linux-kernel - Re: [PATCH v2 5/7] iommu/vt-d: Fix suspicious RCU usage in probe_acpi_namespace

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: <1cf01e76-88d2-c65f-0b54-b85e6da0d720@arm.com>
Date:   Fri, 1 Jul 2022 09:18:08 +0100
From:   Robin Murphy <robin.murphy@....com>
To:     Baolu Lu <baolu.lu@...ux.intel.com>,
        Joerg Roedel <joro@...tes.org>,
        David Woodhouse <dwmw2@...radead.org>
Cc:     kevin.tian@...el.com, ashok.raj@...el.com,
        linux-kernel@...r.kernel.org, iommu@...ts.linux-foundation.org,
        cai@....pw, jacob.jun.pan@...el.com
Subject: Re: [PATCH v2 5/7] iommu/vt-d: Fix suspicious RCU usage in
 probe_acpi_namespace_devices()

On 2022-07-01 08:19, Baolu Lu wrote:
> On 2022/6/29 21:03, Robin Murphy wrote:
>> On 2019-06-12 01:28, Lu Baolu wrote:
>>> The drhd and device scope list should be iterated with the
>>> iommu global lock held. Otherwise, a suspicious RCU usage
>>> message will be displayed.
>>>
>>> [    3.695886] =============================
>>> [    3.695917] WARNING: suspicious RCU usage
>>> [    3.695950] 5.2.0-rc2+ #2467 Not tainted
>>> [    3.695981] -----------------------------
>>> [    3.696014] drivers/iommu/intel-iommu.c:4569 suspicious 
>>> rcu_dereference_check() usage!
>>> [    3.696069]
>>>                 other info that might help us debug this:
>>>
>>> [    3.696126]
>>>                 rcu_scheduler_active = 2, debug_locks = 1
>>> [    3.696173] no locks held by swapper/0/1.
>>> [    3.696204]
>>>                 stack backtrace:
>>> [    3.696241] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.2.0-rc2+ 
>>> #2467
>>> [    3.696370] Call Trace:
>>> [    3.696404]  dump_stack+0x85/0xcb
>>> [    3.696441]  intel_iommu_init+0x128c/0x13ce
>>> [    3.696478]  ? kmem_cache_free+0x16b/0x2c0
>>> [    3.696516]  ? __fput+0x14b/0x270
>>> [    3.696550]  ? __call_rcu+0xb7/0x300
>>> [    3.696583]  ? get_max_files+0x10/0x10
>>> [    3.696631]  ? set_debug_rodata+0x11/0x11
>>> [    3.696668]  ? e820__memblock_setup+0x60/0x60
>>> [    3.696704]  ? pci_iommu_init+0x16/0x3f
>>> [    3.696737]  ? set_debug_rodata+0x11/0x11
>>> [    3.696770]  pci_iommu_init+0x16/0x3f
>>> [    3.696805]  do_one_initcall+0x5d/0x2e4
>>> [    3.696844]  ? set_debug_rodata+0x11/0x11
>>> [    3.696880]  ? rcu_read_lock_sched_held+0x6b/0x80
>>> [    3.696924]  kernel_init_freeable+0x1f0/0x27c
>>> [    3.696961]  ? rest_init+0x260/0x260
>>> [    3.696997]  kernel_init+0xa/0x110
>>> [    3.697028]  ret_from_fork+0x3a/0x50
>>>
>>> Fixes: fa212a97f3a36 ("iommu/vt-d: Probe DMA-capable ACPI name space 
>>> devices")
>>> Signed-off-by: Lu Baolu <baolu.lu@...ux.intel.com>
>>> ---
>>>   drivers/iommu/intel-iommu.c | 2 ++
>>>   1 file changed, 2 insertions(+)
>>>
>>> diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
>>> index 19c4c387a3f6..84e650c6a46d 100644
>>> --- a/drivers/iommu/intel-iommu.c
>>> +++ b/drivers/iommu/intel-iommu.c
>>> @@ -4793,8 +4793,10 @@ int __init intel_iommu_init(void)
>>>       cpuhp_setup_state(CPUHP_IOMMU_INTEL_DEAD, "iommu/intel:dead", 
>>> NULL,
>>>                 intel_iommu_cpu_dead);
>>> +    down_read(&dmar_global_lock);
>>>       if (probe_acpi_namespace_devices())
>>>           pr_warn("ACPI name space devices didn't probe correctly\n");
>>> +    up_read(&dmar_global_lock);
>>
>> Doing a bit of archaeology here, is this actually broken? If any ANDD 
>> entries exist, we'd end up doing:
>>
>>    down_read(&dmar_global_lock)
>>    probe_acpi_namespace_devices()
>>    -> iommu_probe_device()
>>       -> iommu_create_device_direct_mappings()
>>          -> iommu_get_resv_regions()
>>             -> intel_iommu_get_resv_regions()
>>                -> down_read(&dmar_global_lock)
>>
>> I'm wondering whether this might explain why my bus_set_iommu series 
>> prevented Baolu's machine from booting, since "iommu: Move bus setup 
>> to IOMMU device registration" creates the same condition where we end 
>> up in get_resv_regions (via bus_iommu_probe() this time) from the same 
>> task that already holds dmar_global_lock. Of course that leaves me 
>> wondering how it *did* manage to boot OK on my Xeon box, but maybe 
>> there's a config difference or dumb luck at play?
> 
> This is really problematic. Where does the latest bus_set_iommu series
> locate? I'd like to take a closer look at what happened here. Perhaps
> two weeks later? I'm busy with preparing Intel IOMMU patches for v5.20
> these days.

I've prepared an up-to-date series here:

https://gitlab.arm.com/linux-arm/linux-rm/-/tree/bus-set-iommu-v3

but I've been hesitant to post it without trying to make *some* progress 
on your breakage. I think last time I was just testing with 
x86_64_defconfig, so I'll double-check it with lockdep this afternoon.

Thanks,
Robin.