On Fri, Jan 14, 2011 at 11:16 PM, Kirill A. Shutemov <kas@xxxxxxxxxx> wrote: > This patch partly reverts 2c6755988a. > > spin_[un]lock_bucket() uses bit_spin_[un]lock() to set/clear first > bit of ->head.first. bit_spin_[un]lock() touches the bit on UP if > DEBUG_SPINLOCK enabled. > > So we have to define LIST_BL_LOCKMASK to 1, if DEBUG_SPINLOCK enabled. > Other way is reimplementation of spin_[un]lock_bucket(). > > [ 0.140008] BUG: unable to handle kernel NULL pointer dereference at 00000005 > [ 0.140008] IP: [<c10a7cdf>] __d_rehash+0x6f/0xc0 > [ 0.140008] *pde = 00000000 > [ 0.140008] Oops: 0002 [#1] > [ 0.140008] last sysfs file: > [ 0.140008] Modules linked in: > [ 0.140008] > [ 0.140008] Pid: 1, comm: swapper Not tainted 2.6.37+ #150 /Bochs > [ 0.140008] EIP: 0060:[<c10a7cdf>] EFLAGS: 00010202 CPU: 0 > [ 0.140008] EIP is at __d_rehash+0x6f/0xc0 > [ 0.140008] EAX: 00000022 EBX: c7402460 ECX: 00000000 EDX: c7402468 > [ 0.140008] ESI: c7ee157c EDI: 00000001 EBP: c782bedc ESP: c782bec4 > [ 0.140008] DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 > [ 0.140008] Process swapper (pid: 1, ti=c782a000 task=c7830000 task.ti=c782a000) > [ 0.140008] Stack: > [ 0.140008] c13b698f 00000001 c7402468 02eef038 c7402460 c7402460 c782bee8 c10a7d69 > [ 0.140008] c74024b0 c782bef8 c10a7d90 c7402460 c7400430 c782bf04 c10b32f4 c1304380 > [ 0.140008] c782bf1c c109dcd2 00000000 00000000 c74023c0 c7400430 c782bf3c c109df85 > [ 0.140008] Call Trace: > [ 0.140008] [<c10a7d69>] _d_rehash+0x39/0x40 > [ 0.140008] [<c10a7d90>] d_rehash+0x20/0x40 > [ 0.140008] [<c10b32f4>] simple_lookup+0x34/0x40 > [ 0.140008] [<c109dcd2>] d_alloc_and_lookup+0x32/0x60 > [ 0.140008] [<c109df85>] __lookup_hash+0xc5/0x160 > [ 0.140008] [<c109f2ae>] lookup_one_len+0x8e/0xe0 > [ 0.140008] [<c111e997>] debugfs_create_file+0x77/0x200 > [ 0.140008] [<c11bc7c9>] ? __class_create+0x29/0x80 > [ 0.140008] [<c143f6ec>] ? kernel_init+0x0/0x120 > [ 0.140008] [<c111ebb1>] debugfs_create_dir+0x21/0x30 > [ 0.140008] [<c144c256>] bdi_class_init+0x33/0x3c > [ 0.140008] [<c1001100>] do_one_initcall+0x30/0x160 > [ 0.140008] [<c144c223>] ? bdi_class_init+0x0/0x3c > [ 0.140008] [<c143f6ec>] ? kernel_init+0x0/0x120 > [ 0.140008] [<c143f778>] kernel_init+0x8c/0x120 > [ 0.140008] [<c1003096>] kernel_thread_helper+0x6/0x10 > [ 0.140008] Code: 9e d0 3d c1 e8 19 14 25 00 8b 3e 8d 53 08 89 55 f0 89 7c 24 04 c7 04 24 8f 69 3b c1 e8 01 14 25 00 85 ff 8b 55 f0 89 7b 08 74 03 <89> 57 04 89 73 0c 89 16 c7 04 24 6c dc 3d c1 e8 e3 13 25 00 8b > [ 0.140008] EIP: [<c10a7cdf>] __d_rehash+0x6f/0xc0 SS:ESP 0068:c782bec4 > [ 0.140008] CR2: 0000000000000005 > [ 0.144008] ---[ end trace 4eaa2a86a8e2da22 ]--- > > Signed-off-by: Kirill A. Shutemov <kas@xxxxxxxxxx> > --- > include/linux/list_bl.h | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/include/linux/list_bl.h b/include/linux/list_bl.h > index b2adbb4..5bad17d 100644 > --- a/include/linux/list_bl.h > +++ b/include/linux/list_bl.h > @@ -16,7 +16,7 @@ > * some fast and compact auxiliary data. > */ > > -#if defined(CONFIG_SMP) > +#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK) > #define LIST_BL_LOCKMASK 1UL > #else > #define LIST_BL_LOCKMASK 0UL Thank you. Russell reported this too. I need a brown paper bag... -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
