Date: Mon, 1 Dec 1997 15:55:30 -0800 (PST) From: Sean Eric Fagan <sef@kithrup.com> To: jlemon@americantv.com Cc: current@FreeBSD.ORG Subject: Re: cvs commit: src/sys/kern vfs_aio.c Message-ID: <199712012355.PAA11311@kithrup.com> In-Reply-To: <19971201155116.18825.kithrup.freebsd.current@right.PCS> References: <199712012017.PAA01433@khavrinen.lcs.mit.edu>; from Garrett Wollman on Dec 12, 1997 at 03:17:41PM -0500
next in thread | previous in thread | raw e-mail | index | archive | help
In article <19971201155116.18825.kithrup.freebsd.current@right.PCS> you write: >> This is an example of why said change was a mistake... >I never did get around to asking, so I guess this might be a good time as >any; why were the retvals moved into the proc structure in the first place? >I assume for performance reasons, but I don't recall seeing any discussion >about it on the lists. When I saw the commit message by PHK, I replied, to him and to cvs-all@freebsd.org (which is apparently not archived, so I can't get the exact date). I questioned the change, and said I "strongly objected" to it. I got neither acknowledgment nor response, from phk or anyone else. I saw no "Reviewed by" in the commit message, nor anything else to indicate that anyone else had seen the changes or approved of them. I disapprove of them for a variety of reasons -- they're overly intrusive, introduce too much incompatibility for no real gain, needlessly change the proc structure, don't allow for future system calls that might need to return more than two words, and make some of the stuff done in the computability libraries harder. About the only thing they did, other than causing a need to recompile libkvm, ps, and possibly gdb, was to allow the return arguments to not be on trap()'s stack. I did not, and still don't, see any reason for that -- I happen to much prefer the *retval method to the old method of using u.u_ar0 and u.u_ar1. Yes, I'm just a bit bitter about my complaint having gone unheeded and ignored, and having great delight in saying, "Hey, I told you so!" :) Without a compelling reason for them, I would prefer to see PHK's changes backed out. Sorry :(.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199712012355.PAA11311>