Skip to contentSkip to conversation

address violations of MISRA C rules

PATCH Superseded Nicola Vetrini 4 participants 26 messages Jun 1, 2024 – Jun 20, 2024

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover address violations of MISRA C rules

Patches 1 to 4 address violations of MISRA C Rule 20.12 by deviating certain uses of some macros, while the last patch addresses some regressions introduced by the latest bitops series

Nicola Vetrini (5):
xen/domain: deviate violation of MISRA C Rule 20.12 x86/domain: deviate violation of MISRA C Rule 20.12
x86: deviate violation of MISRA C Rule 20.12 automation/eclair_analysis: address remaining violations of MISRA C

    Rule 20.12

xen: fix MISRA regressions on rule 20.9 and 20.12

 automation/eclair_analysis/ECLAIR/deviations.ecl | 8 +++++++-
 docs/misra/safe.json                             | 8 ++++++++
 xen/arch/x86/domain.c                            | 1 +
 xen/arch/x86/domctl.c                            | 1 +
 xen/arch/x86/include/asm/shared.h                | 1 +
 xen/common/domain.c                              | 1 +
 xen/include/xen/self-tests.h                     | 2 +-
 7 files changed, 20 insertions(+), 2 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH 1/5 xen/domain: deviate violation of MISRA C Rule 20.12

MISRA C Rule 20.12 states: "A macro parameter used as an operand to the # or ## operators, which is itself subject to further macro replacement, shall only be used as an operand to these operators".

In this case, builds where CONFIG_DEBUG_LOCK_PROFILE=y the domain_lock macro is used both as a regular macro argument and as an operand for stringification in the expansion of macro spin_lock_init_prof. A SAF-x-safe deviation is introduced to justify this.

No functional change.

---
 docs/misra/safe.json | 8 ++++++++
 xen/common/domain.c  | 1 +
 2 files changed, 9 insertions(+)
docs/misra/safe.json +8 -0
index 9b13bcf71706..c213e0a0be3b 100644
@@ -52,6 +52,14 @@
         },
         {
             "id": "SAF-6-safe",
+            "analyser": {
+                "eclair": "MC3R1.R20.12"
+            },
+            "name": "MC3R1.R20.12: use of a macro argument that deliberately violates the Rule",
+            "text": "A macro parameter that is itself a macro is intentionally used within the macro both as a regular parameter and for text replacement."
+        },
+        {
+            "id": "SAF-7-safe",
             "analyser": {},
             "name": "Sentinel",
             "text": "Next ID to be used"
xen/common/domain.c +1 -0
index 67cadb7c3f4f..2c7168093734 100644
@@ -632,6 +632,7 @@ struct domain *domain_create(domid_t domid,
 
     atomic_set(&d->refcnt, 1);
     RCU_READ_LOCK_INIT(&d->rcu_lock);
+    /* SAF-6-safe Rule 20.12 expansion of macro domain_lock in debug builds */
     rspin_lock_init_prof(d, domain_lock);
     rspin_lock_init_prof(d, page_alloc_lock);
     spin_lock_init(&d->hypercall_deadlock_mutex);
Signature · 1 line
2.34.1
#2 lore raw

Re: [XEN PATCH 1/5] xen/domain: deviate violation of MISRA C Rule 20.12

Quoted 7 lines · Nicola Vetrini

MISRA C Rule 20.12 states: "A macro parameter used as an operand to the # or ## operators, which is itself subject to further macro replacement, shall only be used as an operand to these operators".

In this case, builds where CONFIG_DEBUG_LOCK_PROFILE=y the domain_lock macro is used both as a regular macro argument and as an operand for stringification in the expansion of macro spin_lock_init_prof.

The shouldn't the marker be on the definition of spin_lock_init_prof(), rather than ...

xen/common/domain.c struct domain *domain_create(domid_t domid, · line 638

Quoted 10 lines
xen/common/domain.c +1 -0
@@ -632,6 +632,7 @@ struct domain *domain_create(domid_t domid,
 
     atomic_set(&d->refcnt, 1);
     RCU_READ_LOCK_INIT(&d->rcu_lock);
+    /* SAF-6-safe Rule 20.12 expansion of macro domain_lock in debug builds */
     rspin_lock_init_prof(d, domain_lock);
     rspin_lock_init_prof(d, page_alloc_lock);
     spin_lock_init(&d->hypercall_deadlock_mutex);

... actually just one of the two uses here (and presumably several more elsewhere)?

Jan

#3 lore raw

Re: [XEN PATCH 1/5] xen/domain: deviate violation of MISRA C Rule 20.12

Quoted 27 lines · Jan Beulich
Quoted 7 lines · Nicola Vetrini

MISRA C Rule 20.12 states: "A macro parameter used as an operand to the # or ## operators, which is itself subject to further macro replacement, shall only be used as an operand to these operators".

In this case, builds where CONFIG_DEBUG_LOCK_PROFILE=y the domain_lock macro is used both as a regular macro argument and as an operand for stringification in the expansion of macro spin_lock_init_prof.

The shouldn't the marker be on the definition of spin_lock_init_prof(), rather than ...

Quoted 11 lines
xen/common/domain.c +1 -0
@@ -632,6 +632,7 @@ struct domain *domain_create(domid_t domid,
 
     atomic_set(&d->refcnt, 1);
     RCU_READ_LOCK_INIT(&d->rcu_lock);
+    /* SAF-6-safe Rule 20.12 expansion of macro domain_lock in debug 

builds */

     rspin_lock_init_prof(d, domain_lock);
     rspin_lock_init_prof(d, page_alloc_lock);
     spin_lock_init(&d->hypercall_deadlock_mutex);

... actually just one of the two uses here (and presumably several more elsewhere)?

Quoted 1 line

Jan

Actually it seems that this violation went away with some refactorings, so this patch is no longer needed other than for the addition to safe.json, so it can be folded into the next one. I'll make the adjustment.

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

PATCH 2/5 x86/domain: deviate violation of MISRA C Rule 20.12

In reply to Nicola Vetrini

MISRA C Rule 20.12 states: "A macro parameter used as an operand to the # or ## operators, which is itself subject to further macro replacement, shall only be used as an operand to these operators".

In this case, builds where CONFIG_COMPAT=y the fpu_ctxt macro is used both as a regular macro argument and as an operand for stringification in the expansion of CHECK_FIELD_.
This is deviated using a SAF-x-safe comment.

No functional change.

---
 xen/arch/x86/domain.c | 1 +
 xen/arch/x86/domctl.c | 1 +
 2 files changed, 2 insertions(+)
xen/arch/x86/domain.c +1 -0
index 536542841ef5..ccadfe0c9e70 100644
@@ -1084,6 +1084,7 @@ void arch_domain_creation_finished(struct domain *d)
 #ifdef CONFIG_COMPAT
 #define xen_vcpu_guest_context vcpu_guest_context
 #define fpu_ctxt fpu_ctxt.x
+/* SAF-6-safe Rule 20.12 expansion of macro fpu_ctxt with CONFIG_COMPAT */
 CHECK_FIELD_(struct, vcpu_guest_context, fpu_ctxt);
 #undef fpu_ctxt
 #undef xen_vcpu_guest_context
xen/arch/x86/domctl.c +1 -0
index 9a72d57333e9..335aedf46d03 100644
@@ -1326,6 +1326,7 @@ long arch_do_domctl(
 #ifdef CONFIG_COMPAT
 #define xen_vcpu_guest_context vcpu_guest_context
 #define fpu_ctxt fpu_ctxt.x
+/* SAF-6-safe Rule 20.12 expansion of macro fpu_ctxt with CONFIG_COMPAT */
 CHECK_FIELD_(struct, vcpu_guest_context, fpu_ctxt);
 #undef fpu_ctxt
 #undef xen_vcpu_guest_context
Signature · 1 line
2.34.1
Acked-by on v1 2/5
#5 lore raw

Re: [XEN PATCH 2/5] x86/domain: deviate violation of MISRA C Rule 20.12

Quoted 4 lines · Nicola Vetrini

MISRA C Rule 20.12 states: "A macro parameter used as an operand to the # or ## operators, which is itself subject to further macro replacement, shall only be used as an operand to these operators".

Quoted 8 lines

In this case, builds where CONFIG_COMPAT=y the fpu_ctxt macro is used both as a regular macro argument and as an operand for stringification in the expansion of CHECK_FIELD_.
This is deviated using a SAF-x-safe comment.

No functional change.

At least for the time being; plans that Andrew vaguely described my render this unnecessary down the road.

Jan

Quoted 22 lines
xen/arch/x86/domain.c +1 -0
@@ -1084,6 +1084,7 @@ void arch_domain_creation_finished(struct domain *d)
 #ifdef CONFIG_COMPAT
 #define xen_vcpu_guest_context vcpu_guest_context
 #define fpu_ctxt fpu_ctxt.x
+/* SAF-6-safe Rule 20.12 expansion of macro fpu_ctxt with CONFIG_COMPAT */
 CHECK_FIELD_(struct, vcpu_guest_context, fpu_ctxt);
 #undef fpu_ctxt
 #undef xen_vcpu_guest_context
xen/arch/x86/domctl.c +1 -0
index 9a72d57333e9..335aedf46d03 100644
@@ -1326,6 +1326,7 @@ long arch_do_domctl(
 #ifdef CONFIG_COMPAT
 #define xen_vcpu_guest_context vcpu_guest_context
 #define fpu_ctxt fpu_ctxt.x
+/* SAF-6-safe Rule 20.12 expansion of macro fpu_ctxt with CONFIG_COMPAT */
 CHECK_FIELD_(struct, vcpu_guest_context, fpu_ctxt);
 #undef fpu_ctxt
 #undef xen_vcpu_guest_context
#6 lore raw

PATCH 3/5 x86: deviate violation of MISRA C Rule 20.12

In reply to Nicola Vetrini

MISRA C Rule 20.12 states: "A macro parameter used as an operand to the # or ## operators, which is itself subject to further macro replacement, shall only be used as an operand to these operators".

When the second parameter of GET_SET_SHARED is a macro and is used as both a regular parameter and for token pasting the rule deliberately violated. A SAF-x-safe comment is used to deviate the usage.

No functional change.

---
 xen/arch/x86/include/asm/shared.h | 1 +
 1 file changed, 1 insertion(+)
xen/arch/x86/include/asm/shared.h +1 -0
index 60b67fa4b427..c26d4b2b3f0f 100644
@@ -76,6 +76,7 @@ static inline void arch_set_##field(struct vcpu *v,         \
 
 GET_SET_SHARED(unsigned long, max_pfn)
 GET_SET_SHARED(xen_pfn_t, pfn_to_mfn_frame_list_list)
+/* SAF-6-safe Rule 20.12: expansion of macro nmi_reason */
 GET_SET_SHARED(unsigned long, nmi_reason)
 
 GET_SET_VCPU(unsigned long, cr2)
Signature · 1 line
2.34.1
#7 lore raw

Re: [XEN PATCH 3/5] x86: deviate violation of MISRA C Rule 20.12

xen/arch/x86/include/asm/shared.h static inline void arch_set_##field(struct vcpu *v, \ · line 80

Quoted 8 lines · Nicola Vetrini
xen/arch/x86/include/asm/shared.h +1 -0
@@ -76,6 +76,7 @@ static inline void arch_set_##field(struct vcpu *v,         \
 
 GET_SET_SHARED(unsigned long, max_pfn)
 GET_SET_SHARED(xen_pfn_t, pfn_to_mfn_frame_list_list)
+/* SAF-6-safe Rule 20.12: expansion of macro nmi_reason */
 GET_SET_SHARED(unsigned long, nmi_reason)

Before we go this route, were alternatives at least considered? Plus didn't we special-case function-like macros already, when used in situations where only object-like macros would be expanded anyway?

As to alternatives: nmi_reason() is used in exactly one place. Dropping the #define and expanding the one use instead would be an option. I further wonder whether moving the #define-s past the piece of code you actually modify would also be an option (i.e. the tool then no longer complaining).

Jan

#8 lore raw

Re: [XEN PATCH 3/5] x86: deviate violation of MISRA C Rule 20.12

Quoted 6 lines · Jan Beulich
Quoted 5 lines · Nicola Vetrini
xen/arch/x86/include/asm/shared.h +0 -0
@@ -76,6 +76,7 @@ static inline void arch_set_##field(struct vcpu *v,  
       \
 
Quoted 9 lines

xen/arch/x86/include/asm/shared.h static inline void arch_set_##field(struct vcpu *v, \ · line 80

Quoted 4 lines
xen/arch/x86/include/asm/shared.h +1 -0
 GET_SET_SHARED(unsigned long, max_pfn)
 GET_SET_SHARED(xen_pfn_t, pfn_to_mfn_frame_list_list)
+/* SAF-6-safe Rule 20.12: expansion of macro nmi_reason */
 GET_SET_SHARED(unsigned long, nmi_reason)

Before we go this route, were alternatives at least considered? Plus didn't we special-case function-like macros already, when used in situations where only object-like macros would be expanded anyway?

It may be the case that this is already deviated, hence meaning that the patch can be dropped: I'll recheck.

In that case, thanks for pointing this out.

Quoted 7 lines

As to alternatives: nmi_reason() is used in exactly one place. Dropping the #define and expanding the one use instead would be an option. I further wonder whether moving the #define-s past the piece of code you actually modify would also be an option (i.e. the tool then no longer complaining).

Jan

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

PATCH 4/5 automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

In reply to Nicola Vetrini

The DEFINE macro in asm-offsets.c (for all architectures) still generates violations despite the file(s) being excluded from compliance, due to the fact that in its expansion it sometimes refers entities in non-excluded files. These corner cases are deviated by the configuration.

No functional change.

---
 automation/eclair_analysis/ECLAIR/deviations.ecl | 6 ++++++
 1 file changed, 6 insertions(+)
automation/eclair_analysis/ECLAIR/deviations.ecl +6 -0
index cf62a874d928..f29db9e08248 100644
@@ -483,6 +483,12 @@ leads to a violation of the Rule are deviated."
 -config=MC3R1.R20.12,macros+={deliberate, "name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
 -doc_end
 
+-doc_begin="The macro DEFINE is defined and used in excluded files asm-offsets.c.
+This may still cause violations if entities outside these files are referred to
+in the expansion."
+-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end
+
 #
 # Series 21.
 #
Signature · 1 line
2.34.1
#10 lore raw

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

automation/eclair_analysis/ECLAIR/deviations.ecl leads to a violation of the Rule are deviated." · line 490

Quoted 11 lines · Nicola Vetrini
automation/eclair_analysis/ECLAIR/deviations.ecl +5 -0
@@ -483,6 +483,12 @@ leads to a violation of the Rule are deviated."
 -config=MC3R1.R20.12,macros+={deliberate, "name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
 -doc_end
 
+-doc_begin="The macro DEFINE is defined and used in excluded files asm-offsets.c.
+This may still cause violations if entities outside these files are referred to
+in the expansion."
+-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end

Can you give an example of such a reference? Nothing _in_ asm-offsets.c should be referenced, I'd think. Only stuff in asm-offsets.h as _generated from_ asm-offsets.c will, of course, be.

Jan

#11 lore raw

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

Quoted 8 lines · Jan Beulich
Quoted 7 lines · Nicola Vetrini
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
@@ -483,6 +483,12 @@ leads to a violation of the Rule are deviated."
 -config=MC3R1.R20.12,macros+={deliberate, 
 "name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
 -doc_end
 
Quoted 13 lines
Quoted 8 lines
automation/eclair_analysis/ECLAIR/deviations.ecl +1 -0
+-doc_begin="The macro DEFINE is defined and used in excluded files 

asm-offsets.c.
+This may still cause violations if entities outside these files are referred to
+in the expansion." +-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end

Can you give an example of such a reference? Nothing _in_ asm-offsets.c should be referenced, I'd think. Only stuff in asm-offsets.h as _generated from_ asm-offsets.c will, of course, be.

Perhaps I could have expressed that more clearly. What I meant is that there are some arguments to DEFINE that are not part of asm-offsets.c, therefore they end up in the violation report, but are not actually relevant, because the macro DEFINE is actually what we want to exclude.

See for instance at the link below VCPU_TRAP_{NMI,MCE}, which are defined in asm/domain.h and used as arguments to DEFINE inside asm-offsets.c.

https://saas.eclairit.com:3787/fs/var/local/eclair/XEN.ecdf/ECLAIR_normal/staging/X86_64-BUGSENG/676/PROJECT.ecd;/by_service/MC3R1.R20.12.html

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

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

Quoted 29 lines · Nicola Vetrini
Quoted 8 lines · Jan Beulich
Quoted 7 lines · Nicola Vetrini
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
@@ -483,6 +483,12 @@ leads to a violation of the Rule are deviated."
 -config=MC3R1.R20.12,macros+={deliberate, 
 "name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
 -doc_end
 
Quoted 14 lines
Quoted 8 lines
automation/eclair_analysis/ECLAIR/deviations.ecl +1 -0
+-doc_begin="The macro DEFINE is defined and used in excluded files 

asm-offsets.c.
+This may still cause violations if entities outside these files are referred to
+in the expansion." +-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end

Can you give an example of such a reference? Nothing _in_ asm-offsets.c should be referenced, I'd think. Only stuff in asm-offsets.h as _generated
from_ asm-offsets.c will, of course, be.

Perhaps I could have expressed that more clearly. What I meant is that there are some arguments to DEFINE that are not part of asm-offsets.c, therefore they end up in the violation report, but are not actually relevant, because the macro DEFINE is actually what we want to exclude.

Quoted 5 lines

See for instance at the link below VCPU_TRAP_{NMI,MCE}, which are defined in asm/domain.h and used as arguments to DEFINE inside asm-offsets.c.

https://saas.eclairit.com:3787/fs/var/local/eclair/XEN.ecdf/ECLAIR_normal/staging/X86_64-BUGSENG/676/PROJECT.ecd;/by_service/MC3R1.R20.12.html

I'm afraid I still don't understand: The file being supposed to be excluded from scanning, why does it even show up in that report?

Jan

#13 lore raw

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

Quoted 34 lines · Jan Beulich
Quoted 33 lines · Nicola Vetrini
Quoted 8 lines · Jan Beulich
Quoted 7 lines · Nicola Vetrini
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
@@ -483,6 +483,12 @@ leads to a violation of the Rule are deviated."
 -config=MC3R1.R20.12,macros+={deliberate,
 "name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
 -doc_end
 
Quoted 14 lines
Quoted 8 lines
automation/eclair_analysis/ECLAIR/deviations.ecl +1 -0
+-doc_begin="The macro DEFINE is defined and used in excluded files

asm-offsets.c.
+This may still cause violations if entities outside these files are referred to
+in the expansion." +-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end

Can you give an example of such a reference? Nothing _in_ asm-offsets.c should be referenced, I'd think. Only stuff in asm-offsets.h as _generated
from_ asm-offsets.c will, of course, be.

Perhaps I could have expressed that more clearly. What I meant is that there are some arguments to DEFINE that are not part of asm-offsets.c, therefore they end up in the violation report, but are not actually relevant, because the macro DEFINE is actually what we want to exclude.

See for instance at the link below VCPU_TRAP_{NMI,MCE}, which are defined in asm/domain.h and used as arguments to DEFINE inside asm-offsets.c.

Quoted 6 lines
Quoted 1 line

https://saas.eclairit.com:3787/fs/var/local/eclair/XEN.ecdf/ECLAIR_normal/staging/X86_64-BUGSENG/676/PROJECT.ecd;/by_service/MC3R1.R20.12.html

I'm afraid I still don't understand: The file being supposed to be excluded from scanning, why does it even show up in that report?

Jan

The report is made up of several source code locations. Three of them are within asm-offsets.c, which is excluded from compliance but still analyzed, but one references a macro definition in another file (e.g., VCPU_TRAP_NMI from asm/domain.h). So in this case the exclusion of asm-offsets.c is not enough for the report not to be shown.

The configuration is telling the tool that the macro DEFINE from asm-offsets is meant to be deviated, which is what is needed in this case.

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

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

Quoted 39 lines · Nicola Vetrini
Quoted 38 lines · Jan Beulich
Quoted 36 lines · Nicola Vetrini
Quoted 8 lines · Jan Beulich
Quoted 7 lines · Nicola Vetrini
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
@@ -483,6 +483,12 @@ leads to a violation of the Rule are deviated."
 -config=MC3R1.R20.12,macros+={deliberate,
 "name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
 -doc_end
 
Quoted 15 lines
Quoted 8 lines
automation/eclair_analysis/ECLAIR/deviations.ecl +1 -0
+-doc_begin="The macro DEFINE is defined and used in excluded files

asm-offsets.c.
+This may still cause violations if entities outside these files are referred to
+in the expansion." +-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end

Can you give an example of such a reference? Nothing _in_ asm-offsets.c
should be referenced, I'd think. Only stuff in asm-offsets.h as _generated
from_ asm-offsets.c will, of course, be.

Perhaps I could have expressed that more clearly. What I meant is that there are some arguments to DEFINE that are not part of asm-offsets.c, therefore they end up in the violation report, but are not actually relevant, because the macro DEFINE is actually what we want to exclude.

See for instance at the link below VCPU_TRAP_{NMI,MCE}, which are defined in asm/domain.h and used as arguments to DEFINE inside asm-offsets.c.

https://saas.eclairit.com:3787/fs/var/local/eclair/XEN.ecdf/ECLAIR_normal/staging/X86_64-BUGSENG/676/PROJECT.ecd;/by_service/MC3R1.R20.12.html

Quoted 8 lines
Quoted 2 lines

I'm afraid I still don't understand: The file being supposed to be excluded from scanning, why does it even show up in that report?

The report is made up of several source code locations. Three of them are within asm-offsets.c, which is excluded from compliance but still analyzed, but one references a macro definition in another file (e.g., VCPU_TRAP_NMI from asm/domain.h). So in this case the exclusion of asm-offsets.c is not enough for the report not to be shown.

But the (would-be-)violation is in asm-offsets.c. The other locations pointed at are providing context. To report a violation, it should be enough to exclude the file where the violation itself is?

Jan

#15 lore raw

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

Quoted 52 lines · Jan Beulich
Quoted 42 lines · Nicola Vetrini
Quoted 41 lines · Jan Beulich
Quoted 39 lines · Nicola Vetrini
Quoted 9 lines · Jan Beulich
Quoted 8 lines · Nicola Vetrini
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
@@ -483,6 +483,12 @@ leads to a violation of the Rule are 

deviated."
-config=MC3R1.R20.12,macros+={deliberate,
"name(GENERATE_CASE)&&loc(file(deliberate_generate_case))"}
-doc_end

Quoted 17 lines
Quoted 10 lines

+-doc_begin="The macro DEFINE is defined and used in excluded files
asm-offsets.c.
+This may still cause violations if entities outside these files are
referred to
+in the expansion." +-config=MC3R1.R20.12,macros+={deliberate, "name(DEFINE)&&loc(file(asm_offsets))"}
+-doc_end

Can you give an example of such a reference? Nothing _in_ asm-offsets.c
should be referenced, I'd think. Only stuff in asm-offsets.h as _generated
from_ asm-offsets.c will, of course, be.

Perhaps I could have expressed that more clearly. What I meant is that there are some arguments to DEFINE that are not part of asm-offsets.c, therefore they end up in the violation report, but are not actually relevant, because the macro DEFINE is actually what we want to exclude.

See for instance at the link below VCPU_TRAP_{NMI,MCE}, which are defined in asm/domain.h and used as arguments to DEFINE inside asm-offsets.c.

https://saas.eclairit.com:3787/fs/var/local/eclair/XEN.ecdf/ECLAIR_normal/staging/X86_64-BUGSENG/676/PROJECT.ecd;/by_service/MC3R1.R20.12.html

Quoted 8 lines
Quoted 2 lines

I'm afraid I still don't understand: The file being supposed to be excluded from scanning, why does it even show up in that report?

The report is made up of several source code locations. Three of them are within asm-offsets.c, which is excluded from compliance but still analyzed, but one references a macro definition in another file (e.g., VCPU_TRAP_NMI from asm/domain.h). So in this case the exclusion of asm-offsets.c is not enough for the report not to be shown.

Quoted 4 lines

But the (would-be-)violation is in asm-offsets.c. The other locations pointed at are providing context. To report a violation, it should be enough to exclude the file where the violation itself is?

In general that may not be the safest choice, or it can limit the granularity of the configuration (not in the specific case, but the mechanism is general). It's a design aspect of the tool to show a report unless all its locations are not meant to be shown.

Signature · 2 lines
Nicola Vetrini, BSc
Software Engineer, BUGSENG srl (https://bugseng.com)
Acked-by on v1 4/5
#16 lore raw

Re: [XEN PATCH 4/5] automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12

In reply to Nicola Vetrini
Quoted 8 lines · Nicola Vetrini

The DEFINE macro in asm-offsets.c (for all architectures) still generates violations despite the file(s) being excluded from compliance, due to the fact that in its expansion it sometimes refers entities in non-excluded files. These corner cases are deviated by the configuration.

No functional change.

#17 lore raw

PATCH 5/5 xen: fix MISRA regressions on rule 20.9 and 20.12

In reply to Nicola Vetrini

ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements") introduced new violations on previously clean rules 20.9 and 20.12.

The first is introduced because CONFIG_CC_IS_CLANG in xen/self-tests.h is not defined in the configuration under analysis. Using "defined()" instead avoids relying on the preprocessor's behaviour upon encountering an undedfined identifier and addresses the violation.

The violation of Rule 20.12 is due to "val" being used both as an ordinary argument in macro RUNTIME_CHECK, and as a stringification operator.

No functional change.

Fixes ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements")
---
 automation/eclair_analysis/ECLAIR/deviations.ecl | 2 +-
 xen/include/xen/self-tests.h                     | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
automation/eclair_analysis/ECLAIR/deviations.ecl +1 -1
index f29db9e08248..e2653f77eb2c 100644
@@ -473,7 +473,7 @@ deliberate."
 -doc_begin="Uses of a macro parameter for ordinary expansion and as an operand
 to the # or ## operators within the following macros are deliberate, to provide
 useful diagnostic messages to the user."
--config=MC3R1.R20.12,macros+={deliberate, "name(ASSERT||BUILD_BUG_ON||BUILD_BUG_ON_ZERO)"}
+-config=MC3R1.R20.12,macros+={deliberate, "name(ASSERT||BUILD_BUG_ON||BUILD_BUG_ON_ZERO||RUNTIME_CHECK)"}
 -doc_end
 
 -doc_begin="The helper macro GENERATE_CASE may use a macro parameter for ordinary
xen/include/xen/self-tests.h +1 -1
index 8410db7aaaae..42a4cc4d17fe 100644
@@ -16,7 +16,7 @@
  * Clang < 8 can't fold constants through static inlines, causing this to
  * fail.  Simply skip it for incredibly old compilers.
  */
-#if !CONFIG_CC_IS_CLANG || CONFIG_CLANG_VERSION >= 80000
+#if !defined(CONFIG_CC_IS_CLANG) || CONFIG_CLANG_VERSION >= 80000
 #define COMPILE_CHECK(fn, val, res)                                     \
     do {                                                                \
         typeof(fn(val)) real = fn(val);                                 \
Signature · 1 line
2.34.1
#18 lore raw

Re: [XEN PATCH 5/5] xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 8 lines · Nicola Vetrini

ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements") introduced new violations on previously clean rules 20.9 and 20.12.

The first is introduced because CONFIG_CC_IS_CLANG in xen/self-tests.h is not defined in the configuration under analysis. Using "defined()" instead avoids relying on the preprocessor's behaviour upon encountering an undedfined identifier and addresses the violation.

Quoted 7 lines

The violation of Rule 20.12 is due to "val" being used both as an ordinary argument in macro RUNTIME_CHECK, and as a stringification operator.

No functional change.

Fixes ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements")

Thankyou for this patch.  I'd seen that I'd broken something.  (Entirely my fault - I've done a lot of testing in Gitlab for the series, but never manually ran the Eclair jobs.  I'll try to remember better next time.)

One question though.
https://gitlab.com/xen-project/xen/-/jobs/6994213979 says:

Failure: 1 regressions found for clean guidelines service MC3R1.R20.9: (required) All identifiers used in the controlling expression of #if' or #elif' preprocessing directives shall be #define'd before evaluation:
   violation: 1

While there is a report for 20.12, it's not clean yet (so the first sentence wants adjusting), and RUNTIME_CHECK doesn't show up newly in it.

So, while I agree that RUNTIME_CHECK() should be included in the 20.12 exclusions, why is current Gitlab Run not reporting it?

~Andrew

#19 lore raw

Re: [XEN PATCH 5/5] xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 30 lines · Andrew Cooper
Quoted 8 lines · Nicola Vetrini

ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements") introduced new violations on previously clean rules 20.9 and 20.12.

The first is introduced because CONFIG_CC_IS_CLANG in xen/self-tests.h is not defined in the configuration under analysis. Using "defined()" instead avoids relying on the preprocessor's behaviour upon encountering an undedfined identifier and addresses the violation.

Quoted 7 lines

The violation of Rule 20.12 is due to "val" being used both as an ordinary argument in macro RUNTIME_CHECK, and as a stringification operator.

No functional change.

Fixes ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements")

Thankyou for this patch.  I'd seen that I'd broken something.  (Entirely my fault - I've done a lot of testing in Gitlab for the series, but never manually ran the Eclair jobs.  I'll try to remember better next time.)

One question though.
https://gitlab.com/xen-project/xen/-/jobs/6994213979 says:

Failure: 1 regressions found for clean guidelines service MC3R1.R20.9: (required) All identifiers used in the controlling expression of #if' or #elif' preprocessing directives shall be #define'd before evaluation:
   violation: 1

Quoted 7 lines

While there is a report for 20.12, it's not clean yet (so the first sentence wants adjusting), and RUNTIME_CHECK doesn't show up newly in it.

So, while I agree that RUNTIME_CHECK() should be included in the 20.12 exclusions, why is current Gitlab Run not reporting it?

~Andrew

While Rule 20.12 wasn't clean on x86, but it was on arm, so in the arm64 run it is reported as such

https://gitlab.com/xen-project/xen/-/jobs/6994213980

with the other patches in the series the rule should be clean on both architectures, so this imbalance should disappear rather shortly.

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

Re: [XEN PATCH 5/5] xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 47 lines · Nicola Vetrini
Quoted 37 lines · Andrew Cooper
Quoted 16 lines · Nicola Vetrini

ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements")
introduced new violations on previously clean rules 20.9 and 20.12.

The first is introduced because CONFIG_CC_IS_CLANG in xen/self-tests.h is not
defined in the configuration under analysis. Using "defined()" instead avoids
relying on the preprocessor's behaviour upon encountering an undedfined identifier
and addresses the violation.

The violation of Rule 20.12 is due to "val" being used both as an ordinary argument
in macro RUNTIME_CHECK, and as a stringification operator.

Quoted 5 lines

No functional change.

Fixes ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure

ahead of rearrangements")

Thankyou for this patch.  I'd seen that I'd broken something.  (Entirely my fault - I've done a lot of testing in Gitlab for the series, but never manually ran the Eclair jobs.  I'll try to remember better next time.)

One question though.
https://gitlab.com/xen-project/xen/-/jobs/6994213979 says:

Failure: 1 regressions found for clean guidelines service MC3R1.R20.9: (required) All identifiers used in the controlling expression of #if' or #elif' preprocessing directives shall be #define'd before evaluation:
   violation: 1

Quoted 8 lines

While there is a report for 20.12, it's not clean yet (so the first sentence wants adjusting), and RUNTIME_CHECK doesn't show up newly in it.

So, while I agree that RUNTIME_CHECK() should be included in the 20.12 exclusions, why is current Gitlab Run not reporting it?

~Andrew

Quoted 8 lines

While Rule 20.12 wasn't clean on x86, but it was on arm, so in the arm64 run it is reported as such

https://gitlab.com/xen-project/xen/-/jobs/6994213980

with the other patches in the series the rule should be clean on both architectures, so this imbalance should disappear rather shortly.

Thanks - I'd just found the ARM report and was replying - I'll rebase onto this thread.

I still don't understand why the violation doesn't show up on the x86 build.  Specifically, why it's missing from here:
https://saas.eclairit.com:3787/fs/var/local/eclair/xen-project.ecdf/xen-project/xen/ECLAIR_normal/staging/X86_64/6994213979/prev/PROJECT.ecd;/by_service/MC3R1.R20.12.html

From the ARM report, one is xen/lib/generic-ffsl.c:61 and checking with the x86 build log, we are building generic-ffsl.c too (which is expected.)

~Andrew

#21 lore raw

Re: [XEN PATCH 5/5] xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 57 lines · Andrew Cooper
Quoted 47 lines · Nicola Vetrini
Quoted 37 lines · Andrew Cooper
Quoted 16 lines · Nicola Vetrini

ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements")
introduced new violations on previously clean rules 20.9 and 20.12.

The first is introduced because CONFIG_CC_IS_CLANG in xen/self-tests.h is not
defined in the configuration under analysis. Using "defined()" instead avoids
relying on the preprocessor's behaviour upon encountering an undedfined identifier
and addresses the violation.

The violation of Rule 20.12 is due to "val" being used both as an ordinary argument
in macro RUNTIME_CHECK, and as a stringification operator.

Quoted 5 lines

No functional change.

Fixes ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure

ahead of rearrangements")

Thankyou for this patch.  I'd seen that I'd broken something.  (Entirely my fault - I've done a lot of testing in Gitlab for the series, but never manually ran the Eclair jobs.  I'll try to remember better next time.)

One question though.
https://gitlab.com/xen-project/xen/-/jobs/6994213979 says:

Failure: 1 regressions found for clean guidelines service MC3R1.R20.9: (required) All identifiers used in the controlling expression of #if' or #elif' preprocessing directives shall be #define'd before evaluation:
   violation: 1

Quoted 8 lines

While there is a report for 20.12, it's not clean yet (so the first sentence wants adjusting), and RUNTIME_CHECK doesn't show up newly in it.

So, while I agree that RUNTIME_CHECK() should be included in the 20.12 exclusions, why is current Gitlab Run not reporting it?

~Andrew

Quoted 8 lines

While Rule 20.12 wasn't clean on x86, but it was on arm, so in the arm64 run it is reported as such

https://gitlab.com/xen-project/xen/-/jobs/6994213980

with the other patches in the series the rule should be clean on both architectures, so this imbalance should disappear rather shortly.

Quoted 7 lines

Thanks - I'd just found the ARM report and was replying - I'll rebase onto this thread.

I still don't understand why the violation doesn't show up on the x86 build.  Specifically, why it's missing from here:
https://saas.eclairit.com:3787/fs/var/local/eclair/xen-project.ecdf/xen-project/xen/ECLAIR_normal/staging/X86_64/6994213979/prev/PROJECT.ecd;/by_service/MC3R1.R20.12.html

Note the "prev" here in the URL: this means you're looking at the analysis run before your series was merged (on 03147e6837ff045db)

this is the latest run for x86 on staging:

https://saas.eclairit.com:3787/fs/var/local/eclair/xen-project.ecdf/xen-project/xen/ECLAIR_normal/staging/X86_64/6994213979/PROJECT.ecd;/by_service/MC3R1.R20.12.html

Quoted 4 lines

From the ARM report, one is xen/lib/generic-ffsl.c:61 and checking with the x86 build log, we are building generic-ffsl.c too (which is expected.)

~Andrew

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

Re: [XEN PATCH 5/5] xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 68 lines · Nicola Vetrini
Quoted 58 lines · Andrew Cooper
Quoted 48 lines · Nicola Vetrini
Quoted 38 lines · Andrew Cooper
Quoted 16 lines · Nicola Vetrini

ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure ahead of rearrangements")
introduced new violations on previously clean rules 20.9 and 20.12.

The first is introduced because CONFIG_CC_IS_CLANG in xen/self-tests.h is not
defined in the configuration under analysis. Using "defined()" instead avoids
relying on the preprocessor's behaviour upon encountering an undedfined identifier
and addresses the violation.

The violation of Rule 20.12 is due to "val" being used both as an ordinary argument
in macro RUNTIME_CHECK, and as a stringification operator.

Quoted 5 lines

No functional change.

Fixes ea59e7d780d9 ("xen/bitops: Cleanup and new infrastructure

ahead of rearrangements")

Thankyou for this patch.  I'd seen that I'd broken something. (Entirely
my fault - I've done a lot of testing in Gitlab for the series, but never manually ran the Eclair jobs.  I'll try to remember better next time.)

One question though.
https://gitlab.com/xen-project/xen/-/jobs/6994213979 says:

Failure: 1 regressions found for clean guidelines service MC3R1.R20.9: (required) All identifiers used in the controlling expression of #if' or #elif' preprocessing directives shall be #define'd before evaluation:
   violation: 1

Quoted 8 lines

While there is a report for 20.12, it's not clean yet (so the first sentence wants adjusting), and RUNTIME_CHECK doesn't show up newly in it.

So, while I agree that RUNTIME_CHECK() should be included in the 20.12 exclusions, why is current Gitlab Run not reporting it?

~Andrew

Quoted 8 lines

While Rule 20.12 wasn't clean on x86, but it was on arm, so in the arm64 run it is reported as such

https://gitlab.com/xen-project/xen/-/jobs/6994213980

with the other patches in the series the rule should be clean on both architectures, so this imbalance should disappear rather shortly.

Quoted 8 lines

Thanks - I'd just found the ARM report and was replying - I'll rebase onto this thread.

I still don't understand why the violation doesn't show up on the x86 build.  Specifically, why it's missing from here:
https://saas.eclairit.com:3787/fs/var/local/eclair/xen-project.ecdf/xen-project/xen/ECLAIR_normal/staging/X86_64/6994213979/prev/PROJECT.ecd;/by_service/MC3R1.R20.12.html

Quoted 6 lines

Note the "prev" here in the URL: this means you're looking at the analysis run before your series was merged (on 03147e6837ff045db)

this is the latest run for x86 on staging:

https://saas.eclairit.com:3787/fs/var/local/eclair/xen-project.ecdf/xen-project/xen/ECLAIR_normal/staging/X86_64/6994213979/PROJECT.ecd;/by_service/MC3R1.R20.12.html

Oh - thankyou for explaining.

~Andrew

#23 lore raw
In reply to Nicola Vetrini
Quoted 11 lines · Nicola Vetrini

Patches 1 to 4 address violations of MISRA C Rule 20.12 by deviating certain uses of some macros, while the last patch addresses some regressions introduced by the latest bitops series

Nicola Vetrini (5):
xen/domain: deviate violation of MISRA C Rule 20.12 x86/domain: deviate violation of MISRA C Rule 20.12
x86: deviate violation of MISRA C Rule 20.12 automation/eclair_analysis: address remaining violations of MISRA C

    Rule 20.12

xen: fix MISRA regressions on rule 20.9 and 20.12

I've committed patch 5 because it fixes a blocking failure in Gitlab CI from content already accepted for Xen 4.19.

The others look fine to me, but you'll need to negotiate with Oleksii (CC'd) to get them in, at this point in the release.

Given that this series makes x86 clean to Rule 20.12, shouldn't there be a final patch making it blocking, to bring x86 in line with ARM?

~Andrew

#24 lore raw
Quoted 13 lines · Andrew Cooper
Quoted 11 lines · Nicola Vetrini

Patches 1 to 4 address violations of MISRA C Rule 20.12 by deviating certain uses of some macros, while the last patch addresses some regressions introduced by the latest bitops series

Nicola Vetrini (5):
xen/domain: deviate violation of MISRA C Rule 20.12 x86/domain: deviate violation of MISRA C Rule 20.12
x86: deviate violation of MISRA C Rule 20.12 automation/eclair_analysis: address remaining violations of MISRA C

    Rule 20.12

xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 3 lines

I've committed patch 5 because it fixes a blocking failure in Gitlab CI from content already accepted for Xen 4.19.

Thanks

Quoted 3 lines

The others look fine to me, but you'll need to negotiate with Oleksii (CC'd) to get them in, at this point in the release.

Well, having one more clean rule does look better for Xen, and the changes to the codebase are harmless enough, but ultimately given the closeness with the deadline I didn't really see a need to.

Quoted 3 lines

Given that this series makes x86 clean to Rule 20.12, shouldn't there be a final patch making it blocking, to bring x86 in line with ARM?

Good point, I should have done that in patch 4. I'll do a follow-up patch.

Quoted 1 line

~Andrew

Signature · 2 lines
Nicola Vetrini, BSc
Software Engineer, BUGSENG srl (https://bugseng.com)
#25 lore raw
Quoted 30 lines · Nicola Vetrini
Quoted 15 lines · Andrew Cooper
Quoted 6 lines · Nicola Vetrini

Patches 1 to 4 address violations of MISRA C Rule 20.12 by deviating certain
uses of some macros, while the last patch addresses some regressions introduced
by the latest bitops series

Quoted 7 lines

Nicola Vetrini (5):
  xen/domain: deviate violation of MISRA C Rule 20.12 x86/domain: deviate violation of MISRA C Rule 20.12
  x86: deviate violation of MISRA C Rule 20.12 automation/eclair_analysis: address remaining violations of MISRA C Rule 20.12
  xen: fix MISRA regressions on rule 20.9 and 20.12

Quoted 3 lines

I've committed patch 5 because it fixes a blocking failure in Gitlab CI from content already accepted for Xen 4.19.

Thanks

Quoted 3 lines

The others look fine to me, but you'll need to negotiate with Oleksii (CC'd) to get them in, at this point in the release.

Well, having one more clean rule does look better for Xen, and the changes to the codebase are harmless enough, but ultimately given the closeness with the deadline I didn't really see a need to.

Quoted 6 lines
Quoted 3 lines

Given that this series makes x86 clean to Rule 20.12, shouldn't there be a final patch making it blocking, to bring x86 in line with ARM?

Good point, I should have done that in patch 4. I'll do a follow-up patch.

FWIW, given how simple this series is, I'm +1 for including it in 4.19, even at this point.  It is definitely nicer to have the disposition of Rule 20.12 the same between ARM and x86.

Still - it's very much Oleksii's call.

~Andrew