This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Patch] Unbreak build on mips (was Re: Passes are now C++ classes (was Re: [PATCH 03/11] Handwritten part of conversion of passes to C++ classes))


On Tue, 2013-08-06 at 08:16 -0700, Steve Ellcey wrote:
> On Mon, 2013-08-05 at 17:03 -0400, David Malcolm wrote:
> 
> > Given all of the above testing I'm reasonably confident that this works.
> > However this is such a large change [1] that there's a non-zero chance
> > of at least one glitch - let me know if you see any breakages.
> 
> The mips*-*-* targets are not building.  It looks like the mips reorg
> pass (pass_mips_machine_reorg2) in config/mips/mips.c was not converted
> and/or was not converted correctly.

Sorry about this.

I was able to reproduce this build error with configure
--target=mips-linux-elf:
../../src/gcc/config/mips/mips.c:16379:28: error: expected
primary-expression before â.â token

I'm attaching a patch which fixes that issue for me.  Only lightly
tested (build&host=x86_64, target as above) - I'm able to build stage1,
and cc1 appears to generate assembler on a simple test case.  I stepped
through the changed code in the debugger and it appears to do the right
thing.  However I'm not familiar with the internals of the pass in
question.

commit 11d46884e8bd9802b0f528a16b3970b4076fe8a9
Author: David Malcolm <dmalcolm@redhat.com>
Date:   Tue Aug 6 13:48:59 2013 -0400

    gcc/
    	* config/mips/mips.c (insert_pass_mips_machine_reorg2): Move
    	into...
    	(mips_option_override): ...here, porting to new C++ API for
    	passes.

diff --git a/gcc/config/mips/mips.c b/gcc/config/mips/mips.c
index 05ba003..4da80f4 100644
--- a/gcc/config/mips/mips.c
+++ b/gcc/config/mips/mips.c
@@ -56,6 +56,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "target-globals.h"
 #include "opts.h"
 #include "tree-pass.h"
+#include "context.h"
 
 /* True if X is an UNSPEC wrapper around a SYMBOL_REF or LABEL_REF.  */
 #define UNSPEC_ADDRESS_P(X)					\
@@ -16374,13 +16375,6 @@ make_pass_mips_machine_reorg2 (gcc::context *ctxt)
   return new pass_mips_machine_reorg2 (ctxt);
 }
 
-struct register_pass_info insert_pass_mips_machine_reorg2 =
-{
-  &pass_mips_machine_reorg2.pass,	/* pass */
-  "dbr",				/* reference_pass_name */
-  1,					/* ref_pass_instance_number */
-  PASS_POS_INSERT_AFTER			/* po_op */
-};
 
 /* Implement TARGET_ASM_OUTPUT_MI_THUNK.  Generate rtl rather than asm text
    in order to avoid duplicating too much logic from elsewhere.  */
@@ -17174,6 +17168,14 @@ mips_option_override (void)
   /* We register a second machine specific reorg pass after delay slot
      filling.  Registering the pass must be done at start up.  It's
      convenient to do it here.  */
+  opt_pass *new_pass = make_pass_mips_machine_reorg2 (g);
+  struct register_pass_info insert_pass_mips_machine_reorg2 =
+    {
+      new_pass,		/* pass */
+      "dbr",			/* reference_pass_name */
+      1,			/* ref_pass_instance_number */
+      PASS_POS_INSERT_AFTER	/* po_op */
+    };
   register_pass (&insert_pass_mips_machine_reorg2);
 
   if (TARGET_HARD_FLOAT_ABI && TARGET_MIPS5900)

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]