- l4-hackers - example.com Site

Hello!
by Blaine Garst 17 Feb '14

17 Feb '14

Greetings, all! I am beginning to explore nano/micro-kernel issues again after a long hiatus (24 years). I don’t know how active this list is nor its followers, so I will be brief. At the moment I am discussing kernel vs user scheduling speed bumps in the now prevalent multi-core cpu architectures in the context of the ISO C11 subcommittee CPLEX which is considering adding parallelization language extensions. I have already demonstrated feature equivalence using closures (Apple’s Blocks, which I introduced and have previously proposed for C11) and some new synchronization and scheduling primitives that I am already using as part of some Actor research. I am looking for a non-GPL kernel/executive/nano-kernel to redo some nano-kernel work I did as part of a joint Bell Labs-Sun collaboration that gave rise to SVR4 and, in opposition, the Open Software Foundation. Essentially I built a capability based nano-kernel somewhat along the lines of the V Kernel of David Cheriton (the Sun folks did “Spring”) My prototype at the time had IPC transfer speed of 10x vtable dispatch cost. At the time I wrote my own IDL, but after joining NeXT we added such a concept as a first class language feature, and it was copied into Java as “interfaces”. This is still an active interest of mine also in the context of Actors. Is this an appropriate list to discuss whether L4 is an appropriate starting point and such issues that might arise from my exploration? (My friends and former colleagues at Apple have IP restrictions such that I would prefer to not risk even the appearance of possible improprieties in discussing such matters) Blaine Garst

5 12

FOSDEM microkernel dinner
by Julian Stecklina 13 Feb '14

13 Feb '14

Hello fellow microkernel hackers, for those of you at FOSDEM: Some of us want to meet up for dinner tonight. I suggest we meet on Grand Place / Grote Markt at 20:30. This should leave everyone enough time to drop off their stuff at the hotel and get to the city center. Julian

2 1

building_l4re_pandaboard
by Комаров Тимофей 11 Feb '14

11 Feb '14

Hello, guys! I've successfully built the latest L4RE snapshot for beagleboard. Any apps and even L4Linux work like a charm. It simply rocks! Right now I'm trying to run the latest snapshot on pandaboard. I've successfully built it. Helloworld example is OK, but L4Linux gets a certain error while booting: l4linux | libio: Warning: Query of 'vbus' failed! l4linux | PH 0 (t: 1) offs=00008000 vaddr=02000000 vend=02362f1c l4linux | phys=02000000 ephys=02362f1c l4linux | f_sz=002de3b6 memsz=00362f1c flgs=rwx l4linux | PH 1 (t: 4) offs=002a16ec vaddr=022996ec vend=02299710 l4linux | phys=022996ec ephys=02299710 l4linux | f_sz=00000024 memsz=00000024 flgs=r-x l4linux | PH 2 (t: 1685382481) offs=00000000 vaddr=00000000 vend=00000000 l4linux | phys=00000000 ephys=00000000 l4linux | f_sz=00000000 memsz=00000000 flgs=rwx l4linux | Starting binary at 0x20002a0, argc=7 argv=0x80007f8c *argv=0xb1007ff0m l4linux | External resolver is at 0xa8000798 l4linux | ======> L4Linux starting... <======== l4linux | Linux version 3.12.0-l4 (user1@debian) (gcc version 4.7.3 (Sourcery C4 l4linux | Binary name: rom/vmlinuz.arm l4linux | This is an AEABI build. l4linux | Linux kernel command line (6 args): mem=64M console=ttyLv0 l4x_rd=romh l4linux | CPU mapping (l:p)[1]: 0:0 l4linux | Image: 02000000 - 02400000 [4096 KiB]. l4linux | Areas: Text: 02000000 - 022b8000 [2784kB] (a bit longer) l4linux | Data: 022b8000 - 022d9200 [132kB] l4linux | Initdata: 0229a000 - 022b635c [112kB] l4linux | BSS: 022de3b6 - 02362f1c [530kB] l4linux | Device scan: l4linux | Device scan done. l4linux | l4lx_thread_create: Created thread 413 (cpu0) (u:b3000e00, v:b3000c00) l4linux | main thread will be 413 l4linux | l4x_register_pointer_section: addr = 02000000 size = 3551232 l4linux | section-with-init: Virt: 0x2000000 to 0x2362f1b [3467 KiB] l4linux | section-with-init: Phys: 0x8031e000 to 0x80680f1b, [3467 KiB] l4linux | Main thread running, waiting... l4linux | L4x: Memory size: 64MB l4linux | L4x: Setting superpages for main memory l4linux | L4x: Adjusted memory start: 02000000 l4linux | Main memory: Virt: 0x2400000 to 0x63fffff [65536 KiB] l4linux | Main memory: Phys: 0x81900000 to 0x858fffff, [65536 KiB] l4linux | l4x: vmalloc area: 06400000 - 0e400000 l4linux | l4x_register_pointer_section: addr = 02000000 size = 3551232 l4linux | text: Virt: 0x2000000 to 0x2362f1b [3467 KiB] l4linux | text: Phys: 0x8031e000 to 0x80680f1b, [3467 KiB] l4linux | l4x_rd_path: rom/ramdisk-arm.rd l4linux | Loading: rom/ramdisk-arm.rd l4linux | INITRD: Size of RAMdisk is 3072KiB l4linux | RAMdisk from 00002000 to 00302000 [3072KiB] l4linux | 'swp(b)' instruction at 021ec5a0 and faulting. l4linux | Linux built for the wrong ARM version? The last 2 lines are going to be repeated forever :-) I've tried both 4.7 and 4.8 gcc by Linaro and CodeSourcery and always get the same error. What can be a problem? Greetings! Looking forward for your help!

3 3

Passing Compilation Flag from CMake file
by Irvanda Kurniadi 11 Feb '14

11 Feb '14

Hi, I'm in the middle of porting OpenMPI. I got some trouble that I have to use CMakelists.txt to pass the Compilation Flag through Makefile. I got this error message while trying to compile: error: 'OMPI_MPIHANDLES_DLL_PREFIX' undeclared (first use in this function) error: 'OMPI_MSGQ_DLL_PREFIX' undeclared (first use in this function) And I found the OMPI_MPIHANDLES_DLL_PREFIX in CMakelist.txt as shown below: SET_TARGET_PROPERTIES(libmpi PROPERTIES COMPILE_FLAGS "${OMPI_C_DEF_PRE}OMPI_ MPIHANDLES_DLL_PREFIX=libompi_dbg_mpihandles ${OMPI_C_DEF_PRE}OMPI_MSGQ_DLL_PREFIX=libompi_dbg_msgq") Does anybody know how to use this CMake file? Regards, Irvanda

1 0

Re: Re: Re: Network Server Ankh
by Ajith Whowe 10 Feb '14

10 Feb '14

Dear Martin, These commands indicate that the link is always down with the "NO-CARRIER" flag. Please kindly refer to the outputs given below. This flag is set by the function "update_carrier", which checks 3 conditions, namely the i) priv->remote_attached ii) l4shmc_is_chunk_ready(&priv->tx_chunk) iii) l4shmc_is_chunk_ready(&priv->rx_chunk). Unfortunately the problem is originated at check point "l4shmc_is_chunk_ready". This function returns 0, the if clause fails and sets the link down. I just removed this check point from the if clause, set the link_up to 1 and observed that the ping request showed that the l4x_l4shmc_xmit_frame function is called. Nevertheless still no answer back from the second instance. What can be the main reason for the l4x_l4shmc_xmit_frame to return 0? Best Regards and many Thanks in Advance, The result for the "ip addr show" is as follows: XXXXX - XXXX: Inside l4x_l4shmc_get_stats function - start 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo 2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 52:54:00:b0:cf:02 brd ff:ff:ff:ff:ff:ff inet 192.168.124.116/24 scope global eth0 inet 192.168.0.5/24 scope global eth0:0 The result for the "ip link show" is as follows: XXXXX - XXXX: Inside l4x_l4shmc_get_stats function - start 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 52:54:00:b0:cf:02 brd ff:ff:ff:ff:ff:ff Hi Ajith, Am 07.02.2014 17:58, schrieb Ajith Whowe: >* Dear Martin, *>>* Thanks a lot. It did the job. I have the network interfaces on both *>* sides. Nevertheless I can not ping one another. The strange thing is the *>* ping request clearly does invoke any of the net_device_ops: *>>* static const struct net_device_ops l4shmnet_netdev_ops = { *>* .ndo_open = l4x_l4shmc_open, *>* .ndo_start_xmit = l4x_l4shmc_xmit_frame, *>* .ndo_stop = l4x_l4shmc_close, *>* .ndo_change_mtu = l4x_l4shmc_change_mtu, *>* .ndo_get_stats = l4x_l4shmc_get_stats, *>>* }; *>>* I put in all these functions some printk's and I can not see it once the *>* ping starts. *>>* I also did with -I option like: ping 192.168.1.2 -I eth0 but still no *>* success. *>* Do You have any idea? * If none of the functions are called, the packet get not through the IP stack. Can you post the IP configuration (ip addr show and ip link show) of of both partitions? Martin

2 1

Re: Re: Network Server Ankh
by Ajith Whowe 07 Feb '14

07 Feb '14

Dear Martin, Thanks a lot. It did the job. I have the network interfaces on both sides. Nevertheless I can not ping one another. The strange thing is the ping request clearly does invoke any of the net_device_ops: static const struct net_device_ops l4shmnet_netdev_ops = { .ndo_open = l4x_l4shmc_open, .ndo_start_xmit = l4x_l4shmc_xmit_frame, .ndo_stop = l4x_l4shmc_close, .ndo_change_mtu = l4x_l4shmc_change_mtu, .ndo_get_stats = l4x_l4shmc_get_stats, }; I put in all these functions some printk's and I can not see it once the ping starts. I also did with -I option like: ping 192.168.1.2 -I eth0 but still no success. Do You have any idea? Best Regards, Hi, Am 07.02.2014 15:22, schrieb Ajith Whowe: >* Dear Adam, *>>* Thank You very much for the response. Do You mean that the l4shmnet can be used as a communication channel between two (or more ?) instances of l4linux over TCP/IP? * Yes, never tried with more, but two instances are working fine. >* By the way, how should then the start scripts seem? Would it be again like a client - server type interaction between both instances? For instance would it work with the following set up and would I see for example a ping request and response of each side? * The l4shmnet negotiates client and server automatically. l4shmnet needs some namespace to find the other side: - local my_channel = l:new_channel(); + local test_l4shm = l:create_namespace({test_l4shm = "shmns_1"}); >* local linux_caps1 = { *>* log = L4.Env.log:m("rws"); *>* vbus = io_caps.linux_bus1, *>* rtc = rtc_caps.rtc, *>* rom = L4.Env.rom, * - myL4shm = my_channel:svr(), + test_l4shm = test_l4shm:mode("rw"); >* }; *>>* l:start( *>* { *>* caps = linux_caps1, *>* }, *>* "rom/vmlinuz1 mem=512M console=ttyLv0 l4shmnet.add=test_l4shm,macpart=01 ramdisk=rom/ramdisk1 root=1:0 ramdisk_size=5000 l4x_rd=rom/ramdisk1 init=/sbin/init"); *>>>* local linux_caps2 = { *>* log = L4.Env.log:m("r"); *>* vbus = io_caps.linux_bus1, *>* rtc = rtc_caps.rtc, *>* rom = L4.Env.rom, * - myL4shm = my_channel, + test_l4shm = test_l4shm:mode("rw"); >* }; *>>* l:start( *>* { *>* caps = linux_caps2, *>* }, *>* "rom/vmlinuz2 mem=512M console=ttyLv0 l4shmnet.add=test_l4shm,macpart=02 ramdisk=rom/ramdisk2 root=1:0 ramdisk_size=5000 l4x_rd=rom/ramdisk2 init=/sbin/init"); *> Do not forget to enable CONFIG_L4_NET_L4SHM Martin

2 1

Re: Network Server Ankh
by Ajith Whowe 07 Feb '14

07 Feb '14

Dear Adam, Thank You very much for the response. Do You mean that the l4shmnet can be used as a communication channel between two (or more ?) instances of l4linux over TCP/IP? By the way, how should then the start scripts seem? Would it be again like a client - server type interaction between both instances? For instance would it work with the following set up and would I see for example a ping request and response of each side? local my_channel = l:new_channel(); local linux_caps1 = { log = L4.Env.log:m("rws"); vbus = io_caps.linux_bus1, rtc = rtc_caps.rtc, rom = L4.Env.rom, myL4shm = my_channel:svr(), }; l:start( { caps = linux_caps1, }, "rom/vmlinuz1 mem=512M console=ttyLv0 l4shmnet.add=test_l4shm,macpart=01 ramdisk=rom/ramdisk1 root=1:0 ramdisk_size=5000 l4x_rd=rom/ramdisk1 init=/sbin/init"); local linux_caps2 = { log = L4.Env.log:m("r"); vbus = io_caps.linux_bus1, rtc = rtc_caps.rtc, rom = L4.Env.rom, myL4shm = my_channel, }; l:start( { caps = linux_caps2, }, "rom/vmlinuz2 mem=512M console=ttyLv0 l4shmnet.add=test_l4shm,macpart=02 ramdisk=rom/ramdisk2 root=1:0 ramdisk_size=5000 l4x_rd=rom/ramdisk2 init=/sbin/init"); Many thanks in advance, Regards, On Thu Feb 06, 2014 at 17:44:58 +0100, Ajith Whowe wrote: >* I coincidentally met the ankh server in the mail archives and could not *>* figure out well what the ankh exactly does. Is it an ethernet server for *>* more than one application running on the fiasco? * Yes, basically a software network switch with an (optional) network driver. >* By the way it is not existing for the newer versions of l4linux. Was it *>* replaced with the l4shmnet? * No, l4shmnet is only between two partners. The driver is there but disabled because it's not working. Adam -- Adam adam at os.inf.tu-dresden.de <http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers> Lackorzynski http://os.inf.tu-dresden.de/~adam/

2 1

Network Server Ankh
by Ajith Whowe 07 Feb '14

07 Feb '14

Dear all, I coincidentally met the ankh server in the mail archives and could not figure out well what the ankh exactly does. Is it an ethernet server for more than one application running on the fiasco? By the way it is not existing for the newer versions of l4linux. Was it replaced with the l4shmnet? Best Regards,

2 1

nothing after make for snapshot
by Emanuel Berg 03 Feb '14

03 Feb '14

I have not been able to install from l4re-snapshot-2013123021.tar.bz2 The toolchain test says everything is fine (though I had to install dialog and doxygen from the repos as well). With 'make setup' I get the dialog, and I can pick between x86-32 (ia32), UX-32, or amd64/x86_64 (or arm, though I didn't test that). I have a 64-bit CPU but it runs a 32-bit Debian Linux. What happens next is whatever I pick, I get l4/, l4linux/, and fiasco/ in obj/. It says I should set the PATH to include bin/, but except for the arm stuff, there is only setup.d/04-setup there to execute (is that it?). Anyway, I did that, and on 'make' it says: make[1]: Entering directory `/home/incal/core/l4re-snapshot-2013123021' =============== Building all Fiasco configurations ============ =============== Building all L4Re configurations ============ =============== Building all L4Linux configurations =========== =============== Building Image ================================ =============== Build done ==================================== make[1]: Leaving directory `/home/incal/core/l4re-snapshot-2013123021' However, as there is nothing in those directories, nothing happens. -- underground experts united: http://user.it.uu.se/~embe8573

3 6

Data streaming / sharing from client side in an IPC example
by Ajith Whowe 30 Jan '14

30 Jan '14

Dear Martin, thank you very much for the kind answer. The solution You provided worked well. The only concern is as You know the data to be send can not have a greater size than the L4_UTCB_GENERIC_DATA_SIZE (due to UTCB size). So the question is do you have also another solution for larger data streaming? (Off course other than sending the data by means of various blocks with L4_UTCB_GENERIC_DATA_SIZE) Best Regards, Hi, Am 27.01.2014 15:03, schrieb Ajith Whowe: >* Dear all, *>>* I have reviewed available examples for the client server based *>* applications. Unfortunately could not find any example for which a *>* client sends a char array to the server side. Each time the client *>* sends a request and the server shares the data. Would it be possible *>* to find an example for that? *>>* For instance in the streammap example, the clients does the following *>* (writes basic protocal and message tag to the UTCB and waits for the *>* stream from the server would it be possible for the client to perform *>* the stream whereas the server waits for the input?: *>>* s << l4_umword_t *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga1c2c4b3… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga1c2c4b3…>>(Opcode::Do_map) *>* << (l4_addr_t *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99…>>)addr; *Just insert the following line here: s << "message from client"; >* s << L4::Ipc::Rcv_fpage::mem((l4_addr_t *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99…>>)addr, *>* L4_PAGESHIFT *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__memory__api.html#gad2f7e71d… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__memory__api.html#gad2f7e71d…>>, *>* 0); *>>* The server on the other hand gets the protocol and base adress and *>* then does the streaming: *>* l4_addr_t *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99…>> *>* snd_base; *>* ios >> snd_base; *and insert here: unsigned long len; char buf[100]; len = sizeof(buf); ios >> L4::Ipc::Buf_cp_in<char>(buf, len); buf[len]='\0'; printf ("Received from client: %s\n", buf); >* // put something into the page to read it out at the other side *>* snprintf(page_to_map, sizeof(page_to_map), "Hello from the server!"); *>* printf("Sending to client\n"); *>* // send page *>* ios << L4::Ipc::Snd_fpage::mem((l4_addr_t *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__basic__types.html#ga4087b99…>>)page_to_map, *>* L4_PAGESHIFT *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__memory__api.html#gad2f7e71d… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__memory__api.html#gad2f7e71d…>>, *>* L4_FPAGE_RO *>* <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__fpage__api.html#gga5d4f4972… <http://os.inf.tu-dresden.de/L4Re/doc/group__l4__fpage__api.html#gga5d4f4972…>>, *>* snd_base) *>

2 1