Re: [PATCH v3 0/2] iommu/iova: Make the rcache depot properly flexible

[Ethan Zhao <haifeng.zhao@linux.intel.com>]

On 1/9/2024 1:54 PM, Ethan Zhao wrote:
>
> On 1/9/2024 1:35 AM, Robin Murphy wrote:
>> On 2023-12-28 12:23 pm, Ido Schimmel wrote:
>>> On Tue, Sep 12, 2023 at 05:28:04PM +0100, Robin Murphy wrote:
>>>> v2: 
>>>> https://lore.kernel.org/linux-iommu/cover.1692641204.git.robin.murphy@arm.com/
>>>>
>>>> Hi all,
>>>>
>>>> I hope this is good to go now, just fixed the locking (and threw
>>>> lockdep at it to confirm, which of course I should have done to begin
>>>> with...) and picked up tags.
>>>
>>> Hi,
>>>
>>> After pulling the v6.7 changes we started seeing the following memory
>>> leaks [1] of 'struct iova_magazine'. I'm not sure how to reproduce it,
>>> which is why I didn't perform bisection. However, looking at the
>>> mentioned code paths, they seem to have been changed in v6.7 as part of
>>> this patchset. I reverted both patches and didn't see any memory leaks
>>> when running a full regression (~10 hours), but I will repeat it to be
>>> sure.
>>>
>>> Any idea what could be the problem?
>>
>> Hmm, we've got what looks to be a set of magazines forming a 
>> plausible depot list (or at least the tail end of one):
>>
>> ffff8881411f9000 -> ffff8881261c1000
>>
>> ffff8881261c1000 -> ffff88812be26400
>>
>> ffff88812be26400 -> ffff8188392ec000
>>
>> ffff8188392ec000 -> ffff8881a5301000
>>
>> ffff8881a5301000 -> NULL
>>
>> which I guess has somehow become detached from its rcache->depot 
>> without being freed properly? However I'm struggling to see any 
>> conceivable way that could happen which wouldn't already be more 
>> severely broken in other ways as well (i.e. either general memory 
>> corruption or someone somehow still trying to use the IOVA domain 
>> while it's being torn down).
>>
>> Out of curiosity, does reverting just patch #2 alone make a 
>> difference? And is your workload doing anything "interesting" in 
>> relation to IOVA domain lifetimes, like creating and destroying 
>> SR-IOV virtual functions, changing IOMMU domain types via sysfs, or 
>> using that horrible vdpa thing, or are you seeing this purely from 
>> regular driver DMA API usage?
>
> There no lock held when free_iova_rcaches(), is it possible 
> free_iova_rcaches() race with the delayed cancel_delayed_work_sync() ?
>
> I don't know why not call cancel_delayed_work_sync(&rcache->work); 
> first in free_iova_rcaches() to avoid possible race.
>
between following functions pair, race possible ? if called cocurrently.

1. free_iova_rcaches() with iova_depot_work_func()

    free_iova_rcaches() holds no lock, iova_depot_work_func() holds 
rcache->lock.

2. iova_cpuhp_dead() with iova_depot_work_func()

   iova_cpuhp_dead() holds per cpu lock cpu_rcache->lock, 
iova_depot_work_func() holds rcache->lock.

3. iova_cpuhp_dead() with free_iova_rcaches()

    iova_cpuhp_dead() holds per cpu lock cpu_rcache->lock, 
free_iova_rcaches() holds no lock.

4. iova_cpuhp_dead() with free_global_cached_iovas()

    iova_cpuhp_dead() holds per cpu lock cpu_rcache->lock and 
free_global_cached_iovas() holds rcache->lock.

......

Thanks,

Ethan


>
> Thanks,
>
> Ethan
>
>>
>> Thanks,
>> Robin.
>>
>>>
>>> Thanks
>>>
>>> [1]
>>> unreferenced object 0xffff8881a5301000 (size 1024):
>>>    comm "softirq", pid 0, jiffies 4306297099 (age 462.991s)
>>>    hex dump (first 32 bytes):
>>>      00 00 00 00 00 00 00 00 e7 7d 05 00 00 00 00 00 .........}......
>>>      0f b4 05 00 00 00 00 00 b4 96 05 00 00 00 00 00 ................
>>>    backtrace:
>>>      [<ffffffff819f5f08>] __kmem_cache_alloc_node+0x1e8/0x320
>>>      [<ffffffff818a239a>] kmalloc_trace+0x2a/0x60
>>>      [<ffffffff8231d31e>] free_iova_fast+0x28e/0x4e0
>>>      [<ffffffff82310860>] fq_ring_free_locked+0x1b0/0x310
>>>      [<ffffffff8231225d>] fq_flush_timeout+0x19d/0x2e0
>>>      [<ffffffff813e95ba>] call_timer_fn+0x19a/0x5c0
>>>      [<ffffffff813ea16b>] __run_timers+0x78b/0xb80
>>>      [<ffffffff813ea5bd>] run_timer_softirq+0x5d/0xd0
>>>      [<ffffffff82f1d915>] __do_softirq+0x205/0x8b5
>>>
>>> unreferenced object 0xffff8881392ec000 (size 1024):
>>>    comm "softirq", pid 0, jiffies 4306326731 (age 433.359s)
>>>    hex dump (first 32 bytes):
>>>      00 10 30 a5 81 88 ff ff 50 ff 0f 00 00 00 00 00 ..0.....P.......
>>>      f3 99 05 00 00 00 00 00 87 b7 05 00 00 00 00 00 ................
>>>    backtrace:
>>>      [<ffffffff819f5f08>] __kmem_cache_alloc_node+0x1e8/0x320
>>>      [<ffffffff818a239a>] kmalloc_trace+0x2a/0x60
>>>      [<ffffffff8231d31e>] free_iova_fast+0x28e/0x4e0
>>>      [<ffffffff82310860>] fq_ring_free_locked+0x1b0/0x310
>>>      [<ffffffff8231225d>] fq_flush_timeout+0x19d/0x2e0
>>>      [<ffffffff813e95ba>] call_timer_fn+0x19a/0x5c0
>>>      [<ffffffff813ea16b>] __run_timers+0x78b/0xb80
>>>      [<ffffffff813ea5bd>] run_timer_softirq+0x5d/0xd0
>>>      [<ffffffff82f1d915>] __do_softirq+0x205/0x8b5
>>>
>>> unreferenced object 0xffff8881411f9000 (size 1024):
>>>    comm "softirq", pid 0, jiffies 4306708887 (age 51.459s)
>>>    hex dump (first 32 bytes):
>>>      00 10 1c 26 81 88 ff ff 2c 96 05 00 00 00 00 00 ...&....,.......
>>>      ac fe 0f 00 00 00 00 00 a6 fe 0f 00 00 00 00 00 ................
>>>    backtrace:
>>>      [<ffffffff819f5f08>] __kmem_cache_alloc_node+0x1e8/0x320
>>>      [<ffffffff818a239a>] kmalloc_trace+0x2a/0x60
>>>      [<ffffffff8231d31e>] free_iova_fast+0x28e/0x4e0
>>>      [<ffffffff82310860>] fq_ring_free_locked+0x1b0/0x310
>>>      [<ffffffff8231225d>] fq_flush_timeout+0x19d/0x2e0
>>>      [<ffffffff813e95ba>] call_timer_fn+0x19a/0x5c0
>>>      [<ffffffff813ea16b>] __run_timers+0x78b/0xb80
>>>      [<ffffffff813ea5bd>] run_timer_softirq+0x5d/0xd0
>>>      [<ffffffff82f1d915>] __do_softirq+0x205/0x8b5
>>>
>>> unreferenced object 0xffff88812be26400 (size 1024):
>>>    comm "softirq", pid 0, jiffies 4306710027 (age 50.319s)
>>>    hex dump (first 32 bytes):
>>>      00 c0 2e 39 81 88 ff ff 32 ab 05 00 00 00 00 00 ...9....2.......
>>>      e3 ac 05 00 00 00 00 00 1f b6 05 00 00 00 00 00 ................
>>>    backtrace:
>>>      [<ffffffff819f5f08>] __kmem_cache_alloc_node+0x1e8/0x320
>>>      [<ffffffff818a239a>] kmalloc_trace+0x2a/0x60
>>>      [<ffffffff8231d31e>] free_iova_fast+0x28e/0x4e0
>>>      [<ffffffff82310860>] fq_ring_free_locked+0x1b0/0x310
>>>      [<ffffffff8231225d>] fq_flush_timeout+0x19d/0x2e0
>>>      [<ffffffff813e95ba>] call_timer_fn+0x19a/0x5c0
>>>      [<ffffffff813ea16b>] __run_timers+0x78b/0xb80
>>>      [<ffffffff813ea5bd>] run_timer_softirq+0x5d/0xd0
>>>      [<ffffffff82f1d915>] __do_softirq+0x205/0x8b5
>>>
>>> unreferenced object 0xffff8881261c1000 (size 1024):
>>>    comm "softirq", pid 0, jiffies 4306711547 (age 48.799s)
>>>    hex dump (first 32 bytes):
>>>      00 64 e2 2b 81 88 ff ff c0 7c 05 00 00 00 00 00 .d.+.....|......
>>>      87 a5 05 00 00 00 00 00 0e 9a 05 00 00 00 00 00 ................
>>>    backtrace:
>>>      [<ffffffff819f5f08>] __kmem_cache_alloc_node+0x1e8/0x320
>>>      [<ffffffff818a239a>] kmalloc_trace+0x2a/0x60
>>>      [<ffffffff8231d31e>] free_iova_fast+0x28e/0x4e0
>>>      [<ffffffff82310860>] fq_ring_free_locked+0x1b0/0x310
>>>      [<ffffffff8231225d>] fq_flush_timeout+0x19d/0x2e0
>>>      [<ffffffff813e95ba>] call_timer_fn+0x19a/0x5c0
>>>      [<ffffffff813ea16b>] __run_timers+0x78b/0xb80
>>>      [<ffffffff813ea5bd>] run_timer_softirq+0x5d/0xd0
>>>      [<ffffffff82f1d915>] __do_softirq+0x205/0x8b5
>>
>