Date: Mon, 4 Jun 2007 22:56:52 -0400 From: "Andrey Shuvikov" <mr.hyro@gmail.com> To: freebsd-acpi@freebsd.org Subject: Re: Computer hangs on reboot Message-ID: <24393ae80706041956g286b1e97jfa649a3a1f530f5a@mail.gmail.com> In-Reply-To: <200706041119.26958.jhb@freebsd.org> References: <24393ae80705172026u6194f11esc525d52470aa71f9@mail.gmail.com> <24393ae80705290916i7962bfaak62ea867494d2e77b@mail.gmail.com> <24393ae80706012112he0e2a96x1e0da9e8f5f6d1b0@mail.gmail.com> <200706041119.26958.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 6/4/07, John Baldwin <jhb@freebsd.org> wrote:
> On Saturday 02 June 2007 12:12:14 am Andrey Shuvikov wrote:
> > On 5/29/07, Andrey Shuvikov <mr.hyro@gmail.com> wrote:
> > > On 5/29/07, John Baldwin <jhb@freebsd.org> wrote:
> > > > On Saturday 26 May 2007 01:14:52 pm Andrey Shuvikov wrote:
> > > > > I tried to add some debug output to cpu_reset_real() code:
> > > > >
> > > > > printf("cpu_reset_real(): before keyboard reset\n");
> > > > > outb(IO_KBD + 4, 0xFE);
> > > > > printf("cpu_reset_real(): after keyboard reset\n");
> > > > >
> > > > > The last two lines I see on the screen when I try to reboot are:
> > > > >
> > > > > cpu_reset_real(): before keyboard reset
> > > > > c
> > > > >
> > > > > So the computer is halted but doesn't go to reboot.
> > > >
> > > > So what if you disable the keyboard reset and let the other methods
> happen?
> > > >
> > > > --
> > > > John Baldwin
> > > >
> > >
> > > I tried to set BROKEN_KEYBOARD_RESET option but nothing changed
> > > (except that I didn't see these debugging messages, obviously). Is it
> > > possible that keyboard controller issues the reset signal but the CPU
> > > waits for something? Or reset pin is disabled somehow?
> > >
> >
> > Tried to investigate further... Added code to print keyboard
> > controller output port to make sure A20 is open. Got 0x0B. Does it
> > mean that A20 is really open? If A20 is controlled by some kind of
> > "Fast A20" option, will keyboard controller still show the right
> > status? And what else can I check?
>
> I believe I/O port 0x92 can gate A20 as well, and that it is only open if both
> say it is open. Grab a datasheet for an ICH from developer.intel.com to find
> more details on port 0x92.
>
> --
> John Baldwin
>
According to Intel documentation keyboard controller and port 0x92 are
OR-ed together, which means it's enough to use either one. Anyway I
checked 0x92 (which was 0x00) and set it to 0x02. Didn't help. :-(
>From the other hand my chipset is not Intel and I found some
references to reboot problem for my motherboard under linux in 2002.
These references mention some LPC configuration parameters of Sis630
which enable/disable A20 and allow software reset. I downloaded Sis630
datasheet and will check it tomorrow.
What I don't understand though is that OpenBSD, which is installed on
the same machine doesn't have any problem rebooting...
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?24393ae80706041956g286b1e97jfa649a3a1f530f5a>