debugfs_create_{u,x,s}{8,16,32,64}

debugfs_create_{u,x,s}{8,16,32,64}

[Matt Longnecker]

Greg,

Back in December 2007 Mattias Nissler proposed a patch defining 
debugfs_create_s32 and friends. In it he reworked the already existent u 
and x versions. The subject of Mattias's mail was "[PATCH] debugfs: 
Revamp debugfs_create_{u,x,s}{8,16,32,64} to support signed integers".

As I'm writing a new driver related to thermal management, I'm finding 
the absence of debugfs_create_s32 to be a stumbling block. I see at 
least two other folks who have stumbled on the absence of 
debugfs_create_s32 in the past few years.

Why didn't we merge Mattias's patch? Would you entertain a similar patch 
today?

Thanks,
Matt Longnecker

Re: debugfs_create_{u,x,s}{8,16,32,64}

[Greg KH]

On Wed, Aug 13, 2014 at 02:57:05PM -0700, Matt Longnecker wrote:
> Greg,
> 
> Back in December 2007 Mattias Nissler proposed a patch defining
> debugfs_create_s32 and friends. In it he reworked the already existent u and
> x versions. The subject of Mattias's mail was "[PATCH] debugfs: Revamp
> debugfs_create_{u,x,s}{8,16,32,64} to support signed integers".
> 
> As I'm writing a new driver related to thermal management, I'm finding the
> absence of debugfs_create_s32 to be a stumbling block. I see at least two
> other folks who have stumbled on the absence of debugfs_create_s32 in the
> past few years.
> 
> Why didn't we merge Mattias's patch?

I can't remember why I didn't merge a patch from last week, let alone
from 2007, are you crazy?  :)

> Would you entertain a similar patch today?

I always entertain all patches, you never need to ask if you are allowed
to send a patch, that's like asking "Am I allowed to ask a question?"...

greg k-h

Re: debugfs_create_{u,x,s}{8,16,32,64}

[Mattias Nissler]

On Wed, Aug 13, 2014 at 11:57 PM, Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Wed, Aug 13, 2014 at 02:57:05PM -0700, Matt Longnecker wrote:
> > Greg,
> >
> > Back in December 2007 Mattias Nissler proposed a patch defining
> > debugfs_create_s32 and friends. In it he reworked the already existent u and
> > x versions. The subject of Mattias's mail was "[PATCH] debugfs: Revamp
> > debugfs_create_{u,x,s}{8,16,32,64} to support signed integers".
> >
> > As I'm writing a new driver related to thermal management, I'm finding the
> > absence of debugfs_create_s32 to be a stumbling block. I see at least two
> > other folks who have stumbled on the absence of debugfs_create_s32 in the
> > past few years.
> >
> > Why didn't we merge Mattias's patch?
>
> I can't remember why I didn't merge a patch from last week, let alone
> from 2007, are you crazy?  :)

Needless to say, I don't remember details either. It might just have
been a case of not adding further debugfs API as long as there's
little demand to justify the addition. The latter may have changed in
the last 7 years (Matt hints at other potential consumers?), but I
personally can't tell as I haven't done any kernel work in more recent
times.

>
>
> > Would you entertain a similar patch today?
>
> I always entertain all patches, you never need to ask if you are allowed
> to send a patch, that's like asking "Am I allowed to ask a question?"...
>
> greg k-h