Date: Sun, 10 Nov 2002 13:46:34 -0600 From: "Alan L. Cox" <alc@imimic.com> To: mjacob@feral.com Cc: alpha@FreeBSD.ORG, John Baldwin <jhb@FreeBSD.ORG> Subject: Re: on the same note.. Message-ID: <3DCEB79A.69EB752D@imimic.com> References: <Pine.BSF.4.21.0211101125460.60718-100000@beppo>
next in thread | previous in thread | raw e-mail | index | archive | help
Matthew Jacob wrote: > > Err, well, this *is* in vm object allocation... Ah, I see what you're talking about now. Until the vm object has been returned by the allocation routine it is not shared data. The generation count is used to detect changes in the list of resident pages by sleeping processes, not to detect that the same storage is being recycled for a new object. Alan > On Sun, 10 Nov 2002, Alan L. Cox wrote: > > > Matthew Jacob wrote: > > > > > > .... > > > > > > In vm_object.c you or somebody had a question about atomicity in bumping > > > the object generation count. What is wrong with just making that > > > atomic_add_int for now? I see no locks otherwise covering it. > > > > It and the other vm_object fields are generally covered by Giant. I > > believe that vm_object allocation and initialization are the only > > activities that aren't. > > > > Alan > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3DCEB79A.69EB752D>