Re: Keyboard Jammed error patch 2.4.35-pre4

[David W Studeman <avionicsdv@aim.com>]

Willy Tarreau wrote:
> Hello David,
> 
> first, please always take care of keeping persons in CC when replying to
> LKML, otherwise you risk your mail being missed.
> 
> On Sun, Sep 21, 2008 at 06:00:50AM -0700, David W Studeman wrote:
>> Ok, I finally applied the patch manually to 2.4.36.7 and then made my own
>> diff from that. Your patch would likely have applied if I increased the
>> fuzz factor as it was merely off by a number of lines. When trying to
>> compile with no keyboard or VT support, I still get this error:
>>
>> arch/i386/kernel/kernel.o(.text.init+0x33e6): In function
>> `disable_console_keyboard':
>> : undefined reference to `keyboard_controller_present'
>> make: *** [vmlinux] Error 1
> 
> I know what causes this. The fix is merged into 2.4.37-rc1. I've been
> using it for a long time without even noticing that it was still not
> merged. The fix is trivial as it simply removes the disable_console_keyboard
> function when CONFIG_VT is not set. This is needed because the variable
> keyboard_controller_present is only defined when CONFIG_VT is set.
> 
> Here's the patch if it can help you.
> 
> Also, next time you encounter problems, please post your config. It's
> very hard to troubleshoot build bugs without a config. I would have
> immediately spotted this missing patch on the first time had I got
> that config.
> 
> Regards,
> Willy
> 
> ----
> From 90e228888a59d630c8094d5d1b273b00a2c76cc0 Mon Sep 17 00:00:00 2001
> From: Willy Tarreau <w@1wt.eu>
> Date: Sun, 7 Sep 2008 12:37:51 +0200
> Subject: x86 would not build without CONFIG_VT
> 
> I've been using this patch for a while without noticing it never
> went into mainline. It is required to build i386 without CONFIG_VT.
> 
> Signed-off-by: Willy Tarreau <w@1wt.eu>
> ---
>  arch/i386/kernel/dmi_scan.c |    6 +++++-
>  1 files changed, 5 insertions(+), 1 deletions(-)
> 
> diff --git a/arch/i386/kernel/dmi_scan.c b/arch/i386/kernel/dmi_scan.c
> index 07c6d94..9c27ca2 100644
> --- a/arch/i386/kernel/dmi_scan.c
> +++ b/arch/i386/kernel/dmi_scan.c
> @@ -524,6 +524,7 @@ static __init int disable_acpi_pci(struct dmi_blacklist *d)
>   */
>   
>  
> +#ifdef CONFIG_VT
>  /*      IBM bladeservers have a USB console switch. The keyboard type is USB
>   *      and the hardware does not have a console keyboard. We disable the
>   *      console keyboard so the kernel does not try to initialize one and
> @@ -538,6 +539,7 @@ static __init int disable_console_keyboard(struct dmi_blacklist *d)
>          keyboard_controller_present = 0;
>          return 0;
>  }
> +#endif
>  
>  /*
>   *	This will be expanded over time to force things like the APM 
> @@ -825,7 +827,8 @@ static __initdata struct dmi_blacklist dmi_blacklist[]={
>  			MATCH(DMI_SYS_VENDOR, "IBM"),
>  			NO_MATCH, NO_MATCH, NO_MATCH
>  			} },
> - /*
> +#ifdef CONFIG_VT
> +        /*
>           *      IBM Bladeservers
>           */
>  
> @@ -834,6 +837,7 @@ static __initdata struct dmi_blacklist dmi_blacklist[]={
>                          MATCH(DMI_BOARD_NAME, "Server Blade"),
>                          NO_MATCH, NO_MATCH
>                          } },
> +#endif
>  
>  #ifdef	CONFIG_ACPI_BOOT
>  	/*


Ok, Thanks Willy! The changes to the dmi_scan.c as you gave above from 
2.4.37rc1 fixed the remaining problem. I have a 2.4.36.7 Cobalt patched 
kernel running in a Raq550 as I type this and it compiled with no 
keyboard or VT support cleanly with your dmi_scan and pc_keybd patch so 
I no longer am using the old regression patch I made previously.

  Thanks for your effort in solving this.


Dave Studeman