gnrc: integrate gnrc_netif2#7424
gnrc: integrate gnrc_netif2#7424miri64 merged 1 commit intoRIOT-OS:gnrc_netif2_integration/masterfrom
Conversation
ede4a9c to
2522259
Compare
f45f37b to
61e9ef0
Compare
50ccef3 to
9f1a304
Compare
|
Rebased and adapted to current master and current dependencies |
9f1a304 to
65ae715
Compare
c02ef0c to
eba909a
Compare
|
Rebased to current master and dependencies |
The function `_update_nce_ar_state()` was introduced during the review of \RIOT-OS#7424, but it's return value never used, causing faulty behavior.
|
I changed the base branch now to the freshly created |
Seems like I already did that ^^" Ready for review. |
| USEMODULE += ipv6_addr | ||
| endif | ||
|
|
||
| ifneq (,$(filter gnrc_ipv6_nib_6lbr,$(USEMODULE))) |
There was a problem hiding this comment.
why do you remove the nib dependency information?
|
Mhhhhh problem with all the non-unintrusive porting now is: Murdock fails to build (actually only to link), since old NDP isn't ported (and I don't plan to). This is now merged into a seperate branch. Do we accept the temporary broken state? |
8f4a5dd to
f4b9474
Compare
|
Rebased to current gnrc_netif2_integration/master. |
bergzand
left a comment
There was a problem hiding this comment.
First part of the PR revieweded (Up to (and including) sys/net/gnrc/network_layer/ipv6/gnrc_ipv6.c).
So far only three minor things. Until now everything looks like obvious old-new code replacements.
examples/gnrc_minimal/main.c
Outdated
| ipv6_addr_to_str(ipv6_addr, &entry->addrs[i].addr, IPV6_ADDR_MAX_STR_LEN); | ||
| printf("My address is %s\n", ipv6_addr); | ||
| } | ||
| /* get interface and print their addresses */ |
There was a problem hiding this comment.
Small grammatical error between "interface" (singular) and "their" (plural)
There was a problem hiding this comment.
"their" can be singular, but I guess this does not apply for things 😛
examples/gnrc_minimal/main.c
Outdated
| int res = gnrc_netapi_get(netif->pid, NETOPT_IPV6_ADDR, 0, ipv6_addrs, | ||
| sizeof(ipv6_addrs)); | ||
|
|
||
| for (int i = 0; i < (res / sizeof(ipv6_addr_t)); i++) { |
There was a problem hiding this comment.
I'm getting a signed vs unsigned comparison error here (gcc-5.4.0)
examples/gnrc_networking/Makefile
Outdated
| # Add a routing protocol | ||
| USEMODULE += gnrc_rpl | ||
| USEMODULE += auto_init_gnrc_rpl | ||
| # USEMODULE += auto_init_gnrc_rpl |
There was a problem hiding this comment.
I assume this depends on the RPL port to gnrc_ipv6_nib?
|
@bergzand addressed your comments |
| gnrc_ipv6_netif_addr_t *netif_addr) | ||
| bool gnrc_rpl_dodag_init(gnrc_rpl_instance_t *instance, ipv6_addr_t *dodag_id, kernel_pid_t iface) | ||
| { | ||
| /* TODO: check if netif_addr belongs to iface */ |
There was a problem hiding this comment.
This todo looks obsolete now that the netif_addr argument is removed
There was a problem hiding this comment.
I would like to keep it in for the person that is porting RPL (@cgundogan) so he can decide how they get the interface address.
sys/shell/commands/sc_fib.c
Outdated
| && (strcmp("lifetime", argv[5]) == 0)) { | ||
| kernel_pid_t ifs[GNRC_NETIF_NUMOF]; | ||
| size_t ifnum = gnrc_netif_get(ifs); | ||
| size_t ifnum = gnrc_netif2_numof(); |
sys/shell/commands/sc_gnrc_6ctx.c
Outdated
| gnrc_ndp_internal_send_rtr_adv(ifs[i], NULL, NULL, false); | ||
| } | ||
| } | ||
| #endif |
|
Just to confirm, I'm getting compilation errors (not only linking errors) with the gnrc_networking example. All due to missing They are gone when testing with #7456, so I don't really mind as long as it is known. |
It is known :-) |
|
Addressed comments |
|
Thanks for the changes and clarifications. Looks good to me now |
Not link-able, since NDP and NC are missing (intentionally)
5bf981b to
861035f
Compare
|
Squashed |
|
@smlng can we dismiss your review? |
|
@miri64 general question (and suggestion if you want): netif2 will replace netif, right? As you now PRed against a feature branch it would IMHO make perfect sense to use |
I will do that, but not in this PR. In fact that's intended to be the very last PR against #7925 before reintegrating it into master ;) |
changes addressed, but haven't had time for 2 second review round
|
Merged intentionally with broken Murdock state |
Does not link ATM, since I did not bother to port the old NC and NDP (will provide a port for #7014 in a separate PR though [#7456]) and those modules now complain that they don't find the old interfaces functions. Since this change is unintrusive (and I do not provide special new applications for this), this should be okay.
Depends on #7410 and #7409 and all their dependencies.This PR is part of the network layer remodelling effort:
