Re: [PATCH v3 9/9] LoongArch: KVM: INTC: Add address alignment check

[Bibo Mao <maobibo@loongson.cn>]

On 2025/6/19 下午4:47, Huacai Chen wrote:
> Hi, Bibo,
> 
> On Wed, Jun 11, 2025 at 9:51 AM Bibo Mao <maobibo@loongson.cn> wrote:
>>
>> IOCSR instruction supports 1/2/4/8 bytes access, the address should
>> be naturally aligned with its access size. Here address alignment
>> check is added in eiointc kernel emulation.
>>
>> At the same time len must be 1/2/4/8 bytes from iocsr exit emulation
>> function kvm_emu_iocsr(), remove the default case in switch case
>> statements.
> Robust code doesn't depend its callers do things right, so I suggest
> keeping the default case, which means we just add the alignment check
> here.
ok, will keep the default case.
> 
> And I think this patch should also Cc stable and add a Fixes tag.
ok, will add Cc stabe and Fixes tag.

Regards
Bibo Mao
> 
> 
> Huacai
> 
>>
>> Signed-off-by: Bibo Mao <maobibo@loongson.cn>
>> ---
>>   arch/loongarch/kvm/intc/eiointc.c | 21 +++++++++++++--------
>>   1 file changed, 13 insertions(+), 8 deletions(-)
>>
>> diff --git a/arch/loongarch/kvm/intc/eiointc.c b/arch/loongarch/kvm/intc/eiointc.c
>> index 8b0d9376eb54..4e9d12300cc4 100644
>> --- a/arch/loongarch/kvm/intc/eiointc.c
>> +++ b/arch/loongarch/kvm/intc/eiointc.c
>> @@ -311,6 +311,12 @@ static int kvm_eiointc_read(struct kvm_vcpu *vcpu,
>>                  return -EINVAL;
>>          }
>>
>> +       /* len must be 1/2/4/8 from function kvm_emu_iocsr() */
>> +       if (addr & (len - 1)) {
>> +               kvm_err("%s: eiointc not aligned addr %llx len %d\n", __func__, addr, len);
>> +               return -EINVAL;
>> +       }
>> +
>>          vcpu->stat.eiointc_read_exits++;
>>          spin_lock_irqsave(&eiointc->lock, flags);
>>          switch (len) {
>> @@ -323,12 +329,9 @@ static int kvm_eiointc_read(struct kvm_vcpu *vcpu,
>>          case 4:
>>                  ret = loongarch_eiointc_readl(vcpu, eiointc, addr, val);
>>                  break;
>> -       case 8:
>> +       default:
>>                  ret = loongarch_eiointc_readq(vcpu, eiointc, addr, val);
>>                  break;
>> -       default:
>> -               WARN_ONCE(1, "%s: Abnormal address access: addr 0x%llx, size %d\n",
>> -                                               __func__, addr, len);
>>          }
>>          spin_unlock_irqrestore(&eiointc->lock, flags);
>>
>> @@ -682,6 +685,11 @@ static int kvm_eiointc_write(struct kvm_vcpu *vcpu,
>>                  return -EINVAL;
>>          }
>>
>> +       if (addr & (len - 1)) {
>> +               kvm_err("%s: eiointc not aligned addr %llx len %d\n", __func__, addr, len);
>> +               return -EINVAL;
>> +       }
>> +
>>          vcpu->stat.eiointc_write_exits++;
>>          spin_lock_irqsave(&eiointc->lock, flags);
>>          switch (len) {
>> @@ -694,12 +702,9 @@ static int kvm_eiointc_write(struct kvm_vcpu *vcpu,
>>          case 4:
>>                  ret = loongarch_eiointc_writel(vcpu, eiointc, addr, val);
>>                  break;
>> -       case 8:
>> +       default:
>>                  ret = loongarch_eiointc_writeq(vcpu, eiointc, addr, val);
>>                  break;
>> -       default:
>> -               WARN_ONCE(1, "%s: Abnormal address access: addr 0x%llx, size %d\n",
>> -                                               __func__, addr, len);
>>          }
>>          spin_unlock_irqrestore(&eiointc->lock, flags);
>>
>> --
>> 2.39.3
>>