Note that -flto doesn't actually LTO system libraries, because we do not yet provide LLVM IR libraries. If you put an explicit_bzero implementation in your test code though you can get it to be eliminated. As it stands today for normal use (linking against libc as usual) it will not be eliminated.

See also Oracle image guidelines at https://docs.oracle.com/en/cloud/marketplace/partner-portal/ocmpd/how-to-publish-an-oci-image-listing.pdf

If you upload again for changes, please include full context e.g. git diff -U99999 .... or git show -U99999 ...

Can we include the command(s) used to generate the certs somewhere, maybe a comment in the test?

There are developers who want to run these tasks on their own forks - at least, I do. Users could also configure Cirrus-CI not to run on their fork, if so desired (i.e., by not giving Cirrus-CI access to the specific repo).

I like Ravi's first, but you can also drop "can be used to"
The following sysctl variables control how the kernel ...

Reported by: Thomas Hurst https://twitter.com/blaagh/status/1521857962297507842

Will change the comment to

# Restart sshd if running (PR263489).  Note that this does not
# affect child sshd processes handling existing sessions.

Dumb question, I'm sure, but doesn't restarting sshd kill current ssh sessions?

I changed the WITH_PROFILE description in f768ecf247b7acd98fd0ee395522e559769e3e51 as it may not be gone in 14.

Is it time to look at deprecation plans for firewire?

Regenerate after:

• eb45bc6829157b4bd8367a7fac96acc0e740f116
• 613b4b79713e294140757270f02a8aa6273be3d4
• 19780592633e50efca39454d1ecf029bd7d87868
• 835ee05f3c754d905099a3500f421dc01fab028f
• 6e24fe61287b02600be19042a6f6ce939aabfaa6

There seems to be no change here or in https://github.com/trixirt/libc11ext in the last five years. @trix_juniper.net are you still planning to pursue this?

Note to self: for upstreaming, see also D33934

LGTM except style issue, look at the SANITIZER_PPC64V2 case above, space after the 1 and line up the string indentation.

I'm reasonably confident that this patch has addressed the issue I observed.

there are still a few things controlled by WITH_/WITHOUT_GAMES:

$ git grep MK_GAMES
Makefile.inc1:.if ${MK_GAMES} != "no"
release/Makefile:           MK_GAMES=no \
tools/build/mk/OptionalObsoleteFiles.inc:.if ${MK_GAMES} == no
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   caesar
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   factor
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   fortune
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   grdc
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   morse
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   number
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   pom
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   primes
usr.bin/Makefile:SUBDIR.${MK_GAMES}+=   random

I've applied this to my local tree and so far it appears to address the livelock issue I was observed somewhat regularly under load (video/keyboard unresponsive, although mouse cursor still moved).

looks reasonable to me