Hi Bartlomiej, On Wed, Jun 14, 2017 at 12:54:58PM +0200, Bartlomiej Zolnierkiewicz wrote: > > Hi, > > On Monday, May 08, 2017 08:11:16 AM Christophe JAILLET wrote: > > According to surrounding goto, it is likely that 'goto err_probed_panel' is > > expected here. > > This change is just done in order to silence some coccinelle scripts > > which try to detect wrongly ordered goto. > > > > If 'info->fb[HEAD_PANEL]' and' 'info->fb[HEAD_CRT]' are both NULL, this > > means that no 'framebuffer_alloc' has been performed, so 'goto err_alloc' > > looks fine. > > Anyway, it is also harmless in this case to call 'framebuffer_release'. > > The code looks more straight forward. > > > > Signed-off-by: Christophe JAILLET <christophe.jail...@wanadoo.fr> > > It seems that the code for supporting only selected framebuffers > (only HEAD_PANEL fb or only HEAD_CRT fb) is broken anyway as at least > the suspend/resume support assumes that both framebuffers are always > present. Also all sm501fb driver users always try try to initialize > both framebuffers. Therefore I would prefer the removal of non-working > individual framebuffers support (the code that your patch modifies > would be removed as well). Could you please look into it?
Looks like it will be easier to fix the suspend/resume functions and the part of probe which creates the device files for both. Let me have a look and I will try to send you a patch over the weekend. -- Regards Sudip
