svn commit: r289222 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Tue Oct 13 16:25:02 2015
New Revision: 289222
URL: https://svnweb.freebsd.org/changeset/base/289222

Log:
  Add llvm patch corresponding to r289221.

Added:
  head/contrib/llvm/patches/patch-08-llvm-r250085-fix-avx-crash.diff

Added: head/contrib/llvm/patches/patch-08-llvm-r250085-fix-avx-crash.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-08-llvm-r250085-fix-avx-crash.diff  Tue Oct 
13 16:25:02 2015(r289222)
@@ -0,0 +1,142 @@
+Pull in r250085 from upstream llvm trunk (by Andrea Di Biagio):
+
+  [x86] Fix wrong lowering of vsetcc nodes (PR25080).
+
+  Function LowerVSETCC (in X86ISelLowering.cpp) worked under the wrong
+  assumption that for non-AVX512 targets, the source type and destination type
+  of a type-legalized setcc node were always the same type.
+
+  This assumption was unfortunately incorrect; the type legalizer is not always
+  able to promote the return type of a setcc to the same type as the first
+  operand of a setcc.
+
+  In the case of a vsetcc node, the legalizer firstly checks if the first input
+  operand has a legal type. If so, then it promotes the return type of the 
vsetcc
+  to that same type. Otherwise, the return type is promoted to the 'next legal
+  type', which, for vectors of MVT::i1 is always a 128-bit integer vector type.
+
+  Example (-mattr=+avx):
+
+%0 = trunc <8 x i32> %a to <8 x i23>
+%1 = icmp eq <8 x i23> %0, zeroinitializer
+
+  The initial selection dag for the code above is:
+
+  v8i1 = setcc t5, t7, seteq:ch
+t5: v8i23 = truncate t2
+  t2: v8i32,ch = CopyFromReg t0, Register:v8i32 %vreg1
+  t7: v8i32 = build_vector of all zeroes.
+
+  The type legalizer would firstly check if 't5' has a legal type. If so, then 
it
+  would reuse that same type to promote the return type of the setcc node.
+  Unfortunately 't5' is of illegal type v8i23, and therefore it cannot be used 
to
+  promote the return type of the setcc node. Consequently, the setcc return 
type
+  is promoted to v8i16. Later on, 't5' is promoted to v8i32 thus leading to the
+  following dag node:
+v8i16 = setcc t32, t25, seteq:ch
+
+where t32 and t25 are now values of type v8i32.
+
+  Before this patch, function LowerVSETCC would have wrongly expanded the setcc
+  to a single X86ISD::PCMPEQ. Surprisingly, ISel was still able to match an
+  instruction. In our case, ISel would have matched a VPCMPEQWrr:
+t37: v8i16 = X86ISD::VPCMPEQWrr t36, t25
+
+  However, t36 and t25 are both VR256, while the result type is instead of 
class
+  VR128. This inconsistency ended up causing the insertion of COPY instructions
+  like this:
+%vreg7 = COPY %vreg3; VR128:%vreg7 VR256:%vreg3
+
+  Which is an invalid full copy (not a sub register copy).
+  Eventually, the backend would have hit an UNREACHABLE "Cannot emit physreg 
copy
+  instruction" in the attempt to expand the malformed pseudo COPY instructions.
+
+  This patch fixes the problem adding the missing logic in LowerVSETCC to 
handle
+  the corner case of a setcc with 128-bit return type and 256-bit operand type.
+
+  This problem was originally reported by Dimitry as PR25080. It has been 
latent
+  for a very long time. I have added the minimal reproducible from that 
bugzilla
+  as test setcc-lowering.ll.
+
+  Differential Revision: http://reviews.llvm.org/D13660
+
+This should fix the "Cannot emit physreg copy instruction" errors when
+compiling contrib/wpa/src/common/ieee802_11_common.c, and CPUTYPE is set
+to a CPU supporting AVX (e.g. sandybridge, ivybridge).
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/289221
+
+Index: lib/Target/X86/X86ISelLowering.cpp
+===
+--- lib/Target/X86/X86ISelLowering.cpp
 lib/Target/X86/X86ISelLowering.cpp
+@@ -13573,6 +13573,35 @@ static SDValue LowerVSETCC(SDValue Op, const X86Su
+DAG.getConstant(SSECC, dl, MVT::i8));
+   }
+ 
++  MVT VTOp0 = Op0.getSimpleValueType();
++  assert(VTOp0 == Op1.getSimpleValueType() &&
++ "Expected operands with same type!");
++  assert(VT.getVectorNumElements() == VTOp0.getVectorNumElements() &&
++ "Invalid number of packed elements for source and destination!");
++
++  if (VT.is128BitVector() && VTOp0.is256BitVector()) {
++// On non-AVX512 targets, a vector of MVT::i1 is promoted by the type
++// legalizer to a wider vector type.  In the case of 'vsetcc' nodes, the
++// legalizer firstly checks if the first operand in input to the setcc has
++// a legal type. If so, then it promotes the return type to that same 
type.
++// Otherwise, the return type is promoted to the 'next legal type' which,
++// for a vector of MVT::i1 is always a 128-bit integer vector type.
++//
++// We reach this code only if the following two conditions are met:
++

Re: svn commit: r283526 - in head: . contrib/llvm/include/llvm/Target contrib/llvm/lib/Analysis contrib/llvm/lib/CodeGen contrib/llvm/lib/CodeGen/SelectionDAG contrib/llvm/lib/ExecutionEngine/RuntimeD

[Dimitry Andric]

On 10 Oct 2015, at 23:32, Ivan Klymenko <fi...@ukr.net> wrote:
> On Mon, 25 May 2015 13:43:03 + (UTC)
> Dimitry Andric <d...@freebsd.org> wrote:
> 
>> Author: dim
>> Date: Mon May 25 13:43:03 2015
>> New Revision: 283526
>> URL: https://svnweb.freebsd.org/changeset/base/283526
>> 
>> Log:
>>  Upgrade our copy of clang and llvm to 3.6.1 release.

This is some time ago already, you probably meant the upgrade to 3.7.0?
:-)


> Cannot emit physreg copy instruction
> UNREACHABLE executed at 
> /usr/src/lib/clang/libllvmx86codegen/../../../contrib/llvm/lib/Target/X86/X86InstrInfo.cpp:3935!
> cc: error: unable to execute command: Abort trap
> cc: error: clang frontend command failed due to signal (use -v to see 
> invocation)
> FreeBSD clang version 3.7.0 (tags/RELEASE_370/final 246257) 20150906
> Target: x86_64-unknown-freebsd11.0
> Thread model: posix
> cc: note: diagnostic msg: PLEASE submit a bug report to 
> https://bugs.freebsd.org/submit/ and include the crash backtrace, 
> preprocessed source, and associated run script.
> cc: note: diagnostic msg:
> 
> 
> PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:
> Preprocessed source(s) and associated run script(s) are located at:
> cc: note: diagnostic msg: /tmp/ieee802_11_common-974cbf.c
> cc: note: diagnostic msg: /tmp/ieee802_11_common-974cbf.sh

Yes, this error has already been reported; please see the thread
starting here:

https://lists.freebsd.org/pipermail/freebsd-current/2015-October/057701.html

It has been submitted as an upstream bug, and a fix is being worked on.
For now, you can work around it by unsetting CPUTYPE.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r289082 - head/contrib/libc++/include

[Dimitry Andric]

Author: dim
Date: Fri Oct  9 21:04:28 2015
New Revision: 289082
URL: https://svnweb.freebsd.org/changeset/base/289082

Log:
  Pull in r242623 from upstream libc++ trunk (by Eric Fiselier):
  
Enable and fix warnings during the build.
  
Although CMake adds warning flags, they are ignored in the libc++ headers
because the headers '#pragma system header' themselves.
  
This patch disables the system header pragma when building libc++ and fixes
the warnings that arose.
  
The warnings fixed were:
1.  - anonymous structs are a GNU extension
2.  - anonymous structs are a GNU extension.
3. <__hash_table> - Embedded preprocessor directives have undefined 
behavior.
4.  - Definition is missing noexcept from declaration.
5. <__std_stream> - Unused variable.
  
  This should fix building world (in particular libatf-c++) with -std=c++11.
  
  Reported by:  Oliver Hartmann 

Modified:
  head/contrib/libc++/include/__hash_table
  head/contrib/libc++/include/__std_stream
  head/contrib/libc++/include/functional
  head/contrib/libc++/include/memory
  head/contrib/libc++/include/string

Modified: head/contrib/libc++/include/__hash_table
==
--- head/contrib/libc++/include/__hash_tableFri Oct  9 21:00:04 2015
(r289081)
+++ head/contrib/libc++/include/__hash_tableFri Oct  9 21:04:28 2015
(r289082)
@@ -984,15 +984,17 @@ public:
 __equal_range_multi(const _Key& __k) const;
 
 void swap(__hash_table& __u)
+#if _LIBCPP_STD_VER <= 11
 _NOEXCEPT_(
 __is_nothrow_swappable::value && 
__is_nothrow_swappable::value
-#if _LIBCPP_STD_VER <= 11
 && 
(!allocator_traits<__pointer_allocator>::propagate_on_container_swap::value
   || __is_nothrow_swappable<__pointer_allocator>::value)
 && (!__node_traits::propagate_on_container_swap::value
   || __is_nothrow_swappable<__node_allocator>::value)
-#endif
 );
+#else
+ _NOEXCEPT_(__is_nothrow_swappable::value && 
__is_nothrow_swappable::value);
+#endif
 
 _LIBCPP_INLINE_VISIBILITY
 size_type max_bucket_count() const _NOEXCEPT
@@ -2351,15 +2353,17 @@ __hash_table<_Tp, _Hash, _Equal, _Alloc>
 template 
 void
 __hash_table<_Tp, _Hash, _Equal, _Alloc>::swap(__hash_table& __u)
+#if _LIBCPP_STD_VER <= 11
 _NOEXCEPT_(
 __is_nothrow_swappable::value && 
__is_nothrow_swappable::value
-#if _LIBCPP_STD_VER <= 11
 && 
(!allocator_traits<__pointer_allocator>::propagate_on_container_swap::value
   || __is_nothrow_swappable<__pointer_allocator>::value)
 && (!__node_traits::propagate_on_container_swap::value
   || __is_nothrow_swappable<__node_allocator>::value)
-#endif
 )
+#else
+  _NOEXCEPT_(__is_nothrow_swappable::value && 
__is_nothrow_swappable::value)
+#endif
 {
 {
 __node_pointer_pointer __npp = __bucket_list_.release();

Modified: head/contrib/libc++/include/__std_stream
==
--- head/contrib/libc++/include/__std_streamFri Oct  9 21:00:04 2015
(r289081)
+++ head/contrib/libc++/include/__std_streamFri Oct  9 21:04:28 2015
(r289082)
@@ -276,7 +276,6 @@ __stdoutbuf<_CharT>::overflow(int_type _
 codecvt_base::result __r;
 char_type* pbase = &__1buf;
 char_type* pptr = pbase + 1;
-char_type* epptr = pptr;
 do
 {
 const char_type* __e;

Modified: head/contrib/libc++/include/functional
==
--- head/contrib/libc++/include/functional  Fri Oct  9 21:00:04 2015
(r289081)
+++ head/contrib/libc++/include/functional  Fri Oct  9 21:04:28 2015
(r289082)
@@ -2402,14 +2402,14 @@ struct _LIBCPP_TYPE_VIS_ONLY hash::operator()(__v);
 #endif

Modified: head/contrib/libc++/include/memory
==
--- head/contrib/libc++/include/memory  Fri Oct  9 21:00:04 2015
(r289081)
+++ head/contrib/libc++/include/memory  Fri Oct  9 21:04:28 2015
(r289082)
@@ -3420,7 +3420,7 @@ struct __scalar_hash<_Tp, 2>
 {
 size_t __a;
 size_t __b;
-};
+} __s;
 } __u;
 __u.__t = __v;
 return __murmur2_or_cityhash()(&__u, sizeof(__u));
@@ -3442,7 +3442,7 @@ struct __scalar_hash<_Tp, 3>
 size_t __a;
 size_t __b;
 size_t __c;
-};
+} __s;
 } __u;
 __u.__t = __v;
 return __murmur2_or_cityhash()(&__u, sizeof(__u));
@@ -3465,7 +3465,7 @@ struct __scalar_hash<_Tp, 4>
 size_t __b;
 size_t __c;
 size_t __d;
- 

svn commit: r289081 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Fri Oct  9 21:00:04 2015
New Revision: 289081
URL: https://svnweb.freebsd.org/changeset/base/289081

Log:
  Add llvm patch corresponding to r289072.

Added:
  head/contrib/llvm/patches/patch-07-undo-llvm-r240144-iostream-sigbus.diff

Added: head/contrib/llvm/patches/patch-07-undo-llvm-r240144-iostream-sigbus.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-07-undo-llvm-r240144-iostream-sigbus.diff   
Fri Oct  9 21:00:04 2015(r289081)
@@ -0,0 +1,50 @@
+Temporarily revert upstream llvm trunk r240144 (by Michael Zolotukhin):
+
+  [SLP] Vectorize for all-constant entries.
+
+This should fix libc++'s iostream initialization SIGBUSing whenever the
+global cout symbol is not aligned to 16 bytes.
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/289072
+
+Index: lib/Transforms/Vectorize/SLPVectorizer.cpp
+===
+--- lib/Transforms/Vectorize/SLPVectorizer.cpp
 lib/Transforms/Vectorize/SLPVectorizer.cpp
+@@ -1642,10 +1642,8 @@ bool BoUpSLP::isFullyVectorizableTinyTree() {
+   if (VectorizableTree.size() != 2)
+ return false;
+ 
+-  // Handle splat and all-constants stores.
+-  if (!VectorizableTree[0].NeedToGather &&
+-  (allConstant(VectorizableTree[1].Scalars) ||
+-   isSplat(VectorizableTree[1].Scalars)))
++  // Handle splat stores.
++  if (!VectorizableTree[0].NeedToGather && 
isSplat(VectorizableTree[1].Scalars))
+ return true;
+ 
+   // Gathering cost would be too much for tiny trees.
+Index: test/Transforms/SLPVectorizer/X86/tiny-tree.ll
+===
+--- test/Transforms/SLPVectorizer/X86/tiny-tree.ll
 test/Transforms/SLPVectorizer/X86/tiny-tree.ll
+@@ -153,19 +153,3 @@ define void @store_splat(float*, float) {
+   store float %1, float* %6, align 4
+   ret void
+ }
+-
+-
+-; CHECK-LABEL: store_const
+-; CHECK: store <4 x i32>
+-define void @store_const(i32* %a) {
+-entry:
+-  %ptr0 = getelementptr inbounds i32, i32* %a, i64 0
+-  store i32 10, i32* %ptr0, align 4
+-  %ptr1 = getelementptr inbounds i32, i32* %a, i64 1
+-  store i32 30, i32* %ptr1, align 4
+-  %ptr2 = getelementptr inbounds i32, i32* %a, i64 2
+-  store i32 20, i32* %ptr2, align 4
+-  %ptr3 = getelementptr inbounds i32, i32* %a, i64 3
+-  store i32 40, i32* %ptr3, align 4
+-  ret void
+-}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r289072 - in head/contrib: libc++/src llvm/lib/Transforms/Vectorize

[Dimitry Andric]

Author: dim
Date: Fri Oct  9 18:21:45 2015
New Revision: 289072
URL: https://svnweb.freebsd.org/changeset/base/289072

Log:
  Temporarily revert upstream llvm trunk r240144 (by Michael Zolotukhin):
  
[SLP] Vectorize for all-constant entries.
  
  This should fix libc++'s iostream initialization SIGBUSing on amd64,
  whenever the global cout symbol is not aligned to 16 bytes.
  
  Some further explanation: libc++'s iostream.cpp contains the definitions
  of std::cout, std::cerr and so on.  These global objects are effectively
  declared with an alignment of 8 bytes.  When an executable is linked
  against libc++.so, it can sometimes get a copy of the global object,
  which is then at the same alignment.
  
  However, with clang 3.7.0, the initialization of these global objects
  will incorrectly use SSE instructions (e.g. movdqa), whenever the
  optimization level is high enough, and SSE is enabled, such as on amd64.
  When any of these objects is not aligned to 16 bytes, this will result
  in a SIGBUS during iostream initialization.  In contrast, clang 3.6.x
  and earlier took the 8 byte alignment into consideration, and avoided
  SSE for those particular operations.
  
  After bisecting of upstream changes, I found that the above revision
  caused the change of this behavior, so I am reverting it now as a
  workaround, while a discussion and test case is being prepared for
  upstream.

Modified:
  head/contrib/libc++/src/iostream.cpp
  head/contrib/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp

Modified: head/contrib/libc++/src/iostream.cpp
==
--- head/contrib/libc++/src/iostream.cppFri Oct  9 17:46:05 2015
(r289071)
+++ head/contrib/libc++/src/iostream.cppFri Oct  9 18:21:45 2015
(r289072)
@@ -86,3 +86,4 @@ ios_base::Init::~Init()
 }
 
 _LIBCPP_END_NAMESPACE_STD
+

Modified: head/contrib/llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
==
--- head/contrib/llvm/lib/Transforms/Vectorize/SLPVectorizer.cppFri Oct 
 9 17:46:05 2015(r289071)
+++ head/contrib/llvm/lib/Transforms/Vectorize/SLPVectorizer.cppFri Oct 
 9 18:21:45 2015(r289072)
@@ -1642,10 +1642,8 @@ bool BoUpSLP::isFullyVectorizableTinyTre
   if (VectorizableTree.size() != 2)
 return false;
 
-  // Handle splat and all-constants stores.
-  if (!VectorizableTree[0].NeedToGather &&
-  (allConstant(VectorizableTree[1].Scalars) ||
-   isSplat(VectorizableTree[1].Scalars)))
+  // Handle splat stores.
+  if (!VectorizableTree[0].NeedToGather && 
isSplat(VectorizableTree[1].Scalars))
 return true;
 
   // Gathering cost would be too much for tiny trees.
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r289073 - head/contrib/libc++/src

[Dimitry Andric]

Author: dim
Date: Fri Oct  9 18:23:10 2015
New Revision: 289073
URL: https://svnweb.freebsd.org/changeset/base/289073

Log:
  Remove empty line again from libc++'s iostream.cpp.  This was used to
  force updates to this file, so it will be rebuilt by the fixed clang
  from r289072.

Modified:
  head/contrib/libc++/src/iostream.cpp

Modified: head/contrib/libc++/src/iostream.cpp
==
--- head/contrib/libc++/src/iostream.cppFri Oct  9 18:21:45 2015
(r289072)
+++ head/contrib/libc++/src/iostream.cppFri Oct  9 18:23:10 2015
(r289073)
@@ -86,4 +86,3 @@ ios_base::Init::~Init()
 }
 
 _LIBCPP_END_NAMESPACE_STD
-
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r288951 - head/lib/clang

[Dimitry Andric]

Author: dim
Date: Tue Oct  6 19:49:53 2015
New Revision: 288951
URL: https://svnweb.freebsd.org/changeset/base/288951

Log:
  For llvm/clang libraries, skip including tablegen-produced .d files when
  the target is "make depend".  This works around errors during
  incremental make depend of some clang libraries, for example "don't know
  how to make contrib/llvm/include/llvm/IR/IntrinsicsR600.td".
  
  Reported by:  emaste

Modified:
  head/lib/clang/clang.build.mk

Modified: head/lib/clang/clang.build.mk
==
--- head/lib/clang/clang.build.mk   Tue Oct  6 19:31:07 2015
(r288950)
+++ head/lib/clang/clang.build.mk   Tue Oct  6 19:49:53 2015
(r288951)
@@ -247,9 +247,11 @@ Checkers.inc.h: ${CLANG_SRCS}/lib/Static
-I ${CLANG_SRCS}/include -d ${.TARGET:C/\.h$/.d/} -o ${.TARGET} \
${CLANG_SRCS}/lib/StaticAnalyzer/Checkers/Checkers.td
 
-.for dep in ${TGHDRS:C/$/.inc.d/}
-. sinclude "${dep}"
-.endfor
+.if !make(depend)
+. for dep in ${TGHDRS:C/$/.inc.d/}
+.  sinclude "${dep}"
+. endfor
+.endif
 
 SRCS+= ${TGHDRS:C/$/.inc.h/}
 DPSRCS+=   ${TGHDRS:C/$/.inc.h/}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r288099 - head/contrib/binutils/opcodes

[Dimitry Andric]

On 22 Sep 2015, at 16:23, Ian Lepore <i...@freebsd.org> wrote:
> 
> On Tue, 2015-09-22 at 11:53 +0200, Dimitry Andric wrote:
>> On 22 Sep 2015, at 11:43, Slawa Olhovchenkov <s...@zxy.spb.ru> wrote:
>>> 
>>> On Tue, Sep 22, 2015 at 09:35:35AM +, Dimitry Andric wrote:
>> ...
>>>> -imm |= (-1 << 7);
>>>> +imm |= -(1 << 7);
>>> 
>>> May be (~0 << 7) is more simple to understund?
>> 
>> This will give the same warning.  ~0 will implicitly convert to -1.
>> 
>> (It would be better to convert all these masks to unsigned, where such
>> shifting is always defined, but that gives a lot more churn.)
>> 
>> -Dimitry
>> 
> 
> How is (-1U << 7) more churn?

For this one line, it isn't, of course.  Writing -(1 << 7) is simply my
personal preference, and it matches the type of 'imm'.  But what I meant
was changing all mask-containing int variables to unsigned, and
suffixing all mask literals U, and so on.  This is what should have been
done by GNU in the first place. :)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r288099 - head/contrib/binutils/opcodes

[Dimitry Andric]

Author: dim
Date: Tue Sep 22 09:35:35 2015
New Revision: 288099
URL: https://svnweb.freebsd.org/changeset/base/288099

Log:
  In binutils' arm-dis.c, avoid left-shifting a negative number.
  
  Submitted by: dan.mcgregor_usask.ca (Dan McGregor)
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D3376

Modified:
  head/contrib/binutils/opcodes/arm-dis.c

Modified: head/contrib/binutils/opcodes/arm-dis.c
==
--- head/contrib/binutils/opcodes/arm-dis.c Tue Sep 22 07:40:55 2015
(r288098)
+++ head/contrib/binutils/opcodes/arm-dis.c Tue Sep 22 09:35:35 2015
(r288099)
@@ -1767,7 +1767,7 @@ print_insn_coprocessor (bfd_vma pc, stru
 
/* Is ``imm'' a negative number?  */
if (imm & 0x40)
- imm |= (-1 << 7);
+ imm |= -(1 << 7);
 
func (stream, "%d", imm);
  }
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r288099 - head/contrib/binutils/opcodes

[Dimitry Andric]

On 22 Sep 2015, at 11:43, Slawa Olhovchenkov <s...@zxy.spb.ru> wrote:
> 
> On Tue, Sep 22, 2015 at 09:35:35AM +0000, Dimitry Andric wrote:
...
>> -  imm |= (-1 << 7);
>> +  imm |= -(1 << 7);
> 
> May be (~0 << 7) is more simple to understund?

This will give the same warning.  ~0 will implicitly convert to -1.

(It would be better to convert all these masks to unsigned, where such
shifting is always defined, but that gives a lot more churn.)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r287679 - in head/contrib/libc++: . include

[Dimitry Andric]

Author: dim
Date: Fri Sep 11 17:14:58 2015
New Revision: 287679
URL: https://svnweb.freebsd.org/changeset/base/287679

Log:
  Since contrib/libc++'s ancestry was never correct, subversion 1.8 and
  higher cannot merge to it from the vendor area.
  
  Re-bootstrap the ancestry by doing (apologies to your retinas):
  * svn rm contrib/libc++
  * svn cp ^/vendor/libc++/dist@276792 contrib/libc++
  * svn cp ^/head/contrib/libc++/FREEBSD-upgrade@287629 contrib/libc++
  * svn cp ^/head/contrib/libc++/include/__bit_reference@287629 
contrib/libc++/include
  * svn cp ^/head/contrib/libc++/include/__config@287629 contrib/libc++/include
  * svn cp ^/head/contrib/libc++/include/__tree@287629 contrib/libc++/include
  * svn cp ^/head/contrib/libc++/include/algorithm@287629 contrib/libc++/include
  * svn cp ^/head/contrib/libc++/include/stdexcept@287629 contrib/libc++/include
  * svn cp ^/head/contrib/libc++/include/type_traits@287629 
contrib/libc++/include
  * svn pd -R svn:mergeinfo contrib/libc++
  * svn ps -F mergeinfo.txt contrib/libc++

Added:
  head/contrib/libc++/FREEBSD-upgrade
 - copied unchanged from r287629, head/contrib/libc++/FREEBSD-upgrade
Replaced:
 - copied from r276792, vendor/libc++/dist/
  head/contrib/libc++/include/__bit_reference
 - copied unchanged from r287629, 
head/contrib/libc++/include/__bit_reference
  head/contrib/libc++/include/__config
 - copied unchanged from r287629, head/contrib/libc++/include/__config
  head/contrib/libc++/include/__tree
 - copied unchanged from r287629, head/contrib/libc++/include/__tree
  head/contrib/libc++/include/algorithm
 - copied unchanged from r287629, head/contrib/libc++/include/algorithm
  head/contrib/libc++/include/stdexcept
 - copied unchanged from r287629, head/contrib/libc++/include/stdexcept
  head/contrib/libc++/include/type_traits
 - copied unchanged from r287629, head/contrib/libc++/include/type_traits
Directory Properties:
  head/contrib/libc++/   (props changed)

Copied: head/contrib/libc++/FREEBSD-upgrade (from r287629, 
head/contrib/libc++/FREEBSD-upgrade)
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/libc++/FREEBSD-upgrade Fri Sep 11 17:14:58 2015
(r287679, copy of r287629, head/contrib/libc++/FREEBSD-upgrade)
@@ -0,0 +1,8 @@
+$FreeBSD$
+
+The FreeBSD copy of libc++.  This contains everything from the include and 
+source directories upstream, with the exception of the support/win32 
directory, 
+which is only required for building on Windows (FreeBSD is not Windows).
+
+To update, bring in anything new in source or include that is not
+platform-dependent (unless the platform in question is FreeBSD, obviously).

Copied: head/contrib/libc++/include/__bit_reference (from r287629, 
head/contrib/libc++/include/__bit_reference)
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/libc++/include/__bit_reference Fri Sep 11 17:14:58 2015
(r287679, copy of r287629, head/contrib/libc++/include/__bit_reference)
@@ -0,0 +1,1286 @@
+// -*- C++ -*-
+//===--===//
+//
+// The LLVM Compiler Infrastructure
+//
+// This file is dual licensed under the MIT and the University of Illinois Open
+// Source Licenses. See LICENSE.TXT for details.
+//
+//===--===//
+
+#ifndef _LIBCPP___BIT_REFERENCE
+#define _LIBCPP___BIT_REFERENCE
+
+#include <__config>
+#include 
+
+#include <__undef_min_max>
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#pragma GCC system_header
+#endif
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template  class 
__bit_iterator;
+template  class __bit_const_reference;
+
+template 
+struct __has_storage_type
+{
+static const bool value = false;
+};
+
+template ::value>
+class __bit_reference
+{
+typedef typename _Cp::__storage_type__storage_type;
+typedef typename _Cp::__storage_pointer __storage_pointer;
+
+__storage_pointer __seg_;
+__storage_type__mask_;
+
+#if defined(__clang__) || defined(__IBMCPP__) || defined(_LIBCPP_MSVC)
+friend typename _Cp::__self;
+#else
+friend class _Cp::__self;
+#endif
+friend class __bit_const_reference<_Cp>;
+friend class __bit_iterator<_Cp, false>;
+public:
+_LIBCPP_INLINE_VISIBILITY operator bool() const _NOEXCEPT
+{return static_cast(*__seg_ & __mask_);}
+_LIBCPP_INLINE_VISIBILITY bool operator ~() const _NOEXCEPT
+{return !static_cast(*this);}
+
+_LIBCPP_INLINE_VISIBILITY
+__bit_reference& operator=(bool __x) _NOEXCEPT
+{
+if (__x)
+*__seg_ |= __mask_;
+else
+*__seg_ &= ~__mask_;
+return *this;
+}
+
+_LIBCPP_INLINE_VISIBILITY
+__bit_reference& 

svn commit: r287541 - head/lib/libz

[Dimitry Andric]

Author: dim
Date: Mon Sep  7 20:55:14 2015
New Revision: 287541
URL: https://svnweb.freebsd.org/changeset/base/287541

Log:
  In libz's inflateMark(), avoid left-shifting a negative integer, which
  is undefined.
  
  Reviewed by:  delphij
  Differential Revision: https://reviews.freebsd.org/D3344
  MFC after:3 days

Modified:
  head/lib/libz/inflate.c

Modified: head/lib/libz/inflate.c
==
--- head/lib/libz/inflate.c Mon Sep  7 20:05:56 2015(r287540)
+++ head/lib/libz/inflate.c Mon Sep  7 20:55:14 2015(r287541)
@@ -1504,7 +1504,7 @@ z_streamp strm;
 {
 struct inflate_state FAR *state;
 
-if (strm == Z_NULL || strm->state == Z_NULL) return -1L << 16;
+if (strm == Z_NULL || strm->state == Z_NULL) return -(1L << 16);
 state = (struct inflate_state FAR *)strm->state;
 return ((long)(state->back) << 16) +
 (state->mode == COPY ? state->length :
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287501 - head/contrib/llvm/tools/lldb

[Dimitry Andric]

Author: dim
Date: Sun Sep  6 11:48:50 2015
New Revision: 287501
URL: https://svnweb.freebsd.org/changeset/base/287501

Log:
  Update lldb's FREEBSD-Xlist to match reality.

Modified:
  head/contrib/llvm/tools/lldb/FREEBSD-Xlist

Modified: head/contrib/llvm/tools/lldb/FREEBSD-Xlist
==
--- head/contrib/llvm/tools/lldb/FREEBSD-Xlist  Sun Sep  6 11:23:01 2015
(r287500)
+++ head/contrib/llvm/tools/lldb/FREEBSD-Xlist  Sun Sep  6 11:48:50 2015
(r287501)
@@ -3,6 +3,7 @@
 .clang-format
 .gitignore
 CMakeLists.txt
+CODE_OWNERS.txt
 INSTALL.txt
 Makefile
 cmake/
@@ -27,6 +28,7 @@ include/lldb/Host/msvc/
 include/lldb/Host/windows/
 include/lldb/Makefile
 lib/
+lit/
 lldb.xcodeproj/
 lldb.xcworkspace/
 resources/
@@ -46,12 +48,15 @@ source/Expression/CMakeLists.txt
 source/Expression/Makefile
 source/Host/CMakeLists.txt
 source/Host/Makefile
+source/Host/android/
 source/Host/common/Makefile
 source/Host/freebsd/Makefile
 source/Host/linux/
 source/Host/macosx/
 source/Host/posix/Makefile
 source/Host/windows/
+source/Initialization/CMakeLists.txt
+source/Initialization/Makefile
 source/Interpreter/CMakeLists.txt
 source/Interpreter/Makefile
 source/Makefile
@@ -62,8 +67,18 @@ source/Plugins/ABI/MacOSX-arm64/CMakeLis
 source/Plugins/ABI/MacOSX-arm64/Makefile
 source/Plugins/ABI/MacOSX-i386/CMakeLists.txt
 source/Plugins/ABI/MacOSX-i386/Makefile
+source/Plugins/ABI/SysV-arm/CMakeLists.txt
+source/Plugins/ABI/SysV-arm/Makefile
+source/Plugins/ABI/SysV-arm64/CMakeLists.txt
+source/Plugins/ABI/SysV-arm64/Makefile
 source/Plugins/ABI/SysV-hexagon/CMakeLists.txt
 source/Plugins/ABI/SysV-hexagon/Makefile
+source/Plugins/ABI/SysV-i386/CMakeLists.txt
+source/Plugins/ABI/SysV-i386/Makefile
+source/Plugins/ABI/SysV-mips/CMakeLists.txt
+source/Plugins/ABI/SysV-mips/Makefile
+source/Plugins/ABI/SysV-mips64/CMakeLists.txt
+source/Plugins/ABI/SysV-mips64/Makefile
 source/Plugins/ABI/SysV-ppc/CMakeLists.txt
 source/Plugins/ABI/SysV-ppc/Makefile
 source/Plugins/ABI/SysV-ppc64/CMakeLists.txt
@@ -88,6 +103,10 @@ source/Plugins/Instruction/ARM/Makefile
 source/Plugins/Instruction/ARM64/CMakeLists.txt
 source/Plugins/Instruction/ARM64/Makefile
 source/Plugins/Instruction/CMakeLists.txt
+source/Plugins/Instruction/MIPS/CMakeLists.txt
+source/Plugins/Instruction/MIPS/Makefile
+source/Plugins/Instruction/MIPS64/CMakeLists.txt
+source/Plugins/Instruction/MIPS64/Makefile
 source/Plugins/InstrumentationRuntime/AddressSanitizer/CMakeLists.txt
 source/Plugins/InstrumentationRuntime/AddressSanitizer/Makefile
 source/Plugins/InstrumentationRuntime/CMakeLists.txt
@@ -99,6 +118,9 @@ source/Plugins/LanguageRuntime/CPlusPlus
 source/Plugins/LanguageRuntime/CPlusPlus/ItaniumABI/CMakeLists.txt
 source/Plugins/LanguageRuntime/CPlusPlus/ItaniumABI/Makefile
 source/Plugins/LanguageRuntime/ObjC/
+source/Plugins/LanguageRuntime/RenderScript/CMakeLists.txt
+source/Plugins/LanguageRuntime/RenderScript/RenderScriptRuntime/CMakeLists.txt
+source/Plugins/LanguageRuntime/RenderScript/RenderScriptRuntime/Makefile
 source/Plugins/Makefile
 source/Plugins/MemoryHistory/CMakeLists.txt
 source/Plugins/MemoryHistory/asan/CMakeLists.txt
@@ -117,6 +139,7 @@ source/Plugins/ObjectFile/PECOFF/
 source/Plugins/OperatingSystem/CMakeLists.txt
 source/Plugins/OperatingSystem/Python/CMakeLists.txt
 source/Plugins/OperatingSystem/Python/Makefile
+source/Plugins/Platform/Android/
 source/Plugins/Platform/CMakeLists.txt
 source/Plugins/Platform/FreeBSD/CMakeLists.txt
 source/Plugins/Platform/FreeBSD/Makefile
@@ -168,6 +191,7 @@ source/Utility/Makefile
 test/
 tools/CMakeLists.txt
 tools/Makefile
+tools/argdumper/CMakeLists.txt
 tools/darwin-debug/
 tools/darwin-threads/
 tools/debugserver/
@@ -180,7 +204,9 @@ tools/lldb-mi/CMakeLists.txt
 tools/lldb-mi/Makefile
 tools/lldb-mi/lldb-Info.plist
 tools/lldb-perf/
-tools/lldb-platform/CMakeLists.txt
-tools/lldb-platform/Makefile
+tools/lldb-platform/
+tools/lldb-server/CMakeLists.txt
+tools/lldb-server/Makefile
+unittests/
 utils/
 www/
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287007 - head/share/mk

[Dimitry Andric]

On 21 Aug 2015, at 21:51, Warner Losh i...@freebsd.org wrote:
 
 Author: imp
 Date: Fri Aug 21 19:51:27 2015
 New Revision: 287007
 URL: https://svnweb.freebsd.org/changeset/base/287007
 
 Log:
  Document bsd.compiler.mk and the variables it defines.

Thanks!!



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r286733 - head/sys/dev/bxe

[Dimitry Andric]

Author: dim
Date: Thu Aug 13 18:24:41 2015
New Revision: 286733
URL: https://svnweb.freebsd.org/changeset/base/286733

Log:
  Avoid left-shifting negative signed values in bxe(4).
  
  Reviewed by:  davidcs
  MFC after:3 days

Modified:
  head/sys/dev/bxe/ecore_hsi.h

Modified: head/sys/dev/bxe/ecore_hsi.h
==
--- head/sys/dev/bxe/ecore_hsi.hThu Aug 13 18:01:50 2015
(r286732)
+++ head/sys/dev/bxe/ecore_hsi.hThu Aug 13 18:24:41 2015
(r286733)
@@ -2536,9 +2536,9 @@ struct shmem2_region {
#define SHMEM_EEE_SUPPORTED_MASK   0x000f
#define SHMEM_EEE_SUPPORTED_SHIFT  16
#define SHMEM_EEE_ADV_STATUS_MASK  0x00f0
-   #define SHMEM_EEE_100M_ADV (10)
-   #define SHMEM_EEE_1G_ADV   (11)
-   #define SHMEM_EEE_10G_ADV  (12)
+   #define SHMEM_EEE_100M_ADV (1U0)
+   #define SHMEM_EEE_1G_ADV   (1U1)
+   #define SHMEM_EEE_10G_ADV  (1U2)
#define SHMEM_EEE_ADV_STATUS_SHIFT 20
#define SHMEM_EEE_LP_ADV_STATUS_MASK   0x0f00
#define SHMEM_EEE_LP_ADV_STATUS_SHIFT  24
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r286699 - head/contrib/gcclibs/libcpp

[Dimitry Andric]

Author: dim
Date: Wed Aug 12 20:16:13 2015
New Revision: 286699
URL: https://svnweb.freebsd.org/changeset/base/286699

Log:
  In gcc's libcpp, stop using the INTTYPE_MAXIMUM() macro, which relies on
  undefined behavior.  The code used this macro to avoid problems on some
  broken systems which define SSIZE_MAX incorrectly, but this is not
  needed on FreeBSD, obviously.
  
  MFC after: 3 days

Modified:
  head/contrib/gcclibs/libcpp/files.c

Modified: head/contrib/gcclibs/libcpp/files.c
==
--- head/contrib/gcclibs/libcpp/files.c Wed Aug 12 20:08:54 2015
(r286698)
+++ head/contrib/gcclibs/libcpp/files.c Wed Aug 12 20:16:13 2015
(r286699)
@@ -567,7 +567,7 @@ read_file_guts (cpp_reader *pfile, _cpp_
 SSIZE_MAX to be much smaller than the actual range of the
 type.  Use INTTYPE_MAXIMUM unconditionally to ensure this
 does not bite us.  */
-  if (file-st.st_size  INTTYPE_MAXIMUM (ssize_t))
+  if (file-st.st_size  SSIZE_MAX)
{
  cpp_error (pfile, CPP_DL_ERROR, %s is too large, file-path);
  return false;
@@ -581,7 +581,7 @@ read_file_guts (cpp_reader *pfile, _cpp_
file-path);
  return false;
}
-  else if (offset  INTTYPE_MAXIMUM (ssize_t) || (ssize_t)offset  size)
+  else if (offset  SSIZE_MAX || (ssize_t)offset  size)
{
  cpp_error (pfile, CPP_DL_ERROR, current position of %s is too large,
file-path);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r286702 - head/sbin/ipfw

[Dimitry Andric]

Author: dim
Date: Wed Aug 12 21:07:57 2015
New Revision: 286702
URL: https://svnweb.freebsd.org/changeset/base/286702

Log:
  In ipfw2, avoid left-shifting negative integers, which is undefined.
  While here, make some other arguments to htonl(3) unsigned too.
  
  MFC after:3 days

Modified:
  head/sbin/ipfw/ipfw2.c

Modified: head/sbin/ipfw/ipfw2.c
==
--- head/sbin/ipfw/ipfw2.c  Wed Aug 12 20:50:20 2015(r286701)
+++ head/sbin/ipfw/ipfw2.c  Wed Aug 12 21:07:57 2015(r286702)
@@ -2869,14 +2869,14 @@ fill_ip(ipfw_insn_ip *cmd, char *av, int
case '/':
masklen = atoi(p);
if (masklen == 0)
-   d[1] = htonl(0);/* mask */
+   d[1] = htonl(0U);   /* mask */
else if (masklen  32)
errx(EX_DATAERR, bad width ``%s'', p);
else
-   d[1] = htonl(~0  (32 - masklen));
+   d[1] = htonl(~0U  (32 - masklen));
break;
case '{':   /* no mask, assume /24 and put back the '{' */
-   d[1] = htonl(~0  (32 - 24));
+   d[1] = htonl(~0U  (32 - 24));
*(--p) = md;
break;
 
@@ -2885,7 +2885,7 @@ fill_ip(ipfw_insn_ip *cmd, char *av, int
/* FALLTHROUGH */
case 0: /* initialization value */
default:
-   d[1] = htonl(~0);   /* force /32 */
+   d[1] = htonl(~0U);  /* force /32 */
break;
}
d[0] = d[1];   /* mask base address with mask */
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r286515 - head/lib/msun/src

[Dimitry Andric]

Author: dim
Date: Sun Aug  9 10:00:13 2015
New Revision: 286515
URL: https://svnweb.freebsd.org/changeset/base/286515

Log:
  In libm's exp2(3), avoid left-shifting a negative integer, which is
  undefined.  Replace it with the intended value, in a defined way.
  
  Reviewed by:  bde
  MFC after:3 days

Modified:
  head/lib/msun/src/s_exp2.c

Modified: head/lib/msun/src/s_exp2.c
==
--- head/lib/msun/src/s_exp2.c  Sun Aug  9 09:54:29 2015(r286514)
+++ head/lib/msun/src/s_exp2.c  Sun Aug  9 10:00:13 2015(r286515)
@@ -376,14 +376,14 @@ exp2(double x)
/* Compute r = exp2(y) = exp2t[i0] * p(z - eps[i]). */
t = tbl[i0];/* exp2t[i0] */
z -= tbl[i0 + 1];   /* eps[i0]   */
-   if (k = -1021  20)
+   if (k = -(1021  20))
INSERT_WORDS(twopk, 0x3ff0 + k, 0);
else
INSERT_WORDS(twopkp1000, 0x3ff0 + k + (1000  20), 0);
r = t + t * z * (P1 + z * (P2 + z * (P3 + z * (P4 + z * P5;
 
/* Scale by 2**(k20). */
-   if(k = -1021  20) {
+   if(k = -(1021  20)) {
if (k == 1024  20)
return (r * 2.0 * 0x1p1023);
return (r * twopk);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r286519 - head/contrib/binutils/gas/config

[Dimitry Andric]

Author: dim
Date: Sun Aug  9 11:06:40 2015
New Revision: 286519
URL: https://svnweb.freebsd.org/changeset/base/286519

Log:
  In GNU as, avoid left-shifting negative integers, which is undefined.
  
  MFC after:3 days

Modified:
  head/contrib/binutils/gas/config/tc-i386.c

Modified: head/contrib/binutils/gas/config/tc-i386.c
==
--- head/contrib/binutils/gas/config/tc-i386.c  Sun Aug  9 10:36:25 2015
(r286518)
+++ head/contrib/binutils/gas/config/tc-i386.c  Sun Aug  9 11:06:40 2015
(r286519)
@@ -914,8 +914,8 @@ fits_in_signed_long (offsetT num ATTRIBU
 #ifndef BFD64
   return 1;
 #else
-  return (!(((offsetT) -1  31)  num)
- || (((offsetT) -1  31)  num) == ((offsetT) -1  31));
+  return (!(-((offsetT) 1  31)  num)
+ || (-((offsetT) 1  31)  num) == -((offsetT) 1  31));
 #endif
 }  /* fits_in_signed_long() */
 
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r285644 - head/contrib/sqlite3

[Dimitry Andric]

On 18 Jul 2015, at 00:53, Pedro Giffuni p...@freebsd.org wrote:
 
 On 07/17/15 17:26, Peter Jeremy wrote:
 On 2015-Jul-16 22:07:14 +, Pedro F. Giffuni p...@freebsd.org wrote:
 Log:
 ...
  sqlite: clean a couple of invocations of memcpy(3)
Found almost accidentally by our native gcc when enhanced with
  FORTIFY_SOURCE.
 ...
 -  memcpy((void *)aHdr[1], (void *)pWal-hdr, sizeof(WalIndexHdr));
 +  memcpy((void *)aHdr[1], (const void *)pWal-hdr, sizeof(WalIndexHdr));
   walShmBarrier(pWal);
 -  memcpy((void *)aHdr[0], (void *)pWal-hdr, sizeof(WalIndexHdr));
 +  memcpy((void *)aHdr[0], (const void *)pWal-hdr, sizeof(WalIndexHdr));
 If the compiler complained about that, the compiler is broken.
 
 The change was rather collateral (read cosmetical) to the real warning
 which remains unfixed.
 
 The compiler warning is this:
 ...
 === lib/libsqlite3 (obj,depend,all,install)
 cc1: warnings being treated as errors
 /scratch/tmp/pfg/head/lib/libsqlite3/../../contrib/sqlite3/sqlite3.c: In 
 function 'walIndexWriteHdr':
 /scratch/tmp/pfg/head/lib/libsqlite3/../../contrib/sqlite3/sqlite3.c:49490: 
 warning: passing argument 1 of 'memcpy' discards qualifiers from pointer 
 target type
 /scratch/tmp/pfg/head/lib/libsqlite3/../../contrib/sqlite3/sqlite3.c:49492: 
 warning: passing argument 1 of 'memcpy' discards qualifiers from pointer 
 target type
 --- sqlite3.So ---
 *** [sqlite3.So] Error code 1
 ...
 
 make[6]: stopped in /scratch/tmp/pfg/head/lib/libsqlite3
 
 It only happens when using the experimental FORTIFY_SOURCE support [1],
 and it only happens with gcc (the base one is the only tested).
 
 Yes, I am suspecting a compiler bug but gcc has been really useful to find
 bugs in fortify_source.

Actually, it is right to warn about this, as the casts drop a volatile
specifier:

 49482  static void walIndexWriteHdr(Wal *pWal){
 49483volatile WalIndexHdr *aHdr = walIndexHdr(pWal);
...
 49490memcpy((void *)aHdr[1], (void *)pWal-hdr, sizeof(WalIndexHdr));
 49491walShmBarrier(pWal);
 49492memcpy((void *)aHdr[0], (void *)pWal-hdr, sizeof(WalIndexHdr));

If you compile this with clang, and a high enough WARNS level, it tells
you where the problem is:

contrib/sqlite3/sqlite3.c:49490:18: error: cast from 'volatile struct 
WalIndexHdr *' to 'void *' drops volatile qualifier [-Werror,-Wcast-qual]
  memcpy((void *)aHdr[1], (void *)pWal-hdr, sizeof(WalIndexHdr));
 ^
contrib/sqlite3/sqlite3.c:49492:18: error: cast from 'volatile struct 
WalIndexHdr *' to 'void *' drops volatile qualifier [-Werror,-Wcast-qual]
  memcpy((void *)aHdr[0], (void *)pWal-hdr, sizeof(WalIndexHdr));
 ^

Similar with newer versions of gcc:

contrib/sqlite3/sqlite3.c:49490:10: error: cast discards 'volatile' qualifier 
from pointer target type [-Werror=cast-qual]
   memcpy((void *)aHdr[1], (void *)pWal-hdr, sizeof(WalIndexHdr));
  ^
contrib/sqlite3/sqlite3.c:49492:10: error: cast discards 'volatile' qualifier 
from pointer target type [-Werror=cast-qual]
   memcpy((void *)aHdr[0], (void *)pWal-hdr, sizeof(WalIndexHdr));
  ^

Removing all the casts does not help, with base gcc it still gives:

contrib/sqlite3/sqlite3.c:49490: warning: passing argument 1 of 'memcpy' 
discards qualifiers from pointer target type
contrib/sqlite3/sqlite3.c:49492: warning: passing argument 1 of 'memcpy' 
discards qualifiers from pointer target type

and with clang:

contrib/sqlite3/sqlite3.c:49490:10: error: passing 'volatile WalIndexHdr *' 
(aka 'volatile struct WalIndexHdr *') to parameter of type 'void *' discards 
qualifiers
  [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
  memcpy(aHdr[1], pWal-hdr, sizeof(WalIndexHdr));
 ^~~~
/usr/include/string.h:62:31: note: passing argument to parameter here
void*memcpy(void * __restrict, const void * __restrict, size_t);
 ^
contrib/sqlite3/sqlite3.c:49492:10: error: passing 'volatile WalIndexHdr *' 
(aka 'volatile struct WalIndexHdr *') to parameter of type 'void *' discards 
qualifiers
  [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
  memcpy(aHdr[0], pWal-hdr, sizeof(WalIndexHdr));
 ^~~~
/usr/include/string.h:62:31: note: passing argument to parameter here
void*memcpy(void * __restrict, const void * __restrict, size_t);
 ^

  'const'
 is a promise to the caller that the given parameter will not be modified
 by the callee.  There's no requirement that the passed argument be const.
 As bde commented, the casts are all spurious.
 
 And since this is contrib'ed code I am not going through the
 code churn but upstream has been notified.

I don't think it is likely that upstream will go through the trouble of
removing -Wcast-qual warnings.  There are 152 of them in sqlite3.c. :-)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r285340 - head/sys/dev/cxgb/ulp/iw_cxgb

[Dimitry Andric]

Author: dim
Date: Thu Jul  9 22:13:23 2015
New Revision: 285340
URL: https://svnweb.freebsd.org/changeset/base/285340

Log:
  Fix swapped copyin(9) arguments in cxgb's iwch_arm_cq() function.
  Detected by clang 3.7.0 with the warning:
  
  sys/dev/cxgb/ulp/iw_cxgb/iw_cxgb_provider.c:309:18: error: variable
  'rptr' is uninitialized when used here [-Werror,-Wuninitialized]
  chp-cq.rptr = rptr;
 ^~~~
  
  MFC after:1 week

Modified:
  head/sys/dev/cxgb/ulp/iw_cxgb/iw_cxgb_provider.c

Modified: head/sys/dev/cxgb/ulp/iw_cxgb/iw_cxgb_provider.c
==
--- head/sys/dev/cxgb/ulp/iw_cxgb/iw_cxgb_provider.cThu Jul  9 21:53:40 
2015(r285339)
+++ head/sys/dev/cxgb/ulp/iw_cxgb/iw_cxgb_provider.cThu Jul  9 22:13:23 
2015(r285340)
@@ -303,7 +303,7 @@ iwch_arm_cq(struct ib_cq *ibcq, enum ib_
else
cq_op = CQ_ARM_AN;
if (chp-user_rptr_addr) {
-   if (copyin(rptr, chp-user_rptr_addr, 4))
+   if (copyin(chp-user_rptr_addr, rptr, sizeof(rptr)))
return (-EFAULT);
mtx_lock(chp-lock);
chp-cq.rptr = rptr;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r285284 - head/lib/liblzma

[Dimitry Andric]

Check whether the path starts with /usr/bin, maybe?  Normally, you would
check for the existence of a random header in a configure script.  But
from within a C source file, it's not that easy.

That said, immintrin.h is available for all usable versions of clang,
and should be available in all versions of gcc = 4.4 (at least, if I
read gcc's commit history correctly).  And gcc in base is definitely not
4.4. :-)

-Dimitry

 On 09 Jul 2015, at 00:04, Adrian Chadd adrian.ch...@gmail.com wrote:
 
 Is there a blessed way to see whether the compiler we're using is an
 external compiler, or an internal one?
 
 eg, the version check isn't enough - it's just a number. how do I know
 if it's freebsd clang versus upstream clang?
 (Or in my instance, freebsd-gcc versus upstream-gcc.)
 
 
 -a
 
 
 On 8 July 2015 at 14:09, Pedro Giffuni p...@freebsd.org wrote:
 
 
 On 07/08/15 13:36, Luigi Rizzo wrote:
 
 Author: luigi
 Date: Wed Jul  8 18:36:37 2015
 New Revision: 285284
 URL: https://svnweb.freebsd.org/changeset/base/285284
 
 Log:
   only enable immintrin when clang is used. The base gcc does not support
 it.
  Reviewed by:  delphij
 
 Modified:
   head/lib/liblzma/config.h
 
 Modified: head/lib/liblzma/config.h
 
 ==
 --- head/lib/liblzma/config.h   Wed Jul  8 18:12:24 2015(r285283)
 +++ head/lib/liblzma/config.h   Wed Jul  8 18:36:37 2015(r285284)
 @@ -150,7 +150,8 @@
  #define HAVE_ICONV 1
/* Define to 1 if you have the immintrin.h header file. */
 -#if defined(__FreeBSD__)  defined(__amd64__)
 +/* FreeBSD - only with clang because the base gcc does not support it */
 +#if defined(__clang__)  defined(__FreeBSD__)  defined(__amd64__)
  #define HAVE_IMMINTRIN_H 1
  #endif
 
 
 FWIW, gcc 4.3+ does have it so this may some undesired (but hidden)
 effect when building with an external gcc.
 
 Pedro.
 
 



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r285149 - in head/contrib/llvm/lib/Transforms: Scalar Utils

[Dimitry Andric]

Author: dim
Date: Sat Jul  4 20:07:37 2015
New Revision: 285149
URL: https://svnweb.freebsd.org/changeset/base/285149

Log:
  Pull in r241142 from upstream llvm trunk (by David Majnemer):
  
[SCCP] Turn loads of null into undef instead of zero initialized values
  
Surprisingly, this is a correctness issue: the mmx type exists for
calling convention purposes, LLVM doesn't have a zero representation for
them.
  
This partially fixes PR23999.
  
  Pull in r241143 from upstream llvm trunk (by David Majnemer):
  
[LoopUnroll] Use undef for phis with no value live
  
We would create a phi node with a zero initialized operand instead of
undef in the case where no value was originally available.  This was
problematic for x86_mmx which has no null value.
  
  These fix a Cannot create a null constant of that type! error when
  compiling the graphics/sdl2_gfx port with MMX enabled.
  
  Reported by:  amdmi3

Modified:
  head/contrib/llvm/lib/Transforms/Scalar/SCCP.cpp
  head/contrib/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp

Modified: head/contrib/llvm/lib/Transforms/Scalar/SCCP.cpp
==
--- head/contrib/llvm/lib/Transforms/Scalar/SCCP.cppSat Jul  4 19:00:38 
2015(r285148)
+++ head/contrib/llvm/lib/Transforms/Scalar/SCCP.cppSat Jul  4 20:07:37 
2015(r285149)
@@ -1054,7 +1054,7 @@ void SCCPSolver::visitLoadInst(LoadInst 
 
   // load null - null
   if (isaConstantPointerNull(Ptr)  I.getPointerAddressSpace() == 0)
-return markConstant(IV, I, Constant::getNullValue(I.getType()));
+return markConstant(IV, I, UndefValue::get(I.getType()));
 
   // Transform load (constant global) into the value loaded.
   if (GlobalVariable *GV = dyn_castGlobalVariable(Ptr)) {

Modified: head/contrib/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp
==
--- head/contrib/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cppSat Jul 
 4 19:00:38 2015(r285148)
+++ head/contrib/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cppSat Jul 
 4 20:07:37 2015(r285149)
@@ -81,7 +81,7 @@ static void ConnectProlog(Loop *L, Value
   if (L-contains(PN)) {
 NewPN-addIncoming(PN-getIncomingValueForBlock(NewPH), OrigPH);
   } else {
-NewPN-addIncoming(Constant::getNullValue(PN-getType()), OrigPH);
+NewPN-addIncoming(UndefValue::get(PN-getType()), OrigPH);
   }
 
   Value *V = PN-getIncomingValueForBlock(Latch);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r285150 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Sat Jul  4 20:09:24 2015
New Revision: 285150
URL: https://svnweb.freebsd.org/changeset/base/285150

Log:
  Add llvm patch corresponding to r285149.

Added:
  head/contrib/llvm/patches/patch-10-llvm-r241142-r241143-mmx-undef.diff

Added: head/contrib/llvm/patches/patch-10-llvm-r241142-r241143-mmx-undef.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-10-llvm-r241142-r241143-mmx-undef.diff  
Sat Jul  4 20:09:24 2015(r285150)
@@ -0,0 +1,88 @@
+Pull in r241142 from upstream llvm trunk (by David Majnemer):
+
+  [SCCP] Turn loads of null into undef instead of zero initialized values
+
+  Surprisingly, this is a correctness issue: the mmx type exists for
+  calling convention purposes, LLVM doesn't have a zero representation for
+  them.
+
+  This partially fixes PR23999.
+
+Pull in r241143 from upstream llvm trunk (by David Majnemer):
+
+  [LoopUnroll] Use undef for phis with no value live
+
+  We would create a phi node with a zero initialized operand instead of
+  undef in the case where no value was originally available.  This was
+  problematic for x86_mmx which has no null value.
+
+These fix a Cannot create a null constant of that type! error when
+compiling the graphics/sdl2_gfx port with MMX enabled.
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/285149
+
+Index: lib/Transforms/Scalar/SCCP.cpp
+===
+--- lib/Transforms/Scalar/SCCP.cpp
 lib/Transforms/Scalar/SCCP.cpp
+@@ -1054,7 +1054,7 @@
+ 
+   // load null - null
+   if (isaConstantPointerNull(Ptr)  I.getPointerAddressSpace() == 0)
+-return markConstant(IV, I, Constant::getNullValue(I.getType()));
++return markConstant(IV, I, UndefValue::get(I.getType()));
+ 
+   // Transform load (constant global) into the value loaded.
+   if (GlobalVariable *GV = dyn_castGlobalVariable(Ptr)) {
+Index: lib/Transforms/Utils/LoopUnrollRuntime.cpp
+===
+--- lib/Transforms/Utils/LoopUnrollRuntime.cpp
 lib/Transforms/Utils/LoopUnrollRuntime.cpp
+@@ -81,7 +81,7 @@
+   if (L-contains(PN)) {
+ NewPN-addIncoming(PN-getIncomingValueForBlock(NewPH), OrigPH);
+   } else {
+-NewPN-addIncoming(Constant::getNullValue(PN-getType()), OrigPH);
++NewPN-addIncoming(UndefValue::get(PN-getType()), OrigPH);
+   }
+ 
+   Value *V = PN-getIncomingValueForBlock(Latch);
+Index: test/Transforms/LoopUnroll/X86/mmx.ll
+===
+--- test/Transforms/LoopUnroll/X86/mmx.ll
 test/Transforms/LoopUnroll/X86/mmx.ll
+@@ -0,0 +1,21 @@
++; RUN: opt  %s -S -loop-unroll | FileCheck %s
++target datalayout = e-m:e-i64:64-f80:128-n8:16:32:64-S128
++target triple = x86_64-unknown-linux-gnu
++
++define x86_mmx @f() #0 {
++entry:
++  br label %for.body
++
++for.body: ; preds = %for.body, %entry
++  %phi = phi i32 [ 1, %entry ], [ %add, %for.body ]
++  %add = add i32 %phi, 1
++  %cmp = icmp eq i32 %phi, 0
++  br i1 %cmp, label %exit, label %for.body
++
++exit: ; preds = %for.body
++  %ret = phi x86_mmx [ undef, %for.body ]
++  ; CHECK: ret x86_mmx %ret
++  ret x86_mmx %ret
++}
++
++attributes #0 = { target-cpu=x86-64 }
+Index: test/Transforms/SCCP/crash.ll
+===
+--- test/Transforms/SCCP/crash.ll
 test/Transforms/SCCP/crash.ll
+@@ -27,3 +27,8 @@
+   %B = extractvalue [4 x i32] %A, 1
+   ret i32 %B
+ }
++
++define x86_mmx @test3() {
++  %load = load x86_mmx* null
++  ret x86_mmx %load
++}
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r284898 - in head: . share/mk

[Dimitry Andric]

On 28 Jun 2015, at 11:55, Baptiste Daroussin b...@freebsd.org wrote:
 
 On Sun, Jun 28, 2015 at 11:48:48AM +0200, Dimitry Andric wrote:
 On 28 Jun 2015, at 10:57, Baptiste Daroussin b...@freebsd.org wrote:
 
 On Sun, Jun 28, 2015 at 01:44:21AM -0700, NGie Cooper wrote:
 On Sun, Jun 28, 2015 at 12:49 AM, Baptiste Daroussin b...@freebsd.org 
 wrote:
 ...
 Not at all I have this on a both where I haven't yet r284898 iirc it is 
 like
 this since the beginning I do not remember seeing those ld scripts with 
 absolute
 path.
 
 $ cat /usr/lib/libc.so
 /* $FreeBSD$ */
 GROUP ( /lib/libc.so.7 /usr/lib/libc_nonshared.a 
 /usr/lib/libssp_nonshared.a )
 $ what -q /boot/GENERIC.r283337+9c333ed/kernel
 FreeBSD 11.0-CURRENT #4 r283337+9c333ed(isilon-atf): Tue May 26
 21:49:09 PDT 2015
 
 Yes you are right, I was looking at the wrong place.
 
 What is actually the perceived problem with having paths in those linker
 scripts?  If you use --sysroot, the libraries are searched relative to
 that sysroot, right?
 
 (And yes, I know our gcc's sysroot implementation is broken.  So please
 fix that instead. :-)
 
 WHat is the point in having absolute path in the linker script?

Of course, the point is to know exactly *which* libraries you are going
to link in.  E.g., those with the specified paths, and not any other.


 having an
 absolute patch (or even no path at all) will make the compiler looking in its
 search path (and respecting sysroot).

I think you meant relative path here?  I'm not sure if relative names
like ../../foo/libbar.so will work with path searches, though.


 The only case where a path is needed seems
 to be when the lib you want to link to is not in the search path. Am I missing
 something?

You're not missing anything.  Some people just seem to prefer exact
paths, while others trust in search mechanisms (and the risk that the
wrong library is accidentally picked up).

FWIW, I don't mind removing the absolute paths in these scripts, but I
was simply interested in what problems people encountered due to them.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: svn commit: r284898 - in head: . share/mk

[Dimitry Andric]

On 28 Jun 2015, at 10:57, Baptiste Daroussin b...@freebsd.org wrote:
 
 On Sun, Jun 28, 2015 at 01:44:21AM -0700, NGie Cooper wrote:
 On Sun, Jun 28, 2015 at 12:49 AM, Baptiste Daroussin b...@freebsd.org 
 wrote:
 ...
 Not at all I have this on a both where I haven't yet r284898 iirc it is like
 this since the beginning I do not remember seeing those ld scripts with 
 absolute
 path.
 
 $ cat /usr/lib/libc.so
 /* $FreeBSD$ */
 GROUP ( /lib/libc.so.7 /usr/lib/libc_nonshared.a /usr/lib/libssp_nonshared.a 
 )
 $ what -q /boot/GENERIC.r283337+9c333ed/kernel
 FreeBSD 11.0-CURRENT #4 r283337+9c333ed(isilon-atf): Tue May 26
 21:49:09 PDT 2015
 
 Yes you are right, I was looking at the wrong place.

What is actually the perceived problem with having paths in those linker
scripts?  If you use --sysroot, the libraries are searched relative to
that sysroot, right?

(And yes, I know our gcc's sysroot implementation is broken.  So please
fix that instead. :-)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r284732 - head/lib/libcxxrt

[Dimitry Andric]

Author: dim
Date: Tue Jun 23 17:54:24 2015
New Revision: 284732
URL: https://svnweb.freebsd.org/changeset/base/284732

Log:
  Add __cxa_deleted_virtual to libcxxrt's version map.
  
  This symbol can sometimes be emitted by clang++, and was not yet
  exported from libcxxrt.  Attempt to be compatible with libsupc++ by
  using the same CXXABI_1.3.6 symbol version.
  
  Reported by:  y...@rawbw.com
  PR:   200863
  Reviewed by:  emaste
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D2850

Modified:
  head/lib/libcxxrt/Version.map

Modified: head/lib/libcxxrt/Version.map
==
--- head/lib/libcxxrt/Version.map   Tue Jun 23 17:26:46 2015
(r284731)
+++ head/lib/libcxxrt/Version.map   Tue Jun 23 17:54:24 2015
(r284732)
@@ -254,6 +254,10 @@ CXXABI_1.3.1 {
 __cxa_get_exception_ptr;
 } CXXABI_1.3;
 
+CXXABI_1.3.6 {
+__cxa_deleted_virtual;
+} CXXABI_1.3.1;
+
 
 CXXRT_1.0 {
 
@@ -286,7 +290,7 @@ CXXRT_1.0 {
 __cxa_increment_exception_refcount;
 __cxa_rethrow_primary_exception;
 
-} CXXABI_1.3.1;
+} CXXABI_1.3.6;
 
 
 GLIBCXX_3.4 {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r284736 - head/sys/net

[Dimitry Andric]

Author: dim
Date: Tue Jun 23 18:48:41 2015
New Revision: 284736
URL: https://svnweb.freebsd.org/changeset/base/284736

Log:
  Fix endless recursion in sys/net/if.c's drbr_inuse_drv(), found by clang
  3.7.0.
  
  Reviewed by:  marcel

Modified:
  head/sys/net/if.c

Modified: head/sys/net/if.c
==
--- head/sys/net/if.c   Tue Jun 23 18:46:29 2015(r284735)
+++ head/sys/net/if.c   Tue Jun 23 18:48:41 2015(r284736)
@@ -4005,7 +4005,7 @@ if_setgetcounterfn(if_t ifp, if_get_coun
 int
 drbr_inuse_drv(if_t ifh, struct buf_ring *br)
 {
-   return drbr_inuse_drv(ifh, br);
+   return drbr_inuse(ifh, br);
 }
 
 struct mbuf*
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r284721 - head/contrib/elftoolchain/libdwarf

[Dimitry Andric]

Author: dim
Date: Tue Jun 23 06:42:30 2015
New Revision: 284721
URL: https://svnweb.freebsd.org/changeset/base/284721

Log:
  Fix endless recursion in dwarf_get_section_max_offsets(), found by clang
  3.7.0.
  
  Reviewed by:  emaste

Modified:
  head/contrib/elftoolchain/libdwarf/dwarf_sections.c

Modified: head/contrib/elftoolchain/libdwarf/dwarf_sections.c
==
--- head/contrib/elftoolchain/libdwarf/dwarf_sections.c Tue Jun 23 06:30:36 
2015(r284720)
+++ head/contrib/elftoolchain/libdwarf/dwarf_sections.c Tue Jun 23 06:42:30 
2015(r284721)
@@ -104,8 +104,8 @@ dwarf_get_section_max_offsets(Dwarf_Debu
 Dwarf_Unsigned *debug_ranges, Dwarf_Unsigned *debug_pubtypes)
 {
 
-   return (dwarf_get_section_max_offsets(dbg, debug_info, debug_abbrev,
+   return (dwarf_get_section_max_offsets_b(dbg, debug_info, debug_abbrev,
debug_line, debug_loc, debug_aranges, debug_macinfo,
debug_pubnames, debug_str, debug_frame, debug_ranges,
-   debug_pubtypes));
+   debug_pubtypes, NULL));
 }
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r284724 - head/sys/dev/ti

[Dimitry Andric]

Author: dim
Date: Tue Jun 23 06:59:46 2015
New Revision: 284724
URL: https://svnweb.freebsd.org/changeset/base/284724

Log:
  Fix r284722, by making it actually compile.
  
  Pointy hat to:dim

Modified:
  head/sys/dev/ti/if_ti.c

Modified: head/sys/dev/ti/if_ti.c
==
--- head/sys/dev/ti/if_ti.c Tue Jun 23 06:50:03 2015(r284723)
+++ head/sys/dev/ti/if_ti.c Tue Jun 23 06:59:46 2015(r284724)
@@ -3350,7 +3350,7 @@ ti_ifmedia_upd(struct ifnet *ifp)
 
sc = ifp-if_softc;
TI_LOCK(sc);
-   error = ti_ifmedia_upd_locked(ifp);
+   error = ti_ifmedia_upd_locked(sc);
TI_UNLOCK(sc);
 
return (error);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r284722 - head/sys/dev/ti

[Dimitry Andric]

Author: dim
Date: Tue Jun 23 06:48:02 2015
New Revision: 284722
URL: https://svnweb.freebsd.org/changeset/base/284722

Log:
  Fix endless recursion in ti(4)'s ti_ifmedia_upd(), found by clang 3.7.0.

Modified:
  head/sys/dev/ti/if_ti.c

Modified: head/sys/dev/ti/if_ti.c
==
--- head/sys/dev/ti/if_ti.c Tue Jun 23 06:42:30 2015(r284721)
+++ head/sys/dev/ti/if_ti.c Tue Jun 23 06:48:02 2015(r284722)
@@ -3350,7 +3350,7 @@ ti_ifmedia_upd(struct ifnet *ifp)
 
sc = ifp-if_softc;
TI_LOCK(sc);
-   error = ti_ifmedia_upd(ifp);
+   error = ti_ifmedia_upd_locked(ifp);
TI_UNLOCK(sc);
 
return (error);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r284644 - head/usr.sbin/fstyp

[Dimitry Andric]

On 20 Jun 2015, at 17:09, Allan Jude allanj...@freebsd.org wrote:
 
 On 2015-06-20 11:05, Benjamin Kaduk wrote:
...
 Are the warnings emitted by gcc posted somewhere so that I can convince
 myself dropping WARNS is better than changing the code to work around them?
 
 -Ben
 
 The warnings come from the ZFS contrib code, so I was under the
 impression we didn't want to fix those and diverge form upstream.
 
 via kib on irc:
 
 === usr.sbin/fstyp (all)
 cc1: warnings being treated as errors
 In file included from
 /scratch/tmp/kib/src/usr.sbin/fstyp/../../sys/cddl/contrib
 from /scratch/tmp/kib/src/usr.sbin/fstyp/zfs.c:41:
 n/fs/zfs/sys/dmu.h:638: warning: function declaration isn't a prototype
 n/fs/zfs/sys/dmu.h:639: warning: function declaration isn't a prototype

Those functions are declared like this, without any parameters, as if
they were KR functions:

void xuio_stat_wbuf_copied();
void xuio_stat_wbuf_nocopy();

and also their definitions lack any parameters:

void
xuio_stat_wbuf_copied()
{
XUIOSTAT_BUMP(xuiostat_wbuf_copied);
}

void
xuio_stat_wbuf_nocopy()
{
XUIOSTAT_BUMP(xuiostat_wbuf_nocopy);
}

Obviously, these should just have (void) as their parameter list.  For
such a trivial change, it should be no problem to modify it locally, and
then notify upstream.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: svn commit: r284546 - head/contrib/gcc

[Dimitry Andric]

On 18 Jun 2015, at 15:40, Adrian Chadd adr...@freebsd.org wrote:
 
 Author: adrian
 Date: Thu Jun 18 13:40:08 2015
 New Revision: 284546
 URL: https://svnweb.freebsd.org/changeset/base/284546
 
 Log:
  Fix compilation of this macro under gcc-4.9 for MIPS32.
 
  Some point after gcc-4.2 the MIPS inline assembly restrictions changed -
  =h (hi register) disappeared from the list of restrictions and can no
  longer be used.

Just for reference, this happened here (though the commit message is
completely unreadable, as usual):

https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=208414e


  So, until someone requires an assembly version of this function,
  just use a non-assembly version and let the compiler sort it out.

This is also precisely what upstream gcc has done. :)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: svn commit: r268137 - head/sys/sys

[Dimitry Andric]

On 19 Jun 2015, at 17:02, Pedro Giffuni p...@freebsd.org wrote:
 
 On 19/06/2015 05:16 a.m., David Chisnall wrote:
 I only just caught this (having seen the fallout from NetBSD doing the same 
 thing in a shipping release and the pain that it’s caused):
 
 __weak is a reserved keyword in Objective-C, please pick another name for 
 this.  This in cdefs.h makes it impossible to include any FreeBSD standard 
 headers in Objective-C programs (of which we have a couple of hundred in 
 ports) if they use any of the modern Objective-C language modes.
...
 Closely related to this, we are redefining _Noreturn, which is a reserved 
 keyword in C11.

No, sys/cdefs.h has:

   254  /*
   255   * Keywords added in C11.
   256   */
   257
   258  #if !defined(__STDC_VERSION__) || __STDC_VERSION__  201112L || 
defined(lint)
[...]
   284  #if defined(__cplusplus)  __cplusplus = 201103L
   285  #define _Noreturn   [[noreturn]]
   286  #else
   287  #define _Noreturn   __dead2
   288  #endif
[...]
   320  #endif /* __STDC_VERSION__ || __STDC_VERSION__  201112L */

So the whole block redefining all the _Xxx identifiers is skipped for
C11 and higher.

E.g.:

$ cpp -std=c99
#include sys/cdefs.h
_Noreturn void foo(void);
^D
# 1 stdin
# 1 built-in 1
# 1 built-in 3
# 306 built-in 3
# 1 command line 1
# 1 built-in 2
# 1 stdin 2
# 1 /usr/include/sys/cdefs.h 1 3 4
# 2 stdin 2
__attribute__((__noreturn__)) void foo(void);

$ cpp -std=c11
#include sys/cdefs.h
_Noreturn void foo(void);
^D
# 1 stdin
# 1 built-in 1
# 1 built-in 3
# 306 built-in 3
# 1 command line 1
# 1 built-in 2
# 1 stdin 2
# 1 /usr/include/sys/cdefs.h 1 3 4
# 2 stdin 2
_Noreturn void foo(void);

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: svn commit: r284370 - head/usr.bin/kdump

[Dimitry Andric]

On 14 Jun 2015, at 05:30, Simon J. Gerraty s...@freebsd.org wrote:
 
 Author: sjg
 Date: Sun Jun 14 03:30:39 2015
 New Revision: 284370
 URL: https://svnweb.freebsd.org/changeset/base/284370
 
 Log:
  Create proper targets for linux*syscalls.c
 
 Modified:
  head/usr.bin/kdump/Makefile

Is this supposed to fix things?  Even at r284381, I get:

=== usr.bin/kdump (depend)
sh /usr/src/usr.bin/kdump/mksubr /usr/obj/usr/src/tmp/usr/include kdump_subr.c
env MACHINE=i386 CPP=cpp   sh /usr/src/usr.bin/kdump/mkioctls print 
/usr/obj/usr/src/tmp/usr/include  ioctl.c
stdin:1:10: fatal error: 'cam/cam_compat.h' file not found
#include cam/cam_compat.h
 ^
1 error generated.
rm -f .depend
CC='cc  ' mkdep -f .depend -a-I/usr/src/usr.bin/kdump/../ktrace 
-I/usr/src/usr.bin/kdump -I/usr/src/usr.bin/kdump/../.. -I. -DHAVE_LIBCAPSICUM 
-DPF -std=gnu99   kdump_subr.c /usr/src/usr.bin/kdump/kdump.c ioctl.c 
/usr/src/usr.bin/kdump/../ktrace/subr.c
/usr/src/usr.bin/kdump/kdump.c:153:10: fatal error: 'linux_syscalls.c' file not 
found
#include linux_syscalls.c
 ^
1 error generated.
mkdep: compile failed
*** Error code 1

Stop.

Any idea?  I obviously do not see any linux_syscalls.c in the objdir.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: svn commit: r284345 - in head: . bin/cat bin/chflags bin/chio bin/chmod bin/cp bin/csh bin/date bin/dd bin/df bin/domainname bin/echo bin/ed bin/expr bin/freebsd-version bin/getfacl bin/hostname b

[Dimitry Andric]

On 13 Jun 2015, at 21:20, Simon J. Gerraty s...@freebsd.org wrote:
 
 Author: sjg
 Date: Sat Jun 13 19:20:56 2015
 New Revision: 284345
 URL: https://svnweb.freebsd.org/changeset/base/284345
 
 Log:
  Add META_MODE support.
 
  Off by default, build behaves normally.
  WITH_META_MODE we get auto objdir creation, the ability to
  start build from anywhere in the tree.
 
  Still need to add real targets under targets/ to build packages.
 
  Differential Revision:   D2796
  Reviewed by: brooks imp
 
 Added:
  head/bin/cat/Makefile.depend   (contents, props changed)
  head/bin/chflags/Makefile.depend   (contents, props changed)

Some questions about this commit, which I'm sure more people must have:
* Was it really necessary to commit hundreds of clearly generated files?
* Couldn't these be generated on the fly, or with some make depend like 
command?
* How to update these files, if you change anything in the 'real' Makefiles?
* Or are you now the maintainer of these .depend files? :-)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r284346 - head/lib/libfetch

[Dimitry Andric]

Author: dim
Date: Sat Jun 13 19:26:48 2015
New Revision: 284346
URL: https://svnweb.freebsd.org/changeset/base/284346

Log:
  Fix the following clang 3.7.0 warnings in lib/libfetch/http.c:
  
  lib/libfetch/http.c:1628:26: error: address of array 'purl-user'
  will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  aparams.user = purl-user ?
 ~~^~~~ ~
  lib/libfetch/http.c:1630:30: error: address of array 'purl-pwd'
  will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  aparams.password = purl-pwd?
 ~~^~~~
  lib/libfetch/http.c:1657:25: error: address of array 'url-user'
  will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  aparams.user = url-user ?
 ~^~~~ ~
  lib/libfetch/http.c:1659:29: error: address of array 'url-pwd'
  will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  aparams.password = url-pwd ?
 ~^~~ ~
  lib/libfetch/http.c:1669:25: error: address of array 'url-user'
  will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  aparams.user = url-user ?
 ~^~~~ ~
  lib/libfetch/http.c:1671:29: error: address of array 'url-pwd'
  will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  aparams.password = url-pwd ?
 ~^~~ ~
  
  Since url-user and url-pwd are arrays, they can never be NULL, so the
  checks can be removed.
  
  Reviewed by:  bapt
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D2673

Modified:
  head/lib/libfetch/http.c

Modified: head/lib/libfetch/http.c
==
--- head/lib/libfetch/http.cSat Jun 13 19:20:56 2015(r284345)
+++ head/lib/libfetch/http.cSat Jun 13 19:26:48 2015(r284346)
@@ -1625,10 +1625,8 @@ http_request_body(struct url *URL, const
http_auth_params_t aparams;
init_http_auth_params(aparams);
if (*purl-user || *purl-pwd) {
-   aparams.user = purl-user ?
-   strdup(purl-user) : strdup();
-   aparams.password = purl-pwd?
-   strdup(purl-pwd) : strdup();
+   aparams.user = strdup(purl-user);
+   aparams.password = strdup(purl-pwd);
} else if ((p = getenv(HTTP_PROXY_AUTH)) != NULL 
   *p != '\0') {
if (http_authfromenv(p, aparams)  0) {
@@ -1654,10 +1652,8 @@ http_request_body(struct url *URL, const
http_auth_params_t aparams;
init_http_auth_params(aparams);
if (*url-user || *url-pwd) {
-   aparams.user = url-user ?
-   strdup(url-user) : strdup();
-   aparams.password = url-pwd ?
-   strdup(url-pwd) : strdup();
+   aparams.user = strdup(url-user);
+   aparams.password = strdup(url-pwd);
} else if ((p = getenv(HTTP_AUTH)) != NULL 
   *p != '\0') {
if (http_authfromenv(p, aparams)  0) {
@@ -1666,10 +1662,8 @@ http_request_body(struct url *URL, const
}
} else if (fetchAuthMethod 
   fetchAuthMethod(url) == 0) {
-   aparams.user = url-user ?
-   strdup(url-user) : strdup();
-   aparams.password = url-pwd ?
-   strdup(url-pwd) : strdup();
+   aparams.user = strdup(url-user);
+   aparams.password = strdup(url-pwd);
} else {
http_seterr(HTTP_NEED_AUTH);
goto ouch;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r284274 - head/share/mk

[Dimitry Andric]

On 11 Jun 2015, at 18:49, Andrew Turner and...@freebsd.org wrote:
 
 Author: andrew
 Date: Thu Jun 11 16:49:14 2015
 New Revision: 284274
 URL: https://svnweb.freebsd.org/changeset/base/284274
 
 Log:
  Enable clang on armeb, it is now able to build targeting armeb. This is
  the last arm platform to move away from gcc.
 
  Tested by:   jmg

Woohoo! :-)



signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: svn commit: r284167 - head/sys/i386/i386

[Dimitry Andric]

On 09 Jun 2015, at 08:44, Bruce Evans b...@optusnet.com.au wrote:
 
 On Mon, 8 Jun 2015, Dimitry Andric wrote:
 
 Log:
 Merge r283870 from amd64:
 
 Remove unneeded NULL checks in trap_fatal().
 
 Since td_name is an array member of struct thread, it can never be NULL,
 so the check can be removed.  In addition, curproc can never be NULL,
 so remove the if statement, and splice the two printfs() together.
 
 While here, remove the u_long cast, and use the correct printf format
 specifier for curproc-p_pid.
 
 Requested by:jhb
 
 Er. I gave a longer review which implicity requested not doing all of
 this.  The format was correct (it matched the cast), and the cast was
 less wrong than not casting.

Please read https://reviews.freebsd.org/D2695, where Kostik argued
pid_t is int32_t on all arches, and I agreed with that.  The previous
obfuscation is unnecessary now.


 Both amd64/trap.c i386/trap.c still print pids portably (by casting
 to long) in one place.  They each had 2 unportable printings of pids;
 now they each have 3 unportable printings of pids.

I wasn't updating the other parts of the code, so I stayed out of there
for now.  Feel free to put a review in Phabricator to make everything
consistent.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r284167 - head/sys/i386/i386

[Dimitry Andric]

Author: dim
Date: Mon Jun  8 20:12:44 2015
New Revision: 284167
URL: https://svnweb.freebsd.org/changeset/base/284167

Log:
  Merge r283870 from amd64:
  
  Remove unneeded NULL checks in trap_fatal().
  
  Since td_name is an array member of struct thread, it can never be NULL,
  so the check can be removed.  In addition, curproc can never be NULL,
  so remove the if statement, and splice the two printfs() together.
  
  While here, remove the u_long cast, and use the correct printf format
  specifier for curproc-p_pid.
  
  Requested by: jhb
  MFC after:3 days

Modified:
  head/sys/i386/i386/trap.c

Modified: head/sys/i386/i386/trap.c
==
--- head/sys/i386/i386/trap.c   Mon Jun  8 20:03:15 2015(r284166)
+++ head/sys/i386/i386/trap.c   Mon Jun  8 20:12:44 2015(r284167)
@@ -998,12 +998,8 @@ trap_fatal(frame, eva)
if (frame-tf_eflags  PSL_VM)
printf(vm86, );
printf(IOPL = %d\n, (frame-tf_eflags  PSL_IOPL)  12);
-   printf(current process = );
-   if (curproc) {
-   printf(%lu (%s)\n, (u_long)curproc-p_pid, 
curthread-td_name);
-   } else {
-   printf(Idle\n);
-   }
+   printf(current process = %d (%s)\n,
+   curproc-p_pid, curthread-td_name);
 
 #ifdef KDB
if (debugger_on_panic || kdb_active) {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r283526 - in head: . contrib/llvm/include/llvm/Target contrib/llvm/lib/Analysis contrib/llvm/lib/CodeGen contrib/llvm/lib/CodeGen/SelectionDAG contrib/llvm/lib/ExecutionEngine/RuntimeDy...

[Dimitry Andric]

Author: dim
Date: Mon May 25 13:43:03 2015
New Revision: 283526
URL: https://svnweb.freebsd.org/changeset/base/283526

Log:
  Upgrade our copy of clang and llvm to 3.6.1 release.
  
  This release contains the following cherry-picked revisions from
  upstream trunk:
  
226124 226151 226164 226165 226166 226407 226408 226409 226652
226905 226983 227084 227087 227089 227208 227209 227210 227211
227212 227213 227214 227269 227430 227482 227503 227519 227574
227822 227986 227987 227988 227989 227990 228037 228038 228039
228040 228188 228189 228190 228273 228372 228373 228374 228403
228765 228848 228918 229223 229225 229226 229227 229228 229230
229234 229235 229236 229238 229239 229413 229507 229680 229750
229751 229752 229911 230146 230147 230235 230253 230255 230469
230500 230564 230603 230657 230742 230748 230956 231219 231237
231245 231259 231280 231451 231563 231601 231658 231659 231662
231984 231986 232046 232085 232142 232176 232179 232189 232382
232386 232389 232425 232438 232443 232675 232786 232797 232943
232957 233075 233080 233351 233353 233409 233410 233508 233584
233819 233904 234629 234636 234891 234975 234977 235524 235641
235662 235931 236099 236306 236307
  
  Please note that from 3.5.0 onwards, clang and llvm require C++11
  support to build; see UPDATING for more information.

Added:
  head/contrib/llvm/patches/patch-08-llvm-r230348-arm-fix-bad-ha.diff
 - copied, changed from r283525, 
head/contrib/llvm/patches/patch-10-llvm-r230348-arm-fix-bad-ha.diff
  head/contrib/llvm/patches/patch-09-clang-r227115-constantarraytype.diff
 - copied unchanged from r283313, 
head/contrib/llvm/patches/patch-12-clang-r227115-constantarraytype.diff
Deleted:
  head/contrib/llvm/patches/patch-08-llvm-r227089-fix-mips-i128.diff
  head/contrib/llvm/patches/patch-09-llvm-r230058-indirectbrs-assert.diff
  head/contrib/llvm/patches/patch-10-llvm-r230348-arm-fix-bad-ha.diff
  head/contrib/llvm/patches/patch-11-llvm-r231227-aarch64-tls-relocs.diff
  head/contrib/llvm/patches/patch-12-clang-r227115-constantarraytype.diff
  head/contrib/llvm/patches/patch-13-llvm-r229911-uleb128-commas.diff
Modified:
  head/ObsoleteFiles.inc
  head/UPDATING
  head/contrib/llvm/include/llvm/Target/TargetCallingConv.h
  head/contrib/llvm/include/llvm/Target/TargetLowering.h
  head/contrib/llvm/lib/Analysis/ScalarEvolutionExpander.cpp
  head/contrib/llvm/lib/CodeGen/MachineCopyPropagation.cpp
  head/contrib/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
  head/contrib/llvm/lib/CodeGen/SelectionDAG/FastISel.cpp
  head/contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp
  head/contrib/llvm/lib/CodeGen/SelectionDAG/ScheduleDAGRRList.cpp
  head/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
  head/contrib/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
  head/contrib/llvm/lib/ExecutionEngine/RuntimeDyld/RuntimeDyldMachO.cpp
  head/contrib/llvm/lib/IR/ConstantFold.cpp
  head/contrib/llvm/lib/IR/GCOV.cpp
  head/contrib/llvm/lib/Support/Unix/Memory.inc
  head/contrib/llvm/lib/Support/Windows/explicit_symbols.inc
  head/contrib/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
  head/contrib/llvm/lib/Target/ARM/ARMISelLowering.cpp
  head/contrib/llvm/lib/Target/Mips/AsmParser/MipsAsmParser.cpp
  head/contrib/llvm/lib/Target/Mips/Disassembler/MipsDisassembler.cpp
  head/contrib/llvm/lib/Target/Mips/Mips.td
  head/contrib/llvm/lib/Target/Mips/Mips16InstrInfo.cpp
  head/contrib/llvm/lib/Target/Mips/Mips32r6InstrInfo.td
  head/contrib/llvm/lib/Target/Mips/MipsCCState.cpp
  head/contrib/llvm/lib/Target/Mips/MipsCallingConv.td
  head/contrib/llvm/lib/Target/Mips/MipsISelLowering.cpp
  head/contrib/llvm/lib/Target/Mips/MipsISelLowering.h
  head/contrib/llvm/lib/Target/Mips/MipsInstrFPU.td
  head/contrib/llvm/lib/Target/Mips/MipsRegisterInfo.td
  head/contrib/llvm/lib/Target/Mips/MipsSEISelDAGToDAG.cpp
  head/contrib/llvm/lib/Target/Mips/MipsSEISelLowering.cpp
  head/contrib/llvm/lib/Target/Mips/MipsSEInstrInfo.cpp
  head/contrib/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
  head/contrib/llvm/lib/Target/R600/AMDGPU.td
  head/contrib/llvm/lib/Target/R600/AMDGPUAlwaysInlinePass.cpp
  head/contrib/llvm/lib/Target/R600/AMDGPUAsmPrinter.cpp
  head/contrib/llvm/lib/Target/R600/AMDGPUISelDAGToDAG.cpp
  head/contrib/llvm/lib/Target/R600/AMDGPUISelLowering.cpp
  head/contrib/llvm/lib/Target/R600/AMDGPUInstrInfo.h
  head/contrib/llvm/lib/Target/R600/AMDGPUInstrInfo.td
  head/contrib/llvm/lib/Target/R600/AMDGPUInstructions.td
  head/contrib/llvm/lib/Target/R600/AMDGPUIntrinsics.td
  head/contrib/llvm/lib/Target/R600/AMDGPUSubtarget.cpp
  head/contrib/llvm/lib/Target/R600/AMDGPUSubtarget.h
  head/contrib/llvm/lib/Target/R600/CaymanInstructions.td
  head/contrib/llvm/lib/Target/R600/EvergreenInstructions.td
  head/contrib/llvm/lib/Target/R600/InstPrinter/AMDGPUInstPrinter.cpp
  head/contrib/llvm/lib/Target/R600/Processors.td
  

Re: svn commit: r283424 - in head/sys: amd64/linux modules/linux64

[Dimitry Andric]

On 25 May 2015, at 16:04, Chagin Dmitry dcha...@freebsd.org wrote:
 
 On Mon, May 25, 2015 at 07:48:49AM -0400, John Baldwin wrote:
 On Sunday, May 24, 2015 04:07:12 PM Dmitry Chagin wrote:
 Author: dchagin
 Date: Sun May 24 16:07:11 2015
 New Revision: 283424
 URL: https://svnweb.freebsd.org/changeset/base/283424
 
 Log:
  Add preliminary support for x86-64 Linux binaries.
...
 Second, __FreeBSD_version bump perhaps?
 dim@ bumped, is that enough?

I've added notes for both r283424 and r283526 to the porter's handbook:

https://svnweb.freebsd.org/doc?view=revisionrevision=46725

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r283254 - head/sys/sys

[Dimitry Andric]

Author: dim
Date: Thu May 21 17:40:53 2015
New Revision: 283254
URL: https://svnweb.freebsd.org/changeset/base/283254

Log:
  Ensure that the static structs emitted by the MODULE_DEPEND() and
  MODULE_VERSION() macros don't end up as .comm symbols, when all the
  version fields are zero.
  
  Normally, such symbols will end up in .bss, but for kernel module
  version objects, this can lead to garbage version numbers.
  
  Fix this by instructing the compiler to always put these structs in the
  .data segment instead.
  
  Reported by:  delphij, ae
  MFC after:1 week

Modified:
  head/sys/sys/module.h

Modified: head/sys/sys/module.h
==
--- head/sys/sys/module.h   Thu May 21 17:39:42 2015(r283253)
+++ head/sys/sys/module.h   Thu May 21 17:40:53 2015(r283254)
@@ -107,7 +107,8 @@ struct mod_metadata {
DATA_SET(modmetadata_set, _mod_metadata##uniquifier)
 
 #defineMODULE_DEPEND(module, mdepend, vmin, vpref, vmax)   
\
-   static struct mod_depend _##module##_depend_on_##mdepend = {\
+   static struct mod_depend _##module##_depend_on_##mdepend\
+   __section(.data) = {  \
vmin,   \
vpref,  \
vmax\
@@ -147,7 +148,8 @@ struct mod_metadata {
DECLARE_MODULE_WITH_MAXVER(name, data, sub, order, __FreeBSD_version)
 
 #defineMODULE_VERSION(module, version) 
\
-   static struct mod_version _##module##_version = {   \
+   static struct mod_version _##module##_version   \
+   __section(.data) = {  \
version \
};  \
MODULE_METADATA(_##module##_version, MDT_VERSION,   \
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r282994 - head/tools/build

[Dimitry Andric]

Author: dim
Date: Fri May 15 22:19:35 2015
New Revision: 282994
URL: https://svnweb.freebsd.org/changeset/base/282994

Log:
  Provide reallocarray() in -legacy, if needed, to allow building head on
  previous releases.
  
  Also add a stdlib.h wrapper, which declares the function, otherwise the
  compiler may assume it returns int, which can cause segfaults on LP64
  architectures.
  
  Reviewed by: bapt
  Differential Revision: https://reviews.freebsd.org/D2558

Added:
  head/tools/build/stdlib.h   (contents, props changed)
Modified:
  head/tools/build/Makefile

Modified: head/tools/build/Makefile
==
--- head/tools/build/Makefile   Fri May 15 22:14:42 2015(r282993)
+++ head/tools/build/Makefile   Fri May 15 22:19:35 2015(r282994)
@@ -25,6 +25,14 @@ CFLAGS+= -I${.CURDIR}/../../contrib/libc
-I${.CURDIR}/../../lib/libc/include
 .endif
 
+_WITH_REALLOCARRAY!= grep -c reallocarray /usr/include/stdlib.h || true
+.if ${_WITH_REALLOCARRAY} == 0
+.PATH: ${.CURDIR}/../../lib/libc/stdlib
+INCS+= stdlib.h
+SRCS+= reallocarray.c
+CFLAGS+=   -I${.CURDIR}/../../lib/libc/include
+.endif
+
 .if empty(SRCS)
 SRCS=  dummy.c
 .endif

Added: head/tools/build/stdlib.h
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/tools/build/stdlib.h   Fri May 15 22:19:35 2015(r282994)
@@ -0,0 +1,42 @@
+/*-
+ * Copyright (c) 2015 Dimitry Andric d...@freebsd.org
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *notice, this list of conditions and the following disclaimer in the
+ *documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef _LEGACY_STDLIB_H_
+#define_LEGACY_STDLIB_H_
+
+#include_next stdlib.h
+
+__BEGIN_DECLS
+   
+#if __BSD_VISIBLE
+void   *reallocarray(void *, size_t, size_t);
+#endif /* __BSD_VISIBLE */
+
+__END_DECLS
+
+#endif /* !_LEGACY_STDLIB_H_ */
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r281451 - head/sys/vm

[Dimitry Andric]

On 24 Apr 2015, at 13:01, Chris Torek to...@elf.torek.net wrote:
 
 The problem seems likely to be related to odd compiler handling of
 alignment.  Consider this code bit, which extracts the essentials:
 
   struct x {
   int x;
   } __attribute__((__aligned__(32)));
 
   struct s1 {
   int a;
   struct x b[1];
   };
 
   struct s2 {
   int a;
   struct x b[];
   };
 
   extern void test2(int);
   void test(void) {
   test2(sizeof(struct s1));
   test2(sizeof(struct s2));
   }
 
 Compiled, here are the two sizeof values (this particular compiler
 output is from clang but gcc and clang both agree on sizeof here):
 
   movl$64, %edi
   callq   test2
   movl$32, %edi
   popq%rbp
   jmp test2   # TAILCALL
 
 With the flexible array, (sizeof(struct uma_cache)) is going to be
 32 bytes smaller than without it.
 
 (I have not looked closely enough to determine what the size should be.)

I'm not sure I would consider this odd.  For purposes of sizeof(), the
flexible array member declaration can be thought of as:

struct x b[0];

so while both struct s1 and s2 have members that must be aligned at 32
bytes, the former has one extra, while the latter doesn't.  Ergo, 2x32
bytes and 1x32 bytes.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r281775 - head/contrib/llvm/lib/MC/MCParser

[Dimitry Andric]

Author: dim
Date: Mon Apr 20 17:36:35 2015
New Revision: 281775
URL: https://svnweb.freebsd.org/changeset/base/281775

Log:
  Pull in r229911 from upstream llvm trunk (by Benjamin Kramer):
  
MC: Allow multiple comma-separated expressions on the .uleb128 directive.
  
For compatiblity with GNU as. Binutils documents this as
'.uleb128 expressions'. Subtle, isn't it?
  
  Reported by:  sbruno
  PR:   199554
  MFC after:3 days

Modified:
  head/contrib/llvm/lib/MC/MCParser/AsmParser.cpp

Modified: head/contrib/llvm/lib/MC/MCParser/AsmParser.cpp
==
--- head/contrib/llvm/lib/MC/MCParser/AsmParser.cpp Mon Apr 20 17:30:13 
2015(r281774)
+++ head/contrib/llvm/lib/MC/MCParser/AsmParser.cpp Mon Apr 20 17:36:35 
2015(r281775)
@@ -3636,21 +3636,27 @@ bool AsmParser::parseDirectiveSpace(Stri
 }
 
 /// parseDirectiveLEB128
-/// ::= (.sleb128 | .uleb128) expression
+/// ::= (.sleb128 | .uleb128) [ expression (, expression)* ]
 bool AsmParser::parseDirectiveLEB128(bool Signed) {
   checkForValidSection();
   const MCExpr *Value;
 
-  if (parseExpression(Value))
-return true;
+  for (;;) {
+if (parseExpression(Value))
+  return true;
 
-  if (getLexer().isNot(AsmToken::EndOfStatement))
-return TokError(unexpected token in directive);
+if (Signed)
+  getStreamer().EmitSLEB128Value(Value);
+else
+  getStreamer().EmitULEB128Value(Value);
 
-  if (Signed)
-getStreamer().EmitSLEB128Value(Value);
-  else
-getStreamer().EmitULEB128Value(Value);
+if (getLexer().is(AsmToken::EndOfStatement))
+  break;
+
+if (getLexer().isNot(AsmToken::Comma))
+  return TokError(unexpected token in directive);
+Lex();
+  }
 
   return false;
 }
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r281777 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Mon Apr 20 17:37:37 2015
New Revision: 281777
URL: https://svnweb.freebsd.org/changeset/base/281777

Log:
  Add llvm patch corresponding to r281775.

Added:
  head/contrib/llvm/patches/patch-13-llvm-r229911-uleb128-commas.diff

Added: head/contrib/llvm/patches/patch-13-llvm-r229911-uleb128-commas.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-13-llvm-r229911-uleb128-commas.diff Mon Apr 
20 17:37:37 2015(r281777)
@@ -0,0 +1,77 @@
+Pull in r229911 from upstream llvm trunk (by Benjamin Kramer):
+
+  MC: Allow multiple comma-separated expressions on the .uleb128 directive.
+
+  For compatiblity with GNU as. Binutils documents this as
+  '.uleb128 expressions'. Subtle, isn't it?
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/281775
+
+Index: lib/MC/MCParser/AsmParser.cpp
+===
+--- lib/MC/MCParser/AsmParser.cpp
 lib/MC/MCParser/AsmParser.cpp
+@@ -3636,22 +3636,28 @@ bool AsmParser::parseDirectiveSpace(StringRef IDVa
+ }
+ 
+ /// parseDirectiveLEB128
+-/// ::= (.sleb128 | .uleb128) expression
++/// ::= (.sleb128 | .uleb128) [ expression (, expression)* ]
+ bool AsmParser::parseDirectiveLEB128(bool Signed) {
+   checkForValidSection();
+   const MCExpr *Value;
+ 
+-  if (parseExpression(Value))
+-return true;
++  for (;;) {
++if (parseExpression(Value))
++  return true;
+ 
+-  if (getLexer().isNot(AsmToken::EndOfStatement))
+-return TokError(unexpected token in directive);
++if (Signed)
++  getStreamer().EmitSLEB128Value(Value);
++else
++  getStreamer().EmitULEB128Value(Value);
+ 
+-  if (Signed)
+-getStreamer().EmitSLEB128Value(Value);
+-  else
+-getStreamer().EmitULEB128Value(Value);
++if (getLexer().is(AsmToken::EndOfStatement))
++  break;
+ 
++if (getLexer().isNot(AsmToken::Comma))
++  return TokError(unexpected token in directive);
++Lex();
++  }
++
+   return false;
+ }
+ 
+Index: test/MC/ELF/uleb.s
+===
+--- test/MC/ELF/uleb.s
 test/MC/ELF/uleb.s
+@@ -11,16 +11,17 @@ foo:
+   .uleb128128
+   .uleb12816383
+   .uleb12816384
++.uleb128  23, 42
+ 
+ // ELF_32:   Name: .text
+ // ELF_32:   SectionData (
+-// ELF_32: : 00017F80 01FF7F80 8001
++// ELF_32: : 00017F80 01FF7F80 8001172A
+ // ELF_32:   )
+ // ELF_64:   Name: .text
+ // ELF_64:   SectionData (
+-// ELF_64: : 00017F80 01FF7F80 8001
++// ELF_64: : 00017F80 01FF7F80 8001172A
+ // ELF_64:   )
+ // MACHO_32: ('section_name', 
'__text\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00')
+-// MACHO_32: ('_section_data', '00017f80 01ff7f80 8001')
++// MACHO_32: ('_section_data', '00017f80 01ff7f80 8001172a')
+ // MACHO_64: ('section_name', 
'__text\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00')
+-// MACHO_64: ('_section_data', '00017f80 01ff7f80 8001')
++// MACHO_64: ('_section_data', '00017f80 01ff7f80 8001172a')
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r281396 - in head/sys: contrib/dev/acpica contrib/dev/acpica/common contrib/dev/acpica/compiler contrib/dev/acpica/components/debugger contrib/dev/acpica/components/disassembler contri

[Dimitry Andric]

On 11 Apr 2015, at 05:23, Jung-uk Kim j...@freebsd.org wrote:
 
 Author: jkim
 Date: Sat Apr 11 03:23:41 2015
 New Revision: 281396
 URL: https://svnweb.freebsd.org/changeset/base/281396
 
 Log:
  Merge ACPICA 20150410.

Shall we MFC this at some point? :)

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r281047 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Fri Apr  3 18:42:38 2015
New Revision: 281047
URL: https://svnweb.freebsd.org/changeset/base/281047

Log:
  Add clang patch corresponding to r281046.

Added:
  head/contrib/llvm/patches/patch-12-clang-r227115-constantarraytype.diff

Added: head/contrib/llvm/patches/patch-12-clang-r227115-constantarraytype.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-12-clang-r227115-constantarraytype.diff 
Fri Apr  3 18:42:38 2015(r281047)
@@ -0,0 +1,50 @@
+Pull in r227115 from upstream clang trunk (by Ben Langmuir):
+
+  Fix assert instantiating string init of static variable
+
+  ... when the variable's type is a typedef of a ConstantArrayType. Just
+  look through the typedef (and any other sugar).  We only use the
+  constant array type here to get the element count.
+
+This fixes an assertion failure when building the games/redeclipse port.
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/281046
+
+Index: tools/clang/lib/Sema/SemaInit.cpp
+===
+--- tools/clang/lib/Sema/SemaInit.cpp
 tools/clang/lib/Sema/SemaInit.cpp
+@@ -149,10 +149,10 @@ static void updateStringLiteralType(Expr *E, QualT
+ static void CheckStringInit(Expr *Str, QualType DeclT, const ArrayType *AT,
+ Sema S) {
+   // Get the length of the string as parsed.
+-  uint64_t StrLength =
+-castConstantArrayType(Str-getType())-getSize().getZExtValue();
++  auto *ConstantArrayTy =
++  castConstantArrayType(Str-getType()-getUnqualifiedDesugaredType());
++  uint64_t StrLength = ConstantArrayTy-getSize().getZExtValue();
+ 
+-
+   if (const IncompleteArrayType *IAT = dyn_castIncompleteArrayType(AT)) {
+ // C99 6.7.8p14. We have an array of character type with unknown size
+ // being initialized to a string literal.
+Index: tools/clang/test/SemaTemplate/instantiate-static-var.cpp
+===
+--- tools/clang/test/SemaTemplate/instantiate-static-var.cpp
 tools/clang/test/SemaTemplate/instantiate-static-var.cpp
+@@ -114,3 +114,15 @@ namespace PR6449 {
+   template class X1char;
+ 
+ }
++
++typedef char MyString[100];
++template typename T
++struct StaticVarWithTypedefString {
++  static MyString str;
++};
++template typename T
++MyString StaticVarWithTypedefStringT::str = ;
++
++void testStaticVarWithTypedefString() {
++  (void)StaticVarWithTypedefStringint::str;
++}
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r281046 - head/contrib/llvm/tools/clang/lib/Sema

[Dimitry Andric]

Author: dim
Date: Fri Apr  3 18:38:37 2015
New Revision: 281046
URL: https://svnweb.freebsd.org/changeset/base/281046

Log:
  Pull in r227115 from upstream clang trunk (by Ben Langmuir):
  
Fix assert instantiating string init of static variable
  
... when the variable's type is a typedef of a ConstantArrayType. Just
look through the typedef (and any other sugar).  We only use the
constant array type here to get the element count.
  
  This fixes an assertion failure when building the games/redeclipse port.
  
  Reported by:  amdmi3

Modified:
  head/contrib/llvm/tools/clang/lib/Sema/SemaInit.cpp

Modified: head/contrib/llvm/tools/clang/lib/Sema/SemaInit.cpp
==
--- head/contrib/llvm/tools/clang/lib/Sema/SemaInit.cpp Fri Apr  3 18:12:11 
2015(r281045)
+++ head/contrib/llvm/tools/clang/lib/Sema/SemaInit.cpp Fri Apr  3 18:38:37 
2015(r281046)
@@ -149,9 +149,9 @@ static void updateStringLiteralType(Expr
 static void CheckStringInit(Expr *Str, QualType DeclT, const ArrayType *AT,
 Sema S) {
   // Get the length of the string as parsed.
-  uint64_t StrLength =
-castConstantArrayType(Str-getType())-getSize().getZExtValue();
-
+  auto *ConstantArrayTy =
+  castConstantArrayType(Str-getType()-getUnqualifiedDesugaredType());
+  uint64_t StrLength = ConstantArrayTy-getSize().getZExtValue();
 
   if (const IncompleteArrayType *IAT = dyn_castIncompleteArrayType(AT)) {
 // C99 6.7.8p14. We have an array of character type with unknown size
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r281049 - in head: contrib/llvm/tools/llvm-cov contrib/llvm/tools/llvm-profdata usr.bin/clang usr.bin/clang/llvm-cov usr.bin/clang/llvm-profdata

[Dimitry Andric]

Author: dim
Date: Fri Apr  3 19:43:39 2015
New Revision: 281049
URL: https://svnweb.freebsd.org/changeset/base/281049

Log:
  Add the llvm-cov and llvm-profdata tools, when WITH_CLANG_EXTRAS is
  defined.  These help with processing coverage and profile data.

Added:
  head/contrib/llvm/tools/llvm-cov/
 - copied from r280968, vendor/llvm/dist/tools/llvm-cov/
  head/contrib/llvm/tools/llvm-profdata/
 - copied from r280968, vendor/llvm/dist/tools/llvm-profdata/
  head/usr.bin/clang/llvm-cov/
  head/usr.bin/clang/llvm-cov/Makefile   (contents, props changed)
  head/usr.bin/clang/llvm-cov/llvm-cov.1   (contents, props changed)
  head/usr.bin/clang/llvm-profdata/
  head/usr.bin/clang/llvm-profdata/Makefile   (contents, props changed)
  head/usr.bin/clang/llvm-profdata/llvm-profdata.1   (contents, props changed)
Deleted:
  head/contrib/llvm/tools/llvm-cov/CMakeLists.txt
  head/contrib/llvm/tools/llvm-cov/LLVMBuild.txt
  head/contrib/llvm/tools/llvm-cov/Makefile
  head/contrib/llvm/tools/llvm-profdata/CMakeLists.txt
  head/contrib/llvm/tools/llvm-profdata/LLVMBuild.txt
  head/contrib/llvm/tools/llvm-profdata/Makefile
Modified:
  head/usr.bin/clang/Makefile

Modified: head/usr.bin/clang/Makefile
==
--- head/usr.bin/clang/Makefile Fri Apr  3 19:33:26 2015(r281048)
+++ head/usr.bin/clang/Makefile Fri Apr  3 19:43:39 2015(r281049)
@@ -13,12 +13,14 @@ SUBDIR+=bugpoint \
llvm-as \
llvm-bcanalyzer \
llvm-diff \
+   llvm-cov \
llvm-dis \
llvm-extract \
llvm-link \
llvm-mc \
llvm-nm \
llvm-objdump \
+   llvm-profdata \
llvm-rtdyld \
llvm-symbolizer \
macho-dump \

Added: head/usr.bin/clang/llvm-cov/Makefile
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/usr.bin/clang/llvm-cov/MakefileFri Apr  3 19:43:39 2015
(r281049)
@@ -0,0 +1,26 @@
+# $FreeBSD$
+
+.include bsd.own.mk
+
+PROG_CXX=llvm-cov
+
+SRCDIR=tools/llvm-cov
+SRCS=  CodeCoverage.cpp \
+   CoverageFilters.cpp \
+   CoverageReport.cpp \
+   CoverageSummary.cpp \
+   CoverageSummaryInfo.cpp \
+   SourceCoverageView.cpp \
+   TestingSupport.cpp \
+   gcov.cpp \
+   llvm-cov.cpp
+
+LIBDEPS=llvmprofiledata \
+   llvmobject \
+   llvmmcparser \
+   llvmmc \
+   llvmbitreader \
+   llvmcore \
+   llvmsupport
+
+.include ../clang.prog.mk

Added: head/usr.bin/clang/llvm-cov/llvm-cov.1
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/usr.bin/clang/llvm-cov/llvm-cov.1  Fri Apr  3 19:43:39 2015
(r281049)
@@ -0,0 +1,165 @@
+.\ $FreeBSD$
+.\ Man page generated from reStructuredText.
+.
+.TH LLVM-COV 1 2015-04-01 3.6 LLVM
+.SH NAME
+llvm-cov \- emit coverage information
+.
+.nr rst2man-indent-level 0
+.
+.de1 rstReportMargin
+\\$1 \\n[an-margin]
+level \\n[rst2man-indent-level]
+level margin: \\n[rst2man-indent\\n[rst2man-indent-level]]
+-
+\\n[rst2man-indent0]
+\\n[rst2man-indent1]
+\\n[rst2man-indent2]
+..
+.de1 INDENT
+.\ .rstReportMargin pre:
+. RS \\$1
+. nr rst2man-indent\\n[rst2man-indent-level] \\n[an-margin]
+. nr rst2man-indent-level +1
+.\ .rstReportMargin post:
+..
+.de UNINDENT
+. RE
+.\ indent \\n[an-margin]
+.\ old: \\n[rst2man-indent\\n[rst2man-indent-level]]
+.nr rst2man-indent-level -1
+.\ new: \\n[rst2man-indent\\n[rst2man-indent-level]]
+.in \\n[rst2man-indent\\n[rst2man-indent-level]]u
+..
+.SH SYNOPSIS
+.sp
+\fBllvm\-cov\fP [options] SOURCEFILE
+.SH DESCRIPTION
+.sp
+The \fBllvm\-cov\fP tool reads code coverage data files and displays the
+coverage information for a specified source file. It is compatible with the
+\fBgcov\fP tool from version 4.2 of \fBGCC\fP and may also be compatible with
+some later versions of \fBgcov\fP\.
+.sp
+To use llvm\-cov, you must first build an instrumented version of your
+application that collects coverage data as it runs. Compile with the
+\fB\-fprofile\-arcs\fP and \fB\-ftest\-coverage\fP options to add the
+instrumentation. (Alternatively, you can use the \fB\-\-coverage\fP option, 
which
+includes both of those other options.) You should compile with debugging
+information (\fB\-g\fP) and without optimization (\fB\-O0\fP); otherwise, the
+coverage data cannot be accurately mapped back to the source code.
+.sp
+At the time you compile the instrumented code, a \fB\.gcno\fP data file will 
be
+generated for each object file. These \fB\.gcno\fP files contain half of the
+coverage data. The other half of the data comes from \fB\.gcda\fP files that 
are
+generated when you run the instrumented program, with a separate \fB\.gcda\fP
+file for each object file. Each time you run the program, the execution counts
+are summed 

svn commit: r281050 - head/contrib/llvm

[Dimitry Andric]

Author: dim
Date: Fri Apr  3 19:49:39 2015
New Revision: 281050
URL: https://svnweb.freebsd.org/changeset/base/281050

Log:
  Update FREEBSD-Xlist for llvm.

Modified:
  head/contrib/llvm/FREEBSD-Xlist

Modified: head/contrib/llvm/FREEBSD-Xlist
==
--- head/contrib/llvm/FREEBSD-Xlist Fri Apr  3 19:43:39 2015
(r281049)
+++ head/contrib/llvm/FREEBSD-Xlist Fri Apr  3 19:49:39 2015
(r281050)
@@ -397,7 +397,9 @@ tools/llvm-bcanalyzer/LLVMBuild.txt
 tools/llvm-bcanalyzer/Makefile
 tools/llvm-c-test/
 tools/llvm-config/
-tools/llvm-cov/
+tools/llvm-cov/CMakeLists.txt
+tools/llvm-cov/LLVMBuild.txt
+tools/llvm-cov/Makefile
 tools/llvm-diff/CMakeLists.txt
 tools/llvm-diff/LLVMBuild.txt
 tools/llvm-diff/Makefile
@@ -424,7 +426,9 @@ tools/llvm-nm/Makefile
 tools/llvm-objdump/CMakeLists.txt
 tools/llvm-objdump/LLVMBuild.txt
 tools/llvm-objdump/Makefile
-tools/llvm-profdata/
+tools/llvm-profdata/CMakeLists.txt
+tools/llvm-profdata/LLVMBuild.txt
+tools/llvm-profdata/Makefile
 tools/llvm-readobj/CMakeLists.txt
 tools/llvm-readobj/LLVMBuild.txt
 tools/llvm-readobj/Makefile
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r280980 - in head: . lib/csu/i386-elf sys/boot/i386/boot2 sys/boot/i386/gptboot sys/boot/i386/gptzfsboot sys/boot/i386/zfsboot sys/boot/mips/beri/boot2 sys/boot/pc98/boot0 sys/boot/pc98...

[Dimitry Andric]

Author: dim
Date: Thu Apr  2 06:58:17 2015
New Revision: 280980
URL: https://svnweb.freebsd.org/changeset/base/280980

Log:
  Ensure the cross assembler, linker and objcopy are used for the build32
  stage, just like for the regular world stage.
  
  Reviewed by:  rodrigc, imp, bapt, emaste
  MFC after:1 week
  Differential Revision: https://reviews.freebsd.org/D2187

Modified:
  head/Makefile.inc1
  head/lib/csu/i386-elf/Makefile
  head/sys/boot/i386/boot2/Makefile
  head/sys/boot/i386/gptboot/Makefile
  head/sys/boot/i386/gptzfsboot/Makefile
  head/sys/boot/i386/zfsboot/Makefile
  head/sys/boot/mips/beri/boot2/Makefile
  head/sys/boot/pc98/boot0.5/Makefile
  head/sys/boot/pc98/boot0/Makefile
  head/sys/boot/pc98/boot2/Makefile

Modified: head/Makefile.inc1
==
--- head/Makefile.inc1  Thu Apr  2 03:25:35 2015(r280979)
+++ head/Makefile.inc1  Thu Apr  2 06:58:17 2015(r280980)
@@ -411,8 +411,9 @@ LIB32CPUFLAGS=  -march=${TARGET_CPUTYPE}
 LIB32WMAKEENV= MACHINE=i386 MACHINE_ARCH=i386 \
MACHINE_CPU=i686 mmx sse sse2
 LIB32WMAKEFLAGS=   \
-   AS=${AS} --32 \
-   LD=${LD} -m elf_i386_fbsd -Y P,${LIB32TMP}/usr/lib32
+   AS=${XAS} --32 \
+   LD=${XLD} -m elf_i386_fbsd -Y P,${LIB32TMP}/usr/lib32 \
+   OBJCOPY=${XOBJCOPY}
 
 .elif ${TARGET_ARCH} == powerpc64
 .if empty(TARGET_CPUTYPE)
@@ -422,7 +423,8 @@ LIB32CPUFLAGS=  -mcpu=${TARGET_CPUTYPE}
 .endif
 LIB32WMAKEENV= MACHINE=powerpc MACHINE_ARCH=powerpc
 LIB32WMAKEFLAGS=   \
-   LD=${LD} -m elf32ppc_fbsd
+   LD=${XLD} -m elf32ppc_fbsd \
+   OBJCOPY=${XOBJCOPY}
 .endif
 
 

Modified: head/lib/csu/i386-elf/Makefile
==
--- head/lib/csu/i386-elf/Makefile  Thu Apr  2 03:25:35 2015
(r280979)
+++ head/lib/csu/i386-elf/Makefile  Thu Apr  2 06:58:17 2015
(r280980)
@@ -35,7 +35,7 @@ crt1_c.o: crt1_c.s
 
 crt1.o:crt1_c.o crt1_s.o
${LD} ${_LDFLAGS} -o crt1.o -r crt1_s.o crt1_c.o
-   objcopy --localize-symbol _start1 crt1.o
+   ${OBJCOPY} --localize-symbol _start1 crt1.o
 
 Scrt1_c.s: crt1_c.c
${CC} ${CFLAGS} -fPIC -DPIC -S -o ${.TARGET} ${.CURDIR}/crt1_c.c
@@ -46,7 +46,7 @@ Scrt1_c.o: Scrt1_c.s
 
 Scrt1.o: Scrt1_c.o crt1_s.o
${LD} ${_LDFLAGS} -o Scrt1.o -r crt1_s.o Scrt1_c.o
-   objcopy --localize-symbol _start1 Scrt1.o
+   ${OBJCOPY} --localize-symbol _start1 Scrt1.o
 
 realinstall:
${INSTALL} -o ${LIBOWN} -g ${LIBGRP} -m ${LIBMODE} \

Modified: head/sys/boot/i386/boot2/Makefile
==
--- head/sys/boot/i386/boot2/Makefile   Thu Apr  2 03:25:35 2015
(r280979)
+++ head/sys/boot/i386/boot2/Makefile   Thu Apr  2 06:58:17 2015
(r280980)
@@ -61,7 +61,7 @@ boot: boot1 boot2
 CLEANFILES+=   boot1 boot1.out boot1.o
 
 boot1: boot1.out
-   objcopy -S -O binary boot1.out ${.TARGET}
+   ${OBJCOPY} -S -O binary boot1.out ${.TARGET}
 
 boot1.out: boot1.o
${LD} ${LD_FLAGS} -e start -Ttext ${ORG1} -o ${.TARGET} boot1.o
@@ -82,7 +82,7 @@ boot2.ldr:
dd if=/dev/zero of=${.TARGET} bs=512 count=1
 
 boot2.bin: boot2.out
-   objcopy -S -O binary boot2.out ${.TARGET}
+   ${OBJCOPY} -S -O binary boot2.out ${.TARGET}
 
 boot2.out: ${BTXCRT} boot2.o sio.o
${LD} ${LD_FLAGS} -Ttext ${ORG2} -o ${.TARGET} ${.ALLSRC}

Modified: head/sys/boot/i386/gptboot/Makefile
==
--- head/sys/boot/i386/gptboot/Makefile Thu Apr  2 03:25:35 2015
(r280979)
+++ head/sys/boot/i386/gptboot/Makefile Thu Apr  2 06:58:17 2015
(r280980)
@@ -55,7 +55,7 @@ gptboot: gptldr.bin gptboot.bin ${BTXKER
 CLEANFILES+=   gptldr.bin gptldr.out gptldr.o
 
 gptldr.bin: gptldr.out
-   objcopy -S -O binary gptldr.out ${.TARGET}
+   ${OBJCOPY} -S -O binary gptldr.out ${.TARGET}
 
 gptldr.out: gptldr.o
${LD} ${LD_FLAGS} -e start -Ttext ${ORG1} -o ${.TARGET} gptldr.o
@@ -64,7 +64,7 @@ CLEANFILES+=  gptboot.bin gptboot.out gpt
cons.o util.o
 
 gptboot.bin: gptboot.out
-   objcopy -S -O binary gptboot.out ${.TARGET}
+   ${OBJCOPY} -S -O binary gptboot.out ${.TARGET}
 
 gptboot.out: ${BTXCRT} gptboot.o sio.o gpt.o crc32.o drv.o cons.o util.o
${LD} ${LD_FLAGS} -Ttext ${ORG2} -o ${.TARGET} ${.ALLSRC} ${LIBSTAND}

Modified: head/sys/boot/i386/gptzfsboot/Makefile
==
--- head/sys/boot/i386/gptzfsboot/Makefile  Thu Apr  2 03:25:35 2015
(r280979)
+++ head/sys/boot/i386/gptzfsboot/Makefile  Thu Apr  2 06:58:17 2015
(r280980)
@@ -53,7 +53,7 @@ gptzfsboot: gptldr.bin gptzfsboot.bin ${
 CLEANFILES+=   gptldr.bin gptldr.out 

svn commit: r280867 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Mon Mar 30 20:23:06 2015
New Revision: 280867
URL: https://svnweb.freebsd.org/changeset/base/280867

Log:
  Add llvm patch corresponding to r280865.

Added:
  head/contrib/llvm/patches/patch-11-llvm-r231227-aarch64-tls-relocs.diff

Added: head/contrib/llvm/patches/patch-11-llvm-r231227-aarch64-tls-relocs.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-11-llvm-r231227-aarch64-tls-relocs.diff 
Mon Mar 30 20:23:06 2015(r280867)
@@ -0,0 +1,811 @@
+Pull in r231227 from upstream llvm trunk (by Kristof Beyls):
+
+  Fix PR22408 - LLVM producing AArch64 TLS relocations that GNU linkers
+  cannot handle yet.
+
+  As is described at http://llvm.org/bugs/show_bug.cgi?id=22408, the
+  GNU linkers ld.bfd and ld.gold currently only support a subset of the
+  whole range of AArch64 ELF TLS relocations. Furthermore, they assume
+  that some of the code sequences to access thread-local variables are
+  produced in a very specific sequence. When the sequence is not as the
+  linker expects, it can silently mis-relaxe/mis-optimize the
+  instructions.
+  Even if that wouldn't be the case, it's good to produce the exact
+  sequence, as that ensures that linkers can perform optimizing
+  relaxations.
+
+  This patch:
+
+  * implements support for 16MiB TLS area size instead of 4GiB TLS area
+size. Ideally clang would grow an -mtls-size option to allow
+support for both, but that's not part of this patch.
+  * by default doesn't produce local dynamic access patterns, as even
+modern ld.bfd and ld.gold linkers do not support the associated
+relocations. An option (-aarch64-elf-ldtls-generation) is added to
+enable generation of local dynamic code sequence, but is off by
+default.
+  * makes sure that the exact expected code sequence for local dynamic
+and general dynamic accesses is produced, by making use of a new
+pseudo instruction. The patch also removes two
+(AArch64ISD::TLSDESC_BLR, AArch64ISD::TLSDESC_CALL) pre-existing
+AArch64-specific pseudo SDNode instructions that are superseded by
+the new one (TLSDESC_CALLSEQ).
+
+Introduced here: https://svnweb.freebsd.org/changeset/base/280865
+
+Index: lib/Target/AArch64/AArch64AsmPrinter.cpp
+===
+--- lib/Target/AArch64/AArch64AsmPrinter.cpp
 lib/Target/AArch64/AArch64AsmPrinter.cpp
+@@ -12,6 +12,8 @@
+ //
+ 
//===--===//
+ 
++#include MCTargetDesc/AArch64AddressingModes.h
++#include MCTargetDesc/AArch64MCExpr.h
+ #include AArch64.h
+ #include AArch64MCInstLower.h
+ #include AArch64MachineFunctionInfo.h
+@@ -494,12 +496,47 @@ void AArch64AsmPrinter::EmitInstruction(const Mach
+ EmitToStreamer(OutStreamer, TmpInst);
+ return;
+   }
+-  case AArch64::TLSDESC_BLR: {
+-MCOperand Callee, Sym;
+-MCInstLowering.lowerOperand(MI-getOperand(0), Callee);
+-MCInstLowering.lowerOperand(MI-getOperand(1), Sym);
++  case AArch64::TLSDESC_CALLSEQ: {
++/// lower this to:
++///adrp  x0, :tlsdesc:var
++///ldr   x1, [x0, #:tlsdesc_lo12:var]
++///add   x0, x0, #:tlsdesc_lo12:var
++///.tlsdesccall var
++///blr   x1
++///(TPIDR_EL0 offset now in x0)
++const MachineOperand MO_Sym = MI-getOperand(0);
++MachineOperand MO_TLSDESC_LO12(MO_Sym), MO_TLSDESC(MO_Sym);
++MCOperand Sym, SymTLSDescLo12, SymTLSDesc;
++MO_TLSDESC_LO12.setTargetFlags(AArch64II::MO_TLS | AArch64II::MO_PAGEOFF |
++   AArch64II::MO_NC);
++MO_TLSDESC.setTargetFlags(AArch64II::MO_TLS | AArch64II::MO_PAGE);
++MCInstLowering.lowerOperand(MO_Sym, Sym);
++MCInstLowering.lowerOperand(MO_TLSDESC_LO12, SymTLSDescLo12);
++MCInstLowering.lowerOperand(MO_TLSDESC, SymTLSDesc);
+ 
+-// First emit a relocation-annotation. This expands to no code, but 
requests
++MCInst Adrp;
++Adrp.setOpcode(AArch64::ADRP);
++Adrp.addOperand(MCOperand::CreateReg(AArch64::X0));
++Adrp.addOperand(SymTLSDesc);
++EmitToStreamer(OutStreamer, Adrp);
++
++MCInst Ldr;
++Ldr.setOpcode(AArch64::LDRXui);
++Ldr.addOperand(MCOperand::CreateReg(AArch64::X1));
++Ldr.addOperand(MCOperand::CreateReg(AArch64::X0));
++Ldr.addOperand(SymTLSDescLo12);
++Ldr.addOperand(MCOperand::CreateImm(0));
++EmitToStreamer(OutStreamer, Ldr);
++
++MCInst Add;
++Add.setOpcode(AArch64::ADDXri);
++Add.addOperand(MCOperand::CreateReg(AArch64::X0));
++Add.addOperand(MCOperand::CreateReg(AArch64::X0));
++Add.addOperand(SymTLSDescLo12);
++Add.addOperand(MCOperand::CreateImm(AArch64_AM::getShiftValue(0)));
++EmitToStreamer(OutStreamer, Add);
++
++// Emit a relocation-annotation. This expands to no code, but requests
+ // the following 

Re: svn commit: r280865 - in head/contrib/llvm/lib/Target/AArch64: . Utils

[Dimitry Andric]

On 30 Mar 2015, at 22:01, Ed Maste ema...@freebsd.org wrote:
 
 Author: emaste
 Date: Mon Mar 30 20:01:41 2015
 New Revision: 280865
 URL: https://svnweb.freebsd.org/changeset/base/280865
 
 Log:
  llvm: Backport upstream r229195 to fix arm64 TLS relocations

Actually, this was upstream r231227:
http://llvm.org/viewvc/llvm-project?view=revisionrevision=231227

Upstream r229195 was only a temporary workaround so llvm.org's
clang-native-aarch64-full buildbot would not keep failing, and it got
reverted after r231227 went in.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r280864 - head/contrib/libc++/include

[Dimitry Andric]

Author: dim
Date: Mon Mar 30 19:36:33 2015
New Revision: 280864
URL: https://svnweb.freebsd.org/changeset/base/280864

Log:
  Pull in r233552 from upstream libc++ trunk (by Eric Fiselier):
  
[libcxx] Fix PR22771 - Support access control SFINAE in the library
version of is_convertible.
  
Summary:
Currently the conversion check does not take place in a context where
access control SFINAE is applied. This patch changes the context of
the test expression so that SFINAE occurs if access control does not
permit the conversion.
  
Related bug: https://llvm.org/bugs/show_bug.cgi?id=22771
  
Reviewers: mclow.lists, rsmith, dim
  
Reviewed By: dim
  
Subscribers: dim, rodrigc, emaste, cfe-commits
  
Differential Revision: http://reviews.llvm.org/D8461
  
  This fixes building clang, and other programs using libc++, with newer
  versions of gcc (specifically, gcc 4.8 and higher).
  
  Reported by:  rodrigc
  MFC after:1 week

Modified:
  head/contrib/libc++/include/type_traits

Modified: head/contrib/libc++/include/type_traits
==
--- head/contrib/libc++/include/type_traits Mon Mar 30 19:15:43 2015
(r280863)
+++ head/contrib/libc++/include/type_traits Mon Mar 30 19:36:33 2015
(r280864)
@@ -842,7 +842,16 @@ template class _T1, class _T2 struct _
 
 namespace __is_convertible_imp
 {
-template class _Tp char  __test(_Tp);
+template class _Tp void  __test_convert(_Tp);
+
+template class _From, class _To, class = void
+struct __is_convertible_test : public false_type {};
+
+template class _From, class _To
+struct __is_convertible_test_From, _To,
+decltype(__test_convert_To(_VSTD::declval_From())) : public true_type
+{};
+
 template class _Tp __two __test(...);
 #ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES
 template class _Tp _Tp __source();
@@ -877,10 +886,8 @@ template class _T1, class _T2,
 unsigned _T2_is_array_function_or_void = 
__is_convertible_imp::__is_array_function_or_void_T2::value
 struct __is_convertible
 : public integral_constantbool,
-#ifndef _LIBCPP_HAS_NO_RVALUE_REFERENCES
-
sizeof(__is_convertible_imp::__test_T2(__is_convertible_imp::__source_T1()))
 == 1
-#else
-
sizeof(__is_convertible_imp::__test_T2(__is_convertible_imp::__source_T1()))
 == 1
+__is_convertible_imp::__is_convertible_test_T1, _T2::value
+#if defined(_LIBCPP_HAS_NO_RVALUE_REFERENCES)
   !(!is_function_T1::value  !is_reference_T1::value  
is_reference_T2::value
(!is_consttypename remove_reference_T2::type::value
   || is_volatiletypename remove_reference_T2::type::value)
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r280784 - head/share/mk

[Dimitry Andric]

Author: dim
Date: Sat Mar 28 12:23:15 2015
New Revision: 280784
URL: https://svnweb.freebsd.org/changeset/base/280784

Log:
  Re-enable using -mllvm -enable-load-pre=false in CLANG_OPT_SMALL for
  clang versions before 3.5.0.  This should enable building head's version
  of sys/boot/i386/boot2 on 9.x and 10.x again.
  
  Reported by:  bz

Modified:
  head/share/mk/bsd.sys.mk

Modified: head/share/mk/bsd.sys.mk
==
--- head/share/mk/bsd.sys.mkSat Mar 28 10:20:20 2015(r280783)
+++ head/share/mk/bsd.sys.mkSat Mar 28 12:23:15 2015(r280784)
@@ -131,6 +131,8 @@ CLANG_OPT_SMALL= -mstack-alignment=8 -ml
 -mllvm -simplifycfg-dup-ret
 .if ${COMPILER_VERSION} = 30500
 CLANG_OPT_SMALL+= -mllvm -enable-gvn=false
+.else
+CLANG_OPT_SMALL+= -mllvm -enable-load-pre=false
 .endif
 CFLAGS.clang+=  -Qunused-arguments
 .if ${MACHINE_CPUARCH} == sparc64
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r280636 - head/include

[Dimitry Andric]

On 26 Mar 2015, at 14:20, Tijl Coosemans t...@freebsd.org wrote:
 
 On Thu, 26 Mar 2015 17:37:53 +1100 (EST) Bruce Evans b...@optusnet.com.au 
 wrote:
 On Wed, 25 Mar 2015, Pedro Giffuni wrote:
...
 The reason why I had to revert the change is actually a systematic
 bug in gcc: during it's build process gcc generates a new cdefs.h
 from our headers. Attempting to use an older gcc from ports
 that was build with the broken mono-parameter __nonnull() ended
 up causing breakage in any code using signal.h or pthreads.h.
 
 I see.  gcc's fixed headers cause lots of problems.
 
 I've complained about this multiple times in the past.  The gcc ports
 should not install these fixed headers.

Indeed.  See also this recent discussion on -current:

https://lists.freebsd.org/pipermail/freebsd-current/2015-March/055111.html

where a fixed stdio.h (from a gcc port) causes trouble.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r280400 - in head/contrib/llvm: include/llvm/CodeGen lib/CodeGen/SelectionDAG lib/Target/ARM

[Dimitry Andric]

Author: dim
Date: Mon Mar 23 21:13:29 2015
New Revision: 280400
URL: https://svnweb.freebsd.org/changeset/base/280400

Log:
  Pull in r230348 from upstream llvm trunk (by Tim Northover):
  
ARM: treat [N x i32] and [N x i64] as AAPCS composite types
  
The logic is almost there already, with our special homogeneous
aggregate handling. Tweaking it like this allows front-ends to emit
AAPCS compliant code without ever having to count registers or add
discarded padding arguments.
  
Only arrays of i32 and i64 are needed to model AAPCS rules, but I
decided to apply the logic to all integer arrays for more consistency.
  
  This fixes a possible Unexpected member type for HA error when
  compiling lib/msun/bsdsrc/b_tgamma.c for armv6.
  
  Reported by:  Jakub Palider j...@semihalf.com

Modified:
  head/contrib/llvm/include/llvm/CodeGen/CallingConvLower.h
  head/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
  head/contrib/llvm/lib/Target/ARM/ARMCallingConv.h
  head/contrib/llvm/lib/Target/ARM/ARMCallingConv.td
  head/contrib/llvm/lib/Target/ARM/ARMISelLowering.cpp

Modified: head/contrib/llvm/include/llvm/CodeGen/CallingConvLower.h
==
--- head/contrib/llvm/include/llvm/CodeGen/CallingConvLower.h   Mon Mar 23 
20:51:35 2015(r280399)
+++ head/contrib/llvm/include/llvm/CodeGen/CallingConvLower.h   Mon Mar 23 
21:13:29 2015(r280400)
@@ -122,8 +122,8 @@ public:
   // There is no need to differentiate between a pending CCValAssign and other
   // kinds, as they are stored in a different list.
   static CCValAssign getPending(unsigned ValNo, MVT ValVT, MVT LocVT,
-LocInfo HTP) {
-return getReg(ValNo, ValVT, 0, LocVT, HTP);
+LocInfo HTP, unsigned ExtraInfo = 0) {
+return getReg(ValNo, ValVT, ExtraInfo, LocVT, HTP);
   }
 
   void convertToReg(unsigned RegNo) {
@@ -146,6 +146,7 @@ public:
 
   unsigned getLocReg() const { assert(isRegLoc()); return Loc; }
   unsigned getLocMemOffset() const { assert(isMemLoc()); return Loc; }
+  unsigned getExtraInfo() const { return Loc; }
   MVT getLocVT() const { return LocVT; }
 
   LocInfo getLocInfo() const { return HTP; }

Modified: head/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
==
--- head/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp  Mon Mar 
23 20:51:35 2015(r280399)
+++ head/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp  Mon Mar 
23 21:13:29 2015(r280400)
@@ -7429,11 +7429,8 @@ TargetLowering::LowerCallTo(TargetLoweri
   }
   if (Args[i].isNest)
 Flags.setNest();
-  if (NeedsRegBlock) {
+  if (NeedsRegBlock)
 Flags.setInConsecutiveRegs();
-if (Value == NumValues - 1)
-  Flags.setInConsecutiveRegsLast();
-  }
   Flags.setOrigAlign(OriginalAlignment);
 
   MVT PartVT = getRegisterType(CLI.RetTy-getContext(), VT);
@@ -7482,6 +7479,9 @@ TargetLowering::LowerCallTo(TargetLoweri
 CLI.Outs.push_back(MyFlags);
 CLI.OutVals.push_back(Parts[j]);
   }
+
+  if (NeedsRegBlock  Value == NumValues - 1)
+CLI.Outs[CLI.Outs.size() - 1].Flags.setInConsecutiveRegsLast();
 }
   }
 
@@ -7696,11 +7696,8 @@ void SelectionDAGISel::LowerArguments(co
   }
   if (F.getAttributes().hasAttribute(Idx, Attribute::Nest))
 Flags.setNest();
-  if (NeedsRegBlock) {
+  if (NeedsRegBlock)
 Flags.setInConsecutiveRegs();
-if (Value == NumValues - 1)
-  Flags.setInConsecutiveRegsLast();
-  }
   Flags.setOrigAlign(OriginalAlignment);
 
   MVT RegisterVT = TLI-getRegisterType(*CurDAG-getContext(), VT);
@@ -7715,6 +7712,8 @@ void SelectionDAGISel::LowerArguments(co
   MyFlags.Flags.setOrigAlign(1);
 Ins.push_back(MyFlags);
   }
+  if (NeedsRegBlock  Value == NumValues - 1)
+Ins[Ins.size() - 1].Flags.setInConsecutiveRegsLast();
   PartBase += VT.getStoreSize();
 }
   }

Modified: head/contrib/llvm/lib/Target/ARM/ARMCallingConv.h
==
--- head/contrib/llvm/lib/Target/ARM/ARMCallingConv.h   Mon Mar 23 20:51:35 
2015(r280399)
+++ head/contrib/llvm/lib/Target/ARM/ARMCallingConv.h   Mon Mar 23 21:13:29 
2015(r280400)
@@ -160,6 +160,8 @@ static bool RetCC_ARM_AAPCS_Custom_f64(u
State);
 }
 
+static const uint16_t RRegList[] = { ARM::R0,  ARM::R1,  ARM::R2,  ARM::R3 };
+
 static const uint16_t SRegList[] = { ARM::S0,  ARM::S1,  ARM::S2,  ARM::S3,
  ARM::S4,  ARM::S5,  ARM::S6,  ARM::S7,
  ARM::S8,  ARM::S9,  ARM::S10, ARM::S11,
@@ -168,81 +170,114 @@ static const uint16_t DRegList[] = { ARM
 

svn commit: r280401 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Mon Mar 23 21:15:07 2015
New Revision: 280401
URL: https://svnweb.freebsd.org/changeset/base/280401

Log:
  Add llvm patch corresponding to r280400.

Added:
  head/contrib/llvm/patches/patch-10-llvm-r230348-arm-fix-bad-ha.diff

Added: head/contrib/llvm/patches/patch-10-llvm-r230348-arm-fix-bad-ha.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-10-llvm-r230348-arm-fix-bad-ha.diff Mon Mar 
23 21:15:07 2015(r280401)
@@ -0,0 +1,419 @@
+Pull in r230348 from upstream llvm trunk (by Tim Northover):
+
+  ARM: treat [N x i32] and [N x i64] as AAPCS composite types
+
+  The logic is almost there already, with our special homogeneous
+  aggregate handling. Tweaking it like this allows front-ends to emit
+  AAPCS compliant code without ever having to count registers or add
+  discarded padding arguments.
+
+  Only arrays of i32 and i64 are needed to model AAPCS rules, but I
+  decided to apply the logic to all integer arrays for more consistency.
+
+This fixes a possible Unexpected member type for HA error when
+compiling lib/msun/bsdsrc/b_tgamma.c for armv6.
+
+Reported by:   Jakub Palider j...@semihalf.com
+
+Introduced here: https://svnweb.freebsd.org/changeset/base/280400
+
+Index: include/llvm/CodeGen/CallingConvLower.h
+===
+--- include/llvm/CodeGen/CallingConvLower.h
 include/llvm/CodeGen/CallingConvLower.h
+@@ -122,8 +122,8 @@ class CCValAssign {
+   // There is no need to differentiate between a pending CCValAssign and other
+   // kinds, as they are stored in a different list.
+   static CCValAssign getPending(unsigned ValNo, MVT ValVT, MVT LocVT,
+-LocInfo HTP) {
+-return getReg(ValNo, ValVT, 0, LocVT, HTP);
++LocInfo HTP, unsigned ExtraInfo = 0) {
++return getReg(ValNo, ValVT, ExtraInfo, LocVT, HTP);
+   }
+ 
+   void convertToReg(unsigned RegNo) {
+@@ -146,6 +146,7 @@ class CCValAssign {
+ 
+   unsigned getLocReg() const { assert(isRegLoc()); return Loc; }
+   unsigned getLocMemOffset() const { assert(isMemLoc()); return Loc; }
++  unsigned getExtraInfo() const { return Loc; }
+   MVT getLocVT() const { return LocVT; }
+ 
+   LocInfo getLocInfo() const { return HTP; }
+Index: lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
+===
+--- lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
 lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
+@@ -7429,11 +7429,8 @@ TargetLowering::LowerCallTo(TargetLowering::CallLo
+   }
+   if (Args[i].isNest)
+ Flags.setNest();
+-  if (NeedsRegBlock) {
++  if (NeedsRegBlock)
+ Flags.setInConsecutiveRegs();
+-if (Value == NumValues - 1)
+-  Flags.setInConsecutiveRegsLast();
+-  }
+   Flags.setOrigAlign(OriginalAlignment);
+ 
+   MVT PartVT = getRegisterType(CLI.RetTy-getContext(), VT);
+@@ -7482,6 +7479,9 @@ TargetLowering::LowerCallTo(TargetLowering::CallLo
+ CLI.Outs.push_back(MyFlags);
+ CLI.OutVals.push_back(Parts[j]);
+   }
++
++  if (NeedsRegBlock  Value == NumValues - 1)
++CLI.Outs[CLI.Outs.size() - 1].Flags.setInConsecutiveRegsLast();
+ }
+   }
+ 
+@@ -7696,11 +7696,8 @@ void SelectionDAGISel::LowerArguments(const Functi
+   }
+   if (F.getAttributes().hasAttribute(Idx, Attribute::Nest))
+ Flags.setNest();
+-  if (NeedsRegBlock) {
++  if (NeedsRegBlock)
+ Flags.setInConsecutiveRegs();
+-if (Value == NumValues - 1)
+-  Flags.setInConsecutiveRegsLast();
+-  }
+   Flags.setOrigAlign(OriginalAlignment);
+ 
+   MVT RegisterVT = TLI-getRegisterType(*CurDAG-getContext(), VT);
+@@ -7715,6 +7712,8 @@ void SelectionDAGISel::LowerArguments(const Functi
+   MyFlags.Flags.setOrigAlign(1);
+ Ins.push_back(MyFlags);
+   }
++  if (NeedsRegBlock  Value == NumValues - 1)
++Ins[Ins.size() - 1].Flags.setInConsecutiveRegsLast();
+   PartBase += VT.getStoreSize();
+ }
+   }
+Index: lib/Target/ARM/ARMCallingConv.h
+===
+--- lib/Target/ARM/ARMCallingConv.h
 lib/Target/ARM/ARMCallingConv.h
+@@ -160,6 +160,8 @@ static bool RetCC_ARM_AAPCS_Custom_f64(unsigned V
+State);
+ }
+ 
++static const uint16_t RRegList[] = { ARM::R0,  ARM::R1,  ARM::R2,  ARM::R3 };
++
+ static const uint16_t SRegList[] = { ARM::S0,  ARM::S1,  ARM::S2,  ARM::S3,
+  ARM::S4,  ARM::S5,  ARM::S6,  ARM::S7,
+  ARM::S8,  ARM::S9,  ARM::S10, ARM::S11,
+@@ -168,81 +170,114 @@ static const uint16_t DRegList[] = { ARM::D0, ARM:
+  ARM::D4, ARM::D5, ARM::D6, ARM::D7 };

svn commit: r280350 - in head/contrib/llvm: . tools/clang tools/lldb

[Dimitry Andric]

Author: dim
Date: Sun Mar 22 17:56:49 2015
New Revision: 280350
URL: https://svnweb.freebsd.org/changeset/base/280350

Log:
  Add FREEBSD-Xlist files for llvm, clang and lldb.
  
  These are generated, and not optimized in any way, since I am not
  entirely sure of the syntax or format of this type of file.  Feel free
  to suggest ways of shortening these lists.
  
  The general idea is the same for all three files, though:
  * Get rid of upstream build infrastructure (CMakeLists, Makefiles, etc)
  * Delete tests, tools and utilities we don't want or use (including
samples)
  * Remove various bits of upstream metadata files that we don't want or
use (.arcconfig, .gitignore, etc)

Added:
  head/contrib/llvm/FREEBSD-Xlist   (contents, props changed)
  head/contrib/llvm/tools/clang/FREEBSD-Xlist   (contents, props changed)
  head/contrib/llvm/tools/lldb/FREEBSD-Xlist   (contents, props changed)

Added: head/contrib/llvm/FREEBSD-Xlist
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/FREEBSD-Xlist Sun Mar 22 17:56:49 2015
(r280350)
@@ -0,0 +1,514 @@
+# $FreeBSD$
+.arcconfig
+.clang-format
+.clang-tidy
+.gitignore
+CMakeLists.txt
+CODE_OWNERS.TXT
+CREDITS.TXT
+LLVMBuild.txt
+Makefile
+Makefile.common
+Makefile.config.in
+Makefile.rules
+README.txt
+autoconf/
+bindings/
+cmake/
+configure
+docs/
+examples/
+include/llvm/CMakeLists.txt
+include/llvm/Config/
+include/llvm/IR/CMakeLists.txt
+include/llvm/Support/DataTypes.h.cmake
+include/llvm/Support/LICENSE.TXT
+lib/Analysis/CMakeLists.txt
+lib/Analysis/IPA/CMakeLists.txt
+lib/Analysis/IPA/LLVMBuild.txt
+lib/Analysis/IPA/Makefile
+lib/Analysis/LLVMBuild.txt
+lib/Analysis/Makefile
+lib/Analysis/README.txt
+lib/AsmParser/CMakeLists.txt
+lib/AsmParser/LLVMBuild.txt
+lib/AsmParser/Makefile
+lib/Bitcode/CMakeLists.txt
+lib/Bitcode/LLVMBuild.txt
+lib/Bitcode/Makefile
+lib/Bitcode/Reader/CMakeLists.txt
+lib/Bitcode/Reader/LLVMBuild.txt
+lib/Bitcode/Reader/Makefile
+lib/Bitcode/Writer/CMakeLists.txt
+lib/Bitcode/Writer/LLVMBuild.txt
+lib/Bitcode/Writer/Makefile
+lib/CMakeLists.txt
+lib/CodeGen/AsmPrinter/CMakeLists.txt
+lib/CodeGen/AsmPrinter/LLVMBuild.txt
+lib/CodeGen/AsmPrinter/Makefile
+lib/CodeGen/CMakeLists.txt
+lib/CodeGen/LLVMBuild.txt
+lib/CodeGen/Makefile
+lib/CodeGen/README.txt
+lib/CodeGen/SelectionDAG/CMakeLists.txt
+lib/CodeGen/SelectionDAG/LLVMBuild.txt
+lib/CodeGen/SelectionDAG/Makefile
+lib/DebugInfo/CMakeLists.txt
+lib/DebugInfo/LLVMBuild.txt
+lib/DebugInfo/Makefile
+lib/ExecutionEngine/CMakeLists.txt
+lib/ExecutionEngine/IntelJITEvents/CMakeLists.txt
+lib/ExecutionEngine/IntelJITEvents/LLVMBuild.txt
+lib/ExecutionEngine/IntelJITEvents/Makefile
+lib/ExecutionEngine/Interpreter/CMakeLists.txt
+lib/ExecutionEngine/Interpreter/LLVMBuild.txt
+lib/ExecutionEngine/Interpreter/Makefile
+lib/ExecutionEngine/LLVMBuild.txt
+lib/ExecutionEngine/MCJIT/CMakeLists.txt
+lib/ExecutionEngine/MCJIT/LLVMBuild.txt
+lib/ExecutionEngine/MCJIT/Makefile
+lib/ExecutionEngine/Makefile
+lib/ExecutionEngine/OProfileJIT/CMakeLists.txt
+lib/ExecutionEngine/OProfileJIT/LLVMBuild.txt
+lib/ExecutionEngine/OProfileJIT/Makefile
+lib/ExecutionEngine/RuntimeDyld/CMakeLists.txt
+lib/ExecutionEngine/RuntimeDyld/LLVMBuild.txt
+lib/ExecutionEngine/RuntimeDyld/Makefile
+lib/IR/CMakeLists.txt
+lib/IR/LLVMBuild.txt
+lib/IR/Makefile
+lib/IRReader/CMakeLists.txt
+lib/IRReader/LLVMBuild.txt
+lib/IRReader/Makefile
+lib/LLVMBuild.txt
+lib/LTO/CMakeLists.txt
+lib/LTO/LLVMBuild.txt
+lib/LTO/Makefile
+lib/LineEditor/CMakeLists.txt
+lib/LineEditor/LLVMBuild.txt
+lib/LineEditor/Makefile
+lib/Linker/CMakeLists.txt
+lib/Linker/LLVMBuild.txt
+lib/Linker/Makefile
+lib/MC/CMakeLists.txt
+lib/MC/LLVMBuild.txt
+lib/MC/MCDisassembler/CMakeLists.txt
+lib/MC/MCDisassembler/LLVMBuild.txt
+lib/MC/MCDisassembler/Makefile
+lib/MC/MCParser/CMakeLists.txt
+lib/MC/MCParser/LLVMBuild.txt
+lib/MC/MCParser/Makefile
+lib/MC/Makefile
+lib/Makefile
+lib/Object/CMakeLists.txt
+lib/Object/LLVMBuild.txt
+lib/Object/Makefile
+lib/Option/CMakeLists.txt
+lib/Option/LLVMBuild.txt
+lib/Option/Makefile
+lib/ProfileData/CMakeLists.txt
+lib/ProfileData/LLVMBuild.txt
+lib/ProfileData/Makefile
+lib/Support/CMakeLists.txt
+lib/Support/LLVMBuild.txt
+lib/Support/Makefile
+lib/Support/README.txt.system
+lib/TableGen/CMakeLists.txt
+lib/TableGen/LLVMBuild.txt
+lib/TableGen/Makefile
+lib/Target/AArch64/AsmParser/CMakeLists.txt
+lib/Target/AArch64/AsmParser/LLVMBuild.txt
+lib/Target/AArch64/AsmParser/Makefile
+lib/Target/AArch64/CMakeLists.txt
+lib/Target/AArch64/Disassembler/CMakeLists.txt
+lib/Target/AArch64/Disassembler/LLVMBuild.txt
+lib/Target/AArch64/Disassembler/Makefile
+lib/Target/AArch64/InstPrinter/CMakeLists.txt
+lib/Target/AArch64/InstPrinter/LLVMBuild.txt
+lib/Target/AArch64/InstPrinter/Makefile
+lib/Target/AArch64/LLVMBuild.txt
+lib/Target/AArch64/MCTargetDesc/CMakeLists.txt

svn commit: r280328 - head/share/mk

[Dimitry Andric]

Author: dim
Date: Sat Mar 21 19:13:13 2015
New Revision: 280328
URL: https://svnweb.freebsd.org/changeset/base/280328

Log:
  Correctly pass the -mllvm -enable-gvn=false flag in CLANG_OPT_SMALL
  (this has to be passed as a combination of two flags).  Should fix the
  case where the clang version is before 3.5.0.
  
  Submitted by: Pedro Arthur bygran...@gmail.com
  X-MFC-With:   r279018, r279378

Modified:
  head/share/mk/bsd.sys.mk

Modified: head/share/mk/bsd.sys.mk
==
--- head/share/mk/bsd.sys.mkSat Mar 21 17:56:55 2015(r280327)
+++ head/share/mk/bsd.sys.mkSat Mar 21 19:13:13 2015(r280328)
@@ -128,9 +128,9 @@ CWARNFLAGS+=-Wno-unknown-pragmas
 CLANG_NO_IAS=   -no-integrated-as
 .endif
 CLANG_OPT_SMALL= -mstack-alignment=8 -mllvm -inline-threshold=3\
--mllvm -simplifycfg-dup-ret -mllvm
+-mllvm -simplifycfg-dup-ret
 .if ${COMPILER_VERSION} = 30500
-CLANG_OPT_SMALL+= -enable-gvn=false
+CLANG_OPT_SMALL+= -mllvm -enable-gvn=false
 .endif
 CFLAGS.clang+=  -Qunused-arguments
 .if ${MACHINE_CPUARCH} == sparc64
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279980 - head/share/mk

[Dimitry Andric]

Author: dim
Date: Sat Mar 14 12:29:44 2015
New Revision: 279980
URL: https://svnweb.freebsd.org/changeset/base/279980

Log:
  Allow relative pathnames in SRCS, so as to enable building software
  which includes more than one file with the same name, in different
  directories.
  
  For example, setting:
  
  SRCS+=foo/foo.c bar/foo.c baz/foo.c
  
  will now create separate objdirs 'foo', 'bar' and 'baz' for each of the
  sources in the list, and use those objdirs for the corresponding object
  files.
  
  Reviewed by:  brooks, imp
  MFC after:1 week
  Differential Revision: https://reviews.freebsd.org/D1984

Modified:
  head/share/mk/bsd.obj.mk
  head/share/mk/sys.mk

Modified: head/share/mk/bsd.obj.mk
==
--- head/share/mk/bsd.obj.mkSat Mar 14 12:18:26 2015(r279979)
+++ head/share/mk/bsd.obj.mkSat Mar 14 12:29:44 2015(r279980)
@@ -89,6 +89,16 @@ obj: .PHONY
fi; \
${ECHO} ${CANONICALOBJDIR} created for ${.CURDIR}; \
fi
+.for dir in ${SRCS:H:O:u}
+   @if ! test -d ${CANONICALOBJDIR}/${dir}/; then \
+   mkdir -p ${CANONICALOBJDIR}/${dir}; \
+   if ! test -d ${CANONICALOBJDIR}/${dir}/; then \
+   ${ECHO} Unable to create ${CANONICALOBJDIR}/${dir}.; \
+   exit 1; \
+   fi; \
+   ${ECHO} ${CANONICALOBJDIR}/${dir} created for ${.CURDIR}; \
+   fi
+.endfor
 .endif
 
 .if !target(objlink)

Modified: head/share/mk/sys.mk
==
--- head/share/mk/sys.mkSat Mar 14 12:18:26 2015(r279979)
+++ head/share/mk/sys.mkSat Mar 14 12:29:44 2015(r279980)
@@ -242,21 +242,21 @@ YFLAGS?=  -d
${CTFCONVERT_CMD}
 
 .c.o:
-   ${CC} ${CFLAGS} -c ${.IMPSRC}
+   ${CC} ${CFLAGS} -c ${.IMPSRC} -o ${.TARGET}
${CTFCONVERT_CMD}
 
 .cc .cpp .cxx .C:
${CXX} ${CXXFLAGS} ${LDFLAGS} ${.IMPSRC} ${LDLIBS} -o ${.TARGET}
 
 .cc.o .cpp.o .cxx.o .C.o:
-   ${CXX} ${CXXFLAGS} -c ${.IMPSRC}
+   ${CXX} ${CXXFLAGS} -c ${.IMPSRC} -o ${.TARGET}
 
 .m.o:
-   ${OBJC} ${OBJCFLAGS} -c ${.IMPSRC}
+   ${OBJC} ${OBJCFLAGS} -c ${.IMPSRC} -o ${.TARGET}
${CTFCONVERT_CMD}
 
 .p.o:
-   ${PC} ${PFLAGS} -c ${.IMPSRC}
+   ${PC} ${PFLAGS} -c ${.IMPSRC} -o ${.TARGET}
${CTFCONVERT_CMD}
 
 .e .r .F .f:
@@ -264,14 +264,15 @@ YFLAGS?=  -d
-o ${.TARGET}
 
 .e.o .r.o .F.o .f.o:
-   ${FC} ${RFLAGS} ${EFLAGS} ${FFLAGS} -c ${.IMPSRC}
+   ${FC} ${RFLAGS} ${EFLAGS} ${FFLAGS} -c ${.IMPSRC} -o ${.TARGET}
 
 .S.o:
-   ${CC} ${CFLAGS} ${ACFLAGS} -c ${.IMPSRC}
+   ${CC} ${CFLAGS} ${ACFLAGS} -c ${.IMPSRC} -o ${.TARGET}
${CTFCONVERT_CMD}
 
 .asm.o:
-   ${CC} -x assembler-with-cpp ${CFLAGS} ${ACFLAGS} -c ${.IMPSRC}
+   ${CC} -x assembler-with-cpp ${CFLAGS} ${ACFLAGS} -c ${.IMPSRC} \
+   -o ${.TARGET}
${CTFCONVERT_CMD}
 
 .s.o:
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279981 - in head: contrib/compiler-rt/lib/builtins lib/libcompiler_rt

[Dimitry Andric]

Author: dim
Date: Sat Mar 14 12:40:19 2015
New Revision: 279981
URL: https://svnweb.freebsd.org/changeset/base/279981

Log:
  Pull in r231965 from upstream compiler-rt trunk (by Jörg Sonnenberger):
  
Refactor float to integer conversion to share the same code.
80bit Intel/PPC long double is excluded due to lacking support
for the abstraction. Consistently provide saturation logic.
Extend to long double on 128bit IEEE extended platforms.
  
Initial patch with test cases from GuanHong Liu.
Reviewed by Steve Canon.
  
Differential Revision: http://reviews.llvm.org/D2804
  
  Pull in r232107 from upstream compiler-rt trunk (by Ed Maste):
  
Use signed int implementation for __fixint
  
  Requested by: emaste

Added:
  head/contrib/compiler-rt/lib/builtins/fixtfdi.c   (contents, props changed)
  head/contrib/compiler-rt/lib/builtins/fixtfsi.c   (contents, props changed)
  head/contrib/compiler-rt/lib/builtins/fixtfti.c   (contents, props changed)
  head/contrib/compiler-rt/lib/builtins/fixunstfdi.c   (contents, props changed)
  head/contrib/compiler-rt/lib/builtins/fixunstfsi.c   (contents, props changed)
  head/contrib/compiler-rt/lib/builtins/fixunstfti.c   (contents, props changed)
  head/contrib/compiler-rt/lib/builtins/fp_fixint_impl.inc   (contents, props 
changed)
  head/contrib/compiler-rt/lib/builtins/fp_fixuint_impl.inc   (contents, props 
changed)
Modified:
  head/contrib/compiler-rt/lib/builtins/fixdfdi.c
  head/contrib/compiler-rt/lib/builtins/fixdfsi.c
  head/contrib/compiler-rt/lib/builtins/fixdfti.c
  head/contrib/compiler-rt/lib/builtins/fixsfdi.c
  head/contrib/compiler-rt/lib/builtins/fixsfsi.c
  head/contrib/compiler-rt/lib/builtins/fixsfti.c
  head/contrib/compiler-rt/lib/builtins/fixunsdfdi.c
  head/contrib/compiler-rt/lib/builtins/fixunsdfsi.c
  head/contrib/compiler-rt/lib/builtins/fixunsdfti.c
  head/contrib/compiler-rt/lib/builtins/fixunssfdi.c
  head/contrib/compiler-rt/lib/builtins/fixunssfsi.c
  head/contrib/compiler-rt/lib/builtins/fixunssfti.c
  head/contrib/compiler-rt/lib/builtins/fixunsxfdi.c
  head/contrib/compiler-rt/lib/builtins/fixunsxfsi.c
  head/contrib/compiler-rt/lib/builtins/fixunsxfti.c
  head/contrib/compiler-rt/lib/builtins/fixxfdi.c
  head/contrib/compiler-rt/lib/builtins/fixxfti.c
  head/lib/libcompiler_rt/Makefile

Modified: head/contrib/compiler-rt/lib/builtins/fixdfdi.c
==
--- head/contrib/compiler-rt/lib/builtins/fixdfdi.c Sat Mar 14 12:29:44 
2015(r279980)
+++ head/contrib/compiler-rt/lib/builtins/fixdfdi.c Sat Mar 14 12:40:19 
2015(r279981)
@@ -6,40 +6,17 @@
  * Source Licenses. See LICENSE.TXT for details.
  *
  * ===--===
- *
- * This file implements __fixdfdi for the compiler_rt library.
- *
- * ===--===
  */
 
-#include int_lib.h
-
-/* Returns: convert a to a signed long long, rounding toward zero. */
-
-/* Assumption: double is a IEEE 64 bit floating point type 
- *su_int is a 32 bit integral type
- *value in double is representable in di_int (no range checking 
performed)
- */
-
-/* seee        |       
  */
-
+#define DOUBLE_PRECISION
+#include fp_lib.h
 ARM_EABI_FNALIAS(d2lz, fixdfdi)
 
+typedef di_int fixint_t;
+typedef du_int fixuint_t;
+#include fp_fixint_impl.inc
+
 COMPILER_RT_ABI di_int
-__fixdfdi(double a)
-{
-double_bits fb;
-fb.f = a;
-int e = ((fb.u.s.high  0x7FF0)  20) - 1023;
-if (e  0)
-return 0;
-di_int s = (si_int)(fb.u.s.high  0x8000)  31;
-dwords r;
-r.s.high = (fb.u.s.high  0x000F) | 0x0010;
-r.s.low = fb.u.s.low;
-if (e  52)
-r.all = (e - 52);
-else
-r.all = (52 - e);
-return (r.all ^ s) - s;
-} 
+__fixdfdi(fp_t a) {
+return __fixint(a);
+}

Modified: head/contrib/compiler-rt/lib/builtins/fixdfsi.c
==
--- head/contrib/compiler-rt/lib/builtins/fixdfsi.c Sat Mar 14 12:29:44 
2015(r279980)
+++ head/contrib/compiler-rt/lib/builtins/fixdfsi.c Sat Mar 14 12:40:19 
2015(r279981)
@@ -1,50 +1,22 @@
-//===-- lib/fixdfsi.c - Double-precision - integer conversion *- C 
-*-===//
-//
-// The LLVM Compiler Infrastructure
-//
-// This file is dual licensed under the MIT and the University of Illinois Open
-// Source Licenses. See LICENSE.TXT for details.
-//
-//===--===//
-//
-// This file implements double-precision to integer conversion for the
-// compiler-rt library.  No range checking is performed; the behavior of this
-// conversion is undefined for out of range values in the C standard.
-//

svn commit: r279994 - head/sys/conf

[Dimitry Andric]

Author: dim
Date: Sat Mar 14 17:19:48 2015
New Revision: 279994
URL: https://svnweb.freebsd.org/changeset/base/279994

Log:
  Amend r277940, by also disabling -Wcast-qual warnings for a few specific
  aesni files on i386.

Modified:
  head/sys/conf/files.i386

Modified: head/sys/conf/files.i386
==
--- head/sys/conf/files.i386Sat Mar 14 17:08:28 2015(r279993)
+++ head/sys/conf/files.i386Sat Mar 14 17:19:48 2015(r279994)
@@ -118,12 +118,12 @@ crypto/aesni/aeskeys_i386.S   optional aes
 crypto/aesni/aesni.c   optional aesni
 aesni_ghash.o  optional aesni  \
dependency  $S/crypto/aesni/aesni_ghash.c \
-   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${PROF} -mmmx -msse -msse4 -maes -mpclmul ${.IMPSRC} \
+   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${NO_WCAST_QUAL} ${PROF} -mmmx -msse -msse4 -maes -mpclmul ${.IMPSRC} \
no-implicit-rule\
clean   aesni_ghash.o
 aesni_wrap.o   optional aesni  \
dependency  $S/crypto/aesni/aesni_wrap.c  \
-   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${PROF} -mmmx -msse -msse4 -maes ${.IMPSRC} \
+   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${NO_WCAST_QUAL} ${PROF} -mmmx -msse -msse4 -maes ${.IMPSRC} \
no-implicit-rule\
clean   aesni_wrap.o
 crypto/des/arch/i386/des_enc.S optional crypto | ipsec | netsmb
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279760 - head/lib/libnv/tests

[Dimitry Andric]

Author: dim
Date: Sun Mar  8 00:30:52 2015
New Revision: 279760
URL: https://svnweb.freebsd.org/changeset/base/279760

Log:
  Fix lib/libnv tests compilation with -std=c++11, by adding appropriate
  casts for NULL to invocations of the ATF_REQUIER_EQ() macro.
  
  Reviewed by:  rstone, jmmv
  Differential Revision: https://reviews.freebsd.org/D2027

Modified:
  head/lib/libnv/tests/dnv_tests.cc
  head/lib/libnv/tests/nv_tests.cc

Modified: head/lib/libnv/tests/dnv_tests.cc
==
--- head/lib/libnv/tests/dnv_tests.cc   Sat Mar  7 23:01:27 2015
(r279759)
+++ head/lib/libnv/tests/dnv_tests.cc   Sun Mar  8 00:30:52 2015
(r279760)
@@ -450,7 +450,7 @@ ATF_TEST_CASE_BODY(dnvlist_take_nvlist__
nvl = nvlist_create(0);
 
actual_val = dnvlist_take_nvlist(nvl, 123, NULL);
-   ATF_REQUIRE_EQ(actual_val, NULL);
+   ATF_REQUIRE_EQ(actual_val, static_castnvlist_t *(NULL));
 
free(actual_val);
nvlist_destroy(nvl);

Modified: head/lib/libnv/tests/nv_tests.cc
==
--- head/lib/libnv/tests/nv_tests.ccSat Mar  7 23:01:27 2015
(r279759)
+++ head/lib/libnv/tests/nv_tests.ccSun Mar  8 00:30:52 2015
(r279760)
@@ -54,7 +54,7 @@ ATF_TEST_CASE_BODY(nvlist_create__is_emp
ATF_REQUIRE(nvlist_empty(nvl));
 
it = NULL;
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type, it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type, it), static_castconst char 
*(NULL));
 
nvlist_destroy(nvl);
 }
@@ -85,7 +85,7 @@ ATF_TEST_CASE_BODY(nvlist_add_null__sing
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_NULL);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), static_castconst char 
*(NULL));
 
nvlist_destroy(nvl);
 }
@@ -118,7 +118,7 @@ ATF_TEST_CASE_BODY(nvlist_add_bool__sing
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_BOOL);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), static_castconst char 
*(NULL));
 
nvlist_destroy(nvl);
 }
@@ -153,7 +153,7 @@ ATF_TEST_CASE_BODY(nvlist_add_number__si
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_NUMBER);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), static_castconst char 
*(NULL));
 
nvlist_destroy(nvl);
 }
@@ -191,7 +191,7 @@ ATF_TEST_CASE_BODY(nvlist_add_string__si
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_STRING);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), static_castconst char 
*(NULL));
 
nvlist_destroy(nvl);
 }
@@ -237,7 +237,7 @@ ATF_TEST_CASE_BODY(nvlist_add_nvlist__si
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_NVLIST);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), static_castconst char 
*(NULL));
 
nvlist_destroy(sublist);
nvlist_destroy(nvl);
@@ -303,7 +303,7 @@ ATF_TEST_CASE_BODY(nvlist_add_binary__si
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_BINARY);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type,it), static_castconst char 
*(NULL));
 
nvlist_destroy(nvl);
free(value);
@@ -352,7 +352,7 @@ ATF_TEST_CASE_BODY(nvlist_clone__nonempt
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(clone, type, it), key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_NUMBER);
-   ATF_REQUIRE_EQ(nvlist_next(clone, type, it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(clone, type, it), static_castconst char 
*(NULL));
 
nvlist_destroy(clone);
nvlist_destroy(nvl);
@@ -400,13 +400,13 @@ verify_test_nvlist(const nvlist_t *nvl)
ATF_REQUIRE_EQ(strcmp(nvlist_next(value, type, it),
test_string_key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_STRING);
-   ATF_REQUIRE_EQ(nvlist_next(value, type, it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(value, type, it), static_castconst char 
*(NULL));
 
it = NULL;
ATF_REQUIRE_EQ(strcmp(nvlist_next(nvl, type, it),
test_subnvlist_key), 0);
ATF_REQUIRE_EQ(type, NV_TYPE_NVLIST);
-   ATF_REQUIRE_EQ(nvlist_next(nvl, type, it), NULL);
+   ATF_REQUIRE_EQ(nvlist_next(nvl, type, it), static_castconst char 
*(NULL));
 }
 
 ATF_TEST_CASE_WITHOUT_HEAD(nvlist_clone__nested_nvlist);

svn commit: r279757 - head/contrib/libc++/include

[Dimitry Andric]

Author: dim
Date: Sat Mar  7 22:53:15 2015
New Revision: 279757
URL: https://svnweb.freebsd.org/changeset/base/279757

Log:
  Pull in r228344 from upstream libc++ trunk (by Eric Fiselier):
  
Get tests running with warnings. Fix warnings in headers and tests
  
  This fixes a number of -Wunused-local-typedef warnings in libc++ headers.
  
  MFC after:3 days

Modified:
  head/contrib/libc++/include/__bit_reference
  head/contrib/libc++/include/algorithm

Modified: head/contrib/libc++/include/__bit_reference
==
--- head/contrib/libc++/include/__bit_reference Sat Mar  7 22:46:35 2015
(r279756)
+++ head/contrib/libc++/include/__bit_reference Sat Mar  7 22:53:15 2015
(r279757)
@@ -906,7 +906,6 @@ rotate(__bit_iterator_Cp, false __firs
 {
 typedef __bit_iterator_Cp, false _I1;
 typedef  typename _I1::difference_type difference_type;
-typedef typename _I1::__storage_type __storage_type;
 difference_type __d1 = __middle - __first;
 difference_type __d2 = __last - __middle;
 _I1 __r = __first + __d2;

Modified: head/contrib/libc++/include/algorithm
==
--- head/contrib/libc++/include/algorithm   Sat Mar  7 22:46:35 2015
(r279756)
+++ head/contrib/libc++/include/algorithm   Sat Mar  7 22:53:15 2015
(r279757)
@@ -4365,8 +4365,6 @@ __buffered_inplace_merge(_BidirectionalI
 typename iterator_traits_BidirectionalIterator::value_type* 
__buff)
 {
 typedef typename iterator_traits_BidirectionalIterator::value_type 
value_type;
-typedef typename iterator_traits_BidirectionalIterator::difference_type 
difference_type;
-typedef typename iterator_traits_BidirectionalIterator::pointer pointer;
 __destruct_n __d(0);
 unique_ptrvalue_type, __destruct_n __h2(__buff, __d);
 if (__len1 = __len2)
@@ -4400,7 +4398,6 @@ __inplace_merge(_BidirectionalIterator _
  typename 
iterator_traits_BidirectionalIterator::difference_type __len2,
 typename iterator_traits_BidirectionalIterator::value_type* 
__buff, ptrdiff_t __buff_size)
 {
-typedef typename iterator_traits_BidirectionalIterator::value_type 
value_type;
 typedef typename iterator_traits_BidirectionalIterator::difference_type 
difference_type;
 while (true)
 {
@@ -4799,7 +4796,6 @@ void
 __sift_up(_RandomAccessIterator __first, _RandomAccessIterator __last, 
_Compare __comp,
   typename iterator_traits_RandomAccessIterator::difference_type 
__len)
 {
-typedef typename iterator_traits_RandomAccessIterator::difference_type 
difference_type;
 typedef typename iterator_traits_RandomAccessIterator::value_type 
value_type;
 if (__len  1)
 {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279758 - head/contrib/libc++/include

[Dimitry Andric]

Author: dim
Date: Sat Mar  7 22:55:31 2015
New Revision: 279758
URL: https://svnweb.freebsd.org/changeset/base/279758

Log:
  Fix another -Wunused-local-typedef warning in libc++, in include/__tree.
  
  MFC after:3 days

Modified:
  head/contrib/libc++/include/__tree

Modified: head/contrib/libc++/include/__tree
==
--- head/contrib/libc++/include/__tree  Sat Mar  7 22:53:15 2015
(r279757)
+++ head/contrib/libc++/include/__tree  Sat Mar  7 22:55:31 2015
(r279758)
@@ -2069,7 +2069,6 @@ template class _Key
 typename __tree_Tp, _Compare, _Allocator::size_type
 __tree_Tp, _Compare, _Allocator::__count_multi(const _Key __k) const
 {
-typedef pairconst_iterator, const_iterator _Pp;
 __node_const_pointer __result = __end_node();
 __node_const_pointer __rt = __root();
 while (__rt != nullptr)
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279598 - in head/sys/boot: i386/boot2 pc98/boot2

[Dimitry Andric]

Author: dim
Date: Wed Mar  4 20:33:15 2015
New Revision: 279598
URL: https://svnweb.freebsd.org/changeset/base/279598

Log:
  When compiling boot2 with gcc on i386 and pc98, only use the custom flag
  -mno-align-long-strings when compiling with base gcc.  This is checked
  by comparing the version number against 4.2.1, which is not exactly
  right, but good enough.  (There is no other way to check whether we are
  using the non-standard gcc in base, as far as I know.)
  
  Reported by:  rodrigc
  MFC after:3 days

Modified:
  head/sys/boot/i386/boot2/Makefile
  head/sys/boot/pc98/boot2/Makefile

Modified: head/sys/boot/i386/boot2/Makefile
==
--- head/sys/boot/i386/boot2/Makefile   Wed Mar  4 20:04:23 2015
(r279597)
+++ head/sys/boot/i386/boot2/Makefile   Wed Mar  4 20:33:15 2015
(r279598)
@@ -41,8 +41,10 @@ CFLAGS=  -fomit-frame-pointer \
 CFLAGS.gcc+=   -Os \
-fno-guess-branch-probability \
-fno-unit-at-a-time \
-   -mno-align-long-strings \
--param max-inline-insns-single=100
+.if ${COMPILER_TYPE} == gcc  ${COMPILER_VERSION} = 40201
+CFLAGS.gcc+=   -mno-align-long-strings
+.endif
 
 CFLAGS.clang+= -Oz ${CLANG_OPT_SMALL}
 

Modified: head/sys/boot/pc98/boot2/Makefile
==
--- head/sys/boot/pc98/boot2/Makefile   Wed Mar  4 20:04:23 2015
(r279597)
+++ head/sys/boot/pc98/boot2/Makefile   Wed Mar  4 20:33:15 2015
(r279598)
@@ -39,8 +39,10 @@ CFLAGS=  -fomit-frame-pointer \
 CFLAGS.gcc+=   -Os \
-fno-guess-branch-probability \
-fno-unit-at-a-time \
-   -mno-align-long-strings \
--param max-inline-insns-single=100
+.if ${COMPILER_TYPE} == gcc  ${COMPILER_VERSION} = 40201
+CFLAGS.gcc+=   -mno-align-long-strings
+.endif
 
 # Set machine type to PC98_SYSTEM_PARAMETER
 #CFLAGS+=  -DSET_MACHINE_TYPE
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279307 - head/contrib/libcxxrt

[Dimitry Andric]

Author: dim
Date: Thu Feb 26 07:20:05 2015
New Revision: 279307
URL: https://svnweb.freebsd.org/changeset/base/279307

Log:
  Make libcxxrt's parsing of DWARF exception handling tables work on
  architectures with strict alignment, by using memcpy() instead of
  directly reading fields.
  
  Reported by:  Daisuke Aoyama aoy...@peach.ne.jp
  Reviewed by:  imp, bapt
  Tested by:bapt
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D1967

Modified:
  head/contrib/libcxxrt/dwarf_eh.h

Modified: head/contrib/libcxxrt/dwarf_eh.h
==
--- head/contrib/libcxxrt/dwarf_eh.hThu Feb 26 02:22:47 2015
(r279306)
+++ head/contrib/libcxxrt/dwarf_eh.hThu Feb 26 07:20:05 2015
(r279307)
@@ -218,15 +218,17 @@ static int64_t read_sleb128(dw_eh_ptr_t 
 static uint64_t read_value(char encoding, dw_eh_ptr_t *data)
 {
enum dwarf_data_encoding type = get_encoding(encoding);
-   uint64_t v;
switch (type)
{
// Read fixed-length types
 #define READ(dwarf, type) \
case dwarf:\
-   v = 
static_castuint64_t(*reinterpret_casttype*(*data));\
-   *data += sizeof(type);\
-   break;
+   {\
+   type t;\
+   memcpy(t, *data, sizeof t);\
+   *data += sizeof t;\
+   return static_castuint64_t(t);\
+   }
READ(DW_EH_PE_udata2, uint16_t)
READ(DW_EH_PE_udata4, uint32_t)
READ(DW_EH_PE_udata8, uint64_t)
@@ -237,15 +239,11 @@ static uint64_t read_value(char encoding
 #undef READ
// Read variable-length types
case DW_EH_PE_sleb128:
-   v = read_sleb128(data);
-   break;
+   return read_sleb128(data);
case DW_EH_PE_uleb128:
-   v = read_uleb128(data);
-   break;
+   return read_uleb128(data);
default: abort();
}
-
-   return v;
 }
 
 /**
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279312 - in head/sys/arm: arm ti ti/am335x ti/omap4

[Dimitry Andric]

Author: dim
Date: Thu Feb 26 07:47:35 2015
New Revision: 279312
URL: https://svnweb.freebsd.org/changeset/base/279312

Log:
  Fix a number of -Wcast-qual warnings under sys/arm.  No functional
  change.
  
  Submitted by: andrew
  MFC after:3 days

Modified:
  head/sys/arm/arm/disassem.c
  head/sys/arm/ti/am335x/am335x_scm_padconf.c
  head/sys/arm/ti/omap4/omap4_scm_padconf.c
  head/sys/arm/ti/ti_scm.h

Modified: head/sys/arm/arm/disassem.c
==
--- head/sys/arm/arm/disassem.c Thu Feb 26 07:45:40 2015(r279311)
+++ head/sys/arm/arm/disassem.c Thu Feb 26 07:47:35 2015(r279312)
@@ -289,7 +289,7 @@ static void disassemble_printaddr(u_int 
 vm_offset_t
 disasm(const disasm_interface_t *di, vm_offset_t loc, int altfmt)
 {
-   struct arm32_insn *i_ptr = (struct arm32_insn *)arm32_i;
+   const struct arm32_insn *i_ptr = arm32_i;
 
u_int insn;
int matchp;

Modified: head/sys/arm/ti/am335x/am335x_scm_padconf.c
==
--- head/sys/arm/ti/am335x/am335x_scm_padconf.c Thu Feb 26 07:45:40 2015
(r279311)
+++ head/sys/arm/ti/am335x/am335x_scm_padconf.c Thu Feb 26 07:47:35 2015
(r279312)
@@ -298,6 +298,6 @@ const static struct ti_scm_padconf ti_pa
 const struct ti_scm_device ti_scm_dev = {
.padconf_muxmode_mask   = 0x7,
.padconf_sate_mask  = 0x78,
-   .padstate   = (struct ti_scm_padstate *) 
ti_padstate_devmap,
-   .padconf= (struct ti_scm_padconf *) ti_padconf_devmap,
+   .padstate   = ti_padstate_devmap,
+   .padconf= ti_padconf_devmap,
 };

Modified: head/sys/arm/ti/omap4/omap4_scm_padconf.c
==
--- head/sys/arm/ti/omap4/omap4_scm_padconf.c   Thu Feb 26 07:45:40 2015
(r279311)
+++ head/sys/arm/ti/omap4/omap4_scm_padconf.c   Thu Feb 26 07:47:35 2015
(r279312)
@@ -298,6 +298,6 @@ const static struct ti_scm_padconf ti_pa
 const struct ti_scm_device ti_scm_dev = {
.padconf_muxmode_mask   = CONTROL_PADCONF_MUXMODE_MASK,
.padconf_sate_mask  = CONTROL_PADCONF_SATE_MASK,
-   .padstate   = (struct ti_scm_padstate *) 
ti_padstate_devmap,
-   .padconf= (struct ti_scm_padconf *) ti_padconf_devmap,
+   .padstate   = ti_padstate_devmap,
+   .padconf= ti_padconf_devmap,
 };

Modified: head/sys/arm/ti/ti_scm.h
==
--- head/sys/arm/ti/ti_scm.hThu Feb 26 07:45:40 2015(r279311)
+++ head/sys/arm/ti/ti_scm.hThu Feb 26 07:47:35 2015(r279312)
@@ -59,8 +59,8 @@ struct ti_scm_padstate {
 struct ti_scm_device {
uint16_tpadconf_muxmode_mask;
uint16_tpadconf_sate_mask;
-   struct ti_scm_padstate  *padstate;
-   struct ti_scm_padconf   *padconf;
+   const struct ti_scm_padstate*padstate;
+   const struct ti_scm_padconf *padconf;
 };
 
 struct ti_scm_softc {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279310 - head/contrib/libcxxrt

[Dimitry Andric]

Author: dim
Date: Thu Feb 26 07:42:16 2015
New Revision: 279310
URL: https://svnweb.freebsd.org/changeset/base/279310

Log:
  Since newer versions of compiler-rt require unwind.h, and we want to use
  the copy in libcxxrt for it, fix the arm-specific header to define the
  _Unwind_Action type.
  
  Submitted by: andrew
  MFC after:3 days

Modified:
  head/contrib/libcxxrt/unwind-arm.h

Modified: head/contrib/libcxxrt/unwind-arm.h
==
--- head/contrib/libcxxrt/unwind-arm.h  Thu Feb 26 07:26:56 2015
(r279309)
+++ head/contrib/libcxxrt/unwind-arm.h  Thu Feb 26 07:42:16 2015
(r279310)
@@ -36,6 +36,8 @@
_URC_FATAL_PHASE1_ERROR = _URC_FAILURE
 } _Unwind_Reason_Code;
 
+typedef int _Unwind_Action;
+
 typedef uint32_t _Unwind_State;
 #ifdef __clang__
 static const _Unwind_State _US_VIRTUAL_UNWIND_FRAME  = 0;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279311 - head/sys/arm/ti/am335x

[Dimitry Andric]

Author: dim
Date: Thu Feb 26 07:45:40 2015
New Revision: 279311
URL: https://svnweb.freebsd.org/changeset/base/279311

Log:
  In sys/arm/ti/am335x/am335x_rtc.c, fix a clang 3.6.0 warning about
  am33x_rtc_softc::sc_irq_res (which is an array) never being NULL.
  
  Submitted by: andrew
  MFC after:3 days

Modified:
  head/sys/arm/ti/am335x/am335x_rtc.c

Modified: head/sys/arm/ti/am335x/am335x_rtc.c
==
--- head/sys/arm/ti/am335x/am335x_rtc.c Thu Feb 26 07:42:16 2015
(r279310)
+++ head/sys/arm/ti/am335x/am335x_rtc.c Thu Feb 26 07:45:40 2015
(r279311)
@@ -137,7 +137,7 @@ am335x_rtc_detach(device_t dev)
struct am335x_rtc_softc *sc;
 
sc = device_get_softc(dev);
-   if (sc-sc_irq_res)
+   if (sc-sc_irq_res[0] != NULL)
bus_release_resources(dev, am335x_rtc_irq_spec, sc-sc_irq_res);
if (sc-sc_mem_res)
bus_release_resource(dev, SYS_RES_MEMORY, 0, sc-sc_mem_res);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279162 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Sun Feb 22 15:56:16 2015
New Revision: 279162
URL: https://svnweb.freebsd.org/changeset/base/279162

Log:
  Add llvm patch corresponding to r279161.

Added:
  head/contrib/llvm/patches/patch-32-llvm-r230058-indirectbrs-assert.diff

Added: head/contrib/llvm/patches/patch-32-llvm-r230058-indirectbrs-assert.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-32-llvm-r230058-indirectbrs-assert.diff 
Sun Feb 22 15:56:16 2015(r279162)
@@ -0,0 +1,55 @@
+Pull in r230058 from upstream llvm trunk (by Benjamin Kramer):
+
+  LoopRotate: When reconstructing loop simplify form don't split edges
+  from indirectbrs.
+
+  Yet another chapter in the endless story. While this looks like we
+  leave the loop in a non-canonical state this replicates the logic in
+  LoopSimplify so it doesn't diverge from the canonical form in any way.
+
+  PR21968
+
+This fixes a Cannot split critical edge from IndirectBrInst assertion
+failure when building the devel/radare2 port.
+
+Introduced here: https://svnweb.freebsd.org/changeset/base/279161
+
+Index: lib/Transforms/Scalar/LoopRotation.cpp
+===
+--- lib/Transforms/Scalar/LoopRotation.cpp
 lib/Transforms/Scalar/LoopRotation.cpp
+@@ -498,6 +498,8 @@ bool LoopRotate::rotateLoop(Loop *L, bool Simplifi
+   Loop *PredLoop = LI-getLoopFor(*PI);
+   if (!PredLoop || PredLoop-contains(Exit))
+ continue;
++  if (isaIndirectBrInst((*PI)-getTerminator()))
++continue;
+   SplitLatchEdge |= L-getLoopLatch() == *PI;
+   BasicBlock *ExitSplit = SplitCriticalEdge(*PI, Exit, this);
+   ExitSplit-moveBefore(Exit);
+Index: test/Transforms/LoopRotate/crash.ll
+===
+--- test/Transforms/LoopRotate/crash.ll
 test/Transforms/LoopRotate/crash.ll
+@@ -153,3 +153,21 @@ entry:
+ 5:  ; preds = %3, %entry
+   ret void
+ }
++
++; PR21968
++define void @test8(i1 %C, i8* %P) #0 {
++entry:
++  br label %for.cond
++
++for.cond: ; preds = %for.inc, %entry
++  br i1 %C, label %l_bad, label %for.body
++
++for.body: ; preds = %for.cond
++  indirectbr i8* %P, [label %for.inc, label %l_bad]
++
++for.inc:  ; preds = %for.body
++  br label %for.cond
++
++l_bad:; preds = %for.body, 
%for.cond
++  ret void
++}
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r279161 - head/contrib/llvm/lib/Transforms/Scalar

[Dimitry Andric]

Author: dim
Date: Sun Feb 22 15:51:49 2015
New Revision: 279161
URL: https://svnweb.freebsd.org/changeset/base/279161

Log:
  Pull in r230058 from upstream llvm trunk (by Benjamin Kramer):
  
LoopRotate: When reconstructing loop simplify form don't split edges
from indirectbrs.
  
Yet another chapter in the endless story. While this looks like we
leave the loop in a non-canonical state this replicates the logic in
LoopSimplify so it doesn't diverge from the canonical form in any way.
  
http://llvm.org/PR21968
  
  This fixes a Cannot split critical edge from IndirectBrInst assertion
  failure when building the devel/radare2 port.
  
  PR:   195480, 196987
  MFC after:3 days

Modified:
  head/contrib/llvm/lib/Transforms/Scalar/LoopRotation.cpp

Modified: head/contrib/llvm/lib/Transforms/Scalar/LoopRotation.cpp
==
--- head/contrib/llvm/lib/Transforms/Scalar/LoopRotation.cppSun Feb 22 
15:48:36 2015(r279160)
+++ head/contrib/llvm/lib/Transforms/Scalar/LoopRotation.cppSun Feb 22 
15:51:49 2015(r279161)
@@ -498,6 +498,8 @@ bool LoopRotate::rotateLoop(Loop *L, boo
   Loop *PredLoop = LI-getLoopFor(*PI);
   if (!PredLoop || PredLoop-contains(Exit))
 continue;
+  if (isaIndirectBrInst((*PI)-getTerminator()))
+continue;
   SplitLatchEdge |= L-getLoopLatch() == *PI;
   BasicBlock *ExitSplit = SplitCriticalEdge(*PI, Exit, this);
   ExitSplit-moveBefore(Exit);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278438 - head/sys/dev/drm2/radeon

[Dimitry Andric]

Author: dim
Date: Mon Feb  9 07:56:50 2015
New Revision: 278438
URL: https://svnweb.freebsd.org/changeset/base/278438

Log:
  After r278004 was committed, Bruce Evans noted that the casts were
  actually completely unnecessary, here:
  
  https://lists.freebsd.org/pipermail/svn-src-all/2015-February/098478.html
  
  Remove the casts, and just assign xxx_io_mc_regs[0][0] directly.
  
  Reviewed by:  dumbbell
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D1748

Modified:
  head/sys/dev/drm2/radeon/ni.c
  head/sys/dev/drm2/radeon/si.c

Modified: head/sys/dev/drm2/radeon/ni.c
==
--- head/sys/dev/drm2/radeon/ni.c   Mon Feb  9 07:52:45 2015
(r278437)
+++ head/sys/dev/drm2/radeon/ni.c   Mon Feb  9 07:56:50 2015
(r278438)
@@ -190,23 +190,23 @@ int ni_mc_load_microcode(struct radeon_d
 
switch (rdev-family) {
case CHIP_BARTS:
-   io_mc_regs = (const u32 *)barts_io_mc_regs;
+   io_mc_regs = barts_io_mc_regs[0][0];
ucode_size = BTC_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;
case CHIP_TURKS:
-   io_mc_regs = (const u32 *)turks_io_mc_regs;
+   io_mc_regs = turks_io_mc_regs[0][0];
ucode_size = BTC_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;
case CHIP_CAICOS:
default:
-   io_mc_regs = (const u32 *)caicos_io_mc_regs;
+   io_mc_regs = caicos_io_mc_regs[0][0];
ucode_size = BTC_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;
case CHIP_CAYMAN:
-   io_mc_regs = (const u32 *)cayman_io_mc_regs;
+   io_mc_regs = cayman_io_mc_regs[0][0];
ucode_size = CAYMAN_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;

Modified: head/sys/dev/drm2/radeon/si.c
==
--- head/sys/dev/drm2/radeon/si.c   Mon Feb  9 07:52:45 2015
(r278437)
+++ head/sys/dev/drm2/radeon/si.c   Mon Feb  9 07:56:50 2015
(r278438)
@@ -190,18 +190,18 @@ static int si_mc_load_microcode(struct r
 
switch (rdev-family) {
case CHIP_TAHITI:
-   io_mc_regs = (const u32 *)tahiti_io_mc_regs;
+   io_mc_regs = tahiti_io_mc_regs[0][0];
ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_PITCAIRN:
-   io_mc_regs = (const u32 *)pitcairn_io_mc_regs;
+   io_mc_regs = pitcairn_io_mc_regs[0][0];
ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_VERDE:
default:
-   io_mc_regs = (const u32 *)verde_io_mc_regs;
+   io_mc_regs = verde_io_mc_regs[0][0];
ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278349 - head/contrib/llvm/lib/Target/X86

[Dimitry Andric]

Author: dim
Date: Sat Feb  7 12:50:33 2015
New Revision: 278349
URL: https://svnweb.freebsd.org/changeset/base/278349

Log:
  Pull in r224884 from upstream llvm trunk (by Keno Fischer):
  
[FastIsel][X86] Fix invalid register replacement for bool args
  
Summary:
Consider the following IR:
  
 %3 = load i8* undef
 %4 = trunc i8 %3 to i1
 %5 = call %jl_value_t.0* @foo(..., i1 %4, ...)
 ret %jl_value_t.0* %5
  
Bools (that are the result of direct truncs) are lowered as whatever
the argument to the trunc was and a and 1, causing the part of the
MBB responsible for this argument to look something like this:
  
 %vreg8def,tied1 = AND8ri %vreg7kill,tied0, 1, %EFLAGSimp-def; 
GR8:%vreg8,%vreg7
  
Later, when the load is lowered, it will insert
  
 %vreg15def = MOV8rm %vreg14, 1, %noreg, 0, %noreg; mem:LD1[undef] 
GR8:%vreg15 GR64:%vreg14
  
but remember to (at the end of isel) replace vreg7 by vreg15. Now for
the bug. In fast isel lowering, we mistakenly mark vreg8 as the result
of the load instead of the trunc. This adds a fixup to have
vreg8 replaced by whatever the result of the load is as well, so
we end up with
  
 %vreg15def,tied1 = AND8ri %vreg15kill,tied0, 1, %EFLAGSimp-def; 
GR8:%vreg15
  
which is an SSA violation and causes problems later down the road.
  
This fixes PR21557.
  
Test Plan: Test test case from PR21557 is added to the test suite.
  
Reviewers: ributzka
  
Reviewed By: ributzka
  
Subscribers: llvm-commits
  
Differential Revision: http://reviews.llvm.org/D6245
  
  This fixes a possible assertion failure when compiling toolbox.cxx from
  LibreOffice 4.3.5.
  
  Reported by:  kwm

Modified:
  head/contrib/llvm/lib/Target/X86/X86FastISel.cpp

Modified: head/contrib/llvm/lib/Target/X86/X86FastISel.cpp
==
--- head/contrib/llvm/lib/Target/X86/X86FastISel.cppSat Feb  7 12:20:33 
2015(r278348)
+++ head/contrib/llvm/lib/Target/X86/X86FastISel.cppSat Feb  7 12:50:33 
2015(r278349)
@@ -2699,6 +2699,9 @@ bool X86FastISel::FastLowerCall(CallLowe
TM.Options.GuaranteedTailCallOpt))
 return false;
 
+  SmallVectorMVT, 16 OutVTs;
+  SmallVectorunsigned, 16 ArgRegs;
+
   // If this is a constant i1/i8/i16 argument, promote to i32 to avoid an extra
   // instruction. This is safe because it is common to all FastISel supported
   // calling conventions on x86.
@@ -2716,28 +2719,34 @@ bool X86FastISel::FastLowerCall(CallLowe
 
 // Passing bools around ends up doing a trunc to i1 and passing it.
 // Codegen this as an argument + and 1.
-if (auto *TI = dyn_castTruncInst(Val)) {
-  if (TI-getType()-isIntegerTy(1)  CLI.CS 
-  (TI-getParent() == CLI.CS-getInstruction()-getParent()) 
-  TI-hasOneUse()) {
-Val = castTruncInst(Val)-getOperand(0);
-unsigned ResultReg = getRegForValue(Val);
+MVT VT;
+auto *TI = dyn_castTruncInst(Val);
+unsigned ResultReg;
+if (TI  TI-getType()-isIntegerTy(1)  CLI.CS 
+  (TI-getParent() == CLI.CS-getInstruction()-getParent()) 
+  TI-hasOneUse()) {
+  Value *PrevVal = TI-getOperand(0);
+  ResultReg = getRegForValue(PrevVal);
+
+  if (!ResultReg)
+return false;
 
-if (!ResultReg)
-  return false;
+  if (!isTypeLegal(PrevVal-getType(), VT))
+return false;
 
-MVT ArgVT;
-if (!isTypeLegal(Val-getType(), ArgVT))
-  return false;
+  ResultReg =
+FastEmit_ri(VT, VT, ISD::AND, ResultReg, hasTrivialKill(PrevVal), 1);
 
-ResultReg =
-  FastEmit_ri(ArgVT, ArgVT, ISD::AND, ResultReg, Val-hasOneUse(), 1);
-
-if (!ResultReg)
-  return false;
-UpdateValueMap(Val, ResultReg);
-  }
+  if (!ResultReg)
+return false;
+} else {
+  if (!isTypeLegal(Val-getType(), VT))
+return false;
+  ResultReg = getRegForValue(Val);
 }
+
+ArgRegs.push_back(ResultReg);
+OutVTs.push_back(VT);
   }
 
   // Analyze operands of the call, assigning locations to each operand.
@@ -2749,13 +2758,6 @@ bool X86FastISel::FastLowerCall(CallLowe
   if (IsWin64)
 CCInfo.AllocateStack(32, 8);
 
-  SmallVectorMVT, 16 OutVTs;
-  for (auto *Val : OutVals) {
-MVT VT;
-if (!isTypeLegal(Val-getType(), VT))
-  return false;
-OutVTs.push_back(VT);
-  }
   CCInfo.AnalyzeCallOperands(OutVTs, OutFlags, CC_X86);
 
   // Get a count of how many bytes are to be pushed on the stack.
@@ -2777,9 +2779,7 @@ bool X86FastISel::FastLowerCall(CallLowe
 if (ArgVT == MVT::x86mmx)
   return false;
 
-unsigned ArgReg = getRegForValue(ArgVal);
-if (!ArgReg)
-  return false;
+unsigned ArgReg = ArgRegs[VA.getValNo()];
 
 // Promote the value if needed.
 switch (VA.getLocInfo()) {
___

svn commit: r278348 - head/sys/contrib/dev/ath/ath_hal/ar9300

[Dimitry Andric]

Author: dim
Date: Sat Feb  7 12:20:33 2015
New Revision: 278348
URL: https://svnweb.freebsd.org/changeset/base/278348

Log:
  Fix a number of -Wcast-qual warnings in ath's ar9300_attach.c, by making
  the ia_array field of struct ar9300_ini_array const, and removing the
  const-dropping casts.  No functional change.
  
  Reviewed by:  adrian
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D1725

Modified:
  head/sys/contrib/dev/ath/ath_hal/ar9300/ar9300.h

Modified: head/sys/contrib/dev/ath/ath_hal/ar9300/ar9300.h
==
--- head/sys/contrib/dev/ath/ath_hal/ar9300/ar9300.hSat Feb  7 08:47:15 
2015(r278347)
+++ head/sys/contrib/dev/ath/ath_hal/ar9300/ar9300.hSat Feb  7 12:20:33 
2015(r278348)
@@ -317,12 +317,12 @@ typedef struct {
 
 /* Support for multiple INIs */
 struct ar9300_ini_array {
-u_int32_t *ia_array;
+const u_int32_t *ia_array;
 u_int32_t ia_rows;
 u_int32_t ia_columns;
 };
 #define INIT_INI_ARRAY(iniarray, array, rows, columns) do { \
-(iniarray)-ia_array = (u_int32_t *)(array);\
+(iniarray)-ia_array = (const u_int32_t *)(array);\
 (iniarray)-ia_rows = (rows);   \
 (iniarray)-ia_columns = (columns); \
 } while (0)
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278350 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Sat Feb  7 12:52:34 2015
New Revision: 278350
URL: https://svnweb.freebsd.org/changeset/base/278350

Log:
  Add llvm patch corresponding to r278349.

Added:
  head/contrib/llvm/patches/patch-32-llvm-r224884-invalid-reg-replacement.diff

Added: 
head/contrib/llvm/patches/patch-32-llvm-r224884-invalid-reg-replacement.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ 
head/contrib/llvm/patches/patch-32-llvm-r224884-invalid-reg-replacement.diff
Sat Feb  7 12:52:34 2015(r278350)
@@ -0,0 +1,164 @@
+Pull in r224884 from upstream llvm trunk (by Keno Fischer):
+
+  [FastIsel][X86] Fix invalid register replacement for bool args
+
+  Summary:
+  Consider the following IR:
+
+   %3 = load i8* undef
+   %4 = trunc i8 %3 to i1
+   %5 = call %jl_value_t.0* @foo(..., i1 %4, ...)
+   ret %jl_value_t.0* %5
+
+  Bools (that are the result of direct truncs) are lowered as whatever
+  the argument to the trunc was and a and 1, causing the part of the
+  MBB responsible for this argument to look something like this:
+
+   %vreg8def,tied1 = AND8ri %vreg7kill,tied0, 1, %EFLAGSimp-def; 
GR8:%vreg8,%vreg7
+
+  Later, when the load is lowered, it will insert
+
+   %vreg15def = MOV8rm %vreg14, 1, %noreg, 0, %noreg; mem:LD1[undef] 
GR8:%vreg15 GR64:%vreg14
+
+  but remember to (at the end of isel) replace vreg7 by vreg15. Now for
+  the bug. In fast isel lowering, we mistakenly mark vreg8 as the result
+  of the load instead of the trunc. This adds a fixup to have
+  vreg8 replaced by whatever the result of the load is as well, so
+  we end up with
+
+   %vreg15def,tied1 = AND8ri %vreg15kill,tied0, 1, %EFLAGSimp-def; 
GR8:%vreg15
+
+  which is an SSA violation and causes problems later down the road.
+
+  This fixes PR21557.
+
+  Test Plan: Test test case from PR21557 is added to the test suite.
+
+  Reviewers: ributzka
+
+  Reviewed By: ributzka
+
+  Subscribers: llvm-commits
+
+  Differential Revision: http://reviews.llvm.org/D6245
+
+This fixes a possible assertion failure when compiling toolbox.cxx from
+LibreOffice 4.3.5.
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/278349
+
+Index: lib/Target/X86/X86FastISel.cpp
+===
+--- lib/Target/X86/X86FastISel.cpp
 lib/Target/X86/X86FastISel.cpp
+@@ -2699,6 +2699,9 @@ bool X86FastISel::FastLowerCall(CallLoweringInfo 
+TM.Options.GuaranteedTailCallOpt))
+ return false;
+ 
++  SmallVectorMVT, 16 OutVTs;
++  SmallVectorunsigned, 16 ArgRegs;
++
+   // If this is a constant i1/i8/i16 argument, promote to i32 to avoid an 
extra
+   // instruction. This is safe because it is common to all FastISel supported
+   // calling conventions on x86.
+@@ -2716,28 +2719,34 @@ bool X86FastISel::FastLowerCall(CallLoweringInfo 
+ 
+ // Passing bools around ends up doing a trunc to i1 and passing it.
+ // Codegen this as an argument + and 1.
+-if (auto *TI = dyn_castTruncInst(Val)) {
+-  if (TI-getType()-isIntegerTy(1)  CLI.CS 
+-  (TI-getParent() == CLI.CS-getInstruction()-getParent()) 
+-  TI-hasOneUse()) {
+-Val = castTruncInst(Val)-getOperand(0);
+-unsigned ResultReg = getRegForValue(Val);
++MVT VT;
++auto *TI = dyn_castTruncInst(Val);
++unsigned ResultReg;
++if (TI  TI-getType()-isIntegerTy(1)  CLI.CS 
++  (TI-getParent() == CLI.CS-getInstruction()-getParent()) 
++  TI-hasOneUse()) {
++  Value *PrevVal = TI-getOperand(0);
++  ResultReg = getRegForValue(PrevVal);
+ 
+-if (!ResultReg)
+-  return false;
++  if (!ResultReg)
++return false;
+ 
+-MVT ArgVT;
+-if (!isTypeLegal(Val-getType(), ArgVT))
+-  return false;
++  if (!isTypeLegal(PrevVal-getType(), VT))
++return false;
+ 
+-ResultReg =
+-  FastEmit_ri(ArgVT, ArgVT, ISD::AND, ResultReg, Val-hasOneUse(), 1);
++  ResultReg =
++FastEmit_ri(VT, VT, ISD::AND, ResultReg, hasTrivialKill(PrevVal), 1);
+ 
+-if (!ResultReg)
+-  return false;
+-UpdateValueMap(Val, ResultReg);
+-  }
++  if (!ResultReg)
++return false;
++} else {
++  if (!isTypeLegal(Val-getType(), VT))
++return false;
++  ResultReg = getRegForValue(Val);
+ }
++
++ArgRegs.push_back(ResultReg);
++OutVTs.push_back(VT);
+   }
+ 
+   // Analyze operands of the call, assigning locations to each operand.
+@@ -2749,13 +2758,6 @@ bool X86FastISel::FastLowerCall(CallLoweringInfo 
+   if (IsWin64)
+ CCInfo.AllocateStack(32, 8);
+ 
+-  SmallVectorMVT, 16 OutVTs;
+-  for (auto *Val : OutVals) {
+-MVT VT;
+-if (!isTypeLegal(Val-getType(), VT))
+-  return false;
+-OutVTs.push_back(VT);
+-  }
+   CCInfo.AnalyzeCallOperands(OutVTs, OutFlags, CC_X86);
+ 
+   // Get a count of how many bytes are 

Re: svn commit: r278349 - head/contrib/llvm/lib/Target/X86

[Dimitry Andric]

On 07 Feb 2015, at 17:42, Matthew D. Fuller fulle...@over-yonder.net wrote:
 
 On Sat, Feb 07, 2015 at 12:50:34PM + I heard the voice of
 Dimitry Andric, and lo! it spake thus:
 
 Log:
  Pull in r224884 from upstream llvm trunk (by Keno Fischer):
 
 This blows up the kernel build for me.  Apparently reproducible, as my
 first build was -j'd and so unreadable, and another non--j build
 yielded the below.  With the rev patch -R'd out, build completes fine.
...
 Cannot emit physreg copy instruction
 UNREACHABLE executed at 
 /usr/src/lib/clang/libllvmx86codegen/../../../contrib/llvm/lib/Target/X86/X86InstrInfo.cpp:3176!
 Stack dump:
 0.Program arguments: /usr/obj/usr/src/tmp/usr/bin/cc -cc1 -triple 
 x86_64-unknown-freebsd11.0 -emit-obj -mrelax-all -disable-free 
 -main-file-name hptproc.c -mrelocation-model static -mdisable-fp-elim 
 -masm-verbose -mconstructor-aliases -mcode-model kernel -target-cpu x86-64 
 -target-feature -mmx -target-feature -sse -target-feature -aes 
 -target-feature -avx -disable-red-zone -no-implicit-float -gdwarf-2 
 -dwarf-column-info -coverage-file 
 /usr/obj/usr/src/sys/MORTIS/modules/usr/src/sys/modules/hptmv/hptproc.o 
 -nostdsysteminc -nobuiltininc -resource-dir 
 /usr/obj/usr/src/tmp/usr/bin/../lib/clang/3.5.1 -include 
 /usr/obj/usr/src/sys/MORTIS/opt_global.h -D _KERNEL -D KLD_MODULE -D 
 HAVE_KERNEL_OPTION_HEADERS -I . -I /usr/src/sys -I /usr/src/sys/contrib/altq 
 -I /usr/obj/usr/src/sys/MORTIS -isysroot /usr/obj/usr/src/tmp -Werror 
 -Wno-error-tautological-compare -Wno-error-empty-body 
 -Wno-error-parentheses-equality -Wno-error-unused-function 
 -Wno-error-pointer-sign -Wall -Wredundant-decls -Wnested-externs 
 -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual 
 -Wundef -Wno-pointer-sign -Wmissing-include-dirs -Wno-unknown-pragmas 
 -Wno-error-tautological-compare -Wno-error-empty-body 
 -Wno-error-parentheses-equality -Wno-error-unused-function 
 -Wno-error-pointer-sign -std=iso9899:1999 -fdebug-compilation-dir 
 /usr/obj/usr/src/sys/MORTIS/modules/usr/src/sys/modules/hptmv -ferror-limit 
 19 -fmessage-length 80 -ffreestanding -fformat-extensions -fwrapv 
 -stack-protector 1 -mstackrealign -fobjc-runtime=gnustep -fno-common 
 -fdiagnostics-show-option -fcolor-diagnostics -o hptproc.o -x c 
 /usr/src/sys/modules/hptmv/../../dev/hptmv/hptproc.c
 1.eof parser at end of file
 2.Code generation
 3.Running pass 'Function Pass Manager' on module 
 '/usr/src/sys/modules/hptmv/../../dev/hptmv/hptproc.c'.
 4.Running pass 'Post-RA pseudo instruction expansion pass' on function 
 '@hpt_proc_in'

Right, I also got the same report from Sean Bruno.  I've backed this out
in r278361 for now, since building head is more important than one
particular file from LibreOffice.  Sorry for the breakage.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r278361 - in head/contrib/llvm: lib/Target/X86 patches

[Dimitry Andric]

Author: dim
Date: Sat Feb  7 16:57:32 2015
New Revision: 278361
URL: https://svnweb.freebsd.org/changeset/base/278361

Log:
  Back out r278349 and r278350 for now, since this apparently blows up the
  kernel build in sys/dev/hptmv/hptproc.c for some people.
  
  Reported by:  sbruno, Matthew Fuller fulle...@over-yonder.net

Deleted:
  head/contrib/llvm/patches/patch-32-llvm-r224884-invalid-reg-replacement.diff
Modified:
  head/contrib/llvm/lib/Target/X86/X86FastISel.cpp

Modified: head/contrib/llvm/lib/Target/X86/X86FastISel.cpp
==
--- head/contrib/llvm/lib/Target/X86/X86FastISel.cppSat Feb  7 14:31:51 
2015(r278360)
+++ head/contrib/llvm/lib/Target/X86/X86FastISel.cppSat Feb  7 16:57:32 
2015(r278361)
@@ -2699,9 +2699,6 @@ bool X86FastISel::FastLowerCall(CallLowe
TM.Options.GuaranteedTailCallOpt))
 return false;
 
-  SmallVectorMVT, 16 OutVTs;
-  SmallVectorunsigned, 16 ArgRegs;
-
   // If this is a constant i1/i8/i16 argument, promote to i32 to avoid an extra
   // instruction. This is safe because it is common to all FastISel supported
   // calling conventions on x86.
@@ -2719,34 +2716,28 @@ bool X86FastISel::FastLowerCall(CallLowe
 
 // Passing bools around ends up doing a trunc to i1 and passing it.
 // Codegen this as an argument + and 1.
-MVT VT;
-auto *TI = dyn_castTruncInst(Val);
-unsigned ResultReg;
-if (TI  TI-getType()-isIntegerTy(1)  CLI.CS 
-  (TI-getParent() == CLI.CS-getInstruction()-getParent()) 
-  TI-hasOneUse()) {
-  Value *PrevVal = TI-getOperand(0);
-  ResultReg = getRegForValue(PrevVal);
-
-  if (!ResultReg)
-return false;
+if (auto *TI = dyn_castTruncInst(Val)) {
+  if (TI-getType()-isIntegerTy(1)  CLI.CS 
+  (TI-getParent() == CLI.CS-getInstruction()-getParent()) 
+  TI-hasOneUse()) {
+Val = castTruncInst(Val)-getOperand(0);
+unsigned ResultReg = getRegForValue(Val);
 
-  if (!isTypeLegal(PrevVal-getType(), VT))
-return false;
+if (!ResultReg)
+  return false;
 
-  ResultReg =
-FastEmit_ri(VT, VT, ISD::AND, ResultReg, hasTrivialKill(PrevVal), 1);
+MVT ArgVT;
+if (!isTypeLegal(Val-getType(), ArgVT))
+  return false;
 
-  if (!ResultReg)
-return false;
-} else {
-  if (!isTypeLegal(Val-getType(), VT))
-return false;
-  ResultReg = getRegForValue(Val);
-}
+ResultReg =
+  FastEmit_ri(ArgVT, ArgVT, ISD::AND, ResultReg, Val-hasOneUse(), 1);
 
-ArgRegs.push_back(ResultReg);
-OutVTs.push_back(VT);
+if (!ResultReg)
+  return false;
+UpdateValueMap(Val, ResultReg);
+  }
+}
   }
 
   // Analyze operands of the call, assigning locations to each operand.
@@ -2758,6 +2749,13 @@ bool X86FastISel::FastLowerCall(CallLowe
   if (IsWin64)
 CCInfo.AllocateStack(32, 8);
 
+  SmallVectorMVT, 16 OutVTs;
+  for (auto *Val : OutVals) {
+MVT VT;
+if (!isTypeLegal(Val-getType(), VT))
+  return false;
+OutVTs.push_back(VT);
+  }
   CCInfo.AnalyzeCallOperands(OutVTs, OutFlags, CC_X86);
 
   // Get a count of how many bytes are to be pushed on the stack.
@@ -2779,7 +2777,9 @@ bool X86FastISel::FastLowerCall(CallLowe
 if (ArgVT == MVT::x86mmx)
   return false;
 
-unsigned ArgReg = ArgRegs[VA.getValNo()];
+unsigned ArgReg = getRegForValue(ArgVal);
+if (!ArgReg)
+  return false;
 
 // Promote the value if needed.
 switch (VA.getLocInfo()) {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278297 - head/usr.sbin/syslogd

[Dimitry Andric]

Author: dim
Date: Thu Feb  5 22:28:00 2015
New Revision: 278297
URL: https://svnweb.freebsd.org/changeset/base/278297

Log:
  Fix two clang 3.6.0 warnings in usr.sbin/syslogd:
  
  usr.sbin/syslogd/syslogd.c:1023:10: error: address of array 'f-f_prevline' 
will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  f-f_prevline  !strcmp(msg, f-f_prevline) 
  ~~~^~
  usr.sbin/syslogd/syslogd.c:1178:16: error: address of array 'f-f_prevline' 
will always evaluate to 'true' [-Werror,-Wpointer-bool-conversion]
  } else if (f-f_prevline) {
 ~~  ~~~^~
  
  In both cases, the f_prevline field of struct filed is a char array, so
  it can never be null.  Remove the checks.
  
  Reviewed by:  jilles
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D1716

Modified:
  head/usr.sbin/syslogd/syslogd.c

Modified: head/usr.sbin/syslogd/syslogd.c
==
--- head/usr.sbin/syslogd/syslogd.c Thu Feb  5 22:24:22 2015
(r278296)
+++ head/usr.sbin/syslogd/syslogd.c Thu Feb  5 22:28:00 2015
(r278297)
@@ -1020,7 +1020,7 @@ logmsg(int pri, const char *msg, const c
 */
if (no_compress - (f-f_type != F_PIPE)  1 
(flags  MARK) == 0  msglen == f-f_prevlen 
-   f-f_prevline  !strcmp(msg, f-f_prevline) 
+   !strcmp(msg, f-f_prevline) 
!strcasecmp(from, f-f_prevhost)) {
(void)strlcpy(f-f_lasttime, timestamp,
sizeof(f-f_lasttime));
@@ -1175,11 +1175,9 @@ fprintlog(struct filed *f, int flags, co
v-iov_base = repbuf;
v-iov_len = snprintf(repbuf, sizeof repbuf,
last message repeated %d times, f-f_prevcount);
-   } else if (f-f_prevline) {
+   } else {
v-iov_base = f-f_prevline;
v-iov_len = f-f_prevlen;
-   } else {
-   return;
}
v++;
 
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278222 - head/sys/sys

[Dimitry Andric]

Author: dim
Date: Wed Feb  4 20:55:21 2015
New Revision: 278222
URL: https://svnweb.freebsd.org/changeset/base/278222

Log:
  Mark typedefs for manually implementing _Static_assert() as unused, so
  they won't show up unecessarily for -Wunused-local-typedefs.
  
  MFC after:3 days

Modified:
  head/sys/sys/cdefs.h

Modified: head/sys/sys/cdefs.h
==
--- head/sys/sys/cdefs.hWed Feb  4 20:03:57 2015(r278221)
+++ head/sys/sys/cdefs.hWed Feb  4 20:55:21 2015(r278222)
@@ -293,7 +293,8 @@
 #elif defined(__COUNTER__)
 #define_Static_assert(x, y)__Static_assert(x, __COUNTER__)
 #define__Static_assert(x, y)   ___Static_assert(x, y)
-#define___Static_assert(x, y)  typedef char __assert_ ## y[(x) ? 1 : 
-1]
+#define___Static_assert(x, y)  typedef char __assert_ ## y[(x) ? 1 : 
-1] \
+   __unused
 #else
 #define_Static_assert(x, y)struct __hack
 #endif
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r278223 - head/lib/clang

[Dimitry Andric]

On 04 Feb 2015, at 22:09, NGie Cooper yaneurab...@gmail.com wrote:
 
 On Wed, Feb 4, 2015 at 1:00 PM, Dimitry Andric d...@freebsd.org wrote:
 Author: dim
 Date: Wed Feb  4 21:00:29 2015
 New Revision: 278223
 URL: https://svnweb.freebsd.org/changeset/base/278223
 
 Log:
  For now, add -stdlib=libc++ to the flags for building clang, since that
  makes it easier to build head on stable/9, where libstdc++ is still the
  default.  We can revisit this when somebody will try to build base with
  gcc 4.8.1 or higher, and its included libstdc++.
 
 Maybe this should be conditionalized on COMPILER_TYPE?

COMPILER_TYPE doesn't say anything about the C++ library used.  Maybe we
need a CXX_LIBRARY_TYPE setting for it?

That said, you may be right that it should be conditionalized on the
compiler type anyway, since gcc does not support the -stdlib= option.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r278225 - head/lib/clang

[Dimitry Andric]

Author: dim
Date: Wed Feb  4 21:48:50 2015
New Revision: 278225
URL: https://svnweb.freebsd.org/changeset/base/278225

Log:
  Followup to r278223, by only using -stdlib=libc++ when the compiler is
  clang; not even recent versions of gcc support the -stdlib flag.
  
  Noticed by:   ngie

Modified:
  head/lib/clang/clang.build.mk

Modified: head/lib/clang/clang.build.mk
==
--- head/lib/clang/clang.build.mk   Wed Feb  4 21:08:28 2015
(r278224)
+++ head/lib/clang/clang.build.mk   Wed Feb  4 21:48:50 2015
(r278225)
@@ -34,7 +34,8 @@ BUILD_TRIPLE?=${BUILD_ARCH:C/amd64/x86_
 CFLAGS+=   -DLLVM_DEFAULT_TARGET_TRIPLE=\${TARGET_TRIPLE}\ \
-DLLVM_HOST_TRIPLE=\${BUILD_TRIPLE}\ \
-DDEFAULT_SYSROOT=\${TOOLS_PREFIX}\
-CXXFLAGS+= -std=c++11 -stdlib=libc++ -fno-exceptions -fno-rtti
+CXXFLAGS+= -std=c++11 -fno-exceptions -fno-rtti
+CXXFLAGS.clang+= -stdlib=libc++
 
 .PATH: ${LLVM_SRCS}/${SRCDIR}
 
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278223 - head/lib/clang

[Dimitry Andric]

Author: dim
Date: Wed Feb  4 21:00:29 2015
New Revision: 278223
URL: https://svnweb.freebsd.org/changeset/base/278223

Log:
  For now, add -stdlib=libc++ to the flags for building clang, since that
  makes it easier to build head on stable/9, where libstdc++ is still the
  default.  We can revisit this when somebody will try to build base with
  gcc 4.8.1 or higher, and its included libstdc++.
  
  Reported by:  rpaulo

Modified:
  head/lib/clang/clang.build.mk

Modified: head/lib/clang/clang.build.mk
==
--- head/lib/clang/clang.build.mk   Wed Feb  4 20:55:21 2015
(r278222)
+++ head/lib/clang/clang.build.mk   Wed Feb  4 21:00:29 2015
(r278223)
@@ -34,7 +34,7 @@ BUILD_TRIPLE?=${BUILD_ARCH:C/amd64/x86_
 CFLAGS+=   -DLLVM_DEFAULT_TARGET_TRIPLE=\${TARGET_TRIPLE}\ \
-DLLVM_HOST_TRIPLE=\${BUILD_TRIPLE}\ \
-DDEFAULT_SYSROOT=\${TOOLS_PREFIX}\
-CXXFLAGS+= -std=c++11 -fno-exceptions -fno-rtti
+CXXFLAGS+= -std=c++11 -stdlib=libc++ -fno-exceptions -fno-rtti
 
 .PATH: ${LLVM_SRCS}/${SRCDIR}
 
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r278223 - head/lib/clang

[Dimitry Andric]

On 04 Feb 2015, at 22:23, NGie Cooper yaneurab...@gmail.com wrote:
 
 On Wed, Feb 4, 2015 at 1:16 PM, Dimitry Andric d...@freebsd.org wrote:
 On 04 Feb 2015, at 22:09, NGie Cooper yaneurab...@gmail.com wrote:
 
 On Wed, Feb 4, 2015 at 1:00 PM, Dimitry Andric d...@freebsd.org wrote:
 Author: dim
 Date: Wed Feb  4 21:00:29 2015
 New Revision: 278223
 URL: https://svnweb.freebsd.org/changeset/base/278223
 
 Log:
 For now, add -stdlib=libc++ to the flags for building clang, since that
 makes it easier to build head on stable/9, where libstdc++ is still the
 default.  We can revisit this when somebody will try to build base with
 gcc 4.8.1 or higher, and its included libstdc++.
 
 Maybe this should be conditionalized on COMPILER_TYPE?
 
 COMPILER_TYPE doesn't say anything about the C++ library used.  Maybe we
 need a CXX_LIBRARY_TYPE setting for it?
 
 That said, you may be right that it should be conditionalized on the
 compiler type anyway, since gcc does not support the -stdlib= option.
 
Yeah... true.
My original point was coming more from a perspective that this was
 being done strictly for the GNU toolchain, so conditionalizing the
 CFLAGS on the compiler being used seemed reasonable.

This flag is only relevant when you are 1) using clang to compile, and
2) on freebsd 9.x.  You cannot compile clang in head with base gcc
anyway.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r278109 - head/contrib/llvm/patches

[Dimitry Andric]

Author: dim
Date: Mon Feb  2 20:05:52 2015
New Revision: 278109
URL: https://svnweb.freebsd.org/changeset/base/278109

Log:
  Belatedly add the clang patch corresponding to r277423.

Added:
  head/contrib/llvm/patches/patch-29-clang-add-mips-triples.diff
  head/contrib/llvm/patches/patch-30-llvm-r226664-aarch64-x18.diff
 - copied unchanged from r278063, 
head/contrib/llvm/patches/patch-29-llvm-r226664-aarch64-x18.diff
  head/contrib/llvm/patches/patch-31-clang-r227062-fixes-x18.diff
 - copied unchanged from r278063, 
head/contrib/llvm/patches/patch-30-clang-r227062-fixes-x18.diff
Deleted:
  head/contrib/llvm/patches/patch-29-llvm-r226664-aarch64-x18.diff
  head/contrib/llvm/patches/patch-30-clang-r227062-fixes-x18.diff

Added: head/contrib/llvm/patches/patch-29-clang-add-mips-triples.diff
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-29-clang-add-mips-triples.diff  Mon Feb 
 2 20:05:52 2015(r278109)
@@ -0,0 +1,33 @@
+Allow clang to be built for mips/mips64 backend types by adding our mips
+triple ids
+
+This only allows testing and does not change the defaults for mips/mips64.
+They still build/use gcc by default.
+
+Differential Revision: https://reviews.freebsd.org/D1190
+Reviewed by:   dim
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/277423
+
+Index: tools/clang/lib/Driver/Tools.cpp
+===
+--- tools/clang/lib/Driver/Tools.cpp
 tools/clang/lib/Driver/Tools.cpp
+@@ -6592,6 +6592,17 @@ void freebsd::Link::ConstructJob(Compilation C, c
+ CmdArgs.push_back(elf32ppc_fbsd);
+   }
+ 
++  if (Arg *A = Args.getLastArg(options::OPT_G)) {
++if (ToolChain.getArch() == llvm::Triple::mips ||
++  ToolChain.getArch() == llvm::Triple::mipsel ||
++  ToolChain.getArch() == llvm::Triple::mips64 ||
++  ToolChain.getArch() == llvm::Triple::mips64el) {
++  StringRef v = A-getValue();
++  CmdArgs.push_back(Args.MakeArgString(-G + v));
++  A-claim();
++}
++  }
++
+   if (Output.isFilename()) {
+ CmdArgs.push_back(-o);
+ CmdArgs.push_back(Output.getFilename());

Copied: head/contrib/llvm/patches/patch-30-llvm-r226664-aarch64-x18.diff (from 
r278063, head/contrib/llvm/patches/patch-29-llvm-r226664-aarch64-x18.diff)
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/contrib/llvm/patches/patch-30-llvm-r226664-aarch64-x18.diffMon Feb 
 2 20:05:52 2015(r278109, copy of r278063, 
head/contrib/llvm/patches/patch-29-llvm-r226664-aarch64-x18.diff)
@@ -0,0 +1,83 @@
+Pull in r226664 from upstream llvm trunk (by Tim Northover):
+
+  AArch64: add backend option to reserve x18 (platform register)
+
+  AAPCS64 says that it's up to the platform to specify whether x18 is
+  reserved, and a first step on that way is to add a flag controlling
+  it.
+
+  From: Andrew Turner and...@fubar.geek.nz
+
+Introduced here: http://svnweb.freebsd.org/changeset/base/24
+
+Index: lib/Target/AArch64/AArch64RegisterInfo.cpp
+===
+--- lib/Target/AArch64/AArch64RegisterInfo.cpp
 lib/Target/AArch64/AArch64RegisterInfo.cpp
+@@ -33,6 +33,10 @@ using namespace llvm;
+ #define GET_REGINFO_TARGET_DESC
+ #include AArch64GenRegisterInfo.inc
+ 
++static cl::optbool
++ReserveX18(aarch64-reserve-x18, cl::Hidden,
++  cl::desc(Reserve X18, making it unavailable as GPR));
++
+ AArch64RegisterInfo::AArch64RegisterInfo(const AArch64InstrInfo *tii,
+  const AArch64Subtarget *sti)
+ : AArch64GenRegisterInfo(AArch64::LR), TII(tii), STI(sti) {}
+@@ -90,7 +94,7 @@ AArch64RegisterInfo::getReservedRegs(const Machine
+ Reserved.set(AArch64::W29);
+   }
+ 
+-  if (STI-isTargetDarwin()) {
++  if (STI-isTargetDarwin() || ReserveX18) {
+ Reserved.set(AArch64::X18); // Platform register
+ Reserved.set(AArch64::W18);
+   }
+@@ -117,7 +121,7 @@ bool AArch64RegisterInfo::isReservedReg(const Mach
+ return true;
+   case AArch64::X18:
+   case AArch64::W18:
+-return STI-isTargetDarwin();
++return STI-isTargetDarwin() || ReserveX18;
+   case AArch64::FP:
+   case AArch64::W29:
+ return TFI-hasFP(MF) || STI-isTargetDarwin();
+@@ -379,7 +383,7 @@ unsigned AArch64RegisterInfo::getRegPressureLimit(
+   case AArch64::GPR64commonRegClassID:
+ return 32 - 1  // XZR/SP
+- (TFI-hasFP(MF) || STI-isTargetDarwin()) // FP
+-   - STI-isTargetDarwin() // X18 reserved as platform register
++   - (STI-isTargetDarwin() || ReserveX18) // X18 reserved as 
platform register
+- hasBasePointer(MF);   // X19
+   case AArch64::FPR8RegClassID:
+   case AArch64::FPR16RegClassID:
+Index: 

Re: svn commit: r278004 - head/sys/dev/drm2/radeon

[Dimitry Andric]

On 01 Feb 2015, at 06:00, Bruce Evans b...@optusnet.com.au wrote:
 
 On Sat, 31 Jan 2015, Dimitry Andric wrote:
 
 Log:
 Constify a number of accesses in drm2's radeon drivers to avoid
 -Wcast-qual warnings.  No functional change.
 
 This is much better than using __DECONST(), but still has bogus casts.
...
 
 Anyway, it is clearer to point to the
 first element:
 
   io_mc_regs = barts_io_mc_regs[0][0];

Thanks, I've put that up for review, here:

https://reviews.freebsd.org/D1748

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r277974 - head/sys/dev/usb/input

[Dimitry Andric]

Author: dim
Date: Sat Jan 31 14:18:46 2015
New Revision: 277974
URL: https://svnweb.freebsd.org/changeset/base/277974

Log:
  Fix a bunch of -Wcast-qual warnings in sys/dev/usb/input/uhid.c, by
  using __DECONST.  No functional change.
  
  Reviewed by:  hselasky
  Differential Revision: https://reviews.freebsd.org/D1743

Modified:
  head/sys/dev/usb/input/uhid.c

Modified: head/sys/dev/usb/input/uhid.c
==
--- head/sys/dev/usb/input/uhid.c   Sat Jan 31 13:53:29 2015
(r277973)
+++ head/sys/dev/usb/input/uhid.c   Sat Jan 31 14:18:46 2015
(r277974)
@@ -734,7 +734,7 @@ uhid_attach(device_t dev)
if (uaa-info.idProduct == USB_PRODUCT_WACOM_GRAPHIRE) {
 
sc-sc_repdesc_size = 
sizeof(uhid_graphire_report_descr);
-   sc-sc_repdesc_ptr = (void 
*)uhid_graphire_report_descr;
+   sc-sc_repdesc_ptr = __DECONST(void *, 
uhid_graphire_report_descr);
sc-sc_flags |= UHID_FLAG_STATIC_DESC;
 
} else if (uaa-info.idProduct == 
USB_PRODUCT_WACOM_GRAPHIRE3_4X5) {
@@ -755,7 +755,7 @@ uhid_attach(device_t dev)
usbd_errstr(error));
}
sc-sc_repdesc_size = 
sizeof(uhid_graphire3_4x5_report_descr);
-   sc-sc_repdesc_ptr = (void 
*)uhid_graphire3_4x5_report_descr;
+   sc-sc_repdesc_ptr = __DECONST(void *, 
uhid_graphire3_4x5_report_descr);
sc-sc_flags |= UHID_FLAG_STATIC_DESC;
}
} else if ((uaa-info.bInterfaceClass == UICLASS_VENDOR) 
@@ -775,7 +775,7 @@ uhid_attach(device_t dev)
}
/* the Xbox 360 gamepad has no report descriptor */
sc-sc_repdesc_size = sizeof(uhid_xb360gp_report_descr);
-   sc-sc_repdesc_ptr = (void *)uhid_xb360gp_report_descr;
+   sc-sc_repdesc_ptr = __DECONST(void *, 
uhid_xb360gp_report_descr);
sc-sc_flags |= UHID_FLAG_STATIC_DESC;
}
if (sc-sc_repdesc_ptr == NULL) {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278010 - head/contrib/libcxxrt

[Dimitry Andric]

Author: dim
Date: Sat Jan 31 23:08:29 2015
New Revision: 278010
URL: https://svnweb.freebsd.org/changeset/base/278010

Log:
  Revert r256642, not only to reduce diffs against upstream libcxxrt, but
  also because it is the wrong approach: comparing typeinfo names deeply
  causes trouble if two loaded DSOs use independent types of the same
  name.
  
  In addition, this particular change was never merged to FreeBSD 10.x and
  9.x, so let's get rid of it before it ends up in an 11.x release.
  
  Discussed with:   theraven, joerg@netbsd

Modified:
  head/contrib/libcxxrt/typeinfo.cc

Modified: head/contrib/libcxxrt/typeinfo.cc
==
--- head/contrib/libcxxrt/typeinfo.cc   Sat Jan 31 23:02:27 2015
(r278009)
+++ head/contrib/libcxxrt/typeinfo.cc   Sat Jan 31 23:08:29 2015
(r278010)
@@ -35,23 +35,15 @@ type_info::~type_info() {}
 
 bool type_info::operator==(const type_info other) const
 {
-#ifdef LIBCXXRT_MERGED_TYPEINFO
return __type_name == other.__type_name;
-#else
-   return __type_name == other.__type_name || strcmp(__type_name, 
other.__type_name) == 0;
-#endif
 }
 bool type_info::operator!=(const type_info other) const
 {
-   return !operator==(other);
+   return __type_name != other.__type_name;
 }
 bool type_info::before(const type_info other) const
 {
-#ifdef LIBCXXRT_MERGED_TYPEINFO
return __type_name  other.__type_name;
-#else
-   return strcmp(__type_name, other.__type_name)  0;
-#endif
 }
 const char* type_info::name() const
 {
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278004 - head/sys/dev/drm2/radeon

[Dimitry Andric]

Author: dim
Date: Sat Jan 31 22:18:52 2015
New Revision: 278004
URL: https://svnweb.freebsd.org/changeset/base/278004

Log:
  Constify a number of accesses in drm2's radeon drivers to avoid
  -Wcast-qual warnings.  No functional change.
  
  Reviewed by:  dumbbell
  MFC after:3 days
  Differential Revision: https://reviews.freebsd.org/D1727

Modified:
  head/sys/dev/drm2/radeon/ni.c
  head/sys/dev/drm2/radeon/si.c

Modified: head/sys/dev/drm2/radeon/ni.c
==
--- head/sys/dev/drm2/radeon/ni.c   Sat Jan 31 22:07:38 2015
(r278003)
+++ head/sys/dev/drm2/radeon/ni.c   Sat Jan 31 22:18:52 2015
(r278004)
@@ -182,7 +182,7 @@ int ni_mc_load_microcode(struct radeon_d
 {
const __be32 *fw_data;
u32 mem_type, running, blackout = 0;
-   u32 *io_mc_regs;
+   const u32 *io_mc_regs;
int i, ucode_size, regs_size;
 
if (!rdev-mc_fw)
@@ -190,23 +190,23 @@ int ni_mc_load_microcode(struct radeon_d
 
switch (rdev-family) {
case CHIP_BARTS:
-   io_mc_regs = (u32 *)barts_io_mc_regs;
+   io_mc_regs = (const u32 *)barts_io_mc_regs;
ucode_size = BTC_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;
case CHIP_TURKS:
-   io_mc_regs = (u32 *)turks_io_mc_regs;
+   io_mc_regs = (const u32 *)turks_io_mc_regs;
ucode_size = BTC_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;
case CHIP_CAICOS:
default:
-   io_mc_regs = (u32 *)caicos_io_mc_regs;
+   io_mc_regs = (const u32 *)caicos_io_mc_regs;
ucode_size = BTC_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;
case CHIP_CAYMAN:
-   io_mc_regs = (u32 *)cayman_io_mc_regs;
+   io_mc_regs = (const u32 *)cayman_io_mc_regs;
ucode_size = CAYMAN_MC_UCODE_SIZE;
regs_size = BTC_IO_MC_REGS_SIZE;
break;

Modified: head/sys/dev/drm2/radeon/si.c
==
--- head/sys/dev/drm2/radeon/si.c   Sat Jan 31 22:07:38 2015
(r278003)
+++ head/sys/dev/drm2/radeon/si.c   Sat Jan 31 22:18:52 2015
(r278004)
@@ -182,7 +182,7 @@ static int si_mc_load_microcode(struct r
 {
const __be32 *fw_data;
u32 running, blackout = 0;
-   u32 *io_mc_regs;
+   const u32 *io_mc_regs;
int i, ucode_size, regs_size;
 
if (!rdev-mc_fw)
@@ -190,18 +190,18 @@ static int si_mc_load_microcode(struct r
 
switch (rdev-family) {
case CHIP_TAHITI:
-   io_mc_regs = (u32 *)tahiti_io_mc_regs;
+   io_mc_regs = (const u32 *)tahiti_io_mc_regs;
ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_PITCAIRN:
-   io_mc_regs = (u32 *)pitcairn_io_mc_regs;
+   io_mc_regs = (const u32 *)pitcairn_io_mc_regs;
ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
case CHIP_VERDE:
default:
-   io_mc_regs = (u32 *)verde_io_mc_regs;
+   io_mc_regs = (const u32 *)verde_io_mc_regs;
ucode_size = SI_MC_UCODE_SIZE;
regs_size = TAHITI_IO_MC_REGS_SIZE;
break;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r278016 - head/contrib/libcxxrt

[Dimitry Andric]

Author: dim
Date: Sat Jan 31 23:31:45 2015
New Revision: 278016
URL: https://svnweb.freebsd.org/changeset/base/278016

Log:
  Import libcxxrt master 1cb607e89f6135bbc10f3d3b6fba1f983e258dcc.
  
  Interesting fixes:
  1cb607e   Correct gcc version check for __cxa_begin_catch() declaration
with or without throw()
  
  MFC after:3 days

Modified:
  head/contrib/libcxxrt/exception.cc
Directory Properties:
  head/contrib/libcxxrt/   (props changed)

Modified: head/contrib/libcxxrt/exception.cc
==
--- head/contrib/libcxxrt/exception.cc  Sat Jan 31 23:24:25 2015
(r278015)
+++ head/contrib/libcxxrt/exception.cc  Sat Jan 31 23:31:45 2015
(r278016)
@@ -673,7 +673,7 @@ static _Unwind_Reason_Code trace(struct 
  * If the failure happened by falling off the end of the stack without finding
  * a handler, prints a back trace before aborting.
  */
-#if __GNUC__  3  __GNUC_MINOR__  2
+#if __GNUC__  4 || (__GNUC__ == 4  __GNUC_MINOR__ = 4)
 extern C void *__cxa_begin_catch(void *e) throw();
 #else
 extern C void *__cxa_begin_catch(void *e);
@@ -1191,7 +1191,7 @@ BEGIN_PERSONALITY_FUNCTION(__gxx_persona
  * pointer to the caught exception, which is either the adjusted pointer (for
  * C++ exceptions) of the unadjusted pointer (for foreign exceptions).
  */
-#if __GNUC__  3  __GNUC_MINOR__  2
+#if __GNUC__  4 || (__GNUC__ == 4  __GNUC_MINOR__ = 4)
 extern C void *__cxa_begin_catch(void *e) throw()
 #else
 extern C void *__cxa_begin_catch(void *e)
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r277951 - head/sys/netgraph

[Dimitry Andric]

Author: dim
Date: Fri Jan 30 21:59:53 2015
New Revision: 277951
URL: https://svnweb.freebsd.org/changeset/base/277951

Log:
  Fix a bunch of -Wcast-qual warnings in netgraph's ng_parse.c, by using
  __DECONST.  No functional change.
  
  MFC after:3 days

Modified:
  head/sys/netgraph/ng_parse.c

Modified: head/sys/netgraph/ng_parse.c
==
--- head/sys/netgraph/ng_parse.cFri Jan 30 21:22:18 2015
(r277950)
+++ head/sys/netgraph/ng_parse.cFri Jan 30 21:59:53 2015
(r277951)
@@ -1122,7 +1122,7 @@ ng_bytearray_parse(const struct ng_parse
struct ng_parse_type subtype;
 
subtype = ng_parse_bytearray_subtype;
-   *(const void **)subtype.private = type-info;
+   subtype.private = __DECONST(void *, type-info);
return ng_array_parse(subtype, s, off, start, buf, buflen);
}
 }
@@ -1134,7 +1134,7 @@ ng_bytearray_unparse(const struct ng_par
struct ng_parse_type subtype;
 
subtype = ng_parse_bytearray_subtype;
-   *(const void **)subtype.private = type-info;
+   subtype.private = __DECONST(void *, type-info);
return ng_array_unparse(subtype, data, off, cbuf, cbuflen);
 }
 
@@ -1145,7 +1145,7 @@ ng_bytearray_getDefault(const struct ng_
struct ng_parse_type subtype;
 
subtype = ng_parse_bytearray_subtype;
-   *(const void **)subtype.private = type-info;
+   subtype.private = __DECONST(void *, type-info);
return ng_array_getDefault(subtype, start, buf, buflen);
 }
 
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r277952 - head/sys/fs/udf

[Dimitry Andric]

Author: dim
Date: Fri Jan 30 22:01:45 2015
New Revision: 277952
URL: https://svnweb.freebsd.org/changeset/base/277952

Log:
  Fix a -Wcast-qual warning in udf_vnops.c, by using __DECONST.  No
  functional change.
  
  MFC after:3 days

Modified:
  head/sys/fs/udf/udf_vnops.c

Modified: head/sys/fs/udf/udf_vnops.c
==
--- head/sys/fs/udf/udf_vnops.c Fri Jan 30 21:59:53 2015(r277951)
+++ head/sys/fs/udf/udf_vnops.c Fri Jan 30 22:01:45 2015(r277952)
@@ -526,8 +526,9 @@ udf_transname(char *cs0string, char *des
}
 
while (unilen  0  destleft  0) {
-   udf_iconv-conv(udfmp-im_d2l, (const char **)unibuf,
-   (size_t *)unilen, (char **)destname, 
destleft);
+   udf_iconv-conv(udfmp-im_d2l, __DECONST(const char **,
+   unibuf), (size_t *)unilen, (char **)destname,
+   destleft);
/* Unconverted character found */
if (unilen  0  destleft  0) {
*destname++ = '?';
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r277953 - head/sys/fs/smbfs

[Dimitry Andric]

Author: dim
Date: Fri Jan 30 22:02:32 2015
New Revision: 277953
URL: https://svnweb.freebsd.org/changeset/base/277953

Log:
  Fix a -Wcast-qual warning in smbfs_subr.c, by using __DECONST.  No
  functional change.
  
  MFC after:3 days

Modified:
  head/sys/fs/smbfs/smbfs_subr.c

Modified: head/sys/fs/smbfs/smbfs_subr.c
==
--- head/sys/fs/smbfs/smbfs_subr.c  Fri Jan 30 22:01:45 2015
(r277952)
+++ head/sys/fs/smbfs/smbfs_subr.c  Fri Jan 30 22:02:32 2015
(r277953)
@@ -170,8 +170,8 @@ smbfs_fname_tolocal(struct smb_vc *vcp, 
if (error) return error;
*/
 
-   error = iconv_conv_case
-   (vcp-vc_tolocal, (const char **)ibuf, ilen, obuf, 
olen, copt);
+   error = iconv_conv_case(vcp-vc_tolocal,
+   __DECONST(const char **, ibuf), ilen, obuf, olen, copt);
if (error  SMB_UNICODE_STRINGS(vcp)) {
/*
 * If using unicode, leaving a file name as it was when
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r277940 - in head/sys: conf modules/aesni

[Dimitry Andric]

Author: dim
Date: Fri Jan 30 18:17:17 2015
New Revision: 277940
URL: https://svnweb.freebsd.org/changeset/base/277940

Log:
  For clang, disable -Wcast-qual warnings for specific aesni files, since
  clang 3.6.0 will emit a number of such warnings for those files, and
  they are partially contributed code.

Modified:
  head/sys/conf/files.amd64
  head/sys/conf/kern.mk
  head/sys/modules/aesni/Makefile

Modified: head/sys/conf/files.amd64
==
--- head/sys/conf/files.amd64   Fri Jan 30 18:07:46 2015(r277939)
+++ head/sys/conf/files.amd64   Fri Jan 30 18:17:17 2015(r277940)
@@ -131,12 +131,12 @@ crypto/aesni/aeskeys_amd64.S  optional ae
 crypto/aesni/aesni.c   optional aesni
 aesni_ghash.o  optional aesni  \
dependency  $S/crypto/aesni/aesni_ghash.c \
-   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${PROF} -mmmx -msse -msse4 -maes -mpclmul ${.IMPSRC} \
+   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${NO_WCAST_QUAL} ${PROF} -mmmx -msse -msse4 -maes -mpclmul ${.IMPSRC} \
no-implicit-rule\
clean   aesni_ghash.o
 aesni_wrap.o   optional aesni  \
dependency  $S/crypto/aesni/aesni_wrap.c  \
-   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${PROF} -mmmx -msse -msse4 -maes ${.IMPSRC} \
+   compile-with${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} 
${NO_WCAST_QUAL} ${PROF} -mmmx -msse -msse4 -maes ${.IMPSRC} \
no-implicit-rule\
clean   aesni_wrap.o
 crypto/blowfish/bf_enc.c   optionalcrypto | ipsec

Modified: head/sys/conf/kern.mk
==
--- head/sys/conf/kern.mk   Fri Jan 30 18:07:46 2015(r277939)
+++ head/sys/conf/kern.mk   Fri Jan 30 18:17:17 2015(r277940)
@@ -23,6 +23,7 @@ NO_WSHIFT_COUNT_OVERFLOW= -Wno-shift-cou
 NO_WSELF_ASSIGN=   -Wno-self-assign
 NO_WUNNEEDED_INTERNAL_DECL=-Wno-unneeded-internal-declaration
 NO_WSOMETIMES_UNINITIALIZED=   -Wno-error-sometimes-uninitialized
+NO_WCAST_QUAL= -Wno-cast-qual
 # Several other warnings which might be useful in some cases, but not severe
 # enough to error out the whole kernel build.  Display them anyway, so there is
 # some incentive to fix them eventually.

Modified: head/sys/modules/aesni/Makefile
==
--- head/sys/modules/aesni/Makefile Fri Jan 30 18:07:46 2015
(r277939)
+++ head/sys/modules/aesni/Makefile Fri Jan 30 18:17:17 2015
(r277940)
@@ -22,3 +22,6 @@ aesni_wrap.o: aesni_wrap.c
${CTFCONVERT_CMD}
 
 .include bsd.kmod.mk
+
+CWARNFLAGS.aesni_ghash.c=  ${NO_WCAST_QUAL}
+CWARNFLAGS.aesni_wrap.c=   ${NO_WCAST_QUAL}
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

svn commit: r277944 - head/contrib/libcxxrt

[Dimitry Andric]

Author: dim
Date: Fri Jan 30 18:26:38 2015
New Revision: 277944
URL: https://svnweb.freebsd.org/changeset/base/277944

Log:
  Partially revert r273382, to reduce diffs against upstream.  This was a
  temporary fix to solve a conflict with an older version of libc++, and
  it is no longer relevant.
  
  MFC after:3 days

Modified:
  head/contrib/libcxxrt/stdexcept.cc

Modified: head/contrib/libcxxrt/stdexcept.cc
==
--- head/contrib/libcxxrt/stdexcept.cc  Fri Jan 30 18:25:53 2015
(r277943)
+++ head/contrib/libcxxrt/stdexcept.cc  Fri Jan 30 18:26:38 2015
(r277944)
@@ -82,19 +82,14 @@ const char* bad_typeid::what() const thr
return std::bad_typeid;
 }
 
-__attribute__((weak))
 bad_array_new_length::bad_array_new_length() throw() {}
-__attribute__((weak))
 bad_array_new_length::~bad_array_new_length() {}
-__attribute__((weak))
 bad_array_new_length::bad_array_new_length(const bad_array_new_length) 
throw() {}
-__attribute__((weak))
 bad_array_new_length bad_array_new_length::operator=(const 
bad_array_new_length) throw()
 {
return *this;
 }
 
-__attribute__((weak))
 const char* bad_array_new_length::what() const throw()
 {
return std::bad_array_new_length;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org

Re: svn commit: r277860 - head/usr.sbin/sa

[Dimitry Andric]

On 29 Jan 2015, at 09:23, John-Mark Gurney j...@funkthat.com wrote:
 
 Dimitry Andric wrote this message on Wed, Jan 28, 2015 at 22:22 +:
  Replace the VERSION_KEY define with a writable char array, so no const
  qualifier needs to be dropped anymore.
 
 You do realize that you should juse use __DECONST for this instead?
 
 This moves the data from .rodata into .data making it writable, and
 also means that bad programs could end up stop working, instead of
 faulting when someone tries to change the version string...

I don't see the risk here, the db api does not actually modify the
incoming key (the API is just badly designed).  But if you think it is
nicer to use __DECONST, feel free to change it.

-Dimitry



signature.asc
Description: Message signed with OpenPGP using GPGMail

svn commit: r277883 - head/sys/sys

[Dimitry Andric]

Author: dim
Date: Thu Jan 29 18:44:21 2015
New Revision: 277883
URL: https://svnweb.freebsd.org/changeset/base/277883

Log:
  Ensure that lint does not pick up C11 keywords (e.g.  _Noreturn), even
  if C11 mode is used.  It does not support any C11 constructs.
  
  MFC after:3 days

Modified:
  head/sys/sys/cdefs.h

Modified: head/sys/sys/cdefs.h
==
--- head/sys/sys/cdefs.hThu Jan 29 18:08:50 2015(r277882)
+++ head/sys/sys/cdefs.hThu Jan 29 18:44:21 2015(r277883)
@@ -252,7 +252,7 @@
  * Keywords added in C11.
  */
 
-#if !defined(__STDC_VERSION__) || __STDC_VERSION__  201112L
+#if !defined(__STDC_VERSION__) || __STDC_VERSION__  201112L || defined(lint)
 
 #if !__has_extension(c_alignas)
 #if (defined(__cplusplus)  __cplusplus = 201103L) || \
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org