> On Tue, 22 May 2001, Richard Earnshaw wrote:
> 
> > > What is the effect of an instruction here?
> >
> > The processor can prefetch-abort when it shouldn't.
> 
> Can that not be worked around in the prefetch-abort handler? (I assume
> not, but I'm interested to know why)
> 
I think it might be possible but I haven't figured out how yet. The last time I 
had a problem with this bug was when it hit the cc1plus binary (gcc's c++ 
compiler). I saw a massive storm of page faults (when the bug hit) but 
eventually this stopped and the compile completed as expected. Patching 
cc1plus prevented these page fault storms so I presume they were caused by the 
SA bug.

Since the binary recovered and carried on there must be something that 
can be done to prevent the repeated page faults and allow the program to 
continue. The next time I get a binary which faults in this way I will try and 
figure out a reliable fix.

Mike