On 01/07/2012 01:13 PM, Stefan Weil wrote: > There is a regression since commit c5705a7728b4a6bc9e4f2d35911adbaf28042b25 > or some other recent change. > > System emulation with a flash device raises an assertion in function > qemu_ram_set_idstr because no new_block is found at the requested addr. > > The address of the memory region (mr->ram_addr) is set by > memory_region_init_rom_device: > > mr->ram_addr = qemu_ram_alloc(size, mr); > mr->ram_addr |= cpu_register_io_memory(mr); > > The 2nd line adds a small offset to mr->ram_addr, and this offset > makes ram_set_idstr fail with an assertion. > > Masking mr->ram_addr with the target's page mask removes the small > offset and makes flash devices work again. > > > diff --git a/memory.c b/memory.c > index 394cbab..1b2f9e8 100644 > --- a/memory.c > +++ b/memory.c > @@ -1386,7 +1386,7 @@ void memory_region_set_alias_offset(MemoryRegion *mr, > target_phys_addr_t offset) > > ram_addr_t memory_region_get_ram_addr(MemoryRegion *mr) > { > - return mr->ram_addr; > + return mr->ram_addr & TARGET_PAGE_MASK; > } >
This makes get_ram_addr() lie about the return value - it's no longer the ram_addr. Doesn't matter much since it's temporary, but I patched vmstate_register_ram() instead. -- error compiling committee.c: too many arguments to function