Re: [PATCH v4 05/16] iommu: Move bus setup to IOMMU device registration

From: Dmitry Osipenko <digetx@gmail.com>
To: Jon Hunter <jonathanh@nvidia.com>,
	Robin Murphy <robin.murphy@arm.com>,
	joro@8bytes.org, Thierry Reding <treding@nvidia.com>
Cc: will@kernel.org, iommu@lists.linux.dev,
	linux-arm-kernel@lists.infradead.org, baolu.lu@linux.intel.com,
	kevin.tian@intel.com, suravee.suthikulpanit@amd.com,
	vasant.hegde@amd.com, mjrosato@linux.ibm.com,
	schnelle@linux.ibm.com, linux-kernel@vger.kernel.org,
	"linux-tegra@vger.kernel.org" <linux-tegra@vger.kernel.org>
Subject: Re: [PATCH v4 05/16] iommu: Move bus setup to IOMMU device registration
Date: Wed, 19 Oct 2022 00:12:58 +0300	[thread overview]
Message-ID: <712c9e51-f589-b566-6194-259393180e43@gmail.com> (raw)
In-Reply-To: <41293685-54aa-6fa3-ca41-189fdbb7e8b0@nvidia.com>

18.10.2022 09:13, Jon Hunter пишет:
> 
> On 06/10/2022 16:27, Robin Murphy wrote:
>> On 2022-10-06 15:01, Jon Hunter wrote:
>>> Hi Robin,
>>>
>>> On 15/08/2022 17:20, 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.
>>>>
>>>> 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.
>>>
>>>
>>> Since this change, I have noticed that the DRM driver on Tegra20 is
>>> failing to probe and I am seeing ...
>>>
>>>   tegra-gr2d 54140000.gr2d: failed to attach to domain: -19
>>>   drm drm: failed to initialize 54140000.gr2d: -19
>>>
>>> Bisect points to this change and reverting it fixes it. Let me know
>>> if you have any thoughts.
>>
>> Oh, apparently what's happened is that I've inadvertently enabled the
>> tegra-gart driver, since it seems that *wasn't* calling
>> bus_set_iommu() before. Looking at the history, it appears to have
>> been that way since c7e3ca515e78 ("iommu/tegra: gart: Do not register
>> with bus"), so essentially that driver has been broken and useless for
>> close to 8 years now :(
>>
>> Given that, I'd be inclined to "fix" it as below, or just give up and
>> delete the whole thing.
>>
>> Thanks,
>> Robin.
>>
>> ----->8-----
>> diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig
>> index 5c5cb5bee8b6..7b3f7fd6e527 100644
>> --- a/drivers/iommu/Kconfig
>> +++ b/drivers/iommu/Kconfig
>> @@ -230,6 +230,7 @@ config TEGRA_IOMMU_GART
>>       bool "Tegra GART IOMMU Support"
>>       depends on ARCH_TEGRA_2x_SOC
>>       depends on TEGRA_MC
>> +    depends on BROKEN
>>       select IOMMU_API
>>       help
>>         Enables support for remapping discontiguous physical memory
> 
> 
> Thanks Robin. This works for me.
> 
> Thierry, Dmitry, we need a fix for v6.1 and so OK with the above?

To me it is more a problem of the DRM driver that it doesn't support
GART. GART will require a special handling from the DRM driver anyways [1].

[1]
https://github.com/grate-driver/linux/blob/master/drivers/gpu/drm/grate/drm.c#L460

The GART driver itself isn't broken, it's working perfectly fine. It's
the DRM driver that should start caring about the GART presence, IMO.