drm/radeon: deal with errors from framebuffer init path.

We've been getting occasional oops running a 32-bit kernel on a certain
system in our RHEL test hw. It appears that we fail to get sufficent ioremap
space for the framebuffer, and this leads to an oops.

This patch should fix the oops and leave a message in the logs we can
check for.

A future fix would probably to resize the console to a size that we can
ioremap.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

1 files changed, 10 insertions, 1 deletions

diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index cf2bf35b56b8..195471cf65d3 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -209,6 +209,11 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
                                                           sizes->surface_depth);
 
         ret = radeonfb_create_pinned_object(rfbdev, &mode_cmd, &gobj);
+        if (ret) {
+                DRM_ERROR("failed to create fbcon object %d\n", ret);
+                return ret;
+        }
+
         rbo = gem_to_radeon_bo(gobj);
 
         /* okay we have an object now allocate the framebuffer */
@@ -220,7 +225,11 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 
         info->par = rfbdev;
 
-        radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj);
+        ret = radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj);
+        if (ret) {
+                DRM_ERROR("failed to initalise framebuffer %d\n", ret);
+                goto out_unref;
+        }
 
         fb = &rfbdev->rfb.base;