Bug#755739: VGA-mode newcons very slow
Steven Chamberlain <steven@pyro.eu.org> writes:
> On 11/09/14 00:27, Christoph Egger wrote:
>> MAKEFLAGS=-j4 make MACHINE_ARCH=amd64 -DWITHOUT_SOURCELESS
>> -DWITHOUT_CLANG_IS_CC WERROR= -DWITHOUT_CTF -DWITHOUT_KERNEL_SYMBOLS
>> -DWITHOUT_SSP CC=clang-3.4 -C
>> flavor-10.1-0-amd64/sys/modules/aic7xxx/ firmware
>> ( cd aicasm ; make aicasm )
>> ( cd ahd ; make ahdfirmware )
>> ( cd ahc ; make ahcfirmware )
>> ../aicasm/aicasm ...
>
> That's due to -j4, right? It would have built aicasm first otherwise.
$ echo $DEB_BUILD_OPTIONS
parallel=4
so jep
>> I think this should be fixed by
>>
>> Index: sys/modules/aic7xxx/Makefile
>> ===================================================================
>> --- sys/modules/aic7xxx/Makefile (revision 271306)
>> +++ sys/modules/aic7xxx/Makefile (working copy)
>> @@ -8,10 +8,10 @@
>> aicasm/aicasm:
>> ( cd ${.CURDIR}/aicasm ; ${MAKE} aicasm )
>>
>> -ahdfirmware:
>> +ahdfirmware: aicasm/aicasm
>> ( cd ${.CURDIR}/ahd ; ${MAKE} ahdfirmware )
>>
>> -ahcfirmware:
>> +ahcfirmware: aicasm/aicasm
>> ( cd ${.CURDIR}/ahc ; ${MAKE} ahcfirmware )
>>
>> cleanfirmware:
>
> Looks right, yes. If you're able to finish building with -j4, please
> continue with the upload and commit that fix.
Should be uploaded let's see when the archive reacts. Do we have a way
to forward that patch upstream? seems to be the right thing there as
well.
Christoph
Reply to: