[PATCH V3 1/2] usb: gadget: u_ether: Convert prints to device prints

[PATCH V3 1/2] usb: gadget: u_ether: Convert prints to device prints

[Jon Hunter]

The USB ethernet gadget driver implements its own print macros which
call printk. Device drivers should use the device prints that print the
device name. Fortunately, the same macro names are defined in the header
file 'linux/usb/composite.h' and these use the device prints. Therefore,
remove the local definitions in the USB ethernet gadget driver and use
those in 'linux/usb/composite.h'. The only difference is that now the
device name is printed instead of the ethernet interface name.

Tested using ethernet gadget on Jetson AGX Orin.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
---
V3: Added this patch and dropped the patch in V2 that improved some of
    the prints.

 drivers/usb/gadget/function/u_ether.c | 36 +--------------------------
 1 file changed, 1 insertion(+), 35 deletions(-)

diff --git a/drivers/usb/gadget/function/u_ether.c b/drivers/usb/gadget/function/u_ether.c
index 8f12f3f8f6ee..740331882e8d 100644
--- a/drivers/usb/gadget/function/u_ether.c
+++ b/drivers/usb/gadget/function/u_ether.c
@@ -17,6 +17,7 @@
 #include <linux/etherdevice.h>
 #include <linux/ethtool.h>
 #include <linux/if_vlan.h>
+#include <linux/usb/composite.h>
 
 #include "u_ether.h"
 
@@ -103,41 +104,6 @@ static inline int qlen(struct usb_gadget *gadget, unsigned qmult)
 
 /*-------------------------------------------------------------------------*/
 
-/* REVISIT there must be a better way than having two sets
- * of debug calls ...
- */
-
-#undef DBG
-#undef VDBG
-#undef ERROR
-#undef INFO
-
-#define xprintk(d, level, fmt, args...) \
-	printk(level "%s: " fmt , (d)->net->name , ## args)
-
-#ifdef DEBUG
-#undef DEBUG
-#define DBG(dev, fmt, args...) \
-	xprintk(dev , KERN_DEBUG , fmt , ## args)
-#else
-#define DBG(dev, fmt, args...) \
-	do { } while (0)
-#endif /* DEBUG */
-
-#ifdef VERBOSE_DEBUG
-#define VDBG	DBG
-#else
-#define VDBG(dev, fmt, args...) \
-	do { } while (0)
-#endif /* DEBUG */
-
-#define ERROR(dev, fmt, args...) \
-	xprintk(dev , KERN_ERR , fmt , ## args)
-#define INFO(dev, fmt, args...) \
-	xprintk(dev , KERN_INFO , fmt , ## args)
-
-/*-------------------------------------------------------------------------*/
-
 /* NETWORK DRIVER HOOKUP (to the layer above this driver) */
 
 static void eth_get_drvinfo(struct net_device *net, struct ethtool_drvinfo *p)
-- 
2.34.1

[PATCH V3 2/2] usb: gadget: u_ether: Don't warn in gether_setup_name_default()

[Jon Hunter]

The function gether_setup_name_default() is called by various USB
ethernet gadget drivers. Calling this function will select a random
host and device MAC addresses. A properly working driver should be
silent and not warn the user about default MAC addresses selection.
Given that the MAC addresses are also printed when the function
gether_register_netdev() is called, remove these unnecessary warnings.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
V3: Updated patch to remove the prints completely.
V2: Changed print to debug instead of info.

 drivers/usb/gadget/function/u_ether.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/usb/gadget/function/u_ether.c b/drivers/usb/gadget/function/u_ether.c
index 740331882e8d..953d936fbae6 100644
--- a/drivers/usb/gadget/function/u_ether.c
+++ b/drivers/usb/gadget/function/u_ether.c
@@ -811,13 +811,11 @@ struct net_device *gether_setup_name_default(const char *netname)
 	snprintf(net->name, sizeof(net->name), "%s%%d", netname);
 
 	eth_random_addr(dev->dev_mac);
-	pr_warn("using random %s ethernet address\n", "self");
 
 	/* by default we always have a random MAC address */
 	net->addr_assign_type = NET_ADDR_RANDOM;
 
 	eth_random_addr(dev->host_mac);
-	pr_warn("using random %s ethernet address\n", "host");
 
 	net->netdev_ops = &eth_netdev_ops;
 
-- 
2.34.1

Re: [PATCH V3 2/2] usb: gadget: u_ether: Don't warn in gether_setup_name_default()

[Andrzej Pietrasiewicz]

W dniu 9.02.2023 o 13:53, Jon Hunter pisze:
> The function gether_setup_name_default() is called by various USB
> ethernet gadget drivers. Calling this function will select a random
> host and device MAC addresses. A properly working driver should be
> silent and not warn the user about default MAC addresses selection.
> Given that the MAC addresses are also printed when the function
> gether_register_netdev() is called, remove these unnecessary warnings.
> 
> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>

Reviewed-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>

> ---
> V3: Updated patch to remove the prints completely.
> V2: Changed print to debug instead of info.
> 
>   drivers/usb/gadget/function/u_ether.c | 2 --
>   1 file changed, 2 deletions(-)
> 
> diff --git a/drivers/usb/gadget/function/u_ether.c b/drivers/usb/gadget/function/u_ether.c
> index 740331882e8d..953d936fbae6 100644
> --- a/drivers/usb/gadget/function/u_ether.c
> +++ b/drivers/usb/gadget/function/u_ether.c
> @@ -811,13 +811,11 @@ struct net_device *gether_setup_name_default(const char *netname)
>   	snprintf(net->name, sizeof(net->name), "%s%%d", netname);
>   
>   	eth_random_addr(dev->dev_mac);
> -	pr_warn("using random %s ethernet address\n", "self");
>   
>   	/* by default we always have a random MAC address */
>   	net->addr_assign_type = NET_ADDR_RANDOM;
>   
>   	eth_random_addr(dev->host_mac);
> -	pr_warn("using random %s ethernet address\n", "host");
>   
>   	net->netdev_ops = &eth_netdev_ops;
>   

Re: [PATCH V3 1/2] usb: gadget: u_ether: Convert prints to device prints

[Andrzej Pietrasiewicz]

Hi Jon,

W dniu 9.02.2023 o 13:53, Jon Hunter pisze:
> The USB ethernet gadget driver implements its own print macros which
> call printk. Device drivers should use the device prints that print the
> device name. Fortunately, the same macro names are defined in the header
> file 'linux/usb/composite.h' and these use the device prints. Therefore,
> remove the local definitions in the USB ethernet gadget driver and use
> those in 'linux/usb/composite.h'. The only difference is that now the
> device name is printed instead of the ethernet interface name.
> 
> Tested using ethernet gadget on Jetson AGX Orin.
> 
> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
> Tested-by: Jon Hunter <jonathanh@nvidia.com>
> ---
> V3: Added this patch and dropped the patch in V2 that improved some of
>      the prints.
> 
>   drivers/usb/gadget/function/u_ether.c | 36 +--------------------------
>   1 file changed, 1 insertion(+), 35 deletions(-)
> 
> diff --git a/drivers/usb/gadget/function/u_ether.c b/drivers/usb/gadget/function/u_ether.c
> index 8f12f3f8f6ee..740331882e8d 100644
> --- a/drivers/usb/gadget/function/u_ether.c
> +++ b/drivers/usb/gadget/function/u_ether.c
> @@ -17,6 +17,7 @@
>   #include <linux/etherdevice.h>
>   #include <linux/ethtool.h>
>   #include <linux/if_vlan.h>
> +#include <linux/usb/composite.h>
>   
>   #include "u_ether.h"
>   
> @@ -103,41 +104,6 @@ static inline int qlen(struct usb_gadget *gadget, unsigned qmult)
>   
>   /*-------------------------------------------------------------------------*/
>   
> -/* REVISIT there must be a better way than having two sets
> - * of debug calls ...
> - */
> -
> -#undef DBG
> -#undef VDBG
> -#undef ERROR
> -#undef INFO
> -
> -#define xprintk(d, level, fmt, args...) \
> -	printk(level "%s: " fmt , (d)->net->name , ## args)
> -
> -#ifdef DEBUG
> -#undef DEBUG
> -#define DBG(dev, fmt, args...) \
> -	xprintk(dev , KERN_DEBUG , fmt , ## args)
> -#else
> -#define DBG(dev, fmt, args...) \
> -	do { } while (0)
> -#endif /* DEBUG */

Actually there are more (at least hypothetical) changes than the declared
change of printed device name.

If DEBUG is not set there can be _more_ messages printed
when your patch is applied as-is (i.e. all DBG() invocations will
expand into some dev_dbg(), whereas before the patch is applied
they compile into nothing).

> -
> -#ifdef VERBOSE_DEBUG
> -#define VDBG	DBG
> -#else
> -#define VDBG(dev, fmt, args...) \
> -	do { } while (0)
> -#endif /* DEBUG */

In the same spirit, the above block could be considered valid.

Not sure, how much of an issue it is, though.

Regards,

Andrzej

> -
> -#define ERROR(dev, fmt, args...) \
> -	xprintk(dev , KERN_ERR , fmt , ## args)
> -#define INFO(dev, fmt, args...) \
> -	xprintk(dev , KERN_INFO , fmt , ## args)
> -
> -/*-------------------------------------------------------------------------*/
> -
>   /* NETWORK DRIVER HOOKUP (to the layer above this driver) */
>   
>   static void eth_get_drvinfo(struct net_device *net, struct ethtool_drvinfo *p)

Re: [PATCH V3 1/2] usb: gadget: u_ether: Convert prints to device prints

[Jon Hunter]

On 13/02/2023 13:49, Andrzej Pietrasiewicz wrote:
> Hi Jon,
> 
> W dniu 9.02.2023 o 13:53, Jon Hunter pisze:
>> The USB ethernet gadget driver implements its own print macros which
>> call printk. Device drivers should use the device prints that print the
>> device name. Fortunately, the same macro names are defined in the header
>> file 'linux/usb/composite.h' and these use the device prints. Therefore,
>> remove the local definitions in the USB ethernet gadget driver and use
>> those in 'linux/usb/composite.h'. The only difference is that now the
>> device name is printed instead of the ethernet interface name.
>>
>> Tested using ethernet gadget on Jetson AGX Orin.
>>
>> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
>> Tested-by: Jon Hunter <jonathanh@nvidia.com>
>> ---
>> V3: Added this patch and dropped the patch in V2 that improved some of
>>      the prints.
>>
>>   drivers/usb/gadget/function/u_ether.c | 36 +--------------------------
>>   1 file changed, 1 insertion(+), 35 deletions(-)
>>
>> diff --git a/drivers/usb/gadget/function/u_ether.c 
>> b/drivers/usb/gadget/function/u_ether.c
>> index 8f12f3f8f6ee..740331882e8d 100644
>> --- a/drivers/usb/gadget/function/u_ether.c
>> +++ b/drivers/usb/gadget/function/u_ether.c
>> @@ -17,6 +17,7 @@
>>   #include <linux/etherdevice.h>
>>   #include <linux/ethtool.h>
>>   #include <linux/if_vlan.h>
>> +#include <linux/usb/composite.h>
>>   #include "u_ether.h"
>> @@ -103,41 +104,6 @@ static inline int qlen(struct usb_gadget *gadget, 
>> unsigned qmult)
>>   
>> /*-------------------------------------------------------------------------*/
>> -/* REVISIT there must be a better way than having two sets
>> - * of debug calls ...
>> - */
>> -
>> -#undef DBG
>> -#undef VDBG
>> -#undef ERROR
>> -#undef INFO
>> -
>> -#define xprintk(d, level, fmt, args...) \
>> -    printk(level "%s: " fmt , (d)->net->name , ## args)
>> -
>> -#ifdef DEBUG
>> -#undef DEBUG
>> -#define DBG(dev, fmt, args...) \
>> -    xprintk(dev , KERN_DEBUG , fmt , ## args)
>> -#else
>> -#define DBG(dev, fmt, args...) \
>> -    do { } while (0)
>> -#endif /* DEBUG */
> 
> Actually there are more (at least hypothetical) changes than the declared
> change of printed device name.
> 
> If DEBUG is not set there can be _more_ messages printed
> when your patch is applied as-is (i.e. all DBG() invocations will
> expand into some dev_dbg(), whereas before the patch is applied
> they compile into nothing).

So dev_dbg() messages will only be printed if DEBUG is defined or if 
dynamic debug is enabled and the prints are enabled.

Greg, let me know what your thoughts on this version are?

Thanks
Jon

-- 
nvpublic

Re: [PATCH V3 1/2] usb: gadget: u_ether: Convert prints to device prints

[Greg Kroah-Hartman]

On Mon, Feb 13, 2023 at 02:49:46PM +0100, Andrzej Pietrasiewicz wrote:
> Hi Jon,
> 
> W dniu 9.02.2023 o 13:53, Jon Hunter pisze:
> > The USB ethernet gadget driver implements its own print macros which
> > call printk. Device drivers should use the device prints that print the
> > device name. Fortunately, the same macro names are defined in the header
> > file 'linux/usb/composite.h' and these use the device prints. Therefore,
> > remove the local definitions in the USB ethernet gadget driver and use
> > those in 'linux/usb/composite.h'. The only difference is that now the
> > device name is printed instead of the ethernet interface name.
> > 
> > Tested using ethernet gadget on Jetson AGX Orin.
> > 
> > Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
> > Tested-by: Jon Hunter <jonathanh@nvidia.com>
> > ---
> > V3: Added this patch and dropped the patch in V2 that improved some of
> >      the prints.
> > 
> >   drivers/usb/gadget/function/u_ether.c | 36 +--------------------------
> >   1 file changed, 1 insertion(+), 35 deletions(-)
> > 
> > diff --git a/drivers/usb/gadget/function/u_ether.c b/drivers/usb/gadget/function/u_ether.c
> > index 8f12f3f8f6ee..740331882e8d 100644
> > --- a/drivers/usb/gadget/function/u_ether.c
> > +++ b/drivers/usb/gadget/function/u_ether.c
> > @@ -17,6 +17,7 @@
> >   #include <linux/etherdevice.h>
> >   #include <linux/ethtool.h>
> >   #include <linux/if_vlan.h>
> > +#include <linux/usb/composite.h>
> >   #include "u_ether.h"
> > @@ -103,41 +104,6 @@ static inline int qlen(struct usb_gadget *gadget, unsigned qmult)
> >   /*-------------------------------------------------------------------------*/
> > -/* REVISIT there must be a better way than having two sets
> > - * of debug calls ...
> > - */
> > -
> > -#undef DBG
> > -#undef VDBG
> > -#undef ERROR
> > -#undef INFO
> > -
> > -#define xprintk(d, level, fmt, args...) \
> > -	printk(level "%s: " fmt , (d)->net->name , ## args)
> > -
> > -#ifdef DEBUG
> > -#undef DEBUG
> > -#define DBG(dev, fmt, args...) \
> > -	xprintk(dev , KERN_DEBUG , fmt , ## args)
> > -#else
> > -#define DBG(dev, fmt, args...) \
> > -	do { } while (0)
> > -#endif /* DEBUG */
> 
> Actually there are more (at least hypothetical) changes than the declared
> change of printed device name.
> 
> If DEBUG is not set there can be _more_ messages printed
> when your patch is applied as-is (i.e. all DBG() invocations will
> expand into some dev_dbg(), whereas before the patch is applied
> they compile into nothing).

But if you do not manually set the call to dev_dbg() to be printed out,
it will not, so there should not be any functional change here.

thanks,

greg k-h

Re: [PATCH V3 1/2] usb: gadget: u_ether: Convert prints to device prints

[Greg Kroah-Hartman]

On Tue, Feb 14, 2023 at 11:55:56AM +0000, Jon Hunter wrote:
> 
> On 13/02/2023 13:49, Andrzej Pietrasiewicz wrote:
> > Hi Jon,
> > 
> > W dniu 9.02.2023 o 13:53, Jon Hunter pisze:
> > > The USB ethernet gadget driver implements its own print macros which
> > > call printk. Device drivers should use the device prints that print the
> > > device name. Fortunately, the same macro names are defined in the header
> > > file 'linux/usb/composite.h' and these use the device prints. Therefore,
> > > remove the local definitions in the USB ethernet gadget driver and use
> > > those in 'linux/usb/composite.h'. The only difference is that now the
> > > device name is printed instead of the ethernet interface name.
> > > 
> > > Tested using ethernet gadget on Jetson AGX Orin.
> > > 
> > > Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
> > > Tested-by: Jon Hunter <jonathanh@nvidia.com>
> > > ---
> > > V3: Added this patch and dropped the patch in V2 that improved some of
> > >      the prints.
> > > 
> > >   drivers/usb/gadget/function/u_ether.c | 36 +--------------------------
> > >   1 file changed, 1 insertion(+), 35 deletions(-)
> > > 
> > > diff --git a/drivers/usb/gadget/function/u_ether.c
> > > b/drivers/usb/gadget/function/u_ether.c
> > > index 8f12f3f8f6ee..740331882e8d 100644
> > > --- a/drivers/usb/gadget/function/u_ether.c
> > > +++ b/drivers/usb/gadget/function/u_ether.c
> > > @@ -17,6 +17,7 @@
> > >   #include <linux/etherdevice.h>
> > >   #include <linux/ethtool.h>
> > >   #include <linux/if_vlan.h>
> > > +#include <linux/usb/composite.h>
> > >   #include "u_ether.h"
> > > @@ -103,41 +104,6 @@ static inline int qlen(struct usb_gadget
> > > *gadget, unsigned qmult)
> > > /*-------------------------------------------------------------------------*/
> > > -/* REVISIT there must be a better way than having two sets
> > > - * of debug calls ...
> > > - */
> > > -
> > > -#undef DBG
> > > -#undef VDBG
> > > -#undef ERROR
> > > -#undef INFO
> > > -
> > > -#define xprintk(d, level, fmt, args...) \
> > > -    printk(level "%s: " fmt , (d)->net->name , ## args)
> > > -
> > > -#ifdef DEBUG
> > > -#undef DEBUG
> > > -#define DBG(dev, fmt, args...) \
> > > -    xprintk(dev , KERN_DEBUG , fmt , ## args)
> > > -#else
> > > -#define DBG(dev, fmt, args...) \
> > > -    do { } while (0)
> > > -#endif /* DEBUG */
> > 
> > Actually there are more (at least hypothetical) changes than the declared
> > change of printed device name.
> > 
> > If DEBUG is not set there can be _more_ messages printed
> > when your patch is applied as-is (i.e. all DBG() invocations will
> > expand into some dev_dbg(), whereas before the patch is applied
> > they compile into nothing).
> 
> So dev_dbg() messages will only be printed if DEBUG is defined or if dynamic
> debug is enabled and the prints are enabled.
> 
> Greg, let me know what your thoughts on this version are?

Looks good to me, let me queue it up now.

thanks,

greg k-h