Re: [PATCH v2 2/2] x86/cpu: Change query logic so cpuid is enabled before testing

From: Andy Lutomirski
Date: Fri Sep 21 2018 - 17:54:59 EST


On Fri, Sep 21, 2018 at 2:21 PM Matthew Whitehead <tedheadster@xxxxxxxxx> wrote:
>
> Presently we check for cpuid to be enabled first. If it is not already
> enabled, then we next call identify_cpu_without_cpuid() and clear
> X86_FEATURE_CPUID.
>
> Unfortunately, identify_cpu_without_cpuid() is the function where cpuid
> becomes _enabled_ on Cyrix 6x86/6x86L cpus. So we must reverse the
> calling sequence so that cpuid is first enabled, and then check a second
> time to see if the feature has now been activated.
>
> Reviewed-by: Andy Lutomirski <luto@xxxxxxxxxxxxxx>

How about Suggested-and-reviewed-by: Andy Lutomirski <luto@xxxxxxxxxx> instead?

(No need to resubmit for just that. Ingo or Thomas, if you like
these, can you just make that substitution?)

> Signed-off-by: Matthew Whitehead <tedheadster@xxxxxxxxx>
> ---
> arch/x86/kernel/cpu/common.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
> index eb4cb3efd20e..60c7c5ce7e55 100644
> --- a/arch/x86/kernel/cpu/common.c
> +++ b/arch/x86/kernel/cpu/common.c
> @@ -1040,6 +1040,9 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c)
> memset(&c->x86_capability, 0, sizeof c->x86_capability);
> c->extended_cpuid_level = 0;
>
> + if (!have_cpuid_p())
> + identify_cpu_without_cpuid(c);
> +
> /* cyrix could have cpuid enabled via c_identify()*/
> if (have_cpuid_p()) {
> cpu_detect(c);
> @@ -1057,7 +1060,6 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c)
> if (this_cpu->c_bsp_init)
> this_cpu->c_bsp_init(c);
> } else {
> - identify_cpu_without_cpuid(c);
> setup_clear_cpu_cap(X86_FEATURE_CPUID);
> }
>
> --
> 2.16.4
>


--
Andy Lutomirski
AMA Capital Management, LLC