On Sun, Jan 2, 2011 at 3:03 PM, H.J. Lu <[email protected]> wrote:
> On Sun, Jan 2, 2011 at 2:05 PM, H.J. Lu <[email protected]> wrote:
>> On Sun, Jan 2, 2011 at 1:18 PM, Ian Lance Taylor <[email protected]> wrote:
>>> Richard Guenther <[email protected]> writes:
>>>
>>>> On Sun, Jan 2, 2011 at 9:24 PM, Ian Lance Taylor <[email protected]> wrote:
>>>>> Richard Guenther <[email protected]> writes:
>>>>>
>>>>>> Your small patch removing have_o || is ok I guess.
>>>>>
>>>>> Wait. That will change the behaviour of
>>>>> gcc -o foo.o -c f1.c f2.c f3.c
>>>>> Is that what we want?
>>>>
>>>> Does it? I don't think so. Most of the combine handling was removed by
>>>> the patch that caused the regression, so -o and -c doesn't combine anymore
>>>> (with multiple sources).
>>>
>>> Sorry, you're right. The difference is that @c has 0 for the combinable
>>> field, and @assembler has 1. Before H.J.'s change, this worked
>>> gcc -c -o f.o f1.s f2.s
>>> After his change, it does not. That is probably not a big deal.
>>>
>>> I wonder why @assembler has 1 for combinable? It seems to have been set
>>> to 1 when the combinable field was added in 2004-04-05 with -combine.
>>> Now that -combine has been removed, if the combinable field for
>>> @assembler were 0, it seems to me that H.J.'s problem would also be
>>> fixed. And it seems to me that it should be 0.
>>>
>>>
>>>>> Also, right now the gccgo driver depends on the -o behaviour to combine
>>>>> inputs. If that changes, the driver will need to provide some other way
>>>>> to let the frontend force inputs to be combined.
>>>>
>>>> For go it isn't equivalent to do gcgo -c t1.go; gcgo -c t2.go; gcgo t1.o
>>>> t2.o
>>>> compared to gcgo t1.go t2.go?
>>>
>>> No, it is not. All .go input files must be passed to go1 at once.
>>> H.J.'s patch has indeed broken gccgo.
>>>
>>
>> Can you try this patch?
>>
>> Thanks.
>>
>>
>> --
>> H.J.
>> ---
>> diff --git a/gcc/gcc.c b/gcc/gcc.c
>> index 0d633a4..d0b2c96 100644
>> --- a/gcc/gcc.c
>> +++ b/gcc/gcc.c
>> @@ -6582,7 +6582,20 @@ warranty; not even for MERCHANTABILITY or FITNESS FOR
>> A P
>> ARTICULAR PURPOSE.\n\n"
>>
>> explicit_link_files = XCNEWVEC (char, n_infiles);
>>
>> + /* Check if we should combine inputs. */
>> combine_inputs = flag_wpa;
>> + if (!combine_inputs)
>> + for (i = 1; i < decoded_options_count; i++)
>> + {
>> + if (decoded_options[i].opt_index == OPT_x)
>> + {
>> + struct compiler *compiler
>> + = lookup_compiler (NULL, 0, decoded_options[i].arg);
>> + if (compiler)
>> + combine_inputs = compiler->combinable;
>> + break;
>> + }
>> + }
>>
>> for (i = 0; (int) i < n_infiles; i++)
>> {
>>
>
> This doesn't work for go since -xgo isn't used with gccgo. Is there
> a way to tell what the default language is for a gcc driver?
>
I am testing this patch with Go and LTO.
--
H.J.
---
diff --git a/gcc/gcc.c b/gcc/gcc.c
index 0d633a4..6e26b1d 100644
--- a/gcc/gcc.c
+++ b/gcc/gcc.c
@@ -2817,6 +2817,10 @@ static int n_infiles_alloc;
static bool combine_inputs;
+/* Default language. */
+
+static struct cl_decoded_option *default_language;
+
/* This counts the number of libraries added by lang_specific_driver, so that
we can tell if there were any user supplied any files or libraries. */
@@ -3714,6 +3718,9 @@ process_command (unsigned int decoded_options_count,
for (j = 1; j < decoded_options_count; j++)
{
+ if (decoded_options[j].opt_index == OPT_x)
+ default_language = &decoded_options[j];
+
if (decoded_options[j].opt_index == OPT_SPECIAL_input_file)
{
const char *arg = decoded_options[j].arg;
@@ -6582,7 +6589,15 @@ warranty; not even for MERCHANTABILITY or FITNESS FOR A P
ARTICULAR PURPOSE.\n\n"
explicit_link_files = XCNEWVEC (char, n_infiles);
+ /* Check if we should combine inputs. */
combine_inputs = flag_wpa;
+ if (!combine_inputs && default_language)
+ {
+ struct compiler *compiler
+ = lookup_compiler (NULL, 0, default_language->arg);
+ if (compiler)
+ combine_inputs = compiler->combinable;
+ }
for (i = 0; (int) i < n_infiles; i++)
{
diff --git a/gcc/go/gospec.c b/gcc/go/gospec.c
index 7d21ace..926dec1 100644
--- a/gcc/go/gospec.c
+++ b/gcc/go/gospec.c
@@ -214,7 +214,7 @@ lang_specific_driver (struct cl_decoded_option **in_decoded_
options,
#endif
/* Make sure to have room for the trailing NULL argument. */
- num_args = argc + need_math + shared_libgcc + (library > 0) * 5 + 5;
+ num_args = argc + need_math + shared_libgcc + (library > 0) * 5 + 6;
new_decoded_options = XNEWVEC (struct cl_decoded_option, num_args);
i = 0;
@@ -223,6 +223,10 @@ lang_specific_driver (struct cl_decoded_option **in_decoded
_options,
/* Copy the 0th argument, i.e., the name of the program itself. */
new_decoded_options[j++] = decoded_options[i++];
+ /* Add -xgo. */
+ generate_option (OPT_x, "go", 1, CL_DRIVER, &new_decoded_options[j]);
+ j++;
+
/* If we are linking, pass -fsplit-stack if it is supported. */
#ifdef TARGET_CAN_SPLIT_STACK
if (library >= 0)
