Linux-USB Archive on lore.kernel.org
 help / color / Atom feed
* I get a new Thunderbolt domain UUID on every boot
@ 2020-07-05 21:52 Kenneth R. Crudup
  2020-07-20 10:07 ` Mika Westerberg
  0 siblings, 1 reply; 5+ messages in thread
From: Kenneth R. Crudup @ 2020-07-05 21:52 UTC (permalink / raw)
  To: mika.westerberg; +Cc: linux-usb, Kenneth R. Crudup


Does anyone know why my machine (Dell XPS 7390 2-in-1, i7-1065G7 CPU) gets a
new Thunderbolt domain UUID on every boot?

----
$ sudo boltctl domains | wc -l
3320
$ sudo boltctl domains | tail -20
 o domain 815fe31a-99fd-8680-ffff-ffffffffffff
   |- bootacl:  0/0
   `- security: unknown

 o domain 703bafdf-6a33-8680-ffff-ffffffffffff
   |- bootacl:  0/0
   `- security: unknown

 o domain 91c55303-9244-8680-ffff-ffffffffffff
   |- bootacl:  0/0
   `- security: unknown

 * domain0 800e6342-557e-8680-ffff-ffffffffffff
   |- bootacl:  0/0
   `- security: iommu

 * domain1 9128366f-c373-8680-ffff-ffffffffffff
   |- bootacl:  0/0
   `- security: iommu
$
----
$ egrep '(THUNDER|USB4)' .config
# CONFIG_MDIO_THUNDER is not set
CONFIG_USB4_NET=m
# CONFIG_USB_HSIC_USB4604 is not set
CONFIG_INTEL_WMI_THUNDERBOLT=y
CONFIG_USB4=y
$
----
I did see this, too:
$  dmesg | fgrep -i thunderbolt
[    1.114106] thunderbolt 0000:00:0d.2: 0: uid crc8 mismatch (expected: 0x8e, got: 0xe7)
[    1.657866] thunderbolt 0000:00:0d.3: 0: uid crc8 mismatch (expected: 0x8e, got: 0xe7)
----

Thanks,

	-Kenny

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Orange County CA

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: I get a new Thunderbolt domain UUID on every boot
  2020-07-05 21:52 I get a new Thunderbolt domain UUID on every boot Kenneth R. Crudup
@ 2020-07-20 10:07 ` Mika Westerberg
  2020-07-20 12:43   ` Kenneth R. Crudup
  0 siblings, 1 reply; 5+ messages in thread
From: Mika Westerberg @ 2020-07-20 10:07 UTC (permalink / raw)
  To: Kenneth R. Crudup; +Cc: linux-usb

Hi,

On Sun, Jul 05, 2020 at 02:52:04PM -0700, Kenneth R. Crudup wrote:
> 
> Does anyone know why my machine (Dell XPS 7390 2-in-1, i7-1065G7 CPU) gets a
> new Thunderbolt domain UUID on every boot?
> 
> ----
> $ sudo boltctl domains | wc -l
> 3320
> $ sudo boltctl domains | tail -20
>  o domain 815fe31a-99fd-8680-ffff-ffffffffffff
>    |- bootacl:  0/0
>    `- security: unknown
> 
>  o domain 703bafdf-6a33-8680-ffff-ffffffffffff
>    |- bootacl:  0/0
>    `- security: unknown
> 
>  o domain 91c55303-9244-8680-ffff-ffffffffffff
>    |- bootacl:  0/0
>    `- security: unknown
> 
>  * domain0 800e6342-557e-8680-ffff-ffffffffffff
>    |- bootacl:  0/0
>    `- security: iommu
> 
>  * domain1 9128366f-c373-8680-ffff-ffffffffffff
>    |- bootacl:  0/0
>    `- security: iommu
> $

It is done on purpose. The BIOS generates new UUID every boot.

> ----
> $ egrep '(THUNDER|USB4)' .config
> # CONFIG_MDIO_THUNDER is not set
> CONFIG_USB4_NET=m
> # CONFIG_USB_HSIC_USB4604 is not set
> CONFIG_INTEL_WMI_THUNDERBOLT=y
> CONFIG_USB4=y
> $
> ----
> I did see this, too:
> $  dmesg | fgrep -i thunderbolt
> [    1.114106] thunderbolt 0000:00:0d.2: 0: uid crc8 mismatch (expected: 0x8e, got: 0xe7)
> [    1.657866] thunderbolt 0000:00:0d.3: 0: uid crc8 mismatch (expected: 0x8e, got: 0xe7)

This is harmless. Host router (the TBT host controller) on ICL does not
have DROM so reading UID returns 0. Let me see if we can simply skip
this check on such systems.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: I get a new Thunderbolt domain UUID on every boot
  2020-07-20 10:07 ` Mika Westerberg
@ 2020-07-20 12:43   ` Kenneth R. Crudup
  2020-07-21 10:28     ` Mika Westerberg
  0 siblings, 1 reply; 5+ messages in thread
From: Kenneth R. Crudup @ 2020-07-20 12:43 UTC (permalink / raw)
  To: Mika Westerberg; +Cc: linux-usb


Thank you for getting back to me.

On Mon, 20 Jul 2020, Mika Westerberg wrote:

> It is done on purpose. The BIOS generates new UUID every boot.

I see. This leads to a couple of follow-on questions, then:

There are two "competing" Linux packages for controlling/admining TB,
"bolt" (which includes "boltctl") and "thunderbolt-tools" (which includes
"tbtadm". I used to have both installed but I think they sometimes get
in each other's way and settled on "thunderbolt-tools" as "bolt" was
leaving large numbers of UUID-related files over time.

- Do you have a preference for either?

I've noticed that I can't seem to get IOMMU protection unless I boot
with Secure Boot on in my BIOS (which I never use, as I don't sign my
own kernels). IOMMU/DMAR is on in my kernel.

- Is there any way to force this, or is this expected? I don't get the
"iommu_dma_protection" set on my controllers either.

Finally, every now and then on a reboot, my TB dock "flaps" (repeatedly
connects and disconnects) and I have to either connect/reconnect the TB3
cable, or initate a power-off. I suspect this is a BIOS bug (I also have
to set "intel_iommu=igfx_off" otherwise I get "DMA Hardware is malfunctioning"
errors on power-down (or hibernate)).

- Have you seen this before or have any ideas?

Thanks,

	-Kenny

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Orange County CA

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: I get a new Thunderbolt domain UUID on every boot
  2020-07-20 12:43   ` Kenneth R. Crudup
@ 2020-07-21 10:28     ` Mika Westerberg
  2020-10-17 21:58       ` Kenneth R. Crudup
  0 siblings, 1 reply; 5+ messages in thread
From: Mika Westerberg @ 2020-07-21 10:28 UTC (permalink / raw)
  To: Kenneth R. Crudup; +Cc: linux-usb

On Mon, Jul 20, 2020 at 05:43:05AM -0700, Kenneth R. Crudup wrote:
> 
> Thank you for getting back to me.
> 
> On Mon, 20 Jul 2020, Mika Westerberg wrote:
> 
> > It is done on purpose. The BIOS generates new UUID every boot.
> 
> I see. This leads to a couple of follow-on questions, then:
> 
> There are two "competing" Linux packages for controlling/admining TB,
> "bolt" (which includes "boltctl") and "thunderbolt-tools" (which includes
> "tbtadm". I used to have both installed but I think they sometimes get
> in each other's way and settled on "thunderbolt-tools" as "bolt" was
> leaving large numbers of UUID-related files over time.
> 
> - Do you have a preference for either?

bolt

I suggest to report this to bolt upstream.

> I've noticed that I can't seem to get IOMMU protection unless I boot
> with Secure Boot on in my BIOS (which I never use, as I don't sign my
> own kernels). IOMMU/DMAR is on in my kernel.
> 
> - Is there any way to force this, or is this expected? I don't get the
> "iommu_dma_protection" set on my controllers either.

It depends on the system. The kernel expects that you have ACPI DMAR
table with bit 2 set (platform opt-in). Some systems I have seen have
"Native Security for Thunderbolt" or similar BIOS option that can be
used to enable this and others have it enabled by default.

> Finally, every now and then on a reboot, my TB dock "flaps" (repeatedly
> connects and disconnects) and I have to either connect/reconnect the TB3
> cable, or initate a power-off. I suspect this is a BIOS bug (I also have
> to set "intel_iommu=igfx_off" otherwise I get "DMA Hardware is malfunctioning"
> errors on power-down (or hibernate)).
> 
> - Have you seen this before or have any ideas?

Not seen before, but IIRC some laptops require more power than what the
dock can supply and the symptoms for that are similar that the
connection goes down and then back up several times.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: I get a new Thunderbolt domain UUID on every boot
  2020-07-21 10:28     ` Mika Westerberg
@ 2020-10-17 21:58       ` Kenneth R. Crudup
  0 siblings, 0 replies; 5+ messages in thread
From: Kenneth R. Crudup @ 2020-10-17 21:58 UTC (permalink / raw)
  To: Mika Westerberg; +Cc: linux-usb


> > Finally, every now and then on a reboot, my TB dock "flaps" (repeatedly
> > connects and disconnects) and I have to either connect/reconnect the TB3
> > cable, or initate a power-off. I suspect this is a BIOS bug ...
> > Have you seen this before or have any ideas?

On Tue, 21 Jul 2020, Mika Westerberg wrote:

> Not seen before, but IIRC some laptops require more power than what the
> dock can supply and the symptoms for that are similar that the
> connection goes down and then back up several times.

BTW, apparently this was a Dell BIOS bug- they'd released version 1.6.2 a
couple of weeks ago for my machine (XPS 7390 2-in-1) and I haven't seen it
since. Considering this thing's a year old, I'm not confident in Dell's
ability to be proactive about bugs (but I guess in their defense that did
solve the the last issue I had with this thing).

	-Kenny

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Orange County CA

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, back to index

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-07-05 21:52 I get a new Thunderbolt domain UUID on every boot Kenneth R. Crudup
2020-07-20 10:07 ` Mika Westerberg
2020-07-20 12:43   ` Kenneth R. Crudup
2020-07-21 10:28     ` Mika Westerberg
2020-10-17 21:58       ` Kenneth R. Crudup

Linux-USB Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-usb/0 linux-usb/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-usb linux-usb/ https://lore.kernel.org/linux-usb \
		linux-usb@vger.kernel.org
	public-inbox-index linux-usb

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-usb


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git