net: redirect nftables stdout and stderr to CRIU's log file

[adrianreber]

When using the nftables network locking backend and restoring a process a second time the network locking has already been deleted by the first restore. The second restore will print out to the console text like:

Error: Could not process rule: No such file or directory
delete table inet CRIU-202621

With this change CRIU's log FD is used by libnftables stdout and stderr.

[rst0git]

It looks like the changes in this pull request are also applied in #2550
@adrianreber Would it make sense to close this PR?

[adrianreber]

@avagin Anything missing here from your point of view. It is not exactly the same as #2550. It is related but could be done independently of the discussion in #2550

[adrianreber]

Moved everything to a separate function as suggested by @rst0git and removed the fclose() as suggested by @avagin.

[rst0git]

LGTM

[avagin]

Moved everything to a separate function as suggested by @rst0git and removed the fclose() as suggested by @avagin.

Ooops, I didn't mean that. Without fclose(), it introduces a memory leak, because fdopen allocates a new FILE object. Unfortunately, it is impossible to do fclose() without closing a file description. The only option is to dup a logging file descriptor and operates with a new descriptor.

[adrianreber]

Moved everything to a separate function as suggested by @rst0git and removed the fclose() as suggested by @avagin.

Ooops, I didn't mean that. Without fclose(), it introduces a memory leak, because fdopen allocates a new FILE object. Unfortunately, it is impossible to do fclose() without closing a file description. The only option is to dup a logging file descriptor and operates with a new descriptor.

Thanks for the explanation. Reworked to close the FILE pointers using a dup()ed FD.

[avagin]

LGTM. Thanks for working on this.

[avagin]

Rebasing the commits of this branch on top of the base branch cannot be performed automatically as this would create a different result than a regular merge.

@adrianreber could you rebase it?

[adrianreber]

Rebasing the commits of this branch on top of the base branch cannot be performed automatically as this would create a different result than a regular merge.

@adrianreber could you rebase it?

There is nothing to rebase from what I see. Nothing happens if I try to rebase locally.

[rst0git]

@adrianreber GitHub was showing the message This branch cannot be rebased safely and the "Rebase and merge" button was disabled. I rebased the branch on criu-dev and it seems to work now.