Printing function for VXLAN GPE and NSH

[bugyo]

This pull request is the repost of #487.
I squashed the verbose commit logs and rebased on the current master branch.
Would it be acceptable?

---

I Added verbose output for encapsulated packets with Generic Protocol Extension for VXLAN and Network Service Header.
they are described in following documents.

draft-quinn-vxlan-gpe-04.txt ( https://datatracker.ietf.org/doc/draft-quinn-vxlan-gpe/ )
draft-ietf-sfc-nsh-01.txt ( http://datatracker.ietf.org/doc/draft-ietf-sfc-nsh/ )
I think that researchers and developers related on Service Function Chaining will be happy if this patch is merged.

[infrastation]

draft-quinn-vxlan-gpe has been replaced with draft-ietf-nvo3-vxlan-gpe, does this code still stand for the latest revision?

[bugyo]

Yes, draft-ietf-nvo3-vxlan-gpe did not change the alignment of the header. Only the latest draft allows MPLS payload. I will commit a supported version.

[bugyo]

OK, now this pull request stands for the latest draft.

[infrastation]

Thank you, please allow some time for somebody to review in better detail. Meanwhile it would help if you could add a make check test case for it based on a sample packet capture.

[bugyo]

I also think that. I will add a test case and am waiting for some comments. Thank you.

[bugyo]

OK, I added a test case for my code.

[fxlb]

Build test: there are some warnings.
See CONTRIBUTING file part 4).

[fxlb]

Please update according to f25ed1f.

[bugyo]

OK, I updated the code according to f25ed1f and correct warnings with -Wall -Wmissing-prototypes -Wstrict-prototypes. Thank you.

[bugyo]

This pull request merged a upa's contribution which enable to display MD-type 2 NSH Header.

[fxlb]

Thanks, I will review in a few days.

[fxlb]

Some comments after a first review:

A. A bound test was missing in print-vxlan.c, added in 20c9007. Also added a #define for more readability. Please update print-vxlan-gpe.c in accordance with this commit.
Same for print-nsh.c.

B. print-nsh.c, line 89: The loop is only for MD Type == 0x01 ?

C. Which code for MD Type == 0x02?

D. print-nsh.c, line 90: before extracting 32 bits you need to check for remaining data with f.e. ND_TCHECK2.

E. The usual way to print the VXLAN-GPE flags is:

static const struct tok vxlan_gpe_flags [] = {
    { 0x08, "I" },
    { 0x04, "P" },
    { 0x01, "O" },
    { 0, NULL }
};

[...]

    ND_PRINT((ndo, "flags [%s], ", bittok2str_nosep(vxlan_gpe_flags,
              "none", flags)));

Same for print-nsh.c.

F. print-nsh.c, line 109 & print-vxlan-gpe.c, line 98: Is it truncated? Not just invalid next protocol?

G. As the output depend on vflag, you should add more test cases with f.e.:

nsh-over-vxlan-gpe nsh-over-vxlan-gpe.pcap nsh-over-vxlan-gpe.out -t
nsh-over-vxlan-gpe-v nsh-over-vxlan-gpe.pcap nsh-over-vxlan-gpe-v.out -t -v
nsh-over-vxlan-gpe-vv nsh-over-vxlan-gpe.pcap nsh-over-vxlan-gpe-vv.out -t -vv
nsh-over-vxlan-gpe-vvv nsh-over-vxlan-gpe.pcap nsh-over-vxlan-gpe-vvv.out -t -vvv

H. print-nsh.c & print-vxlan-gpe.c: Remove space at end of line 5.

I. X < ndo->ndo_vflag tests: We usually test with: ndo->ndo_vflag > X

[bugyo]

First, Thank you for your review.

A. A bound test was missing in print-vxlan.c, added in 20c9007. Also added a #define for more readability. Please update print-vxlan-gpe.c in accordance with this commit.
Same for print-nsh.c.

Sure. I will correct.

B. print-nsh.c, line 89: The loop is only for MD Type == 0x01 ?
C. Which code for MD Type == 0x02?

This code is for both MD Types 0x01 and 0x02. They have the same syntax except for MD Type 0x01 fixes Length field == 0x06. But I noticed that the draft specifies the format for Optional Variable Length Metadata. I will support it.

E. The usual way to print the VXLAN-GPE flags is:

I will correct my code using bittok2str_nosep.

F. print-nsh.c, line 109 & print-vxlan-gpe.c, line 98: Is it truncated? Not just invalid next protocol?

Yes, It just invalid or unknown next protocol.
I will correct my code like print-gre.c.

G. As the output depend on vflag, you should add more test cases with f.e.:

Yes.

H. print-nsh.c & print-vxlan-gpe.c: Remove space at end of line 5.

Yes.

I. X < ndo->ndo_vflag tests: We usually test with: ndo->ndo_vflag > X

OK.

[bugyo]

I fixed my code according to the comment.

[fxlb]

Thanks. Please give me a few days to check once again.

[fxlb]

Merged with space editing.
Thanks for your work and patience!