On 8/6/18, 4:15 PM, "[email protected] on behalf of Ben Pfaff"
<[email protected] on behalf of [email protected]> wrote:
On Mon, Aug 06, 2018 at 03:55:55PM -0700, Darrell Ball wrote:
> By making opt_dpif_open() more general, it can be used effectively
> by all potential callers. Also, the error handling is improved by
> having more specific errors.
>
> Signed-off-by: Darrell Ball <[email protected]>
I'm not sure I understand the improvement yet. It looks to me like this
only provides value if the user specifies a dp name that contains @.
I'm not sure that's something that is commonly done; I only do it in
extremis.
Right, name syntax is not enforced; not sure if that is good per se, but
admittedly,
I had not given this much thought b4. I made a correction.
It looks like the algorithm enumerates all the datapaths and opens them
in turn and then checks whether the name is what the user asked for. Is
this better than just trying to open what the user asked for and seeing
if it worked?
Overall, this allows for more correct error handling since we don’t try to open
bogus datapaths provided by the user as b4. The new code tries to open
datapaths that we expect
exist. However, upon revisiting the code, I see no reason to be opening (and
closing) existing
datapaths in dp_exists(). I just removed this useless part.
Thanks,
Ben.
_______________________________________________
dev mailing list
[email protected]
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.openvswitch.org%2Fmailman%2Flistinfo%2Fovs-dev&data=02%7C01%7Cdball%40vmware.com%7C61732b29c57c426ba09908d5fbf1c309%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636691941526057539&sdata=4zj6MzDxfExCkO%2F%2F8ER534ur0142%2Fm3MvPntexUmF8c%3D&reserved=0
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
