Re: [shmem] 7a7256d5f5: WARNING:possible_recursive_locking_detected

From: Randy Dunlap <rdunlap@infradead.org>
To: Ira Weiny <ira.weiny@intel.com>,
	Andrew Morton <akpm@linux-foundation.org>
Cc: Matthew Wilcox <willy@infradead.org>,
	kernel test robot <yujie.liu@intel.com>,
	"Fabio M. De Francesco" <fmdefrancesco@gmail.com>,
	lkp@lists.01.org, lkp@intel.com, linux-kernel@vger.kernel.org,
	linux-mm@kvack.org, Peter Xu <peterx@redhat.com>,
	Andrea Arcangeli <aarcange@redhat.com>
Subject: Re: [shmem] 7a7256d5f5: WARNING:possible_recursive_locking_detected
Date: Fri, 21 Oct 2022 17:02:57 -0700	[thread overview]
Message-ID: <a0e2af81-39de-8c25-e7bd-f2234a0d7f8b@infradead.org> (raw)
In-Reply-To: <Y1MymJ/INb45AdaY@iweiny-desk3>

On 10/21/22 17:00, Ira Weiny wrote:

> 
> I was about to get this patch ready to send when I found this:
> 
> commit b6ebaedb4cb1a18220ae626c3a9e184ee39dd248
> Author: Andrea Arcangeli <aarcange@redhat.com>
> Date:   Fri Sep 4 15:47:08 2015 -0700
> 
>     userfaultfd: avoid mmap_sem read recursion in mcopy_atomic
> 
>     If the rwsem starves writers it wasn't strictly a bug but lockdep
>     doesn't like it and this avoids depending on lowlevel implementation  
>     details of the lock.
>     
>     [akpm@linux-foundation.org: delete weird BUILD_BUG_ON()]
>     Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
>     Acked-by: Pavel Emelyanov <xemul@parallels.com>
> ...
> 
> So I wonder if the true fix is something to lockdep?
> 
> Regardless I'll send the below patch because it will restore things to a
> working order.

with ; added to each + line.

> But I'm CC'ing Andrea for comments.
> 
> Ira
> 
>>
>> Ira
>>
>> diff --git a/mm/shmem.c b/mm/shmem.c
>> index 8280a5cb48df..6c8e99bf5983 100644
>> --- a/mm/shmem.c
>> +++ b/mm/shmem.c
>> @@ -2424,9 +2424,11 @@ int shmem_mfill_atomic_pte(struct mm_struct *dst_mm,
>>
>>                 if (!zeropage) {        /* COPY */
>>                         page_kaddr = kmap_local_folio(folio, 0);
>> +                       pagefault_disable()
>>                         ret = copy_from_user(page_kaddr,
>>                                              (const void __user *)src_addr,
>>                                              PAGE_SIZE);
>> +                       pagefault_enable()
>>                         kunmap_local(page_kaddr);
>>
>>                         /* fallback to copy_from_user outside mmap_lock */
>>
>>

-- 
~Randy