diff mbox

SPU build broken (Re: CFT: [build] Move libgcc2 to toplevel libgcc)

Message ID yddaa82pr0v.fsf@manam.CeBiTec.Uni-Bielefeld.DE
State New
Headers show

Commit Message

Rainer Orth Nov. 11, 2011, 4:23 p.m. UTC 
"Ulrich Weigand" <uweigand@de.ibm.com> writes:

> Rainer Orth wrote:
>
>> diff --git a/gcc/config/spu/t-spu-elf b/gcc/config/spu/t-spu-elf
>
>> -# We exclude those because the libgcc2.c default versions do not support
>> -# the SPU single-precision format (round towards zero).  We provide our
>> -# own versions below and/or via direct expansion.
>> -LIB2FUNCS_EXCLUDE = _floatdisf _floatundisf _floattisf _floatunstisf
>
>> diff --git a/libgcc/config/spu/t-elf b/libgcc/config/spu/t-elf
>   
>> +# We exclude those because the libgcc2.c default versions do not support
>> +# the SPU single-precision format (round towards zero).  We provide our
>> +# own versions below and/or via direct expansion.
>> +LIB2ADD = _floatdisf _floatundisf _floattisf _floatunstisf
>
>
> This seems to have caused:
>
> make[2]: Entering directory `/home/kwerner/dailybuild/spu-tc-2011-11-05/gcc-build/spu/libgcc'
> Makefile:792: *** Unsupported files in LIB2ADD or LIB2ADD_ST..  Stop.

Sorry for the delay.  Indeed, the test at that line only accepts .c, .S,
and .asm files, where the latter should probably be removed now that
we've standardized on .S.

> Shouldn't the variable still be called LIB2FUNCS_EXCLUDE after the
> move to libgcc?  LIB2ADD seems to expect full file names ...

Of course, the change is bogus.  I can only (half) explain this by the
change from LIB2FUNCS_STATIC_EXTRA to LIB2ADD_ST extra.

The trivial patch allowed a x86_64-unknown-linux-gnu x spu-elf cross to
finish the libgcc build, and at least the set of objects built before my
patch series is identical to the set built now.

Ok for mainline?

	Rainer


2011-11-11  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>

	* config/spu/t-elf (LIB2ADD): Use LIB2FUNCS_EXCLUDE instead.

Comments

Paolo Bonzini Nov. 11, 2011, 4:26 p.m. UTC | #1 
On 11/11/2011 05:23 PM, Rainer Orth wrote:
> The trivial patch allowed a x86_64-unknown-linux-gnu x spu-elf cross to
> finish the libgcc build, and at least the set of objects built before my
> patch series is identical to the set built now.
>
> Ok for mainline?

Ok.  Have you checked for other occurrences?

Paolo
Rainer Orth Nov. 11, 2011, 4:41 p.m. UTC | #2 
Paolo Bonzini <bonzini@gnu.org> writes:

> On 11/11/2011 05:23 PM, Rainer Orth wrote:
>> The trivial patch allowed a x86_64-unknown-linux-gnu x spu-elf cross to
>> finish the libgcc build, and at least the set of objects built before my
>> patch series is identical to the set built now.
>>
>> Ok for mainline?
>
> Ok.  Have you checked for other occurrences?

Yup.  LIB2FUNCS_EXCLUDE only lists functions names, and LIB2ADD* only
lists filenames with supported extensions, all of which reference
$(srcdir) with the exception of cris/t-cris and m68k/t-floatlib, where
the sources are generated at build time.

This has been a visual inspection since grep cannot easily deal with the
continuation lines.

	Rainer
diff mbox

Patch

# HG changeset patch
# Parent 4b61b438da8a6a11ab1e06abe67fd26fa715c25e
Fix SPU libgcc build

diff --git a/libgcc/config/spu/t-elf b/libgcc/config/spu/t-elf
--- a/libgcc/config/spu/t-elf
+++ b/libgcc/config/spu/t-elf
@@ -5,7 +5,7 @@  CRTSTUFF_T_CFLAGS =
 # We exclude those because the libgcc2.c default versions do not support
 # the SPU single-precision format (round towards zero).  We provide our
 # own versions below and/or via direct expansion.
-LIB2ADD = _floatdisf _floatundisf _floattisf _floatunstisf
+LIB2FUNCS_EXCLUDE = _floatdisf _floatundisf _floattisf _floatunstisf
 
 LIB2ADD_ST = $(srcdir)/config/spu/float_unssidf.c \
 	     $(srcdir)/config/spu/float_unsdidf.c \