Messages in this thread | | | Date | Fri, 20 Oct 2006 12:54:17 -0700 (PDT) | From | Linus Torvalds <> | Subject | Re: [PATCH 1/3] Fix COW D-cache aliasing on fork |
| |
On Fri, 20 Oct 2006, David Miller wrote: > > You get an asynchronous fault from the L2 cache, and that's also what > happens when the TLB entry is missing during L2 writeback too. You > get a level 15 non-maskable IRQ when these asynchronous errors happen.
Well, sparc always was crud. I can see the missing tlb entry, but if it's been turned read-only, the write-back should still work (it clearly _was_ writable when the write that dirtied the cacheline happened).
Anyway, if you cannot flush a read-only mapping, then the "flush at COW fault time" won't work _either_, since the original mapping is still read-only.
So regardless, the COW-time flush cannot work. But the "flush before the TLB flush, and then flush after in case we had a race" approach should work as well as it can in practice, no?
Linus - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |