Re: [v2 PATCH] mm: shmem: allow split THP when truncating THP partially

From: Hugh Dickins <hughd@google.com>
To: David Hildenbrand <david@redhat.com>
Cc: Hugh Dickins <hughd@google.com>,
	Yang Shi <yang.shi@linux.alibaba.com>,
	 kirill.shutemov@linux.intel.com, aarcange@redhat.com,
	 akpm@linux-foundation.org, linux-mm@kvack.org,
	 linux-kernel@vger.kernel.org
Subject: Re: [v2 PATCH] mm: shmem: allow split THP when truncating THP partially
Date: Tue, 25 Feb 2020 12:31:07 -0800 (PST)	[thread overview]
Message-ID: <alpine.LSU.2.11.2002251216480.7087@eggly.anvils> (raw)
In-Reply-To: <14abd659-1571-8196-202d-d2fcc227a4b0@redhat.com>

On Tue, 25 Feb 2020, David Hildenbrand wrote:
> > 
> > I notice that this thread has veered off into QEMU ballooning
> > territory: which may indeed be important, but there's nothing at all
> > that I can contribute on that.  I certainly do not want to slow down
> > anything important, but remain convinced that the correct filesystem
> > implementation for punching a hole is to punch a hole.
> 
> I am not completely sure I follow all the shmem details (sorry!). But
> trying to "punch a partial hole punch" into a hugetlbfs page will result
> in the very same behavior as with shmem as of now, no?

I believe so.

> 
> FALLOC_FL_PUNCH_HOLE: "Within the specified range, partial filesystem
> blocks are zeroed, and whole filesystem blocks are removed from the
> file." ... After a successful call, subsequent reads from this range
> will return zeros."
> 
> So, as long as we are talking about partial blocks the documented
> behavior seems to be to only zero the memory.
> 
> Does this patch fix "FALLOC_FL_PUNCH_HOLE does not free blocks if called
> in block granularity on shmem" (which would be a valid fix),

Yes. The block size of tmpfs is (talking x86_64 for simplicity) 4KiB;
but when mounted huge, it transparently takes advantage of 2MiB extents
when it can.  Rather like a disk-based filesystem always presenting a
4KiB block interface, but stored on disk in multisector extents.

Whereas hugetlbfs is a different filesystem, which is and always has
been limited to supporting only certain larger block sizes.

> or does it
> try to implement something that is not documented? (removing partial
> blocks when called in sub-block granularity)

No.

> 
> I assume the latter, in which case I would interpret "punching a hole is
> to punch a hole" as "punching sub-blocks will not free blocks".
> 
> (if somebody could enlighten me which important piece I am missing or
> messing up, that would be great :) )
> 
> -- 
> Thanks,
> 
> David / dhildenb