Re: cirrusdrmfb broken with simplefb

* David Herrmann [off-list ref] wrote:

Hi

On Thu, Dec 19, 2013 at 1:31 PM, Ingo Molnar [off-list ref] wrote:

quoted

* David Herrmann [off-list ref] wrote:

quoted

Hi

On Thu, Dec 19, 2013 at 1:03 AM, One Thousand Gnomes
[off-list ref] wrote:

quoted

quoted

quoted

That bug always existed, simplefb is just the first driver to hit it
(vesafb/efifb didn't use resources). I'm aware of the issue but as a
workaround you can simply disable CONFIG_X86_SYSFB. That restores
the old behavior.

This looks like a regression, so we'll either need a fix or we'll have
to mark CONFIG_X86_SYSFB as CONFIG_BROKEN.

Kernel bugzilla has entries for simplefb breaking both vesafb and
matrox mga.

Thanks for the hints. I've read through all I could find and tried
to provide some help.

I'm kind of confused, most of them enable CONFIG_X86_SYSFB (which is
'n' by default) but don't read the help text. I did my best to tell
people that this option requires CONFIG_FB_SIMPLE, but if you don't
read the help-text you won't notice that. Don't know what to do
about that..

People generally don't read the help text - still the kernel should
not break. So please the Kconfig angle (and the bootup logic, etc.)
fool-proof, graphics failures are not fun to debug!

There're dozens of combinations that break gfx-boot. [...]

Then that's a bug too.

[...] Even non-obvious things like disabling VT_HW_CONSOLE_BINDING 
will break gfx-handover. Anyhow, bad examples are no excuse.. I will 
send patches to fool-proof sysfb.

Cool, thanks!

Arguably there's a lot of broken gfx legacy in Linux.

<SoapBox>: 15 years ago we should have merged GGI that really got the 
integrated gfx principles right from the get go IMHO - and we've been 
struggling with fragmented, disjunct gfx concepts since then, but IMHO 
it's getting better gradually, so please don't give up! ;-)

	Ingo