Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | llvm-mc: Match a few X86 instructions. | Daniel Dunbar | 2009-07-31 | 1 | -8/+71 |
| | | | | | | | | | | | - This is "experimental" code, I am feeling my way around and working out the best way to do things (and learning tblgen in the process). Comments welcome, but keep in mind this stuff will change radically. - This is enough to match "subb" and friends, but not much else. The next step is to automatically generate the matchers for individual operands. llvm-svn: 77657 | ||||
* | Manage MachineFunctions with an analysis Pass instead of the Annotable | Dan Gohman | 2009-07-31 | 11 | -104/+119 |
| | | | | | | | mechanism. To support this, make MachineFunctionPass a little more complete. llvm-svn: 77654 | ||||
* | Fix cmake build. | Benjamin Kramer | 2009-07-31 | 1 | -1/+0 |
| | | | | llvm-svn: 77649 | ||||
* | Add getOrInsertNamedMetadata(). | Devang Patel | 2009-07-30 | 1 | -3/+14 |
| | | | | llvm-svn: 77646 | ||||
* | When fp is not eliminated, instructions with T2_i12 modes will be changed to ↵ | Evan Cheng | 2009-07-30 | 2 | -3/+10 |
| | | | | | | T2_i8 ones. Take that into consideration when determining stack size limit for reserving register scavenging slot. llvm-svn: 77642 | ||||
* | Removed the BigBlock register allocator. | Lang Hames | 2009-07-30 | 1 | -892/+0 |
| | | | | llvm-svn: 77640 | ||||
* | Do not use abbrev while writing NamedMDNode name. | Devang Patel | 2009-07-30 | 1 | -9/+4 |
| | | | | llvm-svn: 77637 | ||||
* | Enumerate NamedMDNode elements first. | Devang Patel | 2009-07-30 | 1 | -4/+0 |
| | | | | llvm-svn: 77636 | ||||
* | Move more code back to 2.5 APIs. | Owen Anderson | 2009-07-30 | 41 | -221/+180 |
| | | | | llvm-svn: 77635 | ||||
* | Remove redundant match for frame index from imm8 addrmode, it is handled by ↵ | David Goodwin | 2009-07-30 | 1 | -24/+14 |
| | | | | | | the imm12 addrmode. llvm-svn: 77632 | ||||
* | Darwin assembler now recognizes "orn", so remove workaround. | David Goodwin | 2009-07-30 | 1 | -5/+3 |
| | | | | llvm-svn: 77627 | ||||
* | Darwin assembler now supports "rrx", so remove workaround. | David Goodwin | 2009-07-30 | 1 | -2/+1 |
| | | | | llvm-svn: 77625 | ||||
* | Twine: Directly support int, long, and long long types. | Daniel Dunbar | 2009-07-30 | 1 | -17/+29 |
| | | | | | | - This should resolve Cygwin gcc ambiguities. llvm-svn: 77624 | ||||
* | Use CallbackVH in AliasSetTracker to avoid getting stuck with | Dan Gohman | 2009-07-30 | 1 | -4/+20 |
| | | | | | | dangling Value*s. llvm-svn: 77623 | ||||
* | Cleanup and include code selection for some frame index cases. | David Goodwin | 2009-07-30 | 1 | -20/+49 |
| | | | | llvm-svn: 77622 | ||||
* | Start using DebugInfoFinder. | Devang Patel | 2009-07-30 | 1 | -13/+11 |
| | | | | llvm-svn: 77621 | ||||
* | Do not use getNamedValue() to lookup NamedMDNode. NamedMDNode is not a ↵ | Devang Patel | 2009-07-30 | 1 | -1/+1 |
| | | | | | | | | GlobalValue. Thanks Benjamin Kramer! llvm-svn: 77619 | ||||
* | Twine: Use raw_ostream::write_hex, remove unused itohexstr method. | Daniel Dunbar | 2009-07-30 | 1 | -2/+1 |
| | | | | llvm-svn: 77617 | ||||
* | s/DebugInfoEnumerator/DebugInfoFinder/g | Devang Patel | 2009-07-30 | 1 | -31/+31 |
| | | | | llvm-svn: 77615 | ||||
* | Add raw_ostream::write_hex | Daniel Dunbar | 2009-07-30 | 1 | -4/+7 |
| | | | | llvm-svn: 77614 | ||||
* | Add missing D* register clobbers for Thumb-2 call. | David Goodwin | 2009-07-30 | 1 | -0/+1 |
| | | | | llvm-svn: 77611 | ||||
* | Twines: Don't allow implicit conversion from integers, this is too tricky. | Daniel Dunbar | 2009-07-30 | 4 | -13/+13 |
| | | | | llvm-svn: 77605 | ||||
* | walk DbgRegionStartInst and DbgRegionEndInst | Devang Patel | 2009-07-30 | 1 | -2/+20 |
| | | | | llvm-svn: 77604 | ||||
* | Minor whitespace tidiness. | Dan Gohman | 2009-07-30 | 1 | -3/+0 |
| | | | | llvm-svn: 77602 | ||||
* | Rename GRAD to GR32_AD, to follow the naming convention of other | Dan Gohman | 2009-07-30 | 2 | -2/+4 |
| | | | | | | classes. And define its SubRegClassList. llvm-svn: 77601 | ||||
* | add a random codegen deficiency. | Chris Lattner | 2009-07-30 | 1 | -0/+46 |
| | | | | llvm-svn: 77598 | ||||
* | fix a unitialized pointer in NamedMDNode (and reenable unittest) | Benjamin Kramer | 2009-07-30 | 1 | -1/+1 |
| | | | | llvm-svn: 77597 | ||||
* | Allow targets to define libcall names for mem(cpy,set,move) intrinsics, ↵ | Sanjiv Gupta | 2009-07-30 | 2 | -3/+9 |
| | | | | | | rather than hardcoding them in DAG lowering. llvm-svn: 77586 | ||||
* | Add a note. | Evan Cheng | 2009-07-30 | 1 | -0/+20 |
| | | | | llvm-svn: 77584 | ||||
* | Optimize some common usage patterns of atomic built-ins ↵ | Evan Cheng | 2009-07-30 | 6 | -14/+281 |
| | | | | | | | | | | | | __sync_add_and_fetch() and __sync_sub_and_fetch. When the return value is not used (i.e. only care about the value in the memory), x86 does not have to use add to implement these. Instead, it can use add, sub, inc, dec instructions with the "lock" prefix. This is currently implemented using a bit of instruction selection trick. The issue is the target independent pattern produces one output and a chain and we want to map it into one that just output a chain. The current trick is to select it into a merge_values with the first definition being an implicit_def. The proper solution is to add new ISD opcodes for the no-output variant. DAG combiner can then transform the node before it gets to target node selection. Problem #2 is we are adding a whole bunch of x86 atomic instructions when in fact these instructions are identical to the non-lock versions. We need a way to add target specific information to target nodes and have this information carried over to machine instructions. Asm printer (or JIT) can use this information to add the "lock" prefix. llvm-svn: 77582 | ||||
* | Switch obvious clients to Twine instead of utostr (when they were already using | Daniel Dunbar | 2009-07-30 | 5 | -20/+15 |
| | | | | | | | | | a Twine, e.g., for names). - I am a little ambivalent about this; we don't want the string conversion of utostr, but using overload '+' mixed with string and integer arguments is sketchy. On the other hand, this particular usage is something of an idiom. llvm-svn: 77579 | ||||
* | Keep track of references to mem(cpy,move,set) and then print only one extern | Sanjiv Gupta | 2009-07-30 | 2 | -2/+41 |
| | | | | | | declaration for them. llvm-svn: 77578 | ||||
* | Twine: Provide [u]int{32,64} conversions via implicit constructors instead of | Daniel Dunbar | 2009-07-30 | 1 | -6/+18 |
| | | | | | | explicitly. llvm-svn: 77576 | ||||
* | Equal SCEVs of a subscript give rise to dependence. | Andreas Bolka | 2009-07-30 | 1 | -1/+11 |
| | | | | llvm-svn: 77570 | ||||
* | Typo | Nate Begeman | 2009-07-30 | 1 | -1/+1 |
| | | | | llvm-svn: 77568 | ||||
* | Add a new register class to describe operands that can't be SP, | Dan Gohman | 2009-07-30 | 5 | -38/+189 |
| | | | | | | | due to x86 encoding restrictions. This is currently off by default because it may cause code quality regressions. This is for PR4572. llvm-svn: 77565 | ||||
* | Check null NameMDNode elements. | Devang Patel | 2009-07-30 | 1 | -4/+8 |
| | | | | llvm-svn: 77559 | ||||
* | Eliminate a bunch of redundant tables. | Dan Gohman | 2009-07-30 | 1 | -167/+61 |
| | | | | llvm-svn: 77558 | ||||
* | Lower a 128-bit BUILD_VECTOR with 2 elements to a pair of INSERT_VECTOR_ELTs. | Bob Wilson | 2009-07-30 | 1 | -1/+19 |
| | | | | llvm-svn: 77557 | ||||
* | Use array_endof instead of doing it manually. | Dan Gohman | 2009-07-30 | 1 | -27/+20 |
| | | | | llvm-svn: 77553 | ||||
* | print single NamedMDNode. | Devang Patel | 2009-07-30 | 1 | -3/+45 |
| | | | | llvm-svn: 77549 | ||||
* | tbb / tbh instructions only branch forward, not backwards. | Evan Cheng | 2009-07-29 | 1 | -2/+4 |
| | | | | llvm-svn: 77522 | ||||
* | Add VFP3 D registers to the DPR register class. | Evan Cheng | 2009-07-29 | 1 | -1/+3 |
| | | | | llvm-svn: 77521 | ||||
* | In TrimAllocationToSize, if a block is below the minimum allocation size, | Nicolas Geoffray | 2009-07-29 | 1 | -1/+4 |
| | | | | | | | there is no new block added to the free list. Therefore on the next startFunctionBody call, a new slab must be allocated. llvm-svn: 77520 | ||||
* | Eliminate a few unused-variable warnings | Douglas Gregor | 2009-07-29 | 4 | -7/+0 |
| | | | | llvm-svn: 77519 | ||||
* | Read and write NamedMDNode. | Devang Patel | 2009-07-29 | 3 | -1/+68 |
| | | | | llvm-svn: 77517 | ||||
* | Move types back to the 2.5 API. | Owen Anderson | 2009-07-29 | 53 | -356/+235 |
| | | | | llvm-svn: 77516 | ||||
* | Print named metadata. | Devang Patel | 2009-07-29 | 1 | -0/+25 |
| | | | | llvm-svn: 77513 | ||||
* | Fix PR4645 which was fallout from the fix for PR4641. | Daniel Dunbar | 2009-07-29 | 1 | -8/+15 |
| | | | | | | - Call RAUW to delete all instructions (this is a patch from Nick Lewycky). llvm-svn: 77512 | ||||
* | There is no need to keep name ref in NamedMDNode. | Devang Patel | 2009-07-29 | 2 | -10/+6 |
| | | | | llvm-svn: 77511 |