[PATCH] OMAPDSS: OMAPFB: check for matching memory size early

[PATCH] OMAPDSS: OMAPFB: check for matching memory size early

[Grazvydas Ignotas]

If the size of memory region that is being set up is the same as before,
we don't have to do memory and layer busy checks.

Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
---
I'm not sure if this is semantically correct, but this helps to get
some old programs working, that rely on old DSS version which didn't
check for mmap lock, but incidentally they are using same size too..

 drivers/video/omap2/omapfb/omapfb-ioctl.c |   13 +++++++------
 1 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/video/omap2/omapfb/omapfb-ioctl.c b/drivers/video/omap2/omapfb/omapfb-ioctl.c
index df7bcce..a0ac95c 100644
--- a/drivers/video/omap2/omapfb/omapfb-ioctl.c
+++ b/drivers/video/omap2/omapfb/omapfb-ioctl.c
@@ -225,6 +225,9 @@ static int omapfb_setup_mem(struct fb_info *fbi, struct omapfb_mem_info *mi)
 	down_write_nested(&rg->lock, rg->id);
 	atomic_inc(&rg->lock_count);
 
+	if (rg->size = size && rg->type = mi->type)
+		goto out;
+
 	if (atomic_read(&rg->map_count)) {
 		r = -EBUSY;
 		goto out;
@@ -245,12 +248,10 @@ static int omapfb_setup_mem(struct fb_info *fbi, struct omapfb_mem_info *mi)
 		}
 	}
 
-	if (rg->size != size || rg->type != mi->type) {
-		r = omapfb_realloc_fbmem(fbi, size, mi->type);
-		if (r) {
-			dev_err(fbdev->dev, "realloc fbmem failed\n");
-			goto out;
-		}
+	r = omapfb_realloc_fbmem(fbi, size, mi->type);
+	if (r) {
+		dev_err(fbdev->dev, "realloc fbmem failed\n");
+		goto out;
 	}
 
  out:
-- 
1.7.0.4

Re: [PATCH] OMAPDSS: OMAPFB: check for matching memory size early

[Tomi Valkeinen]

[-- Attachment #1: Type: text/plain, Size: 703 bytes --]

Hi,

On Fri, 2012-03-09 at 02:42 +0200, Grazvydas Ignotas wrote:
> If the size of memory region that is being set up is the same as before,
> we don't have to do memory and layer busy checks.
> 
> Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
> ---
> I'm not sure if this is semantically correct, but this helps to get
> some old programs working, that rely on old DSS version which didn't
> check for mmap lock, but incidentally they are using same size too..

Well, I wouldn't allow code like that if it was a new ioctl, but if, as
you say, the old omapfb worked that way, then I think it's better to do
things the same way.

I can see no harm in it, so I'll apply.

 Tomi


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]