Re: [PATCH 1/4] firmware: generalize "firmware" as "system data" helpers

2015-12-17 Thread Luis R. Rodriguez 
On Thu, Oct 8, 2015 at 1:16 PM, Josh Boyer  wrote:
> On Tue, Oct 6, 2015 at 5:08 AM, Greg KH  wrote:
>> Just responding to one thing at the moment:
>>
>> On Mon, Oct 05, 2015 at 11:22:22PM +0200, Luis R. Rodriguez wrote:
>>>   * we should phase out the usermode helper from firmware_class long term
>>
>> You can "phase out", but you can not delete it as it's a user/kernel api
>> that we have to support for forever, sorry.
>
> Assuming that dell-rbu is the only in-tree legitimate user of the
> userhelper code, I'm curious if the code itself could simply move into
> that driver.  It might help prevent the spread of reliance on an API
> we don't want to see grow in usage.  We'd probably need to evaluate if
> the two new users could migrate off that.

Greg pointed out Daniel might have some uses for this. More on this later.

>> Also, for some devices / use cases, the usermode helper is the solution
>> (think async loading of firmware when the host wants to do it.)
>
> Are any of those use cases in the kernel today, aside from dell-rbu?
> Would Luis' async mode to system data suffice?

We'll have to see based on Daniel's feedback (Daniel, please respond
to the other thread I'll Cc you on).

 Luis
--
To unsubscribe from this list: send the line "unsubscribe 
linux-security-module" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/4] firmware: generalize "firmware" as "system data" helpers

2015-10-08 Thread Josh Boyer 
On Tue, Oct 6, 2015 at 5:08 AM, Greg KH  wrote:
> Just responding to one thing at the moment:
>
> On Mon, Oct 05, 2015 at 11:22:22PM +0200, Luis R. Rodriguez wrote:
>>   * we should phase out the usermode helper from firmware_class long term
>
> You can "phase out", but you can not delete it as it's a user/kernel api
> that we have to support for forever, sorry.

Assuming that dell-rbu is the only in-tree legitimate user of the
userhelper code, I'm curious if the code itself could simply move into
that driver.  It might help prevent the spread of reliance on an API
we don't want to see grow in usage.  We'd probably need to evaluate if
the two new users could migrate off that.

> Also, for some devices / use cases, the usermode helper is the solution
> (think async loading of firmware when the host wants to do it.)

Are any of those use cases in the kernel today, aside from dell-rbu?
Would Luis' async mode to system data suffice?

josh
--
To unsubscribe from this list: send the line "unsubscribe 
linux-security-module" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/4] firmware: generalize "firmware" as "system data" helpers

2015-10-06 Thread Greg KH 
Just responding to one thing at the moment:

On Mon, Oct 05, 2015 at 11:22:22PM +0200, Luis R. Rodriguez wrote:
>   * we should phase out the usermode helper from firmware_class long term

You can "phase out", but you can not delete it as it's a user/kernel api
that we have to support for forever, sorry.

Also, for some devices / use cases, the usermode helper is the solution
(think async loading of firmware when the host wants to do it.)

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe 
linux-security-module" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/4] firmware: generalize "firmware" as "system data" helpers]

2015-10-06 Thread Luis R. Rodriguez 
On Tue, Oct 06, 2015 at 10:08:21AM +0100, Greg KH wrote:
> Just responding to one thing at the moment:
> 
> On Mon, Oct 05, 2015 at 11:22:22PM +0200, Luis R. Rodriguez wrote:
> >   * we should phase out the usermode helper from firmware_class long term
> 
> You can "phase out", but you can not delete it as it's a user/kernel api
> that we have to support for forever, sorry.
>
>
> Also, for some devices / use cases, the usermode helper is the solution
> (think async loading of firmware when the host wants to do it.)

Sure, this can still be kept in a dark corner, no need for it to clutter
or get in the way of creating cleaner APIs. That's one of the goals here,
and going through with these changes should help us get there.

  Luis
--
To unsubscribe from this list: send the line "unsubscribe 
linux-security-module" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/4] firmware: generalize "firmware" as "system data" helpers

2015-10-04 Thread Greg KH 
On Tue, Aug 04, 2015 at 03:00:01PM -0700, Luis R. Rodriguez wrote:
> From: "Luis R. Rodriguez" 
> 
> Historically firmware_class code was added to help
> get device driver firmware binaries but these days
> request_firmware*() helpers are being repurposed for
> general system data needed by the kernel.
> 
> Annotate this before we extend firmare_class more,
> as this is expected. We want to generalize the code
> as much as possible.

No, let's leave this as "firmware", as that is what the code does.

If you want to create a "load a resource from the filesystem into the
kernel" subsystem, then let's do that and then port the firmware loader
code over to use that api.

But until then, let's not try to morph the firmware code into something
that it really is not at all at the moment, just because it looks like
this might be a nice thing to do someday in the future.

sorry,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe 
linux-security-module" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html