I386.md: *_mixed and *_sse
Hi, From i386.md, alternative 1 of *fop_sf_comm_mixed is duplicated with *fop_sf_comm_sse. Why do we define a _mixed pattern here? (define_insn "*fop_sf_comm_mixed" [(set (match_operand:SF 0 "register_operand" "=f,x") (match_operator:SF 3 "binary_fp_operator" [(match_operand:SF 1 "nonimmediate_operand" "%0,0") (match_operand:SF 2 "nonimmediate_operand" "fm,xm")]))] "TARGET_MIX_SSE_I387 && COMMUTATIVE_ARITH_P (operands[3]) && !(MEM_P (operands[1]) && MEM_P (operands[2]))" "* return output_387_binary_op (insn, operands);" [(set (attr "type") (if_then_else (eq_attr "alternative" "1") (if_then_else (match_operand:SF 3 "mult_operator" "") (const_string "ssemul") (const_string "sseadd")) (if_then_else (match_operand:SF 3 "mult_operator" "") (const_string "fmul") (const_string "fop" (set_attr "mode" "SF")]) (define_insn "*fop_sf_comm_sse" [(set (match_operand:SF 0 "register_operand" "=x") (match_operator:SF 3 "binary_fp_operator" [(match_operand:SF 1 "nonimmediate_operand" "%0") (match_operand:SF 2 "nonimmediate_operand" "xm")]))] "TARGET_SSE_MATH && COMMUTATIVE_ARITH_P (operands[3]) && !(MEM_P (operands[1]) && MEM_P (operands[2]))" "* return output_387_binary_op (insn, operands);" [(set (attr "type") (if_then_else (match_operand:SF 3 "mult_operator" "") (const_string "ssemul") (const_string "sseadd"))) (set_attr "mode" "SF")]) Thanks - Joey
Re: How to teach gcc, that registers are clobbered by api calls?
Kai Tietz writes: >I read that too, but how can I teach gcc to do this that registers are >callee-saved? I tried it by use of call_used part in regclass.c, but >this didn't worked as expected. I think you need to modify CALL_USED_REGISTERS and/or CONDITIONAL_REGISTER_USAGE in i386.h. Making any changes to regclass.c is probably not the right thing to do. Ross Ridge
Re: no mul/div instruction
Hi, I wanted support for software floating point on the architecture. I am using fp-bit.c & dp-bit.c and have defined FLOAT_TYPE_SIZE as 32 and DOUBLE_TYPE_SIZE as 64. dp-bit.c requires __muldi3. How do I enable emulation of 64 bit multiply in libgcc2.a ? Thanks in advance, Kunal Parmar
Re: no mul/div instruction
Hi Ian, On Tue, Apr 22, 2008 at 7:12 PM, Ian Lance Taylor <[EMAIL PROTECTED]> wrote: > Looking at libgcc2.h, it seems like you might need to define > LIBGCC2_UNITS_PER_WORD in your tm.h file. That solved my problem. Thanks a ton ! Kunal
Re: no mul/div instruction
"Kunal Parmar" <[EMAIL PROTECTED]> writes: >> Yes, I think __mulsi3 will be built for you automatically. > > I gave a definition of __mulhi3 for my architecture. But I don't get > __mulsi3 in libgcc.a. Do I have to enable some options for this ? Looking at libgcc2.h, it seems like you might need to define LIBGCC2_UNITS_PER_WORD in your tm.h file. Ian
Re: US-CERT Vulnerability Note VU#162289
On Tue, 22 Apr 2008, Mark Mitchell wrote: > Chad Dougherty wrote: > > > The vulnerability note has been significantly reworked to focus on the issue > > of undefined behavior handling in the compiler and the fact that conforming > > implementations are not required to warn of this condition. I've tried to > > incorporate many of the valid concerns that were raise on this list in > > response to the original vulnerability note. > > Thank you for making the update; this is a big improvement. > > However, I'm surprised that only GCC is listed as "vulnerable" at the bottom > of the page. We've provided information about a lot of other compilers that > do the same optimization. Why is the status for compilers from Microsoft, > Intel, IBM, etc. listed as "Unknown" instead of "Vulnerable"? > > -- > Mark Mitchell > CodeSourcery > [EMAIL PROTECTED] > (650) 331-3385 x713 Additionally, the linked to notes for GCC are reflective of the original innaccuracies: http://www.kb.cert.org/vuls/id/CRDY-7DWKWM Vendor Statement No statement is currently available from the vendor regarding this vulnerability. US-CERT Addendum Vendors and developers using the GNU C compiler should consider downgrading their version of gcc or sticking with versions of the gcc compiler (before version 4.1) that do not perform the offending optimization. In the case of gcc, it should be emphasized that this is a change of behavior in the later versions of the compiler. Later, Brad
Re: no mul/div instruction
Hi Ian, > Yes, I think __mulsi3 will be built for you automatically. I gave a definition of __mulhi3 for my architecture. But I don't get __mulsi3 in libgcc.a. Do I have to enable some options for this ? Thanks in advance, Kunal Parmar
Re: no mul/div instruction
"Kunal Parmar" <[EMAIL PROTECTED]> writes: > On Tue, Apr 22, 2008 at 1:24 PM, Ian Lance Taylor <[EMAIL PROTECTED]> wrote: >> It depends on UNITS_PER_WORD. If UNITS_PER_WORD is 4, you need >> __mulsi3. If UNITS_PER_WORD is 2, you need __mulhi3, and, if you have >> 32-bit integer types, you will also need __mulsi3. In the latter case >> you can, if you like, get code for __mulsi3 from longlong.h--see the >> comments at the top of the file. > > UNITS_PER_WORD is 2 and INT_TYPE_SIZE is 16. This means I need to > provide __mulhi3. Does this mean that libgcc will provide the definition of > __mulsi3 ? Yes, I think __mulsi3 will be built for you automatically. Ian
Re: no mul/div instruction
Hi Ian, On Tue, Apr 22, 2008 at 1:24 PM, Ian Lance Taylor <[EMAIL PROTECTED]> wrote: > It depends on UNITS_PER_WORD. If UNITS_PER_WORD is 4, you need > __mulsi3. If UNITS_PER_WORD is 2, you need __mulhi3, and, if you have > 32-bit integer types, you will also need __mulsi3. In the latter case > you can, if you like, get code for __mulsi3 from longlong.h--see the > comments at the top of the file. UNITS_PER_WORD is 2 and INT_TYPE_SIZE is 16. This means I need to provide __mulhi3. Does this mean that libgcc will provide the definition of __mulsi3 ? Thanks, Kunal
Re: US-CERT Vulnerability Note VU#162289
Chad Dougherty wrote: The vulnerability note has been significantly reworked to focus on the issue of undefined behavior handling in the compiler and the fact that conforming implementations are not required to warn of this condition. I've tried to incorporate many of the valid concerns that were raise on this list in response to the original vulnerability note. Thank you for making the update; this is a big improvement. However, I'm surprised that only GCC is listed as "vulnerable" at the bottom of the page. We've provided information about a lot of other compilers that do the same optimization. Why is the status for compilers from Microsoft, Intel, IBM, etc. listed as "Unknown" instead of "Vulnerable"? -- Mark Mitchell CodeSourcery [EMAIL PROTECTED] (650) 331-3385 x713
Re: US-CERT Vulnerability Note VU#162289
Joe Buck wrote: Thanks. I hope that you will correct the advisory promptly to avoid any implication that one should switch from GCC to a different compiler based on this issue, since we've already established that most of GCC's competitors perform similar optimizations under some cicumstances (even if the particular example that appears in the CERT report is not affected, other, similar examples will be, particularly if they appear in a loop). Both CERT and GCC have their reputations to consider here, and I think that this advisory has damaged the reputations of *both*. The vulnerability note has been significantly reworked to focus on the issue of undefined behavior handling in the compiler and the fact that conforming implementations are not required to warn of this condition. I've tried to incorporate many of the valid concerns that were raise on this list in response to the original vulnerability note. The advisory should emphasize the solution of auditing buffer overflow checks to make sure that they are correct C, and should help people write such checks correctly. The vulnerability note itself essentially punts this issue to the corresponding documents in our Secure Coding standard. -Chad
Re: namelookup bug in gcc?
On Tue, Apr 22, 2008 at 02:28:15PM +0200, Balazs Dezso wrote:
> I have tested the following code on g++ 4.3, 4.2, 4.1 and 3.4.
Looks like a bug. Please file a bug report at http://gcc.gnu.org/bugzilla
> #include
>
> struct B {
> static const int x = 1;
> };
>
> struct A {
> static const int x = 0;
> template
> static void f() {
> std::cerr << A::x << std::endl;
> }
> };
>
> int main() {
> A::f();
> return 0;
> }
>
> The gcc result:
> 0
> While icc, msvc result:
> 1
Re: no mul/div instruction
Kunal Parmar <[EMAIL PROTECTED]> writes: > I am porting GCC to a new 16 bit RISC architecture which does not have > multiplication and division instructions. I figured that I have to provide > emulation routines for the multiplication and division which will be > inserted into libgcc2.a. But I am confused about which versions of these > routines to provide i.e. do I provide __mulsi3 or __mulhi3 or both. It depends on UNITS_PER_WORD. If UNITS_PER_WORD is 4, you need __mulsi3. If UNITS_PER_WORD is 2, you need __mulhi3, and, if you have 32-bit integer types, you will also need __mulsi3. In the latter case you can, if you like, get code for __mulsi3 from longlong.h--see the comments at the top of the file. Ian
no mul/div instruction
Hi all, I am porting GCC to a new 16 bit RISC architecture which does not have multiplication and division instructions. I figured that I have to provide emulation routines for the multiplication and division which will be inserted into libgcc2.a. But I am confused about which versions of these routines to provide i.e. do I provide __mulsi3 or __mulhi3 or both. Thanks in advance, Kunal
Re: Google Summer of Code 2008: seven approved applications for gcc
And in our wiki: http://gcc.gnu.org/wiki/SummerOfCode which our new developers are welcome to use to document their progress. Cheers, Manuel. On 22/04/2008, Ian Lance Taylor <[EMAIL PROTECTED]> wrote: > I'm happy to report that Google approved seven applications for Summer > of Code for the gcc project. The approved applications can be found > here: http://code.google.com/soc/2008/gcc/about.html . > > > Ian >
Google Summer of Code 2008: seven approved applications for gcc
I'm happy to report that Google approved seven applications for Summer of Code for the gcc project. The approved applications can be found here: http://code.google.com/soc/2008/gcc/about.html . Ian
namelookup bug in gcc?
Hello all,
I have tested the following code on g++ 4.3, 4.2, 4.1 and 3.4.
#include
struct B {
static const int x = 1;
};
struct A {
static const int x = 0;
template
static void f() {
std::cerr << A::x << std::endl;
}
};
int main() {
A::f();
return 0;
}
The gcc result:
0
While icc, msvc result:
1
Best, Balazs
---
Re: How to insert functions?
Hi Andrew, Yes, I did a bit. it segfaults at cgraphunit.c:cgraph_expand_all_functions:1323 node->lowered = DECL_STRUCT_FUNCTION (node->decl)->cfg != NULL; It seems that the node->decl has been nulled by the time it gets here. It definitely isn't NULL after leaving my code. Can you think of anything that might do that? I figured I must be doing something pretty wrong and that there must be tons of examples for this kind of thing but I haven't been able to find one. Cheers for the help, Hugh. Andrew Haley wrote: Hugh Leather wrote: Hi, I am trying to add a new destructor function to object files I compile. I'm doing this to instrument programs and then, once the program has finished I want to print out the statistics I've gathered. So, just before pass 'remove_useless_stmts' is called on each function I try to create a static destructor which will print stats for the current function (and obviously, I don't do it on the functions just created). Below I've put a simpler form, which should work for a one function program. Calling createFnDecl() creates a new function called __my_new_function. It should print "It worked!" when the program exits. Now, with compiler flag -O0, this works just fine, but with -O1 or above it seg faults the compiler. Can anyone tell me what I'm doing wrong? Nothing obvious to me. Did you debug the point at which the segfault happened? Andrew.
Re: How to insert functions?
Hugh Leather wrote: > Hi, > > I am trying to add a new destructor function to object files I > compile. I'm doing this to instrument programs and then, once the > program has finished I want to print out the statistics I've gathered. > So, just before pass 'remove_useless_stmts' is called on each function > I try to create a static destructor which will print stats for the > current function (and obviously, I don't do it on the functions just > created). > > Below I've put a simpler form, which should work for a one function > program. Calling createFnDecl() creates a new function called > __my_new_function. It should print "It worked!" when the program exits. > Now, with compiler flag -O0, this works just fine, but with -O1 or > above it seg faults the compiler. > > Can anyone tell me what I'm doing wrong? Nothing obvious to me. Did you debug the point at which the segfault happened? Andrew.
How to insert functions?
Hi,
I am trying to add a new destructor function to object files I
compile. I'm doing this to instrument programs and then, once the
program has finished I want to print out the statistics I've gathered.
So, just before pass 'remove_useless_stmts' is called on each
function I try to create a static destructor which will print stats for
the current function (and obviously, I don't do it on the functions just
created).
Below I've put a simpler form, which should work for a one function
program. Calling createFnDecl() creates a new function called
__my_new_function. It should print "It worked!" when the program exits.
Now, with compiler flag -O0, this works just fine, but with -O1 or
above it seg faults the compiler.
Can anyone tell me what I'm doing wrong?
Cheers,
Hugh.
static tree createFnBody() {
tree arg;
tree argList;
tree putsFn;
tree call;
tree bind;
char msg[] = "It worked!";
bind = build3( BIND_EXPR, void_type_node, NULL_TREE, NULL_TREE,
NULL_TREE );
TREE_SIDE_EFFECTS( bind ) = 1;
putsFn = implicit_built_in_decls[ BUILT_IN_PUTS ];
arg = build_string_literal( sizeof( msg ) + 1, msg );
argList = tree_cons( NULL_TREE, arg, NULL_TREE );
call = build_function_call_expr( putsFn, argList );
append_to_statement_list( call, &BIND_EXPR_BODY( bind ));
return bind;
}
static tree createFnDecl() {
tree callTypes;
tree fnName;
tree fnDecl;
tree t;
struct function* oldCfun = cfun;
fnName = get_identifier( "__my_new_function" );
callTypes = build_function_type_list( void_type_node, NULL_TREE );
fnDecl = build_decl( FUNCTION_DECL, fnName, callTypes );
fnDecl = lang_hooks.decls.pushdecl( fnDecl );
TREE_STATIC( fnDecl ) = 1;
TREE_USED( fnDecl ) = 1;
DECL_ARTIFICIAL( fnDecl ) = 1;
DECL_IGNORED_P( fnDecl ) = 0;
TREE_PUBLIC( fnDecl ) = 0;
DECL_UNINLINABLE( fnDecl ) = 1;
DECL_EXTERNAL( fnDecl ) = 0;
DECL_CONTEXT( fnDecl ) = NULL_TREE;
DECL_INITIAL( fnDecl ) = make_node ( BLOCK );
DECL_STATIC_DESTRUCTOR( fnDecl ) = 1;
DECL_NO_INSTRUMENT_FUNCTION_ENTRY_EXIT( fnDecl ) = 1;
t = build_decl( RESULT_DECL, NULL_TREE, void_type_node );
DECL_ARTIFICIAL( t ) = 1;
DECL_IGNORED_P( t ) = 1;
DECL_RESULT( fnDecl ) = t;
DECL_SAVED_TREE( fnDecl ) = createFnBody();
allocate_struct_function( fnDecl );
cgraph_add_new_function( fnDecl );
cfun = oldCfun;
return fnDecl;
}
Re: ada: What is libgnalasup?
> "Arnaud" == Arnaud Charlet <[EMAIL PROTECTED]> writes: Arnaud> Yes, libgnalasup is a bundled version of lapack/blas Arnaud> libraries. Thanks. I've submitted a patch to make lapack/blas the default version. Sam -- Samuel Tardieu -- [EMAIL PROTECTED] -- http://www.rfc1149.net/
