I am trying to set up a server with an onboard Super Micro AIC 7902B U320 controller under the latest Ubuntu version(s), i.e., using the respective Ubuntu supplied kernels (2.6.20 based for "feisty" and 2.6.22 based for "gutsy"). However, so far all kernels based on 2.6.x, x greater or equal to 20 hang/crash when inserting the aic79xx module. Earlier kernels from various distributions, e.g. 2.6.15 (ubuntu dapper), 2.6.18 (opensuse 10.2), 2.6.18 (debian unstable install kernel) work, but the discs are only running at a fraction of the theoretical performance (see below). I fully understand that it's not the job of the developers to take care of users of the various distributions, but I guess if there are experts to be found who have "been there, done this", then this is the place to ask. I am perfectly happy to try to boot the machine with a standard, self-compiled Linux 2.6.2x kernel, but would appreciate hints as to what version plus which patches (if any) to use. So, if anyone is willing to take a look, here are some facts (again, please let me know whatever other info you need!). I have put the output of lspci, dmesg and cat /proc/scsi/aic79xx/0 of the system running a default Ubuntu dapper installation (2.6.15 based kernel) on http://www.mdy.univie.ac.at/de/people/boresch/privat/dapper.lspci.output http://www.mdy.univie.ac.at/de/people/boresch/privat/dapper.dmesg.output http://www.mdy.univie.ac.at/de/people/boresch/privat/dapper.proc_aic79xx.output respectively. Hardware: * Supermicro H8DAR-8 motherboard with two dual-core Opterons * Two Seagate Cheetahs (ST373207LC) attached to the onboard Super Micro (Adaptec) AIC 7902B U320 controller * There is also a LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 08) controller, to which an external RAID is connected. This part of the system seems to be working flawlessly!! The problem(s): With the kernels 2.6.x (x < 20) I have tried so far the system boots, but there are many warnings during the boot when loading the aic79xx controller and the result to me seems nowhere near the speed which the disks should be capable of. E.g., for the first SCSI disk from /proc/scsi/aic79xx/0 Target 0 Negotiation Settings User: 320.000MB/s transfers (160.000MHz RDSTRM|DT|IU|RTI|QAS, 16bit) Goal: 40.000MB/s transfers (40.000MHz) Curr: 40.000MB/s transfers (40.000MHz) Channel A Target 0 Lun 0 Settings Commands Queued 3039 Commands Active 0 Command Openings 32 Max Tagged Openings 32 Device Queue Frozen Count 0 The User: entries correspond to what is set in the SCSI BIOS (Adaptec, version 4.30), and which, according to the vendor of the box (transtec) is correct. I could live with all this if kernels 2.6.2x I have tried would not simply crash. When the aic79xx module loads, I get again endless >>>>>>>>>>>>>>>>>> Dump Card State Begins <<<<<<<<<<<<<<<<< listings, but the warnings / error messages get more and more severe until the system hangs. At this point I am really unexperienced and have no clue how I could capture these pages of messages. Again, I am happy to dig in but don't quite know where to start. According to Seagate Tools the disks themselves are fine. I can easily dig out all the source from which the various Ubuntu kernels are compiled; in fact, i have stared at them, and my feeling is that this may be less an issue with the low level driver itself (which claims to be version 3.0 in all cases) but with the surrounding SCSI subsystem ... Many thanks in advance, Stefan -- Stefan Boresch Institute for Computational Biological Chemistry University of Vienna, Waehringerstr. 17 A-1090 Vienna, Austria Phone: -43-1-427752715 Fax: -43-1-427752790 - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
