- l4-hackers - example.com Site

multiple servers
by Paul Phillips 21 Jan '99

21 Jan '99

I have successfully booted L4-Linux under Fiasco. There are problems, but I will work them out eventually. Thanks for a neat platform to work on. I plan to do lots of work on Fiasco and L4. Right now I'm trying to run two servers concurrently. This is just a "proof of concept" for now. I am trying to run the "hello" server (relocated to 0x2400000) along with the L4-Linux server. I load the "hello"server first and then L4-Linux. L4-linux gets to the "Calibrating delay loop.." and hangs. I am obviously forgetting something here. Is the printf used by the hello server causing a problem? BTW - I ran across the problem of the "relocation truncated ...." The problem is in the oskit/libsmp/x86/boot.S which contains the trampoline code. When the jump is made to the 32 bit code some linkers have a problem resolving the long jump. ljmp $KERNEL_CS,......... The fix is to put the ".code32" before the jump. This problem has been fixed in the latest release of the OSkit. Thanks in advance. Paul Phillips

1 0

Re: TCB size in ix86 L4
by Michael Hohmuth 17 Jan '99

17 Jan '99

"Adam 'WeirdArms' Wiggins" <awiggins(a)cse.unsw.edu.au> writes: > Is L4/Alpha suitable for real time (ie DROPS) or similar to > L4/ix86? I think it's more like L4/x86, but Sebastian would be the definitive source for this information. Michael -- hohmuth(a)innocent.com, hohmuth(a)sax.de http://www.sax.de/~hohmuth/

1 0

Re: Fiasco Bug Tracking System
by Michael Hohmuth 17 Jan '99

17 Jan '99

Tom Linden <tom(a)kednos.com> writes: > Just curious why you chose jitterbug rather than, say, gnats. I am in > the process of setting up gnats and would be intyerested in your thoughts. > Certainly the interface is more modern. Which package's interface is more modern? I don't know any Gnats interface which is very ``modern.'' I chose Jitterbug because it's easy to set up and operate. It's interface is simplistic, but I like that. Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Tyson vs Botha Wannna Bet!
by 5hbi＠yahoo.com 16 Jan '99

16 Jan '99

Feeling Lucky?? CASH IN ON thE FIGHT!!!! - http://209.67.19.99/want_to_bet/ TYSON V.s. BOTHA SATURDAY NIGHT JAN 16 9pm! REAL LIVE LAS VEGAS GAMBLING Place your bet online with up to the minute odds! Mike Tyson -750 Frans Botha +500 http://209.67.19.99/want_to_bet/ <a href="http://209.67.19.99/want_to_bet/index.html">CLICK HERE</a> to enter

1 0

Re: L4-Linux fails to boot - hangs setting up IDT
by Paul Phillips 15 Jan '99

15 Jan '99

Hi Jean, You were right, this was a problem encountered befor. The "magic" in the Makefile wasn't implemented because I did not have cpp in my path. I found a reference to the fix in an email to Sun-Mi Yu. I'm not home free yet, but I' closer. Thanks again! Paul Phillips -----Original Message----- From: Jean Wolter <jw5(a)os.inf.tu-dresden.de> To: Paul Phillips <pphillips(a)ivue.com> Cc: l4-hackers(a)os.inf.tu-dresden.de <l4-hackers(a)os.inf.tu-dresden.de> Date: Friday, January 15, 1999 4:53 AM Subject: Re: L4-Linux fails to boot - hangs setting up IDT >"Paul Phillips" <pphillips(a)ivue.com> writes: > >> 1.) Why are you calculating the address of the table by subtracting >> out EMULIB_CODE_ADDRESS (Which is 0xA0008000?) ? > >The emulation library is relocated to address 0xa0008000 and idt_table >is a symbol of the emulation library. If you have a look at the symbol >table you will find > >a00084a8 A idt_table > >So ((unsigned long)&idt_table - EMULIB_CODE_ADDRESS) is the offset of >the idt within our emulation library. > > >> 2.) How & why is the emulator code moved to "special page". Where is >> the "special page" set up since paging isn't set up until after this >> code is executed? > > >We have to copy the code to ensure that it is page aligned. The >special page is mapped into the address space of every unix process >when it raises its first page fault. Every unix process (L4Linux >process) starts at a fixed address, so we know, that the very first >page fault will request the emulation library page. > >> 3.) Any clue why this is hanging here? I should have a remote >> debugger set up soon, but I have -nokdb set and I don't get any >> kernel panics or reg dumps. The system just hangs. > >If my memory serves right, someone else had the same problem. For some >strange reason L4Linux raised a page fault at this point and the page >fault could not be resolved by the underlying pager. The L4Linux >server now hangs in an endless loop trying to resolve the page >fault. But I'm not sure about that. Michael? > >There is a possibility to monitor all page faults in the system, but I >don't know, how you activate that. > >Jean >-- >I get up each morning, gather my wits. >Pick up the paper, read the obits. >if I'm not there I know I'm not dead. >So I eat a good breakfast and go back to bed. Peete Seeger

1 0

Re: L4-Linux fails to boot - hangs setting up IDT
by Paul Phillips 15 Jan '99

15 Jan '99

Thanks for the input Jean. My symbol table shows: 00000000 A emulib_start 00000328 A user_level_pager 000003a0 A _user_task_init 00000478 A emulib_idt_descr 00000480 A idt_table 00000520 A emulib_debug_idt_descr 00000741 A emulib_end I am using gcc version 2.7.2.3 under Red Hat Linux 5.2. I am compiling the l4-981207 release of the L4-Linux code. Does this shed any light on the problem? Thanks again. Paul Phillips -----Original Message----- From: Jean Wolter <jw5(a)os.inf.tu-dresden.de> To: Paul Phillips <pphillips(a)ivue.com> Cc: l4-hackers(a)os.inf.tu-dresden.de <l4-hackers(a)os.inf.tu-dresden.de> Date: Friday, January 15, 1999 4:53 AM Subject: Re: L4-Linux fails to boot - hangs setting up IDT >"Paul Phillips" <pphillips(a)ivue.com> writes: > >> 1.) Why are you calculating the address of the table by subtracting >> out EMULIB_CODE_ADDRESS (Which is 0xA0008000?) ? > >The emulation library is relocated to address 0xa0008000 and idt_table >is a symbol of the emulation library. If you have a look at the symbol >table you will find > >a00084a8 A idt_table > >So ((unsigned long)&idt_table - EMULIB_CODE_ADDRESS) is the offset of >the idt within our emulation library. > > >> 2.) How & why is the emulator code moved to "special page". Where is >> the "special page" set up since paging isn't set up until after this >> code is executed? > > >We have to copy the code to ensure that it is page aligned. The >special page is mapped into the address space of every unix process >when it raises its first page fault. Every unix process (L4Linux >process) starts at a fixed address, so we know, that the very first >page fault will request the emulation library page. > >> 3.) Any clue why this is hanging here? I should have a remote >> debugger set up soon, but I have -nokdb set and I don't get any >> kernel panics or reg dumps. The system just hangs. > >If my memory serves right, someone else had the same problem. For some >strange reason L4Linux raised a page fault at this point and the page >fault could not be resolved by the underlying pager. The L4Linux >server now hangs in an endless loop trying to resolve the page >fault. But I'm not sure about that. Michael? > >There is a possibility to monitor all page faults in the system, but I >don't know, how you activate that. > >Jean >-- >I get up each morning, gather my wits. >Pick up the paper, read the obits. >if I'm not there I know I'm not dead. >So I eat a good breakfast and go back to bed. Peete Seeger

1 0

Re: L4-Linux fails to boot - hangs setting up IDT
by Jean Wolter 15 Jan '99

15 Jan '99

"Paul Phillips" <pphillips(a)ivue.com> writes: > 1.) Why are you calculating the address of the table by subtracting > out EMULIB_CODE_ADDRESS (Which is 0xA0008000?) ? The emulation library is relocated to address 0xa0008000 and idt_table is a symbol of the emulation library. If you have a look at the symbol table you will find a00084a8 A idt_table So ((unsigned long)&idt_table - EMULIB_CODE_ADDRESS) is the offset of the idt within our emulation library. > 2.) How & why is the emulator code moved to "special page". Where is > the "special page" set up since paging isn't set up until after this > code is executed? We have to copy the code to ensure that it is page aligned. The special page is mapped into the address space of every unix process when it raises its first page fault. Every unix process (L4Linux process) starts at a fixed address, so we know, that the very first page fault will request the emulation library page. > 3.) Any clue why this is hanging here? I should have a remote > debugger set up soon, but I have -nokdb set and I don't get any > kernel panics or reg dumps. The system just hangs. If my memory serves right, someone else had the same problem. For some strange reason L4Linux raised a page fault at this point and the page fault could not be resolved by the underlying pager. The L4Linux server now hangs in an endless loop trying to resolve the page fault. But I'm not sure about that. Michael? There is a possibility to monitor all page faults in the system, but I don't know, how you activate that. Jean -- I get up each morning, gather my wits. Pick up the paper, read the obits. if I'm not there I know I'm not dead. So I eat a good breakfast and go back to bed. Peete Seeger

1 0

Fiasco Bug Tracking System
by Michael Hohmuth 14 Jan '99

14 Jan '99

Hackers, I have set up a new bug tracking system for Fiasco; I hope you'll like it. I've already imported a few of your bug reports, but it's well possible that I've forgotten some. Please visit <URL:http://os.inf.tu-dresden.de/fiasco/bugs/> and tell me what you think. Use this page to report new (or old) bugs, or email your reports to <fiasco-bugs(a)os.inf.tu-dresden.de>, to make sure they won't get lost. Michael -- hohmuth(a)innocent.com, hohmuth(a)sax.de http://www.sax.de/~hohmuth/

2 1

L4-Linux fails to boot - hangs setting up IDT
by Paul Phillips 14 Jan '99

14 Jan '99

I am trying to boot L4-Linux using the Fiasco microkernel. The microkernel boots fine, but when L4Linux is booting, in module chead.c you have the code: /* create idt */ { extern long idt_table; struct entry_t {unsigned short w0, w1,w2,w3;} ; struct entry_t *entries = (struct entry_t *)((unsigned long)emu_lib_text + ((unsigned long)&idt_table - EMULIB_CODE_ADDRESS)); unsigned short int tmp; int i; for (i=0; i<20; i++, entries++) { tmp = entries->w1; entries->w1 = entries->w3; entries->w3 = tmp; } } My system hangs when it attempts to execute this code. I understand why you are swapping the shorts in the table (swapping the offset bytes with the selector bytes) but there are a couple of things I don't understand Several questions: 1.) Why are you calculating the address of the table by subtracting out EMULIB_CODE_ADDRESS (Which is 0xA0008000?) ? 2.) How & why is the emulator code moved to "special page". Where is the "special page" set up since paging isn't set up until after this code is executed? 3.) Any clue why this is hanging here? I should have a remote debugger set up soon, but I have -nokdb set and I don't get any kernel panics or reg dumps. The system just hangs. Thanks in advance!! Paul Phillips

1 0

Re: TCB size in ix86 L4
by Torsten Paul 08 Jan '99

08 Jan '99

Michael Hohmuth <hohmuth(a)innocent.com> writes: > can't meet real-time requirements; the task_delete operation seems to > be especially bad, sometimes disabling interrupts for more than one > second. > Yes, the measurements showed interrupt a latency of sometimes 1.5 seconds (even if the interrupt thread had the highest priority) when in the L4-Linux System had to serve a lot of calls to fork/exec. This was tested with the lat_proc benchmark from the lmbench-suite that tries to measure the latency for creating simple processes (that only call exit) Still there is only one explanation for this: disabled interrupts while flushing the complete address space of a task when doing task_delete. ciao, Torsten. -- Torsten Paul Es ist leichter, einen Atomkern zu paul(a)os.inf.tu-dresden.de // spalten als ein Vorurteil. __________________________ooO_(+ +)_Ooo________Albert Einstein_(1879-1955) U

2 1