For those of you who don't obsessively track cvs commits...
I submitted a patch tonight to clean up the remaining uses of suser() in libafs when compiled on Solaris 10 or better. In all cases except two, all calls to suser() have been replaced with calls to priv_policy (). This insulates libafs in case suser() goes away some day (currently in the Solaris 10 kernel, suser() has been demoted to a macro around priv_policy() anyway)
The two exceptions are in afs_mount() and afs_unmount() in src/afs/ SOLARIS/osi_vfsops.c. For Solaris <= 9, the calls to suser() to determine if the calling user has permission to mount/umount AFS are kept. On Solaris >= 10, secpolicy_fs_mount() and secpolicy_fs_unmount () are called instead. In a basic sense, these are transparent replacements for suser()... with one advantage - it's one more step in the direction of (hopefully) making the afs kernel module zone- friendly... some day.
Still tracking down any more uses of private functions or stucts/ struct members. Right now I'm starting with a dump from 'nm -u libafs.nonfs.o' and going down the list.
Solaris Nevada (aka Solaris 11) is now recognized. My patch to add sun4x_511 and sunx86_511 sysnames was accepted a few days ago. You can now compile OpenAFS cvs HEAD on OpenSolaris or a derivative thereof. It seems to work as of right now, but since Solaris Nevada is ("officially") still a moving target, this support should stay in the 1.5 tree until Solaris 11 is officially released.
/dale -- Dale Ghent UNIX Systems Specialist UMBC - Office of Information Technology ECS 201 - x51705 _______________________________________________ OpenAFS-devel mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-devel
