- l4-hackers - example.com Site

Using IO-Server
by Korbinian Ederer 10 Sep '14

10 Sep '14

Hello Community, I'm relative new to l4RE. My Goal is to get access to an Hardware Device which is Build in the Processor. After some research i came to the conclusion that the IO-Server is the thing I need, but i have no clue how to invoke it. It would be great if someone could supply a short example Thanks Korbinian

2 1

Aborted L4linux due to CLI/STI
by Masti Ramya Jayaram 10 Sep '14

10 Sep '14

Dear all, I get an error during L4linux boot when I enter the debugger (using ESC) and resume: Aborted L4linux due to CLI/STI at 0x534911. When I disassemble this using gdb: (gdb) disassemble 0x534911 Dump of assembler code for function cmpxchg8b_emu: 0x00534910 <+0>: pushf 0x00534911 <+1>: cli 0x00534912 <+2>: cmp (%esi),%eax 0x00534914 <+4>: jne 0x534922 <cmpxchg8b_emu+18> 0x00534916 <+6>: cmp 0x4(%esi),%edx 0x00534919 <+9>: jne 0x534924 <half_same> 0x0053491b <+11>: mov %ebx,(%esi) 0x0053491d <+13>: mov %ecx,0x4(%esi) 0x00534920 <+16>: popf 0x00534921 <+17>: ret 0x00534922 <+18>: mov (%esi),%eax 0x00534924 <+0>: mov 0x4(%esi),%edx 0x00534927 <+3>: popf 0x00534928 <+4>: ret When I built l4linux, I had to disable CONFIG_X86_CMPXCHG, CONFIG_CMPXCHG_LOCAL and CONFIG_X86_CMPXCHG64 despite building for the Pentium because my processor does not support PAE (which is the option that enables this). I am guessing This is the reason it tries to emulate the CMPXCHG instructions. What is the correct l4linux configuration to use for a machine that does not support any of the atomic instructions like CMPXCHG8, ATOMIC, etc.? Thanks, ramya

2 5

GCC 4.9
by Stefan Wallentowitz 09 Sep '14

09 Sep '14

Dear L4 hackers, we recently started porting FIASCO to OpenRISC and our cross-compiler is GCC 4.9. I added the required libraries to our project mirror at github: https://github.com/TUM-LIS/optimsoc-fiasco/tree/for-upstream-gcc4.9 Are you interested in this? Do you want me to submit a patch for the latest public svn tree? Regards, Stefan

2 1

Preventing access to a part of the address space from Sigma0
by Masti Ramya Jayaram 09 Sep '14

09 Sep '14

Dear all, I would like to prevent access to a part of the address space from sigma0 for security reasons on my board. For example, I would like to prevent even sigma from accessing say 0xc0000000 - 0xffffffff of memory. What is the best way to accomplish this and test it? Thanks, ramya

2 1

Checking executables before running them in L4re/fiasco
by Masti Ramya Jayaram 09 Sep '14

09 Sep '14

Dear all, I am using fiasco/l4re to run applications (say hello world) and L4linux. The module list in the two cases looks like this: entry hello-cfg kernel fiasco -serial_esc -freq=533000 roottask moe rom/hello.cfg module l4re module ned module hello.cfg module hello entry l4linux kernel fiasco -serial_esc -freq=533000 roottask moe rom/l4linux.cfg module l4re module ned module l4linux.cfg module io module vmlinuz module ramdisk-x86.rd I would like to add some functionality that performs a few checks on hello and vmlinuz before they are executed. In this context, I have the following questions: a. Where is it best to implement such functionality - moe or ned? Or should I implement a separate server and do this? b. I need access to an special range of IO memory when I perform the checks. i guess I could always use the io module and do "l4io_request_iomem()" . Would that be the right way to do it from moe/ned? c. Is there a way to allocate/reserve a dedicated range of physical memory to these applications (hello/vmlinuz) so that they get allocated the same chunk everytime (of physical memory)? Thanks in advance, ramya

2 12

Adding a custom library to l4/pkg/bootstrap
by Masti Ramya Jayaram 05 Sep '14

05 Sep '14

Hi all, I have managed to compile a custom C library (libscc-sec) as part of the l4 packages. Now I would like to use it in the bootstrap package (make a call to a function in the library). I did the following: a. Added the library as a requirement in pkg/bootstrap/Control requires: drivers_uart drivers_of libc l4util cxx_io libscc-sec b. Modified the pkg/bootstrap/server/src/Make.rules so that it also looks for libscc-sec by: L4_LIBS = -static -nostdlib $(DRV_LIBS) -lcxx_base -lcxx_io -llibscc-sec But on compiling, I get an error that says that -llibscc-sec not found The package is getting compiled and I can see the library libscc-sec.a in the build folder. What am I missing? Thanks, ramya

3 6

Compiling the Fiasco.OC +L4Re for Raspberry PI
by Waldo Paz Rodriguez 03 Sep '14

03 Sep '14

hello hanzel, I have seen in l4hacker list that you have compiled fiasco.oc + l4re for Raspberry PI whithout problems. Iá¸¿ trying to compile it for rpi. I follow the next steps> 1- Compile Fiasco.OC for Broadcom2835 2- Compile L4Re for RaspberryPI model B with MODULE_SEARCH_PATH pointing to FOC builddir 3- Generate the hello example ELF image I have transformed it into a binary using: /usr/local/genode-gcc/bin/genode-arm-objcopy -Obinary bootstrap.elf l4.img I have used the three elf image and have used to the linaro and genode toolchain (both). When I have put the binary on the PI only can see the colored screen. I am using the Fiasco.OC +L4Re from 2014022815 snapshot. I am doing something wrong? Can you help me?

2 1

Passing data to a thread
by Valentin Hauner 03 Sep '14

03 Sep '14

Hi, I'm trying to pass some data to a newly created thread. Since l4_thread_ex_regs only accepts an instruction pointer, but no data pointer, it seems impossible to me to use a parameterized function such as: > thread1_func(void *data) { /* Read the data ... */ } So far, my efforts are: 1. I've tried filling the stack allocated to each newly created thread with > thread_stack = malloc(8 << 10); > thread_stack[0] = my_data; but then I'm not able to pop the data from the stack in thread1_func. 2. I've tried using IPC: The main thread which is creating the new threads sends the data to the destination thread. I'm using l4_ipc_send in the main thread and l4_ipc_receive in thread1_func just like in your utcb-ipc example. It works great, but unfortunately the call of l4_ipc_send leads to an _immediate_ execution of thread1_func (timeouts: L4_IPC_NEVER), so my scheduling policy is not respected. So how can I pass data to a thread in a way that is equivalent to passing arguments to functions? Best regards, Valentin

2 2

Fault in function writev (ulibc)
by Masti Ramya Jayaram 02 Sep '14

02 Sep '14

Dear all, I am running a custom application package on fiasco using a serial line (no keyboard/screen) and I intermittently get faults caused by printf. I notice that on disassembling the code around the error that the function is called writev in ulibc. Has anyone experienced something similar before? Thanks, ramya

2 2

New package linking fails
by Masti Ramya Jayaram 01 Sep '14

01 Sep '14

Dear all, I am trying to create a new package under l4/pkg. The linking of the package fails and it has to do with the linker path but I am not sure how to fix it. Here are the details. Package structure: (relevant parts) l4/pkg/test/server/src/main.cc l4/pkg/test/server/src/folder1/test1.c l4/pkg/test/include/test1.h l4/pkg/test/server/src/Makefile PKGDIR ?= ../.. L4DIR ?= $(PKGDIR)/../.. TARGET = $(PKGNAME) # list requirements of your program here REQUIRES_LIBS = libsigma0 libirq l4util l4re_c-util # list your .c or .cc files here SRC_C = folder1/test1.c SRC_CC = main.cc include $(L4DIR)/mk/prog.mk The compilation of the package fails with an undefined reference in main.cc to a function in test1.c (defined through test1.h). All the object files are generated and I checked under build. What do you think is the issue? Thanks, ramya

2 3