[OE-core] [PATCH] tune-i586-nlp: Add new tune file to support Quark/X1000 CPU

Darren Hart dvhart at linux.intel.com
Tue Jun 23 17:17:19 UTC 2015


On 6/23/15, 9:02 AM, "Richard Purdie" <richard.purdie at linuxfoundation.org>
wrote:

>On Tue, 2015-06-23 at 08:22 -0700, Saul Wold wrote:
>> This tune file is needed to enable a GAS option specific to this cpu
>>family
>> in order to disable the usage of lock prefix instructions.
>> 
>> Signed-off-by: Saul Wold <sgw at linux.intel.com>
>> ---
>>  meta/conf/machine/include/tune-i586-nlp.inc | 20 ++++++++++++++++++++
>>  1 file changed, 20 insertions(+)
>>  create mode 100644 meta/conf/machine/include/tune-i586-nlp.inc
>> 
>> diff --git a/meta/conf/machine/include/tune-i586-nlp.inc
>>b/meta/conf/machine/include/tune-i586-nlp.inc
>> new file mode 100644
>> index 0000000..6815a16
>> --- /dev/null
>> +++ b/meta/conf/machine/include/tune-i586-nlp.inc
>> @@ -0,0 +1,20 @@
>> +#
>> +# Settings for the GAS(1) cpu-type quark/x1000 which has issues
>> +# with lock-prefix so use that option to omit them.
>> +#
>> +
>> +DEFAULTTUNE ?= "i586-nlp"
>> +
>> +# Include the previous tune to pull in PACKAGE_EXTRA_ARCHS
>> +require conf/machine/include/tune-i586.inc
>> +
>> +# x86 with no lock prefix
>> +TUNEVALID[nlp] = "IA32 with Lock Prefix omitted"
>> +TUNE_CCARGS .= "${@bb.utils.contains("TUNE_FEATURES", "nlp", "
>>-Wa,-momit-lock-prefix=yes", "", d)}"
>> +
>> +# Quark tune feature
>> +AVAILTUNES = "i586-nlp"
>> +TUNE_FEATURES_tune-i586-nlp = "${TUNE_FEATURES_tune-i586} nlp"
>> +TUNE_PKGARCH_tune-i586-nlp = "i586-nlp"
>> +PACKAGE_EXTRA_ARCHS_tune-i586-nlp = "${PACKAGE_EXTRA_ARCHS_tune-i586}
>>i586-nlp"
>
>Since i586 doesn't work with this hardware, should this inherit the i586
>tune?

Hrm... The thing we need to avoid is quark machines accepting an i586
package in lieu of a i586-nlp package. I don't think that's a problem here
is it? The PACKAGE_EXTRA_ARCHS makes it so the i586-nlp package can be
used on machines that require i586, but not vice versa - correct?

I don't see a conflict here, but maybe I'm missing something.

-- 
Darren Hart
Intel Open Source Technology Center






More information about the Openembedded-core mailing list