On 25.07.19 14:53, João Reis wrote:
> Even when i set all RAM regions to non rootshared, the problem continues (send
> in attachment the newly adjusted config file). Is the linux image that i am
> using too big (1,2GB)? I've generated the image using the steps described
> on setup-on-zynqmp-zcu102.md (Testing Jailhouse Linux).
> 

If I do the math right, you are even not close to that 1.2 GB when it comes to
the RAM you assigned. It's maybe some 700 MB.

Jan

> 
> quinta-feira, 25 de Julho de 2019 às 11:00:36 UTC+1, Jan Kiszka escreveu:
> 
>     On 25.07.19 11:32, João Reis wrote:
>     > Hello,
>     >
>     > I'm trying to apply cache coloring to Linux and Erika. As root cell 
> Linux
>     cannot
>     > be colored, i intend to use colored Linux in a non-root cell and colored
>     Erika
>     > in another non-root cell (the system stays as: Linux in root cell, 
> Linux and
>     > Erika in non-root cells).
>     >
>     > The problem that is occuring to me is when i issue the command: 
> jailhouse
>     cell
>     > linux ultra96-linux-demo.cell Image -i rootfs.cpio -c "console=ttyS0, 
> 115200"
>     >
>     > The error is the following:
>     >
>     > root@xilinx-ultra96-reva-2018_2:~# jailhouse cell linux
>     ultra96-linux-demo.cell
>     > Image -d inmate-zynqmp-zcu102-2.dtb -i rootfs.cpio -c "console=ttyS0, 
> 115200"
>     > [  127.150370] python invoked oom-killer:
>     > gfp_mask=0x16080c0(GFP_KERNEL|__GFP_ZERO|__GFP_NOTRACK), 
> nodemask=(null), 
>     > order=0, oom_score_adj=0
> 
>     Looks like your inmate configuration grants too little memory for the 
> Linux
>     guest. Note that the in-tree configs may only be happily used with very 
> small
>     Linux images, just as the buildroot we generate via jailhouse-images.
> 
>     Looking at your config...
> 
>             /* RAM */ {
>                     .phys_start = 0x74000000,
>                     .virt_start = 0x74000000,
>                     .size = 0x7ef0000,
>                     .flags = JAILHOUSE_MEM_READ | JAILHOUSE_MEM_WRITE |
>                             JAILHOUSE_MEM_EXECUTE | JAILHOUSE_MEM_DMA |
>                             JAILHOUSE_MEM_LOADABLE,
>             },
> 
>     This one is fine and will be available to the guest.
> 
>             /* RAM */{
>                     .phys_start = 0x3fd00000,
>                     .virt_start = 0x3fd00000,
>                     .size = 0x202f0000, //must be page size aligned
>                     .flags = JAILHOUSE_MEM_READ | JAILHOUSE_MEM_WRITE |
>                             JAILHOUSE_MEM_EXECUTE | JAILHOUSE_MEM_LOADABLE |
>                             JAILHOUSE_MEM_ROOTSHARED, //se tirar
>     JAILHOUSE_MEM_ROOTSHARED da exception
>     fault 0x20
>             },
> 
>     This one isn't because it shares RAM with the root cell that is probably 
> not
>     intended to be shared. But it will also not be presented as RAM to the 
> non-root
>     cell (via the devices tree) because of the ROOTSHARED tag.
> 
>     Jan
> 
>     -- 
>     Siemens AG, Corporate Technology, CT RDA IOT SES-DE
>     Corporate Competence Center Embedded Linux
> 
> -- 
> 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 [email protected]
> <mailto:[email protected]>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/jailhouse-dev/320981f3-9d93-46c5-b95f-ddb68083f7ee%40googlegroups.com
> <https://groups.google.com/d/msgid/jailhouse-dev/320981f3-9d93-46c5-b95f-ddb68083f7ee%40googlegroups.com?utm_medium=email&utm_source=footer>.

-- 
Siemens AG, Corporate Technology, CT RDA IOT SES-DE
Corporate Competence Center Embedded Linux

-- 
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 [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/jailhouse-dev/c1a63d36-2dd0-5b52-bb16-31794ab93b62%40siemens.com.

Reply via email to