Date: Wed, 06 Jun 2001 16:24:20 -0700 (PDT) From: John Baldwin <jhb@FreeBSD.org> To: Alfred Perlstein <alfred@FreeBSD.org> Cc: obrien@FreeBSD.org, Freebsd-alpha@FreeBSD.org, Andrew Gallatin <gallatin@cs.duke.edu> Subject: Re: -current kernel still considered dangerous Message-ID: <XFMail.010606162420.jhb@FreeBSD.org> In-Reply-To: <20010606184523.P1832@superconductor.rush.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 06-Jun-01 Alfred Perlstein wrote: > * John Baldwin <jhb@FreeBSD.org> [010606 12:58] wrote: >> >> > Is there any documentation what the locking requirements of various >> > vm functions are now? I tested osf1 after my initial set of commits to >> > catch >> > alpha up to x86, but an assert must have been added since then. >> >> Not really, and it is in a state of flux right now. On my todo list is to >> change the vm_map's to be locked by a sx lock, and once that is done I will >> change those mtx_assert's to simply require the sx lock rather than the >> vm_mtx >> lock. However, I'm still not sure how vm_page's will be locked. >> vm_object's >> will probably have their own mutex or sx lock though. > > Linux uses a single lock to protect them, most of the splvm()'s that you > removed were placeholders for the vm page queue's mutex. Which use a single lock: pages, objects, or both? Due to the fact that getpages/putpages can block, I'm thinking that objects may actually need a sx lock. -- John Baldwin <jhb@FreeBSD.org> -- http://www.FreeBSD.org/~jhb/ PGP Key: http://www.baldwin.cx/~john/pgpkey.asc "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ 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?XFMail.010606162420.jhb>