- l4-hackers - example.com Site

question about read clock API in ARM
by 박찬주 07 Aug '11

07 Aug '11

Hello, I'm newbie in L4 and now i'm trying to find out how to get system clock info. My test environment is QEMU, ARM realview EB. In fiasco kernel, there is a function to read system's clock Clock::read_count(){ return Io::read<Mword>(SYS_24MHZ); } then how can I use this in l4re application? Is there any related API in l4re? Thanks in advance, david

2 1

Sigma0 raised exception on boot
by Anna Fischer 04 Aug '11

04 Aug '11

I'm running Fiasco on the PandaBoard. On platform boot I'm getting an error that Sigma0 raised an exception and then Fiasco panics and the boot process stops. Any idea what the problem could be? I can't find any more information to debug this properly. I have compile Fiasco.OC and L4Re for the PandaBoard platform. The board has 1GB memory. L4 Bootstrapper Build: #162 Wed Aug 3 12:52:02 BST 2011, 4.5.2 Scanning up to 1024 MB RAM Memory size is 1024MB (80000000 - c0000000) RAM: 0000000080000000 - 00000000bfffffff: 1048576kB Total RAM: 1024MB [...] Relocated mbi to [0x8100f000-0x8100f1b1] Loading fiasco Loading sigma0 Loading moe find kernel info page... found kernel info page at 0x80002000 Regions of list regions [ 80001000, 800019ff] { a00} Kern fiasco [ 80002000, 8005cfff] { 5b000} Kern fiasco [ 80090000, 8009e17b] { e17c} Sigma0 sigma0 [ 80140000, 8017160b] { 3160c} Root moe [ 80178000, 8018eec7] { 16ec8} Root moe [ 81000000, 810143eb] { 143ec} Boot bootstrap [ 8100f000, 8100f2ae] { 2af} Root Multiboot info [ 81190000, 8194dfff] { 7be000} Root Modules Memory API Version: (87) experimental Sigma0 config ip:800900e0 sp:81013044 Roottask config ip:80140178 sp:00000000 Starting kernel fiasco at 80001000 Hello from Startup::stage2 Initialize page table Number of IRQs available at this GIC: 160 Vmem_alloc::init() Cache config: ON ID_PFR0: 00001231 ID_PFR1: 00000011 ID_DFR0: 00010444 ID_AFR0: 00000000 ID_MMFR0: 00100103 ID_MMFR1: 20000000 ID_MMFR2: 01230000 ID_MMFR3: 00102111 SERIAL ESC: allocated IRQ 106 for serial uart Not using serial hack in slow timer handler. Welcome to Fiasco.OC (arm)! L4/Fiasco.OC arm microkernel (C) 1998-2011 TU Dresden Rev: r36 compiled with gcc 4.5.2 for Pandaboard [] Build: #4 Wed Aug 3 12:46:10 BST 2011 Calibrating timer loop... done. KERNEL: Warning: Sigma0 raised an exception --> HALT Panic: ... Return reboots, "k" enters L4 kernel debugger...

2 2

Dependencies
by Bryce CR 03 Aug '11

03 Aug '11

Hi Hackers, I'm attempting to build the Fiasco.OC kernel including several additional files. When I attempt to make, I get endless complaints about multiple definitions of functions. The catch is, every single file I have added has a unique include guard... A few things I want to check: > My files are written in standard c++ format, not Preprocess compatible format. E.g. I make use of namespaces. I haven't added the files to any modules lists / Makerules files, so my thinking is that I'm just trying to link with these files but not use any of their features (for testing purposes). My impression was that namespace usage would be tolerated for files used thus: Preprocess doesn't complain. > The c++ preprocessor is still run, correct? I don't see include guards in kernel source files: are these generated by Preprocess, or simply not needed? > Specifically, the errors start appearing at "==> Archiving into jdb_compound.o," and they complain about finding a definition of a function in my source file which was already defined in an object file, e.g.: jdb-ia32-amd64.o: In function `prio_queue_rel::query_iterator_priority__element::next()': /home/bryce/parsnip/fiasco_modified_src/recovery/src/kernel/fiasco/src/kern/R_test11impl.cpp:499: multiple definition of `prio_queue_rel::query_iterator_priority__element::next()' jdb.o:/home/bryce/parsnip/fiasco_modified_src/recovery/src/kernel/fiasco/src/kern/R_test11impl.cpp:499: first defined here I know that's not a lot to go on, but I'm not sure what else would be useful. Please let me know what I do to help you help me... Thank you! --Bryce

2 2

L4Linux and mmap
by Vasily Sartakov 03 Aug '11

03 Aug '11

Hallo. I try to use a mmap for communicate userspace application and other servers via kernel character driver. and its doesn't work. but, of course its work in native linux. my mmap call return no error, but when i try acces to memmory via returned pointer, i got seg fault Here is driver source and example program, based on mmap example: any of vmalloc or kalloc got seg fault: mmap ok, vaddr = 40034000 Segmentation fault //#include <linux/config.h> #include <linux/version.h> #include <linux/init.h> #include <linux/module.h> #include <linux/fs.h> #include <linux/cdev.h> #include <linux/slab.h> #include <linux/vmalloc.h> #include <linux/mm.h> #ifdef MODVERSIONS # include <linux/modversions.h> #endif #include <asm/io.h> /* character device structures */ static dev_t vhsm_dev; static struct cdev vhsm_cdev; /* methods of the character device */ static int vhsm_open(struct inode *inode, struct file *filp); static int vhsm_release(struct inode *inode, struct file *filp); static int vhsm_mmap(struct file *filp, struct vm_area_struct *vma); /* the file operations, i.e. all character device methods */ static struct file_operations vhsm_fops = { .open = vhsm_open, .release = vhsm_release, .mmap = vhsm_mmap, .owner = THIS_MODULE, }; // internal data // length of the two memory areas #define NPAGES 16 // pointer to the vmalloc'd area - alway page aligned static int *vmalloc_area; // pointer to the kmalloc'd area, rounded up to a page boundary static int *kmalloc_area; // original pointer for kmalloc'd area as returned by kmalloc static void *kmalloc_ptr; /* character device open method */ static int vhsm_open(struct inode *inode, struct file *filp) { return 0; } /* character device last close method */ static int vhsm_release(struct inode *inode, struct file *filp) { return 0; } // helper function, mmap's the kmalloc'd area which is physically contiguous int vhsm_kmem(struct file *filp, struct vm_area_struct *vma) { int ret; long length = vma->vm_end - vma->vm_start; /* check length - do not allow larger mappings than the number of pages allocated */ if (length > NPAGES * PAGE_SIZE) return -EIO; /* map the whole physically contiguous area in one piece */ if ((ret = remap_pfn_range(vma, vma->vm_start, virt_to_phys((void *)kmalloc_area) >> PAGE_SHIFT, length, vma->vm_page_prot)) < 0) { return ret; } return 0; } /* character device mmap method */ static int vhsm_mmap(struct file *filp, struct vm_area_struct *vma) { /* at offset 0 we map the vmalloc'd area */ if (vma->vm_pgoff == 0) { return vhsm_vmem(filp, vma); } /* at offset NPAGES we map the kmalloc'd area */ if (vma->vm_pgoff == NPAGES) { return vhsm_kmem(filp, vma); } /* at any other offset we return an error */ return -EIO; } /* module initialization - called at module load time */ static int __init vhsm_init(void) { int ret = 0; int i; /* allocate a memory area with kmalloc. Will be rounded up to a page boundary */ if ((kmalloc_ptr = kmalloc((NPAGES + 2) * PAGE_SIZE, GFP_KERNEL)) == NULL) { ret = -ENOMEM; goto out; } /* round it up to the page bondary */ kmalloc_area = (int *)((((unsigned long)kmalloc_ptr) + PAGE_SIZE - 1) & PAGE_MASK); /* allocate a memory area with vmalloc. */ if ((vmalloc_area = (int *)vmalloc(NPAGES * PAGE_SIZE)) == NULL) { ret = -ENOMEM; goto out_kfree; } /* get the major number of the character device */ if ((ret = alloc_chrdev_region(&vhsm_dev, 0, 1, "mmap")) < 0) { printk(KERN_ERR "could not allocate major number for mmap\n"); goto out_vfree; } /* initialize the device structure and register the device with the kernel */ cdev_init(&vhsm_cdev, &vhsm_fops); if ((ret = cdev_add(&vhsm_cdev, vhsm_dev, 1)) < 0) { printk(KERN_ERR "could not allocate chrdev for mmap\n"); goto out_unalloc_region; } /* mark the pages reserved */ for (i = 0; i < NPAGES * PAGE_SIZE; i+= PAGE_SIZE) { SetPageReserved(vmalloc_to_page((void *)(((unsigned long)vmalloc_area) + i))); SetPageReserved(virt_to_page(((unsigned long)kmalloc_area) + i)); } /* store a pattern in the memory - the test application will check for it */ for (i = 0; i < (NPAGES * PAGE_SIZE / sizeof(int)); i += 2) { vmalloc_area[i] = (0xaffe << 16) + i; vmalloc_area[i + 1] = (0xbeef << 16) + i; kmalloc_area[i] = (0xdead << 16) + i; kmalloc_area[i + 1] = (0xbeef << 16) + i; } return ret; out_unalloc_region: unregister_chrdev_region(vhsm_dev, 1); out_vfree: vfree(vmalloc_area); out_kfree: kfree(kmalloc_ptr); out: return ret; } /* module unload */ static void __exit vhsm_exit(void) { int i; /* remove the character deivce */ cdev_del(&vhsm_cdev); unregister_chrdev_region(vhsm_dev, 1); /* unreserve the pages */ for (i = 0; i < NPAGES * PAGE_SIZE; i+= PAGE_SIZE) { SetPageReserved(vmalloc_to_page((void *)(((unsigned long)vmalloc_area) + i))); SetPageReserved(virt_to_page(((unsigned long)kmalloc_area) + i)); } /* free the memory areas */ vfree(vmalloc_area); kfree(kmalloc_ptr); } module_init(vhsm_init); module_exit(vhsm_exit); test program: #include <stdio.h> #include <unistd.h> #include <sys/mman.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <stdlib.h> #define NPAGES 16 /* this is a test program that opens the mmap_drv. It reads out values of the kmalloc() and vmalloc() allocated areas and checks for correctness. You need a device special file to access the driver. The device special file is called 'node' and searched in the current directory. To create it - load the driver 'insmod mmap_mod.o' - find the major number assigned to the driver 'grep mmapdrv /proc/devices' - and create the special file (assuming major number 254) 'mknod node c 254 0' */ int main(void) { int fd; unsigned int *vadr; unsigned int *kadr; int len = NPAGES * getpagesize(); if ((fd=open("node", O_RDWR|O_SYNC))<0) { perror("open"); exit(-1); } vadr = mmap(0, len, PROT_READ, MAP_SHARED, fd, 0); if (vadr == MAP_FAILED) { perror("mmap"); exit(-1); } if ((vadr[0]!=0xaffe0000) || (vadr[1]!=0xbeef0000) || (vadr[len/sizeof(int)-2]!=(0xaffe0000+len/sizeof(int)-2)) || (vadr[len/sizeof(int)-1]!=(0xbeef0000+len/sizeof(int)-2))) { printf("0x%x 0x%x\n", vadr[0], vadr[1]); printf("0x%x 0x%x\n", vadr[len/sizeof(int)-2], vadr[len/sizeof(int)-1]); } kadr = mmap(0, len, PROT_READ|PROT_WRITE, MAP_SHARED| MAP_LOCKED, fd, len); if (kadr == MAP_FAILED) { perror("mmap"); exit(-1); } if ((kadr[0]!=0xdead0000) || (kadr[1]!=0xbeef0000) || (kadr[len / sizeof(int) - 2] != (0xdead0000 + len / sizeof(int) - 2)) || (kadr[len / sizeof(int) - 1] != (0xbeef0000 + len / sizeof(int) - 2))) { printf("0x%x 0x%x\n", kadr[0], kadr[1]); printf("0x%x 0x%x\n", kadr[len / sizeof(int) - 2], kadr[len / sizeof(int) - 1]); } close(fd); return(0); } -- Sartakov A. Vasily

2 1

CfP: Workshop on Isolation and Integration in Embedded Systems (IIES)
by Michael Engel 01 Aug '11

01 Aug '11

Dear colleagues, the fourth Workshop on Isolation and Integration in Embedded Systems, organized by Olaf Spinczyk (TU Dortmund), Jörg Nolte (BTU Cottbus) and myself (TU Dortmund), takes place on October 9th, 2011 at ESweek in Taipei, Taiwan. Embedded systems developers today are confronted with a large variety of challenges in the design and implementation of a system. Issues such as co-location of different applications, platform integration of different embedded operating environments as well as safety and protection of embedded software systems require novel approaches to embedded systems development as well as the adaptation of technologies from mainframe and desktop computing, like virtualization and emulation. This, in turn, will change traditional embedded design processes to include hardware and software of growing complexity. The IIES workshop is a platform to discuss recent developments and novel concepts in these areas. We would be happy if you submitted a research or position paper. Submissions may be up to six pages long (ACM SIGOPS Format). Submission deadline: August 15th, 2011 Notification of acceptance: September 12th, 2011 Workshop: October 9th, 2011 Best regards, Michael Engel (michael.engel(a)tu-dortmund.de) -- Dr. Michael Engel michael.engel(a)tu-dortmund.de - engel(a)multicores.org TU Dortmund Phone: +49 231 755 6141 Faculty of Computer Science XII Fax: +49 231 755 6116 Embedded Systems Group - Otto-Hahn-Str. 16 - D-44221 Dortmund, Germany

1 0

L4Re Library Usage
by Wesley Miaw 30 Jul '11

30 Jul '11

I have been able to get my module to compile as an ext-pkg. However I'm a little confused by the description of L4Re listed at http://os.inf.tu-dresden.de/L4Re/overview.html What exactly do I need to replace in my code to use the L4Re C or C++ libraries, or libstdc++ etc.? I didn't get any compilation errors with code that directly calls malloc() and I saw some other modules call malloc() too. But I would have thought all memory allocation/deallocation needs to be changed to use an L4 function. When I looked inside packages like cxx or libstdc++ it doesn't actually look like these are replacements for the GNU headers or libraries. Is the only code I need to have use L4Re functions and types the code that is going to handle message receiving and sending? Everything else could be as if it was not part of L4? Thanks, Wes

2 1

[ANNOUNCE] NOVA Microhypervisor 0.4 prerelease
by Udo A. Steinberg 28 Jul '11

28 Jul '11

Hi, The NOVA project is happy to announce that there is a new prerelease of the NOVA microhypervisor available for download at http://www.hypervisor.org/ under the terms of the GNU General Public License version 2. NOVA is based on a modern microhypervisor written in C++ and assembler. The current implementation runs on multi-core x86 machines that support ACPI, under QEMU (including VM support), and as a microkernel in a virtual machine on top of itself. NOVA makes use of hardware virtualization features such as Intel VT-x/VT-d or AMD-V, and facilitates running unmodified guest operating systems in virtual machines with near-native performance. Among other things, the new prerelease improves the security of device- and interrupt assignment by using capabilities for these operations. There is also a new hypercall for interacting with the scheduler. At the same time, we've released a new version of the NOVA Userland (NUL). While there are few user-visible changes, a lot has changed under the hood. We have integrated a central admission server that is going to implement global policies for CPU time distribution in a future release. Another notable development is rudimentary libvirt support, which will eventually replace the old configuration syntax. Cheers, - Udo

1 0

Re: l4-hackers Digest, Vol 99, Issue 18
by Wesley Miaw 28 Jul '11

28 Jul '11

> On Mon Jul 25, 2011 at 16:43:09 -0700, Wesley Miaw wrote: >> I'm trying to add a new lib/server module to L4 but I'm having some >> trouble. I followed the snapshot top-level README suggestion to >> place the new code in an external directory, so I placed it into >> src/mymodule/ next to src/l4re and src/l4linux. Then I created a >> Makefile under src/mymodule/src/ based on the example with include >> $(L4DIR)/mk/lib.mk. >> >> However, this doesn't appear to be working. $(L4DIR)/mk/Makeconf is >> complaining that files existing inside the obj/l4/... directory are >> not to be found in the obj/mymodule/... directory. (I tried to copy >> the obj/... logic found in the top-level Makefile.) >> >> Is the README's suggestion accurate? > > I think so. Could you show me some of the failed output + surroundings > to better judge what's going on? My bad. I didn't realize the build directory (O=) had to be the l4 obj directory for my own packages. (Missed that part of the README.) It's working for me now. Thanks! -- Wesley Miaw wesley(a)wesman.net

1 0

No configuration file found in build directory
by Wesley Miaw 26 Jul '11

26 Jul '11

I'm trying to add a new lib/server module to L4 but I'm having some trouble. I followed the snapshot top-level README suggestion to place the new code in an external directory, so I placed it into src/ mymodule/ next to src/l4re and src/l4linux. Then I created a Makefile under src/mymodule/src/ based on the example with include $(L4DIR)/mk/ lib.mk. However, this doesn't appear to be working. $(L4DIR)/mk/Makeconf is complaining that files existing inside the obj/l4/... directory are not to be found in the obj/mymodule/... directory. (I tried to copy the obj/... logic found in the top-level Makefile.) Is the README's suggestion accurate? Thanks, -- Wesley Miaw wesley(a)wesman.net

2 1

transfer_msg_items
by Chen Tian 21 Jul '11

21 Jul '11

Hello, Why was a sender's send item (e.g. mr[0] and mr[1] ) merged and stored into the receiver's mr[0]? Basically I have trouble understanding the following code. *rcv_words=(item->b.raw() & ~0x0ff7) | (sfp.raw())&0x0ff0); rcv_word +=2 ; Why didn't we use receiver's mr[1] as well? In the case of page fault handling IPC, the pf address stored in sender's mr[1] is not used ? Thanks. Chen

1 1

what is the difference between the 2 methods of booting hello demo
by cheng guanghui 19 Jul '11

19 Jul '11

Hello: In the modules.list there are 2 entires to boot hello demo as following: entry hello roottask moe --init=rom/hello module l4re module hello entry hello-cfg roottask moe rom/hello.cfg module l4re module ned module hello.cfg module hello The booting logs are following separately : hello MOE: Hello world MOE: found 117648 KByte free memory MOE: found RAM from 1000 to 75c7000 MOE: allocated 117 KByte for the page array @0x1000 MOE: virtual user address space [0-bfffffff] MOE: rom name space cap -> [C:501000] BOOTFS: [136d000-1470e6d] [C:503000] l4re BOOTFS: [1471000-153030b] [C:504000] hello MOE: cmdline: /home/cheng/TUDOS/Src/l4re-snapshot-2011062216/obj/l4/x86/bin/x86_586/l4f/moe --init=rom/hello MOE: Starting: rom/hello MOE: loading 'rom/hello' Hello World! Hello World! Hello World! hello-cfg MOE: Hello world MOE: found 115604 KByte free memory MOE: found RAM from 1000 to 75c7000 MOE: allocated 117 KByte for the page array @0x1000 MOE: virtual user address space [0-bfffffff] MOE: rom name space cap -> [C:501000] BOOTFS: [136d000-1470e6d] [C:503000] l4re BOOTFS: [1471000-166e572] [C:504000] ned BOOTFS: [166f000-166f066] [C:505000] hello.cfg BOOTFS: [1670000-172f30b] [C:506000] hello MOE: cmdline: /home/cheng/TUDOS/Src/l4re-snapshot-2011062216/obj/l4/x86/bin/x86_586/l4f/moe rom/hello.cfg MOE: Starting: rom/ned rom/hello.cfg MOE: loading 'rom/ned' Ned says: Hi World! Ned: loading file: 'rom/hello.cfg' hello | Hello World! hello | Hello World! hello | Hello World! hello | Hello World! hello | Hello World! hello | Hello World! hello | Hello World! So hello demo could be booted whatever from Ned or Moe directly. In the document Ned is described as this "The main thing to do here is to coordinate the startup of services and applications as well as to provide the communication channels for them". Does it mean "We use Ned just because we need to dynamically create IPC gate (channel) between different applications". If there is no need to use IPC gate we can use Moe directly to boot application. is it correct? Thanks. Warm regards. Cheng Guanghui

2 2

how can i understand this description in the l4re-source document?
by cheng guanghui 19 Jul '11

19 Jul '11

Hello: In the page 19 of the l4re-source document there is a description about caps as following: Caps The table with application-specific named capabilities (default is an empty table). If the table does not contain a capability with the name ’rom’, the ’rom’ capability from Ned’s initial caps is inserted into the table. How can i understand "If the table does not contain a capability with the name ’rom’, the ’rom’ capability from Ned’s initial caps is inserted into the table." Thanks. Warm regards. Cheng Guanghui

2 2

writing my own filesystem for L4
by Michal Schulz 19 Jul '11

19 Jul '11

Hello. I'm going to write a filesystem for L4, using the tmpfs (l4/pkg/tmpfs) as a starting point. However, I would like to know whether there are any tools I could use to test the filesystem. The tool I would love to have should display all registered filesystem and should allow me to walk through filesystem's directory trees. Is there such a tool or example application? Are there any packages in L4 which use VFS layer at all? regards, Michal. -- Michal Schulz

2 1

Re: l4-hackers Digest, Vol 99, Issue 11
by Chen Tian 18 Jul '11

18 Jul '11

Hi Adam, I don't have a separate PHDR for ctors. I got this problem when I put ctors into the text section. Now I have moved ctors into the data section and the problem disappeared. Any ideas? Another questions is why ktext=0x90909090 in the ld file of fiasco.oc? Is this value randomly chosen? Thanks. On Sat, Jul 16, 2011 at 3:00 AM, <l4-hackers-request(a)os.inf.tu-dresden.de> wrote: > Send l4-hackers mailing list submissions to > l4-hackers(a)os.inf.tu-dresden.de > > To subscribe or unsubscribe via the World Wide Web, visit > http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers > or, via email, send a message with subject or body 'help' to > l4-hackers-request(a)os.inf.tu-dresden.de > > You can reach the person managing the list at > l4-hackers-owner(a)os.inf.tu-dresden.de > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of l4-hackers digest..." > > > Today's Topics: > > 1. Re: write my own root task. (Adam Lackorzynski) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Fri, 15 Jul 2011 23:20:50 +0200 > From: Adam Lackorzynski <adam(a)os.inf.tu-dresden.de> > To: l4-hackers(a)os.inf.tu-dresden.de > Subject: Re: write my own root task. > Message-ID: <20110715212050.GB5040(a)os.inf.tu-dresden.de> > Content-Type: text/plain; charset=iso-8859-1 > > > On Thu Jul 14, 2011 at 15:15:25 -0700, Chen Tian wrote: >> I am trying to use C++ to write my own root task. I found that if I >> declare a static or global object, bootstrap complains that two >> regions belonging to my root task overlapped and thus cannot proceed. >> Why does my root task have two conflicting regions? According to the >> bootstrap report, they both have the correct starting address, but one >> ends just at the global object's ctor in .ctors section, while the >> other region covers the entire program. >> If the .ctors section is empty then everything is fine. Any ideas? > > I wonder why you have an own PHDR for ctors? How many do you have? > > > Adam > -- > Adam adam(a)os.inf.tu-dresden.de > Lackorzynski http://os.inf.tu-dresden.de/~adam/ > > > > ------------------------------ > > _______________________________________________ > l4-hackers mailing list > l4-hackers(a)os.inf.tu-dresden.de > http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers > > > End of l4-hackers Digest, Vol 99, Issue 11 > ****************************************** >

2 1

write my own root task.
by Chen Tian 15 Jul '11

15 Jul '11

Hello, I am trying to use C++ to write my own root task. I found that if I declare a static or global object, bootstrap complains that two regions belonging to my root task overlapped and thus cannot proceed. Why does my root task have two conflicting regions? According to the bootstrap report, they both have the correct starting address, but one ends just at the global object's ctor in .ctors section, while the other region covers the entire program. If the .ctors section is empty then everything is fine. Any ideas? Thanks. Chen

2 1

TLS on amd64
by Daniel Müller 15 Jul '11

15 Jul '11

Hello, I want to use the __thread keyword gcc provides. I was told this was recently adjusted to work (at least in the private repository which I am using). It appears to be working when building for IA32 but when I compile the same program for amd64 I get a page fault when running it (it is a simple program just incrementing a __thread declared int). It faults at 10010f4 (the first access of the variable). 00000000010010f0 <main>: 10010f0: 48 83 ec 08 sub rsp,0x8 10010f4: 64 c7 04 25 fc ff ff mov DWORD PTR fs:0xfffffffffffffffc,0x0 10010fb: ff 00 00 00 00 1001100: bf 60 48 01 01 mov edi,0x1014860 1001105: e8 36 88 00 00 call 1009940 <puts> 100110a: 64 83 04 25 fc ff ff add DWORD PTR fs:0xfffffffffffffffc,0x1 1001111: ff 01 1001113: bf 01 00 00 00 mov edi,0x1 1001118: e8 d3 71 00 00 call 10082f0 <sleep> 100111d: eb e1 jmp 1001100 <main+0x10> 100111f: 90 nop Is TLS for amd64 supposed to work? Any idea on the problem? Greetings, Daniel

2 3

Running the IO server
by Anna Fischer 12 Jul '11

12 Jul '11

I'm trying to run the ARM-based frame buffer configuration from conf/examples/arm-rv-lcd.cfg with pkg/io/config/arm-rv-eb-mc.devs and QEMU (-M realview-pbx-a9). I can boot the L4 Linux kernel on this (without GUI but just serial I/O) configuration, but I get an error when trying to set up the LCD (see below). Perhaps the LCD device configuration from the .devs file does not match what hardware QEMU emulates? How can I actually enumerate in the IO server what hardware it actually runs on? Is there a way to do that? Thanks, Anna L4 Bootstrapper Build: #34 Mon Jul 11 16:17:10 BST 2011, 4.5.2 Scanning up to 256 MB RAM Memory size is 256MB (00000000 - 10000000) RAM: 0000000000000000 - 000000000fffffff: 262144kB Total RAM: 256MB mod11: 012a9000-012cac7c: ex_fb_spectrum mod10: 0121b000-012a846c: mag mod09: 011f1000-0121a5b8: fb-drv mod08: 0119d000-011f01b8: ned mod07: 010c2000-0119cb58: io mod06: 010a8000-010c1510: l4re mod05: 010a7000-010a70e4: arm-rv-lcd.io mod04: 010a6000-010a659d: arm-rv-eb-mc.devs mod03: 010a5000-010a56bb: arm-rv-lcd.cfg mod02: 0106b000-010a45f0: moe mod01: 01061000-0106a33c: sigma0 mod00: 01015000-0106014c: fiasco Moving 12 modules to 1100000 with offset eb000 moving module 12 { 12a9000-12cac7c } -> { 1394000-13b5c7c } moving module 11 { 121b000-12a846c } -> { 1306000-139346c } moving module 10 { 11f1000-121a5b8 } -> { 12dc000-13055b8 } moving module 09 { 119d000-11f01b8 } -> { 1288000-12db1b8 } moving module 08 { 10c2000-119cb58 } -> { 11ad000-1287b58 } moving module 07 { 10a8000-10c1510 } -> { 1193000-11ac510 } moving module 06 { 10a7000-10a70e4 } -> { 1192000-11920e4 } moving module 05 { 10a6000-10a659d } -> { 1191000-119159d } moving module 04 { 10a5000-10a56bb } -> { 1190000-11906bb } moving module 03 { 106b000-10a45f0 } -> { 1156000-118f5f0 } moving module 02 { 1061000-106a33c } -> { 114c000-115533c } moving module 01 { 1015000-106014c } -> { 1100000-114b14c } Scanning fiasco -serial_esc Scanning sigma0 Scanning moe rom/arm-rv-lcd.cfg Relocated mbi to [0x100f000-0x100f19e] Loading fiasco Loading sigma0 Loading moe find kernel info page... found kernel info page at 0x2000 Regions of list regions [ 1000, 19ff] { a00} Kern fiasco [ 2000, 5cfff] { 5b000} Kern fiasco [ 90000, 9e17b] { e17c} Sigma0 sigma0 [ 140000, 17160b] { 3160c} Root moe [ 178000, 18eec7] { 16ec8} Root moe [ 1000000, 10143eb] { 143ec} Boot bootstrap [ 100f000, 100f29b] { 29c} Root Multiboot info [ 1190000, 13b5c7b] { 225c7c} Root Modules Memory API Version: (87) experimental Sigma0 config ip:000900e0 sp:01013044 Roottask config ip:00140178 sp:00000000 Starting kernel fiasco at 00001000 Realview System ID: Rev=0 HBI=178 Build=0 Arch=5 FPGA=00 L2: ID=00000000 Type=00000000 L2 cache enabled Hello from Startup::stage2 Initialize page table Number of IRQs available at this GIC: 96 Vmem_alloc::init() Cache config: ON ID_PFR0: 00001031 ID_PFR1: 00000011 ID_DFR0: 00000000 ID_AFR0: 00000000 ID_MMFR0: 00100103 ID_MMFR1: 20000000 ID_MMFR2: 01230000 ID_MMFR3: 00002111 SERIAL ESC: allocated IRQ 44 for serial uart Not using serial hack in slow timer handler. Welcome to Fiasco.OC (arm)! L4/Fiasco.OC arm microkernel (C) 1998-2011 TU Dresden Rev: r33 compiled with gcc 4.5.2 for Realview [] Build: #1 Tue Jun 21 12:57:15 BST 2011 Calibrating timer loop... done. SIGMA0: Hello! KIP @ 2000 allocated 4KB for maintenance structures SIGMA0: Dump of all resource maps RAM:------------------------ [0:0;fff] [0:5d000;8ffff] [0:9f000;13ffff] [4:140000;171fff] [0:172000;177fff] [4:178000;18efff] [0:18f000;100efff] [4:100f000;100ffff] [0:1010000;118ffff] [4:1190000;13b5fff] [0:13b6000;effffff] IOMEM:---------------------- [0:10000000;ffffffff] MOE: Hello world MOE: found 242832 KByte free memory MOE: found RAM from 5d000 to f000000 MOE: allocated 239 KByte for the page array @0x9f000 MOE: virtual user address space [0-bfffffff] MOE: rom name space cap -> [C:501000] BOOTFS: [1190000-11906bb] [C:503000] arm-rv-lcd.cfg BOOTFS: [1191000-119159d] [C:504000] arm-rv-eb-mc.devs BOOTFS: [1192000-11920e4] [C:505000] arm-rv-lcd.io BOOTFS: [1193000-11ac510] [C:506000] l4re BOOTFS: [11ad000-1287b58] [C:507000] io BOOTFS: [1288000-12db1b8] [C:508000] ned BOOTFS: [12dc000-13055b8] [C:509000] fb-drv BOOTFS: [1306000-139346c] [C:50a000] mag BOOTFS: [1394000-13b5c7c] [C:50b000] ex_fb_spectrum MOE: cmdline: moe rom/arm-rv-lcd.cfg MOE: Starting: rom/ned rom/arm-rv-lcd.cfg MOE: loading 'rom/ned' Ned says: Hi World! Ned: loading file: 'rom/arm-rv-lcd.cfg' IO | Io service mag | libio: Warning: Finding 'icu' in system bus failed with 'Void capability invoked' mag | Hello from MAG fbdrv | libio: Warning: Finding 'icu' in system bus failed with 'Void capability invoked' fbdrv | Could not find LCD. fbdrv | Failed to setup Framebuffer mag | Error: Exception: Void capability invoked: requesting goos info: backtrace ... mag | mag | MAG: destroy FB svc Ned[lua]: Hm, cpapbility 42d000 does not support the meta protocol: -2004 Ned: ERROR: rom/arm-rv-lcd.cfg:62: attempt to call method 'create' (a nil value)

2 2

Fix IO-APIC/IRQ-Routing Bug (affects GSIs 24, 48, 72, ...)
by Christian Ehrhardt 08 Jul '11

08 Jul '11

Hi, there seems to by an off-by-one error in Io_apic::find_apic() that kills IRQ-routing for GSIs that are a multiple of 24. Please consider the patch below (tested on real hardware with 3 IO-APICS). regards Christian diff --git a/src/kernel/fiasco/src/kern/io_apic.cpp b/src/kernel/fiasco/src/kern/io_apic.cpp index 88ffe68..897a1a0 100644 --- a/src/kernel/fiasco/src/kern/io_apic.cpp +++ b/src/kernel/fiasco/src/kern/io_apic.cpp @@ -331,7 +331,7 @@ Io_apic::find_apic(unsigned irqnum) { for (unsigned i = Max_ioapics; i > 0; --i) { - if (_apics[i-1]._apic && _apics[i-1]._offset < irqnum) + if (_apics[i-1]._apic && _apics[i-1]._offset <= irqnum) return i - 1; } return 0;

1 0

Newbie question: L4Linux as a patch
by Pablo Chacin 07 Jul '11

07 Jul '11

Hi all I'm doing a review of options for virtualization in Linux. From the documentation, it seems that L4Linux is a complete port of the Linux Kernel. I'm wondering if there is a patch (or series of patches) that could be used to modify an standard Linux kernel and make it compatible with the L4 virtualization? The question comes from my scenario: I need to deploy a customized linux kernel and want to do so over L4. This kernel is build from the standard kernel applying some patches. I would like to do the same with L4Linux, so the resulting kernel has my patches plus L4. Is this feasible? has been tested? thanks in advance and please consider I'm new to L4Linux. Regards -- ======================================================================== 0 0 0 Pablo Chacin | Dept. d'Arquitectura de Computadors 0 0 0 e-mail: pchacin(a)ac.upc.edu | UPC-Campus Nord. Modul D6-212 0 0 0 phone: +34 +93 405 4059 | Jordi Girona, 1-3 U P C fax: +34 +93 401 7055 | 08034 Barcelona - SPAIN www: http://personals.ac.upc.es/pchacin ========================================================================

2 3

Fiasco.OC Kernel symbolic debugger
by Daniel Waddington 06 Jul '11

06 Jul '11

Hello, Adam, are you likely to get the source-level symbolic debugging working in jdb in the near future? Using objdump is a bit drastic! ;-) Thanks Daniel

2 1

L4 for amd64
by Daniel Müller 05 Jul '11

05 Jul '11

Hi, I have some problems with running L4 for amd64 (i586 is just fine). I am trying to run (qemu-14.1.) the hello world binary and get: --------- L4 Bootstrapper Build: #1 Mi 29. Jun 17:41:45 CEST 2011, x86-64, 4.4.5 RAM: 0000000000000000 - 000000000009f3ff: 637kB RAM: 0000000000100000 - 0000000007ffcfff: 130036kB Total RAM: 127MB Moving 7 modules to 2000000 with offset fa8000 moving module 07 { 1679000-1739a52 } -> { 2621000-26e1a52 } moving module 06 { 1678000-1678039 } -> { 2620000-2620039 } moving module 05 { 1566000-167779e } -> { 250e000-261f79e } moving module 04 { 132d000-1565b9e } -> { 22d5000-250db9e } moving module 03 { 110c000-132cb58 } -> { 20b4000-22d4b58 } moving module 02 { 10be000-110b397 } -> { 2066000-20b3397 } moving module 01 { 1058000-10bdb00 } -> { 2000000-2065b00 } Scanning /fiasco -serial_esc -comspeed 115200 -comport 1 -nokdb This is an invalid binary, fix it (wrong ELF architecture). Key press reboots... Rebooting. --------- This is done with a clean checkout URL: http://svn.tudos.org/repos/oc/tudos/trunk Repository Root: http://svn.tudos.org/repos/oc/tudos Repository UUID: d050ee49-bd90-4346-b210-929a50b99cfc Revision: 33 configured for K8-type CPU. /usr/bin/file tells me: build/obj/bin/amd64_K8/l4f/sigma0: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped build/obj/bin/amd64_K8/l4f/moe: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped build/obj/bin/amd64_K8/l4f/ned: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped build/obj/bin/amd64_K8/l4f/l4re: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped build/obj/bin/amd64_K8/l4f/loader: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped build/obj/bin/amd64_K8/l4f/hello: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), statically linked, not stripped build/obj/bin/amd64_K8/bootstrap: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), statically linked, not stripped ./../kernel/fiasco/build/fiasco: symbolic link to `main' So why is bootstrap not built as 64 bit binary? Daniel

2 3

libc environment for l4linux
by Vasily Sartakov 04 Jul '11

04 Jul '11

Hello Does l4re snapshot has a libc environment for build a applications for l4linux? what toolchains did you use for build busybox? -- Sartakov A. Vasily

2 1

Fiasco.OC
by Daniel Waddington 03 Jul '11

03 Jul '11

Hi, I want to use 32bit x86 Fiasco.OC with lots of SMP CPUs (48+). Can you tell me where the current 32 (for 32bit) and 64 (for 64bit) limitations exist in the code - and suggest whether this could be changed? Thanks Daniel Waddington

2 1

the booting error of l4linux
by cheng guanghui 02 Jul '11

02 Jul '11

Hello: I downloaded the latest l4re-snapshot-2011062216. All the compilation is ok. The hello demo could boot successfully. But when i try to boot L4Linux with this command : make qemu E=L4Linux-mag-x86 O=/home/cheng/TUDOS/Src/l4re-snapshot-2011062216/obj/l4/x86 There are some errors as following: Allocate cpu_mem @ 0xfc54c000 Local APIC[00]: version=11 max_lvt=5 APIC ESR value before/after enabling: 00000000/00000000 Using the PIT (i8254) on IRQ 2 for scheduling Absolute KIP Syscalls using: Sysenter CPU[0:0]: GenuineIntel (6:3:3:0)[00000633] Model: Pentium II (Klamath) at 2395 MHz 32 KB L1 I Cache (8-way associative, 64 bytes per line) 32 KB L1 D Cache (8-way associative, 64 bytes per line) 2048 KB L2 U Cache (8-way associative, 64 bytes per line) Freeing init code/data: 28672 bytes (7 pages) Calibrating timer loop... done. SIGMA0: Hello! KIP @ 400000 Found Fiasco: KIP syscalls: yes allocated 4KB for maintenance structures SIGMA0: Dump of all resource maps RAM:------------------------ [0:1000;9efff] [0:10b000;13ffff] [4:140000;18bfff] [0:18c000;2dcfff] [4:2dd000;2ddfff] [0:2de000;3fffff] [0:465000;136cfff] [4:136d000;166ffff] [0:1670000;1614bfff] IOMEM:---------------------- [0:0;fff] [0:9f000;fffff] [0:17ffd000;febfffff] [0:fec01000;fedfffff] [0:fee01000;ffffffff] IO PORTS-------------------------- [0:0;fffffff] MOE: Hello world MOE: found 357544 KByte free memory MOE: found RAM from 1000 to 1614c000 MOE: allocated 353 KByte for the page array @0x1000 MOE: virtual user address space [0-bfffffff] MOE: rom name space cap -> [C:501000] BOOTFS: [136d000-1470e6d] [C:503000] l4re BOOTFS: [1471000-166e572] [C:504000] ned BOOTFS: [166f000-166f578] [C:505000] l4lx-gfx.cfg MOE: cmdline: /home/cheng/TUDOS/Src/l4re-snapshot-2011062216/obj/l4/x86/bin/x86_586/l4f/moe rom/l4lx-gfx.cfg MOE: Starting: rom/ned rom/l4lx-gfx.cfg MOE: loading 'rom/ned' Ned says: Hi World! Ned: loading file: 'rom/l4lx-gfx.cfg' io | L4Re: FATAL: Exception No such object found: 'rom/io' io | Exception: No such object found: rom/io: backtrace ... fb-drv | L4Re: FATAL: Exception No such object found: 'rom/fb-drv' fb-drv | Exception: No such object found: rom/fb-drv: backtrace ... mag | L4Re: FATAL: Exception No such object found: 'rom/mag' mag | Exception: No such object found: rom/mag: backtrace ... Ned[lua]: Hm, cpapbility 430000 does not support the meta protocol: -2004 Ned: ERROR: rom/l4lx-gfx.cfg:54: attempt to call method 'create' (a nil value) Do i miss something else? Thanks a lot. Warm regards. Cheng Guanghui

2 6

pthread_exit in L4re
by Chen Tian 30 Jun '11

30 Jun '11

Hello, I want a thread to exit after it finishes its job. The function pthread_exit seems not working. I still can see all threads I created through kernel debugger. Any idea? Best, Chen

2 1