Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <a84ac37c-2b03-35f2-2275-442b448cf6b5@linux.intel.com>
Date:   Fri, 13 Jan 2023 22:12:48 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.1
Cc:     baolu.lu@linux.intel.com
Subject: Re: [PATCH 3/7] iommu/vt-d: Support Enhanced Command Interface
To:     kan.liang@linux.intel.com, joro@8bytes.org, will@kernel.org,
        dwmw2@infradead.org, robin.murphy@arm.com, robert.moore@intel.com,
        rafael.j.wysocki@intel.com, lenb@kernel.org, iommu@lists.linux.dev,
        linux-kernel@vger.kernel.org
References: <20230111202504.378258-1-kan.liang@linux.intel.com>
 <20230111202504.378258-4-kan.liang@linux.intel.com>
 <6783f737-2701-77e3-ce83-8d287be51cf6@linux.intel.com>
Content-Language: en-US
From:   Baolu Lu <baolu.lu@linux.intel.com>
In-Reply-To: <6783f737-2701-77e3-ce83-8d287be51cf6@linux.intel.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 2023/1/13 21:55, Baolu Lu wrote:
>> +/*
>> + * Function to submit a command to the enhanced command interface. The
>> + * valid enhanced command descriptions are defined in Table 47 of the
>> + * VT-d spec. The VT-d hardware implementation may support some but not
>> + * all commands, which can be determined by checking the Enhanced
>> + * Command Capability Register.
>> + *
>> + * Return values:
>> + *  - 0: Command successful without any error;
>> + *  - Negative: software error value;
>> + *  - Nonzero positive: failure status code defined in Table 48.
>> + */
>> +int ecmd_submit_sync(struct intel_iommu *iommu, u8 ecmd,
>> +             u64 oa, bool has_ob, u64 ob)
>> +{
>> +    unsigned long flags;
>> +    u64 res;
>> +    int ret;
>> +
>> +    if (!cap_ecmds(iommu->cap))
>> +        return -ENODEV;
>> +
>> +    raw_spin_lock_irqsave(&iommu->register_lock, flags);
>> +
>> +    res = dmar_readq(iommu->reg + DMAR_ECRSP_REG);
>> +    if (res & DMA_ECMD_ECRSP_IP) {
>> +        ret = -EBUSY;
>> +        goto err;
>> +    }
>> +
>> +    if (has_ob)
>> +        dmar_writeq(iommu->reg + DMAR_ECEO_REG, ob);
> 
> The ecmds that require a Operand B are statically defined in the spec,
> right? What will it look like if we define a static ignore_ob(ecmd)?

Or simply remove has_ob parameter? The least case is an unnecessary
write to a register. It's fine as far as I can see since we should avoid
using it in any critical path.

--
Best regards,
baolu