[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: framebuffer broken on 712 using kernel >> 2.6.4



On Thu, 9 Sep 2004, Stuart Brady wrote:

> On Thu, Aug 19, 2004 at 09:52:32AM +0300, Martin-Éric Racine wrote:
> [Snip]
> > 
> > 2) all 2.6.x since 2.6.5
> > 
> > The PA-RISC Tux appears.  Everything within the lines it occupies is
> > a clean pitch black.  Bootup messages start on the last line of the
> > penguin, overlaping its legs.  Messages up to "Gecko-style soft power
> > switch enabled." scroll on below.  Messages following that appear on
> > the feet of the penguin again and are all printed there, without
> > scrolling down.
> > 
> > Once the "INIT: version 2.86" message appears at that location, things
> > magically resume a few lines below the powerswitch kernel message,
> > starting with the "Setting disc parameters: done." message and scroll
> > down, eventually making the penguin scroll off-screen.
> 
> Try it with:
> 
> STI_CONSOLE=y              (STI text console)
> DUMMY_CONSOLE_COLUMNS=<c>  (Initial number of console screen columns)
> DUMMY_CONSOLE_ROWS=<r>     (Initial number of console screen rows)
> 
> Where <c> is the screen width (in pixels) divided by 8 (E.g. 1024/8=128)
> and <r> is the height divided by 16 (E.g. 768/16=48).  (These options are
> under "Device Drivers" -> "Graphics support" -> "Console display driver
> support".)
> 
> I don't know whether you actually need sticon itself, but you need to
> choose in order to set the columns/rows.  Perhaps that's a bug.

Using 128 and 48 solved it.  This being said, it wasn't needed previously.  I
have no choice but to call that a regression, because it forces the kernel to be
built for a specific framebuffer size; there is no way to have a generic kernel
that will work for everyone, e.g. with the new Debian Installer, for instance.

-- 
Martin-Éric Racine, ICT Consultant
http://www.iki.fi/q-funk/



Reply to: