linux-kernel - Re: [PATCH v1] iommu/vt-d: Remove dead code in intel_iommu_domain_alloc_paging

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <d32ac499-fadb-401c-bb8c-9c3616da302a@gmail.com>
Date: Thu, 5 Jun 2025 11:55:20 +0800
From: Ethan Zhao <etzhao1900@...il.com>
To: Baolu Lu <baolu.lu@...ux.intel.com>, Wei Wang <wei.w.wang@...el.com>,
 kevin.tian@...el.com, yi.l.liu@...el.com, dwmw2@...radead.org,
 jroedel@...e.de, linux-kernel@...r.kernel.org, iommu@...ts.linux.dev
Subject: Re: [PATCH v1] iommu/vt-d: Remove dead code in
 intel_iommu_domain_alloc_paging_flags()



On 6/3/2025 10:39 AM, Baolu Lu wrote:
> On 5/30/25 17:13, Wei Wang wrote:
>> When dirty_tracking is enabled, first_stage is set to false to use the
>> second stage translation table. dmar_domain->use_first_level, which is
>> assigned from first_page, is guaranteed to be false when the execution
>> reaches the location of the code to be removed by this patch. So the
>> handling for dmar_domain->use_first_level being true there will never
>> be executed.
>>
>> Signed-off-by: Wei Wang<wei.w.wang@...el.com>
>> ---
>>   drivers/iommu/intel/iommu.c | 7 +------
>>   1 file changed, 1 insertion(+), 6 deletions(-)
>>
>> diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
>> index cb0b993bebb4..1145567c60f9 100644
>> --- a/drivers/iommu/intel/iommu.c
>> +++ b/drivers/iommu/intel/iommu.c
>> @@ -3418,13 +3418,8 @@ intel_iommu_domain_alloc_paging_flags(struct 
>> device *dev, u32 flags,
>>           spin_lock_init(&dmar_domain->s1_lock);
>>       }
>> -    if (dirty_tracking) {
>> -        if (dmar_domain->use_first_level) {
> 
> This *explicit* check enforces that dirty tracking cannot be supported
> for a domain that relies on first-stage translation due to the lack of
> enabling/disabling dirty tracking support.
> 
> While this might appear redundant, this prevents potential issues
> if related code is modified without awareness of this dependency. 
There would always something sooner or later like this to stop the 
improper configuration/coding , under an assumption that caller path 
always coded perfect, indeed, we could remove a lot of error stopper code.

But that world seems never works like that.

Thanks,
Ethan
> 
>> -            iommu_domain_free(domain);
>> -            return ERR_PTR(-EOPNOTSUPP);
>> -        }
> 
> Thanks,
> baolu
>