linux-kernel - Re: [PATCH v2 03/14] iommu: Move bus setup to IOMMU device registration

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] [thread-next>] [day] [month] [year] [list]
Message-ID: <fa0d0663-5393-c533-105a-85bd2e9e0649@arm.com>
Date:   Fri, 29 Apr 2022 09:50:24 +0100
From:   Robin Murphy <robin.murphy@....com>
To:     Baolu Lu <baolu.lu@...ux.intel.com>, joro@...tes.org,
        will@...nel.org
Cc:     iommu@...ts.linux-foundation.org, sven@...npeter.dev,
        robdclark@...il.com, m.szyprowski@...sung.com,
        yong.wu@...iatek.com, mjrosato@...ux.ibm.com,
        gerald.schaefer@...ux.ibm.com, zhang.lyra@...il.com,
        thierry.reding@...il.com, vdumpa@...dia.com,
        jean-philippe@...aro.org, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 03/14] iommu: Move bus setup to IOMMU device
 registration

On 2022-04-29 07:57, Baolu Lu wrote:
> Hi Robin,
> 
> On 2022/4/28 21:18, Robin Murphy wrote:
>> Move the bus setup to iommu_device_register(). This should allow
>> bus_iommu_probe() to be correctly replayed for multiple IOMMU instances,
>> and leaves bus_set_iommu() as a glorified no-op to be cleaned up next.
> 
> I re-fetched the latest patches on
> 
> https://gitlab.arm.com/linux-arm/linux-rm/-/commits/iommu/bus
> 
> and rolled back the head to "iommu: Cleanup bus_set_iommu".
> 
> The test machine still hangs during boot.
> 
> I went through the code. It seems that the .probe_device for Intel IOMMU
> driver can't handle the probe replay well. It always assumes that the
> device has never been probed.

Hmm, but probe_iommu_group() is supposed to prevent the 
__iommu_probe_device() call even happening if the device *has* already 
been probed before :/

I've still got an old Intel box spare in the office so I'll rig that up 
and see if I can see what might be going on here...

Thanks,
Robin.

> 
> Best regards,
> baolu
> 
>>
>> At this point we can also handle cleanup better than just rolling back
>> the most-recently-touched bus upon failure - which may release devices
>> owned by other already-registered instances, and still leave devices on
>> other buses with dangling pointers to the failed instance. Now it's easy
>> to clean up the exact footprint of a given instance, no more, no less.
>>
>> Tested-by: Marek Szyprowski <m.szyprowski@...sung.com>
>> Signed-off-by: Robin Murphy <robin.murphy@....com>
>> ---
>>   drivers/iommu/iommu.c | 51 +++++++++++++++++++++++--------------------
>>   1 file changed, 27 insertions(+), 24 deletions(-)
>>
>> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
>> index 6c4621afc8cf..c89af4dc54c2 100644
>> --- a/drivers/iommu/iommu.c
>> +++ b/drivers/iommu/iommu.c
>> @@ -175,6 +175,14 @@ static int __init iommu_subsys_init(void)
>>   }
>>   subsys_initcall(iommu_subsys_init);
>> +static int remove_iommu_group(struct device *dev, void *data)
>> +{
>> +    if (dev->iommu && dev->iommu->iommu_dev == data)
>> +        iommu_release_device(dev);
>> +
>> +    return 0;
>> +}
>> +
>>   /**
>>    * iommu_device_register() - Register an IOMMU hardware instance
>>    * @iommu: IOMMU handle for the instance
>> @@ -197,12 +205,29 @@ int iommu_device_register(struct iommu_device 
>> *iommu,
>>       spin_lock(&iommu_device_lock);
>>       list_add_tail(&iommu->list, &iommu_device_list);
>>       spin_unlock(&iommu_device_lock);
>> +
>> +    for (int i = 0; i < ARRAY_SIZE(iommu_buses); i++) {
>> +        struct bus_type *bus = iommu_buses[i];
>> +        int err;
>> +
>> +        WARN_ON(bus->iommu_ops && bus->iommu_ops != ops);
>> +        bus->iommu_ops = ops;
>> +        err = bus_iommu_probe(bus);
>> +        if (err) {
>> +            iommu_device_unregister(iommu);
>> +            return err;
>> +        }
>> +    }
>> +
>>       return 0;
>>   }
>>   EXPORT_SYMBOL_GPL(iommu_device_register);
>>   void iommu_device_unregister(struct iommu_device *iommu)
>>   {
>> +    for (int i = 0; i < ARRAY_SIZE(iommu_buses); i++)
>> +        bus_for_each_dev(iommu_buses[i], NULL, iommu, 
>> remove_iommu_group);
>> +
>>       spin_lock(&iommu_device_lock);
>>       list_del(&iommu->list);
>>       spin_unlock(&iommu_device_lock);
>> @@ -1655,13 +1680,6 @@ static int probe_iommu_group(struct device 
>> *dev, void *data)
>>       return ret;
>>   }
>> -static int remove_iommu_group(struct device *dev, void *data)
>> -{
>> -    iommu_release_device(dev);
>> -
>> -    return 0;
>> -}
>> -
>>   static int iommu_bus_notifier(struct notifier_block *nb,
>>                     unsigned long action, void *data)
>>   {
>> @@ -1884,27 +1902,12 @@ static int iommu_bus_init(struct bus_type *bus)
>>    */
>>   int bus_set_iommu(struct bus_type *bus, const struct iommu_ops *ops)
>>   {
>> -    int err;
>> -
>> -    if (ops == NULL) {
>> -        bus->iommu_ops = NULL;
>> -        return 0;
>> -    }
>> -
>> -    if (bus->iommu_ops != NULL)
>> +    if (bus->iommu_ops && ops && bus->iommu_ops != ops)
>>           return -EBUSY;
>>       bus->iommu_ops = ops;
>> -    /* Do IOMMU specific setup for this bus-type */
>> -    err = bus_iommu_probe(bus);
>> -    if (err) {
>> -        /* Clean up */
>> -        bus_for_each_dev(bus, NULL, NULL, remove_iommu_group);
>> -        bus->iommu_ops = NULL;
>> -    }
>> -
>> -    return err;
>> +    return 0;
>>   }
>>   EXPORT_SYMBOL_GPL(bus_set_iommu);
>