- l4-hackers - example.com Site

Question about dataspace
by Yuxin Ren 15 May '14

15 May '14

I cannot understand dataspace very well. I have some questions about it. Each time I want to alloc a piece of memory, I must create a dataspace first? I look at the example code of memory allocation. I found after create adataspace, we call L4Re::Env::env<http://l4re.org/doc/classL4Re_1_1Env.html#aa1181d6256037deda5b5f83c2006ba60>()->mem_alloc()->alloc() to alloca memory via dataspace. Can I call this function multiple times on the same dataspace instance without free? If so, what is its effect? In addtion we use attach to map the dataspace. Can I attach the same dataspace multiple times without free? In this case, does it mean that the same physical memory is mapped into different virtual address?

2 5

Porting Fiasco+L4re to DaVinci DA850 - prefetch abort in sigma0
by Peter Howard 14 May '14

14 May '14

Dear all, I'm trying to port Fiasco to the DA850. I'm basing this on the 2014022818 snapshot. I've gotten to the point where I've modified the configuration for both fiasco and l4re, added the BSP and bootstrap platform files, and added the extra drivers I need to get started. I can build an image to load on the machine and it boots through the kernel fine. Then it tries to run sigma0 and gets a prefetch abort: EXCEPTION: prefetch abort pfa=c0090100, error=0033000d R[0]: c0002000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 R[8]: 00000000 00000000 00000000 00000000 00000000 c1013de4 04080124 c0090100 (note, I'm including the full boot output at the end of the email) The address, c0090100, is correct as the entry point for sigma0. Now, if I'm reading the FSR right, that's indicating a permission error on the section. But nothing I've added/changed has anything to do with the MMU/mapping. As far as I can see, that's all straight armv5 code, with nothing special to the board (which uses an arm926ejs core). So any idea why sigma0 would be getting mappings with the wrong permission? ------ Full output from boot: U-Boot > bootm 0xc0700000 ## Booting kernel from Legacy Image at c0700000 ... Image Name: L4 Image #2 Image Type: ARM Linux Kernel Image (uncompressed) Data Size: 860160 Bytes = 840 KiB Load Address: c1000000 Entry Point: c1000000 Verifying Checksum ... OK Loading Kernel Image ... OK OK Starting kernel ... L4 Bootstrapper Build: #2 Tue May 13 14:38:28 EST 2014, 4.7.3 Scanning up to 32 MB RAM Memory size is 32MB (c0000000 - c1ffffff) RAM: 00000000c0000000 - 00000000c1ffffff: 32768kB Total RAM: 32MB mod04: c10b8000-c10d1574: hello mod03: c109e000-c10b7454: l4re mod02: c106c000-c109d620: moe mod01: c1062000-c106b338: sigma0 mod00: c1015000-c1061380: fiasco Moving up to 5 modules behind c1100000 moving module 00 { c1015000-c106137f } -> { c11bd000-c120937f } [312192] moving module 01 { c1062000-c106b337 } -> { c120a000-c1213337 } [37688] moving module 02 { c106c000-c109d61f } -> { c1214000-c124561f } [202272] moving module 03 { c109e000-c10b7453 } -> { c1100000-c1119453 } [103508] moving module 04 { c10b8000-c10d1573 } -> { c111a000-c1133573 } [103796] Scanning fiasco -serial_esc Scanning sigma0 Scanning moe --init=rom/hello Relocated mbi to [0xc100e000-0xc100e0e2] Loading fiasco Loading sigma0 Loading moe find kernel info page... found kernel info page at 0xc0002000 Regions of list 'regions' [ c0001000, c0001a3f] { a40} Kern fiasco [ c0002000, c005afff] { 59000} Kern fiasco [ c0090000, c00965bb] { 65bc} Sigma0 sigma0 [ c0098000, c009e17b] { 617c} Sigma0 sigma0 [ c0140000, c016cb87] { 2cb88} Root moe [ c0170000, c0186f0f] { 16f10} Root moe [ c1000000, c10143f3] { 143f4} Boot bootstrap [ c100a000, c100a0a6] { a7} Boot mbi [ c100e000, c100e1df] { 1e0} Root Multiboot info [ c1014048, c101409f] { 58} Boot mbi [ c1100000, c1133573] { 33574} Root Module API Version: (87) experimental Sigma0 config ip:c0090100 sp:c1013de4 Roottask config ip:c0140260 sp:00000000 Starting kernel fiasco at c0001208 Hello from Startup::stage2 Cache config: ON Timer::Timer() SERIAL ESC: allocated IRQ 61 for serial uart Not using serial hack in slow timer handler. Welcome to Fiasco.OC (arm)! L4/Fiasco.OC arm microkernel (C) 1998-2013 TU Dresden Rev: rUnversioned directory compiled with gcc 4.7.3 for Davinci DA850 [] Build: #1 Tue May 13 14:33:35 EST 2014 Calibrating timer loop... done. MDB: use page size: 20 MDB: use page size: 12 EXCEPTION: prefetch abort pfa=c0090100, error=0033000d R[0]: c0002000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 R[8]: 00000000 00000000 00000000 00000000 00000000 c1013de4 04080124 c0090100 KERNEL: Warning: Sigma0 raised an exception --> HALT Panic: ... Return reboots, "k" enters L4 kernel debugger... -- Peter Howard <pjh(a)northern-ridge.com.au>

2 2

Facing l4x-evict_task problem in L4Linux
by Irvanda Kurniadi 10 May '14

10 May '14

Hi hackers, I want to ask you about my problem in running mpi-program over L4Linux. I'm usinng ubuntu 12.04 kernel with gnu libc-2.15. I put the library from /lib/i386-linux-gnu/ to /lib in my custom ramdisk. Then I put the library that is compiled in the kernel's host to the /usr/lib in my custom ramdisk. I also put my mpi hello program and my own mpi program which are compiled in the kernel's host (I mean in ubuntu 12.04 with libc-2.15). I'm using MPICH3 for the mpi library and I use gforker process manager. When I run the mpi Hello program, it can run properly (I invoke "mpiexec -n 4 ./mpi_hello and there are 4 output from different 4 different processor, it means the program run properly). Then I run my own mpi application which also include the tiff, jpeg, amd libstep libraries that I put altogether in /usr/lib/ in my custom ramdisk. But I got this error message when I invoke "mpiexec -n 1 ./mpi_program image.tif" or "mpiexec -n 2 ./mpi_program image.tif" command (shown below). l4x-evict_tasks: 1089 callbacks suppressed l4x-evict: Found no process to free l4x-evict: Found no process to free l4x-evict: Found no process to free INFO: rcu_sched detected stalls on CPUs/tasks: { 1} (detected by 0, t=348877 jiffies, g=4294967184, c=4294967183, q=(5) sending NMI to all cpus: and I can't interrupt the running process. I presume this problem happens because the all of process have been finalized but the l4linux kernel still think that there is a remaining process that should be freed. But I'm not sure about this. If you know about this, plis let me know how to fix it and if you can give me an explanation about this I will be really grateful. Thanks, Irvanda

2 1

Re: How to talk to a specific thread in a server
by Björn Döbel 09 May '14

09 May '14

Could you share your client and server's code? Bjoern Am 08.05.2014 21:02, schrieb Yuxin Ren: > Year, > I did those thing. Each server tread has its own server obj and register > server instance, and it calls register_obj() and server.loop() on their own. > But this does not work. And actually the master thread in server always > receives all requests. When I let the master thread to spin, the server > cannot > receive any request, even though other threads in server block on the > server.loop() function. > I am really confused about it. > > Thank you. > > > On Thu, May 8, 2014 at 12:22 PM, Björn Döbel <doebel(a)os.inf.tu-dresden.de>wrote: > > Am Do 08 Mai 2014 18:17:54 CEST schrieb Yuxin Ren: >>>> HiThank you very much for your reply.I did what you said, but the > behavior is very confusing.I create two channels in lua script, and > correspond each channel to a pair of client and server threads. >>>> In the server side, each thread registers its own server obj to its > own Registry_server, and of course they have different dispatch > functions.Now the result is only one server thread receive all requests > from client, and it selects different dispatch function to run based on the > request. >>>> But I want different threads to run different dispatch function, not > only one server thread process all requests using different dispatch > functions.Do you have any ideas about this?Thank you. >>>> On Thu, May 8, 2014 at 2:07 AM, Björn Döbel <doebel(a)os.inf.tu-dresden.de> > wrote: >>>> >>>> ********* *ANFANG des verschlüsselten oder unterschriebenen Bereichs* > ********* >>>> >>>> Hi, >>>> >>>> On 08.05.2014 03:29, Yuxin Ren wrote: >>>>> I have one client and one server talking to each other, and both of >>>>> them have 2 threads. I want thread 1 in client only talk to thread >>>>> 1 in server, and thread 2 in client only talk to thread 2 in >>>>> server. How can I achieve this? >>>> >>>> The simple client/server example >>>> (http://wiki.tudos.org/L4Re_simple_client_server_example) shows how to >>>> connect a single thread in each process (or task in L4 speak). You >>>> could extend this example to create two channels in the Lua setup >>>> file, pass both channels to the tasks during startup and then let each >>>> thread use one of them for communication with its counterpart. >>>> >>>> Note that does not completely enforce your requirement that each >>>> thread can only talk to exactly one other. Capabilities (e.g., for the >>>> communication channels) are shared per task and technically each >>>> thread could therefore send messages on each channel. If you want >>>> strict isolation, you have to separate the threads into separate tasks. >>>> >>>> Bjoern >>>> >>>> >>>> ********* *ENDE des verschlüsselten oder unterschriebenen Bereichs* > ********* >>>> >>>> _______________________________________________ >>>> l4-hackers mailing list >>>> l4-hackers(a)os.inf.tu-dresden.de >>>> http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers >>>> >>>> >>>> >>>> > > Quick guess: your separate threads need to call the register_obj() > function and the server.loop() functions on their own. (You also need > two independent instances of the Server_object.) > > Bjoern >> >> >

2 1

How to talk to a specific thread in a server
by Yuxin Ren 08 May '14

08 May '14

I have one client and one server talking to each other, and both of them have 2 threads. I want thread 1 in client only talk to thread 1 in server, and thread 2 in client only talk to thread 2 in server. How can I achieve this? Thanks

2 3

Compile library in L4Linux ramdisk
by Irvanda Kurniadi 08 May '14

08 May '14

Hi hackers, Can you help me to check whether my understanding about library in L4Linux is right or wrong? IMO, the library in L4linux ramdisk should be compiled with the same version of compiler. But I found out that every application in l4linux ramdisk is compiled by glibc 2.2.5 for GNU/Linux 2.0.30 which is very old version. Does it mean I have to build the library over that system and then put the built library to the ramdisk of l4linux? Thanks, Irvanda

2 1

What is 'new-res Mmio_ram(0x80000, 0 )' in '.io' file ?
by Taeung 07 May '14

07 May '14

Hi, l4 hackers :-) I'm trying to secure Ethernet on L4 linux using Freescale imx6q. Referencing other 'imx28.io' file which succeeded in securing Ethernet , I'm modifying my 'imx6q.io' file as follows.. =================== hw-root { NIC => new Deviceo() { .hi-d = "imx6q-fec"; new-res Mmio (0x02188000 .. 0x0218BFFF); new-res Irq (150); new-res Irq (151); } DMAMEM => new Deivce() { .hid ="dmamem"; new-res Mmio_ram (0x80000, 0); } 141x => new System_bus() { "imx6q-fec" -> wrap(hw-root.NIC); "dmamem" => wrap(hw-root, DMAMEM); ) } =================== What mean 'DMAMEM' ? And What is the function of which name is 'Mmio_ram()' ? I know that 'Mmio()' need arguments which are start and end address. But, I don't know 'Mmio_ram()'. Please, help me. If you tell a bit me hints, I'd appreciate it. - Taeung -

2 1

How to enable 'CONFIG_COMMON_CLK' ?
by Taeung 30 Apr '14

30 Apr '14

Hello, L4 Hackers :-) I'm trying to build L4 Linux with added option of MMC device driver. When I've tried to compile L4 Linux , I've failed it with error log such as.. ============== drivers/built-in.o: In function `sdhci_pltfm_clk_get_max_clock': /home/taeung/test/source/l4linux/drivers/mmc/host/sdhci-pltfm.c:43: undefined reference to `clk_get_rate' make[1]: *** [vmlinux] Error 1 make: *** [sub-make] Error 2 ============== After I saw the log , I've guessed the cause is that 'drivers/clk/clk.c' was not built. So, I must enable 'CONFIG_COMMON_CLK' , to build 'drivers/clk/clk.c' How to enable 'CONFIG_COMMON_CLK' ? For reference, I've tried to compile L4 Linux with options as follows.. =================================== --- MMC/SD/SDIO card support [ ] MMC debugging [*] Assume MMC/SD cards are non-removable (DANGEROUS) [ ] MMC host clock gating *** MMC/SD/SDIO Card Drivers *** <*> MMC block device driver (8) Number of minors per block device [*] Use bounce buffer for simple hosts < > SDIO UART/GPS class support < > MMC host test driver *** MMC/SD/SDIO Host Controller Drivers *** < > ARM AMBA Multimedia Card Interface support -*- Secure Digital Host Controller Interface support -*- SDHCI platform and OF driver helper <*> Marvell MMP2 SD Host Controller support (PXAV3) < > Marvell PXA9XX SD Host Controller support (PXAV2) < > Synopsys DesignWare Memory Card Interface ==================================== -Taeung -

2 2

L4Linux on amd64 arch
by Irvanda Kurniadi 28 Apr '14

28 Apr '14

Hi hackers, Is there anyone who has tried to build l4linux on amd64? I compiled the new snapshot of L4re, but there is no l4linux for amd64. Can I barely compile this l4linux in l4re-amd64? thanks, Irvanda

2 3

L4linux: cpu and core [comprehension question]
by Irvanda Kurniadi 28 Apr '14

28 Apr '14

Hi hackers, I am a little bit confused about the term of cpu and core in l4linux. I have a system with 8 processor intel xeon E5606 with 4 cores in each cpu. I'm going to maximize it to become the maximum number of cores in all cpus. But I am confused in determining the l4x_cpus that what is the term cpu in here? Is it literally cpu or the number of threads that can be handled in each core? I run an openmp hello world apps over l4linux. I call the omp_get_thread_num() and I got 4 threads in the result. Are these 4 threads obtained from 4 cores in single cpu? So, if I'm right, how can I optimize to use all of the cpu existed in the system? I've tried to use l4x_cpus=8 and l4x_cpus=32, but those commands doesn't work for me. In my opinion, if I can maximize until the maximum number of cores and I have 32 threads that run concurrently, then it will be served by each core in those 8 cpus. In fiasco configuration, there is maximal supported number of CPUs. So referring to my case, the maximum is supposed to be 8, am I right? In case of thread migration in example ex_thread_migrate, I create 16 threads and will be migrated to each cpu. Here is the result: migrate | Migrated Thread00 -> CPU02 migrate | Migrated Thread01 -> CPU03 migrate | Migrated Thread02 -> CPU04 migrate | Migrated Thread03 -> CPU05 migrate | Migrated Thread04 -> CPU06 migrate | Migrated Thread05 -> CPU07 migrate | Migrated Thread06 -> CPU00 migrate | Migrated Thread07 -> CPU01 migrate | Migrated Thread08 -> CPU02 migrate | Migrated Thread09 -> CPU03 migrate | Migrated Thread10 -> CPU04 migrate | Migrated Thread11 -> CPU05 migrate | Migrated Thread12 -> CPU06 migrate | Migrated Thread13 -> CPU07 migrate | Migrated Thread14 -> CPU00 migrate | Migrated Thread15 -> CPU01 My question is do the thread00 and thread08 share the same core in CPU02 or they use different core in CPU02 which run concurrently? And also, is it normal to get the output like cpu[1] to cpu[7] goes to idle loop? Thanks, Irvanda

2 3