From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Mon, 29 Jan 2001 11:11:14 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Mon, 29 Jan 2001 11:10:55 -0500 Received: from md.aacisd.com ([64.23.207.34]:46610 "HELO md.aacisd.com") by vger.kernel.org with SMTP id ; Mon, 29 Jan 2001 11:10:49 -0500 Message-ID: <8FED3D71D1D2D411992A009027711D671859@md> From: Nathan Black To: linux-kernel@vger.kernel.org Subject: Bolck Device problem or Compaq Smart array 2 problem? kernel -2.4 .0+ Date: Mon, 29 Jan 2001 11:06:15 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: multipart/mixed; boundary="----_=_NextPart_000_01C08A0D.680CC850" Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. ------_=_NextPart_000_01C08A0D.680CC850 Content-Type: text/plain; charset="iso-8859-1" Here is my scenario. I have a smart array controller 2. I have 6 logical drives( one on it's own physical drive). I am using the one drive for data aquisition from an atm board. To locate my problem, I removed all of the atm stuff. and I am trying just a read/write. Both show the same symptoms. I have tried many different kernels. Here are my results. 2.2.18- works fine. 24 MBytes/sec at 100+ gigabytes (16GB looped many times ( lseek64(FD,SEEK_SET,0) )). 2.4.0 release SMP and Uniprocessor with NMI on- Kernel oops. I can reproduce if necessary( oops at about 700 MB) sometimes more, sometime less. (In BDFLUSH if I recall) 2.4.0 release UniProcessor NMI off- Works like the 2.2.18 2.4.1-pre10 & 11- Works but system becomes unusable(requires reboot) after completing. Here is my setup. Compaq Proliant 8500R 4GB Ram 4 Hot Swap 18GB hard disk. Compaq Smart Array Controller Linksys Network(10/100 BT) card 8 PIII Xeon w/ 2MB cache Processors. egcs-2.91.66 redhat 6.2 distribution I do have a 512 MB swap partition I am including the code that I have been playing with. It looks very ugly, but I have tried just about everything I could without going in and trying to learn how the kernel is written. I am including the source code for the block write device. and the kernel configuration from 2.4.1-pre11. Has anyone else see this problem. I have tried to write a file to on the filesystem. Looks like a similar problem. Nathan <> <> ------_=_NextPart_000_01C08A0D.680CC850 Content-Type: application/octet-stream; name="write.cpp" Content-Transfer-Encoding: quoted-printable Content-Disposition: attachment; filename="write.cpp" #include =0A= #include =0A= #define __USE_LARGEFILE64=0A= #define __USE_LARGE_OFFSETS=0A= #include =0A= #include =0A= #include =0A= #include =0A= #include =0A= =0A= #include =0A= #define lseek lseek64=0A= #define off_t off64_t=0A= #define open open64=0A= #define MAX_SIZE 32*1024 /* Maximum Block Write Size */=0A= =0A= extern "C" off_t lseek(int FD,off_t offset,int whence);=0A= =0A= #include =0A= #include =0A= #include =0A= #include =0A= #include =0A= int WRITE (int FD,char *buffer, int size)=0A= {=0A= int written=3D0,current=3DMAX_SIZE;=0A= int writelen;=0A= for (;written