l4-hackers

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

2827 discussions

Re: gcc VS g++
by Michael Hohmuth 27 Jan '99

27 Jan '99

"Paul Phillips" <pphillips(a)ivue.com> writes: > FYI gcc is version 2.7.2.3 > g++ is egcs-2.90.29 980515 (egcs-1.0.3 release) > The GNU manual simple says that gcc uses 'C' linkage and g++ calls gcc and > uses C++ linkage. I don't thoroughly understand what might be meant with ``linkage'' here, but I don't think that linkage is a problem unless you combine modules compiled with different compilers. I have successfully used Fiasco kernels compiled with the `gcc' driver from Gcc 2.7.2.3, so certainly its ``linkage'' is alright. One possible other reason for this problem is that you might be trying to compile Fiasco without optimization ("-O" flag). For Gcc 2.7.2.3 I know for sure that this doesn't work: Optimization must be enabled to allow inline-assembler code to work; if it's not, the compiler skips the register-conflict-detection pass which makes sure arguments for inline-assembler code are put into the correct CPU registers. Maybe this has been corrected in Egcs (I haven't tried). > I won't try to solve this problem right now as g++ seems to work OK. Your > makefile uses gcc however, and I am wondering which I should be using. The Makefile shipped with the distribution should be using `egcc', a driver which belongs to Egcs, so it should be using Egcs 1.0 or 1.1. I recommend using a version of Egcs if you have access to it; Egcs 1.1 is best. > One of your correspondence mentions a version of egcs, and I'm not > sure of the relationships between gcc and egcs. Gcc and Egcs are two different compiler distributions which both include a C compiler and a C++ compiler. Normally, only one of the distributions is installed on one system; the front ends (driver programs) `gcc' and `g++' access the same compiler distribution, and one can compile both C and C++ source files with both drivers. However, newer Linux distributions confusingly come with both Gcc and Egcs, and both distributions contain both the C compiler and the C++ compiler. The driver program `gcc' is made to access Gcc, while `g++' and `egcc' use Egcs. As I said, all of these drivers can compile C code and C++ code, but they access different compiler distributions. The reason why both Egcs and Gcc are installed is that the Linux-2.0 kernel cannot be compiled with compilers newer than Gcc 2.7.2.x for political reasons (Linus Torvalds and the Egcs team like to blame and yell at each other for that). However, supplying only this old version of Gcc is unsatisfying because the newer C++ compilers offer more features and are more standards-compliant. Does this clear up some confusion? Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Re: multiple servers
by Michael Hohmuth 27 Jan '99

27 Jan '99

"Paul Phillips" <pphillips(a)ivue.com> writes: > I believe I have found a couple of bugs in Fiasco which were causing the > multiple server problem, and probably some other problems. In > /fiasco/src/time.h : > the code for set( microseconds): >From the first impression, your fix looks right. Thanks for tracking this down! I still wonder why I haven't ever seen this problem... I think I tried booting L4Linux and `hello' concurrently, but maybe the order in which they're started makes a difference. I almost can't believe that I wrote that routine... :-) It must have been late at night. I probably didn't invest much thought into it because I was thinking I was going to rewrite it soon. The timeout handling also suffers from at least two other problems: It disables interrupts while traversing a linear linked list of unknown length, that is, potentially for an unacceptably long time; and it suffers from the ``page fault may re-enable interrupts'' problem because it potentially touches a TCB in a 4-MB kernel region for which a reference to its shared page table hasn't yet been inserted into the current task's page directory (this is done ``on demand'' in the page fault handler). This is just one more example why doing kernel synchronization using interrupt-disabling is a bad idea. [ from a different email ] > >Why did you relocate it to 0x2400000? It should be able to run (even > >concurrently with L4Linux) at the memory location it is linked to > >using the distributed version of l4/server/hello/Makefile, 0x200000. > I thought perhaps GRUB was trying to load both 'hello' and L4Linux at the > same physical address and that was the reasoning behind my relocating it. It > makes no difference. GRUB is never trying to load boot modules to conflicting memory addresses; the modules (in this case, ELF binaries) are not interpreted at all---GRUB just locates them linearly behind what has been loaded as the ``kernel,'' which in our case is Rmgr, the resource and boot manager. Once it has been started by GRUB, Rmgr interprets the ELF binaries and loads them to their destination addresses. Rmgr makes sure these addresses don't conflict---otherwise you will see an ugly (and not very informative, I'm afraid) assertion failure. Thanks again! Michael [ code included for reference so that the bug tracking system sees it, too ] > inline void > timeout_t::set(unsigned long long abs_microsec) > { > // XXX uses cli/sti > > _wakeup = abs_microsec; > _flags.set = 1; > > unsigned flags = get_eflags(); > cli(); > > if (!timer::first_timeout) > { > timer::first_timeout = this; > _prev = _next = 0; > } > else > { > timeout_t *ti = timer::first_timeout; > > for (;;) > { > if (ti->_wakeup >= _wakeup) > { > // insert before ti > _next = ti; > _prev = ti->_prev; > ti->_prev = this; > if (_prev) _prev->_next = this; > /***** here you need to set timer::first_timeout since you just replaced it > */ > else timer::first_timeout = this; > /**********/ > goto done; > } > > if (ti->_next) > { > ti = ti->_next; > continue; > } > /**** here you need to put a break in case ti->wakeup < timeout, > and and ti->_next==0 . > Otherwise you loop here with interrupts disabled forever */ > else break; > /********************/ > } > > // insert as last item after ti > ti->_next = this; > _prev = ti; > _next = 0; > } > > done: > set_eflags(flags); > } -- hohmuth(a)innocent.com, hohmuth(a)sax.de http://www.sax.de/~hohmuth/

1 0

Re: multiple servers
by Paul Phillips 27 Jan '99

27 Jan '99

Michael, Thanks again for responding. I believe I have found a couple of bugs in Fiasco which were causing the multiple server problem, and probably some other problems. In /fiasco/src/time.h : the code for set( microseconds): inline void timeout_t::set(unsigned long long abs_microsec) { // XXX uses cli/sti _wakeup = abs_microsec; _flags.set = 1; unsigned flags = get_eflags(); cli(); if (!timer::first_timeout) { timer::first_timeout = this; _prev = _next = 0; } else { timeout_t *ti = timer::first_timeout; for (;;) { if (ti->_wakeup >= _wakeup) { // insert before ti _next = ti; _prev = ti->_prev; ti->_prev = this; if (_prev) _prev->_next = this; /***** here you need to set timer::first_timeout since you just replaced it */ else timer::first_timeout = this; /**********/ goto done; } if (ti->_next) { ti = ti->_next; continue; } /**** here you need to put a break in case ti->wakeup < timeout, and and ti->_next==0 . Otherwise you loop here with interrupts disabled forever */ else break; /********************/ } // insert as last item after ti ti->_next = this; _prev = ti; _next = 0; } done: set_eflags(flags); } Thanks, Paul Phillips -----Original Message----- From: Michael Hohmuth <hohmuth(a)innocent.com> 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: Tuesday, January 26, 1999 11:21 AM Subject: Re: multiple servers >Paul, > >thanks for trying Fiasco and L4Linux! > >"Paul Phillips" <pphillips(a)ivue.com> writes: > >> 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? > >No, I don't think so. > >I take it that without `hello', L4Linux works correctly. My guess >would be that hello has trouble running at the memory location you >linked it at. Did you see `hello's output ("hello: My thread-id >is...") at least once? > >Did you get `hello' to work without L4Linux? > >Why did you relocate it to 0x2400000? It should be able to run (even >concurrently with L4Linux) at the memory location it is linked to >using the distributed version of l4/server/hello/Makefile, 0x200000. > >> 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 for the hint! > >Michael >-- >hohmuth(a)innocent.com, hohmuth(a)sax.de >http://www.sax.de/~hohmuth/

1 0

Re: gcc VS g++
by Paul Phillips 26 Jan '99

26 Jan '99

Thanks Michael, FYI gcc is version 2.7.2.3 g++ is egcs-2.90.29 980515 (egcs-1.0.3 release) The GNU manual simple says that gcc uses 'C' linkage and g++ calls gcc and uses C++ linkage. I won't try to solve this problem right now as g++ seems to work OK. Your makefile uses gcc however, and I am wondering which I should be using. One of your correspondence mentions a version of egcs, and I'm not sure of the relationships between gcc and egcs. Thanks, Paul Phillips -----Original Message----- From: Michael Hohmuth <hohmuth(a)innocent.com> 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: Tuesday, January 26, 1999 11:22 AM Subject: Re: gcc VS g++ >"Paul Phillips" <pphillips(a)ivue.com> writes: > >> When I compile Fiasco with CC=gcc I get a message : >> >> KERNEL: no page fault handler for region: 0xffff0000, error 0x0. > >> However, when I compile with CC=g++ I get no such message. > >> Any ideas why this is happening? > >Obviously, the `gcc' driver uses a different compiler than the `g++' >driver ("gcc -v" and "g++ -v" should tell you that.) However, without >further context (like a backtrace) it's hard to tell what exactly went >wrong. > >Michael

1 0

Re: multiple servers
by Paul Phillips 26 Jan '99

26 Jan '99

Michael, Thanks for responding, -----Original Message----- From: Michael Hohmuth <hohmuth(a)innocent.com> 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: Tuesday, January 26, 1999 11:21 AM Subject: Re: multiple servers >Paul, > >thanks for trying Fiasco and L4Linux! > >"Paul Phillips" <pphillips(a)ivue.com> writes: > >> 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? > >No, I don't think so. > >I take it that without `hello', L4Linux works correctly. My guess >would be that hello has trouble running at the memory location you >linked it at. Did you see `hello's output ("hello: My thread-id >is...") at least once? > Yes, both 'hello' and L4Linux run correctly (some bugs in L4, but that's OK for now) when running as a single server. I do see the 'hello' output one time only: "hello: My id is 80000" , in the middle of the Linux bootup messages. I tried 'hello' at the default 0x200000 as well as 0x2400000. Both perform the same. >Did you get `hello' to work without L4Linux? > Yes, no problem! >Why did you relocate it to 0x2400000? It should be able to run (even >concurrently with L4Linux) at the memory location it is linked to >using the distributed version of l4/server/hello/Makefile, 0x200000. > I thought perhaps GRUB was trying to load both 'hello' and L4Linux at the same physical address and that was the reasoning behind my relocating it. It makes no difference. I have GDB running, but because the system hangs I can't generate a backtrace. Any ideas? Thanks Paul Phillips

1 0

Re: gcc VS g++
by Michael Hohmuth 26 Jan '99

26 Jan '99

"Paul Phillips" <pphillips(a)ivue.com> writes: > When I compile Fiasco with CC=gcc I get a message : > > KERNEL: no page fault handler for region: 0xffff0000, error 0x0. > However, when I compile with CC=g++ I get no such message. > Any ideas why this is happening? Obviously, the `gcc' driver uses a different compiler than the `g++' driver ("gcc -v" and "g++ -v" should tell you that.) However, without further context (like a backtrace) it's hard to tell what exactly went wrong. Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Re: multiple servers
by Michael Hohmuth 26 Jan '99

26 Jan '99

Paul, thanks for trying Fiasco and L4Linux! "Paul Phillips" <pphillips(a)ivue.com> writes: > 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? No, I don't think so. I take it that without `hello', L4Linux works correctly. My guess would be that hello has trouble running at the memory location you linked it at. Did you see `hello's output ("hello: My thread-id is...") at least once? Did you get `hello' to work without L4Linux? Why did you relocate it to 0x2400000? It should be able to run (even concurrently with L4Linux) at the memory location it is linked to using the distributed version of l4/server/hello/Makefile, 0x200000. > 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 for the hint! Michael -- hohmuth(a)innocent.com, hohmuth(a)sax.de http://www.sax.de/~hohmuth/

1 0

gcc VS g++
by Paul Phillips 26 Jan '99

26 Jan '99

When I compile Fiasco with CC=gcc I get a message : KERNEL: no page fault handler for region: 0xffff0000, error 0x0. However, when I compile with CC=g++ I get no such message. Any ideas why this is happening?

1 0

Alrighty then...
by steven_weiss＠hotmail.com 21 Jan '99

21 Jan '99

--Looking to Go into business?-- Enjoy the Benefits of Working from Home! Build a Financial Future For You and Your Family! Learn How to Make Money with The Internet! --- The Money is NOT in the Technical End, it is IN the Marketing! Complete Training and Support! No Selling! Not an M.L.M or Network Marketing ! ------------------------------------------------------------------ --- Income Potential = $4,500 -$10,000 a month.* *(Based on Appox. 15-20 hrs every wk) ------------------------------------------------------------------ --- Requirements: 10-20 hrs a wk. 18 yrs or Older. Have or Apply for an In Home Business License. Be Ready to Learn--(this is NOT rocket science). (Click Reply) with Name, Address, Phone # and the Best Time to Reach You! Please give brief description of why you are looking for a business, How long you have been looking, What type of Business that you have run before (if any). How much computer experience do you have?

1 0

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

Jump to page:

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