Error for l4android with Genode 14.11

Matthias Lange matthias.lange at ...272...
Mon Feb 2 13:51:29 CET 2015


Hi,

On 02/02/2015 11:46 AM, Stefan Kalkowski wrote:
> Hi,
> 
> On 01/28/2015 02:56 PM, li94575 wrote:
>> Hi all,
>> When I run l4android (3.5) with the latest Genode, I got an error, the
>> args for l4android in run script is "mem=512M console=ttys0
>> l4x_rd=root-lc.gz".  
> 
> I wonder that it compiles for you using Genode's latest release? The
> recent e-mail by Zachary showed, and I can confirm that l4android
> doesn't even compile.
> 
> The simple reason is that we don't use l4android at all, and was not
> tested it in the past :-(. I will investigate whether it is easy to fix
> the current issues, and dependend on that provide a fix. However, the
> current outdated version of L4android does not make sense from my point
> of view. Therefore, we should either remove L4Android from the Genode
> repository, or alternatively it needs to be upgraded and integrated into
> our testing environment - which means resources that are bound to that
> task. Any opinions from other developers? Or is somebody like
> li94575 or Zachary willing to upgrade the current version?

Which version of the L4Android kernel are you using? We've updated it to
3.16 a couple of months ago.

However, if it is about to run Android, a recent L4Linux kernel should
be capable as well, as the Android related subsystems have been moved
back into mainline Linux a couple of revisions ago.

Best,
Matthias.

> 
>> Part of log is as follows:
>>
>> [init -> l4android] Main thread running, waiting...
>> [init -> l4android] l4re_global_env:0
>> [init -> l4android] Main memory size: 512MB
>> unmapping of managed dataspaces not yet supported
>> [init -> l4android] Main memory: virt:20000000 to 3fffffff [524288 KiB]
>> [init -> l4android] Main memory: phys:0x00000000 to 0x20000000, size:
>> 536870912
>> [init -> l4android] Filling lower ptabs...
>> [init -> l4android] Done (0 entries).
>> [init -> l4android] l4x_register_pointer_section: addr = 01000000 size =
>> 5242880
>> [init -> l4android] text: virt: 1000000 to 14fffff [5120 KiB]
>> [init -> l4android] Cannot determine physical address for dataspace Binary!
>> [init -> l4android] error: failed to get physical address for 1000000.
>> virtual void
>> Genode::Signal_session_component::submit(Genode::signal_context_capability,
>> unsigned int): invalid signal-contex capability
>> virtual void Genode::Pager_activation_base::entry(): Could not resolve
>> pf=3fffff000 ip=1000e88
>>
>> It seems that the virtual address(0x20000000 ~ 0x3fffffff) has not been
>> mapped to actual physical address. So I compared the L4Linux support library
>> with previous versions, I found the funtion "l4re_ma_alloc" (in
>> ports-foc/src/lib/l4lx/l4_re_c_mem_alloc.cc) has changed, and l4android
>> will use the
>> function for memory allocation. Can somebody tells me why make that
>> adjustment?  Is it the reason why l4android produces that error?
> 
> Yes there were changes regarding the main memory allocation/mapping of
> the L4Linux guest within the L4Re emulation library. Especially, it
> changed that not all memory is allocated once, but gets allocated on
> demand. Thereby prior patches to L4Linux need to be reverted.
> 
> I've opened an issue at our issue tracker, let's discuss things there:
> 
>   https://github.com/genodelabs/genode/issues/1390
> 
> Regards
> Stefan
> 
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming. The Go Parallel Website,
>> sponsored by Intel and developed in partnership with Slashdot Media, is your
>> hub for all things parallel software development, from weekly thought
>> leadership blogs to news, videos, case studies, tutorials and more. Take a
>> look and join the conversation now. http://goparallel.sourceforge.net/
>>
>>
>>
>> _______________________________________________
>> genode-main mailing list
>> genode-main at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/genode-main
>>
> 


-- 
Matthias Lange, matthias.lange at ...272..., +49 - 351 - 41 88 86 14

Kernkonzept GmbH.  Sitz: Dresden.  Amtsgericht Dresden, HRB 31129.
Geschäftsführer: Dr.-Ing. Michael Hohmuth





More information about the users mailing list