From owner-freebsd-scsi Tue Dec 19 02:53:09 1995 Return-Path: owner-freebsd-scsi Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id CAA07451 for freebsd-scsi-outgoing; Tue, 19 Dec 1995 02:53:09 -0800 (PST) Received: from mail.barrnet.net (mail.barrnet.net [131.119.246.7]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id CAA07445 for ; Tue, 19 Dec 1995 02:53:05 -0800 (PST) Received: from Sysiphos (Sysiphos.MI.Uni-Koeln.DE [134.95.212.10]) by mail.barrnet.net (8.7.1/MAIL-RELAY-LEN) with SMTP id MAA19674 for ; Tue, 12 Dec 1995 12:11:50 -0800 (PST) Received: by Sysiphos id AA26517 (5.67b/IDA-1.5 for scsi@freebsd.org); Tue, 12 Dec 1995 20:55:46 +0100 Message-Id: <199512121955.AA26517@Sysiphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Tue, 12 Dec 1995 20:55:45 +0100 In-Reply-To: Rob Mallory "bootverbose #1" (Dec 10, 11:57) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: Rob Mallory Subject: Re: bootverbose #1 Cc: scsi@freebsd.org Sender: owner-freebsd-scsi@freebsd.org Precedence: bulk On Dec 10, 11:57, Rob Mallory wrote: **** bootverbose #1 } here is the 'broken' mode where the jumper setting tells the } card to advertise itself in the same area as the 825. the 825 } somehow steps harder on the 810 which gets disabled at boot. } the bios/sdms probe only sees the drives on the 825 and boots } scsi id 0 which is the toshiba. } } I'll send you both outputs.. currently, I have to run in this } mode, without the 810 so I can boot the kernel off my toshiba 1GB Ok. Thanks ... } eventualy, i'd like to boot the kernel off my conner4207 on the 810. } (and have it be sd0:) } ncr0 rev 1 int a irq ?? on pci0:6 } mapreg[10] type=1 addr=0000ff00 size=0100. } mapreg[14] type=0 addr=0000ff00 size=0100. The NCR 53c810 got mapped to an absolutely silly address ! This is at the end of the first 64KB, this is in the middle of system RAM ... } pci_map_mem failed: device's memrange 0xff00-0xffff is incompatible with its bridge's memrange 0x4000000-0xffffffff } reg20: virtual=0xf4575f00 physical=0xff00 size=0x100 } CACHE TEST FAILED: reg dstat-sstat2 readback ffffffff. } CACHE INCORRECTLY CONFIGURED. Not surprising at all ... } ncr1 rev 2 int a irq 11 on pci0:11 } mapreg[10] type=1 addr=0000e400 size=0100. } mapreg[14] type=0 addr=f7fff000 size=0100. The 53c825 gets a reasonable address, a little below 2GB. } reg20: virtual=0xf4578000 physical=0xf7fff000 size=0x100 } ncr1: restart (scsi reset). } ncr1 scanning for targets 0..6 and 8..15 (V2 pl23 95/09/07) } pci0: uses 33554688 bytes of memory from f7fff000 upto f9ffffff. } pci0: uses 512 bytes of I/O space from e400 upto ffff. Ok. Everything looks fine, except the 53c810 was ignored since the system BIOS had assigned it an impossible address. **** bootverbose #2 } STefan, } Here is the second bootverbose. it boots a 2.1 kernel off of } the only disk on the 810, and then changes root to sd0, which } is a -current disk. hmmm.. i did not notice the errors in the } non-bootverbose situation. There was a CD in the drive, this may make a difference ... } Oh, the 810 is a 'GA-410'.... Gigabyte making ncr cards now? Yes, starting at the same time as ASUS (early 1994, at least). } ncr0 rev 2 int a irq 11 on pci0:11 } mapreg[10] type=1 addr=0000e400 size=0100. } mapreg[14] type=0 addr=f7fff000 size=0100. The 53c825 is mapped to a reasonable address, as it was before. } reg20: virtual=0xf4372000 physical=0xf7fff000 size=0x100 } ncr0: restart (scsi reset). } ncr0 scanning for targets 0..6 and 8..15 (V2 pl23 95/09/07) } (ncr0:6:0): "PLEXTOR CD-ROM PX-6XCS 1.00" type 5 removable SCSI 2 } cd0(ncr0:6:0): CD-ROM } cd0(ncr0:6:0): asynchronous. } cd present.[264691 x 2048 byte records] } ncr0:6: ERROR (81:1) (8-0-0) (0/13) @ (ffd8f00c:f000ef6f). Will look at this message in detail later ... (Well, I just did: 81: dma fifo empty + illegal instruction (!) 1: SCSI parity error (aha!) (ffd8f00c:f000ef6f) Well, that should be a NCR program counter offset and next instruction. But the offset is wrong, it ought to be a small positive number ... Seems the chip was thrown off its path when a SCSI parity error occured talking to the CDROM. In fact, this is one of the least tested part sof the driver, since we did not have any hardware to test it ... This is a sign of too long a cable for the transfer speed choosen. Since this is caused by a hardware problem, there is not much I can do. You may want to shorten your SCSI cable, check for any problems with your terminators, make sure an active terminator is used at the end of the cable. What kind of connectors and cables are you using (i.e. is the WIDE drive connected to one of the internal connectors, and the rest to the other ? There are some restrictions on combinations of connectors used at a tiem ...) } reg: da 00 00 13 47 00 0c 0f 35 08 82 00 80 00 0f 02. } ncr0: restart (fatal error). } ncr0:15: ERROR (80:7810) (8-0-0) (ef/f0) @ (ffd8f000:00000000). This seems to be a result of the previous error. There is no device with a SCSI ID of 15 in your system ... } reg: ca 00 00 13 47 00 00 0f 71 00 02 00 80 00 0f 02. } ncr0: restart (fatal error). } ncr1 rev 1 int a irq 14 on pci0:12 } mapreg[10] type=1 addr=0000e000 size=0100. } mapreg[14] type=0 addr=f7ffe000 size=0100. The 53c810 is mapped correctly here, too ... } reg20: virtual=0xf4375000 physical=0xf7ffe000 size=0x100 } ncr1: restart (scsi reset). } ncr1 scanning for targets 0..6 (V2 pl23 95/09/07) } ncr1 waiting for scsi devices to settle } (ncr1:0:0): "CONNER CFA540S 13B0" type 0 fixed SCSI 2 } sd4(ncr1:0:0): Direct-Access } sd4(ncr1:0:0): FAST SCSI-2 100ns (10 Mb/sec) offset 8. } 515MB (1056708 512 byte sectors) } sd4(ncr1:0:0): with 2800 cyls, 4 heads, and an average 94 sectors/track And it can be used at the address it got assigned. } Dec 10 11:57:02 kickme /kernel: pci0: uses 33554944 bytes of memory from f7ffe000 upto f9ffffff. } Dec 10 11:57:02 kickme /kernel: pci0: uses 512 bytes of I/O space from e000 upto e4ff. Well, if you get your BIOS to assign a reasonable address to the 53c810 card, then you are set. You may want to use the line options "NCR_IOMAPPED" in your kernel config file. This will make the driver use port I/O instead of memory mapped register accesses. Let me know, if this makes the driver work on your system. But you really should try to get a BIOS that does not assign\ DRAM addresses to PCI chips ... Rergards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/~se