Hi Geliang,
On 22/05/2023 15:11, Geliang Tang wrote:
> v14:
> - drop mptcp_pm_remove_addrs helper in patch 1
> - add two flags in patch 4, the address entry'll be removed from
> userspace_pm_local_addr_list only when both flags are set, by doing this,
> it's now independent of the order of the remove_subflows command and
> the remove_addrs command.
Thank you for this v14.
Please see my replies and questions on the individual patches but in short:
- Patch 1/5:
- Maybe best to keep the helper. You can also modify it to remove just
one address instead of a list of addresses.
- see patch 4/5: you might want to ignore the returned value of
remove_anno_list_by_saddr()
- Patch 3/5:
- see patch 4/5: I think you can remove this one
- Patch 4/5:
- I think it is better to use the version from v12: if the userspace
asks to destroy the subflow, we remove the linked address entry (if not
used by another one). So yes, it means we cannot send a RM_ADDR after
the destruction of the subflow but that sounds OK to me. I don't see why
it is needed to close a subflow (send FIN) and then send a RM_ADDR
- Or am I missing something?
- (if really we want to send a RM_ADDR for an ID we don't have, we can
also create a dummy entry just to be able to send the RM_ADDR but
currently, I don't see why the userspace will need to do that).
- compared to v12, please:
- do not call mptcp_pm_alloc_anno_list() (+
mptcp_pm_remove_anno_list_by_saddr()) from mptcp_nl_cmd_sf_create()
- and move the comment above mptcp_userspace_pm_delete_local_addr()
- Patch 5/5:
- first send a RM_ADDR, then a destroy subflow (or adapt the
verification we do not to expect a remove address)
- Patch 6 (patch 1/7 from your v13 part 2):
- please add it to this series, it is also for -net
Cheers,
Matt
--
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net