Message ID | 1469201948-20700-1-git-send-email-stefan.bader@canonical.com |
---|---|
State | New |
Headers | show |
On 22/07/16 16:39, Stefan Bader wrote: > Otherwise console can be rather slow on higher resolutions. We > had this in Trusty-Vivid but at some point after that the source > file moved into a different directory. Which likely caused it to > get lost again. > > We probably need to be careful to make sure that any Yakkety kernel > post 4.4 still adds it, too. > > -Stefan > > > From 24924c740ea3c5c15a27f391171fad2ca93582a2 Mon Sep 17 00:00:00 2001 > From: Stefan Bader <stefan.bader@canonical.com> > Date: Tue, 21 Apr 2015 14:54:52 +0200 > Subject: [PATCH] UBUNTU: SAUCE: vesafb: Set mtrr:3 (write-combining) as > default > > We had this setting from Oneiric to Saucy[1]. In Saucy it was part > of the combined patch to make vesafb a module, which was unnessesary > in Trusty as we moved back to built-in. > Was added back in Trusty but again lost after Vivid. > > It will still be possible to disable mtrr by putting > video=vesafb:nomtrr or video=vesafb:mtrr:<x> > on the kernel command-line but without that knowledge the default > will be noticably slower (especially on higher resolutions). > > [1] https://lists.ubuntu.com/archives/kernel-team/2011-May/015600.html > > BugLink: http://bugs.launchpad.net/bugs/1605662 > > (forward-ported from trusty by adapting file location) > Signed-off-by: Stefan Bader <stefan.bader@canonical.com> > --- > drivers/video/vesafb.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/video/fbdev/vesafb.c b/drivers/video/fbdev/vesafb.c > index 1c7da3b..d479804 100644 > --- a/drivers/video/fbdev/vesafb.c > +++ b/drivers/video/fbdev/vesafb.c > @@ -47,7 +47,7 @@ static struct fb_fix_screeninfo vesafb_fix = { > }; > > static int inverse __read_mostly; > -static int mtrr __read_mostly; /* disable mtrr */ > +static int mtrr __read_mostly = 3; /* enable mtrr write-comb. */ > static int vram_remap; /* Set amount of memory to be used */ > static int vram_total; /* Set total amount of memory */ > static int pmi_setpal __read_mostly = 1; /* pmi for palette changes ??? */ > Makes sense. Acked-by: Colin Ian King <colin.king@canonical.com>
Is this upstreamable ?
On 22.07.2016 18:05, Tim Gardner wrote: > Is this upstreamable ? > dunno. long term I had the feeling they wanted to get rid of mtrr completely (to be replace pat). So I guess interest might be rather low... -Stefan
On 22/07/16 17:11, Stefan Bader wrote: > On 22.07.2016 18:05, Tim Gardner wrote: >> Is this upstreamable ? >> > dunno. long term I had the feeling they wanted to get rid of mtrr completely (to > be replace pat). So I guess interest might be rather low... Always worth doing IMHO Colin
Applied to X. -Kamal
diff --git a/drivers/video/fbdev/vesafb.c b/drivers/video/fbdev/vesafb.c index 1c7da3b..d479804 100644 --- a/drivers/video/fbdev/vesafb.c +++ b/drivers/video/fbdev/vesafb.c @@ -47,7 +47,7 @@ static struct fb_fix_screeninfo vesafb_fix = { }; static int inverse __read_mostly; -static int mtrr __read_mostly; /* disable mtrr */ +static int mtrr __read_mostly = 3; /* enable mtrr write-comb. */ static int vram_remap; /* Set amount of memory to be used */ static int vram_total; /* Set total amount of memory */ static int pmi_setpal __read_mostly = 1; /* pmi for palette changes ??? */