Ship It!

Ship It!

usr/src/uts/i86pc/io/gfx_private/gfxp_bitmap.c (Diff revision 1)

Show all issues

How did you pick EFAULT for these? It doesn't feel like the only operation that the setmode entry points could ever run is copying in and out memory from userland since we've taken away all the arguments here that could have come from userland.

Since these all return an integer, any reason we shouldn't propagate that actual error?

Note, this applies to all of the setmode() calls.

1. 

   Toomas Soome April 5, 2019, 2:33 p.m.

   The problem is that setmode() does return boolean. true or false and error codes are eaten. So I was in trouble - or still am. I am not sure what would be appropriate error code there.

2. 

   Robert Mustacchi April 16, 2019, 2:06 p.m.

   Maybe we should use EIO, as it's a generic something what wrong with the device request error code and at least would point someone to saying, hmm, maybe there's an issue between the driver and hardware, which is what would be the first place to look if we had a failure here.

Ship It!