2003-03-09 16:22:20

by Andi Kleen

Subject: [PATCH] Work around console initialization ordering problem

Works around the console ordering problem in 2.5.64-bk3. Following
the similar fix I did for x86-64.

-Andi

--- linux-2.5.64-work/arch/i386/kernel/setup.c-o 2003-03-05 10:40:08.000000000 +0100
+++ linux-2.5.64-work/arch/i386/kernel/setup.c 2003-03-09 17:27:57.000000000 +0100
@@ -516,6 +516,9 @@
int len = 0;
int userdef = 0;

+ if (!strstr(saved_command_line, "console="))
+ strcat(saved_command_line, " console=tty0");
+
/* Save unparsed command line copy for /proc/cmdline */
memcpy(saved_command_line, COMMAND_LINE, COMMAND_LINE_SIZE);
saved_command_line[COMMAND_LINE_SIZE-1] = '\0';

2003-03-09 21:34:01

by Andrew Morton

Subject: Re: [PATCH] Work around console initialization ordering problem

Andi Kleen <[email protected]> wrote:
>
>
> Works around the console ordering problem in 2.5.64-bk3. Following
> the similar fix I did for x86-64.
> ...
> + if (!strstr(saved_command_line, "console="))
> + strcat(saved_command_line, " console=tty0");
> +

We can do it by shuffling the link order:


diff -puN drivers/Makefile~console-ordering-fix drivers/Makefile
--- 25/drivers/Makefile~console-ordering-fix 2003-03-09 02:48:33.000000000 -0800
+++ 25-akpm/drivers/Makefile 2003-03-09 02:48:33.000000000 -0800
@@ -11,9 +11,10 @@ obj-$(CONFIG_ACPI) += acpi/
# PnP must come after ACPI since it will eventually need to check if acpi
# was used and do nothing if so
obj-$(CONFIG_PNP) += pnp/
+obj-y += char/
obj-y += serial/
obj-$(CONFIG_PARPORT) += parport/
-obj-y += base/ char/ block/ misc/ net/ media/
+obj-y += base/ block/ misc/ net/ media/
obj-$(CONFIG_NUBUS) += nubus/
obj-$(CONFIG_ATM) += atm/
obj-$(CONFIG_IDE) += ide/

_

2003-03-09 23:24:15

by Andi Kleen

Subject: Re: [PATCH] Work around console initialization ordering problem

On Sun, Mar 09, 2003 at 10:45:06PM +0100, Andrew Morton wrote:
> Andi Kleen <[email protected]> wrote:
> >
> >
> > Works around the console ordering problem in 2.5.64-bk3. Following
> > the similar fix I did for x86-64.
> > ...
> > + if (!strstr(saved_command_line, "console="))
> > + strcat(saved_command_line, " console=tty0");
> > +
>
> We can do it by shuffling the link order:

Yes, but someone will surely break it again. I feel my low tech solution is
less fragile.

-Andi

2003-03-09 23:46:29

by Andries Brouwer

Subject: Re: [PATCH] Work around console initialization ordering problem

On Mon, Mar 10, 2003 at 12:34:47AM +0100, Andi Kleen wrote:
> On Sun, Mar 09, 2003 at 10:45:06PM +0100, Andrew Morton wrote:
> > Andi Kleen <[email protected]> wrote:
> > >
> > >
> > > Works around the console ordering problem in 2.5.64-bk3. Following
> > > the similar fix I did for x86-64.
> > > ...
> > > + if (!strstr(saved_command_line, "console="))
> > > + strcat(saved_command_line, " console=tty0");
> > > +
> >
> > We can do it by shuffling the link order:
>
> Yes, but someone will surely break it again. I feel my low tech solution is
> less fragile.

But what about COMMAND_LINE_SIZE?

2003-03-10 00:15:25

by Andi Kleen

Subject: Re: [PATCH] Work around console initialization ordering problem

On Mon, Mar 10, 2003 at 12:57:01AM +0100, Andries Brouwer wrote:
> But what about COMMAND_LINE_SIZE?

It's normally 2k. Is that likely to be reached?

-Andi

2003-03-10 00:31:22

by Johannes Erdfelt

Subject: Re: [PATCH] Work around console initialization ordering problem

On Mon, Mar 10, 2003, Andi Kleen <[email protected]> wrote:
> On Sun, Mar 09, 2003 at 10:45:06PM +0100, Andrew Morton wrote:
> > Andi Kleen <[email protected]> wrote:
> > >
> > >
> > > Works around the console ordering problem in 2.5.64-bk3. Following
> > > the similar fix I did for x86-64.
> > > ...
> > > + if (!strstr(saved_command_line, "console="))
> > > + strcat(saved_command_line, " console=tty0");
> > > +
> >
> > We can do it by shuffling the link order:
>
> Yes, but someone will surely break it again. I feel my low tech solution is
> less fragile.

Wouldn't a well placed comment solve that?

JE