Re: [PATCH V1] x86, espfix: postpone the initialization of espfix stack for AP

[Borislav Petkov <bp@alien8.de>]

On Wed, Jun 17, 2015 at 12:27:05AM -0700, H. Peter Anvin wrote:
> On 06/04/2015 02:45 AM, Gu Zheng wrote:
> > The following lockdep warning occurrs when running with latest kernel:
> > [    3.178000] ------------[ cut here ]------------
> > [    3.183000] WARNING: CPU: 128 PID: 0 at kernel/locking/lockdep.c:2755 lockdep_trace_alloc+0xdd/0xe0()
> > [    3.193000] DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags))
> > [    3.199000] Modules linked in:
> > 
> > [    3.203000] CPU: 128 PID: 0 Comm: swapper/128 Not tainted 4.1.0-rc3 #70
> > [    3.221000]  0000000000000000 2d6601fb3e6d4e4c ffff88086fd5fc38 ffffffff81773f0a
> > [    3.230000]  0000000000000000 ffff88086fd5fc90 ffff88086fd5fc78 ffffffff8108c85a
> > [    3.238000]  ffff88086fd60000 0000000000000092 ffff88086fd60000 00000000000000d0
> > [    3.246000] Call Trace:
> > [    3.249000]  [<ffffffff81773f0a>] dump_stack+0x4c/0x65
> > [    3.255000]  [<ffffffff8108c85a>] warn_slowpath_common+0x8a/0xc0
> > [    3.261000]  [<ffffffff8108c8e5>] warn_slowpath_fmt+0x55/0x70
> > [    3.268000]  [<ffffffff810ee24d>] lockdep_trace_alloc+0xdd/0xe0
> > [    3.274000]  [<ffffffff811cda0d>] __alloc_pages_nodemask+0xad/0xca0
> > [    3.281000]  [<ffffffff810ec7ad>] ? __lock_acquire+0xf6d/0x1560
> > [    3.288000]  [<ffffffff81219c8a>] alloc_page_interleave+0x3a/0x90
> > [    3.295000]  [<ffffffff8121b32d>] alloc_pages_current+0x17d/0x1a0
> > [    3.301000]  [<ffffffff811c869e>] ? __get_free_pages+0xe/0x50
> > [    3.308000]  [<ffffffff811c869e>] __get_free_pages+0xe/0x50
> > [    3.314000]  [<ffffffff8102640b>] init_espfix_ap+0x17b/0x320
> > [    3.320000]  [<ffffffff8105c691>] start_secondary+0xf1/0x1f0
> > [    3.327000] ---[ end trace 1b3327d9d6a1d62c ]---
> > 
> > As we alloc pages with GFP_KERNEL in init_espfix_ap() which is called
> > before enabled local irq, and the lockdep sub-system considers this
> > behaviour as allocating memory with GFP_FS with local irq disabled,
> > then trigger the warning as mentioned about.
> > 
> > Though we could allocate them on the boot CPU side and hand them over to
> > the secondary CPU, but it seemes a bit waste if some of cpus are offline.
> > As thers is no need to these pages(espfix stack) until we try to run user
> > code, so we postpone the initialization of espfix stack, and let the boot
> > up routine init the espfix stack for the target cpu after it booted to
> > avoid the noise.
> > 
> 
> It isn't *at all* obvious to me at least that if the GFP_KERNEL
> allocation fails we may not get rescheduled on another CPU and/or get stuck.
> 
> I'm starting to think that the right thing to do is to allocate these on
> the CPU that is bringing up the other CPU, at the same time we allocate
> the percpu area.  This won't affect offline CPUs.

Btw, as part of experimenting for something else, I was able to trigger
this even on a guest here. It is an insane guest though: 16 NUMA nodes,
with 8 cores each:

[    0.032000] ------------[ cut here ]------------
[    0.032000] WARNING: CPU: 64 PID: 0 at kernel/locking/lockdep.c:2755 lockdep_trace_alloc+0x10c/0x120()
[    0.032000] DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags))
[    0.032000] Modules linked in:
[    0.032000] CPU: 64 PID: 0 Comm: swapper/64 Not tainted 4.1.0-rc3+ #4
[    0.032000] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.7.5-20140531_083030-gandalf 04/01/2014
[    0.032000]  ffffffff818dd1a1 ffff880006a1fca8 ffffffff816a9685 0000000000000000
[    0.032000]  ffff880006a1fcf8 ffff880006a1fce8 ffffffff81058585 00000000001d74c0
[    0.032000]  0000000000000080 0000000000000046 ffff880047ffcd00 ffff88003e804058
[    0.032000] Call Trace:
[    0.032000]  [<ffffffff816a9685>] dump_stack+0x4f/0x7b
[    0.032000]  [<ffffffff81058585>] warn_slowpath_common+0x95/0xe0
[    0.032000]  [<ffffffff81058616>] warn_slowpath_fmt+0x46/0x50
[    0.032000]  [<ffffffff810a662c>] lockdep_trace_alloc+0x10c/0x120
[    0.032000]  [<ffffffff8113d6ed>] __alloc_pages_nodemask+0xad/0xab0
[    0.032000]  [<ffffffff813442d7>] ? debug_smp_processor_id+0x17/0x20
[    0.032000]  [<ffffffff810a370e>] ? put_lock_stats.isra.19+0xe/0x30
[    0.032000]  [<ffffffff816ae288>] ? mutex_lock_nested+0x2e8/0x420
[    0.032000]  [<ffffffff8117e0cc>] alloc_page_interleave+0x3c/0x90
[    0.032000]  [<ffffffff8117e995>] alloc_pages_current+0xc5/0xd0
[    0.032000]  [<ffffffff81138734>] __get_free_pages+0x14/0x50
[    0.032000]  [<ffffffff8100a484>] init_espfix_ap.part.5+0x164/0x270
[    0.032000]  [<ffffffff8100a5b1>] init_espfix_ap+0x21/0x30
[    0.032000]  [<ffffffff8103cd28>] start_secondary+0xe8/0x180
[    0.032000] ---[ end trace 6a7abdb28fbb7667 ]---

Now I can test the future fix too. :)

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--