From: Vasant Hegde <vasant.hegde@amd.com>
To: Paul Menzel <pmenzel@molgen.mpg.de>
Cc: "Jörg Rödel" <joro@8bytes.org>,
iommu@lists.linux.dev, it+linux-iommu@molgen.mpg.de
Subject: Re: iommu ivhd0: AMD-Vi: Event logged [INVALID_DEVICE_REQUEST device=0000:00:00.0 pasid=0x00000 address=0xfffffffdf8000000 flags=0x0a00]
Date: Mon, 11 Mar 2024 18:50:07 +0530 [thread overview]
Message-ID: <5d469b74-6a47-b478-3360-922e619f567e@amd.com> (raw)
In-Reply-To: <2b424597-0a42-4e89-aacf-6c56b190c520@molgen.mpg.de>
Hi Paul,
On 3/7/2024 1:29 PM, Paul Menzel wrote:
> Dear Vasant,
>
>
> Thank you very much for the quick response.
>
>
> Am 07.03.24 um 06:33 schrieb Vasant Hegde:
>
>> On 3/6/2024 11:39 PM, Paul Menzel wrote:
>
>>> Am 06.03.24 um 17:24 schrieb Vasant Hegde:
>>>
>>>> On 3/5/2024 8:18 PM, Paul Menzel wrote:
>>>
>>>>> On a Dell OptiPlex 5055 Ryzen CPU/0P03DX, BIOS 1.1.20
>>>>> 05/31/2019, Linux 5.15.112, 5.15.131, 6.1.53, 6.6.12, 6.6.20,
>>>>> 6.8-rc6+ and 6.8-rc7+ log the error below:
>>>>>
>>>>> iommu ivhd0: AMD-Vi: Event logged [INVALID_DEVICE_REQUEST
>>>>> device=0000:00:00.0 pasid=0x00000 address=0xfffffffdf8000000 flags=0x0a00]
>>>>
>>>> I did look into attached dmesg. I see there is a mismatch in
>>>> IOMMU Feature bits. This is usually a BIOS bug. Can you check
>>>> with FW folks? Specially I am not sure why they have enabled bit
>>>> 62.
>>>
>>> Unfortunately, this device is out of support, so Dell won’t look
>>> into it.
>>>
>>>> [ 3.527391] AMD-Vi: [Firmware Warn]: EFR mismatch. Use IVHD EFR
>>>> (0xf77ef22294ada : 0x400f77ef22294ada), EFR2 (0x0 : 0x0).
>>>
>>> This is a known issue with the Dell firmware at least of this
>>> version [1].
>>>
>>>> You are hitting fault with device ID 00:00.0... I am not sure
>>>> whether this device is managed by IOMMU or not. Can you add
>>>> `amd_iommu_dump=1` to kernel command line and boot the system. It
>>>> should tell us which devices are managed by the IOMMU.
>>>
>>> Ah, thank you for the hint. Please find the Linux messages with
>>> this parameter attached.
>>
>> Thanks Paul. Looking into the attached dmesg, looks like 00:00.0 is
>> not managed by IOMMU. Hence we are getting above error in dmesg. This
>> needs FW side fix.
>
> Interesting. Could you please share from what line in the dump you deduce your
> findings?
[ 1.528983] AMD-Vi: AMD-Vi: Using IVHD type 0x11
[ 1.533624] AMD-Vi: AMD-Vi: device: 0000:00:00.2 cap: 0040 flags: b0 info 0000
[ 1.540874] AMD-Vi: AMD-Vi: mmio-addr: 00000000fc000000
[ 1.546728] AMD-Vi: AMD-Vi: PCI segment : 0x0, last bdf : 0xffff
[ 1.553012] AMD-Vi: AMD-Vi: DEV_SELECT_RANGE_START devid: 0000:00:01.0
flags: 00
[ 1.560690] AMD-Vi: AMD-Vi: DEV_RANGE_END devid: 0000:ff:1f.6
[ 1.567420] AMD-Vi: AMD-Vi: DEV_ALIAS_RANGE devid: 0000:ff:00.0 flags: 00
devid_to: 0000:00:14.4
[ 1.576574] AMD-Vi: AMD-Vi: DEV_RANGE_END devid: 0000:ff:1f.7
[ 1.582691] AMD-Vi: AMD-Vi: DEV_SPECIAL(HPET[0]) devid: 0000:00:14.0
[ 1.589330] AMD-Vi: AMD-Vi: DEV_SPECIAL(IOAPIC[32]) devid: 0000:00:14.0
[ 1.596225] AMD-Vi: AMD-Vi: DEV_SPECIAL(IOAPIC[33]) devid: 0000:00:00.1
With this we will know what device ID ranges that IOMMU manages.
> Should 00:00.0 be normally managed by IOMMU? Also, excuse my IOMMU
> ignorance, but can you please elaborate, what is happening?
iommu ivhd0: AMD-Vi: Event logged [INVALID_DEVICE_REQUEST device=0000:00:00.0
pasid=0x00000 address=0xfffffffdf8000000 flags=0x0a00]
Basically 00:00.0 is not managed by IOMMU.. but it trying to send request. In
this case its posted interrupt write (I confirmed that based on flags). So IOMMU
logged INVALID_DEVICE_REQUEST.
>
> The other interesting detail was, that it’s not happening when rebooting the
> system. The IOMMU dump in the attached log is the same, but the error is gone.
May be device is not sending any messages. So we didn't see any error message
from IOMMU.
-Vasant
prev parent reply other threads:[~2024-03-11 13:20 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-05 14:48 iommu ivhd0: AMD-Vi: Event logged [INVALID_DEVICE_REQUEST device=0000:00:00.0 pasid=0x00000 address=0xfffffffdf8000000 flags=0x0a00] Paul Menzel
2024-03-06 16:24 ` Vasant Hegde
2024-03-06 18:09 ` Paul Menzel
2024-03-07 5:33 ` Vasant Hegde
2024-03-07 7:59 ` Paul Menzel
2024-03-11 13:20 ` Vasant Hegde [this message]
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=5d469b74-6a47-b478-3360-922e619f567e@amd.com \
--to=vasant.hegde@amd.com \
--cc=iommu@lists.linux.dev \
--cc=it+linux-iommu@molgen.mpg.de \
--cc=joro@8bytes.org \
--cc=pmenzel@molgen.mpg.de \
/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).