Claus Guttesen wrote: >>You need to increase VM_KMEM_SIZE_MAX in your kernel >>config, e.g. >>options VM_KMEM_SIZE_MAX=536870912 > > > Did that, but it panicked, hand-transcribed screen: > > vm_fault on nofault entry, addr : ffffffffb091d000 > KDB stack backtrace > panic() at 0xffffffff8020bf17 = panic+0x2a7 > vm_fault() at 0xffffffff802f57b0 = vm_fault+0x1ce0 > trap_pfault() at 0xffffffff80330350 = > trap_pfault+0x2a0 > alltraps_with_regs_pushed() at 0xffffffff8031c3fb = > alltraps_with_regs_pushed+0x5 > amr_mapcmd() at 0xffffffff80184f34 = amr_mapcmd+0x114 > amr_start() at 0xffffffff801850e0 = amr_start+0x170 > amr_startio() at 0xffffffff801857ce = amr_startio+0x3e > amr_submit_bio() at 0xffffffff80185bb0 = > amr_submit_bio+0x20 > g_disk_start() at 0xffffffff801ed02a = > g_disk_start+0xea > g_io_schedule_down() at 0xffffffff801cf4ca = > g_io_schedule_down+0xea > g_down_procbody() at 0xffffffff801cfa28 = > g_down_procbody+0x28 > fork_exit() at 0xffffffff801f17af = fork_exit+0x8f > fork_trampoline() at 0xffffffff8031c5fe = > fork_trampoline+0xe > --- trap 0, rip = 0, rsp = 0xffffffffba13cd00, rbp = 0 > > Claus > The AMR driver cannot handle >4GB of RAM. Even though you only have 4GB, your chipset is remapping part of it to the >4GB region. The driver incorrectly interfaces with busdma and cannot handle this scenario very well. The panic is to be expected. You can verify this by setting the tunable 'hw.physmem' in the loader to some value under 4GB to artifically limit the amount of RAM that the OS sees. ScottReceived on Thu Sep 30 2004 - 06:32:00 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:14 UTC