Re: [PATCH 06/10] wifi: iwlwifi: replace deprecated PCI functions

[Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>]

On Fri, 25 Oct 2024, Philipp Stanner wrote:

> On Fri, 2024-10-25 at 19:11 +0300, Ilpo Järvinen wrote:
> > On Fri, 25 Oct 2024, Philipp Stanner wrote:
> > 
> > > On Fri, 2024-10-25 at 18:31 +0300, Ilpo Järvinen wrote:
> > > > On Fri, 25 Oct 2024, Philipp Stanner wrote:
> > > > 
> > > > > pcim_iomap_table() and pcim_iomap_regions_request_all() have
> > > > > been
> > > > > deprecated by the PCI subsystem in commit e354bb84a4c1 ("PCI:
> > > > > Deprecate
> > > > > pcim_iomap_table(), pcim_iomap_regions_request_all()").
> > > > > 
> > > > > Replace these functions with their successors, pcim_iomap() and
> > > > > pcim_request_all_regions().
> > > > > 
> > > > > Signed-off-by: Philipp Stanner <pstanner@xxxxxxxxxx>
> > > > > Acked-by: Kalle Valo <kvalo@xxxxxxxxxx>
> > > > > ---
> > > > >  drivers/net/wireless/intel/iwlwifi/pcie/trans.c | 16 ++++-----
> > > > > ----
> > > > > ---
> > > > >  1 file changed, 4 insertions(+), 12 deletions(-)
> > > > > 
> > > > > diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
> > > > > b/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
> > > > > index 3b9943eb6934..4b41613ad89d 100644
> > > > > --- a/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
> > > > > +++ b/drivers/net/wireless/intel/iwlwifi/pcie/trans.c
> > > > > @@ -3533,7 +3533,6 @@ struct iwl_trans
> > > > > *iwl_trans_pcie_alloc(struct
> > > > > pci_dev *pdev,
> > > > >  	struct iwl_trans_pcie *trans_pcie, **priv;
> > > > >  	struct iwl_trans *trans;
> > > > >  	int ret, addr_size;
> > > > > -	void __iomem * const *table;
> > > > >  	u32 bar0;
> > > > >  
> > > > >  	/* reassign our BAR 0 if invalid due to possible
> > > > > runtime
> > > > > PM races */
> > > > > @@ -3659,22 +3658,15 @@ struct iwl_trans
> > > > > *iwl_trans_pcie_alloc(struct pci_dev *pdev,
> > > > >  		}
> > > > >  	}
> > > > >  
> > > > > -	ret = pcim_iomap_regions_request_all(pdev, BIT(0),
> > > > > DRV_NAME);
> > > > > +	ret = pcim_request_all_regions(pdev, DRV_NAME);
> > > > >  	if (ret) {
> > > > > -		dev_err(&pdev->dev,
> > > > > "pcim_iomap_regions_request_all failed\n");
> > > > > +		dev_err(&pdev->dev, "pcim_request_all_regions
> > > > > failed\n");
> > > > >  		goto out_no_pci;
> > > > >  	}
> > > > >  
> > > > > -	table = pcim_iomap_table(pdev);
> > > > > -	if (!table) {
> > > > > -		dev_err(&pdev->dev, "pcim_iomap_table
> > > > > failed\n");
> > > > > -		ret = -ENOMEM;
> > > > > -		goto out_no_pci;
> > > > > -	}
> > > > > -
> > > > > -	trans_pcie->hw_base = table[0];
> > > > > +	trans_pcie->hw_base = pcim_iomap(pdev, 0, 0);
> > > > >  	if (!trans_pcie->hw_base) {
> > > > > -		dev_err(&pdev->dev, "couldn't find IO mem in
> > > > > first
> > > > > BAR\n");
> > > > > +		dev_err(&pdev->dev, "pcim_iomap failed\n");
> > > > 
> > > > This seems a step backwards as a human readable English error
> > > > message
> > > > was 
> > > > replaced with a reference to a function name.
> > > 
> > > I think it's still an improvement because "couldn't find IO mem in
> > > first BAR" is a nonsensical statement. What the author probably
> > > meant
> > > was: "Couldn't find first BAR's IO mem in magic pci_iomap_table" ;)
> > 
> > Well, that's just spelling things on a too low level too. It's
> > irrelevant
> > detail to the _user_ that kernel used some "magic table". Similarly,
> > it's 
> > irrelevant to the user that function called pcim_iomap failed.
> > 
> > > The reason I just wrote "pcim_iomap failed\n" is that this seems to
> > > be
> > > this driver's style for those messages. See the dev_err() above,
> > > there
> > > they also just state that this or that function failed.
> > 
> > The problem in using function names is they have obvious meaning for 
> > developers/coders but dev_err() is presented to user with varying
> > level
> > of knowledge about kernel internals/code.
> > 
> > While users might be able to derive some information from the
> > function 
> > name, it would be simply better to explain on higher level what
> > failed 
> > which is what I think the original message tried to do even if it was
> > a bit clumsy. There is zero need to know about kernel internals to 
> > interpret that message (arguably one needs to know some PCI to
> > understand 
> > BAR, though).
> > 
> > (Developers can find the internals by looking up the error message
> > from
> > the code so it doesn't take away something from developers.)
> 
> Feel free to make a suggestion for a better error message.
> 
> sth like "could not ioremap PCI BAR 0.\n" could satisfy your criteria.

Yes.

-- 
 i.

> (I just now noticed that so far it called BAR 0 the "first bar", which
> is also not gold standard)