Re: [RFC 0/3] Add madvise(..., MADV_WILLWRITE)

From: Andy Lutomirski <luto@amacapital.net>
To: Jan Kara <jack@suse.cz>
Cc: Dave Hansen <dave.hansen@intel.com>,
	linux-mm@kvack.org, linux-ext4@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [RFC 0/3] Add madvise(..., MADV_WILLWRITE)
Date: Thu, 8 Aug 2013 08:56:28 -0700	[thread overview]
Message-ID: <CALCETrX1GXr58ujqAVT5_DtOx+8GEiyb9svK-SGH9d+7SXiNqQ@mail.gmail.com> (raw)
In-Reply-To: <20130808101807.GB4325@quack.suse.cz>

On Thu, Aug 8, 2013 at 3:18 AM, Jan Kara <jack@suse.cz> wrote:
> On Wed 07-08-13 11:00:52, Andy Lutomirski wrote:
>> On Wed, Aug 7, 2013 at 10:40 AM, Dave Hansen <dave.hansen@intel.com> wrote:
>> > On 08/07/2013 06:40 AM, Jan Kara wrote:
>> >>   One question before I look at the patches: Why don't you use fallocate()
>> >> in your application? The functionality you require seems to be pretty
>> >> similar to it - writing to an already allocated block is usually quick.
>> >
>> > One problem I've seen is that it still costs you a fault per-page to get
>> > the PTEs in to a state where you can write to the memory.  MADV_WILLNEED
>> > will do readahead to get the page cache filled, but it still leaves the
>> > pages unmapped.  Those faults get expensive when you're trying to do a
>> > couple hundred million of them all at once.
>>
>> I have grand plans to teach the kernel to use hardware dirty tracking
>> so that (some?) pages can be left clean and writable for long periods
>> of time.  This will be hard.
>   Right that will be tough... Although with your application you could
> require such pages to be mlocked and then I could imagine we would get away
> at least from problems with dirty page accounting.

True.  The nasty part will be all the code that assumes that the acts
of un-write-protecting and dirtying are the same thing, for example
__block_write_begin, which is why I don't really believe in my
willwrite patches...

>
>> Even so, the second write fault to a page tends to take only a few
>> microseconds, while the first one often blocks in fs code.
>   So you wrote blocks are already preallocated with fallocate(). If you
> also preload pages in memory with MADV_WILLNEED is there still big
> difference between the first and subsequent write fault?

I haven't measured it yet, because I suspect that my patches are
rather buggy in their current form.  But the idea is that fallocate
will do the heavy lifting and give me a nice contiguous allocation,
and the MADV_WILLNEED call will take about as long as the first write
fault would have taken.  Then the first write fault after
MADV_WILLNEED will take about as long as the second write fault would
have taken without it.

--Andy