aboutsummaryrefslogtreecommitdiff
Commit message (Expand)AuthorAgeFilesLines
* Remove phi source merging.Linus Torvalds2005-04-071-2/+0
* Remove OP_SETVAL after symbol-pseudo simplification.Linus Torvalds2005-04-071-1/+1
* Fix undefined symbol linearization caseLinus Torvalds2005-04-071-1/+1
* Expose "show_bb()" for debugging, and make it do more appropriateLinus Torvalds2005-04-071-1/+1
* CSE may have caused more unreachable bb's, so do a secondLinus Torvalds2005-04-071-0/+1
* Make OP_PHISOURCE track the OP_PHI instructions that it defines.Linus Torvalds2005-04-071-4/+9
* Don't output code for static/toplevel symbols.Linus Torvalds2005-04-071-0/+4
* Linearize anonymous symbol initializers when we encounter them,Linus Torvalds2005-04-071-11/+9
* Clean up debug output from warning.Linus Torvalds2005-04-071-1/+1
* Fix silly asm input pseudo linearization.Linus Torvalds2005-04-071-1/+1
* Linearize inline asm statementsLinus Torvalds2005-04-071-3/+81
* Fix flow: we only remove one parent at a time. The same blockLinus Torvalds2005-04-071-1/+1
* Make the "entrypoint" be a special OP_ENTRY instruction instead ofLinus Torvalds2005-04-071-7/+15
* Expose "show_pseudo()" to the world.Linus Torvalds2005-04-071-1/+1
* Track argument pseudo lifetimes too.Linus Torvalds2005-04-071-2/+2
* Add pseudo death-note tracking.Linus Torvalds2005-04-071-0/+7
* Be more graceful about missing types and malformed expressions.Linus Torvalds2005-04-071-4/+6
* Be more careful with symbol sizes - don't SIGSEGV on undefined stuff.Linus Torvalds2005-04-071-12/+17
* Remove OP_SETCC, make OP_SEL bigger instead.Linus Torvalds2005-04-071-30/+21
* Fix embarrassing linearized bug with empty iterator post-conditions.Linus Torvalds2005-04-071-9/+6
* Duh! Remove a very very incorrect left-over flow simplification call.Linus Torvalds2005-04-071-1/+0
* Do real flow simplification only after liveness analysis.Linus Torvalds2005-04-071-2/+15
* Simplify trivial casts (and handle pointers specially).Linus Torvalds2005-04-071-1/+25
* Be more forgiving about missing types in linearization.Linus Torvalds2005-04-071-12/+22
* Rename "register.c" into "liveness.c". That's what it does.Linus Torvalds2005-04-071-1/+1
* Associate pseudos with the symbol name whose value they got.Linus Torvalds2005-04-071-3/+15
* Allow multiple levels of verbosity, and print out the _really_Linus Torvalds2005-04-071-21/+24
* Start tracking cross-basic-block pseudo usage.Linus Torvalds2005-04-071-1/+18
* Start using instruction sizes properly.Linus Torvalds2005-04-071-197/+212
* When replacing OP_SWITCH with OP_BR, make sure to removeLinus Torvalds2005-04-071-1/+3
* Add entrypoint pointer to each bb.Linus Torvalds2005-04-071-18/+19
* Add a flow verification thing.Linus Torvalds2005-04-071-0/+2
* Allow CSE to run after bb packing. Linus Torvalds2005-04-071-6/+5
* Make list-ptr remove/replace take a count.Linus Torvalds2005-04-071-1/+1
* Don't add the fake parent to the entryblock - we don't need it.Linus Torvalds2005-04-071-3/+0
* Make linearizer use the proper inc/dec operation value.Linus Torvalds2005-04-071-1/+1
* Add a semblance of sanity to structure member accesses.Linus Torvalds2005-04-071-50/+37
* Kill long-dead pseudo-reuse code.Linus Torvalds2005-04-071-4/+1
* Add a final pseudo usage tracking phase, which keepsLinus Torvalds2005-04-071-0/+3
* Fix pseudo->def on OP_PHI->OP_SEL conversion.Linus Torvalds2005-04-071-1/+7
* Clean up "linearize_symbol()" by making the function linearizerLinus Torvalds2005-04-071-59/+65
* Make the CSE "repeat" logic be more fine-grained than justLinus Torvalds2005-04-071-4/+1
* Clean up OP_STORE kill, and remove the use of the data.Linus Torvalds2005-04-071-0/+12
* More undefined symbol handling - don't oops on printout.Linus Torvalds2005-04-071-0/+4
* Change linearizer debug output syntax for call instructions.Linus Torvalds2005-04-071-1/+3
* Handle undefined symbols a bit more gracefully.Linus Torvalds2005-04-071-1/+5
* Do early CSE before even doing the symbol simplification.Linus Torvalds2005-04-071-6/+15
* Use cleaned-up ptr list removal for removing basic blocksLinus Torvalds2005-04-071-9/+1
* Be a lot more proper about rewriting end branches.Linus Torvalds2005-04-071-11/+26
* Simplify constant "conditional" branches and OP_SETCC/OP_SEL instructions.Linus Torvalds2005-04-071-3/+9