Do not report random errors as table-managing errors in ipfw

[lytboris_gmail.com]

Summary

https://github.com/freebsd/freebsd-src/pull/1221 (commit) is not needed as table_do_modify_record() copies errno into the return value.

Current code grabs errno from a random failed syscall. E.g. in https://lists.freebsd.org/archives/freebsd-current/2025-October/009048.html it gets an Inappropriate ioctl for device error from a isatty() call way up in the calling stack.

The fix is to remove copying completely in table_modify_record()

Test Plan

sh
#!/bin/sh

IPFW=/usr/obj/usr/src/amd64.amd64/sbin/ipfw/ipfw.full

tbl_block=4

${IPFW} table $tbl_block create type addr or-flush
while read -r line
do ${IPFW} table $tbl_block atomic add $line
done  <<-HEREDOC
45.143.220.228/32
68.64.211.117/32
68.64.211.117/32
45.143.220.229/32
HEREDOC

Yields:

added: 45.143.220.228/32 0
added: 68.64.211.117/32 0
exists: 68.64.211.117/32 0
ipfw.full: Adding record failed: record already exists
added: 45.143.220.229/32 0

as expected

[dhw]

Looks believable; will test & report a bit later today.

[dhw]

OK; I took the opportunity during my daily update cycle to (hand-)make the above change for a laptop that attempts to populate 3 ipfw tables once it gets a DHCP lease. This has been working from late 2008 up through stable/14, but was recently seen to be failing for both stable/15 and head.
TL;DR: the change is effective.
Evidence:
First, head, showing failing condition:

g1-96(14.3-R-p3)[6] ssh -xt g1-48 'uname -aUK && sudo ipfw table 1 list | wc -l'
FreeBSD g1-48.catwhisker.org 16.0-CURRENT FreeBSD 16.0-CURRENT #460 main-n281070-1129bf441e99: Sun Oct 12 11:16:31 UTC 2025     [e-mail]:/common/S4/obj/usr/src/amd64.amd64/sys/CANARY amd64 1600001 1600001
Password:
       1
Connection to g1-48.catwhisker.org closed.

After update, table 1 is populated correctly:

g1-96(14.3-R-p3)[7] ssh -xt g1-48 'uname -aUK && sudo ipfw table 1 list | wc -l'
FreeBSD g1-48.catwhisker.org 16.0-CURRENT FreeBSD 16.0-CURRENT #461 main-n281113-c8e077e57b25-dirty: Mon Oct 13 11:29:28 UTC 2025     [e-mail]:/common/S4/obj/usr/src/amd64.amd64/sys/CANARY amd64 1600001 1600001
Password:
   15447
Connection to g1-48.catwhisker.org closed.

And for stable/15; before update:

g1-96(14.3-R-p3)[8] ssh -xt g1-48 'uname -aUK && sudo ipfw table 1 list | wc -l'
FreeBSD g1-48.catwhisker.org 15.0-STABLE FreeBSD 15.0-STABLE #455 stable/15-n280668-998dd4a11a2b: Sun Oct 12 12:46:36 UTC 2025     [e-mail]:/common/S3/obj/usr/src/amd64.amd64/sys/CANARY amd64 1500500 1500500
Password:
       1
Connection to g1-48.catwhisker.org closed.

And after:

g1-96(14.3-R-p3)[9] ssh -xt g1-48 'uname -aUK && sudo ipfw table 1 list | wc -l'
FreeBSD g1-48.catwhisker.org 15.0-STABLE FreeBSD 15.0-STABLE #456 stable/15-n280674-53fcc7f9b0de-dirty: Mon Oct 13 12:11:17 UTC 2025     [e-mail]:/common/S3/obj/usr/src/amd64.amd64/sys/CANARY amd64 1500500 1500500
Password:
   15447
Connection to g1-48.catwhisker.org closed.

I am not prepared to test to validate that there are no failures that could have been caught with the existing code that are not caught with the proposed change, but from the "poking around" I did while investigating the problem, I expect that nothing would turn up even if I were (& did).

Summary: Looks good to me.