[PATCH] PCI: Mark expected switch fall-throughs

[PATCH] PCI: Mark expected switch fall-throughs

[Gustavo A. R. Silva]

In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.

This patch fixes the following warnings:

drivers/pci/proc.c: In function ‘proc_bus_pci_ioctl’:
drivers/pci/proc.c:216:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (arch_can_pci_mmap_wc()) {
      ^
drivers/pci/proc.c:225:2: note: here
  default:
  ^~~~~~~

drivers/pci/xen-pcifront.c: In function ‘pcifront_backend_changed’:
drivers/pci/xen-pcifront.c:1105:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (xdev->state == XenbusStateClosed)
      ^
drivers/pci/xen-pcifront.c:1108:2: note: here
  case XenbusStateClosing:
  ^~~~

Notice that, in this particular case, the /* fall through */
comment is placed at the very bottom of the case statement,
which is what GCC is expecting to find.

Warning level 3 was used: -Wimplicit-fallthrough=3

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
---
 drivers/pci/proc.c         | 1 +
 drivers/pci/xen-pcifront.c | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
index 6fa1627ce08d..445b51db75b0 100644
--- a/drivers/pci/proc.c
+++ b/drivers/pci/proc.c
@@ -222,6 +222,7 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
 		}
 		/* If arch decided it can't, fall through... */
 #endif /* HAVE_PCI_MMAP */
+		/* fall through */
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
index eba6e33147a2..14cf0f41ecf0 100644
--- a/drivers/pci/xen-pcifront.c
+++ b/drivers/pci/xen-pcifront.c
@@ -1104,7 +1104,7 @@ static void __ref pcifront_backend_changed(struct xenbus_device *xdev,
 	case XenbusStateClosed:
 		if (xdev->state == XenbusStateClosed)
 			break;
-		/* Missed the backend's CLOSING state -- fallthrough */
+		/* fall through - Missed the backend's CLOSING state. */
 	case XenbusStateClosing:
 		dev_warn(&xdev->dev, "backend going away!\n");
 		pcifront_try_disconnect(pdev);
-- 
2.21.0

Re: [PATCH] PCI: Mark expected switch fall-throughs

[Bjorn Helgaas]

On Wed, Mar 20, 2019 at 01:27:15PM -0500, Gustavo A. R. Silva wrote:
> In preparation to enabling -Wimplicit-fallthrough, mark switch
> cases where we are expecting to fall through.

Does this fix all the remaining cases in drivers/pci?  I'd like to fix
them all at once.

What's the best way to watch for new warnings being added?  I fiddled
with "make W=2" etc but it didn't seem useful.  Does the 0-day robot
warn when -Wimplicit-fallthrough warnings are added?

Bjorn

> This patch fixes the following warnings:
> 
> drivers/pci/proc.c: In function ‘proc_bus_pci_ioctl’:
> drivers/pci/proc.c:216:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
>    if (arch_can_pci_mmap_wc()) {
>       ^
> drivers/pci/proc.c:225:2: note: here
>   default:
>   ^~~~~~~
> 
> drivers/pci/xen-pcifront.c: In function ‘pcifront_backend_changed’:
> drivers/pci/xen-pcifront.c:1105:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
>    if (xdev->state == XenbusStateClosed)
>       ^
> drivers/pci/xen-pcifront.c:1108:2: note: here
>   case XenbusStateClosing:
>   ^~~~
> 
> Notice that, in this particular case, the /* fall through */
> comment is placed at the very bottom of the case statement,
> which is what GCC is expecting to find.
> 
> Warning level 3 was used: -Wimplicit-fallthrough=3
> 
> This patch is part of the ongoing efforts to enable
> -Wimplicit-fallthrough.
> 
> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
> ---
>  drivers/pci/proc.c         | 1 +
>  drivers/pci/xen-pcifront.c | 2 +-
>  2 files changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
> index 6fa1627ce08d..445b51db75b0 100644
> --- a/drivers/pci/proc.c
> +++ b/drivers/pci/proc.c
> @@ -222,6 +222,7 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
>  		}
>  		/* If arch decided it can't, fall through... */
>  #endif /* HAVE_PCI_MMAP */
> +		/* fall through */
>  	default:
>  		ret = -EINVAL;
>  		break;
> diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
> index eba6e33147a2..14cf0f41ecf0 100644
> --- a/drivers/pci/xen-pcifront.c
> +++ b/drivers/pci/xen-pcifront.c
> @@ -1104,7 +1104,7 @@ static void __ref pcifront_backend_changed(struct xenbus_device *xdev,
>  	case XenbusStateClosed:
>  		if (xdev->state == XenbusStateClosed)
>  			break;
> -		/* Missed the backend's CLOSING state -- fallthrough */
> +		/* fall through - Missed the backend's CLOSING state. */
>  	case XenbusStateClosing:
>  		dev_warn(&xdev->dev, "backend going away!\n");
>  		pcifront_try_disconnect(pdev);
> -- 
> 2.21.0
> 

Re: [Xen-devel] [PATCH] PCI: Mark expected switch fall-throughs

[Andrew Cooper]

On 20/03/2019 18:27, Gustavo A. R. Silva wrote:
> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
> index 6fa1627ce08d..445b51db75b0 100644
> --- a/drivers/pci/proc.c
> +++ b/drivers/pci/proc.c
> @@ -222,6 +222,7 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
>  		}
>  		/* If arch decided it can't, fall through... */
>  #endif /* HAVE_PCI_MMAP */
> +		/* fall through */

Surely it would be better to transpose the #endif and its previous line,
than to add a second fallthrough ?

~Andrew

Re: [PATCH] PCI: Mark expected switch fall-throughs

[Gustavo A. R. Silva]

On 3/20/19 2:24 PM, Bjorn Helgaas wrote:
> On Wed, Mar 20, 2019 at 01:27:15PM -0500, Gustavo A. R. Silva wrote:
>> In preparation to enabling -Wimplicit-fallthrough, mark switch
>> cases where we are expecting to fall through.
> 
> Does this fix all the remaining cases in drivers/pci?  I'd like to fix
> them all at once.
> 

These are actually the last ones.

> What's the best way to watch for new warnings being added?  I fiddled
> with "make W=2" etc but it didn't seem useful.  Does the 0-day robot
> warn when -Wimplicit-fallthrough warnings are added?
> 

0-day robot doesn't know about these warnings yet.  But it certainly
will once we finally enable -Wimplicit-fallthrough.

Add this to your Makefile to see the warnings:

KBUILD_CFLAGS  += $(call cc-option,-Wimplicit-fallthrough=3,)

Thanks
--
Gustavo

> Bjorn
> 
>> This patch fixes the following warnings:
>>
>> drivers/pci/proc.c: In function ‘proc_bus_pci_ioctl’:
>> drivers/pci/proc.c:216:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
>>    if (arch_can_pci_mmap_wc()) {
>>       ^
>> drivers/pci/proc.c:225:2: note: here
>>   default:
>>   ^~~~~~~
>>
>> drivers/pci/xen-pcifront.c: In function ‘pcifront_backend_changed’:
>> drivers/pci/xen-pcifront.c:1105:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
>>    if (xdev->state == XenbusStateClosed)
>>       ^
>> drivers/pci/xen-pcifront.c:1108:2: note: here
>>   case XenbusStateClosing:
>>   ^~~~
>>
>> Notice that, in this particular case, the /* fall through */
>> comment is placed at the very bottom of the case statement,
>> which is what GCC is expecting to find.
>>
>> Warning level 3 was used: -Wimplicit-fallthrough=3
>>
>> This patch is part of the ongoing efforts to enable
>> -Wimplicit-fallthrough.
>>
>> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
>> ---
>>  drivers/pci/proc.c         | 1 +
>>  drivers/pci/xen-pcifront.c | 2 +-
>>  2 files changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
>> index 6fa1627ce08d..445b51db75b0 100644
>> --- a/drivers/pci/proc.c
>> +++ b/drivers/pci/proc.c
>> @@ -222,6 +222,7 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
>>  		}
>>  		/* If arch decided it can't, fall through... */
>>  #endif /* HAVE_PCI_MMAP */
>> +		/* fall through */
>>  	default:
>>  		ret = -EINVAL;
>>  		break;
>> diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
>> index eba6e33147a2..14cf0f41ecf0 100644
>> --- a/drivers/pci/xen-pcifront.c
>> +++ b/drivers/pci/xen-pcifront.c
>> @@ -1104,7 +1104,7 @@ static void __ref pcifront_backend_changed(struct xenbus_device *xdev,
>>  	case XenbusStateClosed:
>>  		if (xdev->state == XenbusStateClosed)
>>  			break;
>> -		/* Missed the backend's CLOSING state -- fallthrough */
>> +		/* fall through - Missed the backend's CLOSING state. */
>>  	case XenbusStateClosing:
>>  		dev_warn(&xdev->dev, "backend going away!\n");
>>  		pcifront_try_disconnect(pdev);
>> -- 
>> 2.21.0
>>

Re: [Xen-devel] [PATCH] PCI: Mark expected switch fall-throughs

[Gustavo A. R. Silva]

On 3/20/19 2:27 PM, Andrew Cooper wrote:
> On 20/03/2019 18:27, Gustavo A. R. Silva wrote:
>> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
>> index 6fa1627ce08d..445b51db75b0 100644
>> --- a/drivers/pci/proc.c
>> +++ b/drivers/pci/proc.c
>> @@ -222,6 +222,7 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
>>  		}
>>  		/* If arch decided it can't, fall through... */
>>  #endif /* HAVE_PCI_MMAP */
>> +		/* fall through */
> 
> Surely it would be better to transpose the #endif and its previous line,
> than to add a second fallthrough ?
> 

I agree.  The thing is that, currently, GCC is expecting to find the
fall-through "annotations" at the very bottom of the case statement,
as I mentioned it in the changelog text.

That's the reason why I decided to left in place the original comment.

Thanks
--
Gustavo

Re: [PATCH] PCI: Mark expected switch fall-throughs

[Bjorn Helgaas]

On Wed, Mar 20, 2019 at 01:27:15PM -0500, Gustavo A. R. Silva wrote:
> In preparation to enabling -Wimplicit-fallthrough, mark switch
> cases where we are expecting to fall through.
> 
> This patch fixes the following warnings:
> 
> drivers/pci/proc.c: In function ‘proc_bus_pci_ioctl’:
> drivers/pci/proc.c:216:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
>    if (arch_can_pci_mmap_wc()) {
>       ^
> drivers/pci/proc.c:225:2: note: here
>   default:
>   ^~~~~~~
> 
> drivers/pci/xen-pcifront.c: In function ‘pcifront_backend_changed’:
> drivers/pci/xen-pcifront.c:1105:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
>    if (xdev->state == XenbusStateClosed)
>       ^
> drivers/pci/xen-pcifront.c:1108:2: note: here
>   case XenbusStateClosing:
>   ^~~~
> 
> Notice that, in this particular case, the /* fall through */
> comment is placed at the very bottom of the case statement,
> which is what GCC is expecting to find.
> 
> Warning level 3 was used: -Wimplicit-fallthrough=3
> 
> This patch is part of the ongoing efforts to enable
> -Wimplicit-fallthrough.
> 
> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>

Applied to pci/misc for v5.2, thanks!

> ---
>  drivers/pci/proc.c         | 1 +
>  drivers/pci/xen-pcifront.c | 2 +-
>  2 files changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/proc.c b/drivers/pci/proc.c
> index 6fa1627ce08d..445b51db75b0 100644
> --- a/drivers/pci/proc.c
> +++ b/drivers/pci/proc.c
> @@ -222,6 +222,7 @@ static long proc_bus_pci_ioctl(struct file *file, unsigned int cmd,
>  		}
>  		/* If arch decided it can't, fall through... */
>  #endif /* HAVE_PCI_MMAP */
> +		/* fall through */
>  	default:
>  		ret = -EINVAL;
>  		break;
> diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
> index eba6e33147a2..14cf0f41ecf0 100644
> --- a/drivers/pci/xen-pcifront.c
> +++ b/drivers/pci/xen-pcifront.c
> @@ -1104,7 +1104,7 @@ static void __ref pcifront_backend_changed(struct xenbus_device *xdev,
>  	case XenbusStateClosed:
>  		if (xdev->state == XenbusStateClosed)
>  			break;
> -		/* Missed the backend's CLOSING state -- fallthrough */
> +		/* fall through - Missed the backend's CLOSING state. */
>  	case XenbusStateClosing:
>  		dev_warn(&xdev->dev, "backend going away!\n");
>  		pcifront_try_disconnect(pdev);
> -- 
> 2.21.0
> 

Re: [PATCH] PCI: Mark expected switch fall-throughs

[Gustavo A. R. Silva]

On 3/20/19 3:13 PM, Bjorn Helgaas wrote:
> On Wed, Mar 20, 2019 at 01:27:15PM -0500, Gustavo A. R. Silva wrote:

[..]

>>
>> Warning level 3 was used: -Wimplicit-fallthrough=3
>>
>> This patch is part of the ongoing efforts to enable
>> -Wimplicit-fallthrough.
>>
>> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
> 
> Applied to pci/misc for v5.2, thanks!
> 

Awesome!

Thanks
--
Gustavo

Re: [Xen-devel] [PATCH] PCI: Mark expected switch fall-throughs

[Jan Beulich]

>>> On 20.03.19 at 19:27, <gustavo@embeddedor.com> wrote:
> In preparation to enabling -Wimplicit-fallthrough, mark switch
> cases where we are expecting to fall through.
> 
> This patch fixes the following warnings:
> 
> drivers/pci/proc.c: In function ‘proc_bus_pci_ioctl’:
> drivers/pci/proc.c:216:6: warning: this statement may fall through 
> [-Wimplicit-fallthrough=]
>    if (arch_can_pci_mmap_wc()) {
>       ^
> drivers/pci/proc.c:225:2: note: here
>   default:
>   ^~~~~~~
> 
> drivers/pci/xen-pcifront.c: In function ‘pcifront_backend_changed’:
> drivers/pci/xen-pcifront.c:1105:6: warning: this statement may fall through 
> [-Wimplicit-fallthrough=]
>    if (xdev->state == XenbusStateClosed)
>       ^
> drivers/pci/xen-pcifront.c:1108:2: note: here
>   case XenbusStateClosing:
>   ^~~~
> 
> Notice that, in this particular case, the /* fall through */
> comment is placed at the very bottom of the case statement,
> which is what GCC is expecting to find.
> 
> Warning level 3 was used: -Wimplicit-fallthrough=3
> 
> This patch is part of the ongoing efforts to enable
> -Wimplicit-fallthrough.

None of the above explains why ...

> --- a/drivers/pci/xen-pcifront.c
> +++ b/drivers/pci/xen-pcifront.c
> @@ -1104,7 +1104,7 @@ static void __ref pcifront_backend_changed(struct xenbus_device *xdev,
>  	case XenbusStateClosed:
>  		if (xdev->state == XenbusStateClosed)
>  			break;
> -		/* Missed the backend's CLOSING state -- fallthrough */
> +		/* fall through - Missed the backend's CLOSING state. */

... the original comment here wasn't good enough.

Jan