> How have you tested this? Seems like some significantly random
> connection fault injection is needed, when doing active multichannel
> load testing to a Windows server sku.
I tested this to Windows by pausing the Windows server VM in HyperV to
force timeout, then resumed the VM to see reconnect (I also tried this
with the Windows client to see how it updated the ChannelSequence).
I also tried this to Samba by dropping the network interface and
letting the requests time out ("ifconfig lo down") and then doing
"ifconfig lo up" and letting it reconnect
On Thu, Aug 24, 2023 at 11:51 PM Steve French <smfrench@xxxxxxxxx> wrote:
>
> The ChannelSequence field in the SMB3 header is supposed to be
> increased after reconnect to allow the server to distinguish
> requests from before and after the reconnect. We had always
> been setting it to zero. There are cases where incrementing
> ChannelSequence on requests after network reconnects can reduce
> the chance of data corruptions.
>
> See MS-SMB2 3.2.4.1 and 3.2.7.1
>
> Note that (as Tom Talpey pointed out) a macro "CIFS_SERVER_IS_CHAN"
> used by this patch is confusing (has a confusing name) since
> multichannel is not supported for older dialects like CIFS. I will
> fix that macro name in a followon patch.
>
> --
> Thanks,
>
> Steve
>
>
> --
> Thanks,
>
> Steve
--
Thanks,
Steve
