RE: [PATCH 1/7] staging: fsl-mc: MC bus IRQ support

[Jose Rivera]

> -----Original Message-----
> From: Wood Scott-B07421
> Sent: Tuesday, May 05, 2015 2:42 PM
> To: Rivera Jose-B46482
> Cc: Dan Carpenter; devel@xxxxxxxxxxxxxxxxxxxx; agraf@xxxxxxx;
> arnd@xxxxxxxx; Sharma Bhupesh-B45370; gregkh@xxxxxxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; Yoder Stuart-B08248; Erez Nir-RM30794; katz Itai-
> RM05202; Hamciuc Bogdan-BHAMCIU1; Marginean Alexandru-R89243; Schmitt
> Richard-B43082
> Subject: Re: [PATCH 1/7] staging: fsl-mc: MC bus IRQ support
> 
> On Tue, 2015-05-05 at 11:08 -0500, Rivera Jose-B46482 wrote:
> > > > > to read what "goto error;" does.  The error handling here calls
> > > > > devm_kfree() which is not needed...  devm_ functions
> > > > > automatically clean up after themselves.  This seems a pattern
> > > > > throughout.  Do a search for
> > > > > devm_free() and see which ones are really needed or not.
> > > > >
> > > > I know that memory allocated with devm_kzalloc() is freed at the
> > > > end of the lifetime of the device it is attached to. However, in
> > > > error paths, why wait until the device is destroyed? Why not free
> > > > the memory earlier so that it can be used for other purposes?
> > >
> > Why then do the devm_kfree() function exist?
> >
> > I will not remove the devm_free() calls unless the upstream maintainer
> > requires me to do so.
> 
> The whole point of devm is to simplify (often poorly tested) error paths.
> You're trying to optimize the error path instead of simplify it, which
> doesn't make sense.
> 
> devm_kfree() is for the uncommon case when you want to free the memory in
> a situation where the device isn't being unbound any time soon, but you
> still want the memory to be handled by devm for some reason.
> 
> Most users of devm_kzalloc() do not use devm_kfree():
> scott@snotra:~/fsl/git/linux/upstream$ git grep devm_kzalloc|wc -l
> 2750
> scott@snotra:~/fsl/git/linux/upstream$ git grep devm_kfree|wc -l
> 118
> 
Ok, I'll remove the devm_kfree() calls.

Thanks,

German

> -Scott
>