[Bug target/69231] New: rtx_cost of subreg is bad for <= wordmode integral subregs of > wordmode integral pseudos

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69231

            Bug ID: 69231
           Summary: rtx_cost of subreg is bad for <= wordmode integral
                    subregs of > wordmode integral pseudos
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: jakub at gcc dot gnu.org
                CC: bernds at gcc dot gnu.org, hjl.tools at gmail dot com,
                    jakub at gcc dot gnu.org, ktkachov at gcc dot gnu.org,
                    law at gcc dot gnu.org, nheghathivhistha at gmail dot com,
                    unassigned at gcc dot gnu.org
        Depends on: 67462
  Target Milestone: ---

+++ This bug was initially created as a clone of Bug #67462 +++

As mentioned in the PR67462, rtx_cost for SUBREG uses !MODES_TIEABLE_P and in
that case uses fairly high cost.  But, for say SImode subreg of DImode if the
pseudo is likely to end up in a GPR on i?86, the actual cost is 0 (the same as
of say SImode REG).  So we should for GCC 7 consider changing this.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67462
[Bug 67462] [6 Regression] FAIL: gcc.dg/ifcvt-3.c scan-rtl-dump ce1 "3 true
changes made"