2016-08-02 06:12:44 |
Stefan Weil |
description |
The QEMU PC emulation of DMA does not behave like real hardware or other virtualization software.
From the original bug report (Benjamin David Lunt):
Back to the READ_DMA command, it is my conclusion that the
READ_DMA command, more precisely, the BUS Master part of QEMU is
in error. The tests that people have done to see if it works, is
probably the guest finding out that DMA doesn't work and defaulting
to PIO, but since the read was successful visually to the user, the
user assumed the READ_DMA command works, where the guest actually
defaulted back to PIO transfers without notice.
My code works on real hardware (numerous machines), Bochs, and Oracle's
Virtual Box.
...
I have a small test suite, zipped and included at:
www.fysnet.net/temp/c8bug.zip
Within this zip file is a.img. This is a freeDOS bootable
floppy. Emulate it with QEMU and then at the DOS prompt, run
c8bug.exe.
It will say that the drive is not ready.
"Drive never became ready"
(along with a few other lines of text)
The source for this test suite is also included.
c8bug.c is the c source code
c8bug.h is the header file
ctype.h is a quick way to get 'bit8u' type defines
timer.h is a delay routine from another project I have
(The base IO addresses are assumed and set at the top of c8bug.c)
(compiled with DJGPP for DPMI)
q.bat is my command line for QEMU
On all other machines and VirtualBox, the controller is ready
for me to receive the sector data.
"Drive is ready to transmit data..."
However, in QEMU, the controller never becomes ready.
"Drive never became ready"
The bug was reported for QEMU for Windows, but I can confirm that QEMU for Linux also shows that
behaviour, while VirtualBox works. |
The QEMU PC emulation of DMA does not behave like real hardware or other virtualization software.
From the original bug report (Benjamin David Lunt):
Back to the READ_DMA command, it is my conclusion that the
READ_DMA command, more precisely, the BUS Master part of QEMU is
in error. The tests that people have done to see if it works, is
probably the guest finding out that DMA doesn't work and defaulting
to PIO, but since the read was successful visually to the user, the
user assumed the READ_DMA command works, where the guest actually
defaulted back to PIO transfers without notice.
My code works on real hardware (numerous machines), Bochs, and Oracle's
Virtual Box.
...
I have a small test suite, zipped and included at:
www.fysnet.net/temp/c8bug.zip
Within this zip file is a.img. This is a freeDOS bootable
floppy. Emulate it with QEMU and then at the DOS prompt, run
c8bug.exe.
It will say that the drive is not ready.
"Drive never became ready"
(along with a few other lines of text)
The source for this test suite is also included.
c8bug.c is the c source code
c8bug.h is the header file
ctype.h is a quick way to get 'bit8u' type defines
timer.h is a delay routine from another project I have
(The base IO addresses are assumed and set at the top of c8bug.c)
(compiled with DJGPP for DPMI)
q.bat is my command line for QEMU
On all other machines and VirtualBox, the controller is ready
for me to receive the sector data.
"Drive is ready to transmit data..."
However, in QEMU, the controller never becomes ready.
"Drive never became ready"
The bug was reported for QEMU for Windows, but I can confirm that QEMU for Linux also shows that behaviour, while VirtualBox works. |
|