Page MenuHomeFreeBSD

FreeBSD base system ssh 9.0p1 diff against upstream
Needs ReviewPublic

Authored by emaste on Sep 2 2021, 3:57 PM.
Tags
None
Referenced Files
Unknown Object (File)
Thu, Nov 21, 12:40 AM
Unknown Object (File)
Fri, Nov 8, 9:42 AM
Unknown Object (File)
Oct 16 2024, 1:41 PM
Unknown Object (File)
Oct 16 2024, 1:41 PM
Unknown Object (File)
Oct 16 2024, 1:40 PM
Unknown Object (File)
Oct 16 2024, 1:40 PM
Unknown Object (File)
Oct 16 2024, 1:39 PM
Unknown Object (File)
Oct 16 2024, 1:18 PM
This revision needs review, but there are no reviewers specified.

Details

Reviewers
None
Summary

This is a diff of the FreeBSD base system ssh (crypto/openssh) against OpenSSH 9.0p1, for reference.

Generated via git diff -U9999 --diff-filter=M vendor/openssh/9.0p1 HEAD:crypto/openssh

Diff Detail

Lint
Lint Skipped
Unit
Tests Skipped

Event Timeline

emaste requested review of this revision.Sep 2 2021, 3:57 PM
kevans added inline comments.
crypto/openssh/pathnames.h
123 ↗(On Diff #94570)

These probably should have been converted to _PATH_LOCALBASE as well for consistency if we're already patching them.

crypto/openssh/pathnames.h
123 ↗(On Diff #94570)

Probably, although we'd have to #include <paths.h> in pathnames.h or in all of its consumers. Such a change won't make the update any easier (or harder), so it may be something to revisit after the update is done.

emaste retitled this revision from FreeBSD base system ssh 7.9p1 diff against upstream to FreeBSD base system ssh 9.0p1 diff against upstream.
emaste edited the summary of this revision. (Show Details)

Update to 9.0p1 diff (i.e., what is in main as of today)

Generated via git diff -U9999 --diff-filter=M vendor/openssh/9.0p1 HEAD:crypto/openssh

channels.c
2604–2610

this is a fix from the V_9_0 branch upstream

sandbox-capsicum.c
72–75

I've sent this to the openssh-unix-dev mailing list

scp.1
298

I've sent this to the openssh-unix-dev mailing list

auth.c
27

Can we stop including these now?

session.c
1017

Can this be moved down to just above the setusercontext() bit? I believe it's here since setusercontext() can override TZ.

1040

Hmm, I wonder why we set TERM here. It can be overridden below.

1295

I'd think this could be upstreamed, it'd be ok even on platforms where login_getcapstr() returns a char * rather than const char *. Maybe there are weird platforms where stat() and fopen() don't specify that the path is const, but I see other uses of those functions with const strings.

auth.c
27

I was planning on dropping all of them after FreeBSD 12 is past EOL, but I think we could keep ones in config files for now and drop the rest.

session.c
1295

This is a leftover from upstreaming: https://lists.mindrot.org/pipermail/openssh-unix-dev/2021-February/039129.html - Damien took the other part but not this.

Even though const is better we might as well reduce diffs against upstream and drop it.

session.c
1295

I presume it's needed to avoid a compile error? Our login_getcapstr() returns a const char *, OpenBSD's returns a char *.

1295

... but presumably it could be declared const on OpenBSD too, now that the free() call is gone.

session.c
1040

Indeed. This comes from ca86bcf2531c7