Re: x86/mce merge, integration hickup + crash, design thoughts

[Tim Hockin <thockin@gmail.com>]

On Wed, Jan 14, 2009 at 10:05 AM, Andi Kleen <ak@linux.intel.com> wrote:
>
>>
>> From my point of view: a single, consistent, easy logging interface
>> for the kernel to send *structured data* about hardware/system events
>> and errors up to userspace.
>
> Which kinds of events were you thinking of?
>
> So far we managed by cramming some other CPU events like thermal
> trip into "pseudo banks" in struct mce. Admittedly it's not the
> most pretty solution in the world, but it worked.

Yeah, no offense, but that's horrible :)

Ideally, I'd rather see a more generic conduit for all sorts of
events.  Polled and exception MCEs.  Thermal interrupts.  MCE
threshold interrupts.  EDAC polled errors.  PCI-express errors.  SATA
disk timeouts.

Now I know there are different conduits for some events - netlink
tells me about netif link up/down events I think.  I would settle for
a small number of interfaces.  What I don't want is what we have today
- EVERYTHING has a different interface.  Some are poll()-able.  Some
have to be actively polled.  Some have to have a daemon listening or
else messages are dropped.  Some have to parse logs.  Puke.

Put it this way:  Given a thousand machines, I want to gather,
collate, and correlate all these events.  I want to be able to produce
a "life story" of sorts for a machine and for a data center.  Once I
can do that, I can start to make predictive diagnoses more accurately,
and I can know how much these things actually COST us.

Tim