Hi! On Fri, 9 Dec 2016 14:08:21 +0100, Martin Jambor <mjam...@suse.cz> wrote: > this is the promised attempt at splitting omp-low.c [...]
Yay! \o/ I have not yet had a chance to review/test this patch, but I plan to. A few initial comments from the "bike shed departement"; I understand in GCC sources it will not be easy to rename stuff (such as files) later, so we should get the names agreed upon early: Generally, I agree with your division of "omp-low.c" parts. > - move everything that is part of pass_oacc_device_lower, > pass_omp_device_lower and pass_omp_target_link to a new file > omp-device.h, Should we call this file "omp-offload.c", as offloading is what this deals with, is the term we agreed to generally use (as far as I can tell)? > - move all pre-lowering gridification stuff to a new file > omp-grid.c. [...] Is that code generic enough to not call this file "omp-hsa.c" or similar? > - I moved stuff that was used from all over the place to a new file > omp-general.c (unless it would mean exposing omp_region or > omp_context types). I'd have called that simply "omp.c". > I am opened to suggestions what to do differently, names of the file > are for example of course subject to discussion, and I absolutely > welcome any review and checking, for one I am not going to pretend I > understand the stuff I put into omp-device.c. If however there is > consensus that we should do something like this, I would like to ask > the community to freeze omp-low.c file until this gets committed, I > hope you understand that I am afraid of any conflicts. I very much understand... :-| When I had worked on the very same thing months ago, and my changes went without review/approval for a long time, I had spent numerous hours on keeping my patch up to date. So, I'm happy to see that this is now near approval! (Even though I don't understand what's different now from when I worked on the same thing back then...) I hope to have time later today to review/test your actual patch. Grüße Thomas