[PATCH] declance: Fix 64-bit compilation warnings

[PATCH] declance: Fix 64-bit compilation warnings

[Maciej W. Rozycki]

This fixes compiler warnings:

drivers/net/ethernet/amd/declance.c: In function 'lance_init_ring':
drivers/net/ethernet/amd/declance.c:478: warning: format '%8.8x' expects type 'unsigned int', but argument 3 has type 'long unsigned int'
drivers/net/ethernet/amd/declance.c:487: warning: format '%8.8x' expects type 'unsigned int', but argument 3 has type 'long unsigned int'
drivers/net/ethernet/amd/declance.c:503: warning: cast from pointer to integer of different size
drivers/net/ethernet/amd/declance.c:520: warning: cast from pointer to integer of different size

in 64-bit compilation.  Where the value printed is an offset (whose range 
will always fit) the cast uses a 32-bit type, otherwise, where it is a 
host memory address, the `long' type is used and the width of the number 
printed adjusted according to the size of this type.

Tested with both 32-bit and 64-bit compilation, as well as at the run time 
(with the debug messages affected enabled).

Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
---
linux-declance-desc-printk.patch
Index: net-next-20140424-4maxp64/drivers/net/ethernet/amd/declance.c
===================================================================
--- net-next-20140424-4maxp64.orig/drivers/net/ethernet/amd/declance.c
+++ net-next-20140424-4maxp64/drivers/net/ethernet/amd/declance.c
@@ -475,7 +475,7 @@ static void lance_init_ring(struct net_d
 	*lib_ptr(ib, rx_ptr, lp->type) = leptr;
 	if (ZERO)
 		printk("RX ptr: %8.8x(%8.8x)\n",
-		       leptr, lib_off(brx_ring, lp->type));
+		       leptr, (uint)lib_off(brx_ring, lp->type));
 
 	/* Setup tx descriptor pointer */
 	leptr = offsetof(struct lance_init_block, btx_ring);
@@ -484,7 +484,7 @@ static void lance_init_ring(struct net_d
 	*lib_ptr(ib, tx_ptr, lp->type) = leptr;
 	if (ZERO)
 		printk("TX ptr: %8.8x(%8.8x)\n",
-		       leptr, lib_off(btx_ring, lp->type));
+		       leptr, (uint)lib_off(btx_ring, lp->type));
 
 	if (ZERO)
 		printk("TX rings:\n");
@@ -499,8 +499,9 @@ static void lance_init_ring(struct net_d
 						/* The ones required by tmd2 */
 		*lib_ptr(ib, btx_ring[i].misc, lp->type) = 0;
 		if (i < 3 && ZERO)
-			printk("%d: 0x%8.8x(0x%8.8x)\n",
-			       i, leptr, (uint)lp->tx_buf_ptr_cpu[i]);
+			printk("%d: 0x%8.8x(%#0*lx)\n",
+			       i, leptr, (int)sizeof(long),
+			       (long)lp->tx_buf_ptr_cpu[i]);
 	}
 
 	/* Setup the Rx ring entries */
@@ -516,8 +517,9 @@ static void lance_init_ring(struct net_d
 							     0xf000;
 		*lib_ptr(ib, brx_ring[i].mblength, lp->type) = 0;
 		if (i < 3 && ZERO)
-			printk("%d: 0x%8.8x(0x%8.8x)\n",
-			       i, leptr, (uint)lp->rx_buf_ptr_cpu[i]);
+			printk("%d: 0x%8.8x(%#0*lx)\n",
+			       i, leptr, (int)sizeof(long),
+			       (long)lp->rx_buf_ptr_cpu[i]);
 	}
 	iob();
 }

Re: [PATCH] declance: Fix 64-bit compilation warnings

[Joe Perches]

On Sat, 2014-06-28 at 23:57 +0100, Maciej W. Rozycki wrote:
> This fixes compiler warnings:
> 
> drivers/net/ethernet/amd/declance.c: In function 'lance_init_ring':
> drivers/net/ethernet/amd/declance.c:478: warning: format '%8.8x' expects type 'unsigned int', but argument 3 has type 'long unsigned int'
> drivers/net/ethernet/amd/declance.c:487: warning: format '%8.8x' expects type 'unsigned int', but argument 3 has type 'long unsigned int'
> drivers/net/ethernet/amd/declance.c:503: warning: cast from pointer to integer of different size
> drivers/net/ethernet/amd/declance.c:520: warning: cast from pointer to integer of different size
> 
> in 64-bit compilation.  Where the value printed is an offset (whose range 
> will always fit) the cast uses a 32-bit type, otherwise, where it is a 
> host memory address, the `long' type is used and the width of the number 
> printed adjusted according to the size of this type.
> 
> Tested with both 32-bit and 64-bit compilation, as well as at the run time 
> (with the debug messages affected enabled).
[]
> Index: net-next-20140424-4maxp64/drivers/net/ethernet/amd/declance.c
[]
> @@ -499,8 +499,9 @@ static void lance_init_ring(struct net_d
>  						/* The ones required by tmd2 */
>  		*lib_ptr(ib, btx_ring[i].misc, lp->type) = 0;
>  		if (i < 3 && ZERO)
> -			printk("%d: 0x%8.8x(0x%8.8x)\n",
> -			       i, leptr, (uint)lp->tx_buf_ptr_cpu[i]);
> +			printk("%d: 0x%8.8x(%#0*lx)\n",
> +			       i, leptr, (int)sizeof(long),
> +			       (long)lp->tx_buf_ptr_cpu[i]);

You need to adjust the "*" and sizeof(long) with +2 for the
0x prefix in the output length here.

Re: [PATCH] declance: Fix 64-bit compilation warnings

[Maciej W. Rozycki]

On Sat, 28 Jun 2014, Joe Perches wrote:

> > @@ -499,8 +499,9 @@ static void lance_init_ring(struct net_d
> >  						/* The ones required by tmd2 */
> >  		*lib_ptr(ib, btx_ring[i].misc, lp->type) = 0;
> >  		if (i < 3 && ZERO)
> > -			printk("%d: 0x%8.8x(0x%8.8x)\n",
> > -			       i, leptr, (uint)lp->tx_buf_ptr_cpu[i]);
> > +			printk("%d: 0x%8.8x(%#0*lx)\n",
> > +			       i, leptr, (int)sizeof(long),
> > +			       (long)lp->tx_buf_ptr_cpu[i]);
> 
> You need to adjust the "*" and sizeof(long) with +2 for the
> 0x prefix in the output length here.

 Good catch, thanks!  MIPS kernel addresses always have their MSB set so 
this is hard to notice in testing.  But also the width has to be doubled 
first, to take two digits per byte into account too.  Sending an updated 
change right away, as soon as it builds and boots.

  Maciej

RE: [PATCH] declance: Fix 64-bit compilation warnings

[David Laight]

From: Joe Perches
> On Sat, 2014-06-28 at 23:57 +0100, Maciej W. Rozycki wrote:
> > This fixes compiler warnings:
> >
> > drivers/net/ethernet/amd/declance.c: In function 'lance_init_ring':
> > drivers/net/ethernet/amd/declance.c:478: warning: format '%8.8x' expects type 'unsigned int', but
> argument 3 has type 'long unsigned int'
> > drivers/net/ethernet/amd/declance.c:487: warning: format '%8.8x' expects type 'unsigned int', but
> argument 3 has type 'long unsigned int'
> > drivers/net/ethernet/amd/declance.c:503: warning: cast from pointer to integer of different size
> > drivers/net/ethernet/amd/declance.c:520: warning: cast from pointer to integer of different size
> >
> > in 64-bit compilation.  Where the value printed is an offset (whose range
> > will always fit) the cast uses a 32-bit type, otherwise, where it is a
> > host memory address, the `long' type is used and the width of the number
> > printed adjusted according to the size of this type.
> >
> > Tested with both 32-bit and 64-bit compilation, as well as at the run time
> > (with the debug messages affected enabled).
> []
> > Index: net-next-20140424-4maxp64/drivers/net/ethernet/amd/declance.c
> []
> > @@ -499,8 +499,9 @@ static void lance_init_ring(struct net_d
> >  						/* The ones required by tmd2 */
> >  		*lib_ptr(ib, btx_ring[i].misc, lp->type) = 0;
> >  		if (i < 3 && ZERO)
> > -			printk("%d: 0x%8.8x(0x%8.8x)\n",
> > -			       i, leptr, (uint)lp->tx_buf_ptr_cpu[i]);
> > +			printk("%d: 0x%8.8x(%#0*lx)\n",
> > +			       i, leptr, (int)sizeof(long),
> > +			       (long)lp->tx_buf_ptr_cpu[i]);
> 
> You need to adjust the "*" and sizeof(long) with +2 for the
> 0x prefix in the output length here.

Do you need all the leading zeros at all?

	David

RE: [PATCH] declance: Fix 64-bit compilation warnings

[Maciej W. Rozycki]

On Mon, 30 Jun 2014, David Laight wrote:

> > > @@ -499,8 +499,9 @@ static void lance_init_ring(struct net_d
> > >  						/* The ones required by tmd2 */
> > >  		*lib_ptr(ib, btx_ring[i].misc, lp->type) = 0;
> > >  		if (i < 3 && ZERO)
> > > -			printk("%d: 0x%8.8x(0x%8.8x)\n",
> > > -			       i, leptr, (uint)lp->tx_buf_ptr_cpu[i]);
> > > +			printk("%d: 0x%8.8x(%#0*lx)\n",
> > > +			       i, leptr, (int)sizeof(long),
> > > +			       (long)lp->tx_buf_ptr_cpu[i]);
> > 
> > You need to adjust the "*" and sizeof(long) with +2 for the
> > 0x prefix in the output length here.
> 
> Do you need all the leading zeros at all?

 For output alignment; plus the change is not intended to change 
semantics, but just to address the lack of 64-bit cleanliness.

  Maciej