Re: [Qemu-devel] qemu-system-sh4 broken again.

[Rob Landley <rob@landley.net>]

On Saturday 13 March 2010 15:11:03 Rob Landley wrote:
> On Saturday 13 March 2010 04:26:47 Aurelien Jarno wrote:
> > On Sat, Mar 13, 2010 at 01:51:35AM -0600, Rob Landley wrote:
> > > Using qemu-system-sh4, this commit:
> > >
> > > e1c09175bc00dd8dfb2ad1b26e1858dcdc109b59 is first bad commit
> > > commit e1c09175bc00dd8dfb2ad1b26e1858dcdc109b59
> > > Author: Gerd Hoffmann <kraxel@redhat.com>
> > > Date:   Tue Dec 8 13:11:44 2009 +0100
> > >
> > >     zap serial_monitor_mux
> > >
> > >     The logic in this code obviously predates the multiple monitor
> > >     capability of qemu and looks increasingly silly these days.
> > >
> > >     I think the intention of this piece of code is to get a reasonable
> > >     default for the -nographic case: have monitor and serial line muxed
> > >     on stdio.
> > >
> > >     With the new default_serial and default_monitor variables we have
> > > now doing just that became much easier ;)
> > >
> > >     Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> > >     Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
> > >
> > > Made "-serial stdio" now do:
> > >
> > >   chardev: opening backend "stdio" failed
> > >   qemu: could not open serial device 'stdio': Inappropriate ioctl for
> > > device
> > >
> > > Am I using it wrong?
> > >
> > > If I don't override it, I instead get:
> > >
> > >   long read to SH7750_WCR1_A7 (0x000000001f800008) ignored
> > >   long read to SH7750_WCR2_A7 (0x000000001f80000c) ignored
> > >   long read to SH7750_WCR3_A7 (0x000000001f800010) ignored
> > >   long read to SH7750_MCR_A7 (0x000000001f800014) ignored
> > >   long read to SH7750_MCR_A7 (0x000000001f800014) ignored
> > >   sh_serial: unsupported read from 0x10
> > >   qemu-system-sh4: /home/landley/qemu/git/hw/sh_serial.c:285:
> > >     sh_serial_ioport_read: Assertion `0' failed.
> > >
> > > Is anybody out there actually using the sh4 emulator?  I've gotten it
> > > to work several times before on various random git snapshots, but never
> > > in a release version...
> >
> > Would be nice to give that you give the actual command line to try to
> > use. I am using the stable-0.12 branch (plus some backporting patches
> > for the MMU) to do development on sh4. HEAD still works correctly here.
> > I am using the following command line:
>
> Mine's:
>
> And the kernel and root filesystem images to use with it are at:
>   http://impactlinux.com/fwl/downloads/binaries/system-image-sh4.tar.bz2
>
> Note the binaries in that tarball used to work, they haven't been rebuilt
> since the last time I had a qemu that could run them.  The problem it was a
> random -git snapshot (no release version has ever worked for me on sh4,
> that I'm aware of) and I don't remember which version.  However, my blog
> says that on November 5, 2009 I had it working with unpatched git, so
> presumably I could grab a qemu from there and bisect...
>
> > ~/git/qemu/sh4-softmmu/qemu-system-sh4 -M r2d -kernel zImage -drive
> > file=sh4.img -serial null -serial stdio -usb -append "root=/dev/sda1
> > console=tty0 console=ttySC0,115200 noiotrap" -usbdevice keyboard
> > -usbdevice mouse
>
> I found my blog entries about getting sh4 to work back in February 2009:
>
>  http://landley.net/notes-2009.html#13-02-2009
>
> Which told me when to look in the archives to find the threads where I
> learned how to make it work in the first place.  This one is where I got
> the -serial null -serial stdio advice:
>
>   http://lists.gnu.org/archive/html/qemu-devel/2009-02/msg00825.html
>
> And here's where I got an actual command prompt:
>
>   http://lists.gnu.org/archive/html/qemu-devel/2009-02/msg00961.html
>
> Meaning I've figured out who I should have cc'd in the first place. :)
>
> Rob

I found out that "-serial stdio" is apparently trying to open /dev/stdio, 
which Ubuntu 9.04 hasn't got.  If I say -serial /dev/tty it works from the 
command line (but not in scripts).

Before the zap_serial_monitor_mux commit above, -serial stdio was being 
recognized as a builtin or some such, and didn't depend on /dev/stdio.

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds