qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] QEMU 0.5.2 testing


From: Herbert Poetzl
Subject: Re: [Qemu-devel] QEMU 0.5.2 testing
Date: Tue, 20 Jan 2004 08:07:41 +0100
User-agent: Mutt/1.4.1i

On Mon, Jan 19, 2004 at 11:49:09AM +0100, Fabrice Bellard wrote:
> Herbert Poetzl wrote:
> >On Mon, Jan 19, 2004 at 12:09:11AM +0100, Fabrice Bellard wrote:
> >
> >>Hi,
> >>
> >>I commited all the current fixes in the CVS. I will make the 0.5.2 
> >>release tomorrow, so it would be interesting if some people could test 
> >>it before it is officially released...
> >
> >texi2html -monolithic -number qemu-doc.texi
> >make: texi2html: Command not found
> >make: *** [qemu-doc.html] Error 127
> 
> In the distribution I add the html et man files, so it should not be a 
> problem to most users (I am relunctant to add texi2html in the 
> distribution).

good

> >usual set of compiler warnings, are you interested
> >in that output too?
> 
> I know they are warnings in translate.c about a function prototype. If 
> you have others I am interested.

will check ...

> >seems to work as expected, with the folowing options
> >
> >qemu-cvs-19.01.2004/i386/qemu-fast -s -nographic -L qemu-0.5.1/pc-bios -m 
> >128 -snapshot -hda IMGs/TEST_32M.img -hdb IMGs/TEST_256M.img -kernel 
> ><kernelpath>/arch/i386/boot/bzImage -append "rw root=/dev/hda1
> >
> >tested with/out -s for 2.4.24, 2.4.25-pre6 and 2.6.0 ...
> >two harddisk images 32M and 256MB in size ...
> 
> So you don't use the '-kernel' option ? I am thinking about modifying 
> this option so that the PC bios is used in all cases (it would allow to 
> have the VGA output even with '-kernel'). A possible way would be to 
> integrate a custom linux loader ROM, as it is done now in plex86.

huh?  -kernel <kernelpath>...

please, I'm _verry_ happy that I do not need the vga 
and interface stuff, and that it works on the console
I'm planning to use it for automated kernel testing,
so _please_ don't do that ...

  -nographic -L qemu-0.5.1/pc-bios -kernel ... -append ...

by the way, I tried to pass commands to qemu in the
following way:

 echo "abcdefghij" | qemu ....

and it seems to work 'somewhat' but there are some 
'weird' deficiencies ...

 - the first two characters are always swallowed
 - depending on the kernel a number of 'padding'
   chars is required, they are displayed but do
   not reach the init process (or might be discarded
   by the busybox I use)

for example:
echo -e "  aaaaaaaaaaaaaaaaaaaaaa\n/tmp/testme.sh\n" | qemu ...

will work for 2.4.25-pre6 ;) and execute /tmp/testme.sh

> >hmm didn't notice that one until now (seems like a bug) ...
> >
> >Program received signal 0, Signal 0.
> >default_idle () at process.c:92
> >92   }
> >(gdb) break vfs_readdir
> >Breakpoint 1 at 0x90152640: file readdir.c, line 17.
> >(gdb) c
> >Continuing.
> >
> ># ls /proc
> ># ls /tmp
> ># <nothing happens> CTRL-C
> >
> >Program received signal 0, Signal 0.
> >default_idle () at process.c:92
> >92   }
> >(gdb) break sys_vserver
> >Breakpoint 2 at 0x901261f0: file vswitch.c, line 39.
> >(gdb) c
> >Continuing.
> >
> ># ls /tmp
> >
> >Program received signal SIGTRAP, Trace/breakpoint trap.
> >vfs_readdir (file=0xc183c189, filler=0x2e9c103, buf=0xde89a5f3) at 
> >readdir.c:17
> >17   {
> >
> >(gdb) disable 1
> >(gdb) c
> >Continuing.
> >
> ># sys_vserver (call)
> >
> >Program received signal SIGTRAP, Trace/breakpoint trap.
> >sys_vserver (cmd=0, id=0, data=0x0) at vswitch.c:39
> >39   {
> 
> Do you have the problem if you don't use gdb ? Maybe my last TLB 
> optimisations broke the gdb stub logic. I will check...

hum, how would I test without gdb? color me confused.

TIA,
Herbert

> Fabrice.
> 
> 
> 
> _______________________________________________
> Qemu-devel mailing list
> address@hidden
> http://mail.nongnu.org/mailman/listinfo/qemu-devel




reply via email to

[Prev in Thread] Current Thread [Next in Thread]