On 04/13/2012 06:21 PM, Idwer Vollering wrote:
>
>
> Op 13 april 2012 23:27 schreef Srini Amble <srini.amble(a)windriver.com
> <mailto:srini.amble@windriver.com>> het volgende:
>
> I have an x86 board which has X58 and ICH10 chipset. If any of you
> know how I can disable the LAN device in ICH chipset (dev ID 25
> and function number 0) in SeaBIOS then please share it with me.
>
>
> How do you start SeaBIOS? E.G. if you use coreboot to start SeaBIOS
> you can (AFAIK) disable the LAN device from devicetree.cb.
SeaBIOS is the payload for coreboot in my case. Looking at the source
files I have for coreboot and SeaBIOS I do not see reference to
'devicetree.cb' any where. Also, any pointer to how to disable the
device in 'devicetree.cb' is much appreciated.
I have an x86 board which has X58 and ICH10 chipset. If any of you know
how I can disable the LAN device in ICH chipset (dev ID 25 and function
number 0) in SeaBIOS then please share it with me.
I'm working towards a goal of having specific devices be bootable, and
*only* those devices. For example, if my bootorder file specifies SATA
drive 3, I do not want it to try SATA drives 0, 1, and 2, nor any other HDD
or floppy that it finds.
My first question is: how do I do this?
If that can't be answered, can someone explain to me the relationship
between the bootorder file and the BEV (Boot Execution Vector) configured in
boot.c?
Thanks!
-- Steve G.
Hi Kevin,
Is there a reason why drive_s structures are allocated from f-segment?
The space for f-segment allocations are barely sufficient for bios
tables.
--
Gleb.
On Thu, Apr 12, 2012 at 04:10:21PM +0900, Daniel Castro wrote:
> On Thu, Apr 12, 2012 at 9:34 AM, Kevin O'Connor <kevin(a)koconnor.net> wrote:
> > On Wed, Apr 11, 2012 at 09:09:44PM +0900, Daniel Castro wrote:
> >> Hello All,
> >>
> >> Since the hypercall on 16bit is not working, how can I transition the
> >> CPU to 32bit while doing the I/O stuff.
> >> And maybe get back to 16 Bit?
> >
> > See the call32() function.
>
> Thanks for the answer Kevin, correct me if I am wrong, but I can only
> pass a single pointer argument to the function?
Yes - one parameter to the target function.
> Also in the pointer argument I get the return value, right?
No, the return code of the target function becomes the return code of
call32.
-Kevin
On Wed, Apr 11, 2012 at 09:09:44PM +0900, Daniel Castro wrote:
> Hello All,
>
> Since the hypercall on 16bit is not working, how can I transition the
> CPU to 32bit while doing the I/O stuff.
> And maybe get back to 16 Bit?
See the call32() function.
-Kevin
Hello All,
Since the hypercall on 16bit is not working, how can I transition the
CPU to 32bit while doing the I/O stuff.
And maybe get back to 16 Bit?
It appears that the 16bit CPU emulation code on Xen side can not
handle the hypercall.
--
+-=====---------------------------+
| +---------------------------------+ | This space intentionally blank
for notetaking.
| | | Daniel Castro, |
| | | Consultant/Programmer.|
| | | U Andes |
+-------------------------------------+
Hello all,
I am have a small problem, when running the BIOS in 16Bit if Y try to
do a hypercall the virtual machines crashes.
Here is the code of the hypercall:
#define _hypercall2(type, name, a1, a2) \
({ \
unsigned long __hentry = xen_hypercall_page+__HYPERVISOR_##name*32; \
long __res, __ign1, __ign2; \
asm volatile ( \
"call *%%eax" \
: "=a" (__res), "=b" (__ign1), "=c" (__ign2) \
: "0" (__hentry), "1" ((long)(a1)), "2" ((long)(a2)) \
: "memory" ); \
(type)__res; \
})
I guess the something is being passed wrong to the call, and the
return address is damaged...
Daniel
--
+-=====---------------------------+
| +---------------------------------+ | This space intentionally blank
for notetaking.
| | | Daniel Castro, |
| | | Consultant/Programmer.|
| | | U Andes |
+-------------------------------------+
On Tue, Apr 10, 2012 at 06:01:52AM -0600, Steve Goodrich wrote:
> > On Mon, Apr 09, 2012 at 10:29:03AM -0600, Steve Goodrich wrote:
> > > Anyway, the code as written was finding the first video mode of the
> > correct
> > > width/height and using that. Unfortunately, the mode was a 5-5-5
> > while the
> > > JPG was being decoded as 5-6-5 (bits per RGB channel).
> >
> > What happened when the 5-5-5 mode was chosen? The code should still
> > have worked (if not as pretty looking).
>
> The 5-5-5 mode showed the correct image, but the colors were off: for
> example, some of the green areas had a significant red component (some
> greens were almost entirely red).
Oops - I was confusing the above with the automatic conversion from
24bpp to 16bpp. You are right that 15bpp isn't supported. It
shouldn't be too hard to alter the jpeg code to support 15bpp, but it
should also be safe to just skip 15bpp modes. Along the same lines,
the code should really try to select a 24bpp mode if one is available.
Looking closer at this, it also appears the SeaVGABIOS code isn't
encoding the 15bpp mode correctly in its vbe tables.
-Kevin
On Mon, Apr 09, 2012 at 10:29:03AM -0600, Steve Goodrich wrote:
> Anyway, the code as written was finding the first video mode of the correct
> width/height and using that. Unfortunately, the mode was a 5-5-5 while the
> JPG was being decoded as 5-6-5 (bits per RGB channel).
What happened when the 5-5-5 mode was chosen? The code should still
have worked (if not as pretty looking).
-Kevin