On 10/12/2007, andrzej zaborowski <[EMAIL PROTECTED]> wrote: > On 02/12/2007, Armin <[EMAIL PROTECTED]> wrote: > > Thiemo, > > > > Thiemo Seufer wrote: > > > Armin wrote: > > > > > >> Hello, > > >> > > >> This includes the previous Mainstone re-org patch I sent earlier plus > > >> flash > > >> support. > > >> This adds two 32MiB flash devices. Mounts from mtdblock2 on flash device > > >> 0 > > >> fine at boot. > > >> > > > > > > I did some guesswork on the flash initialization to make it build with > > > Laurent's -disk patch. Please check if it is still correct. > > > > > > > works fine. > > Note that both chips get mapped at the same offset in phys_ram_base, > I'm quite sure this is a bug and not intentional? It may corrupt data > if the OS reads from both chips. I wanted to convert mainstone.c to > use qemu_ram_alloc like other pxa boards but I want to make sure this > is not intentional. > > The value of mainstone_rom indicates this too.
Sorry, mainstone_rom is not related to this issue, but that means that the ram_size check allows too low values and qemu may crash. It also means that the two flash chips *and* the ROM all overlap. I think gumstix.c registers the flash correctly, you may wnat to look in it. Regards
