Re: [PATCH v2] x86/platform/uv: Rework NMI "action" modparam handling

[Hans de Goede]

Hi,

On 9/14/23 10:42, Andy Shevchenko wrote:
> On Wed, Sep 13, 2023 at 9:01 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
>>
>> Rework NMI "action" modparam handling:
>>
>> 1. Replace the uv_nmi_action string with an enum; and
>> 2. Use sysfs_match_string() for string parsing in param_set_action()
> 
> ...
> 
> Don't you need to include string.h?

Yes (pre-existing problem really, but still). Will fix for v3.

> 
> ...
> 
>> +enum action_t {
>> +       nmi_act_kdump,
>> +       nmi_act_dump,
>> +       nmi_act_ips,
>> +       nmi_act_kdb,
>> +       nmi_act_kgdb,
>> +       nmi_act_health,
>>  };
> 
>> +
>> +static_assert(ARRAY_SIZE(actions) == ARRAY_SIZE(actions_desc));
> 
> I believe with enum in place you don't need this, just add a terminator to it
> 
> enum action_t {
>    ...
>    nmi_act_max
> };
> 
> and use as an array size both arrays.
> 
> ...
> 
>> -       int i;
>> -       int n = ARRAY_SIZE(valid_acts);
> 
>> +       int i, n = ARRAY_SIZE(actions);
> 
> Since you are changing them, why not make them unsigned?

i is used to store the result of sysfs_match_string() which
returns a negative errno.

> 
> ...
> 
>> -       pr_err("UV: Invalid NMI action:%s, valid actions are:\n", arg);
>> +       pr_err("UV: Invalid NMI action:%s, valid actions are:\n", val);
> 
> As mentioned previously the val may or may not have a new line in it.
> I dunno about comma removal, but a new line presence can be easily checked.
> 
> Either way it's not so critical, hence removing comma. or replacing it
> by '-' (dash) may be enough.

I'll fix this for v3 by not printing the wrong val at all as suggested
by Steve Wahl.

Regards,

Hans