Date: Thu, 25 May 1995 10:42:44 +1000 From: Bruce Evans <bde@zeta.org.au> To: bde@zeta.org.au, terry@cs.weber.edu Cc: bde@freefall.cdrom.com, br\ian@med\iac\ity.com, hackers@FreeBSD.org, jkh@freefall.cdrom.com, julian@ref.tfs.com Subject: Re: Cyclades Driver/Multi-port Serial Card Message-ID: <199505250042.KAA07392@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>> large as the original driver). I intend to make even larger changes so >> that all the serial drivers use a common front end. cy.c currently has >> about 30K of code (out of 60K total) in common with sio.c. rc.c and >> cy.c are very similar. They drive similar Cirrus Logic chips and are >> based on sio.c for the non-hardware parts. >If you put a common cannonical processing front end on things, then you >are a god in my book! My personal preference is for a module that is >ldterm type seperable and shared with the console (I have a fetish for >Streams). I meant a front end for open() (there is lots of device-independent code for bidirectional devices), pseudo-DMA interrupt handlers (only the lowest level is very device-dependent), ... I think you actually want non-common not-necessarily-canonical processing: generate maximally efficient routines for certain devices, line disciplines, and internal settings of the line disciplines and load these into the kernel as required. This is too hard for me. I'll settle for maximally efficient slip and ppp disciplines. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199505250042.KAA07392>