Date: Fri, 7 May 1999 17:38:59 +1000 From: Bruce Evans <bde@zeta.org.au> To: bde@zeta.org.au, grog@lemis.com Cc: cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org, phk@critter.freebsd.dk Subject: Re: cvs commit: src/sys/sys disk.h Message-ID: <199905070738.RAA24019@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>Small is beautiful. One less header file is good. disk.h got >superseded by nothing :-) Not true. Things in it got moved to disklabel.h and diskslice.h "temporarily" since I didn't want to redesign the whole disk subsystem. ><parable-mode> > >Novice hacker goes to guru. > >NH: Guru, I want to write a driver. >G: RTFM. >NH: Which FM? >G: Read the source code of another driver, then WTFM. > ></parable-mode> > >Guess which driver I read? NH: Which FD? G: Depends on the type of driver. For disk drivers, read ahc source. >> They all got used but a few (2 or 3) recently became unused and fewer >> (1 or 2) haven't been used for some time. Countless generations of >> driver authors should have known to not even look at most fields in >> struct buf. Only vfs_bio and meta-drivers like ccd and vinum need to >> know about them. > >If these countless generations have known about it, how come buf(9) is >still so woefully inadequate? I had a crack at it a while back, but The existence of buf.9 is (almost) news to me (I vaguely remember the commit mail for it). The driver interface to buffers is mostly very old, so it is mostly documented in the 4.[3-4] BSD books. >I think you'll find that phk will drag up and eliminate a whole lot >more cruft before he's done. I would have thought that you'd applaud >his actions. In parts. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199905070738.RAA24019>