From: Jon Hunter <jonathanh@nvidia.com>
To: Diogo Ivo <diogo.ivo@tecnico.ulisboa.pt>, Jason Gunthorpe <jgg@ziepe.ca>
Cc: thierry.reding@gmail.com, vdumpa@nvidia.com, joro@8bytes.org,
will@kernel.org, robin.murphy@arm.com, baolu.lu@linux.intel.com,
jsnitsel@redhat.com, jroedel@suse.de,
linux-tegra@vger.kernel.org, iommu@lists.linux.dev,
regressions@lists.linux.dev
Subject: Re: [REGRESSION] Failed buffer allocation in Tegra fbdev
Date: Wed, 24 Jan 2024 10:15:34 +0000 [thread overview]
Message-ID: <b0334c5e-3a6c-4b58-b525-e72bed8899b3@nvidia.com> (raw)
In-Reply-To: <1754c86c-704e-4c80-93c3-1b75a9a247df@tecnico.ulisboa.pt>
On 23/01/2024 18:44, Diogo Ivo wrote:
>
> On 1/23/24 15:15, Jason Gunthorpe wrote:
>> On Tue, Jan 23, 2024 at 02:33:15PM +0000, diogo.ivo@tecnico.ulisboa.pt
>> wrote:
>>> Commit c8cc2655cc6c in the recent IOMMU changes breaks Tegra fbdev
>>> at least on the Pixel C with the following error message reporting
>>> a failed buffer allocation:
>>>
>>> [ 1.857660] drm drm: failed to allocate buffer of size 18432000
>>>
>>> This error message is printed from tegra_bo_alloc() which is called by
>>> tegra_bo_create() in tegra_fbdev_probe(), which may indicate that other
>>> allocations would fail as well, not just the framebuffer.
>> Presumably this is because iommu_map_sgtable() under
>> tegra_bo_iommu_map() fails?
>>
>> Which I suspect is because of the logic in
>> host1x_client_iommu_attach().
>>
>> After c8cc2655cc6c iommu_get_domain_for_dev() will never return NULL.
>>
>> So this:
>>
>> if (domain && domain != tegra->domain)
>> return 0;
>>
>> Will happen and the domain will be left at IDENTITY, while I suppose
>> the tegra_bo_iommu_map() assumes the domain was switched to
>> tegra->domain.
>>
>> Does this solve your issue?
>>
>> diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
>> index ff36171c8fb700..15c7910b2e1c76 100644
>> --- a/drivers/gpu/drm/tegra/drm.c
>> +++ b/drivers/gpu/drm/tegra/drm.c
>> @@ -960,7 +960,7 @@ int host1x_client_iommu_attach(struct
>> host1x_client *client)
>> * not the shared IOMMU domain, don't try to attach it to a
>> different
>> * domain. This allows using the IOMMU-backed DMA API.
>> */
>> - if (domain && domain != tegra->domain)
>> + if (domain && domain->type != IOMMU_DOMAIN_IDENTITY && domain
>> != tegra->domain)
>> return 0;
>> if (tegra->domain) {
>>
>>> This may be connected with an error in of_iommu_configure() that
>>> became visible after commit 6ff6e184f1f4d:
>>>
>>> [ 1.200004] host1x drm: iommu configuration for device failed with
>>> -ENOENT
>> Hmmm
>>
>> This is a new logging, so it doesn't necessarily mean something has
>> changed in the code flow.
>>
>> It seems the issue is something in there is returning ENOENT when it
>> probably should be ENODEV, but I haven't been able to guess where it
>> comes from.
>>
>> Can you do some tracing and figure out where under
>> of_iommu_configure() this ENOENT return code is from?
>>
>> Jason
>
> Yes, this does fix the issue!
I am seeing the same error message on Tegra194 AGX Jetson. However, the
above does not seem to fix it. In this case it is slightly different as
I am not seeing the DRM allocation failure. In this case, it appears to
be the of_dma_configure_id() call from the
host1x_memory_context_list_init() function that is failing.
Jon
next prev parent reply other threads:[~2024-01-24 10:15 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-23 14:33 [REGRESSION] Failed buffer allocation in Tegra fbdev diogo.ivo
2024-01-23 15:15 ` Jason Gunthorpe
2024-01-23 18:44 ` Diogo Ivo
2024-01-24 10:15 ` Jon Hunter [this message]
2024-01-24 10:30 ` Diogo Ivo
2024-01-24 10:36 ` Jon Hunter
2024-01-26 19:51 ` Jason Gunthorpe
2024-01-29 12:07 ` Diogo Ivo
2024-01-24 9:13 ` Diogo Ivo
2024-01-24 10:17 ` Jon Hunter
2024-01-24 11:46 ` Robin Murphy
2024-01-24 12:56 ` Diogo Ivo
2024-02-29 14:50 ` Jon Hunter
2024-02-29 16:46 ` Thierry Reding
2024-01-24 17:03 ` Jason Gunthorpe
2024-02-08 1:22 ` Robin Murphy
2024-02-08 2:31 ` Jason Gunthorpe
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b0334c5e-3a6c-4b58-b525-e72bed8899b3@nvidia.com \
--to=jonathanh@nvidia.com \
--cc=baolu.lu@linux.intel.com \
--cc=diogo.ivo@tecnico.ulisboa.pt \
--cc=iommu@lists.linux.dev \
--cc=jgg@ziepe.ca \
--cc=joro@8bytes.org \
--cc=jroedel@suse.de \
--cc=jsnitsel@redhat.com \
--cc=linux-tegra@vger.kernel.org \
--cc=regressions@lists.linux.dev \
--cc=robin.murphy@arm.com \
--cc=thierry.reding@gmail.com \
--cc=vdumpa@nvidia.com \
--cc=will@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).