On Tue, Apr 1, 2008 at 5:45 PM, Tomas Kalibera <[EMAIL PROTECTED]> wrote: > > The stack trace starts getting printed even before the kernel boots - the > first occurrence is below. It then repeats printing so frequently that the > system is unusable (and I could not run the Xenomai task). > > I've attached the highmem_32.c I used.
Ok. You can get rid of old patches. What happens if you make the printk, stack trace and fix up, conditional to type == KM_PTE0 || type == KM_PTE1 ? As in: if (type == KM_PTE0 || type == KM_PTE1) { printk("Wrong address passed to kunmap_atomic!\n"); show_stack(NULL, NULL); kpte_clear_flush(kmap_pte-idx, __fix_to_virt(FIX_KMAP_BEGIN+idx)); } -- Gilles _______________________________________________ Xenomai-core mailing list Xenomai-core@gna.org https://mail.gna.org/listinfo/xenomai-core