Re: [Linux-nvdimm] [PATCH v2] pmem: Initial version of persistent memory driver

From: Dan Williams
Date: Wed Sep 10 2014 - 19:01:47 EST


On Wed, Sep 10, 2014 at 10:47 AM, Boaz Harrosh <boaz@xxxxxxxxxxxxx> wrote:
> On 09/10/2014 08:03 PM, Dan Williams wrote:
>> Hi Boaz,
>>
> <>
>>> We please need to start somewhere, no?
>>
>> Sure, but you used the operative term "start", as in you already
>> expect to enhance this capability down the road, right?
>>
>
> Yes
>
>> It's fine to dismiss this request_firmware() based approach, but don't
>> mis-characterize it in the process. With regards to describing device
>> boundaries, a bus-descriptor-blob handed to the kernel is a superset
>> of the capability provided by the kernel command line. It can be
>> injected statically at compile time, or dynamically loaded from the
>> initrd or the rootfs. It has the added benefit of being flexible to
>> change whereas the kernel command line is a more permanent contract
>> that we will need to maintain compatibility with in perpetuity.
>>
>
> initrd or rootfs means for me "make install". But I want my fedora
> to never make or install. Pre-compiled binary blobs including rootfs and
> it needs to work.
>
>> If you already see this bus description as a "starting" point, then I
>> think we need an interface that is more amenable to ongoing change,
>> that's not the kernel-command-line.
>>
>
> module-command-line. a module can be loaded via udev and/or module param
> can be changed dynamically on the fly. And also be specified via
> kernel-command-line. So it is much less permanent contract API than
> "rootfs"
>
> And yes, I intend to add more interfaces. And No! I do not intend to
> ever extend this module-param interface, that I can see. This one is
> that, which it is right now. Later a sysfs/ objects will enable dynamic
> management of devices. So both: initial device list on load - more devices
> or removal on the fly, unload all on unload. This is my plan. So right
> now I do not see this map= need ever change in the future. Only more
> interfaces added in (the very near) future.
>

Imagine you want to deploy a policy like "use half of the memory
provided by the dimm in slot3, i.e. the only one with a battery".
That sort of thing gets unwieldy in a command line string compared to
a description table format that we can update at will.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/