l4-hackers July 2012

l4-hackers@os.inf.tu-dresden.de

21 participants
21 discussions

L4Re GUI
by Robert Campbell 12 Jul '12

12 Jul '12

Dope seems to offer a subset of the functionality of the features of Mag. What is the motivation to keeping both DoPE and Mag?

5 8

problem about ipc
by gaober 11 Jul '12

11 Jul '12

>> 2.l4_ipc_error( ),return 4,i look up in l4re document it means >> L4_ENOT_EXISTENT.but how can i solve that? >The documentation in the header file (pkg/l4sys/include/ipc.h) is quite clear about the error source: >L4_IPC_ENOT_EXISTENT = 0x04, /**< Non-existing destination or source. >could you please give us a minimal code example (including the lua script and both communcation partners) that >causes this? You communication partner has probably terminated or something like that. i want to use touchscreen on my omap3530 board.then i add REQUIRES_LIBS_arm = libdrivers-input in file pkg/input/lib/src/makefie. then in function tsc_irq_func( ) in file pkg/drivers/input/src/tsc-omap3.c , l4_ipc_err( tag,l4_utcb()) return 4. the arm-omap3.devs # OMAP3 (OMAP3EVM, Beagleboard) hw-root { sys => new Device() { .hid = "System Control"; new-res Mmio(0x48002000 .. 0x48002fff); new-res Mmio(0x48004000 .. 0x48004fff); } prcm => new Device() { .hid = "OMAP PRCM"; new-res Mmio(0x48004000 .. 0x48004fff); new-res Mmio(0x48005000 .. 0x48005fff); } gpio => new Device() { .hid = "GPIO"; new-res Mmio(0x48004000 .. 0x48004fff); new-res Mmio(0x48310000 .. 0x48310fff); new-res Mmio(0x49050000 .. 0x49050fff); new-res Mmio(0x49052000 .. 0x49052fff); new-res Mmio(0x49054000 .. 0x49054fff); new-res Mmio(0x49056000 .. 0x49056fff); new-res Mmio(0x49058000 .. 0x49058fff); } lcd => new Device() { .hid = "OMAP_LCD"; new-res Mmio(0x48050000 .. 0x48050fff); } i2c1 => new Device() { .hid = "I2C"; new-res Mmio(0x48070000 .. 0x48070fff); } kp => new Device() { .hid = "OMAP_KP"; new-res Mmio(0x48070000 .. 0x48070fff); } mcspi => new Device() { .hid = "McSPI"; new-res Mmio(0x48098000 .. 0x48098fff); } tsc => new Device() { .hid = "OMAP_TSC"; new-res Mmio(0x48098000 .. 0x48098fff); new-res Irq(32); } } the l4_android-arm.vbus file is # vim:set ft=ioconfig: # configuration file for io gui => new System_bus() { #KBD => wrap(hw-root.KBD); #MOUSE => wrap(hw-root.MOUSE); tsc => wrap(hw-root.tsc); mcspi => wrap(hw-root.mcspi); kp => wrap(hw-root.kp); i2c1 => wrap(hw-root.i2c1); gpio => wrap(hw-root.gpio); } fbdrv => new System_bus() { sys => wrap(hw-root.sys); lcd => wrap(hw-root.lcd); } android_bus => new System_bus() { } the cfg file is require("L4"); local lxname = "vmlinuz.android"; local l = L4.default_loader; local io_caps = { sigma0 = L4.cast(L4.Proto.Factory, L4.Env.sigma0):create(L4.Proto.Sigma0); gui = L4.default_loader:new_channel():svr(); android_bus = L4.default_loader:new_channel():svr(); fbdrv = L4.default_loader:new_channel():svr(); rom = L4.Env.rom; icu = L4.Env.icu; }; local fb_caps = { vbus = io_caps.fbdrv; rom = L4.Env.rom; fb = L4.default_loader:new_channel():svr(); }; L4.default_loader:startv( { caps = io_caps, log = { "io", "red" }, }, "rom/io", "rom/arm-omap3.devs", "rom/l4_android-arm.vbus"); L4.default_loader:startv( { caps = fb_caps, log = { "fb", "cyan" }, }, "rom/fb-drv"); local mag_caps = { mag = l:new_channel(), svc = l:new_channel(), }; l:start({ caps = { vbus = io_caps.gui, fb = fb_caps.fb, mag = mag_caps.mag:svr(), svc = mag_caps.svc:svr(), }, log = { "mag", "g" }, l4re_dbg = L4.Dbg.Warn, }, "rom/mag"); local android_caps = { log = L4.Env.log:m("rws"); vbus = io_caps.android_bus; rom = L4.Env.rom; fb = mag_caps.svc:create(L4.Proto.Goos, "g=800x476+0+0"); }; L4.default_loader:start( { caps = android_caps, l4re_dbg = L4.Dbg.Warn, log = { "android", "yellow" }, }, "rom/" .. lxname .. " console=ttyLv0 androidboot.console=ttyLv0 mem=128M l4bdds.add=rom/system-arm-ginger.img,rw boot=local android.ril=ttyLv1 ip=dhcp l4x_rd=rom/root-arm-ginger.img rw root=1:0 ramdisk_size=1500 init=/init --debug l4fb.touchscreen=1"); thanks!

2 1

Re: l4linux GPU direct access
by Abhishek Gupta 11 Jul '12

11 Jul '12

Hi Adam, It is not decided yet which graphics card will be used. Basically, we are just trying to check whether the idea of doing this works or not. If I can get xserver running on l4linux on kvm then the graphics card will be decided. Do you have any suggestions?? Abhishek Gupta

2 1

Questions about L4 virtual machines
by ZhangEditing 09 Jul '12

09 Jul '12

Hi! I'm trying to port an OSEK OS to L4. I know that a guest OS runs as a L4 process from Fiasco's point of view. But I have also noticed relevent APIs such as the l4vcpu_* family, VM API for SVM, VMX, and TZ, and the vCPU API of threads. My questions are as follows: 1. What are the differences between a VM and an ordinary process? 2. Is it possible to run a guest OS without utilizing VM/vCPU APIs? 3. What does a guest OS have to do in order to run on L4Re and Fiasco? Is there a guide for doing this? I'm new to L4 and probably need concrete instructions. Thank you in advance!

3 2

Tasks priority and time slice
by Pierre Larus 09 Jul '12

09 Jul '12

Hi, I am running two para-virtualized OS on L4RE. I would like to define tasks priority and time slice. The real time scheduling extensions does not seem to be implemented on ARM. But I have find this code in a Ned example: scheduler = L4.Env.mem_alloc:create(L4.Proto.Scheduler, 0x78, 0x70) I don't understand how to use second and third arguments. Regards, Pierre Larus

2 1

bbclone fbdrv lcd problem
by gaober 09 Jul '12

09 Jul '12

hiHi, On Wed Jun 22, 2011 at 12:54:50 +0400, Vasily Sartakov wrote: >> I try to start ex_fb_spectrum example on beagleboard clone with 7" >> 480x800 LCD and have error in fbdrv//lcd-omap3.c: >> when a i call any l4vbus_i2c* and l4vbus_gpio function i have return >> -38 --- ENOSYS error >> whats may be wrong? >The problem is that 'io' does not currently implement any of the i2c and >gpio functionality. There was some experiment some longer time ago from >which that code is from.now i am doing some work on touchscreen driver.my board is omap3530,with tcs2046.i see there is l4vbus_mcspi* function,but i cann't find the code for the tcs control device operating.maybe the snapshot doesn't have the "server"?can you offer me some example ,maybe some codes of lcd, i think it's great useful for me.thank you very much! >For the beagleboard the LCD is configured by u-boot and can then just be >used. Isn't that approach possible with your board? Wilbert Gao

2 1

Re: l4linux GPU direct access
by Abhishek Gupta 09 Jul '12

09 Jul '12

Hi Adam, I have checked the xserver code and found that the DEFAULT_INT10 was already set to x86emu and it was already using x86emu to execute the int10. Also, there is one more mode vm86 in which int10 can be executed but it uses the vm86 system call which is not supported in l4linux. Is there a way to add vm86 system call to the l4linux? Also, I have tried with video drivers cirrus, qxl but none of them worked. I tried Ubuntu on vmware player and it uses vmware video driver to run xserver and does not uses int10. Abhishek Gupta

2 3

problem about ipc
by gaober 09 Jul '12

09 Jul '12

hi! 1.i find there is l4_ipc_error( ) ops like that in .h files,but where can i find the implementation of that func? 2.l4_ipc_error( ),return 4,i look up in l4re document it means L4_ENOT_EXISTENT.but how can i solve that? thanks! 发自 Android 版 Yahoo! 邮箱

2 1

L4Android effort appreciated
by John van V. 07 Jul '12

07 Jul '12

Hello all, I just want to send out a message of appreciation (speaking for a world of users) thanking you for all this effort to put the user directly onto the OS, that is by removing the long-obsolete Linux monolithic kernel. Ultimately, I would like to see "the app" go away and to see the OS directly to support the human mind to work as it does naturally collaboratively, the whole point behind internetworking. Thanks, again...

1 0

Mount SD card partition as rootfs
by Pierre Larus 04 Jul '12

04 Jul '12

Hi, I am currently trying to mount a SD card partition as L4Linux rootfs. I am using a beagleboard. The first flash card partition (fat32) is mounted by U-boot and contains the L4Linux kernel . The second partition (ext3) contains the rootfs. I have compiled L4Linux to support SD card et Ext3 filesystem and configured L4 to give to L4Linux access to the MMC device through a virtual bus. But the SD card partitions are not available in L4Linux. Do I forget something ? Regards, Pierre Larus ------------------------------------------------------------------- entry l4LinuxUcos kernel fiasco -serial_esc roottask moe rom/l4lx.cfg module l4re module ned module io module l4lx.cfg module vmlinuz.arm module beagle.devs module beagle.io ------------------------------------------------------------------- hw-root { # MultiMediaCard 1 2 3 mmc => new Device() { .hid = "OMAP_HSMMC"; new-res Mmio(0x4809c000 .. 0x4809dfff); new-res Mmio(0x480b4000 .. 0x480b5fff); new-res Mmio(0x480ad000 .. 0x480aefff); } } mmiodev => new System_bus() { mmc => wrap(hw-root.mmc); } ------------------------------------------------------------------- local lxname = "vmlinuz.arm"; local DeviceFile = "beagle.devs" local IOFile = "beagle.io" local loader = L4.default_loader; local io_bus = loader:new_channel(); loader:start({ caps = { mmiodev = io_bus:svr(), icu = L4.Env.icu, sigma0 = L4.cast(L4.Proto.Factory, L4.Env.sigma0):create(L4.Proto.Sigma0), }, }, "rom/io -vv rom/" .. DeviceFile .. " rom/" .. IOFile); loader:start( { caps = { vbus = io_bus, icu = L4.Env.icu, log = L4.Env.log:m("rws"), }, l4re_dbg = L4.Dbg.Warn, log = { "l4linux", "yellow" }, }, "rom/" .. lxname .. " root=/dev/mmcblk0p2 rw rootfstype=ext3 init=/sbin/init console=ttyLv0"); --------------------------------------------------------------------------------------- ... L4 serial driver ttyLv0 at MMIO 0x1 (irq = 210) is a L4 L4 shared mem serial driver l4cdds: No name given, not starting. brd: module loaded l4bdds: No name given, not starting. mousedev: PS/2 mouse device common for all mice sdhci: Secure Digital Host Controller Interface driver sdhci: Copyright(c) Pierre Ossman TCP cubic registered NET: Registered protocol family 17 VFS: Cannot open root device "mmcblk0p2" or unknown-block(0,0) Please append a correct "root=" boot option; here are the available partitions: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0) [<00129d68>] (unwind_backtrace+0x0/0xf0) from [<00335858>] (panic+0x58/0x178) [<00335858>] (panic+0x58/0x178) from [<00100aa0>] (mount_block_root+0x1d4/0x294) [<00100aa0>] (mount_block_root+0x1d4/0x294) from [<00100e24>] (prepare_namespace +0x164/0x1bc) [<00100e24>] (prepare_namespace+0x164/0x1bc) from [<00100794>] (kernel_init+0xe4 /0x118) [<00100794>] (kernel_init+0xe4/0x118) from [<00122410>] (kernel_thread_exit+0x0/ 0x8) panic: going to sleep forever, bye l4linux | panic: going to sleep forever, bye

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

l4-hackers July 2012