bgpd: fix a possible crash issue (code review)

[zice312963205]

The variable 'adj' used in this loop is not assigned a new value within the loop, instead it keeps using the same 'adj' assigned earlier, which may cause an exception in the second iteration of the loop.

[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-PULLREQ2-13240/

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.

[ton31337]

This is the same issue #13625.

[ton31337]

@donaldsharp had some insightful comments regarding this behavior on how adj can be NULL (in what cases).

[riw777]

looks good

[riw777]

This is the same issue #13625.

Seems like maybe we should close this one and get #13625 in?

[zice312963205]

This is the same issue #13625.

Seems like maybe we should close this one and get #13625 in?

Sure. But later on, I reconsidered this issue and believed that a better approach should be used to address it. I have already provided specific comments in PR #13625 .

[zice312963205]

A better approach seems to be moving the route to the withdraw queue when a larger attribute is discovered. Otherwise, if the route was previously sent with a smaller attribute, it cannot be properly updated to other neighbors, and a better way to handle it may be to withdraw the route from the neighbors.(?) Will close this PR and track it in PR #13625