[Top][All Lists]

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

Re: [Qemu-devel] [PATCH] tcg, tci: Add TCG and interpreter for bytecode

From: malc
Subject: Re: [Qemu-devel] [PATCH] tcg, tci: Add TCG and interpreter for bytecode (virtual machine)
Date: Sun, 11 Oct 2009 20:55:34 +0400 (MSD)

On Sun, 11 Oct 2009, Stefan Weil wrote:

> Stuart Brady schrieb:
> > On Mon, Sep 28, 2009 at 06:50:21PM +0200, Stefan Weil wrote:
> >> Please send patches / enhancements based on my latest
> >> version from git://repo.or.cz/qemu/ar7.git.
> >
> > Just bug reports for now, unfortunately...
> >
> > exec-all.h defines 'USE_DIRECT_JUMP' based on the host architecture,
> > which is not relevant under TCI. Also, a tb_set_jmp_target1()
> > implementation is needed for TCI.
> Done. Tested with mips / mipsel host.
> >
> > I think tcg_out64() belongs in tcg.c along with tcg_out8/16/32().
> > There are places where the x86_64, ppc64 and s390 TCG targets call
> > out32() twice, which should be replaced with a single call to out64().
> This is still to be done.
> >
> > tcg_out16(), tcg_out32() and tcg_out64(), etc. rely on the code buffer
> > pointer being aligned on those architectures that require alignment...
> > We'd need definitions of these in terms of stw_p(), stl_p() and stq_p()
> > for TCI... Is there any reason to preserve the old definitions?
> >
> > (BTW, there's no 'ldul' macro, so perhaps one should be added, although
> > the names in that set of macros and functions are rather unclear, so
> > perhaps some more descriptive naming could be used for new code.)
> >
> > Cheers,
> Here is a status update:
> A basic test of tcg + interpreter which includes system emulation
> (pc bios) and linux user mode (simple hello) now works for
> i386, x86_64, mips and mipsel hosts with i386 and x86_64 guests.
> Endianess and 32/64 bit issues should be fixed now.
> Alignment is still to be done (my test platforms don't need it).
> So it is finally possible again to run qemu on mips hosts,
> and any 32 or 64 bit host should be able to run qemu, too.
> Of course the interpreted code is not fast - in my test on
> an emulated mips malta machine it is even terribly slow.
> git://repo.or.cz/qemu/ar7.git contains the latest changes.
> If the maintainers want to integrate tci in the official qemu,
> I can prepare the patches needed to add bytecode generator
> and interpreter and tcg host support for any host.

I'm, at the very least, will definitely look at it.


reply via email to

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