On 20 March 2018 at 21:52, Dibyendu Majumdar <mobile@xxxxxxxxxxxxxxx> wrote: > On 20 March 2018 at 21:49, Luc Van Oostenryck > <luc.vanoostenryck@xxxxxxxxx> wrote: >> On Tue, Mar 20, 2018 at 09:07:47PM +0000, Dibyendu Majumdar wrote: >>> On 20 March 2018 at 00:52, Luc Van Oostenryck >>> <luc.vanoostenryck@xxxxxxxxx> wrote: >>> > This series replace the current SSA conversion by one using >>> > the 'classical' way, via the iterated dominance frontier. >>> > >>> >>> Thank you for posting this patch series. I would like to test it. I >>> will apply my LLVM fixes against your tree so that I can run the set >>> of tests I have prepared. This will take a bit of time but I will >>> report back here on results. >> I have done some initial work on this. My changes are in: https://github.com/dibyendumajumdar/sparse-dev/commits/ssa-v1. Some more work is needed as this is still preliminary. So far 65 out of 85 tests in (https://github.com/dibyendumajumdar/sparse-testing) pass. I haven't yet investigated the failures - they could be due to issues in my merge. In terms of my experience so far: a) The series has new opcodes for floating point values compared to the Sparse master tree. b) The solution for variadic calls is different compared to the value-size patch. I opted not to use the mechanism in this tree as I think the value-size patch is a better approach. Right now therefore this is probably broken. c) In dmrC the macros for var args are disabled in code gen mode - I need to do a similar fix here else any code using var args will crash. I will report further results as I test this more and investigate the test failures. Thanks and Regards Dibyendu -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
