[Bug target/45701] [4.6 Regression] Fail to prefer using r3 for padding a push/pop multiple to 8-byte alignment

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45701

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-12-21 12:30:25 UTC ---
Created attachment 22835
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22835
gcc46-pr45701.patch

Remembering pointers to tail call insns is both wasteful (allocates stuff that
is never used anywhere but on arm), error-prone (nothing adjusts the list if
tail calls are deleted, or duplicated, etc.) and unnecessary, because
the list is trivially available from the IL.

Here is completely untested patch (tested just on the 3 testcases with a
cross).
Could anyone interested in ARM target please bootstrap/regtest this?
We have way too many open P1 bugs.