On 5/3/2016 12:35 AM, Ingo Molnar wrote:
>
> * Mike Travis wrote:
>
>> +config X86_UV1_SUPPORTED
>> +bool "SGI Ultraviolet Series 1 Supported"
>> +depends on X86_UV
>
> So I still think it's much simpler if we simply eliminate this Kconfig
> complication and have it
On 5/3/2016 12:35 AM, Ingo Molnar wrote:
>
> * Mike Travis wrote:
>
>> +config X86_UV1_SUPPORTED
>> +bool "SGI Ultraviolet Series 1 Supported"
>> +depends on X86_UV
>
> So I still think it's much simpler if we simply eliminate this Kconfig
> complication and have it all compatible.
* Mike Travis wrote:
> +config X86_UV1_SUPPORTED
> + bool "SGI Ultraviolet Series 1 Supported"
> + depends on X86_UV
So I still think it's much simpler if we simply eliminate this Kconfig
complication and have it all compatible. AFAICS the runtime impact on newer
* Mike Travis wrote:
> +config X86_UV1_SUPPORTED
> + bool "SGI Ultraviolet Series 1 Supported"
> + depends on X86_UV
So I still think it's much simpler if we simply eliminate this Kconfig
complication and have it all compatible. AFAICS the runtime impact on newer
systems comes down
This patch deletes old code that artificially added and removed APIC
ID bits not available on UV100/UV1000 systems only. An additional
performance hit and code bloat occurs when selecting MMR address or
field defines as the following code example shows. (This is produced
by one of the Verilog to
This patch deletes old code that artificially added and removed APIC
ID bits not available on UV100/UV1000 systems only. An additional
performance hit and code bloat occurs when selecting MMR address or
field defines as the following code example shows. (This is produced
by one of the Verilog to
6 matches
Mail list logo