Re: [PATCH 1/2] i3c: master: Fix confusing cleanup.h syntax

[Alexandre Belloni <alexandre.belloni@bootlin.com>]

Hello Frank,

On 08/12/2025 10:50:50-0500, Frank Li wrote:
> On Mon, Dec 08, 2025 at 03:07:51AM +0100, Krzysztof Kozlowski wrote:
> > Initializing automatic __free variables to NULL without need (e.g.
> > branches with different allocations), followed by actual allocation is
> > in contrary to explicit coding rules guiding cleanup.h:
> >
> > "Given that the "__free(...) = NULL" pattern for variables defined at
> > the top of the function poses this potential interdependency problem the
> > recommendation is to always define and assign variables in one statement
> > and not group variable definitions at the top of the function when
> > __free() is used."
> >
> > Code does not have a bug, but is less readable and uses discouraged
> > coding practice, so fix that by moving declaration to the place of
> > assignment.
> >
> > Not that other existing usage of __free() in this context is a corret
> > exception initialized to NULL, because the actual allocation is branched
> > in if().
> >
> > Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
> > ---
> >  drivers/i3c/master.c | 3 +--
> >  1 file changed, 1 insertion(+), 2 deletions(-)
> >
> > diff --git a/drivers/i3c/master.c b/drivers/i3c/master.c
> > index 823661a81f5e..62437a899cdc 100644
> > --- a/drivers/i3c/master.c
> > +++ b/drivers/i3c/master.c
> > @@ -1742,11 +1742,10 @@ EXPORT_SYMBOL_GPL(i3c_master_do_daa);
> >  struct i3c_dma *i3c_master_dma_map_single(struct device *dev, void *buf,
> >  	size_t len, bool force_bounce, enum dma_data_direction dir)
> >  {
> > -	struct i3c_dma *dma_xfer __free(kfree) = NULL;
> >  	void *bounce __free(kfree) = NULL;
> >  	void *dma_buf = buf;
> >
> > -	dma_xfer = kzalloc(sizeof(*dma_xfer), GFP_KERNEL);
> > +	struct i3c_dma *dma_xfer __free(kfree) = kzalloc(sizeof(*dma_xfer), GFP_KERNEL);
> >  	if (!dma_xfer)
> >  		return NULL;
> 
> According to thread
> https://lore.kernel.org/all/CAHk-=whPZoi03ZwphxiW6cuWPtC3nyKYS8_BThgztCdgPWP1WA@mail.gmail.com/
> "But again: I don't want to make this some kind of hard rule, and I
> think it should be done judiciously and with taste, not some kind of
> crazy conversion thing."
> 
> And there are not logic code between declear and kzalloc.
> https://lore.kernel.org/all/CAHk-=wjMMSAaqTjBSfYenfuzE1bMjLj+2DLtLWJuGt07UGCH_Q@mail.gmail.com/
> 	struct ring_buffer_per_cpu *cpu_buffer __free(kfree);
> 
>         cpu_buffer = kzalloc_node(...);
> 
> as separate things (but probably next to each other, so that the
> "__free(kfree)" part makes sense because the allocation is right
> there)."
> 
> void *bounce __free(kfree) = NULL, still be there. Only cleanup one is
> not enough valuable.
> 
> https://lore.kernel.org/all/CAHk-=wjMMSAaqTjBSfYenfuzE1bMjLj+2DLtLWJuGt07UGCH_Q@mail.gmail.com/#t
> "   would then make that otherwise nasty allocation then become just
> 
>     auto cpu_buffer __free(kfree) = cpu_buffer_alloc();
> "
> 
> Maybe we can defer this cleanup after kernel using "auto" keywords widely.

While you are absolutely correct, I'll apply those just because else
we'll get many people submitting the same patch and it indeed makes the
code nicer.

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com