Date: Fri, 26 Feb 1999 15:47:13 -0500 From: Andrew Heybey <ath@niksun.com> To: Mike Smith <mike@smith.net.au> Cc: freebsd-hackers@freebsd.org, ken@plutotech.com, gibbs@plutotech.com Subject: Re: Advice wanted on tracking down bug (or hw problem?) in 3.1R Message-ID: <199902262047.PAA12740@stiegl.niksun.com> In-Reply-To: Your message of Fri, 26 Feb 1999 12:22:25 -0800. <199902262022.MAA09175@dingo.cdrom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
[cc'd to Ken Merry & Justin Gibbs to ask if they have ever seen such behavior. In short, under heavy network load (40k pkts/sec) my reads from disk sometimes return 8-64 bytes of garbage at the end of a 4k page. The controller is an aic7890 with IBM DRVS09V LVD disks. See PR kern/10243 for more info.] >>On Fri, 26 Feb 1999 12:22:25 -0800, Mike Smith <mike@smith.net.au> said: >> The data is always (in every instance that I have examined) from >> some other part of the file currently being read or some other >> file in my set of test files. How my test setup works is that I >> have 30 50MB files. The files are filled with sequential >> integers (counting over the entire 1.5GB). My test program reads >> from the files (in order, starting over at file #0 when it >> reaches file #29) and compares what read(2) returns to what >> should be there (based on file number and file offset). >> >> One other possible clue: This morning I hooked my disks up to the >> regular Ultra SCSI (40MB/s) port of the 7890 controller rather >> than the Ultra/2 (80MB/s) port and I haven't seen the bug yet. I >> am not 100% positive since I have only run it for a few hours so >> far, but before I could almost always make the bug happen withing >> 10-15 minutes. mike> Could you try bzero'ing your buffers before every read? This mike> sniffs very much like short transfers rather than sniping... Will do. Yes it does seem like a short transfer someplace. If I zero my buffers before every read then what will that tell me? If I get zeros rather than garbage from another file then I guess that somehow the data is not being moved into user space properly. If I still get garbage, then maybe a DMA is stopping short? Also, I still have not been able to reproduce the problem with the disks connected to the 40MB/s SCSI bus. I have cc'd ken & justin (sorry guys--don't know if you read hackers) to ask whether they have ever seen such behavior and whether there are any differences in the ahc driver for ultra vs. ultra2. andrew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199902262047.PAA12740>