All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] xen/x86: Annotate deliberate fallthrough cases from XSA-154
@ 2016-02-18 12:26 Andrew Cooper
  2016-02-18 13:38 ` Jan Beulich
  0 siblings, 1 reply; 3+ messages in thread
From: Andrew Cooper @ 2016-02-18 12:26 UTC (permalink / raw)
  To: Xen-devel; +Cc: Andrew Cooper, Jan Beulich

Coverity objects otherwise.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
---
 xen/arch/x86/mm.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index a05edc3..0bff7dd 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -924,10 +924,15 @@ get_page_from_l1e(
             {
             case 0:
                 break;
+
             case 1:
                 if ( is_hardware_domain(l1e_owner) )
+                {
+                    /* Fallthrough. */
             case -1:
                     return 0;
+                }
+                /* Fallthrough. */
             default:
                 ASSERT_UNREACHABLE();
             }
-- 
2.1.4

^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] xen/x86: Annotate deliberate fallthrough cases from XSA-154
  2016-02-18 12:26 [PATCH] xen/x86: Annotate deliberate fallthrough cases from XSA-154 Andrew Cooper
@ 2016-02-18 13:38 ` Jan Beulich
  2016-02-18 14:05   ` Andrew Cooper
  0 siblings, 1 reply; 3+ messages in thread
From: Jan Beulich @ 2016-02-18 13:38 UTC (permalink / raw)
  To: Andrew Cooper; +Cc: Xen-devel

>>> On 18.02.16 at 13:26, <andrew.cooper3@citrix.com> wrote:
> Coverity objects otherwise.
> 
> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> ---
> CC: Jan Beulich <JBeulich@suse.com>
> ---
>  xen/arch/x86/mm.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
> index a05edc3..0bff7dd 100644
> --- a/xen/arch/x86/mm.c
> +++ b/xen/arch/x86/mm.c
> @@ -924,10 +924,15 @@ get_page_from_l1e(
>              {
>              case 0:
>                  break;
> +
>              case 1:
>                  if ( is_hardware_domain(l1e_owner) )
> +                {
> +                    /* Fallthrough. */
>              case -1:
>                      return 0;
> +                }
> +                /* Fallthrough. */
>              default:

This second fall-through is actually a bug (luckily noticable only
on debug builds).

I'll commit the patch suitably adjusted, albeit I have a hard time
seeing how

            case 1:
                if ( is_hardware_domain(l1e_owner) )
            case -1:

cannot be seen as obviously deliberate. Or did Coverity perhaps
only complain about the second, indeed buggy one?

Jan

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] xen/x86: Annotate deliberate fallthrough cases from XSA-154
  2016-02-18 13:38 ` Jan Beulich
@ 2016-02-18 14:05   ` Andrew Cooper
  0 siblings, 0 replies; 3+ messages in thread
From: Andrew Cooper @ 2016-02-18 14:05 UTC (permalink / raw)
  To: Jan Beulich; +Cc: Xen-devel

On 18/02/16 13:38, Jan Beulich wrote:
>>>> On 18.02.16 at 13:26, <andrew.cooper3@citrix.com> wrote:
>> Coverity objects otherwise.
>>
>> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
>> ---
>> CC: Jan Beulich <JBeulich@suse.com>
>> ---
>>  xen/arch/x86/mm.c | 5 +++++
>>  1 file changed, 5 insertions(+)
>>
>> diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
>> index a05edc3..0bff7dd 100644
>> --- a/xen/arch/x86/mm.c
>> +++ b/xen/arch/x86/mm.c
>> @@ -924,10 +924,15 @@ get_page_from_l1e(
>>              {
>>              case 0:
>>                  break;
>> +
>>              case 1:
>>                  if ( is_hardware_domain(l1e_owner) )
>> +                {
>> +                    /* Fallthrough. */
>>              case -1:
>>                      return 0;
>> +                }
>> +                /* Fallthrough. */
>>              default:
> This second fall-through is actually a bug (luckily noticable only
> on debug builds).
>
> I'll commit the patch suitably adjusted, albeit I have a hard time
> seeing how
>
>             case 1:
>                 if ( is_hardware_domain(l1e_owner) )
>             case -1:
>
> cannot be seen as obviously deliberate.

Many C programmers are not aware that it is even valid syntax.

The point of the MISSING_BREAK check is to second guess what the
programmer has actually written.

>  Or did Coverity perhaps
> only complain about the second, indeed buggy one?

It only complained about the first.  The second is a misaligned
unconditional return when considered in isolation.

~Andrew

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2016-02-18 14:05 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-02-18 12:26 [PATCH] xen/x86: Annotate deliberate fallthrough cases from XSA-154 Andrew Cooper
2016-02-18 13:38 ` Jan Beulich
2016-02-18 14:05   ` Andrew Cooper

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.