Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!hookup!usc!bloom-beacon.mit.edu!eru.mt.luth.se!omega.ludd.luth.se!news.ifm.liu.se!allvis.isy.liu.se!user From: kentp@isy.liu.se (Kent Palmkvist) Newsgroups: comp.os.386bsd.questions Subject: FreeBSD and ST01/02, Future Domain 850/950 Date: 22 Nov 1994 12:28:52 GMT Organization: Dept. EE, Linköping University Lines: 45 Message-ID: <kentp-2211941329020001@allvis.isy.liu.se> Reply-To: kentp@isy.liu.se NNTP-Posting-Host: allvis.isy.liu.se The following is a few hints regarding the use of the Future Domain 950/850, also known as Seagate ST01/02 with FreeBSD. There have been some postings about problems using this driver with different disks. The small documentation available is in /sys/doc/seagate.c. There are a few steps that can be taken if you have problems with the driver: 1) Increase timeout constants in the code. The code has been created on a 386DX 25MHz and the constants might be a bit small. It have despite this been used on a 486DX4-100 without problems. 2) Undefine the SEA_BLINDTRANSFERS and SEA_ASSEMBLER This makes the driver not use blind transfers during the data transfer phase. Blind transfers assume that all transfers in the data transfer phase are done in blocks of 512 bytes and that they are not interrupted. There are also an assumption made about the maximal delay between bytes on the scsibus during the blind transfer. 3) Define SEANOMSGS This definition stops the use of messages. This makes the driver not use disconnects. The driver then waits for the disk to complete the commands before returning to the os. This may cause problems with other drivers as the interrupts are inhibited as long the driver waits for the disk. This may still not be enough. I have found that a Seagate Drive I have does not identify itself correctly, even with the definitions above. I will try to find out more when I have the time. I was thinking of adding support for Future Domain 840/841/880/881, but I have no such card. Is there any interrest in such support? Any volunteers and/or test sites? Feel free to contact me if you have ideas or patches. I like to hear from you. Regards /Kent -- Kent Palmkvist email: kentp@isy.liu.se Dept. of Electrical Engineering phone: +4613281347 Linköping University fax: +4613139282 58183 LINKÖPING Sweden