Hi,
On 20/06/2022 08:02, Michal Orzel wrote:
According to MISRA C 2012 Rule 8.1, types shall be explicitly
specified. Fix all the findings reported by cppcheck with misra addon
by substituting implicit type 'unsigned' to explicit 'unsigned int'.
Signed-off-by: Michal Orzel <michal.or...@arm.com>
---
xen/include/xen/perfc.h | 2 +-
xen/include/xen/sched.h | 6 +++---
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/xen/include/xen/perfc.h b/xen/include/xen/perfc.h
index bb010b0aae..7c5ce537bd 100644
--- a/xen/include/xen/perfc.h
+++ b/xen/include/xen/perfc.h
@@ -49,7 +49,7 @@ enum perfcounter {
#undef PERFSTATUS
#undef PERFSTATUS_ARRAY
-typedef unsigned perfc_t;
+typedef unsigned int perfc_t;
#define PRIperfc ""
DECLARE_PER_CPU(perfc_t[NUM_PERFCOUNTERS], perfcounters);
diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h
index 463d41ffb6..d957b6e11f 100644
--- a/xen/include/xen/sched.h
+++ b/xen/include/xen/sched.h
@@ -518,9 +518,9 @@ struct domain
/* hvm_print_line() and guest_console_write() logging. */
#define DOMAIN_PBUF_SIZE 200
- char *pbuf;
- unsigned pbuf_idx;
- spinlock_t pbuf_lock;
+ char *pbuf;
+ unsigned int pbuf_idx;
+ spinlock_t pbuf_lock;
Looking a "struct domain", we don't often seem to align the name. In
fact, this is something I don't particularly like because it adds a lot
of churn in a patch.
So my preference would be to just change the line "unsigned".
Other than that:
Acked-by: Julien Grall <jgr...@amazon.com>
Cheers,
--
Julien Grall