] ipoib: handle pkey change events

Yosef Etigin yosefe at voltaire.com
Tue May 8 08:11:09 PDT 2007


Michael S. Tsirkin wrote:
>>Quoting Yosef Etigin <yosefe at voltaire.com>:
>>Subject: [PATCHv2 1/2] ipoib: handle pkey change events
>>
>>This issue was found during partitioning & SM fail over testing.
>>
>> * Added flush flag to ipoib_ib_dev_stop(), ipoib_ib_dev_down() alike
>> * Rename the polling thread work to 'pkey_poll_task' to avoid ambiguity
>> * Upon PKEY_CHANGE event, schedule a work that restarts the QP
>> * Restart child interfaces before parent
> 
> 
> What's the reason for this change?
> Is this a separate bugfix?
> You might want to put this info in the log.
> 

The reason is that if the child are restarted after the parent, and the parent is
not up, then the flush function returns immediately due to the INITIALLIZED bit test.
Now I think that we might use a goto statement instead.

> 
>> * Use uncached pkey query upon qp initiallization
>>
>>SM reconfiguration or failover possibly causes a shuffling of the values in the port
>>pkey table. The current implementation only queries for the index of the pkey once,
>>when it creates the device QP and after that moves it into working state, and hence
>>does not address this scenario. Fix this by using the PKEY_CHANGE event as a trigger
>>to reconfigure the device QP.
>>
>>
>>Signed-off-by: Yosef Etigin <yosefe at voltaire.com>
> 
> 
> Btw, pls try making log lines a bit shorter - git log shifts everything
> to the right.
> 
Ok



More information about the general mailing list