On Sun, Mar 29, 2020 at 02:36:51PM +0100, Wei Liu wrote:
> On Sun, Mar 29, 2020 at 10:25:12AM +0530, Simran Singhal wrote:
> > Assignment to a typed pointer is sufficient in C.
> > No cast is needed.
> >
> > Also, changed some u64/u32 to uint64_t/uint32_t.
> >
> > Signed-off-by: Simran Singhal <singhalsimr...@gmail.com>
> > ---
> > Changes in v2:
> > - Took the chance to change some uintX to uintX_t.
> >
> > xen/arch/x86/acpi/cpufreq/powernow.c | 2 +-
> > xen/arch/x86/cpu/vpmu.c | 2 +-
> > xen/arch/x86/hpet.c | 2 +-
> > xen/arch/x86/hvm/save.c | 2 +-
> > xen/arch/x86/hvm/vmx/vvmx.c | 12 ++++++------
> > 5 files changed, 10 insertions(+), 10 deletions(-)
> >
> > diff --git a/xen/arch/x86/acpi/cpufreq/powernow.c
> > b/xen/arch/x86/acpi/cpufreq/powernow.c
> > index 3cf9c6cd05..f620bebc7e 100644
> > --- a/xen/arch/x86/acpi/cpufreq/powernow.c
> > +++ b/xen/arch/x86/acpi/cpufreq/powernow.c
> > @@ -58,7 +58,7 @@ static void transition_pstate(void *pstate)
> >
> > static void update_cpb(void *data)
> > {
> > - struct cpufreq_policy *policy = (struct cpufreq_policy *)data;
> > + struct cpufreq_policy *policy = data;
> >
> > if (policy->turbo != CPUFREQ_TURBO_UNSUPPORTED) {
> > uint64_t msr_content;
> > diff --git a/xen/arch/x86/cpu/vpmu.c b/xen/arch/x86/cpu/vpmu.c
> > index e50d478d23..1ed39ef03f 100644
> > --- a/xen/arch/x86/cpu/vpmu.c
> > +++ b/xen/arch/x86/cpu/vpmu.c
> > @@ -337,7 +337,7 @@ void vpmu_do_interrupt(struct cpu_user_regs *regs)
> >
> > static void vpmu_save_force(void *arg)
> > {
> > - struct vcpu *v = (struct vcpu *)arg;
> > + struct vcpu *v = arg;
> > struct vpmu_struct *vpmu = vcpu_vpmu(v);
> >
> > if ( !vpmu_is_set(vpmu, VPMU_CONTEXT_LOADED) )
> > diff --git a/xen/arch/x86/hpet.c b/xen/arch/x86/hpet.c
> > index 86929b9ba1..c46e7cf4ee 100644
> > --- a/xen/arch/x86/hpet.c
> > +++ b/xen/arch/x86/hpet.c
> > @@ -215,7 +215,7 @@ again:
> > static void hpet_interrupt_handler(int irq, void *data,
> > struct cpu_user_regs *regs)
> > {
> > - struct hpet_event_channel *ch = (struct hpet_event_channel *)data;
> > + struct hpet_event_channel *ch = data;
> >
> > this_cpu(irq_count)--;
> >
> > diff --git a/xen/arch/x86/hvm/save.c b/xen/arch/x86/hvm/save.c
> > index 0fc59d3487..a2c56fbc1e 100644
> > --- a/xen/arch/x86/hvm/save.c
> > +++ b/xen/arch/x86/hvm/save.c
> > @@ -417,7 +417,7 @@ void _hvm_read_entry(struct hvm_domain_context *h, void
> > *dest,
> > memcpy(dest, &h->data[h->cur], d->length);
> >
> > if ( d->length < dest_len )
> > - memset((char *)dest + d->length, 0, dest_len - d->length);
> > + memset(dest + d->length, 0, dest_len - d->length);
>
> I believe you shouldn't drop the cast here either because dest is of
> type void*.
>
> Although the calculation in the end is the same (void* considered of
> size 1), I would still keep the cast such that the semantics is correct.
IMO dropping the case here is fine, as dest is of type void * the
calculation is correct and the cast just obfuscates it.
I usually cast things to void * instead of char * or uint8_t * in
order to do pointer additions with size 1.
Thanks, Roger.
