Date: Thu, 05 Mar 1998 17:19:50 GMT From: jak@cetlink.net (John Kelly) To: John Saunders <john.saunders@scitec.com.au> Cc: freebsd-current@FreeBSD.ORG Subject: Re: silo overflows (Was Re: 3.0-RELEASE?) Message-ID: <34ffdd5f.17323509@mail.cetlink.net> In-Reply-To: <34FE191D.1D676AC3@scitec.com.au> References: <XFMail.980304185844.shimon@simon-shapiro.org> <34FE191D.1D676AC3@scitec.com.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 05 Mar 1998 14:16:45 +1100, John Saunders <john.saunders@scitec.com.au> wrote: >If you use an IDE disk then 10 to 1 that's the problem. Particularly >if the IDE driver is doing multi-sector transfers. Interrupts must be >disabled during IDE PIO transfers (or you can risk data corruption) >and the time taken to transfer, say 16 sectors, is significant. When Sorry, but this is a myth perpetuated by some Linux users. FreeBSD does not have this problem with IDE disks. I have put Linux and FreeBSD side by side and tested this. Linux fails the test, while FreeBSD never misses a serial byte, even under heavy loading of the disk. The problem is related to the 650 UART support in sio.c. The 550 support works fine. >The solution is to use either a SCSI disk system, or support >busmastering IDE. The real problem is the crazy IRQ priorities of >the PC architeture. Nope, not a problem at all for FreeBSD. The IDE flaw is a Linux problem. -- Browser war over, Mozilla now free. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?34ffdd5f.17323509>