Date: Sun, 15 Sep 2013 21:53:47 +0200 From: Davide Italiano <davide@freebsd.org> To: Pedro Giffuni <pfg@freebsd.org> Cc: dtrace@freebsd.org Subject: Re: vmem(9) use in Dtrace Message-ID: <CACYV=-H42pN2TXAp6vCjWRH=8g9Ttpzcex1RPd1Unj1oZWeRGg@mail.gmail.com> In-Reply-To: <52360A3E.40804@FreeBSD.org> References: <52360A3E.40804@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Sep 15, 2013 at 9:27 PM, Pedro Giffuni <pfg@freebsd.org> wrote: > Hi; > > Just noticed this, in the old DtraceTODO wiki: > > https://wiki.freebsd.org/DTraceTODO > > There is mention of some use that Dtrace makes of Solaris's vmem > allocator, specifically vmem_create(), vmem_destroy(), vmem_alloc(), and > vmem_free(). These functions (not exact but similar ones) have > been brought to FreeBSD 10 (r252330). > > According to opengrok we may just have to uncomment some Solaris > code in sys/cddl/contrib/opensolaris/uts/common/dtrace/dtrace.c > We need also to provide at least some sort of wrappers for e.g. flags passed to vmem_alloc which takes VM_SLEEP in solaris while in FreeBSD this is M_WAITOK etc... > This would reduce differences with the Solaris code but it has > to be examined carefully as I think John Birrell had already > solved the issue by providing the resource id's natively (kmem). > > Just thought I would point it out ... It may be that it is not > worth spending too much time on it since the existing code should > just work. > vmem is designed to be a generic resource allocator. I still need to analyze how Dtrace uses it but in general it provides better scalability and less fragmentation wrt home-rolled allocators we have in the kernel. FWIW, Illumos uses it for several things, e.g. pid/tid allocation. It might worth a try. > Pedro. > _______________________________________________ > freebsd-dtrace@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-dtrace > To unsubscribe, send any mail to "freebsd-dtrace-unsubscribe@freebsd.org" Thanks, -- Davide "There are no solved problems; there are only problems that are more or less solved" -- Henri Poincare
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CACYV=-H42pN2TXAp6vCjWRH=8g9Ttpzcex1RPd1Unj1oZWeRGg>