Re: Re: [PATCH v2 3/3] efi: Capsule update with user helper interface

[Andy Lutomirski <luto@amacapital.net>]

On Wed, Feb 25, 2015 at 3:47 AM, Borislav Petkov <bp@alien8.de> wrote:
> On Tue, Feb 24, 2015 at 12:49:09PM +0000, Kweh, Hock Leong wrote:
>> So the process steps basically look like this:
>> 1.) cat capsule_ticket    =======> acquire a number and lock mutex then
>>                                                                      expose firmware_class user helper
>>                                                                      interface as well as start timer for timeout
>>                                                                      counting
>> 2.) repeat step 1 if obtained a "0" number
>> 3.) echo 1 > loading
>> 4.) cat bin > data
>> 5.) echo 0 > loading        =======> stop the timeout counting  then unlock
>>                                                                       mutex at the end of callback routine
>> 6.) cat capsule_report   =======> grep the number acquired from beginning
>>                                                                       for checking succeeded/failed
>
> So this sounds pretty overengineered for no reason, or maybe I'm missing
> the reason.
>
> If I had to give an example from the microcode loader, what we do there
> is put the microcode in /lib/firmware/... and do
>
> echo 1 > /sys/devices/system/cpu/microcode/reload
>
> which goes and calls reload_store() in
> arch/x86/kernel/cpu/microcode/core.c which grabs a mutex, disables CPU
> hotplug, etc, etc...
>
> And this mechanism is as simple as it can get. Maybe capsules can be
> loaded like that too?
>
> Error code can be propagated too, if needed, of course.

How can the error code be propagated?  Would that echo command fail in
case of error?

--Andy