Date: Mon, 26 May 1997 00:47:41 -0700 From: John-Mark Gurney <jmg@hydrogen.nike.efn.org> To: dmaddox@scsn.net Cc: Wolfgang Helbig <helbig@MX.BA-Stuttgart.De>, FreeBSD-current@FreeBSD.ORG Subject: Re: sio1: 64 events for device with no tp Message-ID: <19970526004741.09980@hydrogen.nike.efn.org> In-Reply-To: <19970525161640.02033@cola47.scsn.net>; from Donald J. Maddox on Sun, May 25, 1997 at 04:16:40PM -0400 References: <199705251951.VAA00443@helbig.informatik.ba-stuttgart.de> <19970525161640.02033@cola47.scsn.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Donald J. Maddox scribbled this message on May 25: > On Sun, May 25, 1997 at 09:51:22PM +0200, Wolfgang Helbig wrote: > > Hi, > > > > The above kernel message shows up during booting for both sio0 and sio1. > > Up to now its just annoying -- the messages are repeated once > > syslogd is started. > > > > This behaviour seems buggy to me. These messages were not > > by the release 1.164 sio driver. The kernel is pretty current, > > using release 1.169 of sio.c. (FreeBSD-current) > > > > Only the serial interface 0 is used on this system, connected to > > a mouse. It still works. > > > > Maybe there is some bug in sio.c to be fixed ? > > The following is a patch from phk that will stop the messages... If you > want to enable the COM_BIGFIFO stuff, set FLAGS=0x20000 in the sio lines in > your config file (COM_BIGFIFO doesn't work for my Motorola ModemSurfer Modem. > YMMV). could you redo your patch to reduce the number of style changes... right now I only see about 4-6 lines that are different, and about 10-20 that are removed.. the rest is style changes... and was confusing the heck out of me... of course I'm working on another project that will allow you to set the port's fifo to an arbitrary value using an ioctl... so this patch really wont do much good... all it does is disable the fifo on the port.. it would be better to set it to 8bytes... that way the buffering is actually useful... this will still give you 69us instead of the normal 17us interupt latency... (at 115200bps) this patch will give you 130us... (this assumes normal 16550 16byte fifo)... > + outb(iobase+com_dlbh, 0xff); > + outb(iobase+com_dlbl, 0xff); is setting the port to ~1.757bps really neccessary?? won't this kill the serial console running on the port?? > + outb(iobase+com_mcr, com->mcr_image); > + outb(iobase+com_fifo, 0); I haven't done that much with sio, so I'll have to get back to you on this one... -- John-Mark Gurney Modem/FAX: +1 541 683 6954 Cu Networking Live in Peace, destroy Micro$oft, support free software, run FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970526004741.09980>