Re: mm: ptl is not bloated if it fits in pointer

2013-12-16 Thread Hugh Dickins
On Mon, 16 Dec 2013, Kirill A. Shutemov wrote: > Peter Zijlstra wrote: > > On Mon, Dec 16, 2013 at 01:04:13AM -0800, Hugh Dickins wrote: > > > It's silly to force the 64-bit CONFIG_GENERIC_LOCKBREAK architectures > > > to kmalloc eight bytes for an indirect page table lock: the lock needs > > > to

Re: mm: ptl is not bloated if it fits in pointer

2013-12-16 Thread Kirill A. Shutemov
Peter Zijlstra wrote: > On Mon, Dec 16, 2013 at 01:04:13AM -0800, Hugh Dickins wrote: > > It's silly to force the 64-bit CONFIG_GENERIC_LOCKBREAK architectures > > to kmalloc eight bytes for an indirect page table lock: the lock needs > > to fit in the space that a pointer to it would occupy, not i

Re: mm: ptl is not bloated if it fits in pointer

2013-12-16 Thread Peter Zijlstra
On Mon, Dec 16, 2013 at 01:04:13AM -0800, Hugh Dickins wrote: > It's silly to force the 64-bit CONFIG_GENERIC_LOCKBREAK architectures So yes that's unfortunate, but why are people using that GENERIC_LOCKBREAK stuff to begin with? Its atrocious, a much better path would be to remove it. -- To unsub

Re: mm: ptl is not bloated if it fits in pointer

2013-12-16 Thread Peter Zijlstra
On Mon, Dec 16, 2013 at 01:04:13AM -0800, Hugh Dickins wrote: > It's silly to force the 64-bit CONFIG_GENERIC_LOCKBREAK architectures > to kmalloc eight bytes for an indirect page table lock: the lock needs > to fit in the space that a pointer to it would occupy, not into an int. Ah, no. A spinloc