Page 1 of 2

Bochs 2.4 final released

Posted: Sun May 03, 2009 4:10 am
by stlw
Hi all!

The final release of Bochs 2.4 is out now. The sources and binary package for Linux are available for download. I have also updated the online documentation. The packages for Windows will be available soon.

Brief summary of changes since release 2.3.7:

- Added graphical Bochs debugger frontend for most of the supported
platforms.
- Thanks for Chourdakis Michael and Bruce Ewing.
- Many new CPU features in emulation
- Support for > 32 bit physical address space and configurable MSRs
- VMX, 1G pages in long mode, MOVBE instruction
- Bugfixes for CPU emulation correctness, debugger and CPU instrumentation.
- New config interface 'win32config' with start and runtime menu
- USB: added OHCI support, external hub and cdrom
- Added user plugin interface support.

--
Volker Ruppert

Re: Bochs 2.4 final released

Posted: Sun May 03, 2009 6:52 am
by Zenith
Congrats! You guys FINALLY made it to the final release :lol:

As for me, guess what I'll be doing during the next hour 8)

Re: Bochs 2.4 final released

Posted: Sun May 03, 2009 1:07 pm
by OrOS
Duo! Excellent work guys.

Re: Bochs 2.4 final released

Posted: Sun May 03, 2009 1:13 pm
by Troy Martin
Whee!

Downloaded it this morning, it's really stable and a lot easier to work with!

Re: Bochs 2.4 final released

Posted: Mon May 04, 2009 10:59 am
by Creature
Is it me or did Bochs console output through port 0xE9 disappear? My debug prints suddenly seem to have no effect anymore after updating.

Re: Bochs 2.4 final released

Posted: Mon May 04, 2009 1:23 pm
by stlw
Creature wrote:Is it me or did Bochs console output through port 0xE9 disappear? My debug prints suddenly seem to have no effect anymore after updating.
You should enable them in .bochsrc. And probably read release notes as well. Also checking for configure script errors could be nice.

Stanislav

Re: Bochs 2.4 final released

Posted: Mon May 04, 2009 2:51 pm
by IanSeyler
How do you enable the GUI debugger now?

They still don't compile SMP in by default! Arg!

Re: Bochs 2.4 final released

Posted: Tue May 05, 2009 8:34 am
by Creature
stlw wrote:
Creature wrote:Is it me or did Bochs console output through port 0xE9 disappear? My debug prints suddenly seem to have no effect anymore after updating.
You should enable them in .bochsrc. And probably read release notes as well. Also checking for configure script errors could be nice.

Stanislav
Yes, that was it, sorry. It failed because in the pre-release of Bochs, I never explicitly turned it on and it still worked.

Re: Bochs 2.4 final released

Posted: Fri May 08, 2009 4:38 pm
by AUsername
Works much better, but I'm having issues changing the BIOS. No matter what BIOS I use Boch always seems to throw some sort of error, I even used the latest BIOS version from the SVN and it still seems to throw errors.
00018789910e[CPU0 ] SLDT: not recognized in real or virtual-8086 mode
As for why I am it's because the "A:" is highlighted all the time at the bottom of the screen in my OS. In DOS this turns off, so I made a bootloader that just halts the system and the light still remains there. I'm going to assume that the highlight means activity because it only turns on if you type a command into DOS and it searches the disk for a program.

Don't know if that's a bug in the BIOS or if I'm just doing something wrong. ^-^

Re: Bochs 2.4 final released

Posted: Sat May 09, 2009 12:38 am
by whowhatwhere
It appears USB support is still broken. It fails to build on ~amd64 Gentoo with USB-OHCI enabled.

Re: Bochs 2.4 final released

Posted: Sat May 09, 2009 12:44 am
by stlw
syntropy wrote:It appears USB support is still broken. It fails to build on ~amd64 Gentoo with USB-OHCI enabled.
What are the compilation errors ?

Re: Bochs 2.4 final released

Posted: Sat May 09, 2009 12:47 am
by stlw
AUsername wrote:Works much better, but I'm having issues changing the BIOS. No matter what BIOS I use Boch always seems to throw some sort of error, I even used the latest BIOS version from the SVN and it still seems to throw errors.
00018789910e[CPU0 ] SLDT: not recognized in real or virtual-8086 mode
As for why I am it's because the "A:" is highlighted all the time at the bottom of the screen in my OS. In DOS this turns off, so I made a bootloader that just halts the system and the light still remains there. I'm going to assume that the highlight means activity because it only turns on if you type a command into DOS and it searches the disk for a program.

Don't know if that's a bug in the BIOS or if I'm just doing something wrong. ^-^
You'd better to ask such questions in mailing list. But anyway, I sure nobody will answer you unless you provide some more details.
Like:
- What are you doing ?
- What is the different bios you change to and are you doing it ? Where is the SVN you talking about ?
- I never seen bochs floppy "A:" indicator stays on, so you willl have to provide test example.

Stanislav

Re: Bochs 2.4 final released

Posted: Sat May 09, 2009 10:07 am
by Troy Martin
I think I know the answer, and it's not a bochs issue...

Your bootsector doesn't wait for the floppy motor to shut off (so it keeps going and the light doesn't go out.)

Re: Bochs 2.4 final released

Posted: Sat May 09, 2009 9:25 pm
by JohnnyTheDon
syntropy wrote:It appears USB support is still broken. It fails to build on ~amd64 Gentoo with USB-OHCI enabled.
I have the same problem on amd64 ubuntu, but only when I use it with --enable-long-phy-address. --enable-long-phy-address also causes runtime errors.

--enable-usb-ohci and --enable-long-phy-address build errors:

Code: Select all

usb_ohci.cc: In member function ‘virtual void bx_usb_ohci_c::pci_write_handler(Bit8u, Bit32u, unsigned int)’:
usb_ohci.cc:1407: error: no matching function for call to ‘bx_pci_bridge_stub_c::pci_set_base_mem(bx_usb_ohci_c*&, bx_bool (&)(bx_phy_address, unsigned int, void*, void*), bx_bool (&)(bx_phy_address, unsigned int, void*, void*), bx_phy_address*, Bit8u*, int)’
iodev.h:278: note: candidates are: virtual bx_bool bx_pci_bridge_stub_c::pci_set_base_mem(void*, bx_bool (*)(bx_phy_address, unsigned int, void*, void*), bx_bool (*)(bx_phy_address, unsigned int, void*, void*), Bit32u*, Bit8u*, unsigned int)
--enable-long-phy-address runtime error:

Code: Select all

bx_dbg_read_linear: physical memory read error (phy=0x00000000fffffff0, lin=0x00000000fffffff0)
<bochs:1> c
00000032767i[CPU0 ] LOCK prefix unallowed (op1=0x2, attr=0x0, mod=0x0, nnn=0)
GCC Version: 4.3.3

Re: Bochs 2.4 final released

Posted: Sun May 10, 2009 10:07 am
by quok
I noticed that building on OS X 10.5 is broken because of the use if /rsrc and not ../namedfork/rsrc in the Makefiles. I fixed that, though. Also, iodebug.cc was still using cpu->name when BX_DEBUGGER was #defined, so I also fixed that. I've attached the patches here (they are against 2.4 final), and will also email them to Stanislav shortly.

@JohnnyTheDon: I ran across that same build problem, but only when I was using --enable-plugins and had that specified at the beginning of the ./configure line. I did a make clean, reconfigured with --enable-plugins at the end of the line, and everything built fine.