src.opts.mk: enable OPENSSL_KTLS by default on riscv64

[siva]

Summary

1e649491b8567151270095fda3bce8faea394952 enabled KERN_TLS in
riscv/conf/GENERIC, but didn't enable OPENSSL_KTLS.

This passes all testcases in the sys/kern/ssl_sendfile suite and
fixes CI failures seen here:
https://ci.freebsd.org/job/FreeBSD-main-riscv64-test/16606/testReport/sys.kern/ssl_sendfile/

PR: 293810
Fixes: 1e649491b8567151270095fda3bce8faea394952

[ngie]

This "fixes" the issue, but the tests should also be intelligent enough to understand when the OPENSSL_KTLS feature isn't present and skip the tests if the feature isn't present in the kernel (see ATF_REQUIRE_FEATURE, etc).

[des]

In D57316#1313745, @ngie wrote:

This "fixes" the issue, but the tests should also be intelligent enough to understand when the OPENSSL_KTLS feature isn't present and skip the tests if the feature isn't present in the kernel (see ATF_REQUIRE_FEATURE, etc).

ATF_REQUIRE_FEATURE is awfully unidiomatic, we should implement a require.features property instead.

[ngie]

In D57316#1314576, @des wrote:

In D57316#1313745, @ngie wrote:

This "fixes" the issue, but the tests should also be intelligent enough to understand when the OPENSSL_KTLS feature isn't present and skip the tests if the feature isn't present in the kernel (see ATF_REQUIRE_FEATURE, etc).

ATF_REQUIRE_FEATURE is awfully unidiomatic, we should implement a require.features property instead.

I agree.. just not sure how to handle it yet with other platforms where this feature might not be available. Probably should just push forward with the OS-specific handler behavior in kyua for FreeBSD started by @igoro .

[igoro]

In D57316#1315185, @ngie wrote:

In D57316#1314576, @des wrote:

In D57316#1313745, @ngie wrote:

This "fixes" the issue, but the tests should also be intelligent enough to understand when the OPENSSL_KTLS feature isn't present and skip the tests if the feature isn't present in the kernel (see ATF_REQUIRE_FEATURE, etc).

ATF_REQUIRE_FEATURE is awfully unidiomatic, we should implement a require.features property instead.

I agree.. just not sure how to handle it yet with other platforms where this feature might not be available. Probably should just push forward with the OS-specific handler behavior in kyua for FreeBSD started by @igoro .

Yes, it could be implemented that way. The patch is designed in a way that allows new handlers to be introduced later: https://reviews.freebsd.org/D48087. By the way, it has been ready for landing for quite some time. @ngie , it just waits for your formal approval. All the comments have been addressed there.

[ziaee]

share/man/man5/src.conf.5

1517

why does it say i386 twice?

[imp]

share/man/man5/src.conf.5

1517

because this is MACHINE/MACINE_ARCH notation, just like everywhere else.