Quoting r. Sean Hefty <[EMAIL PROTECTED]>:
> Subject: svn checkin to merge upstream
>
> Roland,
>
> I committed a bug fix in svn rev 8629 to the IB CM that I'd like pushed
> upstream
> for 2.6.18. It was possible for a function to return 0 (success) if an
> allocation occurred.
>
> Thanks,
> - Sean
This one (below)?
Seems to do more than it says, and I now get a warning about the "out:" label
being unused. Is it just me?
Further, wouldn't
if (IS_ERR()) {
ret = PTR_ERR(cm_id_priv->timewait_info);
goto out;
}
be more elegant, and a smaller patch?
Finally, Sean, could you post patches inline on the list please?
It is confusing if the version in SVN differs from was last posted.
----
------------------------------------------------------------------------
r8629 | sean.hefty | 2006-07-22 00:19:38 +0300 (Sat, 22 Jul 2006) | 5 lines
Return error code, rather than success (0), on failure allocating timewait_info
in ib_send_cm_req.
Signed-off-by: Sean Hefty <[EMAIL PROTECTED]>
------------------------------------------------------------------------
Index: xxx/drivers/infiniband/core/cm.c
===================================================================
--- xxx/drivers/infiniband/core/cm.c (revision 8628)
+++ xxx/drivers/infiniband/core/cm.c (revision 8629)
@@ -969,15 +969,14 @@ int ib_send_cm_req(struct ib_cm_id *cm_i
spin_lock_irqsave(&cm_id_priv->lock, flags);
if (cm_id->state != IB_CM_IDLE) {
spin_unlock_irqrestore(&cm_id_priv->lock, flags);
- ret = -EINVAL;
- goto out;
+ return -EINVAL;
}
spin_unlock_irqrestore(&cm_id_priv->lock, flags);
cm_id_priv->timewait_info = cm_create_timewait_info(cm_id_priv->
id.local_id);
if (IS_ERR(cm_id_priv->timewait_info))
- goto out;
+ return PTR_ERR(cm_id_priv->timewait_info);
ret = cm_init_av_by_path(param->primary_path, &cm_id_priv->av);
if (ret)
--
MST
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
