Re: [net-next v7 4/7] net: mtip: The L2 switch driver for imx287

[Lukasz Majewski <lukma@denx.de>]

[-- Attachment #1: Type: text/plain, Size: 4182 bytes --]

Hi Krzysztof,

> On 25/04/2025 09:49, Lukasz Majewski wrote:
> > Hi Krzysztof, Jakub
> >   
> >> On 25/04/2025 08:05, Lukasz Majewski wrote:  
> >>> Hi Krzysztof, Jakub,
> >>>     
> >>>> On 25/04/2025 03:11, Jakub Kicinski wrote:    
> >>>>> On Wed, 23 Apr 2025 09:29:08 +0200 Lukasz Majewski wrote:      
> >>>>>> This patch series provides support for More Than IP L2 switch
> >>>>>> embedded in the imx287 SoC.
> >>>>>>
> >>>>>> This is a two port switch (placed between uDMA[01] and
> >>>>>> MAC-NET[01]), which can be used for offloading the network
> >>>>>> traffic.
> >>>>>>
> >>>>>> It can be used interchangeably with current FEC driver - to be
> >>>>>> more specific: one can use either of it, depending on the
> >>>>>> requirements.
> >>>>>>
> >>>>>> The biggest difference is the usage of DMA - when FEC is used,
> >>>>>> separate DMAs are available for each ENET-MAC block.
> >>>>>> However, with switch enabled - only the DMA0 is used to
> >>>>>> send/receive data to/form switch (and then switch sends them to
> >>>>>> respecitive ports).      
> >>>>>
> >>>>> Lots of sparse warnings and build issues here, at least on x86.
> >>>>>
> >>>>> Could you make sure it's clean with an allmodconfig config, 
> >>>>> something like:
> >>>>>
> >>>>> make C=1 W=1 drivers/net/ethernet/freescale/mtipsw/       
> >>>>
> >>>> ... and W=1 with clang as well.
> >>>>    
> >>>
> >>> The sparse warnings are because of struct switch_t casting and
> >>> register    
> >>
> >> clang W=1 fails on errors, so it is not only sparse:
> >>
> >> error: cast to smaller integer type 'uint' (aka 'unsigned int')
> >> from 'struct cbd_t *' [-Werror,-Wpointer-to-int-cast]
> >>
> >> You probably wanted there kenel_ulong_t.  
> > 
> > This I did not catch earlier (probably because of my testing on
> > imx287). Thanks for spotting it.
> >   
> >>  
> >>> access with this paradigm (as it is done with other drivers).    
> >>
> >> I don't understand. I see code like:
> >>
> >> 	struct switch_t *fecp = fep->hwp;
> >>
> >> But this is not a cast - the same types.  
> > 
> > For example:
> > 
> > The warning:
> > 
> > mtipl2sw.c:208:30: warning: incorrect type in argument 1 (different
> > address spaces) mtipl2sw.c:208:30:    expected void const volatile
> > [noderef] __iomem *addr mtipl2sw.c:208:30:    got unsigned int *
> > 
> > corresponds to:
> >  info->maclo = readl(&fecp->ESW_LREC0);   [*]
> > 
> > where:
> > 
> > struct switch_t {
> >         u32 ESW_REVISION;
> >         u32 ESW_SCRATCH;
> > 	...
> >         /*from 0x420-0x4FC*/
> >         u32 esw_reserved9[57];
> >         /*0xFC0DC500---0xFC0DC508*/
> >         u32 ESW_LREC0;
> >         u32 ESW_LREC1;
> >         u32 ESW_LSR;
> > };
> > 
> > 
> > The 'u32' type seems to be valid here as this register is 32 bit
> > wide.  
> 
> It is not about size, but IOMEM annotation and pointer/non-pointer.
> 

+1

> 
> > 
> > To fix the sparse warnings - I think that I will replace [*] with:
> > 
> > info->maclo = readl((u32 __iomem *)&fecp->ESW_LREC0);  
> 
> I don't understand why are you reading address of ESW_LREC0.

The driver (still) uses the apparently "old" programming paradigm, so
there is struct switch_t with u32 elements cast to the __iomem address.

If I want to have the address - I'm using & on the element of the
struct. That is why sparse is complaining as it in fact gets pointer to
u32.

In the fec.h the set of #defines are used and void __iomem *hwp;
pointer.

It looks like to make the spare happy - I need to use similar approach
with the mtip.

The other option would be to add (u32 __iomem *) explicit cast to
readl()/writel().

Or do you see another solution?

> This is
> MMIO, right? So you are supposes to read base + offset (where base is
> a proper iomem pointer).
> 
> 
> Best regards,
> Krzysztof




Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@denx.de

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]