[Barrelfish] [Barrelfish-users] Weird problem while running slideshow in QEMU!

[Simon Peter]

Hi Minqiu,

It seems you are running out of space in the Multiboot module CNode. An
easy way to find out is to reduce the number of slides you are loading.
If that's the cause, you can increase the size of the module CNode by
increasing MODULECN_SIZE_BITS in include/barrelfish_kpi/init.h. Just
increasing it by 1 should be enough, as that will already double its size.

Simon

tianminqiu wrote:
> Hi,
> I have run slideshow example in QEMU successfully, following the
> instructions given by Andrew Baumann:
> 1. Create '/build/talk' directory and a series of slides talk-*.bmp.gz 
> packages
>     using makeslides.sh, then add them to the menu.lst.
>     Each bitmap is 1024 * 768 created by convert tool of imagemagick.
> 2. Change QEMU config in /build/CMakeFiles/sim.dir/build.make, removing
>     '-nographic' and adding '-vga std -serial stdio' arguments.
> 3. Modify the menu.lst manually:
>    add vbe, lpc_kbd and slideshow module to menu.lst, and the loading     
>    sequence while booting in my menu.lst is like this:
>    timeout 0
>    # menu.lst
>     title    Barrelfish
>     root    (nd)
>     kernel    /kernel/kernel loglevel=4
>     module    /usr/init/init
>     module    /usr/mem_serv/mem_serv
>     module    /usr/monitor/monitor
>     #bootcpus=1
>     module    /usr/chips/chips
>     module  /usr/drivers/vbe/vbe    # I inserted vbe here
>     module  /usr/skb/skb
>     module  /usr/pci/pci
>     module  /usr/drivers/lpc_kbd/lpc_kbd # I inserted kbd here
>     module  /usr/drivers/serial/serial
>     module    /usr/fish/fish
>     module  /usr/slideshow/slideshow 1024 768 32 talk # slideshow is 
> here with 4 arguments passed
>     modulenounzip /talk/talk-0.bmp.gz nospawn # the below are added 
> automatically
>     ...
>     modulenounzip /talk/talk-56.bmp.gz nospawn # 57 slides in total
>     
> At the moment, I can run slideshow with 'make sim' successfully, and
> everything is OK until today, when I needed to port the environment to
> another machine, weird problem occured. I didn't copy the directory to
> the target machine, but rebuilt the whole environment from unzipping the
> source code and reconfigured it like above. However I failed to start the
> slideshow on the target machine, and I got the outputing message:
>         ...
>         #previous messages are omitted
>         Kernel starting at address 0xffffffffc2311000
>         kernel 0 PANIC! kernel assertion "modulecn_slot < (1UL << 
> modulecn->cap.u.cnode.bits)" failed at 
> /media/D/barrelfish/kernel/startup.c:403
>         kernel 0: exception 3 (error code 0x0): breakpoint
>         Faulting instruction pointer (or following instruction): 
> 0xffffffffc23160ad
>         RAX 0xffffffffc232a000 RBX 0xffffff00000298b4 RCX 0x3f8 RDX 
> 0xffffff0000167930 RSP 0xffffffffc237de08
>         Top o' stack:
>         0xffffffffc2317b06 0x3000000020 0xffffffffc237def0 
> 0xffffffffc237de30 0xffffffffc231a4c9 0xffffffffc2318310 
> 0xffffffffc2323da8 0xffffffffc2323c88 0x193 0x0
>         kernel 0: Waiting for GDB connection...
>        
> I have checked every detail carefully and cannot find any difference
> between the two versions, what's wrong with I did? What does the message 
> mean?
> By the way, it seems that commenting out the line of vbe in menu.lst can
> make the system running, of course at the expense of failing to start 
> video mode
> as well as slideshow, either.
> 
> Thanks,
> Minqiu
> -- 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Barrelfish-users mailing list
> Barrelfish-users at lists.inf.ethz.ch
> https://lists.inf.ethz.ch/mailman/listinfo/barrelfish-users
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Barrelfish mailing list
> Barrelfish at lists.inf.ethz.ch
> https://lists.inf.ethz.ch/mailman/listinfo/barrelfish