Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 14 Jan 2017 14:42:23 +0200
From:      Andriy Gapon <avg@FreeBSD.org>
To:        Alexey Egorov <electreg@list.ru>, freebsd-hackers@FreeBSD.org
Subject:   Re: tracing mmap'ed IO with dtrace
Message-ID:  <3ba41915-49dd-6432-d1ea-8c5834ca12cc@FreeBSD.org>
In-Reply-To: <1484310020.407170115@f381.i.mail.ru>
References:  <1484310020.407170115@f381.i.mail.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
On 13/01/2017 14:20, Alexey Egorov via freebsd-hackers wrote:
> Hello,
> 
> I'm trying to use dtrace to trace filesystem IO, and it works great with ordinary read-write syscalls.
> However, it doesn't catch mmap'ed IO. I understand that it is not possible to catch all memory accesses to mmap'ed region, but is it possible to catch page faults that is caused by such accesses?
> dtrace manual says that there is vminfo::fspgin/fspgout probes; but it's not available on FreeBSD. Is it possible to use some of fbt probes here?

Try tracing calls to vm_fault.


-- 
Andriy Gapon



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3ba41915-49dd-6432-d1ea-8c5834ca12cc>