Re: [PATCH v3 1/7] sysfs/cpu: Allow individual architectures to select vulnerabilities

[Suzuki K Poulose]

On 09/01/2019 23:55, Jeremy Linton wrote:
> As suggested on the list, https://lkml.org/lkml/2019/1/4/282, there are
> a number of cases where its useful for a system to avoid exporting a
> sysfs entry for a given vulnerability. This set adds an architecture
> specific callback which returns the bitmap of vulnerabilities the
> architecture would like to advertise.
>
> Signed-off-by: Jeremy Linton <jeremy.linton@xxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: Rafael J. Wysocki <rafael@xxxxxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
> Cc: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Waiman Long <longman@xxxxxxxxxx>
> Cc: Andi Kleen <ak@xxxxxxxxxxxxxxx>
> Cc: Jiri Kosina <jkosina@xxxxxxx>
> ---
>   drivers/base/cpu.c  | 19 +++++++++++++++++++
>   include/linux/cpu.h |  7 +++++++
>   2 files changed, 26 insertions(+)
>
> diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c
> index eb9443d5bae1..35f6dfb24cd6 100644
> --- a/drivers/base/cpu.c
> +++ b/drivers/base/cpu.c
> @@ -561,6 +561,11 @@ static struct attribute *cpu_root_vulnerabilities_attrs[] = {
>   	NULL
>   };
>   
> +uint __weak arch_supported_vuln_attr_fields(void)
> +{
> +	return VULN_MELTDOWN|VULN_SPECTREV1|VULN_SPECTREV2|VULN_SSB|VULN_L1TF;
> +}
> +
>   static const struct attribute_group cpu_root_vulnerabilities_group = {
>   	.name  = "vulnerabilities",
>   	.attrs = cpu_root_vulnerabilities_attrs,
> @@ -568,6 +573,20 @@ static const struct attribute_group cpu_root_vulnerabilities_group = {
>   
>   static void __init cpu_register_vulnerabilities(void)
>   {
> +	int fld;
> +	int max_fields = ARRAY_SIZE(cpu_root_vulnerabilities_attrs) - 1;
> +	struct attribute **hd = cpu_root_vulnerabilities_attrs;
> +	uint enabled_fields = arch_supported_vuln_attr_fields();
> +
> +	/* only enable entries requested by the arch code */
> +	for (fld = 0; fld < max_fields; fld++) {
> +		if (enabled_fields & 1 << fld) {
> +			*hd = cpu_root_vulnerabilities_attrs[fld];
> +			hd++;
> +		}
> +	}
> +	*hd = NULL;
> +

nit: Could we use "is_visible" callback in the attribute group to check this
dynamically ?

>   	if (sysfs_create_group(&cpu_subsys.dev_root->kobj,
>   			       &cpu_root_vulnerabilities_group))
>   		pr_err("Unable to register CPU vulnerabilities\n");
> diff --git a/include/linux/cpu.h b/include/linux/cpu.h
> index 218df7f4d3e1..5e45814bcc24 100644
> --- a/include/linux/cpu.h
> +++ b/include/linux/cpu.h
> @@ -189,4 +189,11 @@ static inline void cpu_smt_check_topology_early(void) { }
>   static inline void cpu_smt_check_topology(void) { }
>   #endif
>   
> +/* generic cpu vulnerability attributes */
> +#define VULN_MELTDOWN  0x01
> +#define VULN_SPECTREV1 0x02
> +#define VULN_SPECTREV2 0x04
> +#define VULN_SSB       0x08
> +#define VULN_L1TF      0x10

nit: May use BIT() ?

Cheers
Suzuki