Date: Sat, 3 Dec 2011 10:19:43 +0100 From: Filippo Sironi <filippo.sironi@gmail.com> To: freebsd-hackers@freebsd.org Cc: Sironi Filippo <filippo.sironi@gmail.com>, Poul-Henning Kamp <phk@phk.freebsd.dk> Subject: Re: mmap implementation for cdev Message-ID: <A842C229-1846-408A-999A-9347770EFE7A@gmail.com> In-Reply-To: <10964.1322903784@critter.freebsd.dk> References: <10964.1322903784@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Ok, that's what I was suspecting. Do you have any ideas on how to get the behavior I described in the = previous mail? On 03/dic/2011, at 10.16, Poul-Henning Kamp wrote: > In message <A51F9B30-A71A-49A5-B684-16EC3C127BC9@gmail.com>, Filippo = Sironi wri > tes: >=20 >> The problem is that when I come =3D >> back to user space all the threads write in memory page mapped by the = =3D >> first thread that called the mmap leaving their memory pages = untouched. >=20 > That's how cdev->mmap works, it is global rather than per-thread. >=20 >=20 > --=20 > Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 > phk@FreeBSD.ORG | TCP/IP since RFC 956 > FreeBSD committer | BSD since 4.3-tahoe =20 > Never attribute to malice what can adequately be explained by = incompetence.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A842C229-1846-408A-999A-9347770EFE7A>