Date: Tue, 19 Feb 2008 21:38:42 +0200 From: Kostik Belousov <kostikbel@gmail.com> To: Robert Watson <rwatson@freebsd.org> Cc: Igor Sysoev <is@rambler-co.ru>, Jason Evans <jasone@freebsd.org>, freebsd-current@freebsd.org Subject: Re: malloc(3) ignores RLIMIT_DATA Message-ID: <20080219193842.GG57756@deviant.kiev.zoral.com.ua> In-Reply-To: <20080219185615.R21494@fledge.watson.org> References: <20080219151809.GF57366@rambler-co.ru> <47BB0D29.5080403@freebsd.org> <20080219185615.R21494@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--NZdDWKHbmnyWnFtu Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, Feb 19, 2008 at 06:58:08PM +0000, Robert Watson wrote: >=20 > On Tue, 19 Feb 2008, Jason Evans wrote: >=20 > >>As sbrk() is less preferable because of framentation and race condition= s,=20 > >>why not to create mmap() flag MMAP_DSS to check RLIMIT_DATA and to use = it=20 > >>in malloc(3) ? > > > >There has been general agreement among the people I've discussed this=20 > >issue with that the correct solution is to add a separate resource limit= =20 > >for anonymously mapped memory, which would provide capabilities similar = to=20 > >what your suggestion would provide. >=20 > Konstantine has updated his patches and reported on them in the recent=20 > status report: >=20 > http://www.freebsd.org/news/status/report-2007-10-2007-12.html#VM-Overc= ommit >=20 > Here's the main site for information on the patch: >=20 > http://people.freebsd.org/~kib/overcommit/ >=20 > He describes a per-uid limit, but I think it might also be useful to have= a=20 > per-process limit tht can also be enforced, although possibly not by=20 > default, so that protecting applications from each other doesn't require= =20 > creating separate users for them. Yes, per-process limits can be added too, although it would require some additional thinking. The persistent objects backed by anonymous memory, like SysV shm or shm_open(2) handles would be billed for the creator only. It is not immediately obvious whether it is right or not. Anyway, I want to get the review first, before doing further modifications. --NZdDWKHbmnyWnFtu Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (FreeBSD) iEYEARECAAYFAke7MEEACgkQC3+MBN1Mb4ikQgCfZrnzQ+FjqJHniRl/1KTOILm7 ymUAoNZmzI/NfVMhASNHC07i2dT4MxDc =7VSO -----END PGP SIGNATURE----- --NZdDWKHbmnyWnFtu--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080219193842.GG57756>