user-mode-linux-user

[User-mode-linux-user] user-mode port 0.29-2.4.0-test6

[Jeff D. <jd...@ka...>]

The user-mode port of 2.4.0-test6 is really available this time.

I fixed a stupid, long-standing bug and it started working a lot better.

I'll recap what went into the last two releases since this is the first 
release in which they are really usable:

Jim Leu's virtual ethernet driver provides a low-level network interface for 
uml.  It's easier to use, more general, and more flexible than the existing 
slip-based umn interface, which I am planning on eliminating at some point.  
Setting up the umn device is fraught with permission problems from allocating 
the pty, to setting the line discipline on it, and configuring the slip 
device.  If all you want is an isolated virtual network, the eth device 
requires no privileges; an external daemon handles packet routing.  If you 
want to put the host on the virtual net or put the virtual net on a physical 
ethernet, all you need is to be able to allocate and configure an ethertap 
device.  See http://user-mode-linux.sourceforge.net/networking.html for more 
details.

Also, there's Lars Brinkhoff's ptrace proxy, which allows gdb to debug kernel 
threads without attaching to them, which they can't because they are already 
being syscall traced.  You can invoke the debugger from kernel startup by 
putting 'debug' on the command line.  You will get an xterm with gdb in it 
stopped at the top of start_kernel().  You can start gdb afterwards by sending 
the tracing thread a SIGUSR1.  You attach to the kernel and do whatever gdb 
things you want.

This doesn't totally work yet.  Breakpoints, examining data, and looking at 
source work fine.  Stepping through code works most of the time, except 
sometimes, it acts like you said 'cont' instead of 'next' or 'step'.  Putting 
conditions or commands on breakpoints doesn't work at all.  Doing that will 
cause segfaults.  More information is available at 
http://user-mode-linux.sourceforge.net/debugging.html

The project's home page is http://user-mode-linux.sourceforge.net

The project's download page is http://sourceforge.net/project/filelist.php?grou
p_id=429

				Jeff

[User-mode-linux-user] [UNSTABLE] user-mode port 0.29-2.4.0-test6

[Jeff D. <jd...@ka...>]

The user-mode port of 2.4.0-test6 is available.

It's been updated to 2.4.0-test6.

The major change in this release is the addition of a clever piece of code 
from Lars Brinkhoff which allows gdb to debug kernel threads without having to 
attach to them.  It is possible to attach gdb to the kernel once it's up, and 
also to start the kernel under the control of gdb by putting 'debug' on the 
command line.

This release is also very unstable.  I've updated only the patch so that 
people can play with it and hopefully find more informative crashes than I've 
been able to get so far.  The debugging interface will come in handy here.  
Put 'debug' on the command line, tell gdb to set it running, and when you get 
a panic, ^C it and get a backtrace and do all the standard post-mortem stuff.  
You can also set breakpoints and step through the code, but adding conditions 
and commands to breakpoints causes crashes, and 'next' and 'step' sometimes 
act as 'cont'.

The project's home page is http://user-mode-linux.sourceforge.net and has been 
completely revamped with lots of new information.
The project's download page is http://sourceforge.net/project/filelist.php?grou
p_id=429

				Jeff

[User-mode-linux-user] Re: [UNSTABLE] user-mode port 0.29-2.4.0-test6

[Pavel M. <pa...@su...>]

Hi!

> The user-mode port of 2.4.0-test6 is available.
> 
> It's been updated to 2.4.0-test6.
> 
> The major change in this release is the addition of a clever piece of code 
> from Lars Brinkhoff which allows gdb to debug kernel threads without having to 
> attach to them.  It is possible to attach gdb to the kernel once it's up, and 
> also to start the kernel under the control of gdb by putting 'debug' on the 
> command line.

This seems to get pretty usefull. Are there any plans to merge to
Linus, soon? After all, it is only "just another architecture".

								Pavel
-- 
I'm pa...@uc.... "In my country we have almost anarchy and I don't care."
Panos Katsaloulis describing me w.r.t. patents at di...@li...

[User-mode-linux-user] Re: [UNSTABLE] user-mode port 0.29-2.4.0-test6

[David W. <dw...@re...>]

jd...@ka... said:
> The user-mode port of 2.4.0-test6 is available.
> It's been updated to 2.4.0-test6. 

Is it now capable of bringing up the umn network device without having to 
run as root?


The 2.4.0-test4 version has the required suid um_ifconfig helper, but if it 
doesn't run as root, then it still doesn't work:

Bringing up interface umn 
	SIOCSIFADDR: No such device
	sl-1: unknown interface: No such device
	SIOCSIFDSTADDR: No such device
	sl-1: unknown interface: No such device
	sl-1: unknown interface: No such device
                                                             [  OK  ]

--
dwmw2

Re: [User-mode-linux-user] [UNSTABLE] user-mode port 0.29-2.4.0-test6

[William S. <wst...@po...>]

Good day, Jeff,

On Thu, 10 Aug 2000, Jeff Dike wrote:

> The user-mode port of 2.4.0-test6 is available.

	I had to make a quick patch to get it to compile; do I have libc.a
in a non-standard location?

diff -bud linux-2.4.0-test6.uml/arch/um/kernel/Makefile.orig linux-2.4.0-test6.uml/arch/um/kernel/Makefile
--- linux-2.4.0-test6.uml/arch/um/kernel/Makefile.orig  Mon Aug 14 19:34:49 2000
+++ linux-2.4.0-test6.uml/arch/um/kernel/Makefile       Mon Aug 14 20:09:03 2000
@@ -41,7 +41,7 @@
        gcc -D__KERNEL__ $(USER_CFLAGS) $(EXTRA_CFLAGS) -c -o $@ $<

unmap_fin.o : unmap.o
-       ld -r -o $@ $< -lc
+       ld -r -L/usr/lib -o $@ $< -lc

 $(LIB): $(OBJS) $(OX_OBJS)
        rm -f $@

	Once applied, I get
...
Initializing random number generator... done.
Recovering nvi editor sessions... done.
INIT: Entering runlevel: 2
Initializing random number generator... done.
Recovering nvi editor sessions... done.
INIT: Entering runlevel: 2
Kernel panic: Kernel mode fault at addr 0x53409f94, ip 0x10034f7d
Hangup

	Adding debug brings up the gdb window, but I'm not good enough
with gdb to figure out what to do.  Sorry.  I do know enough about the
process to know that the addr/ip pair isn't enough to find the problem
without System.map.  I'd be happy to provide that if it's useful, but I
suspect you've got enough crashes of your own.
	Cheers,
	- Bill

---------------------------------------------------------------------------
	"Assuming that piece of ear weighed about 1/2 an ounce, Bill
[Gates] could afford to eat 443.39 pounds of Evander Holyfield if he were
so inclined." 
	-=- http://web.quuxuum.org/~evan/bgnw.html
(Courtesy of James Mastros <ro...@je...>)
--------------------------------------------------------------------------
William Stearns (wst...@po...).  Mason, Buildkernel, named2hosts, 
and ipfwadm2ipchains are at:                http://www.pobox.com/~wstearns
LinuxMonth; articles for Linux Enthusiasts! http://www.linuxmonth.com
--------------------------------------------------------------------------