Comment 24 for bug 6812

Message-Id: <email address hidden>
Date: Tue, 08 Jun 2004 05:14:15 +0900
From: Kenshi Muto <email address hidden>
To: Herbert Xu <email address hidden>
Cc: <email address hidden>, <email address hidden>
Subject: Re: Bug#251861: kernel: vga16fb is broken (at least on vmware)

Hi Herbert,

At Mon, 7 Jun 2004 20:11:13 +1000,
Herbert Xu wrote:
> > Ah..., I've already tried this patch. Unfortunately, this patch
> > doesn't solve a problem at least on my test environment. Herbert, did
> > you succeed by BK patch?
>
> Yes my vmware setup is broken before the patch and it works afterwards.

Ah, sorry I forgot some explainations.

- BK patch fixes screen lock up when I load fbcon.
- But, when I launch framebuffer application (such as console),
  screen still locks up.

Could you install jfbterm and unifont package, and try to run
'jfbterm -q other,EUC-JP,iconv,UTF-8'?

> > On my observation, problem is not from patched point, but is from:
> > --------------------
> > vga16fb.fix.smem_start = VGA_MAP_MEM(vga16fb.fix.smem_start);
> > --------------------
> >
> > When I remove this from code, vga16fb works. I heard this line is for
> > ARM architecture...
>
> Hmm, I can't see how this can make a difference. VGA_MAP_MEM should
> be a noop for the values in question. Can you please do a printk
> before/after that line to see if that's not the case?

I tried to add printk in vga16fb.c l.1375.

+ printk(KERN_INFO "vga16fb: DEBUG: smem_start %lx\n", vga16fb.fix.smem_start);
          vga16fb.fix.smem_start = VGA_MAP_MEM(vga16fb.fix.smem_start);
+ printk(KERN_INFO "vga16fb: DEBUG: smem_start %lx\n", vga16fb.fix.smem_start);

Result:

vga16fb: initializing
vga16fb: mapped to 0xc00a0000
vga16fb: DEBUG: smem_start a0000
vga16fb: DEBUG: smem_start c00a0000
fb0: VGA16 VGA frame buffer device
Console: switching to colour frame buffer device 80x30

Thanks,
--
Kenshi Muto
<email address hidden>