Re: [git pull] drm: previous pull req + 1.

From: Andrew Lutomirski <luto@mit.edu>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: "Thomas Hellström" <thomas@shipmail.org>,
	"Dave Airlie" <airlied@linux.ie>,
	"Alex Deucher" <alexdeucher@gmail.com>,
	dri-devel@lists.sf.net, "Jerome Glisse" <jglisse@redhat.com>,
	"Linux Kernel Mailing List" <linux-kernel@vger.kernel.org>
Subject: Re: [git pull] drm: previous pull req + 1.
Date: Mon, 22 Jun 2009 14:59:07 -0400	[thread overview]
Message-ID: <cb0375e10906221159u56fcf275yff34b93a39e07186@mail.gmail.com> (raw)
In-Reply-To: <alpine.LFD.2.01.0906221121100.3240@localhost.localdomain>

2009/6/22 Linus Torvalds <torvalds@linux-foundation.org>:
>
>
> On Mon, 22 Jun 2009, Thomas Hellström wrote:
>>
>> It would be very helpful if we could introduce an fbdev mutex that protects
>> fbdev accesses to the kernel map and to the fbdev acceleration functions.
>
> Not going to happen.
>
> Why? 'printk'.
>
> If you can't handle printk, then you're basically useless. And printk
> absolutely -has- to work in bad situations (the most important messages
> could happen in any context).

What if we only guaranteed that the framebuffer is mapped when it's
showing on the screen?

printk doesn't need to write to the framebuffer immediately when X
isn't running (since the framebuffer isn't shown) and presumably the
framebuffer needs to be pinned somewhere when it's being displayed
anyway.  This would involve fbcon knowing how to buffer text to be
shown later so that printk still works in interrupt context.

--Andy