[PATCH] x86/gen-cpuid: avoid violations of Misra rule 1.3

Jan Beulich posted 1 patch 9 months ago
Failed in applying to current master (apply log)
[PATCH] x86/gen-cpuid: avoid violations of Misra rule 1.3
Posted by Jan Beulich 9 months ago
Structures or unions without any named members aren't liked by Misra
(nor the C standard). Avoid emitting such for leaves without any known
bits.

At this occasion also add the script to the X86 section in ./MAINTAINERS.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -601,6 +601,7 @@ F:	xen/arch/x86/
 F:	xen/include/public/arch-x86/
 F:	xen/include/xen/lib/x86
 F:	xen/lib/x86
+F:	xen/tools/gen-cpuid.py
 F:	tools/firmware/hvmloader/
 F:	tools/firmware/rombios/
 F:	tools/firmware/vgabios/
--- a/xen/tools/gen-cpuid.py
+++ b/xen/tools/gen-cpuid.py
@@ -367,6 +367,7 @@ def crunch_numbers(state):
     for word in range(state.nr_entries):
 
         names = []
+        empty = 1
         for bit in range(32):
 
             name = state.names.get(word * 32 + bit, "")
@@ -380,9 +381,15 @@ def crunch_numbers(state):
             if name in ("APIC", "OSXSAVE", "OSPKE"):
                 name = ""
 
+            if name:
+                empty = 0
+
             names.append(name.lower())
 
-        state.bitfields.append("bool " + ":1, ".join(names) + ":1")
+        if empty:
+            state.bitfields.append("unsigned int empty_" + str(word) + ":32")
+        else:
+            state.bitfields.append("bool " + ":1, ".join(names) + ":1")
 
 
 def write_results(state):
Re: [PATCH] x86/gen-cpuid: avoid violations of Misra rule 1.3
Posted by Andrew Cooper 9 months ago
On 02/08/2023 10:43 am, Jan Beulich wrote:
> Structures or unions without any named members aren't liked by Misra
> (nor the C standard). Avoid emitting such for leaves without any known
> bits.
>
> At this occasion also add the script to the X86 section in ./MAINTAINERS.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -601,6 +601,7 @@ F:	xen/arch/x86/
>  F:	xen/include/public/arch-x86/
>  F:	xen/include/xen/lib/x86
>  F:	xen/lib/x86
> +F:	xen/tools/gen-cpuid.py
>  F:	tools/firmware/hvmloader/
>  F:	tools/firmware/rombios/
>  F:	tools/firmware/vgabios/
> --- a/xen/tools/gen-cpuid.py
> +++ b/xen/tools/gen-cpuid.py
> @@ -367,6 +367,7 @@ def crunch_numbers(state):
>      for word in range(state.nr_entries):
>  
>          names = []
> +        empty = 1
>          for bit in range(32):
>  
>              name = state.names.get(word * 32 + bit, "")
> @@ -380,9 +381,15 @@ def crunch_numbers(state):
>              if name in ("APIC", "OSXSAVE", "OSPKE"):
>                  name = ""
>  
> +            if name:
> +                empty = 0
> +
>              names.append(name.lower())
>  
> -        state.bitfields.append("bool " + ":1, ".join(names) + ":1")
> +        if empty:
> +            state.bitfields.append("unsigned int empty_" + str(word) + ":32")
> +        else:
> +            state.bitfields.append("bool " + ":1, ".join(names) + ":1")

There's no need for the extra empty variable (this is python, not C).

if any(names):
    state.bitfields.append("bool " + ":1, ".join(names) + ":1")
else:
    state.bitfields.append("uint32_t empty_%s" % (word, ))

But there's a related bug later in the script which MISRA won't notice
unless it happens to run on the cset boundary introducing a new leaf.

I'll see about doing a patch to address both.

~Andrew