Change Details

There are 2 changes in terms of behavior: - mp == NULL is no longer supported. The only place in the tree which calls getnewvnode with mp == NULL does it for vp_crossmp which will never execute this codepath. Any vnode which legally has ->v_mount == NULL is also doomed, which once more wont execute for this code. - VI_ACTIVE is required. It's an invariant that a non-doomed vnode being vdropped has it. Pure refactoring: - move freeing of the vnode away to a routine doing the reverse of getnewvnode. Perhaps both should use uma ctor/dtor instead. - docouple vdrop and vdropl - move returning the vnode to the free list to a separate routine Note this removes an assertion for v_holdcnt from production kernels. If it seems desirable I can put it back to both routines.

There are 2 changes in terms of behavior: - mp == NULL is no longer supported. The only place in the tree which calls getnewvnode with mp == NULL does it for vp_crossmp which will never execute this codepath. Any vnode which legally has ->v_mount == NULL is also doomed, which once more wont execute this code. - VI_ACTIVE is required. It's an invariant that a non-doomed vnode being vdropped has it. Pure refactoring: - move freeing of the vnode away to a routine doing the reverse of getnewvnode. Perhaps both should use uma ctor/dtor instead. - docouple vdrop and vdropl - move returning the vnode to the free list to a separate routine Note this removes an assertion for v_holdcnt from production kernels. If it seems desirable I can put it back to both routines.

There are 2 changes in terms of behavior: - mp == NULL is no longer supported. The only place in the tree which calls getnewvnode with mp == NULL does it for vp_crossmp which will never execute this codepath. Any vnode which legally has ->v_mount == NULL is also doomed, which once more wont execute for this code. - VI_ACTIVE is required. It's an invariant that a non-doomed vnode being vdropped has it. Pure refactoring: - move freeing of the vnode away to a routine doing the reverse of getnewvnode. Perhaps both should use uma ctor/dtor instead. - docouple vdrop and vdropl - move returning the vnode to the free list to a separate routine Note this removes an assertion for v_holdcnt from production kernels. If it seems desirable I can put it back to both routines.