Yes my bad, it is not called SequentiallyConsistentStore but AcquireStore :)
On Thu, Jan 23, 2014 at 11:57 AM, Peter Burka <[email protected]> wrote: > Hi Stephen, > > Honestly I didn't think about it too deeply. I simply copied the same > pattern that the x86-64 SSE2 code is using. If a second barrier is > required, then I think it would be required in that path, too, and would > represent a more general problem in the code. > > Since this function is called Acquire_Store, I think that the single > barrier is sufficient and that it's correctly placed. As an acquire, we > want to make sure that no reads or writes from after the store actually > happen before the store. > > There is complementary function called Release_Store which has a barrier > before its store. > > /peter > > On Thursday, January 23, 2014 11:33:44 AM UTC-5, stephentu wrote: > >> I think you might actually want two memory barriers, something like >> >> MemoryBarrier(); >> *ptr = value; >> MemoryBarrier(); >> >> your current solution prevents compiler reordering the store with >> statements after, but what about statements before? >> >> >> On Thu, Jan 23, 2014 at 11:20 AM, <[email protected]> wrote: >> >>> >>> Comment #1 on issue 599 by [email protected]: Fails to compile for Intel >>> Phi >>> http://code.google.com/p/protobuf/issues/detail?id=599 >>> >>> Thanks Stephen. ICC supports GCC's inline assembly syntax. I've applied >>> your suggestion and tested it. An updated patch is attached. >>> >>> /peter >>> >>> >>> Attachments: >>> atomicops_internals_x86_gcc.h.patch 658 bytes >>> >>> >>> -- >>> You received this message because this project is configured to send all >>> issue notifications to this address. >>> You may adjust your notification preferences at: >>> https://code.google.com/hosting/settings >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Protocol Buffers" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To post to this group, send email to [email protected]. >>> >>> Visit this group at http://groups.google.com/group/protobuf. >>> For more options, visit https://groups.google.com/groups/opt_out. >>> >> >> -- > You received this message because you are subscribed to the Google Groups > "Protocol Buffers" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/protobuf. > For more options, visit https://groups.google.com/groups/opt_out. > -- You received this message because you are subscribed to the Google Groups "Protocol Buffers" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/protobuf. For more options, visit https://groups.google.com/groups/opt_out.
