Looks good. Copyrights, and all changes to zfs_destroy look sane. Looking at the gate it looks like be_create and be_snapshot are the only two files calling a zfs_destroy function.
I think the reason zfs_destroy still "works" is due to it being a parameter on the end of the function which is uninitialized memory. Thank you, Clay > -------- Original Message -------- > Subject: [caiman-discuss] code review for 10809 > Date: Sat, 22 Aug 2009 17:55:05 -0600 > From: Evan Layton <Evan.Layton at Sun.COM> > To: caiman-discuss <caiman-discuss at opensolaris.org> > > I need a code review for bug 10809. > bug: > http://defect.opensolaris.org/bz/show_bug.cgi?id=10809 > > webrev: > http://cr.opensolaris.org/~evanl/10809/ > > > I originally built the fix for 10807 on my machine that had been updated to > build 121, however the changes to libzfs.h that included the addition of a > boolean to the zfs_destroy parameters where not present on this machine. This > appears to be a problem with my test machine since subsequent updates of > other > machines did not show this difference. I can't at this point explain how this > happened but it appears that the change to did make it into build 121 at some > point. While this is causing the builds on indiana_build to fail it does not > stop beadm/libbe from working on a system that has just the fix for 10807 on > it. > I'm not sure at this point why this works since it's missing the boolean > parameter for zfs_destroy. > > I've added the fix for 10809 to all the places where we call zfs_destroy(). > > For testing I've recompiled on a build machine freshly updated to build 121 > and > run not only several manual beadm destroys but the libbe automated test > suite. > > thanks, > -evan > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/caiman-discuss >
