On Tue, Oct 30, 2018 at 5:08 AM <peng.hao2@xxxxxxxxxx> wrote: > >On Mon, Oct 29, 2018 at 12:54 PM <peng.hao2@xxxxxxxxxx> wrote: > >I would consider it as a bug if on some architectures iounmap() is not > >able to take a base got from ioport_map(). > I added TYPE_PVPANIC_MMIO in QEMU to the architecture that does not support IOPORT. > If some architectures does not support IOPORT, set TYPE_PVPANIC_MMIO in qemu. This setting is reflected in the kernel pvpanic driver > and does not invoke ioport_map (calling ioremap). So I think the situation you described above will not happen. > But I should resume one check like this to deal with the failure of ioremap/ioport_map. > if (base) > iounmap(base) I'm a bit confused, how come that base == NULL when you about to call iounmap()? Where you would like to put this check? > The patch of QEMU is still in review. -- With Best Regards, Andy Shevchenko
