Initial TFTP support #3911

DipSwitch · 2015-09-21T10:35:45Z

Options that where thrown up by @authmillenon

Re-Write to use conn API depends on #3921

Con: Re-Writing to conn loses the capability of zero-copy since I now could allocate the packet buffer at maximum size and reallocate it to the size that it actually is after filling it.

OlegHahm · 2015-09-21T11:10:21Z

Nice one!

cgundogan · 2015-09-21T11:46:48Z

sys/net/gnrc/application_layer/tftp/gnrc_tftp.c

+ */
+typedef struct __attribute__((packed)) {
+    tftp_opcodes_t opc          : 16;
+    uint16_t block_nr;


network_uint16_t ?

I can't see the change. push? (:

miri64 · 2015-09-21T19:43:45Z

@DipSwitch may it be possible to use the conn interface for application protocols? This way other future network stacks would also profit from your implementation.

DipSwitch · 2015-09-21T20:42:07Z

I'll check in a sec. How can I implement an timeout msg if I refactor to conn? I'll be on IRC in a minute to avoid spam ;)

miri64 · 2015-09-21T21:06:10Z

I'm only on my phone and about to sleep, so won't be on IRC tonight. Sorry :(

DipSwitch · 2015-09-21T21:10:02Z

Oke np, maybe we could steal an idea of LwIP where we create int gnrc_conn_recvfrom_timeout(conn_t *conn, void *data, size_t max_len, void *addr, size_t *addr_len, uint16_t *port, uint32_t timeout_ms) where the conn API also waits for a predefined timeout message type?

DipSwitch · 2015-09-21T21:20:05Z

This would solve the timeout problem for the TFTP client application. For the server implementation I would need to check, it would probably come down to creating a thread / task for each client.

daniel-k · 2015-10-20T14:02:37Z

sys/include/net/gnrc/tftp.h

+=======
+ * @brief Start an TFTP server
+ *
+>>>>>>> ce6ac59aff53a20b2a06b65b722595e08a9c084a


You have a merge conflict here

miri64 · 2015-12-01T17:57:55Z

Is it sending or receiving? For receiving packets this is the correct order.

DipSwitch · 2015-12-01T18:14:35Z

It is receiving, but the assert was never hit before...

OlegHahm · 2015-12-01T18:24:30Z

Apparently it's happening only for 6LoWPAN.

OlegHahm · 2015-12-01T18:26:26Z

pkt->next->next->next is of type GNRC_NETTYPE_NETIF, pkt is of type GNRC_NETTYPE_UNDEF.

Hence, we have undef->IPv6->UDP->netif which seems odd to me - in both directions. Could this reordering happen during IPHC decompression?

DipSwitch · 2015-12-01T18:26:51Z

But not on my board, on my board the whole network stack seems to stop for a while. But not tested with the latest master. I'll test latest master first thing in the morning.

OlegHahm · 2015-12-01T18:45:44Z

OlegHahm@41569b9 would fix it, but seems to be a workaround. @authmillenon, what's your opinion?

miri64 · 2015-12-01T19:20:20Z

Your way looks definitely more like it was done in GNRC for the most part.

DipSwitch · 2015-12-02T07:41:16Z

Merged and squashed

emmanuelsearch · 2015-12-03T15:02:16Z

@DipSwitch can you open an issue about the header order swap?

emmanuelsearch · 2015-12-03T15:02:20Z

other than that, I'd say: ACK & merge when T is happy?

emmanuelsearch · 2015-12-03T15:57:28Z

One Travis job consistently fails because it exceeds 50min...

DipSwitch · 2015-12-03T16:24:08Z

The build has succeeded without errors in the past if I remember correct.

OlegHahm · 2015-12-03T16:26:52Z

Yes, I saw it, too. Let's give Travis one more chance.

OlegHahm · 2015-12-03T17:23:07Z

Damn! Would you mind rebasing one more time as soon as #4399 is merged? (You may call me paranoid.)

DipSwitch · 2015-12-03T17:28:35Z

Sure no problem :)

cgundogan · 2015-12-03T18:46:45Z

@DipSwitch can you rebase now? #4399 was merged

DipSwitch · 2015-12-03T20:11:38Z

Oops lol, wait, let me fix this xD

Fix: packet size and typo in transfer mode main -> mail

Initial TFTP support

OlegHahm · 2015-12-04T00:36:43Z

Great job, thanks!

daniel-k · 2015-12-04T00:56:37Z

Woohoo, nice! :)

DipSwitch · 2015-12-04T06:02:15Z

Whoohoo! That took a while :D

OlegHahm added this to the Release NEXT MAJOR milestone Sep 21, 2015

OlegHahm added Area: network Area: Networking Type: new feature The issue requests / The PR implemements a new feature for RIOT labels Sep 21, 2015

OlegHahm assigned x3ro Sep 21, 2015

cgundogan reviewed Sep 21, 2015
View reviewed changes

DipSwitch force-pushed the tftp_client branch from bf9f24c to ce6ac59 Compare September 22, 2015 09:22

daniel-k reviewed Oct 20, 2015
View reviewed changes

DipSwitch force-pushed the tftp_client branch from c1398b8 to bbfee68 Compare December 2, 2015 07:41

DipSwitch mentioned this pull request Dec 3, 2015

gnrc_sixlowpan: Header order or UDP and IPv6 swapped #4397

Closed

OlegHahm mentioned this pull request Dec 3, 2015

CI: split up Cortex-M4 in two groups #4399

Merged

DipSwitch force-pushed the tftp_client branch 2 times, most recently from bbfee68 to 274e211 Compare December 3, 2015 20:23

tftp: implement TFTP server and client

ede923b

Fix: packet size and typo in transfer mode main -> mail

DipSwitch force-pushed the tftp_client branch from ef62d6e to ede923b Compare December 3, 2015 20:46

OlegHahm added a commit that referenced this pull request Dec 4, 2015

Merge pull request #3911 from DipSwitch/tftp_client

09aa558

Initial TFTP support

OlegHahm merged commit 09aa558 into RIOT-OS:master Dec 4, 2015

DipSwitch deleted the tftp_client branch December 16, 2015 22:46

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Initial TFTP support #3911

Initial TFTP support #3911

DipSwitch commented Sep 21, 2015

OlegHahm commented Sep 21, 2015

cgundogan Sep 21, 2015

DipSwitch Sep 21, 2015

cgundogan Sep 21, 2015

miri64 commented Sep 21, 2015

DipSwitch commented Sep 21, 2015

miri64 commented Sep 21, 2015

DipSwitch commented Sep 21, 2015

DipSwitch commented Sep 21, 2015

daniel-k Oct 20, 2015

miri64 commented Dec 1, 2015

DipSwitch commented Dec 1, 2015 via email

OlegHahm commented Dec 1, 2015

OlegHahm commented Dec 1, 2015

DipSwitch commented Dec 1, 2015 via email

OlegHahm commented Dec 1, 2015

miri64 commented Dec 1, 2015

DipSwitch commented Dec 2, 2015

emmanuelsearch commented Dec 3, 2015

emmanuelsearch commented Dec 3, 2015

emmanuelsearch commented Dec 3, 2015

DipSwitch commented Dec 3, 2015

OlegHahm commented Dec 3, 2015

OlegHahm commented Dec 3, 2015

DipSwitch commented Dec 3, 2015 via email

cgundogan commented Dec 3, 2015

DipSwitch commented Dec 3, 2015

OlegHahm commented Dec 4, 2015

daniel-k commented Dec 4, 2015

DipSwitch commented Dec 4, 2015 via email

Initial TFTP support #3911

Initial TFTP support #3911

Conversation

DipSwitch commented Sep 21, 2015

OlegHahm commented Sep 21, 2015

cgundogan Sep 21, 2015

Choose a reason for hiding this comment

DipSwitch Sep 21, 2015

Choose a reason for hiding this comment

cgundogan Sep 21, 2015

Choose a reason for hiding this comment

miri64 commented Sep 21, 2015

DipSwitch commented Sep 21, 2015

miri64 commented Sep 21, 2015

DipSwitch commented Sep 21, 2015

DipSwitch commented Sep 21, 2015

daniel-k Oct 20, 2015

Choose a reason for hiding this comment

miri64 commented Dec 1, 2015

DipSwitch commented Dec 1, 2015 via email

OlegHahm commented Dec 1, 2015

OlegHahm commented Dec 1, 2015

DipSwitch commented Dec 1, 2015 via email

OlegHahm commented Dec 1, 2015

miri64 commented Dec 1, 2015

DipSwitch commented Dec 2, 2015

emmanuelsearch commented Dec 3, 2015

emmanuelsearch commented Dec 3, 2015

emmanuelsearch commented Dec 3, 2015

DipSwitch commented Dec 3, 2015

OlegHahm commented Dec 3, 2015

OlegHahm commented Dec 3, 2015

DipSwitch commented Dec 3, 2015 via email

cgundogan commented Dec 3, 2015

DipSwitch commented Dec 3, 2015

OlegHahm commented Dec 4, 2015

daniel-k commented Dec 4, 2015

DipSwitch commented Dec 4, 2015 via email