- User Since
- Oct 2 2015, 1:17 PM (177 w, 13 h)
Tagging kib and jamie to start with, perhaps.
Thu, Feb 21
As we discussed, I no longer have any qualms with dropping jail(3) completely and proxying it all through to jail(8).
Wed, Feb 20
Backpedal, I swapped the -n clauses.
Pass through DESTDIR to certctl, still needs testing
- Gate certctl build/installation on MK_CAROOT && MK_OPENSSL
- Sprinkle some DESTDIR around
- Drop initial certs from this review
- Sprinkled some .PHONY
- mtree update
- Rewrote secure/caroot/trusted/Makefile to just shell out and grab all certs in the current dir instead of makefile churn*
Tue, Feb 19
Mon, Feb 18
Sun, Feb 17
Sat, Feb 16
Fri, Feb 15
Wed, Feb 13
Tue, Feb 12
Any progress/hangups on this? =)
Mon, Feb 11
Sun, Feb 10
Wed, Feb 6
Tue, Feb 5
Sorry, that should have read "just a tad too fat and don't quite implement all of the needed bits"
Sorry, I think I missed some context- exactly what problem are we trying to address here?
Sun, Feb 3
Sat, Feb 2
(Ignore that "KERNELFLAGS" comment -- that was referring to RFKERNELONLY, which for some reason I had thought was used somewhere else that might care... nope)
Fri, Feb 1
New approach, because it looks a lot nicer- add an RFSPAWN flag for rfork(2) that uses vfork(2) semantics but doesn't inherit signal actions.
Thu, Jan 31
LGTM, thanks! =)
Wed, Jan 30
I was attempting to leave the original case as unmolested as possible (hence also the macros -- shims so we don't print things out in the test), but I don't care enough to object.
Address some of the style(9) concerns (bah =)):
Tue, Jan 29
Abandoning in favor of D19019 to start setting up zfsbootcfg vars in loaders properly; I think zfsbootcfg should likely bail out if machdep.bootmethod is unset or != "BIOS" until we teach other loaders pad2 bits, though.
This needs a little bit more consideration -- it would fail miserably if operating on a system that doesn't have a traditional BE setup or the dataset mounted at / isn't at least two levels deep.
Mon, Jan 28
Sun, Jan 27
I went ahead and took the quite literal approach to turning Andrew's test case into an atf test and integrating it into stdlib tests: https://people.freebsd.org/~kevans/dynthr-test.diff -- I guess we could cut out 90% of it, but I'm not sure how much we care to.