[Bug rtl-optimization/26244] [4.2 Regression] FAIL: gcc.c-torture/execute/builtin-bitops-1.c execution, -O3 -fomit-frame-pointer -funroll-loops

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #21 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Subject: Re:  [4.2 Regression] FAIL: gcc.c-torture/execute/builtin-bitops-1.c
execution,  -O3 -fomit-frame-pointer -funroll-loops

> I looked at the documentation for lshiftrt, but it doesn't say
> whether the shift amount range handled by the PA insn for variable
> shifts is ok or not.  On the otherhand, there's no obvious reason
> why we should need to support logical right shifts larger than 31
> bits and I have to wonder why the generated code does this.

I looked at this a bit more and I believe that this behavior is normal
when SHIFT_COUNT_TRUNCATED is defined.

The test doesn't fail with -fno-unroll-loops-all-loops.  It appears
to me that this is a loop unrolling rather than a loop invariant bug,
although I must admit it's not clear where things actually go wrong
because of the complexity of this operation.

I modified the testcase a bit to make debugging easier.  I've attached
the modified testcase and the dumps from loop2_unroll, loop2_done, web
and cse2.  The initialization of "count" in the block after code_label 21
definitely appears wrong after cse2.

I've also attached the final assembly code for reference.

Dave


------- Comment #22 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Created an attachment (id=11748)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11748&action=view)


------- Comment #23 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Created an attachment (id=11749)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11749&action=view)


------- Comment #24 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Created an attachment (id=11750)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11750&action=view)


------- Comment #25 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Created an attachment (id=11751)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11751&action=view)


------- Comment #26 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Created an attachment (id=11752)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11752&action=view)


------- Comment #27 from dave at hiauly1 dot hia dot nrc dot ca  2006-06-25 15:32 -------
Created an attachment (id=11753)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11753&action=view)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26244