From owner-freebsd-hackers Mon Jan 29 17: 0: 5 2001 Delivered-To: freebsd-hackers@freebsd.org Received: from earth.backplane.com (earth-nat-cw.backplane.com [208.161.114.67]) by hub.freebsd.org (Postfix) with ESMTP id 9442A37B400; Mon, 29 Jan 2001 16:59:46 -0800 (PST) Received: (from dillon@localhost) by earth.backplane.com (8.11.1/8.9.3) id f0U0xkI80979; Mon, 29 Jan 2001 16:59:46 -0800 (PST) (envelope-from dillon) Date: Mon, 29 Jan 2001 16:59:46 -0800 (PST) From: Matt Dillon Message-Id: <200101300059.f0U0xkI80979@earth.backplane.com> To: John Baldwin Cc: freebsd-hackers@FreeBSD.org, Drew Eckhardt Subject: Re: Suboptimal mmap of devices on i86 References: Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG : :Oh yuck. Please do rename the darn thing back at some point then. Or if its :only going to be in -current, don't bother. No one expects to be able to use a :4.x module in 5.x, so just change the function w/o the old name. : :-- : :John Baldwin -- http://www.FreeBSD.org/~jhb/ I'd like to see this MFC'd to stable at some point and they should be the same patch. For anything in production, a rename is usually necessary when the API change is not backwards compatible. Even though it's somewhat unsightly, it's worth the time it saves developers (especially developers working on code outside the source tree). When I redid the socket descriptor reference code I pulled the API change & rename trick, and Julian popped out of the woodwork a week later complaining of a kld module refusing to load. I had missed a function. The rename undoubtedly saved him hours of kernel crash debugging. -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message