AMD IOMMU + SME + amdgpu regression

All of lore.kernel.org
 help / color / mirror / Atom feed

* AMD IOMMU + SME + amdgpu regression
       [not found] <1591915710.rakbpzst8h.none.ref@localhost>
  2020-06-11 23:05   ` Alex Xu (Hello71)
@ 2020-06-11 23:05   ` Alex Xu (Hello71)
  0 siblings, 0 replies; 10+ messages in thread
From: Alex Xu (Hello71) @ 2020-06-11 23:05 UTC (permalink / raw)
  To: Joerg Roedel, linux-kernel, David Rientjes, Christoph Hellwig
  Cc: Will Deacon, Robin Murphy, Marek Szyprowski, Kukjin Kim,
	Krzysztof Kozlowski, David Woodhouse, Lu Baolu, Andy Gross,
	Bjorn Andersson, Matthias Brugger, Rob Clark, Heiko Stuebner,
	Gerald Schaefer, Thierry Reding, Jonathan Hunter,
	Jean-Philippe Brucker, Daniel Drake, jonathan.derrick,
	linux-samsung-soc, linux-arm-msm, linux-mediatek, linux-rockchip,
	linux-s390, linux-tegra, virtualization, Joerg Roedel

Hi,

amdgpu + IOMMU + SME is now working for me on 5.7, yay! But, it is 
broken on torvalds master, boo. On boot, depending on which exact commit 
I test, it either hangs immediately (with built-in driver, before 
starting initramfs), displays some errors then hangs, or spams the 
screen with many amdgpu errors.

I bisected the black screen hang to:

commit dce8d6964ebdb333383bacf5e7ab8c27df151218
Author: Joerg Roedel <jroedel@suse.de>
Date:   Wed Apr 29 15:36:53 2020 +0200

    iommu/amd: Convert to probe/release_device() call-backs

    Convert the AMD IOMMU Driver to use the probe_device() and
    release_device() call-backs of iommu_ops, so that the iommu core code
    does the group and sysfs setup.

    Signed-off-by: Joerg Roedel <jroedel@suse.de>
    Link: https://lore.kernel.org/r/20200429133712.31431-16-joro@8bytes.org
    Signed-off-by: Joerg Roedel <jroedel@suse.de>

Testing torvalds master (623f6dc593) with the containing merge 
(98bdc74b36) plus the DMA mapping merge (4e94d08734) reverted allows 
amdgpu + IOMMU + SME to once again work.

I think that nobody is really working on amdgpu + SME, but it would be a 
shame if it was supported and then incidentally broken by a small 
change.

I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
understand this code at all, but let me know what I can do to 
troubleshoot.

Thanks,
Alex.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* AMD IOMMU + SME + amdgpu regression
@ 2020-06-11 23:05   ` Alex Xu (Hello71)
  0 siblings, 0 replies; 10+ messages in thread
From: Alex Xu (Hello71) @ 2020-06-11 23:05 UTC (permalink / raw)
  To: Joerg Roedel, linux-kernel, David Rientjes, Christoph Hellwig
  Cc: Will Deacon, Robin Murphy, Marek Szyprowski, Kukjin Kim,
	Krzysztof Kozlowski, David Woodhouse, Lu Baolu, Andy Gross,
	Bjorn Andersson, Matthias Brugger, Rob Clark, Heiko Stuebner,
	Gerald Schaefer, Thierry Reding, Jonathan Hunter,
	Jean-Philippe Brucker, Daniel Drake, jonathan.derrick,
	linux-samsung-soc, linux-arm-msm

Hi,

amdgpu + IOMMU + SME is now working for me on 5.7, yay! But, it is 
broken on torvalds master, boo. On boot, depending on which exact commit 
I test, it either hangs immediately (with built-in driver, before 
starting initramfs), displays some errors then hangs, or spams the 
screen with many amdgpu errors.

I bisected the black screen hang to:

commit dce8d6964ebdb333383bacf5e7ab8c27df151218
Author: Joerg Roedel <jroedel@suse.de>
Date:   Wed Apr 29 15:36:53 2020 +0200

    iommu/amd: Convert to probe/release_device() call-backs

    Convert the AMD IOMMU Driver to use the probe_device() and
    release_device() call-backs of iommu_ops, so that the iommu core code
    does the group and sysfs setup.

    Signed-off-by: Joerg Roedel <jroedel@suse.de>
    Link: https://lore.kernel.org/r/20200429133712.31431-16-joro@8bytes.org
    Signed-off-by: Joerg Roedel <jroedel@suse.de>

Testing torvalds master (623f6dc593) with the containing merge 
(98bdc74b36) plus the DMA mapping merge (4e94d08734) reverted allows 
amdgpu + IOMMU + SME to once again work.

I think that nobody is really working on amdgpu + SME, but it would be a 
shame if it was supported and then incidentally broken by a small 
change.

I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
understand this code at all, but let me know what I can do to 
troubleshoot.

Thanks,
Alex.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* AMD IOMMU + SME + amdgpu regression
@ 2020-06-11 23:05   ` Alex Xu (Hello71)
  0 siblings, 0 replies; 10+ messages in thread
From: Alex Xu (Hello71) @ 2020-06-11 23:05 UTC (permalink / raw)
  To: Joerg Roedel, linux-kernel, David Rientjes, Christoph Hellwig
  Cc: Heiko Stuebner, Bjorn Andersson, linux-tegra, Thierry Reding,
	Daniel Drake, Will Deacon, Marek Szyprowski,
	Jean-Philippe Brucker, linux-samsung-soc, Krzysztof Kozlowski,
	Jonathan Hunter, linux-rockchip, Andy Gross, Gerald Schaefer,
	Joerg Roedel, linux-s390, linux-arm-msm, linux-mediatek,
	Matthias Brugger, virtualization, jonathan.derrick,
	David Woodhouse, Rob Clark, Kukjin Kim, Robin Murphy, Lu Baolu

Hi,

amdgpu + IOMMU + SME is now working for me on 5.7, yay! But, it is 
broken on torvalds master, boo. On boot, depending on which exact commit 
I test, it either hangs immediately (with built-in driver, before 
starting initramfs), displays some errors then hangs, or spams the 
screen with many amdgpu errors.

I bisected the black screen hang to:

commit dce8d6964ebdb333383bacf5e7ab8c27df151218
Author: Joerg Roedel <jroedel@suse.de>
Date:   Wed Apr 29 15:36:53 2020 +0200

    iommu/amd: Convert to probe/release_device() call-backs

    Convert the AMD IOMMU Driver to use the probe_device() and
    release_device() call-backs of iommu_ops, so that the iommu core code
    does the group and sysfs setup.

    Signed-off-by: Joerg Roedel <jroedel@suse.de>
    Link: https://lore.kernel.org/r/20200429133712.31431-16-joro@8bytes.org
    Signed-off-by: Joerg Roedel <jroedel@suse.de>

Testing torvalds master (623f6dc593) with the containing merge 
(98bdc74b36) plus the DMA mapping merge (4e94d08734) reverted allows 
amdgpu + IOMMU + SME to once again work.

I think that nobody is really working on amdgpu + SME, but it would be a 
shame if it was supported and then incidentally broken by a small 
change.

I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
understand this code at all, but let me know what I can do to 
troubleshoot.

Thanks,
Alex.

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
  2020-06-11 23:05   ` Alex Xu (Hello71)
@ 2020-06-22 10:02     ` Joerg Roedel
  -1 siblings, 0 replies; 10+ messages in thread
From: Joerg Roedel @ 2020-06-22 10:02 UTC (permalink / raw)
  To: Alex Xu (Hello71)
  Cc: Joerg Roedel, linux-kernel, David Rientjes, Christoph Hellwig,
	Will Deacon, Robin Murphy, Marek Szyprowski, Kukjin Kim,
	Krzysztof Kozlowski, David Woodhouse, Lu Baolu, Andy Gross,
	Bjorn Andersson, Matthias Brugger, Rob Clark, Heiko Stuebner,
	Gerald Schaefer, Thierry Reding, Jonathan Hunter,
	Jean-Philippe Brucker, Daniel Drake, jonathan.derrick,
	linux-samsung-soc, linux-arm-msm, linux-mediatek, linux-rockchip,
	linux-s390, linux-tegra, virtualization

Hi Alex,

On Thu, Jun 11, 2020 at 07:05:21PM -0400, Alex Xu (Hello71) wrote:
> I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
> understand this code at all, but let me know what I can do to 
> troubleshoot.

Does it boot without SME enabled?


Regards,

	Joerg

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
@ 2020-06-22 10:02     ` Joerg Roedel
  0 siblings, 0 replies; 10+ messages in thread
From: Joerg Roedel @ 2020-06-22 10:02 UTC (permalink / raw)
  To: Alex Xu (Hello71)
  Cc: Heiko Stuebner, Bjorn Andersson, linux-tegra, Thierry Reding,
	Daniel Drake, Will Deacon, Christoph Hellwig, Marek Szyprowski,
	Jean-Philippe Brucker, linux-samsung-soc, Joerg Roedel,
	Krzysztof Kozlowski, Jonathan Hunter, linux-rockchip, Kukjin Kim,
	Andy Gross, David Rientjes, linux-s390, Gerald Schaefer,
	linux-arm-msm, linux-mediatek, Matthias Brugger, virtualization,
	jonathan.derrick

Hi Alex,

On Thu, Jun 11, 2020 at 07:05:21PM -0400, Alex Xu (Hello71) wrote:
> I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
> understand this code at all, but let me know what I can do to 
> troubleshoot.

Does it boot without SME enabled?


Regards,

	Joerg

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
@ 2020-06-22 15:30       ` Alex Xu (Hello71)
  0 siblings, 0 replies; 10+ messages in thread
From: Alex Xu (Hello71) @ 2020-06-22 15:30 UTC (permalink / raw)
  To: Joerg Roedel
  Cc: Andy Gross, Lu Baolu, Bjorn Andersson, Daniel Drake,
	David Woodhouse, Gerald Schaefer, Christoph Hellwig,
	Heiko Stuebner, Jean-Philippe Brucker, jonathan.derrick,
	Jonathan Hunter, Joerg Roedel, Kukjin Kim, Krzysztof Kozlowski,
	linux-arm-msm, linux-kernel, linux-mediatek, linux-rockchip,
	linux-s390, linux-samsung-soc, linux-tegra, Matthias Brugger,
	Marek Szyprowski, David Rientjes, Rob Clark, Robin Murphy,
	Thierry Reding, virtualization, Will Deacon

Excerpts from Joerg Roedel's message of June 22, 2020 6:02 am:
> Hi Alex,
> 
> On Thu, Jun 11, 2020 at 07:05:21PM -0400, Alex Xu (Hello71) wrote:
>> I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
>> understand this code at all, but let me know what I can do to 
>> troubleshoot.
> 
> Does it boot without SME enabled?
> 
> 
> Regards,
> 
> 	Joerg
> 

Yes, it works with SME off with dbed452a078 ("dma-pool: decouple 
DMA_REMAP from DMA_COHERENT_POOL") applied.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
@ 2020-06-22 15:30       ` Alex Xu (Hello71)
  0 siblings, 0 replies; 10+ messages in thread
From: Alex Xu (Hello71) @ 2020-06-22 15:30 UTC (permalink / raw)
  To: Joerg Roedel
  Cc: Andy Gross, Lu Baolu, Bjorn Andersson, Daniel Drake,
	David Woodhouse, Gerald Schaefer, Christoph Hellwig,
	Heiko Stuebner, Jean-Philippe Brucker,
	jonathan.derrick-ral2JQCrhuEAvxtiuMwx3w, Jonathan Hunter,
	Joerg Roedel, Kukjin Kim, Krzysztof Kozlowski,
	linux-arm-msm-u79uwXL29TY76Z2rM5mHXA,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA,
	linux-mediatek-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	linux-rockchip-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	linux-s390-u79uwXL29TY76Z2rM5mHXA,
	linux-samsung-soc-u79uwXL29TY76Z2rM5mHXA,
	linux-tegra-u79uwXL29TY76Z2rM5mHXA, Matthias Brugger <matt>

Excerpts from Joerg Roedel's message of June 22, 2020 6:02 am:
> Hi Alex,
> 
> On Thu, Jun 11, 2020 at 07:05:21PM -0400, Alex Xu (Hello71) wrote:
>> I am using an ASRock B450 Pro4 with Ryzen 1600 and ASUS RX 480. I don't 
>> understand this code at all, but let me know what I can do to 
>> troubleshoot.
> 
> Does it boot without SME enabled?
> 
> 
> Regards,
> 
> 	Joerg
> 

Yes, it works with SME off with dbed452a078 ("dma-pool: decouple 
DMA_REMAP from DMA_COHERENT_POOL") applied.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
  2020-06-22 15:30       ` Alex Xu (Hello71)
  (?)
@ 2020-07-15  9:17         ` Joerg Roedel
  -1 siblings, 0 replies; 10+ messages in thread
From: Joerg Roedel @ 2020-07-15  9:17 UTC (permalink / raw)
  To: Alex Xu (Hello71)
  Cc: Joerg Roedel, Andy Gross, Lu Baolu, Bjorn Andersson, Daniel Drake,
	David Woodhouse, Gerald Schaefer, Christoph Hellwig,
	Heiko Stuebner, Jean-Philippe Brucker, jonathan.derrick,
	Jonathan Hunter, Kukjin Kim, Krzysztof Kozlowski, linux-arm-msm,
	linux-kernel, linux-mediatek, linux-rockchip, linux-s390,
	linux-samsung-soc, linux-tegra, Matthias Brugger,
	Marek Szyprowski, David Rientjes, Rob Clark, Robin Murphy,
	Thierry Reding, virtualization, Will Deacon

On Mon, Jun 22, 2020 at 11:30:04AM -0400, Alex Xu (Hello71) wrote:
> Yes, it works with SME off with dbed452a078 ("dma-pool: decouple 
> DMA_REMAP from DMA_COHERENT_POOL") applied.

Okay, I can reproduce the problem on my Ryzen System, and the boot log
shows various warnings/bugs from the amdgpu driver. I think this should
be looked at by the AMDGPU folks first, as I didn't really got far
looking into the GPU drivers code.

Regards,

	Joerg

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
@ 2020-07-15  9:17         ` Joerg Roedel
  0 siblings, 0 replies; 10+ messages in thread
From: Joerg Roedel @ 2020-07-15  9:17 UTC (permalink / raw)
  To: Alex Xu (Hello71)
  Cc: Joerg Roedel, Andy Gross, Lu Baolu, Bjorn Andersson, Daniel Drake,
	David Woodhouse, Gerald Schaefer, Christoph Hellwig,
	Heiko Stuebner, Jean-Philippe Brucker,
	jonathan.derrick-ral2JQCrhuEAvxtiuMwx3w, Jonathan Hunter,
	Kukjin Kim, Krzysztof Kozlowski,
	linux-arm-msm-u79uwXL29TY76Z2rM5mHXA,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA,
	linux-mediatek-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	linux-rockchip-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
	linux-s390-u79uwXL29TY76Z2rM5mHXA,
	linux-samsung-soc-u79uwXL29TY76Z2rM5mHXA,
	linux-tegra-u79uwXL29TY76Z2rM5mHXA, Matthias Brugger <matt>

On Mon, Jun 22, 2020 at 11:30:04AM -0400, Alex Xu (Hello71) wrote:
> Yes, it works with SME off with dbed452a078 ("dma-pool: decouple 
> DMA_REMAP from DMA_COHERENT_POOL") applied.

Okay, I can reproduce the problem on my Ryzen System, and the boot log
shows various warnings/bugs from the amdgpu driver. I think this should
be looked at by the AMDGPU folks first, as I didn't really got far
looking into the GPU drivers code.

Regards,

	Joerg

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: AMD IOMMU + SME + amdgpu regression
@ 2020-07-15  9:17         ` Joerg Roedel
  0 siblings, 0 replies; 10+ messages in thread
From: Joerg Roedel @ 2020-07-15  9:17 UTC (permalink / raw)
  To: Alex Xu (Hello71)
  Cc: Heiko Stuebner, Bjorn Andersson, linux-tegra, Thierry Reding,
	Daniel Drake, Will Deacon, Christoph Hellwig, Marek Szyprowski,
	Jean-Philippe Brucker, linux-samsung-soc, Krzysztof Kozlowski,
	Jonathan Hunter, linux-rockchip, Andy Gross, Gerald Schaefer,
	David Rientjes, Joerg Roedel, linux-s390, linux-arm-msm,
	linux-mediatek, Matthias Brugger, virtualization,
	jonathan.derrick, Robin Murphy, linux-kernel, Rob Clark,
	Kukjin Kim, David Woodhouse, Lu Baolu

On Mon, Jun 22, 2020 at 11:30:04AM -0400, Alex Xu (Hello71) wrote:
> Yes, it works with SME off with dbed452a078 ("dma-pool: decouple 
> DMA_REMAP from DMA_COHERENT_POOL") applied.

Okay, I can reproduce the problem on my Ryzen System, and the boot log
shows various warnings/bugs from the amdgpu driver. I think this should
be looked at by the AMDGPU folks first, as I didn't really got far
looking into the GPU drivers code.

Regards,

	Joerg

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2020-07-15  9:17 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <1591915710.rakbpzst8h.none.ref@localhost>
2020-06-11 23:05 ` AMD IOMMU + SME + amdgpu regression Alex Xu (Hello71)
2020-06-11 23:05   ` Alex Xu (Hello71)
2020-06-11 23:05   ` Alex Xu (Hello71)
2020-06-22 10:02   ` Joerg Roedel
2020-06-22 10:02     ` Joerg Roedel
2020-06-22 15:30     ` Alex Xu (Hello71)
2020-06-22 15:30       ` Alex Xu (Hello71)
2020-07-15  9:17       ` Joerg Roedel
2020-07-15  9:17         ` Joerg Roedel
2020-07-15  9:17         ` Joerg Roedel

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.