26 May
2006
26 May
'06
5:44 a.m.
On Thu, 25 May 2006, Lee Revell wrote:
On Thu, 2006-05-25 at 19:57 +0300, Kai Vehmanen
wrote:
Not necessarily much, but I'd like to understand this fully. There might
be other archs affected as well. Again looking at the Linux kernel code -
for instance armv6 atomic_set needs spinning (but not for armv4)... so
it's not just sparc32. There might not be _that_ many sparc32s out
there, but there are over 3 billion arm cores out there. ;)
But, but, even on these platforms, it's just the writes. It'd be
interesting to know if there are any cases where you could be reading
bogus values due to concurrent access to an 'int'.
--
links, my public keys, etc at http://eca.cx/kv
Does someone have a good reference on this? I
think the writes just
Do we still care about 32 bit sparc?