- l4-hackers - example.com Site

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

lwip and C++ compilation/linking issue
by Stark, Josef 01 Sep '14

01 Sep '14

Hello, I want to use TCP/IP (and therefore ankh and lwip) inside a C++ application. The unmodified lwip example (src/l4/pkg/ankh/examples/lwip) compiles fine (and also works without DHCP and with a little tweaking), but it is written and configured as C and not C++. So if I change the filename of main.c to main.cc and the Makefile accordingly, unfortunately it doesn't compile. Following error: ---------------- ... Compiling main.o In file included from /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/contrib/lwip/contrib/src/include/lwip/arch.h:46:0, from /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/contrib/lwip/contrib/src/include/lwip/debug.h:35, from /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/contrib/lwip/contrib/src/include/lwip/opt.h:46, from /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/contrib/lwip/contrib/src/include/lwip/tcpip.h:35, from /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/l4/ankh/lwip-ankh.h:3, from /home/josef/Desktop/test/main.cc:10: /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/contrib/lwip/arch/cc.h:48:13: error: previous declaration of ‘int* __h_errno_location()’ with ‘C++’ linkage extern int *__h_errno_location (void) __THROW __attribute__ ((__const__)); ^ In file included from /home/josef/Desktop/test/main.cc:19:0: /home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/include/uclibc/netdb.h:61:73: error: conflicts with new declaration with ‘C’ linkage extern int *__h_errno_location (void) __THROW __attribute__ ((__const__)); ^ make[1]: *** [main.o] Error 1 make: *** [/home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/ext-pkg/home/josef/Desktop/test/OBJ-x86_586-l4f] Error 2 ----------------- Alright, so I tried surrounding the offending include with 'extern "C" ' (as well as moving it above #include <l4/ankh/lwip-ankh.h> since this file also includes it): extern "C" { #include "arch/cc.h" } I think that it solved the problem, because the error went away, but another error appeared (see below), so I also tried doing the same only to the offending line inside arch/cc.h instead: extern "C" { extern int *__h_errno_location (void) __THROW __attribute__ ((__const__)); } Both approaches led to the same linker error: ------------------ main.o: In function `main': main.cc:(.text.startup+0x11c): undefined reference to `ankhif_init(netif*)' make[1]: *** [test] Error 1 make: *** [/home/josef/Desktop/l4re-snapshot-2014053111/obj/l4/x86/ext-pkg/home/josef/Desktop/test/OBJ-x86_586-l4f] Error 2 ------------------ This is the Makefile (only the line SRC_CC is different to the original example Makefile) ------------------ PKGDIR ?= . L4DIR ?= $(PKGDIR)/../l4re-snapshot-2014053111/src/l4 TARGET = test SRC_CC = main.cc SYSTEMS = x86-l4f arm-l4f REQUIRES_LIBS = lwip libc_be_socket_lwip liblwip_netif_ankh libc_support_misc include $(L4DIR)/mk/prog.mk ------------------ Which library is missing here, or what else could be going wrong? Cheers, Josef

1 1

Minimal requirements to run L4linux using only a serial console
by Masti Ramya Jayaram 30 Aug '14

30 Aug '14

Dear all, I am trying to run L4linux on fiasco/l4re but with the keyboard and VGA disabled. I can get all the way to the command prompt (of the shell) in the initramfs but then, on pressing a key, I get a page fault and I cannot debug the system anymore using the serial_esc. I suspect it is my configuration which I have provided below. I was wondering what the minimal packages are to achieve this. modules.list: 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 drops-rd.rd I have disabled the framebuffer and input server in the L4linux stub drivers configuration and enabled the serial console. l4linux.cfg: --vim:set ft=lua: L4.default_loader:start({ caps = { log = L4.Env.log:m("rws"), }, l4re_dbg = L4.Dbg.Warn, log = { "l4linux", "yellow" }, }, "rom/io -vvvv rom/vmlinuz mem=64M console=ttyLv0 l4x_rd=rom/drops-rd.rd root=1:0 ramdisk$ I also specify a custom frequency for my board to fiasco.oc, do I have to make similar modifications to l4linux (e.g., Timer frequency)? Thanks, Ramya

3 14

RE: Minimal requirements to run L4linux using only a serial console
by Masti Ramya Jayaram 29 Aug '14

29 Aug '14

Hi Martin, That was a very useful tip. I simply changed the default value in kernel/fiasco/src/kern/config.cpp from SERIAL_ESC_IRQ to SERIAL_ESC_NOIRQ which leads to polling (I am guessing) and voila! It works! I even tried on the actual SCC and it works! I can break into the debugger as well as input to l4linux. Thanks a lot for your help for your help. Best, Ramya ________________________________________ From: Masti Ramya Jayaram Sent: 29 August 2014 17:48 To: Martin Schröder; l4-hackers(a)os.inf.tu-dresden.de Subject: RE: Minimal requirements to run L4linux using only a serial console Also with acpi on, it does enter the debugger in qemu. Best, Ramya ________________________________________ From: Masti Ramya Jayaram Sent: 29 August 2014 17:45 To: Martin Schröder; l4-hackers(a)os.inf.tu-dresden.de Subject: RE: Minimal requirements to run L4linux using only a serial console Hey Martin, No I just realized that ESC does not break into the debugger. But I can clearly see that the process instead of printing "Hello world" prints "ello World" :) What could the reason be? Thanks, Ramya ________________________________________ From: l4-hackers [l4-hackers-bounces(a)os.inf.tu-dresden.de] on behalf of Martin Schröder [martin.schroeder(a)openlimit.com] Sent: 29 August 2014 17:35 To: l4-hackers(a)os.inf.tu-dresden.de Subject: Re: Minimal requirements to run L4linux using only a serial console Hi Ramya, a: No, its not a must. The UART could also be polled (by a timer) to see if something arrived. b: At least regarding l4linux there is nothing more to it - afaik. But let's go back to the hello word example on the SCC. What happens if you press the <ESC> key? Does it break into the JDB debugger? Best regards, Martin Am 29.08.2014 um 16:56 schrieb Masti Ramya Jayaram: > Hey Martin, > > I am trying to run l4linux/fiasco on the Intel SCC which does have a pic-8259. The Intel SCC does not have a BIOS and does not have any ACPI tables which I think qemu generates on the fly. It also does not use a Screen or keyboard. > > I obtained a patch from Markus Partheymueller which they used to boot fiasco on the SCC. With it I am able to run hello world both in qemu with no-acpi as well as on the SCC successfully. > > The patch to the best of my knowledge does the following: > > a. Makes the cmdline also parse an argument so that the frequency is an input (533Mhz) > b. Uses the apic timer for scheduling. > c. Disables the pic-8259 because it does not exist. Instead the core in the SCC has two interrupt lines and this patch initializes these lines as interrupts to the core. > d. Disables keyboard and screen and enables a serial line for communication instead > e. Disables super pages because the SCC does not support it. > > So, when I try the patched fiasco/L4linux on qemu, it works when I do not disable acpi. But since the SCC does not have any ACPI/IO-APIC, when I try to run the same thing on qemu with ACPI off, l4linux boots but is not responsive to input. > > In short: > > when I run with I (below) , my l4linux responds to input and with II (below) not. > > qemu-system-i386 -m 256 -serial stdio -kernel bootstrap_l4linux.elf -append "-arg=fiasco,,-serial_esc -freq=533000" > > qemu-system-i386 -no-acpi -m 256 -serial stdio -kernel bootstrap_l4linux.elf -append "-arg=fiasco,,-serial_esc -freq=533000" > > Now the question is: > > a. Is the IO-APIC a must for the serial line? > b. Are ACPI tables just a matter of writing data at the right memory locations (e..g, qemu lets you give the acpi-table as input I noticed)? Or is there more to it? > > Thanks for your patience, > ramya > > > ________________________________________ > From: l4-hackers [l4-hackers-bounces(a)os.inf.tu-dresden.de] on behalf of Martin Schröder [martin.schroeder(a)openlimit.com] > Sent: 29 August 2014 16:29 > To: l4-hackers(a)os.inf.tu-dresden.de > Subject: Re: Minimal requirements to run L4linux using only a serial console > > Hi Ramya, > > Why did you disable the PIC? > What is your aim with ACPI and why should this lead to a change in the UART driver? > And what has the serial console to do with ACPI tables? > > For me the topics have at the moment very little in common or I am missing something. > > Best regards, > Martin > > > Am 29.08.2014 um 15:58 schrieb Masti Ramya Jayaram: >> Also, given that the thing works with ACPI on, wouldn't it be better to include ACPI information in the boot instead of changing the driver? Or is this impossible? >> >> Best, >> Ramya >> >> >> ________________________________________ >> From: Masti Ramya Jayaram >> Sent: 29 August 2014 15:55 >> To: Martin Schröder; l4-hackers(a)os.inf.tu-dresden.de >> Subject: RE: Minimal requirements to run L4linux using only a serial console >> >> Hey Martin, >> >> Back on the list ;-) >> >> Ok, I can confirm that when I busy loop during fiasco init, I can actually see the character that I typed. So would I have to change the l4ser driver in Linux or the main driver in under pkg/l4/bootstrap or under fiasco in vkey.cpp? :) >> >> Thanks, >> ramya >> ________________________________________ >> From: Masti Ramya Jayaram >> Sent: 29 August 2014 12:15 >> To: Martin Schröder; l4-hackers(a)os.inf.tu-dresden.de >> Subject: RE: Minimal requirements to run L4linux using only a serial console >> >> Hey Martin, >> >> I cleaned up my build and it worked - the mistake was mine. I had a question regarding ACPI. The target machine I am using has no BIOS/ACPi support. So, when I boot linux on it my console input does not work.This is similar to booting in qemu with the -no-acpi option. Serial console output works (I can see the prompt of L4Linux) but input does not. >> >> Is there something special to do when booting with a serial console but no ACPI tables? >> >> Thanks, >> ramya >> ________________________________________ >> From: Martin Schröder [martin.schroeder(a)openlimit.com] >> Sent: 27 August 2014 18:32 >> To: Masti Ramya Jayaram; l4-hackers(a)os.inf.tu-dresden.de >> Subject: Re: Minimal requirements to run L4linux using only a serial console >> >> Hi Masti, >> >> you have all of them already! I did not alter any other file beside the files you already sent. I inserted them into a fresh >> snapshot (http://l4re.org/download/snapshots-oc/l4re-snapshot-2014053111.tar.xz) and started building. Since "drops-rd.rd" was not >> found in the snapshot, I just renamed "ramdisk-x86.rd" to "drops-rd.rd". Nothing more. >> >> Martin >> >> >> >> Am 27.08.2014 um 18:02 schrieb Masti Ramya Jayaram: >>> Hey Martin, >>> >>> thanks for the quick response. Could you send me your configs (l4linux, fiasco, l4re) so that I can try it with that? >>> >>> Thanks, >>> Ramya >>> ________________________________________ >>> From: l4-hackers [l4-hackers-bounces(a)os.inf.tu-dresden.de] on behalf of Martin Schröder [martin.schroeder(a)openlimit.com] >>> Sent: 27 August 2014 16:28 >>> To: l4-hackers(a)os.inf.tu-dresden.de >>> Subject: Re: Minimal requirements to run L4linux using only a serial console >>> >>> Hi Masti, >>> >>> I just applied your config, replaced "drops-rd.rd" by "ramdisk-x86.rd" and everything (command prompt, debugger) runs fine as >>> expected in Qemu. Maybe you ran into a similar problem as described in this thread: >>> >>> http://os.inf.tu-dresden.de/pipermail/l4-hackers/2012/thread.html#5165 >>> >>> >>> Best regards, >>> Martin >>> >>> >>> Am 27.08.2014 um 15:18 schrieb Masti Ramya Jayaram: >>>> Sorry, a couple of typos in the l4linux.cfg >>>> --vim:set ft=lua: >>>> >>>> >>>> L4.default_loader:start({ caps = { >>>> log = L4.Env.log:m("rws"), >>>> }, >>>> l4re_dbg = L4.Dbg.Warn, >>>> log = { "l4linux", "yellow" }, >>>> }, "rom/vmlinuz mem=64M console=ttyLv0 l4x_rd=rom/drops-rd.rd root=1:0 ramdisk_size=4000 l4ser.vkey_enable=1"); >>>> >>>> Thanks in advance, >>>> ramya >>>> >>>> ________________________________________ >>>> From: l4-hackers [l4-hackers-bounces(a)os.inf.tu-dresden.de] on behalf of Masti Ramya Jayaram [rmasti(a)inf.ethz.ch] >>>> Sent: 27 August 2014 14:07 >>>> To: l4-hackers(a)os.inf.tu-dresden.de >>>> Subject: Minimal requirements to run L4linux using only a serial console >>>> >>>> Dear all, >>>> >>>> I am trying to run L4linux on fiasco/l4re but with the keyboard and VGA disabled. I can get all the way to the command prompt (of the shell) in the initramfs but then, on pressing a key, I get a page fault and I cannot debug the system anymore using the serial_esc. I suspect it is my configuration which I have provided below. I was wondering what the minimal packages are to achieve this. >>>> >>>> modules.list: >>>> 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 drops-rd.rd >>>> >>>> I have disabled the framebuffer and input server in the L4linux stub drivers configuration and enabled the serial console. >>>> >>>> l4linux.cfg: >>>> >>>> --vim:set ft=lua: >>>> >>>> L4.default_loader:start({ caps = { >>>> log = L4.Env.log:m("rws"), >>>> }, >>>> l4re_dbg = L4.Dbg.Warn, >>>> log = { "l4linux", "yellow" }, >>>> }, "rom/io -vvvv rom/vmlinuz mem=64M console=ttyLv0 l4x_rd=rom/drops-rd.rd root=1:0 ramdisk$ >>>> >>>> >>>> I also specify a custom frequency for my board to fiasco.oc, do I have to make similar modifications to l4linux (e.g., Timer frequency)? >>>> >>>> Thanks, >>>> Ramya _______________________________________________ l4-hackers mailing list l4-hackers(a)os.inf.tu-dresden.de http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers

1 0

Re: Minimal requirements to run L4linux using only a serial console
by Martin Schröder 29 Aug '14

29 Aug '14

Hi Ramya, Am 29.08.2014 um 14:29 schrieb Masti Ramya Jayaram: > Hi Martin, > > Thanks for checking that. I also see that my build works provided I don't apply a special patch meant for my board (Intel SCC). The patch essentially only alters the frequency and disables the PIC. If you disable the PIC you won't get any character into your system since UART reception is normally interrupt based. If you want to change this you have to poll the UART FIFO in the driver every second or so if there is any character available. Best regards, Martin.

1 0

Re: advice on using ddelinux
by Björn Döbel 28 Aug '14

28 Aug '14

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 28.08.2014 12:14, Korbinian Ederer wrote: > Yes the symlinks are build properly, but you're right with the > CFlags. The include dir is not listed. How do I tell the make > process to do so? My bad, the variable is called PRIVATE_INCDIR as you can see in all the example Makefiles around the l4/pkg/dde subdirectory. Cheers, Bjoern - -- Dipl.-Inf. Bjoern Doebel Mail: doebel(a)tudos.org TU Dresden, OS Chair Phone: +49 351 463 38 799 Noethnitzer Str. 46 Fax: +49 351 463 38 284 01187 Dresden, Germany WWW: http://www.tudos.org/~doebel - -- "When the seagulls follow the trawler, it's because they think sardines will be thrown into the sea." (Eric Cantona) -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iEYEARECAAYFAlP/E0sACgkQP5ijxgQLUNmhTgCfbaw/Y+IIysCiwasxjvuVSxEP BdkAoJWEcSHWRfTedvXj0VotJv+1zVvn =o8Pq -----END PGP SIGNATURE-----

1 0

Re: Mapping memory into process address space
by Noah Zentzis 28 Aug '14

28 Aug '14

On 08/17/2014 12:55 AM, Adam Lackorzynski wrote: > So now I'm not sure how you want to proceed. Are your programs special > ones, i.e. that they are quite different from L4 programs or are they L4 > programs? I wondering where to do the paging, however, in any case it > sounds to me like doing it in a different task would be good. > > To address your question: > You map the pager to the task with the dst_task->map() call as shown > above. The pager is probably a thread you implement, for doing your > work. The order does not really matter because the thread only starts > running until the run_thread call, i.e. it does not matter whether the > pager cap (any cap) is mapped before or after it has been registered > with the thread. How would I configure the new task to use the same pager that the current task is? Looking at libloader, I'd think it would be something like: task-map(L4Re::This_task, env->rm().fpage(), env->rm().snd_base()); but I'm not sure if I need to create a new region map or something like that. I think I've got memory mapping working, but the process dies with an unhandled read page fault. - Noah Zentzis

3 3

GPIO Interrupt on L4Linux
by Yunchuan Geng 27 Aug '14

27 Aug '14

Hi, L4 hackers I am implementing a driver using GPIO input as interrupt. When setup handler for gpio, either from "irq_set_chained_handler()" or "requst_irq()" in kernel driver, it finally calls "l4lx_irq_io_startup()" function and return "/l4x-irq: Failed to set type for IRQ x".// / What this error means? Right before this error, it called l4_icu_set_mode() which goes to l4 side, I tried to print out debug information in l4 side, but not very successful. I tried read document about ICU and IRQ handling library about L4Re, but did not find the answer, maybe I overlook it. So I ask help here. Is there an guidance about this interrupt things or could some one please give me some advice. Best Regards, -Yunchuan

2 1