Nouveau - Jan 2012 - [PATCH] drm/nouveau/i2c: Fix i2c initialization on cards with DCB older than 1.5

Signed-off-by: Viktor Novotny <noviktor at seznam.cz>
---
Hi,
I think I found a bug - my NV11 fails to init i2c ports and is then unable to
read EDID.
It's because while i2c_table in nouveau_i2c.c:324 returns null for this
card,
i2c is initialized anyway at line 341 and wrong i2c table format is then
selected at line 350.
This patch fixes the problem on NV11 and was tested also on NV25. I hope no
cards but those
with legacy i2c table have i2c table version set to 0.

Regards
	Viktor

 drivers/gpu/drm/nouveau/nouveau_i2c.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.c
b/drivers/gpu/drm/nouveau/nouveau_i2c.c
index 820ae7f..57bcbc3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_i2c.c
+++ b/drivers/gpu/drm/nouveau/nouveau_i2c.c
@@ -344,14 +344,15 @@ nouveau_i2c_init(struct drm_device *dev)
 		if (i2c[5]) legacy[0][1] = i2c[5];
 		if (i2c[6]) legacy[1][0] = i2c[6];
 		if (i2c[7]) legacy[1][1] = i2c[7];
+		version = 0;
 	}
 
-	if (i2c && version >= 0x30) {
+	if (version >= 0x30) {
 		entry     = i2c[1] + i2c;
 		entries   = i2c[2];
 		recordlen = i2c[3];
 	} else
-	if (i2c) {
+	if (version > 0) {
 		entry     = i2c;
 		entries   = 16;
 		recordlen = 4;
-- 
1.7.8.3

Hi,
did anybody find time to look into this issue? By the way, I think it was
introduced by commit 486a45c2.

Regards
	Viktor

On 18.1.2012 14:28, Viktor Novotn? wrote:
> Signed-off-by: Viktor Novotny<noviktor at seznam.cz>
> ---
> Hi,
> I think I found a bug - my NV11 fails to init i2c ports and is then unable
to read EDID.
> It's because while i2c_table in nouveau_i2c.c:324 returns null for this
card,
> i2c is initialized anyway at line 341 and wrong i2c table format is then
selected at line 350.
> This patch fixes the problem on NV11 and was tested also on NV25. I hope no
cards but those
> with legacy i2c table have i2c table version set to 0.
>