linux-kernel - Re: [PATCH] rcu/nocb: Add Safe checks for access offloaded rdp

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <2c8098f0-2010-4714-97ca-7f46629d67f8@nvidia.com>
Date: Wed, 30 Apr 2025 12:14:47 -0400
From: Joel Fernandes <joelagnelf@...dia.com>
To: Z qiang <qiang.zhang1211@...il.com>
Cc: Frederic Weisbecker <frederic@...nel.org>, paulmck@...nel.org,
 neeraj.upadhyay@...nel.org, joel@...lfernandes.org, urezki@...il.com,
 boqun.feng@...il.com, rcu@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] rcu/nocb: Add Safe checks for access offloaded rdp



On 4/30/2025 10:57 AM, Z qiang wrote:
>>
>>
>>
>> On 4/28/2025 6:59 AM, Z qiang wrote:
>>>>
>>>> Le Mon, Apr 28, 2025 at 05:54:03PM +0800, Zqiang a écrit :
>>>>> For Preempt-RT kernel, when enable CONFIG_PROVE_RCU Kconfig,
>>>>> disable local bh in rcuc kthreads will not affect preempt_count(),
>>>>> this resulted in the following splat:
>>>>>
>>>>> WARNING: suspicious RCU usage
>>>>> kernel/rcu/tree_plugin.h:36 Unsafe read of RCU_NOCB offloaded state!
>>>>> stack backtrace:
>>>>> CPU: 0 UID: 0 PID: 22 Comm: rcuc/0
>>>>> Call Trace:
>>>>> [    0.407907]  <TASK>
>>>>> [    0.407910]  dump_stack_lvl+0xbb/0xd0
>>>>> [    0.407917]  dump_stack+0x14/0x20
>>>>> [    0.407920]  lockdep_rcu_suspicious+0x133/0x210
>>>>> [    0.407932]  rcu_rdp_is_offloaded+0x1c3/0x270
>>>>> [    0.407939]  rcu_core+0x471/0x900
>>>>> [    0.407942]  ? lockdep_hardirqs_on+0xd5/0x160
>>>>> [    0.407954]  rcu_cpu_kthread+0x25f/0x870
>>>>> [    0.407959]  ? __pfx_rcu_cpu_kthread+0x10/0x10
>>>>> [    0.407966]  smpboot_thread_fn+0x34c/0xa50
>>>>> [    0.407970]  ? trace_preempt_on+0x54/0x120
>>>>> [    0.407977]  ? __pfx_smpboot_thread_fn+0x10/0x10
>>>>> [    0.407982]  kthread+0x40e/0x840
>>>>> [    0.407990]  ? __pfx_kthread+0x10/0x10
>>>>> [    0.407994]  ? rt_spin_unlock+0x4e/0xb0
>>>>> [    0.407997]  ? rt_spin_unlock+0x4e/0xb0
>>>>> [    0.408000]  ? __pfx_kthread+0x10/0x10
>>>>> [    0.408006]  ? __pfx_kthread+0x10/0x10
>>>>> [    0.408011]  ret_from_fork+0x40/0x70
>>>>> [    0.408013]  ? __pfx_kthread+0x10/0x10
>>>>> [    0.408018]  ret_from_fork_asm+0x1a/0x30
>>>>> [    0.408042]  </TASK>
>>>>>
>>>>> Currently, triggering an rdp offloaded state change need the
>>>>> corresponding rdp's CPU goes offline, and at this time the rcuc
>>>>> kthreads has already in parking state. this means the corresponding
>>>>> rcuc kthreads can safely read offloaded state of rdp while it's
>>>>> corresponding cpu is online.
>>>>>
>>>>> This commit therefore add rdp->rcu_cpu_kthread_task check for
>>>>> Preempt-RT kernels.
>>>>>
>>>>> Signed-off-by: Zqiang <qiang.zhang1211@...il.com>
>>>>> ---
>>>>>  kernel/rcu/tree_plugin.h | 4 +++-
>>>>>  1 file changed, 3 insertions(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
>>>>> index 003e549f6514..fe728eded36e 100644
>>>>> --- a/kernel/rcu/tree_plugin.h
>>>>> +++ b/kernel/rcu/tree_plugin.h
>>>>> @@ -31,7 +31,9 @@ static bool rcu_rdp_is_offloaded(struct rcu_data *rdp)
>>>>>                 lockdep_is_held(&rcu_state.nocb_mutex) ||
>>>>>                 (!(IS_ENABLED(CONFIG_PREEMPT_COUNT) && preemptible()) &&
>>>>>                  rdp == this_cpu_ptr(&rcu_data)) ||
>>>>> -               rcu_current_is_nocb_kthread(rdp)),
>>>>> +               rcu_current_is_nocb_kthread(rdp) ||
>>>>> +               (IS_ENABLED(CONFIG_PREEMPT_RT) &&
>>>>> +                current == rdp->rcu_cpu_kthread_task)),
>>>>
>>>> Isn't it safe also on !CONFIG_PREEMPT_RT ?
>>>
>>> For !CONFIG_PREEMPT_RT and  in rcuc kthreads, it's also safe,
>>> but the following check will passed :
>>>
>>> (!(IS_ENABLED(CONFIG_PREEMPT_COUNT) && preemptible()) &&
>>>           rdp == this_cpu_ptr(&rcu_data))
>>
>> I think the fact that it already passes for !PREEMPT_RT does not matter, because
>> it simplifies the code so drop the PREEMPT_RT check?
>>
>> Or will softirq_count() not work? It appears to have special casing for
>> PREEMPT_RT's local_bh_disable():
>>
>> (   ( !(IS_ENABLED(CONFIG_PREEMPT_COUNT) && preemptible()) || softirq_count() )
>>    && rdp == this_cpu_ptr(&rcu_data))  )
> 
> Thank you for Joel's reply,  I also willing to accept such
> modifications and resend :) .
Thanks, I am Ok with either approach whichever you and Frederic together decide.
I can then pull this in for the v6.16 merge window once you resend, thanks!

 - Joel