See bug #200127 While installing a Xen guest, we were seeing fs/jbd/transaction.c:do_get_write_access() allocate a buffer which spanned a page boundary: frozen_buffer = jbd_kmalloc(jh2bh(jh)->b_size, GFP_NOFS); This later caused problems for xen's blkfront/blkback drivers which can't handle such a buffer For now we've disabled CONFIG_DEBUG_SLAB with Xen, but it looks like the real fix is for jbd to always allocate properly aligned buffers even when CONFIG_DEBUG_SLAB is enabled.
change QA contact
This report targets FC5, which is now end-of-life. Please re-test against Fedora 7 or later, and if the issue persists, open a new bug. Thanks