On Mon, Dec 03, 2007 at 04:00:05PM -0500, Ed L. Cashin wrote:
...
> I'll keep looking at this, but at a glance it looks like the cpu
> number is valid, because I don't trip a BUG_ON when I make the change
> below (the badval variable is noise, sorry).
>
> --- lx/lib/percpu_counter.c.20071130 2007-12-03 15:43:19.000000000 -0500
> +++ lx/lib/percpu_counter.c 2007-12-03 15:47:38.000000000 -0500
> @@ -33,7 +33,9 @@ void __percpu_counter_add(struct percpu_
> s64 count;
> s32 *pcount;
> int cpu = get_cpu();
> + u64 badval = 0xffffffffffffffffULL;
>
> + BUG_ON(!cpu_possible(cpu));
> pcount = per_cpu_ptr(fbc->counters, cpu);
> count = *pcount + amount;
> if (count >= batch || count <= -batch) {
It appears that the fbc->counters pointer is NULL. I added the line,
BUG_ON(!fbc->counters);
... (on line 39 in my percpu_counter.c), and it results in the trace
below. It looks like when it's NULL, percpu_ptr passes it to
__percpu_disguise, which makes it all ones and then tries to
dereference 0xffffffffffffffff to access to the "ptrs" member of the
struct percpu_data.
------------[ cut here ]------------
kernel BUG at lib/percpu_counter.c:39!
invalid opcode: 0000 [1] SMP
CPU 0
Modules linked in: aoe
Pid: 3354, comm: bash Not tainted 2.6.24-rc3-47dbg #10
RIP: 0010:[<ffffffff8036d5f7>] [<ffffffff8036d5f7>]
__percpu_counter_add+0x2a/0x8f
RSP: 0018:ffff810075031aa8 EFLAGS: 00010046
RAX: 00000000ffffffff RBX: ffff81007fd19bd8 RCX: 0000000000000000
RDX: 0000000000000010 RSI: 0000000000000001 RDI: 0000000000000000
RBP: ffff810075031ac8 R08: ffff81007cc077b0 R09: ffffffff802ae5ee
R10: ffff810075031aa8 R11: ffff8100750318e8 R12: ffff81007c81c380
R13: ffff810073ce8250 R14: 0000000000000200 R15: ffff8100755016b0
FS: 00002b3e5c052db0(0000) GS:ffffffff8078b000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00002b7f44fb64e0 CR3: 000000007c4b1000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process bash (pid: 3354, threadinfo ffff810075030000, task ffff81007b4da040)
Stack: ffff810075031ac8 ffff81007fd19bd8 ffff81007c81c380 0000000000000000
ffff810075031af8 ffffffff802ae682 0000100075031ae8 ffff8100755016b0
0000000000000200 ffff81007fd19bd8 ffff810075031b18 ffffffff802ae75c
Call Trace:
[<ffffffff802ae682>] __set_page_dirty+0xdc/0x121
[<ffffffff802ae75c>] mark_buffer_dirty+0x95/0x99
[<ffffffff802ae7d2>] __block_commit_write+0x72/0xac
[<ffffffff802ae988>] block_write_end+0x4f/0x5b
[<ffffffff802b243f>] blkdev_write_end+0x1b/0x38
[<ffffffff80265d96>] generic_file_buffered_write+0x1c0/0x648
[<ffffffff8023a752>] current_fs_time+0x22/0x29
[<ffffffff80266576>] __generic_file_aio_write_nolock+0x358/0x3c2
[<ffffffff80266c84>] filemap_fault+0x1c4/0x320
[<ffffffff80264cce>] unlock_page+0x2d/0x31
[<ffffffff802666dd>] generic_file_aio_write_nolock+0x3b/0x8d
[<ffffffff8028e40f>] do_sync_write+0xe2/0x126
[<ffffffff802497d0>] autoremove_wake_function+0x0/0x38
[<ffffffff8058e725>] do_page_fault+0x3f8/0x7bb
[<ffffffff8028cae8>] fd_install+0x5f/0x68
[<ffffffff8028eb98>] vfs_write+0xae/0x137
[<ffffffff8028f102>] sys_write+0x47/0x70
[<ffffffff8020b7ae>] system_call+0x7e/0x83
Code: 0f 0b eb fe 0f a3 3d 7e 08 4f 00 19 c0 85 c0 75 04 0f 0b eb
RIP [<ffffffff8036d5f7>] __percpu_counter_add+0x2a/0x8f
RSP <ffff810075031aa8>
BUG: sleeping function called from invalid context at kernel/rwsem.c:20
in_atomic():0, irqs_disabled():1
INFO: lockdep is turned off.
Call Trace:
[<ffffffff802518cb>] debug_show_held_locks+0x1b/0x24
[<ffffffff8022f352>] __might_sleep+0xc7/0xc9
[<ffffffff8024c4d0>] down_read+0x1d/0x4a
[<ffffffff80237bd8>] exit_mm+0x34/0xf7
[<ffffffff8023933e>] do_exit+0x247/0x75b
[<ffffffff8020d01e>] kernel_math_error+0x0/0x7e
[<ffffffff8058d35b>] do_trap+0x101/0x110
[<ffffffff8020d4a6>] do_invalid_op+0x91/0x9a
[<ffffffff8036d5f7>] __percpu_counter_add+0x2a/0x8f
[<ffffffff88002544>] :aoe:aoeblk_make_request+0x1c3/0x1d0
[<ffffffff8058aaeb>] io_schedule+0x28/0x34
[<ffffffff8058cb8d>] error_exit+0x0/0x9a
[<ffffffff802ae5ee>] __set_page_dirty+0x48/0x121
[<ffffffff8036d5f7>] __percpu_counter_add+0x2a/0x8f
[<ffffffff802ae682>] __set_page_dirty+0xdc/0x121
[<ffffffff802ae75c>] mark_buffer_dirty+0x95/0x99
[<ffffffff802ae7d2>] __block_commit_write+0x72/0xac
[<ffffffff802ae988>] block_write_end+0x4f/0x5b
[<ffffffff802b243f>] blkdev_write_end+0x1b/0x38
[<ffffffff80265d96>] generic_file_buffered_write+0x1c0/0x648
[<ffffffff8023a752>] current_fs_time+0x22/0x29
[<ffffffff80266576>] __generic_file_aio_write_nolock+0x358/0x3c2
[<ffffffff80266c84>] filemap_fault+0x1c4/0x320
[<ffffffff80264cce>] unlock_page+0x2d/0x31
[<ffffffff802666dd>] generic_file_aio_write_nolock+0x3b/0x8d
[<ffffffff8028e40f>] do_sync_write+0xe2/0x126
[<ffffffff802497d0>] autoremove_wake_function+0x0/0x38
[<ffffffff8058e725>] do_page_fault+0x3f8/0x7bb
[<ffffffff8028cae8>] fd_install+0x5f/0x68
[<ffffffff8028eb98>] vfs_write+0xae/0x137
[<ffffffff8028f102>] sys_write+0x47/0x70
[<ffffffff8020b7ae>] system_call+0x7e/0x83
--
Ed L Cashin <[EMAIL PROTECTED]>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
