- l4-hackers - example.com Site

Re: l4/log/l4log.h
by Michael Hohmuth 20 May '00

20 May '00

Juho Östman <jostman(a)myrealbox.com> writes: > It seems that the file 'l4log.h' is missing in the CVS distribution. > Where is it found? I have added the "log" package to the CVS distribution. (We recently introduced a dependency on the "log" package in the code we distribute; I think there even are some cross dependencies between packages... the easy fix for now just seems to be distributing the "log" package as well. :-) Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

l4/log/l4log.h
by Juho Östman 15 May '00

15 May '00

It seems that the file 'l4log.h' is missing in the CVS distribution. Where is it found?

1 0

Re: Plaese .. Could you help me ?
by Michael Hohmuth 14 Apr '00

14 Apr '00

Ricky, Jean just went into his well-deserved weekend, but he told me to tell you that he has tried your boot disk, and has been able to reproduce your problem. He got it to run by replacing your Fiasco binary with a newer version of Fiasco. So you are probably using an outdated version of the kernel. You should download a newer version from our remote-CVS server (see <URL:http://os.inf.tu-dresden.de/drops/download.html for instructions). Michael Jean Wolter <jean.wolter(a)inf.tu-dresden.de> writes: > "Ricky(Soohyung) Lee" <onlyjazz(a)cselab.snu.ac.kr> writes: > > > But, I have another problem. > > When I boot from the grub floppy disk with glinux.gz on it, > > after the message that is saying the first task is created, > > the kernel fall into the endless loop with printing out " root: rcvd > > b9002, 4bc10c, src :5.3 ". > > Would it be possible to gzip the boot disk and send it to us? Then we > could try to boot it and see, what fails. > > I´m not quite sure that I understand the scenario good enough to come > up with an explanation and a fix. Somehow the service thread (5.3) > touches a page in the adapter area (a0000-100000) and doesn´t get or > doesn´t have a mapping for it. > > Jean -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Re: Plaese .. Could you help me ?
by Jean Wolter 14 Apr '00

14 Apr '00

"Ricky(Soohyung) Lee" <onlyjazz(a)cselab.snu.ac.kr> writes: > But, I have another problem. > When I boot from the grub floppy disk with glinux.gz on it, > after the message that is saying the first task is created, > the kernel fall into the endless loop with printing out " root: rcvd > b9002, 4bc10c, src :5.3 ". Would it be possible to gzip the boot disk and send it to us? Then we could try to boot it and see, what fails. I´m not quite sure that I understand the scenario good enough to come up with an explanation and a fix. Somehow the service thread (5.3) touches a page in the adapter area (a0000-100000) and doesn´t get or doesn´t have a mapping for it. 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

Plaese .. Could you help me ?
by Ricky(Soohyung) Lee 14 Apr '00

14 Apr '00

Hi, everyone ! First of all, let me introduce myself . My name is Ricky Lee . I'm in the course of master's degree at Seoul National University in Korea. Thesedays I'm reading L4-related papers. And, I'm trying to make experience on L4Linux on Fiasco. But, It is not so easy for me. Today, I finally succeeded in building linux22. I really appreciate Michael Hohmuth' advices. But, I have another problem. When I boot from the grub floppy disk with glinux.gz on it, after the message that is saying the first task is created, the kernel fall into the endless loop with printing out " root: rcvd b9002, 4bc10c, src :5.3 ". I guess that it is printed out from root_pager() in l4_paper.c Can you tell me what's wrong ? And, then, what should I do? I really need to solve this problem. So, I really need your help. Please, Help me .. Thanks for your kindness . Thanks . P.S. I compiled Fiasco and Linux22 server on the Redhat package 5.2. Its kernel version is 2.0.36. Any software requirement that is mentioned in the L4linux homepage was met, I think . Can there be the reason of problem with the linux environment in which I compiled those sources ? Can anyone tell me the kernel version and H/W spec. on which you've built the L4Linux22 Server ? It'll be very helpful to me if you tell me them .. Please !!! Thank you very much !! - Ricky Lee -

1 0

IPC & invalid Thread-ID
by Christian Stueble 04 Mar '00

04 Mar '00

-----BEGIN PGP SIGNED MESSAGE----- Hello, what happens if IPC uses an invalid thread-id? Is the message although forwarded to the subsystem's chief? Ciao, Chris -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Christian Stueble............stueble(a)ls6.cs.uni-dortmund.de PubKey[BF7104F5].......fp=8678C5D3CAD9CD8C F1DDB8EC202F116A To be or not to be is true... (apocrypha of George Boole) -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 5.0i for non-commercial use MessageID: J7gRC3CpAZgh3vXKJpAlEFSfZz45w/ET iQCVAwUBOL5y5fyJjJy/cQT1AQFaywQAhBB4KcN8eHZXMrAPgrvJyQhbEMoRYc6j Wmpt5tBK3Tu1DmcNtTqlcGChzFIDrUiE6+Iw8eN74Bx2Dn+Cn+ohcUmSeaxEjkNl 2F8rTA/13dm/4wojmRz87uAvHWuFx+PycsXm7P3BRRqb+9aC+O1ftau6bNSqsmXC 8A/QvdUxDQg= =PFys -----END PGP SIGNATURE-----

2 3

autoconf
by Juho Östman 28 Feb '00

28 Feb '00

What do you think about adapting L4 (and Fiasco) to use autoconf? The most important immediate benefits are the possibility of using a separate build directory and configuration options. And it would be more compatible with GNU build mechanisms. If you think that is a good idea I could be voluntary to make necessary changes.

1 0

What happened to AIM?
by Gernot Heiser 21 Feb '00

21 Feb '00

Does anyone know what happened to AIM Technology? Their website, www.aim.com, seems to have vanished from the face of the earth. Gernot -- Gernot Heiser ,--_|\ School of Computer Sci. & Engin. Phone: +61 2 9385 5156 / \ The University of NSW Fax: +61 2 9385 5533 \_,--._* Sydney, Australia 2052 E-mail: gernot(a)unsw.edu.au v http://www.cse.unsw.edu.au/~gernot PGP fingerprint: 94 1E B8 28 25 FD 7C 94 20 10 92 E5 0B FF 39 8F

1 0

A few things I'd like to ask
by Juho �stman 04 Feb '00

04 Feb '00

I have become quite familar with L4/x86 Reference manual and L4 User manual (for MIPS) but I think they don't give enough information about designing and writing programs on to the L4 architecture. In my opinion they are a bit unclear and incomplete. Therefore I'd like to ask some questions on that subject. What kind of address space do the initial servers get? I know that sigma0 has the complete physical address space idempotently mapped but what about others? What pages do they have initially mapped? I have understood that the kernel isn't able to free memory it has allocated. Is that true? How is the destination address of an intercepted IPC delivered to the chief? What is sigma1? What should it do? Is sigma0 able to map memory above 0x40000000? How? What happens when the version numbers of a task are used up? Could they be reused? Could tasks donate interrupts like they can donate task numbers? Thanks for everyone who bothers to answer.

4 4

Re: Protected mode I/O driver architecture (Fiasco)
by Michael Hohmuth 28 Jan '00

28 Jan '00

John Nagle <nagle(a)animats.com> writes: > Once I/O flexpages are in, RMGR will be able to issue access to > device space byte by byte. That's right. > It can connect up interrupts to drivers now. Small correction: L4 currently has no interface for safely passing an interrupt assignment. All Rmgr can do when a driver asks for an interrupt is release its attachment so that the driver can attach. This opens a time window for a malicious process racing with the driver to steal the attachment. > The remaining big security problem is DMA. Of course, DMA will be restricted to processes having access to the ports controlling the DMA controller. > In most systems, device drivers are given total access to DMA > controllers. Could that be avoided, with a trusted process that > handles the DMA controllers at the request of the device drivers? > > I'm going to concentrate on PCI bus issues, since ISA is > on the way out and SCSI, USB, and IEEE-1394 devices usually work > through a PCI controller. > > For most devices, DMA is performed by DMA controllers in the > core chipset (such as the Intel 82810), not by the peripherals > themselves. It's seems to be possible to disable the ability of a > PCI device to become a bus master without disabling the ability > of the DMA controllers to execute memory and PCI bus cycles, > at least with the more modern core chipsets. > > So it ought to be possible to have a trusted process that > handles DMA and deals with the memory mapping issues. This > process would be used by device drivers that need DMA. Drivers > would not have access to the DMA controllers directly and would > be untrusted. Yes, all that is correct. A different solution would be to simply trust device drivers also. > There's the related issue of hardware discovery, or > "plug and play". Something is needed at startup to discover > all the hardware, build the hardware tree, and parcel out > hardware resources to drivers. What's planned in that area? We plan to do all this in a device-managing component. As a start, Uwe Dannowski already has added some enhancements to Rmgr that discovers PCI devices and remaps their memory-mapped I/O ports so that only one device has ports in a given high-memory superpage, enabling access control to memory-mapped I/O ports for device drivers. However, I think his changes are currently not available publicly. :( Other than that, not much has been done so far in this area. All drivers currently contain their own PCI-device-discovery code. > Is this consistent with the thinking of the L4/Fiasco/Drops > groups? If not, is there something wrong here? I think you ate pretty much inline with our thinking. Regards, Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Satellite Owners! Must See!
by VWUrkI01t＠please.org 27 Jan '00

27 Jan '00

THIS IS A ONE TIME MAILING You were sent this email because it was believed it would benefit you in some way If you do not want this email - simply delete it - TEST YOUR SATELLITE CHANNELS - DO THEY WORK? TEST THEM AND SEE Your satellite receiver - if you have Direct Tv - came with a little credit card sized programming card known as a smart card. We can, for testing purposes only, program your card to receive any and all channels offered by your satellite company without you having to pay an additional monthly fee. You will receive blacked out sporting events, pay per views, premium channels and unlimited, unrestricted access to ANY channel offered by your satellite company after we program your card. We can only program H cards. These are the cards that say "DSS ACCESS CARD" on them and they are blue. We can NOT program Dish Network cards. This service is provided for testing purposes only. To receive an order form and for more informaton - Please Call 1-770-621-5852 (24hours a day, 7 days a week).

1 0

Re: CVS server
by Michael Hohmuth 26 Jan '00

26 Jan '00

David Welch <david.welch(a)st-edmund-hall.oxford.ac.uk> writes: > The cvs server at os.inf.tu-dresden.de seems to be down, is this a > temporary problem, or has cvs access been moved or removed? Temporary problem -- sorry for the downtime. It's back up now. Michael -- hohmuth(a)innocent.com, hohmuth(a)inf.tu-dresden.de http://home.pages.de/~hohmuth/

1 0

Protected mode I/O driver architecture (Fiasco)
by John Nagle 26 Jan '00

26 Jan '00

A key concept of L4 is that device drivers should be outside the kernel, and not have privileges that allow them to crash the system. This requires support outside the kernel that isn't yet present. I'm writing this to ask some questions about how that will be architected. Once I/O flexpages are in, RMGR will be able to issue access to device space byte by byte. It can connect up interrupts to drivers now. The remaining big security problem is DMA. In most systems, device drivers are given total access to DMA controllers. Could that be avoided, with a trusted process that handles the DMA controllers at the request of the device drivers? I'm going to concentrate on PCI bus issues, since ISA is on the way out and SCSI, USB, and IEEE-1394 devices usually work through a PCI controller. For most devices, DMA is performed by DMA controllers in the core chipset (such as the Intel 82810), not by the peripherals themselves. It's seems to be possible to disable the ability of a PCI device to become a bus master without disabling the ability of the DMA controllers to execute memory and PCI bus cycles, at least with the more modern core chipsets. So it ought to be possible to have a trusted process that handles DMA and deals with the memory mapping issues. This process would be used by device drivers that need DMA. Drivers would not have access to the DMA controllers directly and would be untrusted. There's the related issue of hardware discovery, or "plug and play". Something is needed at startup to discover all the hardware, build the hardware tree, and parcel out hardware resources to drivers. What's planned in that area? Is this consistent with the thinking of the L4/Fiasco/Drops groups? If not, is there something wrong here? John Nagle Animats www.animats.com

1 0

CVS server
by David Welch 20 Jan '00

20 Jan '00

Hi, The cvs server at os.inf.tu-dresden.de seems to be down, is this a temporary problem, or has cvs access been moved or removed? Thanks, David Welch

1 0

main blocks forever (revisited)
by Kermit the Frog 14 Jan '00

14 Jan '00

Hello, after aplying a patch sent by Jean (thanks Jean!) everything seemed to work just fine. Except for one thing. After restarting the stopped thread, with l4_thread_ex_regs, if main() would try to do a getchar() (for instance) the re-started thread would stop ... if I change the getchar() for a for(;;); everyting works just fine. And theres one more thing, please if someone can tell me if this behaviuor is correct or not, when stopping a thread w/l4_thread_ex_regs lets say we stop it by changing its eip to a func. always_release() { for(;;) l4_thread_switch(L4_NIL_ID); } If one were to leave the original esp, when re-starting the original thread, by changing its eip to the original func., all the values contained in the stack were to be clobbered ... now ... is this normal ? should l4_thread_switch() change the values in the stack ? (of course, a workaround to this is to declare a char release_stack[] and use it when stopping the thread as its new stack). Ciao. Cervi~no Ulises <kermit(a)tower.com.ar> <ulises(a)rosario.linux.org.ar> ............................................................................... "Contrary to popular opinion, Unix is user friendly, It just happens to be very selective about who it makes friends with."

1 0

Unknown error
by Christian Stueble 13 Jan '00

13 Jan '00

-----BEGIN PGP SIGNED MESSAGE----- Hello, I hope someone can tell me why the following source produces an error (trapno 14, error 5): ------------------------------------ class BaseClass { public: virtual void print() { printf("BaseClass::print()\n"); }; }; class DerivedClass : public BaseClass { public: virtual void print() { printf("DerivedClass::print()\n"); }; }; DerivedClass dc; void main(void) { BaseClass *bc = &dc; bc->print(); } ---------------------- If dc is defined inside main(), everything works fine. Chris -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Christian Stueble............stueble(a)ls6.cs.uni-dortmund.de PubKey[BF7104F5].......fp=8678C5D3CAD9CD8C F1DDB8EC202F116A To be or not to be is true... (apocrypha of George Boole) -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 5.0i for non-commercial use MessageID: Mqyve75goky7Nb6fefYC/Sh3qq9PuiJi Comment: iQCVAwUBOH3OA/yJjJy/cQT1AQEzOQQAut8lc3ns0EgZiQc80OgiKjlgnLv9g4Np bOdvfgxqspDkUqV6j/D/w0I1tJc3QAwLh8EsF4t6K27dd3yqJN2JVEnvud/7wNGh as2+1S1gqhV6k6dttkY5rfPy+epp/e8ZM5jmhFiS0H/HG+F6Vq+wTAm1f3DnWpoX a1mRhzHjZ+g= =fR1c -----END PGP SIGNATURE-----

2 1

Inside L4
by Gernot Heiser 11 Jan '00

11 Jan '00

A first version of a document entitled "Inside L4/MIPS" containing annotated L4 source code is available from http://www.cse.unsw.edu.au/~disy/L4/index.html. It is still incomplete, so far covering kernel data structures, exception handling and IPC. Any feedback/suggestions/corrections/additions are highly appreciated. Gernot -- Gernot Heiser ,--_|\ School of Computer Sci. & Engin. Phone: +61 2 9385 5156 / \ The University of NSW Fax: +61 2 9385 5533 \_,--._* Sydney, Australia 2052 E-mail: gernot(a)unsw.edu.au v http://www.cse.unsw.edu.au/~gernot PGP fingerprint: 94 1E B8 28 25 FD 7C 94 20 10 92 E5 0B FF 39 8F

1 0

main blocks forever
by Kermit the Frog 11 Jan '00

11 Jan '00

Hello, I have the following scenario were everything goes just wrong. I have only one task running, which creates a thread using l4_thread_ex_regs ... after a 4 secs. sleep using l4_ipc_receive it changes the eip value of the thread to a piece of code that reads: void always_release() { for(;;) l4_thread_switch(L4_NIL_ID); } After doing this the main thread (# 0) tries to wait 4 secs. using l4_ipc_receive again and then change the eip of the created thread to its old eip. The problem is that the main thread never wakes up. I mean, after the 4 secs. have passed ... the main thread is still blocked. Are we in prescence of a bug ? Cheers. Cervi~no Ulises <kermit(a)tower.com.ar> <ulises(a)rosario.linux.org.ar> ............................................................................... "Contrary to popular opinion, Unix is user friendly, It just happens to be very selective about who it makes friends with."

1 0

Re: Its me again :-)
by Jean Wolter 29 Dec '99

29 Dec '99

Kermit the Frog <kermit(a)tower.com.ar> writes: > > Hello, > lets see if I have this thing right. > > Lets say that main() creates a thread that will act as a scheduler. We'll > call it sched. Now main() creates antoher thread that will have sched as > its scheduler. sche() enters an infinite loop with a l4_i386_ipc_wait in > it. Of course, it waits for an preemption RPC message. Currently external and internal preempters are not implemented. They may be released in a future version but I'm not sure about 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

Its me again :-)
by Kermit the Frog 24 Dec '99

24 Dec '99

Hello, lets see if I have this thing right. Lets say that main() creates a thread that will act as a scheduler. We'll call it sched. Now main() creates antoher thread that will have sched as its scheduler. sche() enters an infinite loop with a l4_i386_ipc_wait in it. Of course, it waits for an preemption RPC message. Now, when creating the second thread (lets call it task) with l4_thread_ex_regs, I put sched to be its preempter, anyway, sched() never gets the IPC and task() runs as if it never had any preempter asociated. Theres another thing, if inside main we put l4_thread_ex_regs to change task()'s eip and esp to (dword_t) -1 (this would likely stop task()) nothing happens, task() just runs normally. Any ideas ?? Cheers. PS: I didnt send any source code because its kind of long and I dont want to bother people w/unsolicited mail ;-), if anyone wants it please email me and ill be happy to send it Cervi~no Ulises <kermit(a)tower.com.ar> <ulises(a)rosario.linux.org.ar> ............................................................................... "Contrary to popular opinion, Unix is user friendly, It just happens to be very selective about who it makes friends with."

1 0

requesting mappings from sigma0
by Kermit the Frog 22 Dec '99

22 Dec '99

Hello, Ive been having some problems while trying to request a page from sigma0. I have included the source code ... any help ? many thanks! ------[ source begins here ]------ #include <l4/types.h> #include <l4/syscalls.h> #include <l4/ipc.h> #include <stdio.h> #define SIGMA0_ID 4 void main(void) { l4_threadid_t sigma0; l4_msgdope_t msg, result; dword_t addr, type, foo, *r_add; addr = 0x80000; type = L4_IPC_SHORT_MSG; sigma0.id.task = SIGMA0_ID; sigma0.id.lthread = 0; r_add = 0x0; while(r_add == 0x0) { addr += 4096; /* pasamos a la siguiente pagina de 4K */ printf("Vamos a pedir la region de memoria 0x%x\n", addr); l4_i386_ipc_send(sigma0,(void *) type, addr, addr, L4_IPC_NEVER, &result); printf("ipc enviado ..."); l4_i386_ipc_receive(sigma0, L4_IPC_MAPMSG(addr, 4096), r_add, &foo, L4_IPC_NEVER, &result); printf("%p recivido\n", r_add); } *r_add = 'a'; r_add++; *r_add = 0x7; r_add --; if(*r_add == 'a') printf("BINGO!\n"); printf("Done!\n"); } Cervi~no Ulises <kermit(a)tower.com.ar> <ulises(a)rosario.linux.org.ar> ............................................................................... "Contrary to popular opinion, Unix is user friendly, It just happens to be very selective about who it makes friends with."

1 1

Fiasco & oskit10
by Christian Stueble 21 Dec '99

21 Dec '99

Is it possible to compile fiasco based on oskit10? I think there are some files missing (flux/*/*)?? When compiling l4/pkg/oskit_support/lib/src/ I get an error message: --- ... Making rmgr_stubs.o {standard input}: Assembler messages: {standard input}: Error: Symbol 'rmgr_pager_id' can not be both weak and common --- If I remove '__attribute__ ((weak))' it works. Is this a bug? I use egcs-2.91.66. -Chris- -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Christian Stueble............stueble(a)ls6.cs.uni-dortmund.de PubKey[BF7104F5].......fp=8678C5D3CAD9CD8C F1DDB8EC202F116A To be or not to be is true... (apocrypha of George Boole) -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

2 1

Using g++ to compile hello?
by Christian Stueble 21 Nov '99

21 Nov '99

Hi! Is it possible to compile the test program ŽhelloŽ using the g++ compiler. I get an error message 1) main.cpp:37: undefined reference to `__throw' ...if I link main.o and crt0.o with the ld linker, and 2) crt0.o(.text+0x0): multiple definition of `_start' /usr/lib/crt1.o(.text+0x0): first defined here ...if I try to link the files with g++ -o xxxx. Can somebody help me? Chris -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Christian Stueble............stueble(a)ls6.cs.uni-dortmund.de PubKey[BF7104F5].......fp=8678C5D3CAD9CD8C F1DDB8EC202F116A To be or not to be is true... (apocrypha of George Boole) -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

1 1

Apologies for mispost
by Gernot Heiser 18 Nov '99

18 Nov '99

My last two messages were not meant for l4-hackerßos.inf.tu-dresden.de, but for one of our internal lists. Apologies, Gernot-- Gernot Heiser ,--_|\ School of Computer Sci. & Engin. Phone: +61 2 9385 5156 / \ The University of NSW Fax: +61 2 9385 5995 \_,--._* Sydney, Australia 2052 E-mail: gernot(a)unsw.edu.au v http://www.cse.unsw.edu.au/~gernot PGP fingerprint: 94 1E B8 28 25 FD 7C 94 20 10 92 E5 0B FF 39 8F

1 0

Re: L4 Workshop papers
by Gernot Heiser 18 Nov '99

18 Nov '99

------- Blind-Carbon-Copy From: Gernot Heiser <gernot(a)cse.unsw.edu.au> To: jaegert(a)us.ibm.com cc: Liedtke(a)ira.uka.de, Haertig(a)os.inf.tu-dresden.de, hohmuth(a)innocent.com Subject: Re: L4 Workshop papers In-reply-to: Your message of "Thu, 18 Nov 1999 10:03:10 CDT." <8525682D.0052B697.00(a)D51MTA03.pok.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Reply-to: "Gernot Heiser" <gernot> Date: Fri, 19 Nov 1999 08:04:33 +1100 Sender: gernot(a)zuse.disy.cse.unsw.EDU.AU Thanks, Trent. I really have material for a number of talks, but don't see how I can find the time to write up anything before the 1st (and the students are studying for their exams). Topics: - - StrongARM kernel - - API extensions for SMP scheduling and cacheability - - SMP Alpha kernel - - supporting multiple page sizes - - portable VM implementation? Another problem is that I still haven't got a return flight. I may have to get a business class flight (which means paying the difference out of my own pocket) to be able to make it. Gernot - -- Gernot Heiser ,--_|\ School of Computer Sci. & Engin. Phone: +61 2 9385 5156 / \ The University of NSW Fax: +61 2 9385 5995 \_,--._* Sydney, Australia 2052 E-mail: gernot(a)unsw.edu.au v http://www.cse.unsw.edu.au/~gernot PGP fingerprint: 94 1E B8 28 25 FD 7C 94 20 10 92 E5 0B FF 39 8F ------- End of Blind-Carbon-Copy

1 0