Steven Bosscher <stevenb....@gmail.com> writes:
> On Wed, Mar 21, 2012 at 1:13 AM, Ian Lance Taylor wrote:
>> On Tue, Mar 20, 2012 at 2:06 PM, Steven Bosscher wrote:
>>> This patch splits a couple of pieces of cse_insn out to new functions.
>>> There are no functional changes, and no code generation differences as
>>> far as I could tell on x86_64 (-m64 and -m32).
> Likewise for the attached patch.
>>> The purpose of the patch is and, loto hopefully make cse_insn easier
>>> to understand. In a follow-up patch, I will make canonicalize_insn run
>>> only once per insn (it currently, i.e. before and after this patch,
>>> runs multiple times for CSE on extended basic blocks if a block is in
>>> multiple extended basic blocks).
> That is what the attached patch does.
> Bootstrapped&tested on x86_64-unknown-linux-gnu.
> OK for trunk?
> * cse.c (cse_canonicalized_basic_blocks): New simple bitmap to
> tag basic blocks that have already been traversed at least once,
> so that all insns have been canonicalized.
> (cse_insn): Call canonicalize_insn only if the basic block that
> contains insn is visited for the first time.
> (cse_extended_basic_block): After visiting all insns in a basic
> block, mark the block in cse_canonicalized_basic_blocks.
> (cse_main): Setup and destroy cse_canonicalized_basic_blocks.
OK, thanks (without the microoptimisation, as you say).
Out of curiosity, do you still see this bit as useful:
/* We potentially will process this insn many times. Therefore,
drop the REG_EQUAL note if it is equal to the SET_SRC of the
unique set in INSN.
Do not do so if the REG_EQUAL note is for a STRICT_LOW_PART,
because cse_insn handles those specially. */
? Does "many times" mean in CSE, or later?