On Tue, May 20, 2025 at 09:55:04PM GMT, Xi Ruoyao wrote: > On Tue, 2025-05-20 at 15:44 +0200, Jakub Jelinek wrote: > > > Specifically, the tests told me to extend (thought "truncate" > > > was kind of an equivalent word) the output of left shift, plus/minus, > > > > Truncation is the exact opposite of extension. > > I can understand the need for handling of left shifts, for all the rest > > I'd really like to see testcases. > > I guess the terminology thing is caused by the past experience of the > Loongson team with "a famous !TARGET_TRULY_NOOP_TRUNCATION target." On > that target truncsidi2 is a sign-extension as required by the ISA spec. > > I'm trying to fix an ext-dce bug regarding !TARGET_TRULY_NOOP_TRUNCATION > so I just decided to chime in and explain this :). > > -- > Xi Ruoyao <xry...@xry111.site> > School of Aerospace Science and Technology, Xidian University
Hi Ruoyao, Thank you for this nice comment. I thought this wording was not really a big deal. I used "truncation" because in the code of gimple-lower-bitint.cc, truncation or down-casting is the first actual operation that needs to be done when "info.extended" targets "extends" the output of some operations, which are already in m_limb_type (mostly 64-bit integers). This truncation exists for the "non-extended" targets, and seems also necessary for the "extended" targets in some operations. Again, it would be great if there is a more elegant way to get this done. Yujie
