[ANN] Bochs bios ported to gcc

[ANN] Bochs bios ported to gcc

[Kevin O'Connor]

Hi,

I have been working on a port of bochs bios to gcc.  This port is
nearly complete.  The new code does not rely on bcc or dev86.
Instead, it uses standard gcc and gas.  It should compile on any
recent Linux distribution.

I'm sending this email because I understand kvm has a copy of bochs
bios in its repository.

I am successfully booting freedos, linux, and netbsd with qemu.  I'm
interested in finding people willing to test the bios on other images
and in different environments.  The code is still in a testing phase -
I don't think it will eat your data, but please be careful.

Source and pre-compiled bios images are available (use version 0.2.0)
at:

http://linuxtogo.org/~kevin/legacybios/


More information on the details of the project can be found in the
'README' file, or at:

http://git.linuxtogo.org/?p=kevin/legacybios.git;a=blob_plain;f=README


There is also a source code repository located at:

http://git.linuxtogo.org/


Comments?
-Kevin

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace

Re: [ANN] Bochs bios ported to gcc

[Avi Kivity]

Kevin O'Connor wrote:
> I have been working on a port of bochs bios to gcc.  This port is
> nearly complete.  The new code does not rely on bcc or dev86.
> Instead, it uses standard gcc and gas.  It should compile on any
> recent Linux distribution.
>
> I'm sending this email because I understand kvm has a copy of bochs
> bios in its repository.
>   

While moving away from the horror that is bcc is a blessing, the way to 
really benefit from it is to have this code replace the original bochs 
bios.  This way updates to the bochs bios are not lost for anyone using 
the new code.

Have you proposed replacing the bios to the bochs developers?  What was 
the reaction?

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Register now and save $200. Hurry, offer ends at 11:59 p.m., 
Monday, April 7! Use priority code J8TLD2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [ANN] Bochs bios ported to gcc

[Kevin O'Connor]

On Sun, Apr 06, 2008 at 03:05:11PM +0300, Avi Kivity wrote:
> Kevin O'Connor wrote:
>> I have been working on a port of bochs bios to gcc.
>
> While moving away from the horror that is bcc is a blessing, the way to 
> really benefit from it is to have this code replace the original bochs 
> bios.  This way updates to the bochs bios are not lost for anyone using the 
> new code.

I agree.  I've been interacting with the bochs developers since I
started (about a month and half ago).

> Have you proposed replacing the bios to the bochs developers?  What was the 
> reaction?

The reaction was pretty "quiet".  The bochs project is not nearly as
active as kvm, and the bios is just one small part of that project.

Now that the new code can actually launch a wide variety of payloads,
I'm hoping there will be more interest.

BTW, if anyone is interested in testing, please grab the 0.2.1 version
- the v0.2.0 version had a nasty bug on disk writes to large disks
(assuming one uses an OS that writes to disk via the bios).

-Kevin

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Register now and save $200. Hurry, offer ends at 11:59 p.m., 
Monday, April 7! Use priority code J8TLD2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [ANN] Bochs bios ported to gcc

[H. Peter Anvin]

Avi Kivity wrote:
> Kevin O'Connor wrote:
>> I have been working on a port of bochs bios to gcc.  This port is
>> nearly complete.  The new code does not rely on bcc or dev86.
>> Instead, it uses standard gcc and gas.  It should compile on any
>> recent Linux distribution.
>>
>> I'm sending this email because I understand kvm has a copy of bochs
>> bios in its repository.
>>   
> 
> While moving away from the horror that is bcc is a blessing, the way to 
> really benefit from it is to have this code replace the original bochs 
> bios.  This way updates to the bochs bios are not lost for anyone using 
> the new code.
> 
> Have you proposed replacing the bios to the bochs developers?  What was 
> the reaction?
> 

bcc is truly horrible.  gcc is, of course, best, although it does 
produce bloated 16-bit code.  For "proper" 16-bit code, OpenWatcom is 
currently the best open source compiler, and it's actually getting 
usable even on a Linux host.

	-hpa

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Register now and save $200. Hurry, offer ends at 11:59 p.m., 
Monday, April 7! Use priority code J8TLD2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [ANN] Bochs bios ported to gcc

[Kevin O'Connor]

On Sun, Apr 06, 2008 at 11:42:11PM -0700, H. Peter Anvin wrote:
> bcc is truly horrible.  gcc is, of course, best, although it does produce 
> bloated 16-bit code.

With gcc, the text size has actually been significantly reduced - I'm
not sure if this is because I'm using -fwhole-program and -Os or if it
is just the result of bcc being so bad.

The biggest problem I've run into is stack usage on the 16-bit bios
calls.

>For "proper" 16-bit code, OpenWatcom is currently the 
> best open source compiler, and it's actually getting usable even on a Linux 
> host.

One of the developers on the bochs list also recommended OpenWatcom.

The bios currently has quite a bit of 32bit code in addition to the
16bit code.  It's very convenient to be able to compile the same code
in both modes (eg, printf, inb, etc.).  I'd need to move both 32bit
and 16bit code to OpenWatcom to do this.

-Kevin

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Register now and save $200. Hurry, offer ends at 11:59 p.m., 
Monday, April 7! Use priority code J8TLD2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [ANN] Bochs bios ported to gcc

[H. Peter Anvin]

Kevin O'Connor wrote:
> On Sun, Apr 06, 2008 at 11:42:11PM -0700, H. Peter Anvin wrote:
>> bcc is truly horrible.  gcc is, of course, best, although it does produce 
>> bloated 16-bit code.
> 
> With gcc, the text size has actually been significantly reduced - I'm
> not sure if this is because I'm using -fwhole-program and -Os or if it
> is just the result of bcc being so bad.
> 

The 16-bit code is still horrid.  I have done some experiments with a 
16-bit backend to gcc last summer, and even when it could barely crawl, 
and had effectively no optimizations, it still produced 15% smaller code 
than the 32-bit backend with .code16gcc and the best optimizations I 
could find.

> The biggest problem I've run into is stack usage on the 16-bit bios
> calls.
> 
>> For "proper" 16-bit code, OpenWatcom is currently the 
>> best open source compiler, and it's actually getting usable even on a Linux 
>> host.
> 
> One of the developers on the bochs list also recommended OpenWatcom.
> 
> The bios currently has quite a bit of 32bit code in addition to the
> 16bit code.  It's very convenient to be able to compile the same code
> in both modes (eg, printf, inb, etc.).  I'd need to move both 32bit
> and 16bit code to OpenWatcom to do this.

Yes, of course.  OpenWatcom does do both, although I think gcc is a 
better compiler than OpenWatcom in 32-bit mode.  OpenWatcom isn't bad, 
though.

As long as it's just C, it's not really an issue, though.

	-hpa

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Register now and save $200. Hurry, offer ends at 11:59 p.m., 
Monday, April 7! Use priority code J8TLD2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone