Hi all, I have a x86 (32bit, Intel Core 2 Quad, 4GB RAM) system running fairly stable using 3.6.11-rt25. Now I switched to the latest RT_PREEMPT patch of the 3.6 series (3.6.11-rt30). However, since that I get frequently kernel messages like the one below. The system is not working stable anymore. ------------[ cut here ]------------ WARNING: at arch/x86/mm/highmem_32.c:47 kmap_atomic_prot+0x76/0xe0() Hardware name: HP xw4600 Workstation Modules linked in: kvm Pid: 1640, comm: blkid Not tainted 3.6.11-rt30 #2 Call Trace: [<c1026cd5>] warn_slowpath_common+0x68/0x7d [<c1023e9c>] ? kmap_atomic_prot+0x76/0xe0 [<c1026cfe>] warn_slowpath_null+0x14/0x18 [<c1023e9c>] kmap_atomic_prot+0x76/0xe0 [<c1023f19>] kmap_atomic+0x13/0x15 [<c10909a8>] get_page_from_freelist+0x3f9/0x49a [<c1091324>] __alloc_pages_nodemask+0x11f/0x5b5 [<c108bda7>] ? unlock_page+0x1d/0x20 [<c10b14be>] ? kmem_cache_alloc+0xda/0x18b [<c10a9fca>] ? anon_vma_prepare+0x54/0xf9 [<c10aa034>] ? anon_vma_prepare+0xbe/0xf9 [<c10a3dfb>] handle_pte_fault+0x135/0x58b [<c1091a34>] ? free_pages+0x22/0x24 [<c10a2c0b>] ? tlb_finish_mmu+0x25/0x36 [<c10a4f7a>] handle_mm_fault+0x91/0xb2 [<c1020a9f>] do_page_fault+0x32c/0x369 [<c10a79f6>] ? do_brk+0x1d4/0x207 [<c105b9f0>] ? rt_up_write+0xd/0xf [<c10a7b62>] ? sys_brk+0xf9/0x102 [<c1020773>] ? vmalloc_sync_all+0xf7/0xf7 [<c13a5bfa>] error_code+0x5a/0x60 [<c1020773>] ? vmalloc_sync_all+0xf7/0xf7 ---[ end trace 0000000000000002 ]--- With the 3.6.11-rt25 I never saw this errors. I have attached the full "dmesg" output and the kernel configuration. I have a pure vanilla kernel + the RT_PREEMPT patch. No other patches in the kernel. Any feedback or proposals how to get around that issue are highly welcome. Thanks a lot. Regards Mathias
Attachment:
dmesg.gz
Description: dmesg.gz
Attachment:
config.gz
Description: config.gz