On Mo, 2016-03-14 at 15:22 +0000, Peter Maydell wrote: > On 14 March 2016 at 15:04, Pooja Dhannawat <dhannawatpoo...@gmail.com> wrote: > > Removing support for DEPTH != 32 from blizzard template header > > and file that includes it, as macro DEPTH == 32 only used. > > > > Signed-off-by: Pooja Dhannawat <dhannawatpoo...@gmail.com> > > --- > > hw/display/blizzard.c | 24 ------------------------ > > hw/display/blizzard_template.h | 30 +----------------------------- > > 2 files changed, 1 insertion(+), 53 deletions(-)
> So are there any cases now when > surface_bits_per_pixel(qemu_console_surface(c)) can return > zero? (I don't understand the 'case 0' code in this switch > at all...) Looks like a somewhat strange way to zero-fill the function pointer array. Looks like a historical leftover. As blizzard uses qemu_console_resize() you will always get a surface with the default depth (32bpp). So dropping the whole "switch { ... }" block (except for the "case 32:" lines) and adding an "assert(surface_bits_per_pixel(surface)==32)" should work just fine. cheers, Gerd