"H.J. Lu" <hjl.to...@gmail.com> writes: > On Mon, Aug 29, 2011 at 12:44 PM, Iyer, Balaji V > <balaji.v.i...@intel.com> wrote: >> Hello Everyone, >> >> Here is a link for a new patch (http://software.intel.com/file/38290). This >> patch is for the "cilkplus" branch and includes the following modifications: >> >> 1) Merges changes in the gcc master (SHA1: >> f326eb816922bc183133c09b25564d550ab9a282). >> 2) Adds functionality to allow _Cilk_spawn inside constructors and >> destructors. >> 3) The original code was storing the cilk_for scope as a tree_chain. This >> patch modifies it so that it is stored in the same location as FOR_SCOPE(). >> 4) The presence of a label inside a cilk_for loop was causing an ICE. This >> patch fixes that. >> 5) Spawning a spawned function also was causing an ICE. This patch fixes >> that also. >> 6) Adds the "__cilk" macro and sets it to 200. >> 7) Adds test cases for all the fixes mentioned above. >> 8) Fixes a bug in the libcilkrts runtime during initialization when the >> number of workers is greater than 3x the number of cores on the system. >> > > I suggest you use separate steps to merge with trunk and fix the > bugs. Their order isn't important. You should send the bug fixes > directly to gcc-patches mailing list.
Also it would be good if you cleaned up your original patch a bit. I noticed that it has a lot of white space changes to unrelated code, which makes it hard to figure out what it actually changes. -Andi
