l4-hackers January 2000

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

9 participants
10 discussions

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

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 January 2000