Description of problem: The following kernel.oops shows up with regularity when booting an XO with Snap1. NET: Registered protocol family 1 checking if image is initramfs... it is ------------[ cut here ]------------ WARNING: at arch/x86/mm/pageattr.c:782 change_page_attr_set_clr+0x5e/0x198() (Not tainted) Modules linked in: Pid: 1, comm: swapper Not tainted 2.6.27-1.fc10.i686 #1 [<c06d8fc3>] ? printk+0x14/0x19 [<c042d224>] warn_on_slowpath+0x4c/0x7a [<c04059cb>] ? mcount_call+0x5/0xa [<c04bcd8a>] ? inotify_dentry_parent_queue_event+0xe/0x88 [<c04059cb>] ? mcount_call+0x5/0xa [<c047657d>] ? __free_one_page+0x17f/0x1c4 [<c041c9c1>] change_page_attr_set_clr+0x5e/0x198 [<c046cc28>] ? ftrace_record_ip+0x1ac/0x1c5 [<c046cc28>] ? ftrace_record_ip+0x1ac/0x1c5 [<c041cbae>] set_memory_rw+0x16/0x18 [<c041b7d5>] free_init_pages+0x26/0x96 [<c041b85b>] free_initrd_mem+0x16/0x18 [<c08507c0>] free_initrd+0x8a/0xa4 [<c085085a>] populate_rootfs+0x80/0xd2 [<c040112a>] _stext+0x42/0x11a [<c08507da>] ? populate_rootfs+0x0/0xd2 [<c0467b45>] ? register_irq_proc+0x97/0xb3 [<c04059cb>] ? mcount_call+0x5/0xa [<c0467abc>] ? register_irq_proc+0xe/0xb3 [<c0467ba3>] ? init_irq_proc+0x42/0x4f [<c084b3a8>] kernel_init+0x17c/0x1ca [<c084b22c>] ? kernel_init+0x0/0x1ca [<c04059b3>] kernel_thread_helper+0x7/0x10 ======================= ---[ end trace a7919e7f17c0a725 ]--- Freeing initrd memory: 10033k freed
arch/x86/mm/pageattr.c: /* Ensure we are PAGE_SIZE aligned */ if (addr & ~PAGE_MASK) { addr &= PAGE_MASK; /* * People should not be passing in unaligned addresses: */ WARN_ON_ONCE(1); }
Assuming this is on freeing the ramdisk (whichit looks like), it looks like the ramdisk address is page aligned to me.
*** Bug 470182 has been marked as a duplicate of this bug. ***
*** Bug 470347 has been marked as a duplicate of this bug. ***
Aha, no it wasn't. I've updated http://katzj.fedorapeople.org/olpc/livecd-iso-to-disk.sh -- if you try it now, can you confirm that you don't see this warning anymore? It seems to do the trick for me.
Can somebody please re-test this so that we can confirm the fix?
After several reboots ... not seeing this as of a livecd-creator built image from today.
James, that was using jeremy's livecd-iso-to-disk right? If so, this needs to be re-opened until a package with that change is tagged.
yeah, I grabbed a livecd-iso-to-disk.sh this morning and built a new Live image this morning. # md5sum /tmp/livecd-iso-to-disk.sh 6dc19fe6ed50ea70b7568ba193cdff5a /tmp/livecd-iso-to-disk.sh Apologies ... it slipped my mind that the script (http://katzj.fedorapeople.org/olpc/livecd-iso-to-disk.sh) wasn't yet added to livecd-tools. Should we re-assign this to livecd-tools?
Thanks for the verification jlaska. Pushed changes to the livecd-tools repo and I'm intending to get a release and build of it done tonight (changes are almost entirely livecd-iso-to-disk changes)
And livecd-tools built