Re: [PATCH 4/7] acpi, apei, ghes: Factor out NMI error notification context.

[Tomasz Nowicki <tomasz.nowicki@linaro.org>]

On 13.05.2014 21:41, Borislav Petkov wrote:
> On Wed, Apr 09, 2014 at 05:14:32PM +0200, Tomasz Nowicki wrote:
>> Use CONFIG_ACPI_APEI_NMI to isolate NMI error notification path. NMI related
>> data and functions are grouped so they can be wrapped inside one
I have missed end of sentence. I should goes like that:

Use CONFIG_ACPI_APEI_NMI to isolate NMI error notification path. NMI 
related data and functions are grouped so they can be wrapped inside one
#ifdefs CONFIG_ACPI_APEI_NMI.

>>
>> Signed-off-by: Tomasz Nowicki <tomasz.nowicki@linaro.org>
>> ---
>>   drivers/acpi/apei/ghes.c |   54 +++++++++++++++++++++++++---------------------
>>   1 file changed, 30 insertions(+), 24 deletions(-)
>>
>> diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c
>> index ca8387e..7a0d66e 100644
>> --- a/drivers/acpi/apei/ghes.c
>> +++ b/drivers/acpi/apei/ghes.c
>> @@ -53,7 +53,9 @@
>>   #include <asm/mce.h>
>>   #endif
>>   #include <asm/tlbflush.h>
>> +#ifdef CONFIG_ACPI_APEI_NMI
>>   #include <asm/nmi.h>
>> +#endif
>
> This, again, can be avoided with adding arch-specific versions and empty
> default stubs.
>

I had that thoughts too. Looking at simple MCE calls, yes, it does make 
sense to create corresponding arch-specific version and provide logic as 
needed. I think that NMI is much more complicated. NMI context is 
tightly coupled with the rest of GHES mechanisms like pool memory, list 
locks etc. which are GHES locally available. I am not saying it is 
impossible, I am just concerned if it makes sense to create e.g. 
apei-nmi.c arch-specific file and export necessary GHES mechanisms just 
for NMI purpose. Please let me know your opinion on this.

Regards,
Tomasz