Certainly not, you will not catch a memory _parity_ error by overwriting
current memory contents, especially if it has occurred due to a very rare
event in your machine, (ECC may be different), but by reading it again.
This actually makes the algorithm non destructive! Then you can try to
put this page on the bad memory list and to recover. Don't forget however
that the bad data may have been copied into an L2 and/or L1 cache line.
Perhaps test the whole memory twice just in case...
> > + * (5) test if any NMI is pending: if yes, the
> > + * last page written is bogus, printk its
> > + * address.
> > + * (6) ++ page
> > + * (7) panic() out: we have no more things other that
> > + * raw kernel, running on this machine now.
> > + *
> > + * In (4) we should care not to overwrite the kernel
> > + * because I suspect we need it at least for printk()
> > + * and panic()
> > + */
Gabriel