Skip to contentSkip to conversation

x86: address MISRA C:2012 Rule 5.3

PATCH Superseded Nicola Vetrini 2 participants 11 messages Jul 31, 2023

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover x86: address MISRA C:2012 Rule 5.3

This series aims to address some violations of MISRA C:2012 Rule 5.3:
"An identifier declared in an inner scope shall not hide an identifier declared in an outer scope".

To resolve such violations two strategies have been applied:

1. rename several function parameters or local variables;
2. remove a local variable declaration in an inner scope, if using the outer variable does not change the semantics of the function.

Nicola Vetrini (4):
x86: mechanically rename to address MISRA C:2012 Rule 5.3 x86/mm: remove variable to avoid shadowing
x86emul: avoid shadowing to address MISRA C:2012 Rule 5.3
x86emul: remove local variable to address MISRA C:2012 Rule 5.3

 xen/arch/x86/cpu/mcheck/mce.c          |  8 ++++----
 xen/arch/x86/domain.c                  |  2 +-
 xen/arch/x86/emul-i8254.c              |  2 +-
 xen/arch/x86/include/asm/e820.h        |  2 +-
 xen/arch/x86/include/asm/hvm/save.h    |  8 ++++----
 xen/arch/x86/include/asm/hvm/vpt.h     |  2 +-
 xen/arch/x86/mm/p2m.c                  |  2 --
 xen/arch/x86/psr.c                     |  4 ++--
 xen/arch/x86/traps.c                   |  5 ++---
 xen/arch/x86/x86_emulate/x86_emulate.c | 14 ++++++--------
 10 files changed, 22 insertions(+), 27 deletions(-)

--
2.34.1

#1 lore raw

PATCH v2 1/4 x86: mechanically rename to address MISRA C:2012 Rule 5.3

Rule 5.3 has the following headline:
"An identifier declared in an inner scope shall not hide an identifier declared in an outer scope"

The renames done by this patch avoid shadowing from happening. They are as follows:
- s/i/j/ in 'mce_action' (the shadowing here is due to macro 'x86_mcinfo_lookup' that defines 'i' as a loop counter)
- s/desc/descriptor/ in '_hvm_load_entry'
- s/socket_info/sock_info/ in 'do_write_psr_msrs'

The parameter 'cpu_khz' that causes a violation in 'pit_init' is unused, and hence can be removed.

Parameter 'debug_stack_lines' in 'compat_show_guest_stack' is removed, since the shadowed static variable has the same purpose.

The declaration of 'init_e820' is renamed to be consistent with its definition.

No functional changes.

---
Changes in v2:
- removed renames for 'str' parameters
- removed parameter 'debug_stack_lines' from
'compat_show_guest_stack'

---
 xen/arch/x86/cpu/mcheck/mce.c       | 8 ++++----
 xen/arch/x86/domain.c               | 2 +-
 xen/arch/x86/emul-i8254.c           | 2 +-
 xen/arch/x86/include/asm/e820.h     | 2 +-
 xen/arch/x86/include/asm/hvm/save.h | 8 ++++----
 xen/arch/x86/include/asm/hvm/vpt.h  | 2 +-
 xen/arch/x86/psr.c                  | 4 ++--
 xen/arch/x86/traps.c                | 5 ++---
 8 files changed, 16 insertions(+), 17 deletions(-)
xen/arch/x86/cpu/mcheck/mce.c +4 -4
index c2173cfc78..2dbed604b1 100644
@@ -1746,7 +1746,7 @@ static enum mce_result mce_action(const struct cpu_user_regs *regs,
     struct mcinfo_common *mic = NULL;
     struct mca_binfo binfo;
     const struct mca_error_handler *handlers = mce_dhandlers;
-    unsigned int i, handler_num = mce_dhandler_num;
+    unsigned int j, handler_num = mce_dhandler_num;
 
     /* When in mce context, regs is valid */
     if ( regs )
@@ -1780,11 +1780,11 @@ static enum mce_result mce_action(const struct cpu_user_regs *regs,
         binfo.mib = (struct mcinfo_bank *)mic;
         binfo.bank = binfo.mib->mc_bank;
         bank_result = MCER_NOERROR;
-        for ( i = 0; i < handler_num; i++ )
+        for ( j = 0; j < handler_num; j++ )
         {
-            if ( handlers[i].owned_error(binfo.mib->mc_status) )
+            if ( handlers[j].owned_error(binfo.mib->mc_status) )
             {
-                handlers[i].recovery_handler(&binfo, &bank_result, regs);
+                handlers[j].recovery_handler(&binfo, &bank_result, regs);
                 if ( worst_result < bank_result )
                     worst_result = bank_result;
                 break;
xen/arch/x86/domain.c +1 -1
index 5f66c2ae33..fe86a7f853 100644
@@ -858,7 +858,7 @@ int arch_domain_create(struct domain *d,
     }
 
     /* PV/PVH guests get an emulated PIT too for video BIOSes to use. */
-    pit_init(d, cpu_khz);
+    pit_init(d);
 
     /*
      * If the FPU does not save FCS/FDS then we can always
xen/arch/x86/emul-i8254.c +1 -1
index a81232fc55..41ec4a1ef1 100644
@@ -495,7 +495,7 @@ void pit_reset(struct domain *d)
     spin_unlock(&pit->lock);
 }
 
-void pit_init(struct domain *d, unsigned long cpu_khz)
+void pit_init(struct domain *d)
 {
     PITState *pit = domain_vpit(d);
 
xen/arch/x86/include/asm/e820.h +1 -1
index 92f5efa4f5..213d5b5dd2 100644
@@ -31,7 +31,7 @@ extern int e820_change_range_type(
     uint32_t orig_type, uint32_t new_type);
 extern int e820_add_range(
     struct e820map *, uint64_t s, uint64_t e, uint32_t type);
-extern unsigned long init_e820(const char *, struct e820map *);
+extern unsigned long init_e820(const char *str, struct e820map *raw);
 extern void print_e820_memory_map(const struct e820entry *map,
     unsigned int entries);
 extern struct e820map e820;
xen/arch/x86/include/asm/hvm/save.h +4 -4
index 535cf68ed2..4d0a062b2c 100644
@@ -49,21 +49,21 @@ void _hvm_read_entry(struct hvm_domain_context *h,
  */
 #define _hvm_load_entry(_x, _h, _dst, _strict) ({                       \
     int r;                                                              \
-    struct hvm_save_descriptor *desc                                    \
+    struct hvm_save_descriptor *descriptor                              \
         = (struct hvm_save_descriptor *)&(_h)->data[(_h)->cur];         \
     if ( (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),                 \
                HVM_SAVE_LENGTH(_x), (_strict))) == 0 )                  \
     {                                                                   \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH(_x));             \
         if ( HVM_SAVE_HAS_COMPAT(_x) &&                                 \
-             desc->length != HVM_SAVE_LENGTH(_x) )                      \
-            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);          \
+             descriptor->length != HVM_SAVE_LENGTH(_x) )                \
+            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);    \
     }                                                                   \
     else if (HVM_SAVE_HAS_COMPAT(_x)                                    \
              && (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),          \
                        HVM_SAVE_LENGTH_COMPAT(_x), (_strict))) == 0 ) { \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH_COMPAT(_x));      \
-        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);              \
+        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);        \
     }                                                                   \
     r; })
 
xen/arch/x86/include/asm/hvm/vpt.h +1 -1
index 2af76ca8dc..feb0bf43f1 100644
@@ -179,7 +179,7 @@ void destroy_periodic_time(struct periodic_time *pt);
 int pv_pit_handler(int port, int data, int write);
 void pit_reset(struct domain *d);
 
-void pit_init(struct domain *d, unsigned long cpu_khz);
+void pit_init(struct domain *d);
 void pit_stop_channel0_irq(PITState * pit);
 void pit_deinit(struct domain *d);
 void rtc_init(struct domain *d);
xen/arch/x86/psr.c +2 -2
index a1e0af27c5..90bcb25e60 100644
@@ -1252,7 +1252,7 @@ static void cf_check do_write_psr_msrs(void *data)
 {
     const struct cos_write_info *info = data;
     unsigned int i, index, cos = info->cos;
-    const struct psr_socket_info *socket_info =
+    const struct psr_socket_info *sock_info =
         get_socket_info(cpu_to_socket(smp_processor_id()));
 
     /*
@@ -1261,7 +1261,7 @@ static void cf_check do_write_psr_msrs(void *data)
      */
     for ( index = i = 0; i < ARRAY_SIZE(feat_props); i++ )
     {
-        struct feat_node *feat = socket_info->features[i];
+        struct feat_node *feat = sock_info->features[i];
         const struct feat_props *props = feat_props[i];
         unsigned int cos_num, j;
 
xen/arch/x86/traps.c +2 -3
index 4229bda159..f11239fd0f 100644
@@ -195,8 +195,7 @@ void show_code(const struct cpu_user_regs *regs)
 }
 
 static void compat_show_guest_stack(struct vcpu *v,
-                                    const struct cpu_user_regs *regs,
-                                    int debug_stack_lines)
+                                    const struct cpu_user_regs *regs)
 {
     unsigned int i, *stack, addr, mask = STACK_SIZE;
     void *stack_page = NULL;
@@ -273,7 +272,7 @@ static void show_guest_stack(struct vcpu *v, const struct cpu_user_regs *regs)
 
     if ( is_pv_32bit_vcpu(v) )
     {
-        compat_show_guest_stack(v, regs, debug_stack_lines);
+        compat_show_guest_stack(v, regs);
         return;
     }
 

--
2.34.1

#2 lore raw

Re: [XEN PATCH v2 1/4] x86: mechanically rename to address MISRA C:2012 Rule 5.3

Quoted 8 lines · Nicola Vetrini

Rule 5.3 has the following headline:
"An identifier declared in an inner scope shall not hide an identifier declared in an outer scope"

The renames done by this patch avoid shadowing from happening. They are as follows:
- s/i/j/ in 'mce_action' (the shadowing here is due to macro 'x86_mcinfo_lookup' that defines 'i' as a loop counter)

Then that macro needs adjustment. Especially considering it's in a public header: It's full of name space violations.

Quoted 17 lines
xen/arch/x86/include/asm/hvm/save.h +2 -3
@@ -49,21 +49,21 @@ void _hvm_read_entry(struct hvm_domain_context *h,
  */
 #define _hvm_load_entry(_x, _h, _dst, _strict) ({                       \
     int r;                                                              \
-    struct hvm_save_descriptor *desc                                    \
+    struct hvm_save_descriptor *descriptor                              \
         = (struct hvm_save_descriptor *)&(_h)->data[(_h)->cur];         \
     if ( (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),                 \
                HVM_SAVE_LENGTH(_x), (_strict))) == 0 )                  \
     {                                                                   \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH(_x));             \
         if ( HVM_SAVE_HAS_COMPAT(_x) &&                                 \
-             desc->length != HVM_SAVE_LENGTH(_x) )                      \
-            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);          \
+             descriptor->length != HVM_SAVE_LENGTH(_x) )                \

xen/arch/x86/include/asm/hvm/save.h void _hvm_read_entry(struct hvm_domain_context *h, · line 68

Quoted 10 lines
xen/arch/x86/include/asm/hvm/save.h +2 -1
+            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);    \
     }                                                                   \
     else if (HVM_SAVE_HAS_COMPAT(_x)                                    \
              && (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),          \
                        HVM_SAVE_LENGTH_COMPAT(_x), (_strict))) == 0 ) { \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH_COMPAT(_x));      \
-        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);              \
+        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);        \
     }                                                                   \
     r; })

The macro-local variable gets too long for my taste, to be honest, and it being improperly named anyway suggests it simply wants a trailing underscore added. And then, since for a variable named "r" the risk of shadowing is equally high, that one wants to gain a trailing underscore as well imo. (And while at it, I personally would also drop the leading underscores from the macro parameter names. Furthermore I think it would be nice if at on the lines touched anyway indentation was also corrected. Overall maybe best if I submit a patch.)

Jan

#3 lore raw

Re: [XEN PATCH v2 1/4] x86: mechanically rename to address MISRA C:2012 Rule 5.3

Quoted 30 lines · Jan Beulich
Quoted 17 lines · Nicola Vetrini
xen/arch/x86/include/asm/hvm/save.h +2 -3
@@ -49,21 +49,21 @@ void _hvm_read_entry(struct hvm_domain_context *h,
  */
 #define _hvm_load_entry(_x, _h, _dst, _strict) ({                       \
     int r;                                                              \
-    struct hvm_save_descriptor *desc                                    \
+    struct hvm_save_descriptor *descriptor                              \
         = (struct hvm_save_descriptor *)&(_h)->data[(_h)->cur];         \
     if ( (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),                 \
                HVM_SAVE_LENGTH(_x), (_strict))) == 0 )                  \
     {                                                                   \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH(_x));             \
         if ( HVM_SAVE_HAS_COMPAT(_x) &&                                 \
-             desc->length != HVM_SAVE_LENGTH(_x) )                      \
-            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);          \
+             descriptor->length != HVM_SAVE_LENGTH(_x) )                \

xen/arch/x86/include/asm/hvm/save.h void _hvm_read_entry(struct hvm_domain_context *h, · line 68

Quoted 10 lines
xen/arch/x86/include/asm/hvm/save.h +2 -1
+            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);    \
     }                                                                   \
     else if (HVM_SAVE_HAS_COMPAT(_x)                                    \
              && (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),          \
                        HVM_SAVE_LENGTH_COMPAT(_x), (_strict))) == 0 ) { \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH_COMPAT(_x));      \
-        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);              \
+        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);        \
     }                                                                   \
     r; })

The macro-local variable gets too long for my taste, to be honest,

Quoted 8 lines

and it being improperly named anyway suggests it simply wants a trailing underscore added. And then, since for a variable named "r" the risk of shadowing is equally high, that one wants to gain a trailing underscore as well imo. (And while at it, I personally would also drop the leading underscores from the macro parameter names. Furthermore I think it would be nice if at on the lines touched anyway indentation was also corrected. Overall maybe best if I submit a patch.)

In that replacement patch I would like to mention what "desc" this collides with, but your description didn't say so and I'm afraid I also haven't been able to spot it (grep-ing for "desc", even with a couple of extra restrictions, still yields way too many hits).

Jan

#4 lore raw

Re: [XEN PATCH v2 1/4] x86: mechanically rename to address MISRA C:2012 Rule 5.3

Quoted 63 lines · Jan Beulich
Quoted 53 lines · Jan Beulich
Quoted 43 lines · Nicola Vetrini
xen/arch/x86/include/asm/hvm/save.h +3 -3
@@ -49,21 +49,21 @@ void _hvm_read_entry(struct hvm_domain_context 
 *h,
  */
 #define _hvm_load_entry(_x, _h, _dst, _strict) ({                    
   \
     int r;                                                           
   \
-    struct hvm_save_descriptor *desc                                 
   \
+    struct hvm_save_descriptor *descriptor                           
   \
         = (struct hvm_save_descriptor *)&(_h)->data[(_h)->cur];      
   \
     if ( (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),              
   \
                HVM_SAVE_LENGTH(_x), (_strict))) == 0 )               
   \
     {                                                                
   \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH(_x));          
   \
         if ( HVM_SAVE_HAS_COMPAT(_x) &&                              
   \
-             desc->length != HVM_SAVE_LENGTH(_x) )                   
   \
-            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);       
   \
+             descriptor->length != HVM_SAVE_LENGTH(_x) )             
   \
+            r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length); 
   \
     }                                                                
   \
     else if (HVM_SAVE_HAS_COMPAT(_x)                                 
   \
              && (r = _hvm_check_entry((_h), HVM_SAVE_CODE(_x),       
   \
                        HVM_SAVE_LENGTH_COMPAT(_x), (_strict))) == 0 
) { \
         _hvm_read_entry((_h), (_dst), HVM_SAVE_LENGTH_COMPAT(_x));   

\

xen/arch/x86/include/asm/hvm/save.h

Quoted 7 lines
xen/arch/x86/include/asm/hvm/save.h +1 -1
-        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), desc->length);           
   \
+        r = HVM_SAVE_FIX_COMPAT(_x, (_dst), descriptor->length);     
   \
     }                                                                
   \
     r; })

The macro-local variable gets too long for my taste, to be honest,

Quoted 8 lines

and it being improperly named anyway suggests it simply wants a trailing underscore added. And then, since for a variable named "r" the risk of shadowing is equally high, that one wants to gain a trailing underscore as well imo. (And while at it, I personally would also drop the leading underscores from the macro parameter names. Furthermore I think it would be nice if at on the lines touched anyway indentation was also corrected. Overall maybe best if I submit a patch.)

Quoted 6 lines

In that replacement patch I would like to mention what "desc" this collides with, but your description didn't say so and I'm afraid I also haven't been able to spot it (grep-ing for "desc", even with a couple of extra restrictions, still yields way too many hits).

Jan

It's the local variable in 'xen/arch/x86/hvm/save.c:281', but since the macro is used elsewhere quite a few times, doing a suitable rename inside it (that does not shadow anything) prevents other users from possibly shadowing it in the future. Your points are all agreeable in my opinion. Since your concerns with this macro are also about issues other than MISRA violations, I agree that this patch should be dropped.

Signature · 2 lines
Nicola Vetrini, BSc
Software Engineer, BUGSENG srl (https://bugseng.com)
#5 lore raw

PATCH v2 2/4 x86/mm: remove variable to avoid shadowing

In reply to Nicola Vetrini

The local variable 'p2mt' shadows a declaration of the same variable in the enclosing scope, but removing the inner declaration does not alter the semantics ('p2mt' is an output for the get_gfn call later on) and this resolves a violation of
MISRA C:2012 Rule 5.3.

Both variables were last introduced by commit
d4f699a0df6cea907c1de5c277500b98c0729685.

No functional changes.

---
Change in v2:
- removed the innermost p2mt instead of renaming it.

---
 xen/arch/x86/mm/p2m.c | 2 --
 1 file changed, 2 deletions(-)
xen/arch/x86/mm/p2m.c +0 -2
index 714358f953..f6df35767a 100644
@@ -2439,8 +2439,6 @@ int xenmem_add_to_physmap_one(
 
     case XENMAPSPACE_gmfn:
     {
-        p2m_type_t p2mt;
-
         gfn = idx;
         mfn = get_gfn_unshare(d, gfn, &p2mt);
         /* If the page is still shared, exit early */

--
2.34.1

Reviewed-by on v2 2/4
#6 lore raw

Re: [XEN PATCH v2 2/4] x86/mm: remove variable to avoid shadowing

Quoted 8 lines · Nicola Vetrini

The local variable 'p2mt' shadows a declaration of the same variable in the enclosing scope, but removing the inner declaration does not alter the semantics ('p2mt' is an output for the get_gfn call later on) and this resolves a violation of
MISRA C:2012 Rule 5.3.

Both variables were last introduced by commit
d4f699a0df6cea907c1de5c277500b98c0729685.

No, that commit moved the function (without correcting the issue, perhaps because of not noticing it, but perhaps also intentionally).

Quoted 3 lines

No functional changes.

With the misleading part of the description either dropped (which I'd be okay doing while committing) or made meaningful:

Jan

#7 lore raw

PATCH v2 3/4 x86emul: avoid shadowing to address MISRA C:2012 Rule 5.3

In reply to Nicola Vetrini

Rule 5.3 has the following headline:
"An identifier declared in an inner scope shall not hide an identifier declared in an outer scope"

The declaration of local variable 'cpuid_leaf' causes shadowing with the homonymous function to happen, therefore the variable is renamed to avoid this.

No functional changes.

---
Changes in v2:
- s/res/leaf/ to have a less generic name.
- Split the patch in half (see patch 4/4).

---
 xen/arch/x86/x86_emulate/x86_emulate.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
xen/arch/x86/x86_emulate/x86_emulate.c +6 -6
index 2de1be7996..f811c1c76b 100644
@@ -1483,7 +1483,7 @@ x86_emulate(
     {
         enum x86_segment seg;
         struct segment_register cs, sreg;
-        struct cpuid_leaf cpuid_leaf;
+        struct cpuid_leaf leaf;
         uint64_t msr_val;
         unsigned int i, n;
         unsigned long dummy;
@@ -5024,13 +5024,13 @@ x86_emulate(
         generate_exception_if((msr_val & MSR_MISC_FEATURES_CPUID_FAULTING),
                               X86_EXC_GP, 0); /* Faulting active? (Inc. CPL test) */
 
-        rc = ops->cpuid(_regs.eax, _regs.ecx, &cpuid_leaf, ctxt);
+        rc = ops->cpuid(_regs.eax, _regs.ecx, &leaf, ctxt);
         if ( rc != X86EMUL_OKAY )
             goto done;
-        _regs.r(ax) = cpuid_leaf.a;
-        _regs.r(bx) = cpuid_leaf.b;
-        _regs.r(cx) = cpuid_leaf.c;
-        _regs.r(dx) = cpuid_leaf.d;
+        _regs.r(ax) = leaf.a;
+        _regs.r(bx) = leaf.b;
+        _regs.r(cx) = leaf.c;
+        _regs.r(dx) = leaf.d;
         break;
 
     case X86EMUL_OPC(0x0f, 0xa3): bt: /* bt */

--
2.34.1

Acked-by on v2 3/4
#8 lore raw

Re: [XEN PATCH v2 3/4] x86emul: avoid shadowing to address MISRA C:2012 Rule 5.3

Quoted 11 lines · Nicola Vetrini

Rule 5.3 has the following headline:
"An identifier declared in an inner scope shall not hide an identifier declared in an outer scope"

The declaration of local variable 'cpuid_leaf' causes shadowing with the homonymous function to happen, therefore the variable is renamed to avoid this.

No functional changes.

#9 lore raw

PATCH v2 4/4 x86emul: remove local variable to address MISRA C:2012 Rule 5.3

In reply to Nicola Vetrini

The local variable removed by this commit shadowed another declared in an outer scope and thus violates Rule 5.3. It was introduced as part of a refactoring by commit abf5fb62f73d0dbeabc6a8cd6dbb55174b36c8d1.

No functional changes.

---
 xen/arch/x86/x86_emulate/x86_emulate.c | 2 --
 1 file changed, 2 deletions(-)
xen/arch/x86/x86_emulate/x86_emulate.c +0 -2
index f811c1c76b..60bd2f8d8e 100644
@@ -8408,8 +8408,6 @@ x86_emulate(
         generate_exception(X86_EXC_MF);
     if ( stub_exn.info.fields.trapnr == X86_EXC_XM )
     {
-        unsigned long cr4;
-
         if ( !ops->read_cr || ops->read_cr(4, &cr4, ctxt) != X86EMUL_OKAY )
             cr4 = X86_CR4_OSXMMEXCPT;
         generate_exception(cr4 & X86_CR4_OSXMMEXCPT ? X86_EXC_XM : X86_EXC_UD);
Signature · 1 line
2.34.1
Reviewed-by on v2 4/4
#10 lore raw

Re: [XEN PATCH v2 4/4] x86emul: remove local variable to address MISRA C:2012 Rule 5.3

Quoted 7 lines · Nicola Vetrini

The local variable removed by this commit shadowed another declared in an outer scope and thus violates Rule 5.3. It was introduced as part of a refactoring by commit abf5fb62f73d0dbeabc6a8cd6dbb55174b36c8d1.

No functional changes.

Albeit, just to mention it, the description could have done with making clear why the removal is actually correct (nothing like this can be inferred from pointing at the variable's introduction). But well ...

Jan