Re: [PATCH v2 2/9] mfd: cs5535-mfd: Remove mfd_cell->id hack

[Lee Jones]

On Mon, 21 Oct 2019, Daniel Thompson wrote:

> On Mon, Oct 21, 2019 at 11:58:15AM +0100, Lee Jones wrote:
> > The current implementation abuses the platform 'id' mfd_cell member
> > to index into the correct resources entry.  If we place all cells
> > into their numbered slots, we can cycle through all the cell entries
> > and only process the populated ones which avoids this behaviour.
> > 
> > Signed-off-by: Lee Jones <lee.jones@xxxxxxxxxx>
> > ---
> >  drivers/mfd/cs5535-mfd.c | 31 +++++++++++++------------------
> >  1 file changed, 13 insertions(+), 18 deletions(-)
> > 
> > diff --git a/drivers/mfd/cs5535-mfd.c b/drivers/mfd/cs5535-mfd.c
> > index 2c47afc22d24..9ce6bbcdbda1 100644
> > --- a/drivers/mfd/cs5535-mfd.c
> > +++ b/drivers/mfd/cs5535-mfd.c
> > @@ -62,26 +62,22 @@ static int cs5535_mfd_res_disable(struct platform_device *pdev)
> >  static struct resource cs5535_mfd_resources[NR_BARS];
> >  
> >  static struct mfd_cell cs5535_mfd_cells[] = {
> 
> This array is sized from the initializer...
> 
> > -	{
> > -		.id = SMB_BAR,
> > +	[SMB_BAR] = {
> >  		.name = "cs5535-smb",
> >  		.num_resources = 1,
> >  		.resources = &cs5535_mfd_resources[SMB_BAR],
> >  	},
> > -	{
> > -		.id = GPIO_BAR,
> > +	[GPIO_BAR] = {
> >  		.name = "cs5535-gpio",
> >  		.num_resources = 1,
> >  		.resources = &cs5535_mfd_resources[GPIO_BAR],
> >  	},
> > -	{
> > -		.id = MFGPT_BAR,
> > +	[MFGPT_BAR] = {
> >  		.name = "cs5535-mfgpt",
> >  		.num_resources = 1,
> >  		.resources = &cs5535_mfd_resources[MFGPT_BAR],
> >  	},
> > -	{
> > -		.id = PMS_BAR,
> > +	[PMS_BAR] = {
> >  		.name = "cs5535-pms",
> >  		.num_resources = 1,
> >  		.resources = &cs5535_mfd_resources[PMS_BAR],
> > @@ -89,8 +85,7 @@ static struct mfd_cell cs5535_mfd_cells[] = {
> >  		.enable = cs5535_mfd_res_enable,
> >  		.disable = cs5535_mfd_res_disable,
> >  	},
> > -	{
> > -		.id = ACPI_BAR,
> > +	[ACPI_BAR] = {
> >  		.name = "cs5535-acpi",
> >  		.num_resources = 1,
> >  		.resources = &cs5535_mfd_resources[ACPI_BAR],
> > @@ -115,16 +110,16 @@ static int cs5535_mfd_probe(struct pci_dev *pdev,
> >  		return err;
> >  
> >  	/* fill in IO range for each cell; subdrivers handle the region */
> > -	for (i = 0; i < ARRAY_SIZE(cs5535_mfd_cells); i++) {
> > -		int bar = cs5535_mfd_cells[i].id;
> > -		struct resource *r = &cs5535_mfd_resources[bar];
> > +	for (i = 0; i < NR_BARS; i++) {
> 
> ... which means this translation from ARRAY_SIZE() to NR_BARS
> is rather odd.
> 
> I don't care whether the array is sized using NR_BARS or the loop
> uses ARRAY_SIZE() but IMHO the loop boundary condition must match
> the array declaration.

Sounds reasonable.

> With that fixed free to throw the following onto the next rev:
> Reviewed-by: Daniel Thompson <daniel.thompson@xxxxxxxxxx>

Ta.

-- 
Lee Jones [æçæ]
Linaro Services Technical Lead
Linaro.org â Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog