maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
15,556 Commits 23 Branches 0 Tags 2.2 GiB
Commit Graph

15556 Commits

Author SHA1 Message Date
Reid Spencer 0dbf101c90 Duh, put tools in *bin* directory, not *tools* directory as per 
PR456.

llvm-svn: 17964
 2004-11-18 20:04:39 +00:00
Chris Lattner 7fd27e2c35 Remove debugging code, unneuter this functionality 
llvm-svn: 17963
 2004-11-18 19:42:50 +00:00
Chris Lattner 8cb483b335 remove debugging code 
llvm-svn: 17962
 2004-11-18 19:40:13 +00:00
Tanya Lattner cb0da4f1b7 Fixed to fit in 80 columns. 
llvm-svn: 17961
 2004-11-18 18:47:29 +00:00
Reid Spencer e4190faa0c Add notes about makefile changes and llvm-ar. 
llvm-svn: 17960
 2004-11-18 18:38:58 +00:00
Tanya Lattner 21590db8fd Adding option to llc for ModuloScheduling. By default it is turned off. 
llvm-svn: 17959
 2004-11-18 18:38:01 +00:00
Chris Lattner de5626cd40 These methods are inlined 
llvm-svn: 17958
 2004-11-18 17:47:13 +00:00
Chris Lattner 8a8e988766 Update comments, now that CPR's are gone, inline the methods 
llvm-svn: 17957
 2004-11-18 17:46:57 +00:00
Chris Lattner c08ac110df Check in hook that I forgot 
llvm-svn: 17956
 2004-11-18 17:24:20 +00:00
Reid Spencer 7e9e0b0c82 Fix PR456:\ 
Tools and libraries will be built into $(BUILD_OBJ_ROOT)/$(BuildMode)/bin and \
$(BUILD_OBJ_ROOT)/$(BuildMode)/lib, respectively. Furthermore, the example \
programs will go in $(BUILD_OBJ_ROOT)/$(BuildMode)/examples to keep them \
separate from the tools and hopefully out of the PATH. Install targets \
have not changed.

llvm-svn: 17953
 2004-11-18 10:03:46 +00:00
Reid Spencer 231e4d7772 Include ltdl.h if we have it. 
llvm-svn: 17952
 2004-11-18 09:50:00 +00:00
Reid Spencer 6d854b0b07 Add checks for HAVE_LTDL_H and HAVE_LT_DLOPEN 
llvm-svn: 17951
 2004-11-18 09:49:13 +00:00
Reid Spencer c4b5cfc602 Changes for testing presence of ltdl.h and lt_dlopen(), libtool functions 
for cross-platform dynamic loading of shared objects.

llvm-svn: 17950
 2004-11-18 09:47:37 +00:00
Brian Gaeke de6e536849 Rewrite LoadArgumentsToVirtualRegs, making it match almost exactly how 
visitCallInst works.  Support cast of byte/short/int to long.

llvm-svn: 17949
 2004-11-18 07:43:33 +00:00
Brian Gaeke 2c6b9c914b Update current expected failures list - expand it to include all of 
SingleSource. Update to-do list (open-coding refers to binary operations on
longs, not to passing them into functions, which we already support.)

llvm-svn: 17948
 2004-11-18 07:43:32 +00:00
Nate Begeman 2b575bec80 Eliminate another common source of moves that the register allocator 
shouldn't be forced to coalesce for us: folded GEP operations.  This too
fires thousands of times across the testsuite.

llvm-svn: 17947
 2004-11-18 07:22:46 +00:00
Nate Begeman 5d9c292e7f When accessing the base register for global variables, use the register 
directly rather than making a copy for the register allocator to coalesce.
This kills thousands of live intervals across the testsuite.

llvm-svn: 17946
 2004-11-18 06:51:29 +00:00
Chris Lattner b75e790a2b * There is no reason for SpillWeights to be an instance var 
* Do not put fixed registers into the unhandled set.  This means they will
  never find their way into the inactive, active, or handled sets, so we
  can simplify a bunch of code.

llvm-svn: 17945
 2004-11-18 06:01:45 +00:00
Chris Lattner e3b9cb9959 There is no need to check to see if j overflowed in this loop as we're only 
incrementing i.

llvm-svn: 17944
 2004-11-18 05:28:21 +00:00
Chris Lattner 6e0c3f44ba Moderate head scratching reveals that this conditional is not needed. If 
i->start == j->start, then certainly i->end > j->start.

llvm-svn: 17943
 2004-11-18 05:19:02 +00:00
Nate Begeman 51f6635e42 Clean up and fix cast codegen by removing cases that are handled elsewhere, 
and properly emitting signed short to unsigned int.  This fixes the last
regression vs. the CBE, MultiSource/Applications/hbd.

llvm-svn: 17942
 2004-11-18 04:56:53 +00:00
Reid Spencer dfa81ff3ff Dynamic Library abstraction. This makes the abstraction of a single dynamic 
library (shared library/shared object) whose symbols can be looked up
dynamically. Used for plug-ins.

llvm-svn: 17941
 2004-11-18 04:33:40 +00:00
Reid Spencer 4a763ca854 Dynamic Library abstraction. This makes the abstraction of a single dynamic 
library (shared library/shared object) whose symbols can be looked up
dynamically. Used for plug-ins.

llvm-svn: 17940
 2004-11-18 04:33:39 +00:00
Chris Lattner 08ec603bb0 Fix a couple of bugs where we considered physregs past their range as possibly 
intersecting an interval.

llvm-svn: 17939
 2004-11-18 04:33:31 +00:00
Chris Lattner cf972ff7cb Fix typeo 
llvm-svn: 17938
 2004-11-18 04:31:10 +00:00
Chris Lattner 49ff5f0477 Start using the iterators in the fixed_ intervals to avoid having to binary 
search physreg intervals every time we access it.  This takes another
half second off of linscan.

llvm-svn: 17937
 2004-11-18 04:13:02 +00:00
Chris Lattner 7598c316e5 Take another .7 seconds off of linear scan time. 
llvm-svn: 17936
 2004-11-18 04:02:11 +00:00
Chris Lattner 850852c789 Add a counter for the number of times linscan has to backtrack. Start using 
the iterator hints we have to speed up overlaps().  This speeds linscan up
by about .2s (out of 8.7) on 175.vpr for PPC.

llvm-svn: 17935
 2004-11-18 03:49:30 +00:00
Chris Lattner cb0c9655bf Add ability to give hints to the overlaps routines. 
llvm-svn: 17934
 2004-11-18 03:47:34 +00:00
Chris Lattner a1f7779876 * Improve comments/documentation substantially 
* Eliminate the releaseMemory method, this is not an analysis
* Change the fixed, active, and inactive lists of intervals to maintain an
  iterator for the current position in the interval.  This allows us to do
  constant time increments of the iterator instead of having to do a binary
  search to find our liverange in our liveinterval all of the time, which
  substantially speeds up cases where LiveIntervals have many LiveRanges
  - which is very common for physical registers.  On targets with many
  physregs, this can make a noticable difference.

  With a release build of LLC for PPC, this halves the time in
  processInactiveIntervals and processActiveIntervals, from 1.5s to .75s.

  This also lays the ground for more to come.

llvm-svn: 17933
 2004-11-18 02:42:27 +00:00
Chris Lattner 0624f95175 Add new advanceTo method 
llvm-svn: 17932
 2004-11-18 02:37:31 +00:00
Chris Lattner 7d92fc7e89 Fix a minor bug in expiredAt. endNumber() is the first number that is not valid. 
llvm-svn: 17931
 2004-11-18 01:34:44 +00:00
Chris Lattner a51f5eeac1 Rename some methods, use 'begin' instead of 'start', add new LiveInterval 
iterator/begin/end members.

llvm-svn: 17930
 2004-11-18 01:29:39 +00:00
Brian Gaeke 1553b58f39 Allocate fewer registers and tighten up alignment restrictions. 
llvm-svn: 17929
 2004-11-18 00:25:20 +00:00
Brian Gaeke 2a3e8a7268 Update to-do list. 
llvm-svn: 17927
 2004-11-17 22:33:55 +00:00
Brian Gaeke 3b10744e0f Update list of failing SingleSource Benchmarks 
llvm-svn: 17926
 2004-11-17 22:33:28 +00:00
Brian Gaeke f1233fb84f We were (somehow) getting the wrong branch opcode for setcc float instrs. 
llvm-svn: 17925
 2004-11-17 22:06:56 +00:00
Alkis Evlogimenos 531e901deb Make ReturnInst accept a value of type void as the return value. The 
ReturnInst constructed is the same as if NULL was passed instead of
the void value.

llvm-svn: 17923
 2004-11-17 21:02:25 +00:00
Reid Spencer f8d6f7e973 Add a standards section. 
llvm-svn: 17921
 2004-11-17 20:19:18 +00:00
Reid Spencer 91234ff57a Fix PR458: 
* Don't include Makefile.rules in set of preconditions, it never has to be
  copied to objdir.
* Enable the "update makefile first before executing targets" feature in
  gnu make by *not* using a full path to the Makefile in the rule.

llvm-svn: 17920
 2004-11-17 19:08:44 +00:00
Reid Spencer 67c46c8d70 Fix some things for Mac OSX archives: 
* ensure trailing spaces are eliminated so they don't factor into the
  length of a member's name.
* make sure all the bytes of a name are written even if the name ends in
  multiple null characters (bug in OSX ar)
* make sure we provide the full member name when searching for symbols so
  the module name is not accidentally duplicated.

llvm-svn: 17918
 2004-11-17 18:28:29 +00:00
Reid Spencer b58bf5fa28 Make sure we parse bytecode with a module identifier that reflects the full 
name of the module: "Archive.a(object.o)"

llvm-svn: 17917
 2004-11-17 18:25:21 +00:00
Chris Lattner d5f5288066 Make this more accurate 
llvm-svn: 17916
 2004-11-17 18:19:32 +00:00
Chris Lattner 1c18011de6 New prototype for lowerpacked pass. 
llvm-svn: 17915
 2004-11-17 18:01:49 +00:00
Chris Lattner ba18b9a0da Generalize this code to turn any cast-to-first-element-of into a gep constexpr 
llvm-svn: 17914
 2004-11-17 17:59:35 +00:00
Chris Lattner c368b63515 Simplify conditional and fix LICM/2004-11-17-UndefIndexCrash.ll 
by saying what we mean

llvm-svn: 17913
 2004-11-17 17:39:39 +00:00
Chris Lattner 9d339a75ae New testcase that crashes licm 
llvm-svn: 17912
 2004-11-17 17:39:19 +00:00
Chris Lattner 7a3767521d -disable-opt is not -O0, it's okay for it to disable internalize. 
llvm-svn: 17911
 2004-11-17 16:41:19 +00:00
Reid Spencer 1154bb6b43 Adjust long file name writing to match BSD 4.4 and Mac OSX style archives. 
llvm-svn: 17910
 2004-11-17 16:14:21 +00:00
Reid Spencer 0dea46fe67 Despite documentation to the contrary, Mac OSX and BSD 4.4 archive formats 
*do* include the length of the long file in the length of the member and
they are *not* null terminated.

llvm-svn: 17909
 2004-11-17 16:13:11 +00:00