Re: [Xen-devel] [Not Xen] Facing issues running guest OS on custom Hypervisor

[Julien Grall <julien.grall@xxxxxxx>]

On 20/03/18 20:24, Brijen Raval wrote:
> Hello Julien,
Hello,

> As requested I am moving the conversation to email from IRC
>
> To summarize my setup:
>
> 1. I am running a custom kernel on QEMU ARM64(without KVM) on my linux 
> machine
> 2. I have my custom implementation of Hypervisor
> 3. I am trying to run the same custom kernel as guest OS on top of my 
> Hypervisor
> - I am able to boot my kernel to shell on QEMU
> - I am able to start my guest OS
> - From the logs I see that my guest OS finishes booting up, I can see 
> the $sign for the shell and then it goes into idle state, but I cannot 
> use the shell
> To debug further I enabled tracing in QEMU and printed the exceptions to 
> understand what state is my guest in
> Before I paste some logs here, some more information about my system
>
> IRQ 30 is the physical timer interrupt of my host OS running on QEMU
> IRQ 27 is the virtual timer interrupt of my guest OS
>
> I have added some extra logging in QEMU to print out the VTTBR so as to 
> understand where the exception is coming from
>  From 1st Attachment (GIC 1) I observe that every once in a while the 
> phys timer interrupt occurs (IRQ 30) and its handled by the host VM, and 
> then after about 10-20 times the virtirq 27 level changes to 1 and back 
> to 0 again and again..this is how its looping currently after boot up
>
> Adding a 2nd attachment with extra logging of traps of exceptions as 
> well. It just shows 2 different IRQ exceptions taken, one with VTTBR = 0 
> (IRQ30) and other with a VTTBR value of the guest (IRQ 27, since the irq 
> 27 level is changed to 1 just before it..
>
> Any idea what am I missing, and why my guest OS is not handling the 
> pending interrupt.
> Upon receiving the IRQ 27, I do set the HCR_EL2.VI <http://HCR_EL2.VI> 
> bit to 1 to signal the guest about a pending virual interrupt but I dont 
> think thats working.
HCR_EL2.VI should not be necessary is you are using the GIC HW 
virtualization extension. Can you confirm you are using it?
If so, I would recommend to look at the content of the LRs and checking 
you effectively have IRQ 27 pending in it.
> Another thing I noticed that the qemu logging, never shows anything for 
> the virt_interrupt.
What do you mean by "virt_interrupt"? is it a message QEMU is supposed 
to print when injecting interrupt to the guest?
Cheers,

--
Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel