lib, zebra: Ensure route encoding has enough space

[donaldsharp]

When you have compiled FRR with a large multipath number
then encoding large ecmp routes between zebra and the
routing daemons. There exists a theoritical size
of multipath that will cause the encoding to be larger
than the ZEBRA_MAX_PACKET_SIZ. In the cases where
we have allocated streams that will encode routes
then let's ensure that whatever size we have will
auto-fit what we say we can send.

Signed-off-by: Donald Sharp sharpd@cumulusnetworks.com

[donaldsharp]

I'm going to put this into 6 and 7 and 7.1 dev once approved

[LabN-CI]

💚 Basic BGPD CI results: SUCCESS, 0 tests failed

Results table

_	_
Result	SUCCESS git merge/4435 7c1e168
Date	05/31/2019
Start	09:52:24
Finish	10:15:12
Run-Time	22:48
Total	1813
Pass	1813
Fail	0
Valgrind-Errors	0
Valgrind-Loss	0
Details	vncregress-2019-05-31-09:52:24.txt
Log	autoscript-2019-05-31-09:53:40.log.bz2
Memory	427 434 360

For details, please contact louberger

[LabN-CI]

💚 Basic BGPD CI results: SUCCESS, 0 tests failed

Results table

_	_
Result	SUCCESS git merge/4435 f3f4562
Date	05/31/2019
Start	11:07:00
Finish	11:29:42
Run-Time	22:42
Total	1813
Pass	1813
Fail	0
Valgrind-Errors	0
Valgrind-Loss	0
Details	vncregress-2019-05-31-11:07:00.txt
Log	autoscript-2019-05-31-11:08:07.log.bz2
Memory	435 433 360

For details, please contact louberger

[NetDEF-CI]

Continuous Integration Result: SUCCESSFUL

Congratulations, this patch passed basic tests

Tested-by: NetDEF / OpenSourceRouting.org CI System

CI System Testrun URL: https://ci1.netdef.org/browse/FRR-FRRPULLREQ-7904/

This is a comment from an automated CI system.
For questions and feedback in regards to this CI system, please feel free to email
Martin Winter - mwinter (at) opensourcerouting.org.

---

CLANG Static Analyzer Summary

• Github Pull Request 4435, comparing to Git base SHA 4cc1365

No Changes in Static Analysis warnings compared to base

11 Static Analyzer issues remaining.

See details at
https://ci1.netdef.org/browse/FRR-FRRPULLREQ-7904/artifact/shared/static_analysis/index.html

[mjstapp]

good catch ...
but I'm a bit concerned about having to repeat the correct comparison/max expression. couldn't you do something like:

#define ZEBRA_MAX_PACKET_SIZ MAX(16384U, sizeof(struct zapi_route))

that'd put the operation in just one place?

[donaldsharp]

so for a long time I've been ruminating about revamping the stream allocation sizes across the board. Which is why I tried to point fix this as much as possible. This is especially true when we have a bunch of data to pass back and forth between daemons. We have poorly sized buffers and we end up with a large amount of memory allocated and stored waiting to be sent.

[mjstapp]

so for a long time I've been ruminating about revamping the stream allocation sizes [...]

I'm fine with the change - just wanted to ask the question

[mjstapp]

Approving, understand the issue better now!