Re: [PATCH] xen/arm: smmuv1: remove iommu group when deassign a device

[Mykyta Poturai <mykyta.poturai@xxxxxxxxx>]

> Thanks for testing the patch.
>> But not fixed the "Unexpected Global fault" that occasionally happens when 
>> destroying
>> the domain with an actively working GPU. Although, I am not sure if this 
>> issue
>> is relevant here.
>
> Can you please if possible share the more details and logs so that I can look 
> if this issue is relevant here ?

So in my setup I have a board with IMX8 chip and 2 core Vivante GPU. GPU is 
split between domains.
One core goes to Dom0 and one to DomU.

Steps to trigger this issue:
1. Start DomU
2. Start wayland and glmark2-es2-wayland inside DomU
3. Destroy DomU

Sometimes it destroys fine but roughly 1 of 8 times I get logs like this:

root@dom0:~# xl dest DomU
[12725.412940] xenbr0: port 1(vif8.0) entered disabled state
[12725.671033] xenbr0: port 1(vif8.0) entered disabled state
[12725.689923] device vif8.0 left promiscuous mode
[12725.696736] xenbr0: port 1(vif8.0) entered disabled state
[12725.696989] audit: type=1700 audit(1616594240.068:39): dev=vif8.0 prom=0 
old_prom=256 auid=4294967295 uid=0 gid=0 ses=4294967295
(XEN) smmu: /iommu@51400000: Unexpected global fault, this could be serious
(XEN) smmu: /iommu@51400000:    GFSR 0x00000001, GFSYNR0 0x00000004, GFSYNR1 
0x00001055, GFSYNR2 0x00000000

My guess is that this happens because GPU continues to access memory when the 
context is already invalidated,
and therefore triggers the "Invalid context fault".

Regards,
Mykyta