Am Fri, 23 Feb 2018 02:34:22 -0800
schrieb <a....@matellio.com>:

> On Wednesday, February 21, 2018 at 3:08:35 PM UTC+5:30, Henning
> Schild wrote:
> > Am Tue, 20 Feb 2018 21:19:03 -0800
> > schrieb <a....@matellio.com>:
> >   
> > > On Wednesday, February 21, 2018 at 12:14:47 AM UTC+5:30, Henning
> > > Schild wrote:  
> > > > Am Tue, 20 Feb 2018 06:00:24 -0800
> > > > schrieb <a....@matellio.com>:
> > > >     
> > > > > Hi all,
> > > > > 
> > > > > I'm experimenting jailhouse in QEMU on X86 platform.I'm able
> > > > > to run application and load Linux on non root cell.
> > > > > 
> > > > > And experimenting inter-cell communication between root and
> > > > > non-root cell(as per yours documentation). And additionally i
> > > > > inserted uio_ivshmem.ko driver in root cell (as per jailhouse
> > > > > documentation) for sharing the interrupts between root and non
> > > > > root. And also i'm running the ivshmem-demo.cell along with
> > > > > root and non root cell.
> > > > > 
> > > > > Root cell : 4.15-rc4 (jailhouse-repo) with ubuntu 16.04
> > > > > userland Non-root cell 1: same 
> > > > > Non-root cell 2: ivshmem-demo    
> > > > 
> > > > Ok, three cells but only two configs attached. I am guessing
> > > > you are using the ivshmem-demo default config for cell 2.
> > > >     
> > > > > Following logs i'm getting from console :
> > > > > 
> > > > > 
> > > > > Adding virtual PCI device 00:0f.0 to cell "ivshmem-demo"    
> > > > 
> > > > On a successful connection you should see additional output
> > > > here. 
> > > > > Created cell "ivshmem-demo"
> > > > > Page pool usage after cell creation: mem 741/16328, remap
> > > > > 65605/131072 Cell "ivshmem-demo" can be loaded
> > > > > CPU 1 received SIPI, vector 100
> > > > > Started cell "ivshmem-demo"
> > > > > IVSHMEM: Found 1af4:1110 at 00:0f.0
> > > > > IVSHMEM: shmem is at 0x000000005afff000    
> > > > 
> > > > This indicates that you are not using the default config. But i
> > > > wonder why it prints the end of the region, not the start.
> > > > Please post the config used for that cell.
> > > > It looks like the memory regions do not match, you actually
> > > > need to share memory and map it into two cells at once.
> > > >     
> > > > > IVSHMEM: bar0 is at 0x000000005b000000
> > > > > IVSHMEM: bar2 is at 0x000000005b001000
> > > > > IVSHMEM: mapped the bars got position 0
> > > > > IVSHMEM: 00:0f.0 sending IRQ
> > > > > IVSHMEM: 00:0f.0 sending IRQ
> > > > > IVSHMEM: 00:0f.0 sending IRQ
> > > > > IVSHMEM: 00:0f.0 sending IRQ
> > > > > IVSHMEM: 00:0f.0 sending IRQ
> > > > > 
> > > > > Here it is sending the IRQ's but i'm not getting any
> > > > > acknowledgement for the IRQ's. and also i check "grep
> > > > > ivshemem /proc/interrupts" in non root-cell I'm not receiving
> > > > > any IRQ's. So we need some clarification about this and give
> > > > > some input to us (Added root cell and non root cell config and
> > > > > jailhouse log also). Do we need to change anything in the
> > > > > config?    
> > > > 
> > > > In the configs you did send you establish a link between
> > > > root-cell and "linux-x86-demo". You can probably just run
> > > > ivshmem-demo in that cell and start with two cells.
> > > > 
> > > > If you run ivshmem-demo you will need to set the protocol to
> > > > JAILHOUSE_SHMEM_PROTO_UNDEFINED. Otherwise the root-cell might
> > > > bind virtio-net to the device and the demo-code will refuse to
> > > > drive the device.
> > > > 
> > > > So starting from the original qemu setup, change the proto in
> > > > both configs, run ivshmem-demo instead of non-root-linux (in
> > > > the same cell), and have the uio driver on your root-linux. And
> > > > slowly introduce changes from there.
> > > > 
> > > > Henning
> > > >      
> > > > > Regards,
> > > > > Arun
> > > > >    
> > > 
> > > Hi henning,
> > > 
> > > Thanks for reply.
> > > 
> > > Yes I'm running ivshmem-demo config. Here i attached the config
> > > file. can you   
> > 
> > The config looks good and explains the output. But you are missing a
> > partner-cell that is using the same memory region. That one cell is
> > using 4k behind the shmem that the other two are using.
> > 
> > Henning
> >   
> > > give some input to us if anything wrong in the config?
> > >  
> 
> Hi henning,
> 
> Thanks for the information.
> 
> Now i changed the memory region as per your suggestion. And now i'm
> checking intercell communication between two non root cell linux.
> 
> I created two non root cell and shared memory connection is
> established between this non root cell (i could see the log in
> console). While runnning shared memory application it is sending the
> IRQ's but i'm not receiving acknowledgement for that. 
> 
> Adding virtual PCI device 00:0f.0 to cell "ivshmem-demo"

Again, no connection. Likely because of a protocol mismatch. (NET vs.
UNKNOWN)

> Created cell "ivshmem-demo"
> Page pool usage after cell creation: mem 516/16328, remap 65605/131072
> Cell "ivshmem-demo" can be loaded
> CPU 3 received SIPI, vector 100
> Started cell "ivshmem-demo"
> IVSHMEM: Found 1af4:1110 at 00:0f.0
> IVSHMEM: shmem is at 0x000000005aeff000
> IVSHMEM: bar0 is at 0x000000005af00000
> IVSHMEM: bar2 is at 0x000000005af01000
> IVSHMEM: mapped the bars got position 0
> IVSHMEM: 00:0f.0 sending IRQ
> IVSHMEM: 00:0f.0 sending IRQ
> IVSHMEM: 00:0f.0 sending IRQ
> IVSHMEM: 00:0f.0 sending IRQ
> IVSHMEM: 00:0f.0 sending IRQ
> IVSHMEM: 00:0f.0 sending IRQ
> 
> 
> grep ivshmem /proc/interrupts 
> 
>  24:          1   PCI-MSI 229376-edge      ivshmem-net[0000:00:0e.0]

I do not even need to see the logs. This is not the uio-driver, if this
driver is bound to the device the protocol of the shmem-device is wrong.

Henning

> 
> Still same problem.Here i added the config and log also. Can u tell
> me the reason why i'm not getting acknowledgement?
> 
> 
> Regards,
> Arun  
>  
> 

-- 
You received this message because you are subscribed to the Google Groups 
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to jailhouse-dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to