On 2012-10-11 18:58, Enric Balletbo Serra wrote: > Hi all, > > I see that commit dac8eb5f (OMAPDSS: TFP410: rename dvi files to > tfp410) and commit 2e6f2ee7 (OMAPDSS: TFP410: rename dvi -> tfp410) > changes the panel driver used on some boards. I tested current > linux-next-20101011 kernel with my IGEPv2 board and DSS fails with > following error (see http://pastebin.com/VjPGCQDt for full log) : > > [ 21.222808] omapdss OVERLAY error: check_overlay: paddr cannot be 0 > [ 21.229583] omapdss OVERLAY error: check_overlay: paddr cannot be 0 > [ 21.236236] omapfb omapfb: setup_plane failed > > Before checking what happens there is a known issue with this ? Works fine for me, although only with a quite minimal test environment. Looking at the log, it looks to me that your userspace is trying to configure the overlays, and that somehow leads to the driver getting zero as the fb address. I'm guessing that some wrong value is passed when configuring, or missing some step in configuring. However, It should be the omapfb that rejects configuration that would lead to paddr == 0, not omapdss, so there's definitely bug in the error handling of omapfb. Can you tell me more what's going on there? I presume the error happens when booting? What userspace component is using omapfb in your setup? X? If you can, put some debug prints into omapfb-ioctl.c:omapfb_setup_plane to see what code path it takes, and where it goes wrong. Tomi
Attachment:
signature.asc
Description: OpenPGP digital signature
